Peyton Rous
Nobel Lecture
Nobel Lecture, December 13, 1966
The Challenge to Man of the Neoplastic Cell
Tumors destroy man in a unique and appalling way, as flesh of his own flesh which has somehow been rendered proliferative, rampant, predatory and ungovernable. They are the most concrete and formidable of human maladies, yet despite more than 70 years of experimental study they remain the least understood. This is the more remarkable because they can be evoked at will for scrutiny by any one of a myriad chemical and physical means which are left behind as the tumors grow. These had acted merely as initiation. Few situations are more exasperating to the inquirer than to watch a tiny nodule form on a rabbit’s skin at a spot from which the chemical agent inducing it has long since been gone, and to follow the nodule as it grows, and only too often becomes a destructive epidermal cancer. What can be the why for these happenings?
Every tumor is made up of cells that have been so singularly changed as no longer to obey the fundamental law whereby the cellular constituents of an organism exist in harmony and act together to maintain it. Instead the changed cells multiply at its expense and inflict damage which can be mortal. We term the lawless cells neoplastic because they form new tissue, and the growth itself a neoplasm; but on looking into medical dictionaries, hoping for more information, we are told, in effect, that neoplastic means “of or pertaining to a neoplasm”, and turning to neoplasm learn that it is “a growth which consists of neoplastic cells”. Ignorance could scarcely be more stark.
The chemical and physical initiators are ordinarily called carcinogens; but this is a misleading term because they not only induce the malignant epithelial growths known as carcinomas but other neoplasms of widely various kinds. In the present paper the less used term oncogens will be employed, meaning thereby capable of producing a tumor. It hews precisely to the fact.
Some may exclaim on reading what comes next that it consists mostly of truisms. This does not make these the less vital to my theme.
Tumors occur in vertebrates of so many kinds that it would not be surprising if neoplastic changes took place in them all. Normal cells of any sort capable of multiplying in response to ordinary stimuli are liable to become neoplastic if acted upon by an initiator. It follows that the growths they form are almost incredibly multifarious, and this not only because of their widely various, cellular sources but because tumors of several kind may derive from a single one. Nevertheless the changed cells are all alike in the basic respect that they disobey the law of organism. Obviously what has to be understood is not the tumor but the neoplastic state of its cells. These are all animated by some principle exploiting their capabilities. In this respect the problem they present is a coherent entity.
The range in effectiveness of neoplastic changes in tumors is exceedingly great. Often the obvious alterations undergone by tumor cells are so slight that they look scarcely different from the normal and function in a similar way, retaining the task of storing fat for example. Yet even such cells, when forming tumors possess an abnormal power to multiply, however slowly they divide. They are euphemistically termed “benign”. The cells of some not only retain the normal ability to form hormones needed by the body but may produce these in such quantity as to disturb it greatly; and the hormones themselves are sometimes changed to a pathological extent.
In proportion as neoplastic cells diverge from the normal they ordinarily function less well. As time goes on their ability to multiply usually increases, and often they undergo such alterations in their state as to render them lawless and a threat to life. For the tumors they produce we again use a humanistic term, “malignant”. The changes for the worse undergone by their cells are not gradual but are the result of discontinuous steps according to their magnitude, sometimes several of them following one another at intervals, and they may take place in differing directions with result in heterogeneous growths. With each change in this decensus averni – “progression” as I’ve ventured to term it – the cells leave behind more of the specialized features which have distinguished them when normal, until at last they may have wholly lost all of their normal aspect, becoming so completely “anaplastic” that one cannot tell their source. As the consecutive changes go on they tend to lose their adherence to one another, and not infrequently they become a disorderly mob, penetrating the adjacent normal tissue in groups, or individually, and destroying it. They frequently enter blood vessels or lymph channels and, coming loose within, are borne along on their fluid, to lodge in distant organs and there form secondary tumors of the same sort, “metastases” Sometimes it’s every cell for itself! Not infrequently a primary neoplastic change resulting in malignancy makes the affected cells so fatally ill that they would form no tumors did not their rate of division exceed that of their death. Here one is inclined to ask whether extreme neoplastic changes may not sometimes kill cells at once.
