Renato Dulbecco

The Nature of Cancer

Category: Lectures

Date: 1 July 1981

Duration: 36 min

Quality: HD MD SD

Subtitles: EN

Renato Dulbecco (1981) - The Nature of Cancer

Participating for the first time in a Lindau meeting, Renato Dulbecco began his lecture by explaining two different ways to study cancer: One as a medical doctor looking for causes, symptoms and treatments and the other as a biologist looking into the primary processes in the cell

That early quotation raises many problems, I mean, the main problem is, how can we achieve this type of cooperation? And it’s clear that both sides are to do their job, and the job is, my job as a scientist is to try to understand better the events that lead to cancer in order that I can point out how prevention or therapy can be carried out. And this can be done in two different ways. One would be to be pragmatic and look for things that are agents of cancer or methods for therapy which may be not too general, not too optimal but do some good and use them, and this of course is done every day. Another approach is to be more biological and ask more fundamental questions. What is cancer? What are the mechanisms and the causes, open in hope that the outcome of this type of more biological approach it will be possible to come to indicate consequences that can have a practical application. So by following this second approach, which of course I, as a biologist, prefer, I will now try to show to what kind of consequence this can lead. But before I can do that I must develop a general picture about what cancer is, which is based on the most recent development in various fields of cancer research. I will start asking the question, why is a cancer cell so unique? In fact, because it has properties normal cells do not have. And so I may ask, what are the causes of the uniqueness, what are the differences between the cancer cell and the normal cell? And this question of course is being asked for many years and people have found lots and lots of differences. But these differences are too many, if any, and really they don’t make much sense. Some of them can be explained by saying that the cancer cell is in a persistent state of growth and therefore, because it never shuts off, never reaches that other stage of acquiescence which normal cells always have. But this seems to be only part of the story, there’s a lot more to this. I’ll give you an example, there are many cancers that produce hormones, which are not produced by the normal cells of the same organ. And for instance there are some cancers of the lung that produce the placental hormone, so there is no connection between the two cell types. And if one looks through what is known already, one really comes to the conclusion that there is a very wide alteration over the state of the activity of the genes in the cancer cells compared to normal cells. Now since this seems to be rather general, and seems to be such a striking event, one might think that they might have some kind of profound significance. And maybe one should look in that direction to come to understanding the basic nature of a cancer cell. And I try to do this and I will use results which come from two fields. One is tumour virology, the other is chemical carcinogenesis. I will start with the tumour virology. Of course tomorrow Dr. Temin will talk to you about some of these viruses. And I only will cover whatever is relevant to my argument. There are tumour viruses that produce tumours in animals or can also sometimes cause characteristic alterations of cells in vitro which we call transformation. the transfer of the genetic material of the virus to the DNA of the cell. This occurs in various ways in various steps, but the ultimate result is always the same. Namely that the piece of DNA which is the viral genome will become inserted, integrated into the cellular DNA. This is something that happens constantly and can happen at any part of the DNA of the cell. So that localization at any particular part is very rare generally. Then once the genes of the virus are inserted there, they will multiply together with the genes of the cell, but will have their autonomous expression. Of the viruses that are known to produce tumours in animals there are various kinds, and those I want to mention today are only retroviruses because of the purpose of my talk. The retroviruses, there are two kinds that I want to discuss today. One is sarcoma viruses and the other leukaemia viruses. These viruses, once they become inserted in the cellular DNA, express their own genome. A promoter is where the expression, the transcription of the genes begin, to which are always regularly found at both ends of the genome. And through that there will be a number of messenger RNAs made by variety of mechanisms and finally proteins which will express the function of the various genes of the virus. They produce sarcomas in animals, they produce transformation of cells in vitro. The reason they do this is because they contain a special gene which we call oncogene. Which is not a regular genovirus but it has the ability to transform the cell. This is proven by genetics in a very precise way, I don’t have to go into that. Now one of the questions that is immediately asked is, how does it do it? We know for sure that this is due to the synthesis of a protein which is the product of this particular gene. So now we know from Dr. Watson’s talk yesterday that a cell contains many thousands of different proteins and so why does a single protein added to all these proteins transform a cell? Well the answer is that this protein is very special. It has a special function, it is an enzyme, a protein kinase which phosphorylates other proteins. Of course by phosphorylating them, alters their properties very substantially and presumably their functions as well. Since the evidence says that the function of this enzyme is required for transformation, then we may think that transformation is due to the alteration of a very large number of proteins which are involved in, for instance, growth regulation. This is supported also by the fact that an enzyme with the similar properties is involved in the action of growth factors, the substance that do effect growth regulation. Although many of the detail and the fact that there is only a very sketchy idea of this phenomenon, nevertheless the general conclusion seems clear. The action of this enzyme also can explain probably the widespread alteration of gene expression which occurs in the transformed cells because this very large number of modification of proteins that it causes will alter many cellular macromolecular interactions which are the basis of the regulation. In fact I think we should consider regulation of gene expression in a cell as a kind of a network of interaction. So that anything which is done will have some consequence, and sometimes it will be very profound. Now we can ask again, why does the virus contain an oncogene? Of course one might think that is required by the virus for its own multiplication. Well that