I have seen death as a stranger sees another's world,

Or as a monster sees what the gods created

When they were drunk on wine and had a contest

To show the greatest harm that they could do.

- from the ancient story of Gilgamesh

Nasty Novelties

If we take a stroll through the cancer section of a decent science library we are likely to find books with titles telling us of wayward cells, misguided cells, outlaw cells, conspiracies of cells and other variations on the theme of cells running wild. When we pick up one of those books and check the glossary, we find that in modern medical parlance a cancer is a neoplasm or new growth, that has progressed to malignancy, a condition which literally translates as badly-born although the general connotation is of something that is dangerous and tends to become moreso over time. So without consulting the actual text of any of those books we can determine that the major themes of modern cancer research are new growths that are somehow associated with deviant cells that become increasingly nasty as they go along.

And how do we recognize these nasty novelties? The easiest neoplasms to pick out are those that push beyond the normal boundaries of tissues and organs to make themselves visible and/or palpable as growths such as warts, scales, bumps, lumps, polyps, cysts and tumours. Somewhat more cryptic are the neoplasms that involve dispersed populations of cells which become noticeable due to their abnormal abundance; for example an excess of one or more types of blood, lymphatic or hematopoietic cells is referred to as a leukemia (see the preceding chapter for a description of the hematopoietic system). Thus one thing that all neoplasms have in common is that they are associated with regenerating cells whose proliferation seems to be excessive with regards to the roles they would normally be expected to play in keeping the body's tissues, organs and physiological systems going.

Why do neoplastic cells behave the way they do? The first logical step in exploring what is going on with them is to take samples of neoplastic cells via biopsies or blood samples and examine them more closely, and as soon as we do that we find further signs of abnormality. For instance under the microscope normal tissues and organs can be seen to contain many different kinds of cells that are specialized in form, behaviour and physiological function, and these differentiated cells typically show a considerable degree of organization among themselves; for example the various kinds of skin or epithelium that cover the body's external and internal surfaces contain cells arranged into visibly distinct layers, with the lowermost containing cells that are capable of dividing while those above them have taken on specialized forms (in the outer skin the topmost cells are actually dead). When we look at neoplastic cells under the microscope we typically find more chaotic situations; for example inside colon polyps the neat cell layers of the epithelium are replaced by disorganized masses of undifferentiated cells, most of which appear to be actively dividing. Similar contrasts can be seen between normal hematopoietic cells and leukemic cells, which often show a preference for proliferating in immature states instead of proceeding along the developmental pathways leading to mature blood and lymphatic cells. For example lymphocytic leukemias are associated with an abnormal abundance of lymphoblasts, which under normal circumstances would be expected to give rise to mature lymphocytes such as B-cells and T-cells, and their failure to do so results in lymphocyte deficiencies or lymphocytic anemia. Some leukemias are associated with hematopoietic cells that are capable of completing development, and they produce a different situation; for example when erythroid precursor cells become hyperactive (see the preceding chapter for a description of their development) the result is an overproduction of red cells, or polycythemia, which can cause a dangerous thickening of the blood.

Thus in addition to abnormal growth and proliferation, neoplastic cells can also show unusual patterns of differentiation and organization that mark them out from the normal cells around them, although they can still share some obvious affinities; for example the cells in colon polyps can still be recognized as epithelial cells while lymphoblasts are clearly lymphatic cells. In some cases, however, neoplasms show such extreme divergences from their surroundings that they are obviously foreign. For instance hematopoietic cells do not normally cross the blood-brain barrier and enter cerebral tissue, yet some of the fastest-growing of all brain tumours are called lymphomas because they contain cells that can be recognized as originating in the lymphatic system. Similar indications of wanderlust have been noted in other kinds of neoplastic cells; for example tumours in the liver have been traced to cells originating in late-stage colon epithelium tumours, or carcinomas.

The apparent ability of some neoplastic cells to disperse and colonize other tissues, or metastasize, has obliged pathologists to come up with a descriptive system that recognizes both the location of neoplasms and the origins of the cells within them. For example a brain tumour like the one described above would classed as a cerebral lymphoma, while the tumours that spread from the colon would be metastatic adenocarcinomas of the liver (tumours that arise from liver cells are called hepatomas, and they can wander as well). Some of the other common and easily recognizable types of neoplasms are myelomas, which are derived from myeloid hematopoietic cells; sarcomas, which originate in connective tissue or bone (in the latter case they are called osteosarcomas); and melanomas, which are derived from pigment-producing skin cells.