Many of the chemical and physical initiators already mentioned must be brought to bear for quite a time before their oncogenic effect becomes perceptible. Often those of different kinds, when applied successively, have cumulative action. Since neoplastic change manifests itself by tumor formation as a culminating event the assumption is generally made that it occurs abruptly; but facts speak decisively against this view. Epidermal carcinomas of the penis almost never occur if circumcision has been done at birth, whereas if delayed until adolescence, in persons of the same race, penile cancers not infrequently arise when the man grows old, for no reason that is then perceptible. Evidently in these instances some initiating oncogen must have acted upon the epidermal cells prior to circumcision, though not sufficiently to bring about neoplastic changes at that time, but only what can be termed a preneoplastic condition that has persisted in the cells’ descendants and at last has come to completion, or been brought to it by some imperceptible, new initiating actor then. Similar long-time culminating events have been observed in my laboratory during study of rabbit skin exposed many months previously to an initiating oncogen without any growth occurring in the interval.
Tests have shown that single living cells, taken from notably vigorous tumors and transferred to other inbred animals of the same stock, will give rise to growths of the same sort; yet the traits of neoplastic cells cannot easily be discerned by studying them singly. When many are proliferating together in tumors though, much can be learnt about them.
An exceedingly important trait of most neoplastic cells is their unnatural excitability which sometimes renders them extremely active on what seems slight encouragement. Infection with inflammatory bacteria often has this effect. Indeed merely the healing of a hole a centimeter across punched through a rabbit’s ear that had been swabbed on its smooth inner surface with an oncogenic tar some weeks before may cause several tumors to start forth from the epidermal sheet that is extending in to close the hole, although elsewhere has arisen none. Some of the growths thus elicited actually behave as if malignant, their cells invading and replacing those of the new connective tissue underlying them. But this activity only lasts as long as they are exposed to reparative stimulation. When this ceases, on closure of the hole, the growths disappear leaving an epidermal layer which appears normal microscopically but is not, as shown by recurrence of the spurious malignancy when a new hole is punched inside the boundaries of the old. The hidden cells rendered neoplastic by the tar had not been altered one whit, merely exploited by the stimulus of healing. Many initiating chemicals possess the power to act in this way on the cells that they have caused to undergo neoplastic change but they do so only during the period of their application. Promotion, as Dr. Friedewald and I termed it, can have a great deal to do with the behavior of tumors.
Neoplastic cells never get well if they are self-assertive, meaning thereby capable of forming a tumor without any extraneous help such as promotion. The great majority are of this kind, as is evident from their success of transplantation. Those that have got worse by steps never retrace these. Were this not the case, one would find at autopsy now and then residual nodules composed of cells to all appearance normal marking the place where once there was a tumor. Neoplasms occasionally get smaller for one intercurrent reason or another, and often they are forced to disappear when subjected to strong irradiation. Yet in both instances the cells languish and die as the same neoplastic elements that they previously were. They never revert to the normal and are potentially immortal by way of their progeny if transferred to compatible new hosts before they have slain the old. Not a few rat and mouse tumors have in this way been maintained for study throughout more than fifty years.
A striking trait of actively multiplying, neoplastic cells is their ability to evoke from the adjacent tissue the blood vessels and structural support needed for the production and maintenance of the growths they are capable of forming. In proportion to the rate at which these latter enlarge, their demands of this sort become peremptory, unless indeed their cells have become capable of actively invading and replacing the normal tissue next them and hence are capable of “living off the country”. One can perceive how crucial for most growths help of his kind is by scattering amidst the subcutaneous tissue of a mouse tiny fragments of a murine mammary carcinoma containing several mingled neoplastic components of widely differing capabilities. This can be done by rapidly injecting a suspension of the fragments in salt solution under the dorsal skin of another mouse, together with sufficient air to split the underlying connective tissue horizontally, thus “plating out” the tiny bits of tissue on its broad expanse, much as bacteria are purposely scattered for colony formation on an expanse of nutritive agar. Many of the fragments thus implanted soon give rise to tumors, because their cells call forth blood-vessels and stroma swiftly from the tissue on which they lie, whereas some others do no more than survive because their cells are devoid of such evocative powers. It is as if the normal tissue were acting as a school for iniquity and rapidly promoting its worst scholars.
