Why Malignancy?

So far we have seen that the vast majority of the regenerating cells that keep our bodies going never become cancerous, and when such cells do make abnormal developmental decisions or escape from influences that normally guide their decisions, they do not inevitably follow the evolutionary road malignancy - indeed researchers have recently found that most people contain many small, localized populations of abnormal cells within their bodies that never amount to much in the way of threats. We have also seen that when cancer cells do become malignant they follow unique routes and schedules that can be influenced by a variety of factors such as time, exposure to carcinogenic influences and their past developmental histories. The prominent roles played by chance and contingency mean that what is commonly called the progression to malignancy is not so much progressive, in the sense of following preordained routes or pursuing particular goals, as it is cumulative, in the sense of involving the accumulation of cellular adaptations and mutations. There does, however, seem to be a certain relentlessness in the way some cancer cells evolve. For instance colon polyps have a sufficiently high likelihood of becoming malignant carcinomas that it is worth removing them before they get the opportunity. In the first chapter we encountered other kinds of evolutionary journeys that appear to have inherent directions and momentum - such as the embryonic development of complex organisms like us and the emergence of living communities like rainforests - and we also found that it was possible to make some sense of such journeys by putting them in context. For example many of the evolutionary changes that appear within organism lineages can be understood within the context of differential success in struggles to survive and compete for essential resources (i.e. natural selection), while the evolutionary journeys of regenerating cell lineages within the body can be understood in terms of the cumulative effects of various experiences and the decisions that cells make in response to them. Thus if we want to understand why some cancer cells seem to be bent upon becoming malignant, an obvious place to start is by putting their evolution in context by considering both the changes they undergo and the meaning of those changes to the cells and their surroundings.

We have already noted that the evolutionary journeys of cancer cells differ from those of other regenerating cell lineages not just because they make novel developmental decisions, but because some of those decisions are associated with genomic changes. As is often the case in the living world this distinction is not absolute, because genomic changes do figure prominently in the normal evolution of some regenerating cell lineages. The most extensively studied examples are found in the mammalian immune system, within the lymphocyte lineages that are involved with responses to invaders like parasites, fungi, bacteria and viruses. When such invaders first enter the body the initial response is usually mounted by the amoeba-like macrophages, which are myeloid cells that deal with foreign material by chemically attacking, devouring and digesting it. In many cases this rapid innate response is sufficient, but if a threat persists and spreads through the body it will usually trigger a secondary response in which lymphocytes and a variety of other cells do a highly efficient job of tracking down and eliminating the threat. Once such a secondary response has been activated, future encounters with the same invader - which can be anything from a malaria parasite to a protein molecule - will usually trigger an instantaneous reactivation of the response (i.e. without the delay associated with it initial appearance). This phenomenon is known as acquired immunity, and it is what ensures that we only get some diseases once (e.g. measles) - or indeed not at all if our immune systems are primed by exposure to weakened, killed or fractionated samples of pathogens in vaccines.

The first step in the development of one form of acquired immunity involves populations of lymphocytes that spend their time wandering around the body with no apparent purpose. These surveillance cells express immunoglobulin proteins on their surface that are similar to growth factor receptor proteins in that they are capable of detecting specific molecules or parts thereof, which are referred to as antigens. In contrast to factor receptors, which are typically expressed by many cells, each lineage of surveillance cells carries its own unique set of surface immunoglobulins, which means that each lineage detects a specific antigen (the same antigen may of course be detected by several lineages of surveillance cells, especially in the case of macromolecules that contain an assortment of potential immunoglobulin binding regions on their surface). When a surveillance cell encounters its antigen under the right conditions (more of that in a moment), it proliferates to produce a population of cells sharing the same antigenic specificity - this process is called selective clonal expansion. Cells within that population assume the responsibility of co-ordinating a full-blown secondary immune response, and after the threat has been dealt with a population of sensitized cells remains in the body, ready to raise the attack again if the need arises.

