(CANCER RESEARCH 51, 757-763, February 1, 1991]
Perspectives in Cancer Research
Tumor Transplantation and Tumor Immunity: A Personal View
O. C. A. Scott1
Richard Dimbleby Department of Cancer Research, UMDS, St. Thomas' Hospital, London, SEI 7EH, United Kingdom
These subjects have been discussed in meetings large and
small, in workshops, committees, and over coffee cups. They
have been subjected to numerous overviews and even one underview (1). Could there possibly be any more to be said on the
subject?
What follows could be described as a "side-view," written by
an observer who has followed the scene from a safe distance for
nearly 40 years, and this article is certainly not a review. An
expert wrote, "It is quite impossible to present within reasona
ble space a comprehensive review of all investigation in tumor
immunity." These words were written by E. E. Tyzzer threequarters of a century ago (2).
It has often been remarked that the workers in these fields
have suffered from manic-depressive waves of optimism and
pessimism (3), and an attempt will be made to trace the causes
of this instability. Why have scientists held certain views with
such violence, only to see their position overturned and, if they
lived long enough, overturned again?
A general explanation has been given recently for the course
of events. In a chapter headed "Animal experiments—a failed
technology," Robert Sharpe (4) suggests that animal cancer
resistance of the individual to the growth of engrafted tumor
simply one expression of a resistance to the growth of engrafted
tissues in general?"
It is fascinating to examine the struggles of E. F. Bashford,
the Director of the (Imperial) Cancer Research Fund, described
in his first report published in 1904 (8). Bashford was a re
markable man, gifted with great energy, and his contribution is
sympathetically described by Joan Austoker (9). He confirmed
the earlier work of Leo Loeb, which showed that tumors could
not be transplanted between species. But not for a moment did
Bashford think that this implied immunity to cancer. He wrote,
"The consequences to the transferred tissue are exactly analo
gous to the fate which befalls the blood-corpuscles of one animal
when injected into another of a different species. The trans
ferred cells are treated as albumins strange to the species." He
got the point! Rejection was not due to the cancerous nature of
the graft but to its strangeness. Why did he not give more
weight to the possibility of strangeness within a species? Landsteiner's work on human blood groups had already been pub
tests have proved confusing in developing immunotherapy. But
most scientists would agree that the mouse has proved to be a
superb tool for some transplantation studies. In the preface to
his book, Biology of the Mouse Histocompatibility-2 Complex,
Jan Klein (5) writes, "In recent years, the H-2 system has
become not only the prototype of similar systems in other
species, but also one of the focusing points of modern biology."
In fact, the mouse has been sending us the right signals for
nearly a century, but these signals have not always been cor
rectly interpreted. It is fatally easy for the opponents of animal
experiments to point gleefully to the errors of the past, but are
they aware that scientific investigation is one of the most
difficult activities open to us? The scientist at the frontier may
have to formulate a thought which has never previously entered
the human mind. In the history of transplantation, time and
again we can see with hindsight how the thought processes of
a particular worker have been "blocked," often for want of a
piece of information already available in the literature.
The fact that noninbred animals could frequently be made
immune to transplantable tumors has been known since the
turn of the century. The fundamental question about this im
munity was formulated very clearly by Gorer in 1938 (6). "It is
important to decide whether the stimulus which causes the graft
to be destroyed arises from the presence of iso-antigenic differ
ences or differences between normal and malignant tissues." In
other words, is the immunity directed against any foreign tissue,
cancerous or normal, or against cancer as such? We may be
tempted to say that surely this point must have been obvious
from the very beginning of cancer research. Well, to a man of
genius, Peyton Rous, it was. He wrote in 1910 (7), "Is the
Received 10/10/90; accepted 11/12/90.
1To whom requests for reprints should be addressed.
lished (10). Bashford was presented with all the clues, but he
failed to fit the pieces together and wandered off into a blind
alley. He attributed the variable results observed in transplan
tation studies to variability in the tumors and not to variability
in the noninbred mice with which he worked. Could Bashford
have obtained inbred mice? The breeding of the first artificially
inbred mouse, the dba, commenced in 1909 and the strain was
not "fully" inbred until may years later. [Just how many gen
erations were necessary for inbreeding is still a matter of debate
(11-13).] But by chance, a line of mice was available which
must have been nearly perfectly inbred, the Japanese waltzing
mouse. M. F. W. Festing (14) writes of these mice: "Waltzing
Mice had been bred in the Far East for many centuries, where
it was known that the waltzing characteristic was recessive;
hence the mice were probably highly inbred."
To Leo Loeb goes the credit for realizing the value of the
Japanese mice and their tumors. He found that transplantation
of a Japanese tumor was virtually 100% successful in Japanese
mice and always failed in "white" mice. In view of the impor
tance of this work, it is unfortunate that some authorities give
a 1901 reference for the tumor of the Japanese mouse. In fact,
Loeb tells us with graphic detail that the first tumor transplan
tation took place on March 12, 1904 (15).
Loeb had already been thinking in genetic terms and had
tested a tumor transplant in F, hybrids between rat strains
giving high and low percentages of takes. He only used 2 rats
but obtained a result we would regard today as probably correct.
But further advance for Loeb was blocked by a most unfortunate
conceptual error. He became fixated on the idea of "pedigrees."
It was obvious to him that siblings must always be more closely
related than second cousins. Now this is not a foolish idea; on
the contrary, the man in the street would instinctively agree
with it. Loeb failed to realize that second cousins of his own
inbred Japanese mice were much nearer to genetic identity than
757
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1991 American Association for Cancer Research.
TUMOR TRANSPLANTATION
siblings of his noninbred white mice. It is rather sad to see that
20 years later, C. C. Little (16) found it necessary to devote a
whole article to the destruction of Loeb's intellectual stance.
Nevertheless, Little's article is still worth reading today as a
superb example of controlled invective.