However slowly neoplastic cells divide, they must of course obtain additional food if they are to multiply; and however “benign” the growing tumors, their demands can be peremptory. During old times in China, when surgery was seldom done, benign tumors sometimes reached a prodigious size whereas the body serving them emaciated because of their first claim upon its food. One can witness the same course of events after implanting a small piece of benign mouse tumor at a subcutaneous spot in another mouse where it can grow big without damaging the adjacent tissue. The resulting growth gets huge, while the body wastes.
With the sole exception of the tumor cells initiated and actuated by hormones and viruses – which have yet to be discussed – neoplastic cells neither give off any tell-tale substance indicative of their presence nor elicit any specific reaction from the body, whereby it can be certainly discerned; nor do they form any injurious substance characteristic of themselves even while flourishing in great number or undergoing necrosis. True, the body often reacts against them while they are small, as it does against grafts of incompatible normal tissue, destroying these; yet so rarely do established tumors disappear that such happenings seem miraculous. The generality of them, while growing, do away so completely with the initial resistance the body offers to strange cells that even grafts from animals of alien species may succeed. Pig skin has been known to flourish when transplanted to the skin of volunteers having inoperable lymphoid cancers.
Confronted with the dire challenge offered by neoplastic cells the physician does what he can. Now he is fortunately sometimes aided, and to a life-saving extent, through the discoveries of Dr. Huggins, experimental oncologist and practicing surgeon, who has shown that the existence of not a few cancers in the human body, notably those of the prostatic gland, are initiated, actuated and promoted by hormones formed within the patient’s body, and that many of them can be made to disappear by bringing other hormones or chemically related agents to bear upon them. Dr. Huggins will tell in his lecture about these actual “cures of cancer”, unique achievements in this present day.
Here one fact deserves strong stress, namely the tacit assumption of the medical consultant, based on the experience of his innumerable predecessors that the presence of a growth does not ordinarily imply any significant liability in the same individual to others of differing sorts, unless indeed his patient has tissues ultra-sensitive to initiating agents through inheritance, as in the case of those unfortunates who are albinos or who are subject to intestinal popyposis, both of which render the affected tissues abnormally susceptible to oncogenic changes. The physician’s attitude in this relation signifies how dependent most neoplastic changes are on intercurrent, episodic initiation.
What can be the nature of the generality of neoplastic changes, the reason for their persistence, their irreversibility, and for the discontinuous, steplike alterations that they frequently undergo? A favorite explanation has been that oncogens cause alterations in the genes of the cells of the body, somatic mutations as these are termed. But numerous facts, when taken together, decisively exclude this supposition1.
The number of viruses now realized to cause disease has become great during the last half century, but relatively few have any connection with the production of neoplasms. Yet it should be said at once that what they are now known to do has surpassed all imagination.
Two Danes, Ellermann and Bang, reported the first tumor virus in 19082. It caused leukemia in chickens and they made six successive passages of it from fowl to fowl, producing the same disease each time. They studied it until 1923, meanwhile reporting upon a second virus causing a chicken leukemia of a different sort. Yet though their work was convincing its findings were written off because the leukemias were not then realized to be neoplastic diseases nor indeed until after 1930.
In 1910 I described a malignant chicken sarcoma which could be propagated by transplanting its cells, these multiplying in their new hosts and forming new tumors of the same sort. In other ways the growth showed itself to be a neoplasm of a classical sort, yet, as reported in 1911, its cells yielded a causative virus. Numerous workers had already tried by then to get extraneous causes from transplanted mouse and rat tumors but the transferred cells had held their secret close. Hence the findings with the sarcoma were met with down-right disbelief, though soon several other, morphologically different, “spontaneous” chicken tumors were propagated by transplantation and from each a virus was got causing growths of its kind. Not until after some 15 years of disputation amongst oncologists were the findings with chickens deemed valid, and then they were relegated to a category distinct from that of mammals because from them no viruses could be obtained. Only in 1925, through the efforts of a British worker, W.E. Gye, was much attention given them by scientists.