Not all neoplastic cells are capable of dispersing and aggressively invading other tissues. Many appear in non-threatening locations where they remain static or slow growing, such as scales (e.g. skin keratoses), warts, cysts and the various lumps and solid tumours that are generally classed as benign. Moving up the impact scale, some neoplastic cell populations are capable of interfering modestly with normal body functions. For example chronic leukemias are associated with small, gradual or temporary alterations in the normal mix of hematopoietic cells, while acute leukemias are associated with major disruptions of the hematopoietic system which produce immune system deficiencies and anemias affecting various types of blood cells. Neoplastic hematopoietic cells can also give rise to tumours like lymphomas and myelomas, which along with fast-growing tumours that can arise in other tissues such as the colon can be highly dangerous to their hosts because their aggressive growth often results in the disruption of normal functions in the tissues and organs around them. When an aggressive tumour appears in a vital organ such as the liver or brain its effects can be debilitating and ultimately lethal, and even if a tumour is not life-threatening itself, its ability to serve as a spawning ground for metastatic cells means that it may be only a matter of time before a vital organ or physiological process is affected by its offspring.

Since neoplasms that reside at the nastier end of the malignancy spectrum represent the greatest threats to their hosts, it is understandable that physicians have traditionally concentrated their efforts on dealing with these urgent threats. But before a malignancy can be treated it must be identified and assessed, and this is not always easy to do because populations of neoplastic cells can hide within the body and it is not necessarily a straightforward matter to assess where they stand on the malignancy spectrum. Indeed neoplasms are not obliged to maintain the same status; for example when physicians detect polyps in a patient's colon they will usually recommend surgical removal even if the polyps do not look like carcinomas, because research has shown that some benign growths have the potential to evolve into malignancies - in other words they are not so much benign as premalignant.

These facts of neoplastic life mean that modern physicians are concerned not only with identifying and treating malignancies, but also with identifying and treating populations of cells that have the potential for developing malignant characteristics - or in common parlance of progressing to malignancy. This situation creates a bit of a conceptual problem, because the standard biomedical terminology that is used to distinguish cancers from neoplasms involves a designation that is difficult, arbitrary and often temporary. That is not much of a problem for people who are concerned with classifying and treating cancers, but for those of us who are simply trying to understand what is going on such fuzziness can get in the way, so for the purpose of our exploration I propose to simplify things by using the term "cancer cell" to refer to all neoplastic cells whether they are malignant or not. This is not so much a diagnostic designation as an evolutionary one, since it implies that all cancer cell populations and lineages have the potential to become malignant - which brings us to another key question: how do cancer cells get started on the road malignancy?

Regenerators to Rogues

In the previous chapter we established that the embryonic development, growth and regeneration of complex organisms involves the physical, but not genomic, evolution of cell populations and lineages. For instance the cells that regenerate blood and body tissues tend to follow specific developmental pathways that involve the emergence of distinctive structures, behaviours, functions and gene expression patterns, and those pathways typically end with cells that have lost the ability to divide (i.e. they are terminally differentiated). We also saw that the evolution of such lineages is associated with critical decisions that cells make in response to a variety of experiences and influences. For example fibroblasts stop dividing when they are surrounded by other cells and after they have gone through a specific number of divisions they commit suicide (i.e. undergo apoptosis), as do erythrocyte precursors in the absence of the hormone erythropoietin (in the hormone's presence they give rise to mature red blood cells). We have also seen that the impact of developmental decisions can vary considerably depending upon when and where they are made within a lineage; for example some hematopoietic stem cells are capable of giving rise to populations of many different kinds of cells (i.e. they are multipotential), but their descendants tend to lose options as the generations go by until eventually they are left with the decision faced by erythroid precursors: differentiate or die.

If we apply this knowledge to what we have learned so far about cancer cells, we can make some reasonable inferences about what is going on in neoplasms. To begin with, our understanding of the evolution of cell lineages would lead us to expect that cancer cells start out as regenerating cells, and since we have no reason to assume that one kind is more susceptible than another we would expect to find cancers associated with all sorts of regenerating cells. This is indeed the case; for example there are literally hundreds of different kinds of leukemias, most of which have been linked to a particular class of hematopoietic stem cells. The symptoms of those leukemias tell us something else we might have predicted, which is that the impact of cancer cells varies depending upon what their normal duties are; for instance leukemias that involve multipotential stem cells can have wide-ranging impacts upon the blood and lymphatic systems, while those that are associated with more mature stem cells have more limited effects (e.g. leukemic erythrocyte precursors only cause polycythemia).