The case of E. E. Tyzzer ( 17) is perhaps a greater disappoint
ment, because he so nearly solved the problem of tissue trans
plantation. Unlike Leo Loeb, he had grasped the principles
enunciated by Gregor Mendel and set out to test a mendelian
hypothesis of tumor transplantation. His research tool was the
Japanese waltzing mouse, although he used a different tumor
from that of Loeb. From the fact that 100% takes were obtained
in the F, hybrids between Japanese and "common" mice, he
Laureate George D. Snell (23) is instructive. He was looking
for a genetic project involving mice, and he tells us: "The search
involved both extensive reading and several years of trial and
error in the laboratory." How many post-docs on a 3-year grant
could even consider such a plan of action?
P. B. Medawar drew attention to another inhibitor of prog
ress in science: the fact that there may be very little intercom
munication between different fields, if the two sets of workers
read different journals and attend different conferences. He
wrote in his Harvey Lecture (24), "The study of the grafting of
normal tissues had its own independent origins and history and
for very many years there was little exchange of thought between
those who worked with tumors and those whose interest was
confined to the surgical uses of transplantation."
The truth of Medawar's dictum can be confirmed from a
concluded, quite correctly, that susceptibility is a dominant
character, but he was a careful and conscientious scientist and
he tested the F2 generation, expecting to see 75% tumor takes.
He obtained none in 54 mice. He wrote of this result (p. 557),
"I am unable to accord with Mendel's Law or any other prin
ciple of heredity yet discovered." It might be instructive today
to give Tyzzer's 1909 paper to a group of biology students and
ask them to explain the F2 result. Mendel himself would have
had no problem! It follows from his Law of Independent
Assortment; if 2 factors are involved, 9 of 16 or (0.75)2 of the
F2 generation will contain at least one alÃ-eleof each of two
dominant factors. If there are n factors, we would expect (0.75)".
If n is large, (0.75)" can be very small and Tyzzer's negative
result is explained. Why did Tyzzer not break through this
minor conceptual block? Mendel's work was published in Ger
man but an English translation was available in 1901, and by
1907 the second edition of Punnett's summary of Mendelian
genetics had appeared (18).
Clearly, Tyzzer's mind was fixated on the idea that a single
observable effect (take or no-take) must be controlled by a single
gene. This "block" was swept aside when Tyzzer began to
cooperate with C. C. Little, who clearly had a wider knowledge
of genetics. Together they examined a much larger number of
F2 mice and obtained 3 tumor takes in 183 mice (19) and
suggested that 12 to 14 genetic factors were involved.
What a pity that Tyzzer did not make the breakthrough in
1909. His paper contains several flashes of near genius. For
instance, he wrote, "It is possible to produce by experimental
breeding
a tumor
waltzing
tuations
AND TUMOR IMMUNITY
study of the bibliography of his own classic paper on skin
grafting (25). There was no reference to the work of C. C.
Little, J. J. Bittner, etc. Medawar made generous amends for
this omission in his Harvey Lecture (24), awarding the palm of
priority firmly to C. C. Little, although as we have seen, E. E.
Tyzzer had grasped the main outlines in 1909 (17). In fact, the
artificial division between work on normal tissues and work on
tumors was not a feature of the early days. As we have seen,
Rous bridged this gap in 1910 (7) and in 1921-1922, C. C.
Little and B. W. Johnson (26) transplanted spleens from Jap
anese mice into F, hybrids and confirmed the genetic theory of
transplantation, which had initially been worked out with tu
mors.
In her splendid chapter on the foundations of modern exper
imental cancer research (9), Joan Austoker states that Woglom
demolished the early misconceptions on the nature of tumor
transplantation in his 1929 review (22), but this is only a partial
truth. Certainly, Woglom stated in his incisive style that what
ever the explanation of immunity to transplanted tumors, the
original host could never be immunized, and from this fact he
concluded that immunotherapy was a nonstarter. But he failed
to appreciate the importance of the work of Tyzzer and Little
(17, 19), although he summarized it correctly, and his final
conclusion was as follows: "Immunity to transplantable tu
mours is a generalized refractory condition, which appears to
be unrelated to other forms of immunity." When we remember
waltzing mice which are in every case insusceptible to
that gives very nearly 100% of positive results in other
mice. The alternation of such stocks would give fluc
in infectivity from 100% to 0%." In other words, he
that nearly all the work reviewed by Woglom was carried out
with noninbred mice and nonspecific tumors, i.e., allogeneic
systems, this makes sad reading, particularly when we remem
ber Bashford's (8) words about "strangeness" written a quarter
was proposing the experiment successfully performed by Bittner
2 decades later (20).
After reading the paper of Little and Tyzzer (19), we might
be tempted to suggest a dichotomy between the clever geneti
cists who got it right and the woolly-minded biologists who
confused the issue. However, this simple view would be quite
false. Another geneticist, Maud Slye (21), totally rejected the
interpretation Little and Tyzzer put on their data. She suggested
that the relative failure to obtain tumor takes in the F2 genera
tion was due to the general biological inferiority of these hy
brids.
So, when Woglom wrote his massive review on Tumor Im
munity in 1929 (22), with over 600 references, it is not surpris
ing to find that the truth of the genetic hypothesis, so obvious
to us now, was far from clear to him. Woglom was clearly
suffering from information overload. Perhaps the pendulum
has swung too far the other way, and we no longer read widely
enough. The following passage from the reminiscences of Nobel
of a century before. Here we have another example of the
danger of the Intellectual Block.
It is difficult to point to a particular moment in history when
the genetic basis of immunity to allogeneic grafts was finally
accepted and it would be possible to find papers published
recently where the authors have not quite grasped the principles
involved, but certainly a landmark was the appearance of Bittner's classic paper of 1932 (20). He showed that he could
simulate the results of British workers, who had observed such
variable results in transplantation studies, by judicious choice
of tumor and host, i.e., he performed the experiment proposed
by Tyzzer 23 years before (17). Bittner succeeded in his aim
and his work was accepted by the workers at the Imperial
Cancer Research Fund.
The abandonment of years of work performed with noninbred
animals must have been a painful process for those concerned,
but finally it was almost complete. As Herbert Rapp (27) put
it: "When it was generally recognized that so-called tumor
758
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1991 American Association for Cancer Research.