The virus causing the chicken sarcoma first studied, now generally termed the RSV, has been maintained for more than fifty-five years and is still studied in many countries. Throughout most of this time it would engender growths only in chickens and closely related fowls; but of late several extraneous, non-neoplastic viruses have become associated with it, during its passage in unusual avian hosts; and its scope has thus been so enlarged that now not at few mammals, including monkeys, have been found to develop tumors after inoculation with the enhanced material.
After working out the complex relationship existing between the RSV, the cells it affects and their hosts, I tried for several years to get causative viruses from the transplantable tumors of rodents but with the same failures as previous investigators and hence quit the neoplastic problem for others in pathology which proved rewarding. Not until 1933 did a virus opportunity come my way again. Richard Shope of the Rockefeller Institute (a man already renowned for his discovery of animal diseases with human implications) reported then on a virus causing the giant warts often present on the skin of wild “cottontail” rabbits in the southwestern U.S.A. On inoculation this virus proved effective only in rabbits, and it produced far more vigorous warts on animals of domestic breeds than on its native host, the cottontail. When describing the growths of both species Shope remarked that they might be true tumors. He knew of my fruitless search for a mammalian tumor virus, and he and I had long been friends. Hence he asked me to determine the character of the warts, saying that he knew nothing about tumors and already had more than he could do, what with possessing four new viruses responsible for animal diseases of other sorts. Thus it came about that I experimented as his deputy throughout many later years. Now and again he reported on the peculiarities of the papilloma virus as such, but never concerned himself with its relationship to tumors until after the work in my laboratory had ended. He died – and of cancer – less than a year ago. One of his last papers, written while ill, ranks as a classic3. It is concerned with “the many sly and devious ways that viruses may behave in causing tumors”; and he stressed, as example, an extraordinary finding reported by him early, namely that the papilloma virus can rarely be recovered from the prodigiously active growths it yields in domestic rabbits although immunological tests show it to be present and it can regularly be got in quantity from the cottontail growths.
Experimentation carried out in my laboratory together with Beard, Kidd, Friedewald and MacKenzie showed the “warts” produced by the virus to be genuine tumors, benign epidermal papillomas in which the virus persists although eliciting an antibody capable of neutralizing it on direct exposure. The same anomalous state of affairs had previously been found to exist in chickens carrying the first tumor that yielded a virus and experiments had disclosed the fact that the phagocytes of normal blood can protect ingested bacteria from antibodies present in the surrounding medium, as long as they themselves remain alive and will even shield from a markedly hemolytic serum the phagocyted red cells from a foreign species. Later work with Hudack and McMaster showed that rabbit fibroblasts separated from one another with trypsin and placed in a suspension containing vaccinia virus would, on becoming infected with this, protect it from a strong antibody placed in the surrounding fluid, but only so long as they remained alive.
These findings enable one to understand why the Shope papilloma virus flourishes in the proliferating cells of cottontail growths despite the strong antibody that it engenders, and why it never causes the normal cells around it to become neoplastic but instead produces tumors which grow by intrinsic cell multiplication, “caus sich heraus”, to use the German phrase, like the neoplasms of unknown cause.
After some months carcinomas arose from most of the actively proliferating virus papillomas of both cottontail and domestic rabbits, owing to further changes in their cells resembling those that take place now and again in benign papillomas initiated in rabbits by chemical agents and actuated in some way unknown. Yet they are unlike in certain distinctive, cytological respects; and when the papilloma virus was injected into the blood stream of domestic rabbits carrying papillomas due to chemical initiation it localized in these and urged them on, rendering them much more vigorous, and altering the cells of some in such wise that they became mongrel growths exhibiting merged features referable to both agents, and making others carcinomatous forthwith. Furthermore when fragments of the epidermal carcinomas, arising from papillomas induced in cottontail rabbits by oncogenic hydrocarbons, were exposed to the Shope virus in vitro and reimplanted in the animals from which they had been procured, their cells, on proliferating anew, exhibited the mongrel aspect indicative of viral influence, and their malignancy was also greatly enhanced.