is not the case because it could be cut off and the virus would multiply but will not transform. So the function is purely for transformation. But then why should it be in the virus? And the answer to this question is really a very important point in all fields of cancer. Because the answer is that the oncogene present in the virus is actually a normal cellular gene. and it is incorporated into the virus by some event or a combination as is indicated here. The mechanism does not have to be the same, but essentially somehow a virus which does not contain the oncogene will find this gene present in the normal cell which is generally referred as a proto-oncogene, it will incorporate into its own genome. And out of this the sarcoma virus emerges. This is an extremely rare event but the occurrence of the event, however, is very important for us because it tells us that normal cells contain proto-oncogenes, genes with the potential of becoming cancer genes under proper circumstances. The study of many viruses has shown already that, has allowed the identification of some ten different oncogenes, proto-oncogenes in cells and therefore it is likely that there are more than this number, so there is quite a number of these genes with the cancer potential. This is a very fundamental thing because this shows that cancer really comes from the inside, although outside influences have great importance. Now why do the normal cells continue to be normal although they contain so many oncogenes? Well the reason for this is also quite clear and it is that in the normal cells these oncogenes are expressed at a very low rate. The amount of protein, this specific protein they make is very low, sometimes nothing. And this can be attributed, although not proven, to the fact that they are connected to a promoter which is very weak and therefore they allow only very low levels of expression. And therefore the activation that occurs when recombination happens is due to the fact that the oncogene becomes now under the influence, under the control of the viral promoter, which is of course very active because it must make more and more and more virus. The inactivity of the cellular promoter for the proto-oncogene must be due to the fact the cells must defend themselves against these viruses and they do this by confining the virus in some place. But now what is this oncogene, why does this oncogene exist at all? There are some results which probably clarify this point, namely it is found that two oncogenes which derive one from cats and the other from chickens are actually very similar to each other. So this means that they are conserved and if they are conserved it means that they are very important in some important function which we don’t know. So presumably they are expressed at some stage of ontogeny but we don’t know where we don’t know for what reason. We may think perhaps some kind of embryological development they might have some functions but we don’t know at all. There is another point I want to make that all this construction, the role of the promoter, the role of the oncogene which I presented to you as a result of virological research is actually been conferred in a most beautiful way by the technique of genetic engineering. In fact it has been possible to isolate a proto-oncogene from a normal cell into a vector as a DNA-clone, and then it has been shown that this DNA does not transform cells because, presumably, it’s attached and contains still the normal low grade promoter that is dead in the cell. But for now the gene is disconnected from its own promoter and attached to a viral promoter in a different clone. Then this clone, this DNA is able to transform cells at high efficiency. Therefore there is absolutely no doubt about certain things, namely the existence of the oncogenes, and the role of the promoter, namely the role of the expression of the oncogene, on the effect it has on the cells. Now let’s briefly consider the other group of these retro viruses, namely some of the most typical leukaemia viruses. These are viruses which cause leukaemia very slowly after infecting the animal may take many months to obtain the leukaemia. The generally do not transform cells in vitro. The reason for this is that they do not contain any oncogene. Then the question is, how to they transform? This has been now elucidated at least in one case and in other cases we have some ideas. We see in this slide the first up there is the sarcoma genome with the oncogene which we discussed before and the mechanism of transformation. But with the leukaemia virus one mechanism of transformation is that a piece of the viral genome goes and inserts itself adjacent to the proto-oncogene. And in doing this is gains control of the transcription of the oncogene, in a way of replacing the very active viral promoter to the inactive cellular promoter. Sometimes only the promoter is found in proximity of the oncogene. There is another type of tumours transformed or obtained by these viruses in which it has not been possible to observe this type of combination, that the viral promoter is near the proto-oncogene. Rather the proto-oncogene seemed not to be associated with any viral constituent. The mechanism for that is not known but this mechanism perhaps can be proposed, namely that the proto-oncogene is actually copied into DNA using an enzyme of the virus. And then this new DNA-copy goes back into the cellular DNA but it is inserted at another place where there is no defence against the oncogene, namely a place which is transcribed at a very high rate. Of course all cells have some parts where transcription is very active because all cells make some proteins in large amounts. Now in addition to these differences in mechanism between the sarcoma viruses and the leukaemia viruses, there are other differences. There is one very interesting difference which has to do with the type of cells which these viruses can transform or can render neoplastic. The sarcoma viruses have a broad spectrum, they seem to be able to transform many different kinds of cells, whereas the leukaemia viruses are very specific, they can only cause neoplasia by changing specific cell types, and in fact they produce each a leukaemia of different type. So what does it mean that the differentiation of the cell is important? This means that the repertoire of the genes expressed in a cell is important for the expression of the oncogene. So we can therefore deduce that the oncogene of the sarcoma virus is rather independent of the expression of other cellular genes, whereas the oncogene of the leukaemia, activated by the leukaemia virus, is very strongly dependent and can only be expressed under certain conditions of the repertoire of the cellular genes. Now, so one suggestion to clarify this point would be this; that once the oncogene is transposed to a viral genome, then it becomes very independent of other genes. Whereas when it is activated remaining still in the cellular genome, then it is very dependent. Why that should be so? One possibility comes