TUMOR TRANSPLANTATION
AND TUMOR IMMUNITY
immunologists were actually establishing the basis for trans
plantation immunology, the field collapsed.. . . Tumor immu
nology in the 1930's and 1940's, therefore, was definitely mor
intellectual pursuits, such as the investigation of genetics in all
its manifestations.
Foley's work was confirmed by Baldwin (32) and Prehn and
ibund, and all investigators with even a shred of sanity had
given up the patient for lost."
Main (33). The Prehn and Main paper caught the imagination
of cancer workers and has been quoted widely. They obtained
immunization with 12 of 14 MCA-induced tumors, but isologous immunity was not produced by any of the 7 spontaneous
sarcomas. The discussion section of this paper is particularly
important. It includes the statement: "The observation that 2
MCA-induced tumors were found that did not possess demon
strable iso-antigens and that various tumors had different an
tigens, suggest that the altered antigenicity was not an intrinsic
part of the carcinogenic process but rather an incidental con
comitant change."
Now, a new dogmatism replaced the old certainties concern
ing tumor immunity. It was simply an artifact, seen only when
genetically impure systems were used. This dogma was held
with tenacity and the demonstration by Ludwig Gross (28) that
an inbred line of mice could be immunized against their "own"
tumor passed almost unnoticed. Rapp (27) claims that Gross
"single-handedly established the basis of modern tumor im
munology." In a way, this is all too true, in that it can be shown
that overoptimistic interpretations of Gross's work have been
perpetuated down to the present time. Gross used MCA2induced tumor. On p. 328 of his paper, he tells us, "In 91
animals the intradermal tumors grew progressively. ... In 21,
however, they persisted temporarily, then gradually regressed."
But wait a bit! Woglom, in his two reviews (22, 29), pointed
out with asperity that tumors which spontaneously regress are
poor models of human cancer. So here we have an extraordinary
situation. The paper which, according to Rapp, lies at the basis
of modern tumor immunology made use of a poor model. What
did Ludwig Gross, an outstanding scientist, think about his
own work? He wrote (p. 331), "It should also be emphasized
that various tumors such as Sarcoma I, induced by methylcholanthrene, and a spontaneous mammary tumor, may differ immunologically, although both tumors arose in animals of the
same inbred line." So, right at the start, the suggestion was
made that the immunogenicity of chemically induced tumors
might be a special case. The warning bells should have been
ringing loud and clear.
Ten years later E. J. Foley picked up the story in two excellent
papers (30, 31). He confirmed the antigenicity of MCA-induced
tumors, but he also tested spontaneous mammary tumors and
obtained a negative result. In the second paper (31), he wrote,
"Further study will be required to explain the different behav
iour of the two types of tumors, and the possibility that the
methylcholanthrene-induced
sarcomas represent a mutated tis
sue immunologically distinguishable from that of the strain of
mouse in which they arose, and that a comparable mutation is
not a characteristic of the spontaneous mammary tumors of
C3H-He mice we have studied, should be considered."
Thus, in 1953, the picture was remarkably clear. The chemi
cally induced tumors were indeed immunogenic (although it
was shown later that they are not invariably so) but this im
munogenicity was not a feature of spontaneous tumors. The
chemically induced tumors may spontaneously regress and to
this extent were poor "models" of human cancer.
What went wrong? Why was not the "further study" recom
mended by Foley carried out? Well, the problem was investi
gated, notably by R. W. Baldwin (32) and R. T. Prehn and J.
M. Main (33), but gradually the emphasis changed, and more
and more stress was placed on the positive results obtained with
MCA-induced tumors and later with virus-induced tumors and
less and less on the negative results with spontaneous ones. It
may be that this was due to the baleful influence of the demand
for "relevance" in cancer research. What could be more relevant
than a cancer vaccine? However, the demand for relevance has
again and again led to research of poor quality. It could be
argued that all the great advances have come from disinterested
2The abbreviation used is: MCA, methylcholanthrene.
An incidental concomitant change! Surely a shaky foundation
on which to build the towering structure of modern tumor
immunology. Prehn's discussion clearly implies that chemically
induced tumors might be poor models of spontaneous ones.
Perhaps it was unfortunate that the summary ends with the
words: "These results suggest that immunization against some
types of carcinogenesis, using tumor tissue as antigen, may be
feasible." If a more cautious policy had been adopted, it is not
impossible that by now we might have a vaccine against specific
types of cancer, such as melanoma, or carcinoma of the cervix.
The emotive phrase "poor model" has been used above, and
this may be a dangerous practice. What is a "model"? The
shorter Oxford Dictionary gives us "A representation of a
structure" or, figuratively, "something that accurately repre
sents something else."
Engineers have studied the properties and limitations of
models for many years. But their models are fundamentally
different from those of the biologist. A model bridge and a real
bridge are both man made and therefore largely, if not entirely,
understood. Humans and mice are not man made, and neither
are completely understood. We use inbred mice, but it has even
been suggested that our understanding of the process of inbreed
ing is limited, and the measure of inbreeding, the inbreeding
coefficient, may involve a conceptual confusion (13).
Several general articles on the value of models for the study
of tumor immunology have been published, which emphasize
the difficulty of the subject (3, 34, 35). Baldwin (34) writes,
"The general impression gathered from the many published
studies on experimental tumors is that all, or almost all, tumors
express rejection antigens. This is false, however, and it should
be recognized that the expression of tumor rejection antigens
is not an obligatory response to neoplastic transformation." If
the work of Foley and Prehn had been read more carefully,
perhaps this illusion would never have gained currency.
Martin et al. (35) suggest that systems associated with artifactual immunity may nevertheless provide a valuable facility
for studying phenomena recognized to be immunological. Sup
pose we had direct evidence from human studies that a partic
ular tumor, say chorionepithelioma, was immunogenic and we
wished to raise the level of the immune response with some
agent, then experiments with a chemically induced tumor might
enable us to optimize the effect in a patient. It might not matter
if the immunogenicity of the animal system was due to an
"incidental concomitant change" whereas the immunogenicity
of the chorionepithelioma is due to genetic differences between
mother and child. But if, in fact, a particular human tumor is
not immunogenic, then the use of an animal system associated
with artifactual immunity could indeed raise false hopes. The
validity of any model can be judged only in relation to the
759
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1991 American Association for Cancer Research.