Later tests showed that after the brief application of a hydrocarbon known to be a powerful oncogen, to the lightly scarified skin of domestic rabbits, into which the papilloma virus had just been rubbed, the “warts” that arose during the next few days looked and behaved like those induced as controls elsewhere on the scarified skin; and yet they underwent carcinomatous changes much earlier than did these latter, although the chemical previously applied was known to be devoid of “promotional” effect. The inference seemed justified that it had acted on the same cellular mechanism as the papilloma virus.
Some of the carcinomas arising from the papillomas of domestic rabbits were serially transplanted, and one that soon became anaplastic and exceedingly malignant has now been maintained for 28 years by transfer from rabbit to rabbit. Never has it yielded a virus of any sort, and tests of the blood of its early successive hosts showed that the antigen inducing an antibody against the papilloma virus was gradually disappearing. Now it has been gone for many years. This cancer, known as the V x 2, is studied in many laboratories today because, like the generality of those due to other causes, it yields no sign of what actuates it. Obviously the Shope virus had been merely an initiator when producing the V x 2, though some antigenic remnant of it had persisted for a while. In this relation a remarkable recent discovery by the Kleins of the Royal Caroline Institute of Sweden deserves mention, namely that certain polycyclic hydrocarbons initiating mouse tumors put a specific, antigenic mark on these, which persists despite repeated transplantation of the growths.
Nature sometimes seems possessed of a sardonic humor. In the 1930’s a group of American geneticists, long concerned with the incidence of “spontaneous” tumors in rodents, undertook to determine how a pronounced liability of female mice of certain strains to develop mammary carcinomas as they grow old is inherited; and by breeding tests they came upon the astounding fact that this liability reaches the young from their mothers. How can this be? Was it passed on through the ovum, the placenta, or what? A member of their group made a revealing test soon afterwards. He arranged for some of the young of the strain under study to be suckled from birth by mice of a breed in which the cancers do not occur; and the mice thus wet-nursed never had any! His later work showed that the liability to these cancers had been due to a virus passed on to the young in their very first milk. This virus lay latent within them until they matured, and then it caused orderly, benign tumors in the breasts of the females; and from these growths carcinomas derived by steplike changes like those already described. To this day no virus directly responsible for the cancers has been obtained from these, but only the “milk virus” producing the benign growths. The problem of the actuating cause of the cancers remains unsolved like that of the V x 2 carcinoma of rabbits4. In both instances the virus causing the benign tumor was no more than an initiator of the malignant growth deriving from them.
In 1953 Gross5 discovered a neoplastic virus that has greatly widened knowledge. He obtained it from a carcinoma arising spontaneously in the parotid gland of a mouse and on inoculation into other mice it proved capable of producing tumors of more than 20 kinds, and some as well in rats, rabbits, guinea pigs, hamsters and ferrets. Because of these widely various neoplastic effects it has been aptly termed the polyoma virus. The growths it induces can be maintained indefinitely by transplantation, and nearly always the virus disappears from them as time goes on, yet the activity of the tumor does not lessen. Obviously in such instances its role is no more than that of an initiator, comparable in this respect to the chemical and physical oncogens. Very occasionally though, it is an actuating agent as well, persisting and multiplying in a parotid cancer like the one originally providing it, and from this favorable growth it can be recovered anew and started again on its polyomatous career. Under natural conditions the virus maintains itself as an infectious agent widely prevalent in mice but causing only a trivial, scarcely perceptible illness of non-neoplastic sort save in those rare instances in which it produces a tumor.