to mind is that this has to do with the structure of the chromatin, namely that the chromatin in the viral genome is different from the chromatin in the other part, and this because the chromatin, because the viral genome is transcribed in much more active and that this really is the one difference between the two. Between these two classes of viruses, there are other viruses which have somewhat in between intermediate properties and another useful information is that there are some leukaemia viruses which actually carry an oncogene so they are in effect between the two. They have an oncogene like the sarcoma viruses but induce leukaemias. And so they are kind of differentiation-dependent. So this in effect leads to the generalization that the activity of the oncogene can be dependent to various extent on the expression of the other cellular gene and that this depends perhaps on the oncogene itself. Some are more dependent than others. So in summary now for the viruses we find that there are inactive proto-oncogenes in cells. They are activated by viral promoters and namely that the transcription rate is, the regulation of transcription is very important and that the effect of these genes in causing the alteration of the cells is dependent on the effect, on the balance of the other genes. The fact that there are oncogenes in cells makes one think that perhaps cancer is always the result of the activation of an oncogene. This would include both, also cancers which are induced by chemicals or human cancers. Now do we have some support for this generalization because recently it has been shown that DNA extracted from some cancers, chemical induced or human cancers, at least cancer cells, can transform normal cells. And in a way this the definition of the existence of an activated oncogene. Therefore we can say maybe all cancers are viral, because we know that viruses can activate oncogenes. Or there are different mechanisms of the activation of the oncogene which are not viral. Let’s look a moment at this first possibility. Is it possible that all cancers are viral? Well, there has been such an extensive amount of work over many years, especially in humans and in other species which has not given any evidence so far for the existence of a virus which with a general transforming cancer inducing activity. Of course we do know that in people there are viruses which have some role in some rather special types of cancer, but nothing which has a general role. So this means that we must consider a non-viral mechanism for the activation of proto-oncogenes. We will look at some results from chemical carcinogenesis. And there are two types of results. One is that chemical induced cancers occur after a long lag in humans, 20, 30 years can be the lag, and also that they occur in a series of steps. This is based both on epidemiological evidence, clinical evidence and experimental evidence in animals. An experiment which is very illuminating actually was conducted some 30 years ago by Beerenblum. He applied a weak dose of a carcinogen to the skin of a mouse and he notes that for many months this mouse did not develop any tumour in the area of the skin. But then he treated the same area of the skin some six months later with other substances which I will call facilitator and we will see a little but of them, and then the tumours appeared. Of course now there is lots of evidence which leads to the same general conclusion, and the general conclusion is that tumour is induced by chemicals through two steps, which in this particular experiment can be differentiated, in others may not be differentiable. One is the initiation step which leads to some kind of latent change which persists and then a facilitation step after a long period of latency and the facilitation is due to something which acts also for a long time and this now allows the latent lesion to be revealed. What I call facilitation usually is called tumour promotion, but I don’t want to use the word promoter because I have already used it for the transcription and I want to avoid any possible misunderstanding. Now let’s look at these events. Initiation, what is initiation? For its characteristics I think inevitably almost certainly is some kind of genetic alteration. I think it must be a genetic event which leads directly or indirectly to the activation of a proto-oncogene, but not to the expression of the proto-oncogene, to some kind of activation. Most carcinogens are mutagens and so they can induce a mutation, they can probably also induce transpositions, so either mechanism is open for this initial stage. However mutagenesis is not the whole story. That’s really the important point because there are carcinogens which are not mutagens. And these are serious carcinogens like asbestos or nutritional deficiencies or hormones. What is facilitation? This we know less about, however we know one thing that there are chemicals which are facilitators which alter the differentiation of the cells. And they alter in various ways. So a common action of these substances seems to be to change the repertoire of expressed genes in a cell. Therefore the influence is that this change of this repertoire is what allows the activated oncogene to become expressed. This of course is derived from what I said about viruses. It is also interesting that as I said this phenomenon, these two steps are not necessarily separable in all cases. In fact most carcinogens, the really active powerful carcinogens, are both, are both initiators and facilitators. Now why is that? I think that one possibility is that these substances are really highly reactive compounds, either directly or they generate them in the cells. And therefore they interact with all macromolecules in the cell. They interact with DNA and they produce the initiation phenomenon through that. But they interact within proteins in the RNA and in the past, in fact, there have been times when people have concentrated on these proteins effect of chemical carcinogens which are very widespread and thus quite characteristic sometimes. So it is possible therefore that the duality of action comes depending on the target molecule. Initiation means the target of DNA specific genes which can lead finally to the activation of the oncogene. And promotion is all these other alterations which occur in various macromolecules which somehow alter the network of regulation of cellular genes. The result of this is the disruption of the genomes which contributes to the activation of the oncogene. Now there is even a specific mechanism perhaps involved in this and the fact is that if through the operation of the network some, any gene becomes activated, then it becomes more vulnerable to mutagenesis or to other genetic damage. In the chromatin the inactive genes are all tied up with all these proteins and not very accessible to enzymes or chemicals but in the active genes somehow the structure is more loose that both enzymes and chemicals can reach the DNA more effectively. So this facilitating action may be very intimately