TUMOR TRANSPLANTATION
conclusions we hope to draw (13). In this essay "poor model"
is shorthand for an animal system which is much more immunogenic than the human tumor it is intended to mimic.
Despite the work of Gross, Foley, Baldwin, and Prehn, the
scientific community was still skeptical about the reality of
tumor-specific antigens and considered that the positive results
might be due to transplantation artifacts and strain impurity.
George Klein and his coworkers (36) designed a highly inge
nious experiment to demonstrate that the original host mouse
could be immunized against its own tumor, and the result was
a classic paper which deserves to be read in its entirety. Of
particular interest is a passage on p. 1571: "Isologous mice,
pre-treated in the same way (i.e. with heavily-irradiated cells),
showed often, but not always, a stronger resistance than did the
autochthonous host. One possibility is that some residual heterozygosis may still exist within the inbred strains used and can
contribute to the slightly higher resistance of the isologous
mice"; i.e. a low level of heterozygosis was not excluded in this
AND TUMOR IMMUNITY
is the use of what he calls Strong Inference. Platt would have
given his whole-hearted approval to the 1960 paper of Klein et
al. (36), because it eliminates the possibility that transplantation
artifacts could explain the immunogenicity of MCA-induced
tumors.
Study of the Kleins' 1960 paper makes it clear that at this
period, some workers adopted a severely critical approach to
the use of transplanted tumors. Perhaps the best critical review
of this period is that of N. Kaliss (39). This paper covers many
significant points, but the most important is his attack on the
use of "venerable" tumors, i.e., any long-transplanted tumors.
By 1961, all careful readers of the literature were aware of the
fact that all nonspecific, allogeneic tumors, even if transplanted
in inbred mice, were associated with some degree of artifactual
immunity. However, a sensitive test, the growth of a small
number of cells in a preimmunized animal ("Test 4" (40)], was
often necessary to detect it. But Kaliss drew attention to a more
subtle danger, which is often not recognized even today. A
tumor may begin life under immaculate conditions, as a spon
taneous growth in a strictly inbred mouse. But with the passage
of the years "these tumors tend to grow in more and more
paper.
Now, let us pause for a moment and consider what conclusion
these workers would have drawn if there had been absolutely
no immunity in the original host and very weak immunity on
transplantation to isologous mice. Would they not have sug
gested that the immunogenicity of MCA-induced tumors was
indeed an artifact, caused by residual heterozygosity?
The Kleins' 1960 paper (36) had a considerable effect on
world opinion and helped to shift the old paradigm, "Tumor
unrelated strains of mice, and, eventually, the distinction be
tween specific and non-specific becomes blurred." Kaliss's ad
vice was clear. "The valid material in the search for 'cancer
antigens' is the newly arisen tumor." Yet venerable tumors like
the B16, EL4, LSI78Y, LI210, and Lewis lung are used today
for experiments in which even a trace of artifactual immunity
might invalidate the result. Amazingly, the Lewis lung tumor
showed 4% of spontaneous regression when first described (13,
41). Again, the alarm bells should have been ringing.
It can be very difficult to trace the history of a venerable
tumor to its origin. A case in point is the L5178Y. Some
workers quote Iris Parr (42) for the first description of this
tumor, but she gives an incorrect date, and no verifiable refer
ence. Some quote Dunn and Potter (43), but this paper refers
to a different tumor. Glenn Fischer (44) has been referenced,
but he makes it clear (in a footnote) that the tumor did not
originate in his laboratory; he had received it from Lloyd Law.
Schacter and Law (45) tell us, also in a footnote, that no details
about this tumor have been published!
In the discussion following Fischer's paper (44), Law asked
immunity is an artifact, arising from the use of imperfectly
inbred mice," to the new paradigm, "All transplanted tumors
are immunogenic, and a negative result merely suggests that an
insensitive test has been used."
Evidently, the second view prevailed by 1963, when a second
paper appeared (37) entitled "Demonstration of host resistance
against sarcomas induced by implantation of cellophane films
in isologous (syngeneic) recipients." Now, let us look at the
data in this paper. As stated in the summary, "In the ten
available autochthonous hosts, whose tumors had been operalively removed, no resistance could be demonstrated after pre
treatment with irradiated cells followed by challenge with low
numbers of viable cells from the same tumor. After similar
pretreatment isologous mice could be rendered slightly resist
ant." But wait a bit! Is not this result precisely the negative
result considered above? The logic of the 1960 paper would
lead us to the conclusion that the slight degree of isologous
immunity observed (of 29 tumors, 7 showed no resistance, 12
showed 1 log, 8 showed 2 logs, and 2 showed 3 logs) was due
to residual heterozygosity.
It is very difficult to entertain a new thought and the logic of
the Kleins' 1960 paper was highly original. Perhaps it is not
surprising that after 3 years, the authors reverted to the more
traditional logical structure, when they compared a preimmunized with a control group and assumed that if there is a
difference, then the animals must have been immunized against
the tumor. Other work has suggested that tumors caused by
physical agents may be immunogenic. My object is not so much
to question the conclusion as the route by which the conclusion
was reached.
There is a curious assumption which underlies most of the
teaching of science, that students must be instructed in Physics,
Chemistry and Biology but in some mysterious way, logic is
everybody's property. This comforting idea was challenged by
John R. Platt (38). He suggested that the reason why, for
instance, molecular biology has advanced by leaps and bounds
an interesting question about chromosomes in the tumor. In
1979, one strain of L5178Y had a grossly abnormal chromo
some complement (46). In 1969, Anneliese Wolf (47) wrote,
"The LSI78Y murine lymphoma arose in a DBA/2 mouse, but
in view of its long history of serial transplantation, it cannot be
excluded that immunogenetic changes have developed in the
relationship between the tumor and the DBA/2 subline used
for this study." So she considered that this tumor was "vener
able" 21 years ago. Can any tumor be regarded as syngeneic
with its host if its chromosome complement is abnormal? Can
modern technology be used to do chromosome counts on all
transplanted tumors? Revesz showed clearly in 1960 (48) that
"venerable" tumors could be associated with immunogenicity,
which was lacking in spontaneous tumors of recent origin.