Thus unlike the innate immune response, which relies on cells that recognize and attack a wide range of invaders, acquired immunity involves the selective activation of cell lineages that recognize specific invaders. This is a highly effective strategy - for example acquired immunity can deal with viruses that overwhelm or evade macrophages - and there seems to be no limit to the secondary immune system's ability to recognize invaders ranging from viruses to malaria parasites. Indeed the system is so versatile that it can recognize foreign substances that no human body encountered until very recently in our history, such as the synthetic vaccines that are used to raise acquired immune responses and the synthetic drugs (e.g. antibiotics) which can raise less welcome responses in the form of allergies. It has been estimated that our immune systems are capable of recognizing and acquiring immunity to several million different antigens, which means that our surveillance cells must be capable of producing several million different kinds of immunoglobulins and other detector proteins (e.g. T-cell receptors). While it is impressive, this observation raises something of a biological puzzle, because according to current estimates the human genome contains fewer than a million genes, which means that even if they were all given over to making detector proteins there would still not be nearly enough to produce the observed diversity, assuming that detector proteins are made in the normal way.

As it turns out, a fairly modest portion of the human genome is actually dedicated to the production of immune system detector proteins, but there are some unusual aspects to how those proteins are made. Those aspects become apparent when we examine the production of the smallest and simplest immunoglobulin proteins, the light chains, which consist of two distinct domains, a variable one that recognizes antigens (i.e. the V region) and a joining section (i.e. J region) that attaches the light chain protein to its partners within an immunoglobulin macromolecule (each contains four or more protein chains). If we look inside immunoglobulin-producing cells we find that as we would expect their light chains are produced via the translation of messenger RNA macromolecules which are transcribed from genomic DNA templates (i.e. light chain genes), but a closer look reveals something unusual about the expression of those genes: each lymphocyte only expresses one light chain gene while its allele on the homologous chromosome remains transcriptionally silent. A closer look at the genomic regions near light chain genes turns up something even more unusual: located close by on the same chromosome are two distinct clusters of repeated base sequences, one with a half dozen copies of what appear to be J domain templates, and the other consisting of several hundred V domain templates. The V domain templates within those clusters show considerable variations their base sequences, as do the corresponding regions in the light chain genes carried by different lymphocyte lineages, and indeed different chromosomes in the same cell. The reason for this, to make a long story short, is because immunoglobulin genes are not passed intact from generation to generation, they are assembled during the development of lymphocyte lineages via a specialized form of chromosome recombination that randomly combines templates from the V and J clusters. Other detector protein genes are assembled in much the same way, usually from three or more distinct templates, and in combination with a few other tricks that enhance the diversity of their antigen-binding regions this random assembly ensures that by the end of embryonic development each human body possesses an enormous diversity of immune surveillance cell lineages, with a correspondingly broad capacity for recognizing antigens.

In fact the diversity of our surveillance cells is so great that a sizeable majority never encounter the antigens they have been randomly designed to recognize, and many acquire detector proteins that are primed to detect innocuous aspects of the environment (e.g. food, pollen), or even components of the body itself. Immune systems deal with these potential dangers by restricting the activities of their surveillance cells so that they do not simply react to antigens whenever and wherever they find them, rather the antigens must be presented to them by other cells in a specific context before they respond. There also appear to be controls within developing immune systems that lead surveillance cells which encounter their antigens early and/or often - as happens when the antigens are associated with common proteins in the developing body or environment - to become dormant or commit suicide, a phenomenon known as selective clonal elimination. The importance of these safeguards becomes evident when they fail to prevent inappropriate immune responses from being raised against environmental elements like pollens and foods, resulting in allergies, or against parts of the body, resulting in autoimmune diseases like lupus and type I diabetes.

