What effects do mutations have upon cancer cells? On the surface this is a trivial question, because in the case of something like the ras oncogene mutation the answer is obvious: it makes them cancerous. Why does it do that? Because the mutation causes a specific change in the amino acid sequence of a particular protein. And how does that change turn a well-behaved cell into a cancerous one? Well, the mutation alters the way the protein acts within the cell, so that.........as it turned out, filling in that particular blank for ras took a small army of researchers about fifteen years, and similar situations developed for many of the other cellular oncogenes that turned up. We will consider those findings shortly, but first we will retrace a useful shortcut that allowed researchers to build up a large catalogue of cancer-associated genes and mutations in a fairly short period of time.

Informative Invaders

One of the unique aspects of modern biology is the extent to which researchers have relied upon tools provided by nature. If, for example, if we were to use current technology to recover a ras oncogene mutation from a population of cancer cells, we would first chemically extract the DNA and cut it into pieces using enzymes called restriction endonucleases, which were originally isolated from bacteria. We would then splice the pieces into bacterial or viral DNA using viral enzymes called DNA ligases and offer the recombinant DNA to the appropriate host cells, which might be bacteria, yeast or mammalian cells. We would then grow up large numbers of those cells in culture in order to determine which ones had taken up our "cloned" oncogene using one of several methods involving a variety of enzymes, proteins (e.g. antibodies) and nucleic acid molecules, and once we had isolated the gene we were looking for we could determine its exact base sequence using techniques which employ components derived from the replication and transcription machinery of assorted cells and viruses. If we wanted to, we could also express our cloned gene as protein in a test-tube using tamed translation complexes.

This reliance upon what nature provides has made the history of modern cell biology something of a bootstrap affair, where researchers exploring how viruses, bacteria and other kinds of cells go about the business of survival have discovered new biomolecular and macromolecular tools that have allowed them to probe more deeply into the inner workings of cells and viruses. Indeed looking back, it almost seems as if cells and viruses were designed to present explorers with a series of keys that allowed them to unlock successive mysteries, and this fortuitous strain is especially strong where cancer research is concerned, because much of what has been learned about how regenerating cells become cancer cells has come from experiments involving some of the most mysterious things in the living world. A case in point concerns the acutely transforming retroviruses, which are so important that the researcher who discovered them, F. P. Rous, was awarded a Nobel prize for his efforts, but they are also so unusual and obscure that Rous had to wait for several decades to collect his prize because it took that long before the scientific establishment was convinced his discovery was genuine.

Rous began his work early in the 20th century, and by 1911 he had isolated viruses that had the unusual ability to induce rapidly-growing malignant sarcomas in infected chickens, hence he named the isolate the Rous Sarcoma Virus, or RSV. To make a very long story short, RSV turned out to be the first of what has since become a sizeable family of oncogenic viruses, so named for their apparent ability to transform regenerating cells into cancer cells simply by infecting them. Coincidentally, RSV also turned out to be the first known example of a retrovirus, which is a virus that injects its genome into host cells in the form of RNA macromolecules that are subsequently transcribed into DNA sequences (i.e. backwards or retro to the cellular transcription of DNA templates into RNA sequences). This trick is accomplished by viral enzymes called reverse transcriptases - or more correctly, RNA-dependent DNA polymerases - and once the retrovirus genome has assumed a DNA form it can splice into one of the host cell's chromosomes to become a provirus, which can be replicated and passed on with the rest of the host cell genome during division. The provirus can also provide templates for the production of viral macromolecules which self-assemble into new particles, a reproductive process that generally does not kill the host cell but simply converts it into a source of new retrovirus particles.

In the decades since RSV was isolated several other retroviruses were discovered - with the most famous member of the family being HIV (human immunodeficiency-associated virus) - but few of them have been linked to cancers. Nevertheless, the acutely transforming retroviruses still rank among the most potent oncogenic agents known, which is a bit of a puzzle because it is not immediately obvious why viruses that normally avoid killing host cells during their own reproduction would turn around and start killing host animals with lethal tumours. The first major clue to the link between retroviruses and cancers came with the observation that acutely transforming retroviruses are inevitably found in close association with retroviruses that are not oncogenic, or only weakly so. For example RSV-infected chickens are always co-infected with the Avian Leukosis Virus (ALV), a somewhat misleading name because most animals infected with ALV do not develop the slowly-developing form of leukemia the virus is sometimes associated with.