The conviction spread through the scientific community in
the 1960s that tumor antigens were not artifacts but might even
be found in all tumors if we looked hard enough. In this
atmosphere, the demand for artifact-free tumor systems was
relaxed, as the following passage shows: "One of us has noted
in a few instances regression following autografting of benzpyrine-induced sarcomas in the rat, but this was extremely rare."
In other words, a relatively trivial procedure could touch off
760
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1991 American Association for Cancer Research.
TUMOR TRANSPLANTATION
spontaneous regression of the tumors, not a situation frequently
to be seen in the cancer clinic. If this work had been done 30
years earlier, the authors would have found themselves impaled
on the spears of Woglom's invective, but this was 1964 (49),
and the climate of opinion had changed. One of the authors
was to display a level of intellectual courage quite unique in the
history of this subject. He actually admitted that he had been
wrong! In "Back to the drawing board—the need for more
realistic model systems for immunotherapy" (50), he wrote,
"To me, the most plausible reason why immunotherapy has not
been successfully translated from the experimental to the clin
ical situation is that the animal models in which they were
based were not realistic."
Could this particular deviation have been avoided? It is fatally
easy to be wise after the event but doubts were raised at the
time, as can be seen from H. B. Hewitt's letter in 1965 (51).
Can any general conclusions be drawn? Why did not more
workers read, and read carefully, the two papers by Foley, which
contain so much of the truth (30, 31)?
The comparative failure of the numerous clinical trials of
cancer immunotherapy paved the way for yet another swing of
opinion. In 1976, Hewitt, Blake, and Walder (52) published
their challenging "Critique of the evidence for active defence
against cancer, based on personal studies of 27 murine tumours
of spontaneous origin." They pointed out that spontaneous
regression was never observed using spontaneous tumors in the
home strain, and they went on to question the current assump
tions of tumor immunologists. Hewitt et al. evidently touched
a raw nerve in the immunological community, as can be seen
from the polemic tone of the late Herbert Rapp's reply (53)
("Humans are outbred" and "inbred mice are artifacts, they do
not exist in nature"). These violet clashes of opinion are not
always helpful. Platt (38) suggests that the method of multiple
hypothesis may be a useful antidote to the infatuation of the
scientist with his favorite idea.
In spite of opposition from some quarters, Hewitt's views on
the immunogenicity of spontaneous tumors and the validity of
animal tumors have been accepted quite widely. For instance,
Middle and Embleton (54) wrote of their studies with sponta
neous tumors in rats, "This work, therefore, supports the view
of Hewitt et al., based on their experience in the mouse." And
on the question of the validity of models, Frankel et al. (55) tell
us, "Human tumors may be more analogous to the less fre
quently antigenic 'spontaneous' animal tumors." On the general
question of tumor-specific antigens, Byers and Baldwin (56)
say, "The current consensus is that such antigens do not exist,
in that all "neoantigens" on these tumors are shared to some
extent with normal tissues"; so we are back with Peyton Rous
AND TUMOR IMMUNITY
spontaneous tumors, even in their own tumor spectrum. The
negative result of Middle and Embleton (54) is far more con
vincing. They subjected each of 28 tumors to the most thorough
tests for immunogenicity available to them and obtained nega
tive results. However, other workers using tumors which cannot
be criticized on the grounds of chemical or virus induction, or
"venerability," have found examples of spontaneous tumors
which are immunogenic in the inbred line in which the tumor
arose (3, 59). To the observer on the sidelines, this is hardly a
satisfactory situation. Is the vital question, whether sponta
neous tumors are immunogenic or not, to be settled by a head
count of scientists favoring the two opposing views?
There is a need for crucial experiments which would satisfy
the requirement of John Platt (38). Perhaps a clue to the way
forward is to be found in the work of Middle and Embleton
(54). As stated above, they obtained 28 negative results in a
row, but previously they had found that a proportion of spon
taneous tumors was clearly immunogenic. The only known
difference between the two groups of experiments was the efflux
of time.
The present author suggested a possible explanation, namely
that the original stock of rats was not fully inbred, and residual
heterozygosity might have explained the earlier results (60), but
this is mere hypothesis. Unfortunately, this unique collection
of tumors is no longer available.
To return to Medawar's point (24) that different streams of
research may fail to fuse, there may be a dangerous split between
biologists and mathematicians. The former are sometimes al
most proud of their innumeracy, while the latter are apt to
despise those who are not skilled in their arcane subject.
Most biologists are agreed that the use of inbred lines is
essential for their work but seldom quote a reference for the
definition of an inbred line. The mathematics of inbreeding was
worked out by J. B. S. Haldane (61) and R. A. Fisher (62), who
concluded that more than 20 generations of inbreeding may be
necessary. If we correct the misprints in Haldane's algebra and
recognize that his y„
corresponds to Fisher's (1 —t„)then we
see that their equations for inbreeding at a single locus are
identical. It is not possible to give a precise answer to the
question "How many generations of inbreeding are necessary?"
since the computation involves both the required level of prob
ability for inbreeding at a single locus, and the number of loci
of importance to the experimentalist.
Somehow the idea spread around that 20 generations is
sufficient (11-13). Bailey (12) shows clearly that if we are
interested in a block of histocompatibility genes consisting of
30 loci, assumed to be unlinked inter se and allogeneic in the
original noninbred population, then the probability of encoun
tering at least one H gene segregating in an inbred strain at
F20 would be 0.55, or a little better than one-half of a collection
of independently derived inbred strains would be carrying at
least one segregating H gene in this generation. This sounds
worrying enough, but in practice the situation might well be far
worse. The mathematical theory of inbreeding assumes no
selection of breeding pairs, but that maestro of the inbred mouse
Leonell Strong tells us (63), "The hardiest pair in each descent
was chosen to continue the line." The mathematical conse
in 1910(7).
