Cancer. I. What exactly is cancer?
This article is about cancer and how we can think about it but I am going to begin from a broader perspective concerning science generally.
Most people when thinking about science focus on the results of scientific work, the discoveries that illuminate things in new ways. I would argue, however, that the crucial thing is how one frames the questions. That framing determines not only what is found but, often, how you think about it. Is there any general way to structure one’s questions to insure successful science?
A conventional initial step is to divide the questions into three types: the “what”, the “how”, and the “why” of something, whether that something is a process, phenomenon or entity. The “what” question concerns the basic definition of the subject; the “how” question concerns the manner of operation of the entity or process involved, and; the ”why” question, the trickiest, concerns the underlying conditions that are the source of what is going on. The “why” question, however, is often either ignored or treated with great distrust by many scientists. It seems to verge on the philosophical or even religious, thus being outside the realm of science. Yet it often creeps back in.
That preamble brings us to the specific question of the title, “what is cancer?”, a seemingly simple “what” question. Put simply, cancer is a pathological condition involving the uncontrolled overgrowth of cells in a multicellular organism that leads to great harm, often death, in that individual. This definition may seem all right at first but it soon begins to seem inadequate. It immediately raises the question: Why do cancers occur? and How is normal growth abrogated in cancer? These questions inevitably follow the “what” question and are connected.
A clue to the nature of cancer is that it is often a function of age. Cancer becomes more likely as one grows older. Although we lack good statistics from before the 20th century, we can therefore assume that cancer has become a more frequent problem as human lifetimes have lengthened over the last two centuries to produce older people though cancer as a phenomenon has been known since the dawn of human civilizations. The first published scientific explanation dates to 1914. It was produced by a great German scientist, Theodor Boveri (1862-1915). Boveri was one of two scientists who had independently proposed in 1902-1904 that the chromosomes, found in the cell’s nucleus, were the carriers of the genetic information in all plants and animals. That idea was a major insight and became central to the then new science of genetics.1
In 1914, Boveri extended this idea about chromosomes to try to explain cancer. He was a gifted microscopist and had been looking at the chromosomes of cancer cells, which he discovered often had abnormal structures and number; the more rapidly growing the cancer, the more the chromosomal abnormalities. Boveri hypothesized that there must be genetic factors in normal cells that keep cell growth under control. Correspondingly, in cancers, some of these inhibitory factors must be lost along with sections of chromosomes that were lost during the onset of cancer. The concept of “genes” existed by this time and Boveri was already thinking about there being genes whose normal function was to prevent uncontrolled cell growth.
Decades later, the term “tumor suppressor genes” was coined to denote this phenomenon. Boveri had predicted it.
His basic idea on the genetic basis of cancers was widely known and respected but, effectively, it lay dormant for many years in terms of scientific research. By the 1950s, however, it was getting new attention, and the suggestion was that the connection between ageing and cancer was the accumulation of mutations with time in body (somatic) cells, which at a certain point would trigger uncontrolled, that is cancerous, growth.2
It was a burst of discoveries in the 1970s, however, that established it. It was found that certain viruses that cause cancer do so because of certain specific genes termed “oncogenes” that they carried. Excitingly, the researchers soon found that normal human cells carried homologous versions of these genes, which did not have cancer-inducing potential. These normal forms were labeled “proto-oncogenes”. Altogether, there were deemed now to be two major classes of genes involved in cell growth, tumor suppressor genes and proto-oncogenes that controlled cell growth. Mutations in these, either loss-of-function mutations for the former and gain-of-function for the latter (see Should genes be seen as controllers or nudgers of biological development?) were believed to be at the heart of the genesis of cancers.
By the 1980s, this idea had come to be called the Somatic Mutation Theory (of cancer) or SMT and by the 1990s and early 2000s there was a mountain of data supporting it. Boveri’s bold speculation about cancer had become the reigning concept in the field. The idea was that most cancers are caused by the accumulation over time in ageing individuals of mutations in key genes. Was this really the definitive solution of the mystery of cancer?3
The fact that I am posing the question in this way probably suggests the answer: not really. The SMT is not false but it does not explain all cancers. There are (at least) five kinds of evidence that indicate that some cancers do not have a mutational basis.
Two of these kinds of evidence involve treatments that convert initially normal cells to a state that is a precursor state to cancer. This process, or group of processes, is known as “cell transformation”. Chemicals or treatments that can induce the pre-cancerous state are termed carcinogens. With such substances and treatments, one can investigate how this state comes about. Two sets of findings are not in accord with the SMT.
The first concerns the frequency of cell transformation by carcinogens. With many carcinogens, that frequency can be between 10% and 50%. This is far too high to be explained by mutation. Typically, the frequency of induction of mutations is on the order of 10-6 per gene per replication. If carcinogens act by inducing mutations and if there were 100 genes that give the pre-cancerous state, that might come to 10-4 or 1 in 10,000 cells transformed, not 10% or more. The implication is that a different kind of state is induced by these carcinogens. That state is a so called “epigenetic” one, which involves a stable regulatory change in gene but not a change in the DNA sequences.