interwoven with the initiating action because the two are really kind of helping each other. We have the DNA, a certain point to, the initial state in which there are certain number of genes expressed, the other are not expressed. These produce their own proteins when some can be, when the carcinogen arrive, some can be mutated or altered in other ways, these triangles, others are not, proteins are modified, these proteins are part of the network which interacts with the whole genome of the individual. And as a second subsequent step then we see that some of these interactions with the modified proteins that have lead to expression of new genes, which were not expressed before, and now some of these genes, which are now expressed, can be mutagenized and so there is a sequel of events. And finally this leads to what we consider the final state, in which an oncogene, proto-oncogene has been activated, there are many changes in other cellular genes, and maybe some genes which were originally expressed, cease being expressed and so on. And this phenomenon, this progression of events goes through some kind of path which is determined by two things. One is the initial state of the genome, and the other is all the various parameters which are involved in this network. And of course we know very little about both of them so certainly we cannot make any prediction of this at the present time. So finally, therefore we can understand the process of cancer as a kind of a selective process in which whenever this type of progressive events leads to something which will lead cells to be selected for, cell will grow better or more continuously, then of course it will persist and this will be the general alterations found in cancer, and in addition there will be lots of other changes which will depend on the path followed by this process and this alteration will be different from one cancer to another. So the first consequence is that a carcinogen in order to be an autonomous carcinogen that doesn’t need facilitation must be itself an initiator and facilitator. And most powerful carcinogens as I said are both. Now, most mutagenic agents, most of the carcinogens are mutagenic, as I said before, but some are not. Now why is that possible? One possibility is that these carcinogens are very strong facilitators, and that they cause the expression of oncogene already activated, or maybe by opening up some genes they allow new genes to be hit by background carcinogens. So in fact we can say facilitation is probably the most essential part in the process of carcinogenesis, because there are these carcinogens which are not mutagens. Now the role of facilitation, how important it is, can be really seen well by looking at what has happened in the evaluation of a cigarette smoke which is of course a powerful carcinogen which induces lung cancer. Originally it was thought that it would be carcinogenic because it contains benzopyrene, which is an initiator and probably also a facilitator, but essentially an initiator. But then it was clearly seen that there was too little benzopyrene to cause the effects that it causes, and so other studies then finally showed that cigarette smoke is also a powerful promoter, facilitator. And therefore the combination of the two actions is the one that really leads to the carcinogenic activity. Another second consequence of the model is that if facilitation is really the most important part of the carcinogenic phenomenon, then we should be able to measure facilitation in an easy standard way in order to know which substances are the facilitators which of course they would be those that should be eliminated from the human environment. And it turns out, however, that there is no adequate really good method to study facilitation, because imagine a part of this is because so much attention is being devoted to the mutagenic action in recent years that this part has been completely neglected and I think not rightly. Now, so mutagenicity is not a measure of facilitation, and even induction of cancer in animals is not a measure, an adequate measure. The reason for this is that animals are different from men or from humans in propensity to cancer and we know that this is true because of the, in effect, the incidence of cancer per unit time in humans is much less than it is in a mouse or in a rat. And this obviously is related to longevity. Now it is known that this difference is not due to different susceptibility to mutagenic action, because human cells or mouse cells can be mutagenized at comparable rates, so it has to be something else and it is likely to be just this which has to do with facilitation, namely the whole organization of the genome in its defence against the oncogenes. And of course this must have developed to produce an especially strong defence because of the longevity. An example of our inability really to measure these parameters is the case of saccharin. Saccharin is a good promoter, a facilitator of carcinogenesis in the bladder, in the rat bladder. And it does also cause cancers in the rat bladder. But epidemiologic studies in humans have not born out a carcinogenic effect of saccharin so far, so this again is another possibility that we are by just testing in animals we just don’t have the right answer for what concerns humans. A third consequence has to do with the role of the cell differentiation and this whole operation of this network of regulation which seems to have such a very important role in the expression of the oncogene both in the viral case and in the chemical case. So one possibility is that cells might be, cancer cells might be induced to revert to normality if they can be induced to differentiate. And there is actually an example of this. There is a tumour of mice, teratocarcinoma, it’s a tumour of embryonic cells, very malignant in itself, but since the cells, like all embryonic cells, tend to differentiate, they will produce a variety of tissues like skin or nerve cells or bones and so on. Well these differentiated tissues are not neoplastic anymore. So there is a loss of the cancer stage upon differentiation. So the question really is whether the same results could be achieved with other cancers. And this would require differentiation-inducing substances, and we know that differentiation-inducing substances exist. There are number of them that are being recognized in recent years in the hematopoietic system. So the question is whether there is enough of this substance, I mean we can identify enough to have a great deal of specificity, a great range of specificity to deal with many types of cancer, and of course we would have to purify these substances to have them available in large amount in order to hope to possibly achieve some result. Well it’s obvious that this will require lots of work but it will be interesting anyhow, and I think it is a very promising direction because it might lead to the development of anti-cancer agents which are both effective and physiological, which would be of course the best we could hope. Thank you.