Hewitt appears to have won the battle, but has he won the
war? It is important to criticize the critic and it should be
pointed out that the paper of Hewitt et al. (52) is not beyond
reproach. In the title of their paper, they mention 27 tumors,
but of these only 7 were subjected to a rigorous test of immu
nogenicity, i.e., Test 4 (40). They regarded the results of the 7
experiments they performed as negative, but they could have
been observing the Kaliss effect or immunological enhancement
(57). The authors were aware of this possibility but rejected it
on the grounds that enhancement is observable only with al- quences of various kinds of selection were worked out in the
1950's (64) but are not mentioned today. Graff, Valeriote, and
loantigens. However, enhancement has been observed with
Medoff
(65) found "marked histo-incompatibility between and
systems claimed to be syngeneic (58).
within
sublines
of AKR mice used in a syngeneic leukaemic
The paper of Hewitt et al. cannot therefore be regarded as
providing conclusive evidence for lack of immunogenicity of model." Surely this should give rise to some disquiet? Gross
761
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1991 American Association for Cancer Research.
TUMOR TRANSPLANTATION
AND TUMOR IMMUNITY
in Kaliss's sense of the word, and the stability of an inbred line
failure of inbreeding is most likely to be due to contamination
of the line (12).
The title of the paper of Graff et al. (65) contains the word
"syngeneic". This adjective has become a badge of intellectual
respectability in cancer research (66), but what does it actually
mean? It was defined by Gorer, Loutit, and Micklem (67) as an
alternative to Snell's isograft, to cover the case of "grafts
exchanged between genetically identical individuals." But this
leaves us with a problem; the two hosts are defined but not the
graft. According to the strict definition, any tumor transplanted
within an inbred strain would be "syngeneic," but this imme
diately strikes us as absurd. What did the authors of the defi
nition believe? P. Micklem kindly provided the following infor
mation.' "I don't think we ever regarded syngeneic as an appro
of mice cannot be guaranteed.
The extremely interesting work done by T. Boon and his
coworkers (70), using two of Hewitt's tumors, illustrates the
possible dangers. Clearly, these host-tumor systems were free
from the gross artifacts associated with some others, but the
two tumors used had been transplanted 340 and 205 times.
Technical tricks such as freezing the tumors can be of great
value in preventing genetic drift (71). This and many other
technical points were summarized in "Rodent Tumor Models in
Experimental Cancer Therapy" (72), a book which should be
on every experimentalist's desk.
However, it is the belief of the present writer that a technical
fix is only a short term solution to these problems. The long
term answer has been spelled out for us by George D. Snell:
the necessity for years of reading and thought followed by years
of experiment.
priate word to describe the relationship (autochthonous)
between a tumour and its own host. However, the relationship
to other members of the same inbred strain might (but would
not necessarily) prove to be syngeneic. If an individual animal
in an inbred strain was mutant at a histocompatibility locus
then clearly that individual would not be syngeneic with other
members of the strain. Similarly, if a tumour arose which
proved, on transfer to animals syngeneic with its host, to evoke
a rejection response, then it would not be syngeneic with them."
He added, "Unfortunately, the final sentence of para 6 of the
1961 paper [referring to 'syngeneic immunity'] is ambiguous
and in retrospect, would probably have been better omitted."
Compare Micklem's careful and scholarly words with what
happens in practice today. Kaliss's venerable tumors are cheer
REFERENCES
fully described as syngeneic, even when growing in a different
subline from that in which they arose.
What is the take-home message from this highly personal
view of the field? Perhaps the most significant comment comes
from G. D. Snell, with his history of "extensive reading and
several years of trial and error in the laboratory." He says of
his development of congenie resistant lines (23): "I was working
in unexplored territory, inevitably problems were encountered.
It would be five years before I knew whether the effort was
paying off." Can those who organize and fund research accept
the fact that data accumulate slowly in some biological sciences?
Nothing has been said about the exciting new developments
in the field, due to ignorance on the part of the writer. But the
history of past failures should act as a warning to us to keep an
open mind. Some vigorous treatments are claimed to give
patient benefit, and this benefit is judged to have an immunological basis. But if the treatment provokes hyperthermia, or a
fall in blood pressure, or damage to the tumor vascular system,
can such secondary actions be the cause of the effects observed?
Algire et al. (68) provided evidence that a bacterial polysaccharide, a forerunner of tumor necrosis factor, caused a drop in
blood pressure in a mouse, which in turn led to necrosis in a
tumor. Algire's work was questioned by Shear et al. (69), but
the issue is not unimportant.
The pressures of the research grant system tend to force the
research worker to look for a quick technical fix, rather than
undertake the lengthy and tedious investigation of the funda
mentals of a subject. For instance, a reader of Hewitt's papers
might react by saying, "I will throw away my chemicallyinduced tumors and use one developed by Dr. Hewitt." But
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
such a strategy could be even more disastrous than the previous
one. Hewitt's tumors will sooner or later become "venerable"
24.
3 P. Micklem, personal communication.
25.
Hewitt, H. B. Immunotherapy of cancer: an underview. Mod. Med. Canada,
35: 1352-1361, 1980.
Tyzzer, E. E. Tumor Immunity. J. Cancer Res., /: 125-155, 1916.
Kreider, J. W., and Bartlett, G. L. Is there a host response to metastasis?
Prog. Clin. Biol. Res., 2/2:61-75, 1986.
Sharpe, R. Animal experiments—a failed technology. In: Gill Langley (ed.),
Animal Experimentation—the Consensus Changes, pp. 88-117. London:
Macmillan Press, 1989.
Klein, J. Biology of the Mouse Histocompatibility-2 Complex, 620 pp. New
York: Springer-Verlag, I975.
Gorer, P. A. The antigenic basis of tumour transplantation. J. Pathol.
Bacterio!., 47: 231-252, 1938.
Rous, P. An experimental comparison of transplanted tumor and a trans
planted normal tissue capable of growth. J. Exp. Med., 12: 344-365 (Plates
XXVII-XXX), 1910.
