CANCER SIGNALLING NETWORKS
The major ‘drivers’ of cancer development arise in pathways from receptor tyrosine kinases that ultimately control cell division. Thus, for example, aberrant activity of the RAS-MAPK pathway occurs frequently across a broad spectrum of cancer types. However, cancer-promoting mutations occur in other signalling pathways and, as we have seen, all aspects of cellular behaviour are ultimately affected. The molecular details of most of these pathways and how they can be subverted in cancers are now well established. The integration between diverse pathways creates a picture of a complex ‘information network’ rather than of discrete, linear systems. Despite the multiplicity of signalling pathways that can be involved, the tumour cells that emerge as the result of appropriate groups of mutations are phenotypically similar. This suggests that the diverse pathways converge on a ‘central axis’. We’ll then overlay the major pathways from receptor to nucleus to show how they impact on the central axis and then dissect from that complex map each of the pathways in turn to introduce the key players. The intention is not to show all known detail but to convey the principal features of the pathways with particular emphasis on identified (proto-) oncogenes and tumour suppressors.
One other critical control on cell division is provided by the braking effect of tumour suppressors in the cell cycle, the loss of function of these being examples of the other main category of ‘cancer mutation’. We noted that there are multiple receptors signalling through a considerable number of pathways and indeed one of the unresolved questions in cell biology is how cells make sense of a seemingly complex jumble of input signals. Mysterious though that is, cells clearly do make discrete decisions based on the integration of the messages they receive. An incipient tumour cell faces the problem of overcoming the defence systems built into these signalling pathways that have evolved either to permit repair of damaged DNA or to route irreparably stressed cells into apoptosis or senescence.
Fortunately, it has emerged that the critical cancer resistance mechanisms reside in what might be called a ‘central axis’ into which all the main signalling pathways channel. The essential message is mutations subvert these pathways so that they become hyperactive, in general, all will be well if the central axis players retain their normal function. It is only when they become oncogenically activated or, in the case of the tumour suppressors, lose function that the defences are breached and the prospects for the cell become markedly brighter – brighter, that is, in terms of its chances of founding a tumour clone.
From our point of view, this is very helpful because it means we can begin with a simple map that comprises only the central axis cast (Fig. 1).
1. The central axis of cancer signalling.
Two types of receptor are shown, and for the moment we will ignore the fact that multiple inputs are likely to be involved. Once we’ve considered how this bulwark operates we will take the plunge and put all the major pathways together, knowing that our map will then look somewhat intimidating. We can, however, pass quickly on from that by selecting each major pathway in turn and summarising the key players so that we’ll have met most of the genes and proteins that have substantial roles in cancer. To put them back into context we’ll end with an integrated picture of the phenotypic changes that occur when the defence systems have been overcome and a tumour cell has emerged.
The key players in the central axis of (Fig. 1) are, of course, RAS, MYC, ARF, INK4s, p53 and RB1, together with the E2Fs that are regulated by RB1. These pathways, principally featuring p53 and RB1, lie at the heart of the defence against cancer and in general, they work pretty well: despite the fact that cancers kill huge numbers, they are rare diseases at the molecular level and usually take a long time to develop into overt disease. Activation of either of the oncogenes (MYC and RAS) or loss of function of any of the tumour suppressors (ARF, INK4s, p53 and RB1) would constitute a major breach in the cancer defences and indeed almost all cancers feature at least one of these abnormalities.
The pathways leading to both p53 and RB1 are nearly mirroring images (Fig. 2).
2. Multiple controls of two key tumour suppressors.
In the case of p53, the protein MDM2 functions as a ubiquitin ligase that inhibits p53-mediated transcription. The ubiquitylation of p53 marks it for degradation – a highly efficient way of removing a critical tumour suppressor. By neutralising p53, MDM2 acts as an oncoprotein and as such it is quite often over-expressed in human tumours. There is a second layer of control for p53 in the form of ARF: ARF binds to MDM2, even when it is bound to p53, and inhibits its activity. This makes ARF a tumour suppressor – the protector of p53 – and this key role makes it immensely important as a target for cancer-promoting mutations. Loss of ARF function is one of the most frequent aberrations in human cancers. As well as interacting with p53, MDM2 associates with RB1: the result is relief of RB1 suppression of E2F trans-activating function. Thus MDM2 not only releases a proliferative block by silencing p53 but drives proliferation by stimulating the transcription factor E2F.
The regulator of RB1 that corresponds to ARF for p53 is INK4A. ARF and INK4A are encoded by the same region of DNA so it’s unsurprising that INK4A function is also lost in many cancers. INK4A inhibits cyclin D-CDK4: this is the kinase that phosphorylates RB1, thereby releasing the activity of E2F as a transcription factor and, in effect, switching on the cell cycle – so INK4A is a powerful cell cycle brake.
These pathways lie at the heart of the cellular defence so it is unsurprising that they feature sophisticated levels of control. Each arm includes a double negative in the shape of two tumour suppressors separated by a proto-oncogene. Oncogenic changes (e.g. expression of oncogenic MYC or RAS or loss of RB1 and hence activation of E2F) result in the up-regulation of ARF and hence cause cell cycle arrest or apoptosis. This is achieved not only by the indirect effect of ARF on p53 but also because ARF interacts directly with E2F to block its transcription factor activity. Thus ARF can be viewed as a dual-acting tumour suppressor protein in both the p53 and RB1 pathways – which is why its loss is such a disaster.
In Fig.1, E2F1 is shown as a mixed oncoprotein/tumour suppressor. Up until now, we’ve thought of E2F as a protein that switches on proliferation genes. However, it can also turn on apoptotic genes. The balance between proliferation and apoptosis is regulated by different signalling pathways, particularly those affecting CDK4/cyclin D and PI3K. In cancer cells, this balance is perturbed, frequently by activating mutations in the PI3K pathway.
The MYC protein is essential for normal cells to divide and cells very often lose control of MYC and make it in abnormal amounts, as much as one hundred times the normal level. That can happen when expression of MYC is driven through aberrant activation of a signalling pathway, although sometimes the MYC gene itself is mutated. Because over-expression of MYC is so common in cancers, it’s perhaps a very good example of what has been called ‘oncogene addiction’ – the requirement for a specific function for the continued growth of cancer. Unfortunately, quite often when a drug is used to block such an oncogene, cancer cells find a way around the block (they either mutate the target – again – or increase the level of other signalling molecules and use them instead). MYC may be an exception because of its unique role and it may be possible to make an effective inhibitor that works in humans without drastic side effects.
Remarkably, MYC can also be a tumour suppressor. How does it do that? Critical in this role is the fact that ARF is one of the many genes MYC can switch on. Thus when MYC is significantly over-expressed it activates the ARF–p53 axis. As we also noted above, overexpressed RAS acts similarly to activate both ARF and INK4A. So this sounds as though oncogenic MYC shouldn’t be a threat. All the same, there are two reasons why it is. The first is that the ARF-driven protection system appears to be activated only by large increases in MYC expression: if the oncogenic stimulus only mildly increases MYC expression the defensive radar of the ARF system doesn’t respond. So our main defence system, perhaps unsurprisingly, senses the level of damage and a small stimulus can persist undetected. The second reason why MYC is such a dominant force is that, if the ARF system is compromised (loss of function of ARF or p53), MYC now has a green light to drive unregulated cell proliferation.