The emphasis on self-tolerance is also the main reason why our immune systems are not very effective at recognizing or attacking cancer cells, which may not be a bad thing considering what can happen when they do. For instance some small-cell tumours can stimulate immune cells to respond to them, but if the antigens the immune cells recognize include normal cellular proteins - which is quite likely - the immune system can also start attacking normal cells like neurons with catastrophic consequences. This kind of situation is referred to as a paraneoplastic autoimmune syndrome, and the primary treatment strategy involves locating and removing the tumour that is stimulating the immune system in hopes that the response will subside once the main source of stimulation is removed - which can be tricky because the tumour may be quite small and relatively innocuous on its own.

Our immune cells provide rare examples of regenerating lineages that actively generate genomic novelty - albeit under very specific circumstances - and their evolution has another unique aspect in the specific suppression or deletion of lineages carrying potentially dangerous detector protein variants and the specific stimulation of the proliferation and development of lineages capable of responding to threatening antigens. One way of describing this phenomenon is in terms of a form of selection that operates within the body, or somatic selection, which discourages or encourages immune cells depending upon the specificity of their detector proteins. The variation in those proteins has a genomic basis, but there is no reason why somatic selection cannot work with purely physical variations as well, and an example of this can be found in the early stages of mammalian nervous system development.

When neuron precursors (i.e. neuroblasts) located in the embryonic spinal cord start sending out axons in search of developing muscles there are many chemical and physical cues available to guide the axons in the general direction they should go, but chance plays a significant role in the actual connections that form. Many of those connections turn out to be redundant or functionally useless, but this is not a problem because as the embryo develops only the neurons that have made useful connections become established in the nervous system, while the rest either redirect their axons to make useful connections or commit suicide. This approach may seem chaotic or wasteful when compared to the development of animals like worms and molluscs, where the same identifiable neurons usually make predictable connections with the same muscle and sensory cells, but this "hard-wired" approach is limited to relatively simple neural networks. Whether we are talking about organic evolution or the development of complex organisms, the evolution of large, complex nervous systems is associated with variability and selection - indeed one prominent theory of the development of neural networks is called "neural Darwinism."

It is likely that many of the cells, lineages and populations that appear during human embryonic development are subjected to some kind of selection at some point in their evolution; for example the minority of erythroid precursors that pick up enough erythropoietin to complete development may be said to be selected from the majority that commit suicide. Thus it is not unreasonable to think of somatic selection as playing a role within the body that is analogous to the role natural selection plays in the living world at large, although there are some fundamental differences between these two kinds of selection. As we saw earlier, natural selection is associated with the fates of a wide range of variants that can emerge among organisms, lineages and populations as they struggle to survive and contribute to posterity in the face of natural challenges and competitions for essentials like food, mates and living space. In contrast, somatic selection is associated with the fates of specific types of physical and genomic variations that can emerge among cells in the course of their normal evolution, and this selection requires the active participation of the cells themselves via their developmental decisions. For example the clonal selection of immune cells requires them to recognize and respond to a standard set of activation and suppression signals no matter what detector protein genes they express, while neuronal selection requires cells that make useless connections to try again or die. Thus natural selection influences the evolution of participants in struggles to survive, compete and contribute to posterity, while somatic selection guides the evolution of decision-making components within complex systems.

But what happens when some of the components of a complex system change the way they make certain decisions - for example when regenerating cells start dividing in the absence of normal growth signals or fail to commit suicide on cue? Such changes can disrupt the communication and control networks that normally co-ordinate the activities and development of regenerating cells and they can also give cells the opportunity to become participants in struggles to survive, compete and contribute to posterity. For instance localized populations that are able to access resources needed for growth - such as by attracting blood vessels - will be able to reproduce more rapidly than populations that cannot, while lineages that are able to colonize new sites in the body will become more widespread than those that stay in one place and cells that are able to evade natural and/or man-made defences will survive and proliferate while others are held in check or wiped out. Thus while all regenerating cells can be guided by somatic selection towards becoming useful components of complex organisms, cancer cells are also exposed to natural selection, which can encourage them to develop characteristics that enhance their survival and competitiveness - in other words as far as evolution and organization are concerned, the fundamental difference between normal cells and cancer cells is that the former are components of organisms while the latter are organisms.