Bashford, E. F. The transmissibility of malignant new growths from one
animal to another. Sci. Rep. Cancer Res. Fund (later Imperial Cancer Res.
Fund), London, UK, /: 11-15, 1904.
Austoker, J. A History of the Imperial Cancer Research Fund, 1902-1986,
375 pp. Oxford, England: Oxford University Press, 1988.
Landsteiner, K. On the phenomenon of agglutination in normal human
blood. In: D. J. Bibel (ed.). Milestones in Immunology, pp. 221-224. Madi
son, WI: Science Tech Publishers, 1988.
Bailey, D. W. Sources of subline divergence and their relative importance for
sublines of six major inbred strains of mice. In: H. C. Morse (ed.) Origins of
Inbred Mice, pp. 197-215. New York: Academic Press, 1978.
Bailey, D. W. How pure are inbred strains of mice? Immunol. Today, 3: 210214, 1982.
Scott, O. C. A. The choice of experimental tumour systems. Br. J. Cancer,
41 (Suppl. 4): 112-117, 1980.
Testing, M. F. W. Inbred Strains in Biomedicai Research, 483 pp. London:
Macmillan Press, 1979.
Loeb, L. Further experimental investigations into the growth of tumors.
Development of sarcoma and carcinoma after the inoculation of a carcinomatous tumor of the submaxillary gland in a Japanese mouse. Univ. of Penn.
Med. Bull., 19: 113-116, 1906-7.
Little, C. C. The genetics of tissue transplantation in mammals. J. Cancer
Res., 8: 75-95, 1926.
Tyzzer, E. E. A study of inheritance in mice with reference to their suscep
tibility to transplantable tumors. J. Med. Res., 21 (New Ser. 16): 519-573,
1909.
Punnett, R. C. Mendelism, Ed. 2, 84 pp. Cambridge, United Kingdom:
Macmillan and Bowes, 1907.
Little, C. C., and Tyzzer, E. E. Further experimental studies on the inherit
ance of susceptibility to a transplantable tumor, carcinoma (J.w.A) of the
Japanese waltzing mouse. J. Med. Res., 33 (New Ser. 28): 393-453, 19151916.
Bittner, J. J. Genetic studies on the transplantation of tumors. III. Interpre
tations of apparent rhythms. Am. J. Cancer, 16: 1144-1152, 1932.
Slye, M. The relation of inbreeding to tumor production: studies in the
incidence and inheritability of spontaneous tumors in mice. XIII. Problems
in the behaviour of tumors. J. Cancer Res., 5: 53-79, 1920.
Woglom, W. H. Immunity to transplantable tumours. Cancer Rev., 4: 129214, 1929.
Snell, G. D. Congenie resistant strains of mice. In: H. C. Morse (ed.). Origins
of Inbred Mice, pp. 119-155. New York: Academic Press, 1978.
Medawar, P. B. The immunology of transplantation. Harvey Lectures, 19561957, published 1958, pp. 144-176. New York: Academic Press Inc., 1958.
Medawar, P. B. The behaviour and fate of skin autografts and skin homo-
762
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1991 American Association for Cancer Research.
TUMOR TRANSPLANTATION
grafts in rabbits. J. Anat., 78: 176-199, 1944.
26. Little, C. C., and Johnson, B. W. The influence of susceptibility to implants
of splenic tissue in mice. I. Japanese waltzing mice, albinos, and their I ,
generation hybrids. Proc. Soc. Exp. Biol. Med., 19: 163-167, 1921-1922.
27. Rapp, H. J. Introduction: animal models of immunotherapy of cancer. Nat.
Cancer Inst. Monogr., 39: 1-2, 1973.
28. Gross, L. Intradermal immunization of C3H mice against a sarcoma that
originated in an animal of the same line. Cancer Res., 3: 326-333, 1943.
29. Woglom, W. H. A critique of tumor resistance. J. Cancer Res.. 7: 283-301,
1922.
30. Foley, E. J. Attempts to induce immunity against mammary adenocarcinoma
in inbred mice. Cancer Res., 13: 578-580, 1953.
31. Foley, E. J. Antigenic properties of methylcholanthrene-induced tumours in
mice of the strain of origin. Cancer Res., 13: 835-837, 1953.
32. Baldwin, R. W. Immunity to methylcholanthrene-induced tumours in inbred
rats following atrophy and regression of the implanted tumours. Br. J.
Cancer, 9:652-657, 1955.
33. Prehn, R. T., and Main, J. M. Immunity to methylcholanthrene-induced
sarcomas. J. Nati. Cancer Inst., 18: 769-778, 1957.
34. Baldwin, R. W. Relevant animal models for tumor immunotherapy. Cancer
Immunol. Immunother., 1: 197-198, 1976.
35. Martin, D. S., Stolli, R. L., and Fugmann, R. A. Animal models for tumor
immunotherapy—a commentary. Cancer Immunol. Immunother., 2: 77-79,
1977.
36. Klein, G., Sjögren,H. O., Klein, E., and Hellström, K. E. Demonstration of
resistance against methylcholanthrene-induced sarcomas in the primary au
tochthonous host. Cancer Res., 20: 1561-1572, 1960.
37. Klein, G., Sjögren,H. O., and Klein, E. Demonstration of host resistance
against sarcomas induced by implantation of cellophane films in isologous
(syngeneic) recipients. Cancer Res., 23: 84-92, 1963.
38. Platt, J. R. Strong inference. Science (Washington DC), 146:347-353,1964.
39. Kaliss, N. The transplanted tumor as a research tool in cancer immunology.
Cancer Res., 21: 1203-1208, 1961.
40. Scott, O. C. A. Some observations on the use of transplanted tumors in
radiobiological research. Radiât.Res., 14: 643-652. 1961.
41. Sugiura, K., and Stock, C. C. Studies in a tumor spectrum. III. The effect of
phosphoramides on the growth of a variety of mouse and rat tumors. Cancer
Res., 15: 38-51, 1955.
42. Parr, I. Response of syngeneic murine lymphomata to immunotherapy in
relation to the antigenicity of the tumour. Br. J. Cancer, 26: 174-182, 1972.