No virus has as yet been found that indubitably actuates neoplasms in man. Yet this is not to say that viruses play no part in initiating them occasionally. Now and again a human cancer arises where a virus has persistently wreaked inflammatory injury. The virus causing “fever blisters” often next the mouth of persons who are notably susceptible to its action provides an instance in point, cancer sometimes arising after a while from the epidermal tissue long kept in a disturbed state. One sees the same course of events occasionally on skin where the severe virus causing herpes zoster (“shingles”) has left tissue permanently damaged. Yet these instances tell no more, as concerns causation, than do the cancers that now and then arise on the skin of old people where it was burned in youth. Sunlight provides yet other examples, cancer arising from skin that it has kept inflamed instead of tanned. In all these instances the tumors have been merely initiated. No virus actuating any of them has been discovered.
Of late many gifted investigators have sought new neogiastic viruses, with the aid of technics recently devised, and not a few of these agents have been disclosed in rodents. Certain adeno-viruses of man fail to cause any tumors in their human hosts yet do this in small mammals to which they are foreign. Some viruses of both chickens and mammals, which cause no tumors when acting alone, do it when they are inoculated together. The possibilities in both these directions are boundless, opening vistas of complexity. Yet they will not be dealt with here because a far greater problem presses, that of man’s mortality from cancer today. About one person in six is now slain by neoplastic cells. Can nothing be done to combat these?
During this present century we have learnt enough to take up the task. First there was the observation that the tumors of mammals closely resemble our own. Soon they were maintained by transplantation for study. Next came the realization that what one discovers by experimenting with the tumors of rodents throws light on what happens in man. And then in 1918 came the epoch-making discovery by Yamagiwa and Ichikawa that tarring rabbit skin will cause tumors to arise. This opened an era of rewarding search for other chemical agents – and physical as well – which do the same. It is an era so far from done as to have been the main theme a few months ago of the International Cancer Congress held in Japan; and it was chosen with good reason since some of man’s habits, many of the occupations through which he earns his living, and the mischances taking place in his own body can prove fatal through the growths they induce unless something is done to ward them off. All of these oncogens are initiating in character, and some can be dangerous. promoters too, if the exposure to them is long. The wider and closer one looks the more clearly does one see that similar agents start off nearly all human tumors, and that these are occupational diseases resulting from the exceedingly hazardous occupation of living out a life in this world. Despite all preventive steps – and many are taken in civilized lands – deaths from cancer have scarcely been lessened.
What can be the character of the alteration which renders a cell and its progeny permanently neoplastic, and so self-assertive in its behavior as to kill? On these matters we have as yet only the meager information provided by the neoplastic viruses of animals. Innumerable attempts have been made to obtain evidence of the action of such agents in human tumors and recently to discern them with the electron microscope. Yet such attempts have consistently drawn blanks save in a few highly dubious instances.
In the enormity of our ignorance we have resorted to the tactic of “try-and-see” on a large scale, attempting to destroy the tumors of animals for what any success with these may imply for those of man. In all of the larger civilized countries efforts with this aim are made and they are amply supported with funds, in some instances governmental. Already in the U.S.A. more than half a million chemicals, and materials of biological origin have been tested. Several institutions exist for this special purpose and there is much cooperation by large manufacturing concerns. The most encouraging results thus far have been those of Dr. Huggins, first obtained by studying “on his own” the prostate gland of dogs. As a chemotherapeutic enterprise to save patients with prostatic cancer his discoveries have proved prodigiously effective. The leukemias of children have also been overcome by chemotherapy in some instances and the singular, highly malignant lymphomas in African children known as Burkitt’s disease. Yet the successes thus far obtained have been episodic, and no inkling has been found, save in the case of the small group of viruses, of what happens in a cell when it becomes neoplastic, and how this state of affairs is passed on when it multiplies.
1. P. Rous, Surmise and fact on the nature of cancer, Nature, 183 (1959) 1357.
2. C. Ellerman and O. Bang, Zentr. Bakteriol., 4 (1908) 595.
3. Richard Shope, Evolutionary episodes in the concept of viral oncogenesis, Perspectives Biol. Med., 9 (1966) 258.
4. Despite protracted search, aided by the electron microscope, no tumor virus has ever been found in human milk, and family histories definitely rule it out.
Nobel Prizes and laureates
See them all presented here.