Another obvious inference we can derive from what we have learned so far about cancer cells is that their defining characteristics of unrestrained growth and unwillingness to follow normal developmental pathways indicates that something about their evolution as regenerating cells has gone wrong. Or to be more precise, we can say that among the countless decisions made by the innumerable lineages and populations of cells that are involved in the construction, growth and maintenance of a complex organism like a human body, sometimes some cells go sufficiently astray to produce noticeably abnormal results. Of course we could say this about any number of physical abnormalities and deformities, but the unique thing about cancer cells is that as long as they are capable of dividing and making developmental decisions they have the potential to keep evolving after they become abnormal, and it seems that in many cases their evolution leads towards malignancy. But before we worry about how their journeys end, we should establish how cancer cells get underway.

Since we have already learned some useful things by taking them out of the body for closer examination, we can continue in this vein by looking at what happens when we raise cancer cells in the laboratory, which in many cases is fairly easy to do and produces interesting results. For example like normal fibroblasts, cells derived from epithelial tumours can be grown in culture dishes where they happily spread out to cover the available growing surface. However, unlike normal fibroblasts, tumour cells typically do not stop growing once they have covered the surface with a single layer of cells, rather they continue to proliferate in chaotic layers and clumps. If the cells from such clumps are disaggregated and put on fresh culture dishes they will once again proliferate and spread out, as will normal fibroblasts, but again unlike their normal cousins - which typically die off after a predictable number of generations - cancerous fibroblasts can often be recultured indefinitely; indeed some lineages have been perpetuated for decades in laboratories around the world.

Thus when compared to normal fibroblasts, cancerous ones seem to be both unrestrained and immortal, and the same traits have been observed in many other kinds of cancer cells. For example getting normal hematopoietic cells to grow in culture usually involves finding the exact blend of growth factors and nutrients, but leukemic cells tend to be much less fussy in their growth requirements and they can often be recultured indefinitely. These observations tell us that the abnormal tendencies of cancer cells are so deeply ingrained that the cells behave the same way outside the body as they do within it, and the persistence of those traits from generation to generation indicates that they are hereditary. Further confirmation of that fact comes from experiments where cancer cells have been taken from tumours and transplanted to other tissues in the same host or into different hosts, where they have gone on to give rise to new tumours.

At first glance these observations seem to be consistent with what we know about the normal evolution of regenerating cells within the body, where we also observe the emergence of permanent and heritable changes in physical form, behaviour and gene expression patterns as lineages evolve. However, since cancer cells do not follow the developmental pathways that normal cells pursue as they respond normally to normal influences and experiences, there is reason to suspect that the evolution of cancer cells may involve some unusual aspects. A clue to one of those aspects emerges when we put cancer cells back under the microscope, increase the magnification and use a few cytologists' tricks which allow us to see that some cancer cells have unusual nuclei where some regions of chromosomal DNA - and in some cases entire chromosomes - have become rearranged, duplicated or lost. The genomic rearrangements seen in cancer cells are reminiscent of the gross mutations we encountered in our earlier discussion of meiosis, and similar mutations have been observed to arise in cells that are exposed to nuclear radiation, X-rays and noxious chemicals (e.g. nitrogen mustard) - all of which have long been known to increase the risks of cancers in individuals exposed to them.

Another well-established risk factor for cancer is time: the longer we live the more likely it becomes that we will acquire a cancer; indeed some cancers are virtually exclusive to older people, like those that affect the prostate gland. As we also saw earlier, the passage of cellular generations is associated with the inevitable accumulation of genome replication errors, which are usually point mutations and small rearrangements rather than gross mutations like chromosomal recombinations and non-disjunctions. Yet while such rearrangements may be quite unlikely during normal cell division (i.e. mitosis), simple probability leads us to expect that the longer a lineage runs the more small mutations it will accumulate and the more likely it will become that gross mutations will appear. What all of this adds up to is the suspicion that while the normal evolution of regenerating cells involves changes in form, behaviour and gene expression but not in genomic information, genomic changes may feature in the evolution of cancer cells (along with the other kinds of changes).