43. Dunn, T. B., and Potter, M. A transplantable mast-cell neoplasm in the
mouse. J. Nati. Cancer Inst., 18: 587-595, 1957.
44. Fischer, G. A. Studies of the culture of leukaemic cells in vitro. Ann. NY
Acad. Sci., 76.-673-680, 1958.
45. Shacter, B., and Law, L. W. Azaguanine-deaminase activity of several lymphocytic leukaemias in mice. J. Nati. Cancer Inst., IS: 77-81, 1957.
46. Weinhold, K. J., Miller, D. A., and Wheelock, E. F. The tumor dormant
state. Comparison of L5178Y cells used to establish dormancy with those
that emerge after its termination. J. Exp. Med., 149: 745-757, 1979.
47. Wolf, A. The activity of cell-free tumour fractions in inducing immunity
across a weak histocompatibility barrier. Transplantation (Baltimore). 7:4958, 1969.
48. Revesz, L. Detection of antigenic differences in isologous host-tumor systems
by pretreatment with heavily irradiated tumor cells. Cancer Res., 20: 443451, 1960.
49. Haddow, A., and Alexander, P. An immunological method of increasing the
sensitivity of primary sarcomas to local irradiation with X-rays. Lancet, I:
452-457, 1964.
50. Alexander, P. Back to the drawing board—the need for more realistic model
AND TUMOR IMMUNITY
systems for immunotherapy. Cancer (I'll ¡la.).40: 467-470, 1977.
SI. Hewitt, H. B. Survival curves and autoimmunity against cancer. Br. J.
Radiol., ¿7:714-715, 1964.
52. Hewitt, H. B., Blake, E. R., and Walder, A. S. A critique of the evidence for
active host defence against cancer, based on personal studies of 27 murine
tumours of spontaneous origin. Br. J. Cancer, 33: 241-259, 1976.
S3. Rapp. H. J. Appropriateness of animal models for the immunology of human
cancer. Cancer Res., 39: 4285-4286, 1979.
54. Middle, J. G., and Embleton, M. J. Naturally arising tumors of the WAB/
Not rat strain. II. Immunogenicity of transplanted tumors. J. Nati. Cancer
Inst., 67: 637-643, 1981.
55. Frankel, A. E., Boyer, C. M., and Bast, R. C. Immunobiology of human
breast cancer. In: V. S. Byers and R. W. Baldwin (eds.). Immunology of
Malignant Diseases, pp. 167-224. Lancaster, United Kingdom: MTP Press,
1987.
56. Byers, V. S., and Baldwin. R. W. Preface: In: V. S. Byers and R. W. Baldwin
(eds.). Immunology of Malignant Disease, pp. V-VI. Lancaster, United
Kingdom: MTP Press, 1987.
57. Kaliss, N. Immunological enhancement of tumor homografts in mice: a
review. Cancer Res., IS: 992-1003, 1958.
58. Laursen, M. L., and Laursen, K. Enhancement of tumour growth in two
syngeneic C3H murine systems by immunization via the intracaecal route.
Br. J. Cancer. 37: 1039-1045, 1978.
59. Key, M. E., Brandhorst, J. S., and Hanna, M. G. More on the relevance of
animal tumor models: immunogenicity of transplantable leukaemias of recent
origin in syngeneic strain 2 guinea pigs. J. Biol. Response Modif., .(.-359365, 1984.
60. Scott, O. C. A. Residual heterozygosity in inbred mouse strains. Immunol.
Today, 5:6, 1984.
61. Haldane. J. B. S. Some theoretical results of continued brother-sister mating.
J. Genet., 34: 265-274, 1937.
62. Fisher, R. A. The Theory of Inbreeding, 120 pp. Edinburgh, United Kingdom:
Oliver and Boyd, 1949.
63. Strong, L. Inbred mice in science. In: H. C. Morse (ed.). Origins of Inbred
Strains, pp. 45-67. New York: Academic Press, 1978.
64. Reeve. E. C. R.. and Gower. J. C. Inbreeding with selection and linkage. II.
Sib-mating. Ann. Hum. Genet., 23: 36-49, 1958-1959.
65. Graff. R. J., Valeriote, F.. and Medoff, G. Marked histoincompatibility
between and within sublines of AKR mice used in a syngeneic leukaemia
model. J. Nati. Cancer Inst., 55: 1015-1016, 1975.
66. Scott, O. Syngeneic tumor-host relationship. J. Nati. Cancer Inst., 72: 775,
1984.
67. Gorer, P. A., Loutit, J. F., and Micklem, H. S. Proposed revisions of
"Transplantese." Nature (Lond.), 189: 1024-1025, 1961.
68. Algire, G. H., Legallais, F. Y., and Anderson, B. F. Vascular reactions of
normal and malignant tissues in vivo. V. The role of hypotension in the
action of a bacterial polysaccharide on tumors. J. Nail. Cancer Inst., 12:
1279-1291, 1952.
69. Shear, M. J., Achinstein, B., and Pradhan, S. N. Effect of a bacterial
polysaccharide and several other tumor-necrotizing agents on carotid blood
pressure in mice. J. Nati. Cancer Inst., 21: 585-594, 1958.
70. van Pel, A., Vessière,F., and Boon, T. Protection against two spontaneous
mouse leukaemias conferred by immunogenic variants obtained by mutagenesis. J. Exp. Med., 157: 1992-2001. 1983.
71. Hewitt. H. B., and Blake, E. R. Stability of transplanted murine tumour
systems after storage of cells at —196'Cfor up to 13 years. Br. J. Cancer, 37:
718-722, 1978.
72. Kallman, R. F. (ed.). Rodent Tumor Models in Experimental Cancer Ther
apy, 310 pp. Elmsford, NY: Pergamon Press. 1987.
763
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1991 American Association for Cancer Research.
Tumor Transplantation and Tumor Immunity: A Personal View
O. C. A. Scott
Cancer Res 1991;51:757-763.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/51/3/757.citation
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1991 American Association for Cancer Research.