When virologists set about studying why RSV always appears alongside ALV but not vice versa, they found that this odd state of affairs has a simple explanation that is related to the basic facts of viral life. As we saw in the first chapter viruses are highly efficient survival machines, and retroviruses have carried this efficiency to such an extreme that some of their genomic sequences can serve as templates for more than one protein - in other words their genes overlap. This means that any mutations or rearrangements that appear in a retroviral genome are liable to affect one or more proteins, which leaves little leeway for genomic changes. Another restriction is imposed by the physical limits on the size of the genome that can be packed into new viral particles, which means that if anything is added to a retrovirus genome something else has to be removed - which will of course foul up any number of viral genes. What all of this adds up to is the expectation that retroviral genomes will not change very much from generation to generation except for the odd point mutation, although there is one theoretical possibility for radical change: if an altered genome appears within an infected cell it can be replicated and packaged as long as it is not too large and there are genomes present that are capable of making the necessary viral components.

To once again reduce years of hard work into a glib pronouncement, that slim possibility was responsible for the survival of RSV, which is only capable of reproducing when the necessities of life are provided by ALV infecting the same cell (ALV can of course survive quite happily without RSV). The fact that ALV can serve as a helper virus for RSV tells us something important about the relationship between these two viruses, or rather between these two viral genomes since the particles are the same: RSV must be a mutated form of ALV. That fact was confirmed when biomolecular technology reached the point where it was possible to compare the RSV and ALV genomes directly. They turned out to be similar in size and have identical base sequences at either end, but in between things are quite different. As expected the ALV genome contains templates for a complete set of proteins similar to those found in other retroviruses, but in place of several of those genes the RSV genome contains a sequence that is unlike anything seen in any normal retrovirus. That sequence is, however, expressed in RSV-infected cells, and since its expression is associated with the cancerous behaviour of those cells the RSV "intruder" sequence was dubbed the src retroviral oncogene. src (pronounced "sark") was the first of several oncogenes to be isolated from retroviruses, and in each case the story has been pretty much the same: the oncogene looks like it has been dropped into the viral genome from somewhere to displace normal sequences. And where have the oncogenes dropped in from? Here again the most likely scenario can be deduced from what we know about the retroviral life cycle. Aside from its own genome the only other DNA sequences that a typical retrovirus ever has any intimate contact with are those of its host, which it encounters when it integrates into chromosomes in the form of a provirus. Thus if a retrovirus is going to pick up new genomic information from somewhere the most likely source is the host genome, and sure enough when the base sequence of the src retroviral oncogene was compared to host cell DNA a reasonably close match was found: the src cellular proto-oncogene. Unlike the situation with ras, where the normal and oncogenic alleles differ by a single point mutation, the viral src oncogene turned out to have several point mutations and some bases missing as well, but there was no doubt that it originated in the genome of an infected cell.

Without getting too deep into the details, even what we already know about retrovirus reproduction would not lead us to expect that viral genomes will not pick up cellular DNA very often, nor that many mutant genomes will be just the right size to be packaged, nor that many mutant viral particles will end up in situations where they can reproduce with the assistance of a normal helper. And even if all of these events did come together to produce a viable mutant retrovirus, there is little reason to expect that a genomic sequence picked up at random from the cellular genome will turn out to be capable of turning infected cells into cancerous ones. This chain of low expectations is confirmed by the observation that RSV remained the only acutely oncogenic retrovirus known to science for decades, and by the fact that even with intensive searching only a handful of other examples have turned up in nature. However, in keeping with their willingness to follow nature's lead, researchers did learn how to up the odds in their favour - or I should say in favour of oncogenic viruses - in the laboratory. For example they found that by infecting large numbers of hosts and screening billions of viral particles they could isolate enough rare oncogenic mutants to assemble a catalogue of retroviral oncogenes that now runs into the hundreds, as of course does the list of corresponding cellular proto-oncogenes.

That catalog of genes, supplemented by others isolated directly from cancer cells, has given researchers many leads to follow as they have explored how oncogenic mutations can influence cellular behaviour. The results have yielded an embarrassment of riches concerning what can go wrong in cancer cells, and one way of organizing this information is to begin outside the cell and work our way in.