Second, a prediction of the SMT is that if all cancers start with mutations, then all carcinogens should act by inducing DNA damage. Many carcinogens, in fact, do this and are said to be “genotoxic” but not all. It is believed that between 12% and 23% (two different studies) of all carcinogens are not genotoxic. That, of course, does not predict what percentage of cancers arise from something other than mutation but it does make it certain that the SMT does not explain all cancers.
The third piece of evidence against the SMT involves a particular kind of cancer, so called “epithelial cancers”. Epithelia are thin tissue coverings of many organs and they can develop cancers. They are underlain by another tissue termed stromal tissue. In certain laboratory conditions, the stromal tissue can be treated with carcinogens, then recombined with epithelial tissues, and the latter can then develop cancers. (In such experiments, it was ruled out that there had been carry-over of the carcinogens from the stromal to the epithelial cells.) This result cannot be explained by the SMT. Mutations cannot be transmitted from one tissue type to another. There must have been some cell contact or cell substance transfer from the stromal cells to the epithelial cells that induced cancerous development in the latter.
The fourth type of evidence shows the opposite phenomenon, the suppression of cancer development by normal cells. This involves experiments in which the nuclei of normal cells are replaced by nuclei from cancerous cells, and then small numbers of those cells placed in contact with large numbers of normal cells. The nuclear transplant cells survived and multiplied but they did not grow into patches of cancerous tissues; their descendants contributed to normal tissue development. This experiment shows that it is not enough to have a rogue cell with cancer-promoting mutations as the SMT maintains. Evidently, the tissue “microenvironment” is also crucial to how a cancer cell grows. A large number or normal, non-cancerous cell neighbors, can ensure that the cell behaves like a normal one.
Five, as mentioned above, certain viruses induce cancer. These include human papilloma virus (HPV), Epstein-Barr virus, HIV (the virus that causes AIDS) , and Hepatitus B, C, and D. The ability of a virus to induce cancer is not a new fact. That initial finding was made by a French scientist, Peyton Roux, and published in 1911. (He would receive the Nobel Prize in Medicine & Physiology for this work 50 years later.) However, the existence of other cancer-causing viruses came to be appreciated only much later. It is, of course, not known what percentage of human cancers have a viral origin but the best estimates range between 10 and 20%. The SMT cannot apply: here states of cancer promoted by viruses come about through infection, not mutations.
How does one put all this together? On the one hand, there is much evidence that many cancers are caused by an accumulation of mutations in certain tissue, supporting the SMT. It is also the simplest explanation as to why cancer is an age-related condition, one triggered by the cumulative effect of genetic mutations over time that release the brakes on cell multiplication. On the other hand, the SMT cannot explain all cancers. There are tissue-level effects that either induce or suppress the development of cancers. Furthermore, certain carcinogens and certain viruses can promote cancers and do so without the involvement of mutations. The explanation for the mutation-independent cancers has been called the “the tissue organization field theory” or TOFT.4
In the 1990s and the first decade of this century, there was often heated debate between these two schools of thought about the basis of cancer. Advocates of one position or the other insisted that both could not be right, that it had to be one or the other. The heat has largely gone out of this debate as the field has shifted from questions about causes of cancer to new forms of treatment of cancer, primarily immunological, which do not depend on understanding the underlying causes. One can now see that both explanations might be right, just not for the same cancer case. For some cancers, the SMT might be right, in others, the TOFT might hold. The end-result – totally out of control cell division – might be the same while the starting points might be very different.
We have seen comparable situations before, where there seems to be a shared end-point but different starting points for a condition in different cases. These are two other age-related diseases, Alzheimer’s disease (see The terrible tangle of Alzheimer’s disease) and Parkinson’s disease (see Parkinson’s disease. II. The beginnings of an answer?). There is no reason why cancer, which comes in many different varieties, should not be similar.
I will come back to the subject of cancer in at least two further pieces. In one, I will examine the new immunological methods for defeating cancer and in another, I will look at cancer’s possible evolutionary roots. The latter subject will bring us back to the “why” question about cancer.
The other scientist proposing that chromosomes were the carriers of biological heredity was a young American geneticist, Walter Sutton (1877- 1915). Sutton published first but Boveri had been working on chromosomes longer and had probably had the idea first.
An article discussing Boveri’s cancer idea is that of Hansford, S. and D.G. Huntson (2014). Boveri at 100: Theodor Boveri and genetic predisposition to cancer. J. Pathol. 234(2): 142-145. Doi: 10.1002path.4414
For one review of the SMT, see Bodmer, W.F. (1994). Cancer genetics. Br. Med. Bull 50(3): 517-526. Doi: 10.1093oxfordjournals/bmba072907.
For an exposition of the reasoning behind TOFT, see Soto, A.M. and C. Sonenschein (1998). Society of Cells: Cancer and Control of Cell Proliferation. BIOS Scientiific Publications: Abingdon, Oxfordshire
Really appreciate the nuance here about SMT vs TOFT not being mutually exclusive. The part about carcinogen frequency being way too high for pure mutation-based explanations (10-50% vs 10^-4) is the kind of quantitative argument that should've settled this debate decades ago. I've noticed a similar pattern in other disease research where people get locked into single-mechanism thinking even when the data screams heterogeneity. The tissue microenvironment stuff reminds me of how much we underestimate contextdependent biology in general.