Participating for the first time in a Lindau meeting, Renato Dulbecco began his lecture by explaining two different ways to study cancer: One as a medical doctor looking for causes, symptoms and treatments and the other as a biologist looking into the primary processes in the cell. With his background as both a medical doctor and biologist, Dulbecco could have chosen any of these two ways for his career, but at some point he decided to follow the second one. Since Alfred Nobel in his will wrote that one of his five prizes should be given in physiology or medicine, Dulbecco might very well have received a Nobel Prize even if he had chosen the first way! But with his choice of biology, the Lindau lecture concentrates on the topic that he and his student Howard Temin received their Nobel Prizes for (together with David Baltimore): The interaction between tumour viruses and the genetic material of the cell. That viruses can cause cancer in animals was shown in the early years of the 20th century by Peyton Rous, but his findings were not accepted by the scientific community until the 1950’s. At age 87, Rous finally received a Nobel Prize in Physiology or Medicine in 1966, after one of the longest waiting periods in the history of the Nobel Prizes! Since Alfred Nobel wrote in his will that the prizes should be given for work done during the preceding year, a frequent question is how the Prize-Awarding Institutions can wait for such a long time. The answer is that they didn’t accept the task of awarding the prizes until some more flexible rules were laid down, rules that can be found in the Statutes of the Nobel Foundation. At the time of Dulbecco’s lecture, the study of tumour viruses had made several important breakthroughs, in particular due to the methods of genetic engineering. So it must have been inspiring for Dulbecco to have in the audience two of the pioneers of this technique, Werner Arber and Hamilton Smith. Anders Bárány