Communication Breakdowns

In the first chapter we had a look at how mammalian red blood cell production is regulated by oxygen-sensitive cells in the kidney, which communicate with erythrocyte precursor cells via the hormone erythropoietin. This long-distance monologue is one of many conversations that have been teased out of the enormous and intricate communication networks which extend throughout the body and involve a host of chemical signals ranging from ions like sodium to proteins like erythropoietin. Those signals are capable of influencing vital cellular decisions such as whether to divide, die, differentiate or turn particular genes on or off, and since cancer cells appear to make questionable decisions it is a reasonable bet that for some their problems begin with their communication with other cells. That bet has paid off handsomely for the researchers who have taken it; in fact cancer cells have been linked to so many defects in cellular communication that they have become invaluable tools in efforts to understand how normal cells pass signals between themselves, how those signals penetrate into cells and how they influence cell behaviour.

It is not possible for us to explore all of these aspects of cell communication in any great detail, but we can cover some of the ways that communication breakdowns have been found to allow cancer cells to "get off the leash," and what we have learned so allows us to make some preliminary guesses as to how that might happen. For instance if a specific hormone or growth factor must be present in order for a particular population of cells to divide and/or complete development, then presumably anything that caused a substantial increase in the supply of signal molecules would have a significant impact upon the cells that respond to it. An example of this phenomenon has been observed in some patients afflicted with polycythemia (i.e. chronically high red blood cell counts) which has been traced to abnormally high erythropoietin levels caused by cancerous hormone-secreting cells in the kidney - this condition is referred to as secondary polycythemia because it originates outside of the hematopoietic system proper.

Another obvious way that regenerating cells could escape from reliance upon stimulatory signals coming from outside would be to produce their own signals, which is theoretically possible because each cell in the body carries the same genome and hence any regenerating cell can provide its own growth factors simply by turning on the gene(s) involved in their production. Of course under normal conditions cells do not produce the factors they respond to. For example platelet-derived growth factor-ß (PDGF-ß) is a powerful stimulator of proliferation for several types of target cells, but as its name suggests the expression of this protein is restricted to the hematopoietic lineage that produces the smallest cells in the body, the platelets. Some cancer cells, however, have been found to carry a mutant version of the PDGF-ß gene called the sis oncogene, which is expressed in all cells that carry it and allows them to produce their own supplies of growth factor - in other words they become autocrine cells. This short-circuiting of the normal cellular communication network allows sis mutant cells to proliferate independently of external signals.

While it is certainly an effective way of eliminating reliance upon outside signals, the odds of a regenerating cell becoming autocrine are fairly long because that transition requires a mutation that activates the expression of a specific gene among the hundreds of thousands in the genome. There are, however, other ways of getting around external restraints. For instance cells can change the way they respond to signals, and the opportunities for such changes begin with the molecules that cells use to detect molecular signals: the receptor proteins.

Receptor proteins are often located, appropriately enough, in the outer cell membrane, where they are well positioned to interact with molecules secreted by other cells. Receptors tend to be rather specific in the way they interact with signal molecules; for example erythropoietin molecules fit into three-dimensional pockets within erythropoietin receptors, while other growth factors are ignored. The responses of receptors to the signals they receive are also quite specific; for example when erythropoietin receptors bind to hormone molecules they come together on the cell membrane to form pairs (i.e. dimers), and these complexes behave quite differently from "free" receptors when they interact with other molecules inside the cell. We will get to those interactions in a moment, but we have already encountered an opportunity for mischief: if receptor proteins changed the way they interacted with and/or responded to factors, they could substantially change the way the cells possessing those receptors responded to external signals. An example of such a change has been observed in mouse cells carrying a mutation that changes erythropoietin receptor proteins in a way that leads them to form dimers in the absence of erythropoietin, producing a situation where mutant erythroid precursors do not commit suicide even in the absence of erythropoietin.

And why does such a mutation stop erythroid cells from committing suicide? This brings us back to what happens when normal erythropoietin receptors encounter hormone molecules and bind to each other. The simplest way to describe what happens next is that the receptor complexes located on the cell surface trigger molecular interactions within the membrane that are involved with the passing, or transduction, of signals through the membrane and into the cell. While the physical distance involved may be modest distance, the membrane signal transduction step can be quite complex and involve several different proteins. For instance erythropoietin receptor dimers react with enzymes inside the membrane called protein kinases, which add phosphate groups to amino acids on other proteins in ways that make them chemically active, triggering chains of reactions involving other membrane proteins, including some located on the inner surface of the cell membrane, where they can pass signals into the cytoplasm.