THE FIRST PERSON WHO RECORDED THE SPREAD OF CANCER
Laënnec was the first person to record that cancer could spread to secondary sites; the introduction of the term metastasis is credited to a French surgeon, Récamier, who succeeded him as a member of the Collége de France. The initial demonstration that the metastasis of tumours derived from epithelial cells is caused by malignant cells leaving the primary site and spreading through the body was provided by the German surgeon, Karl Thiersch. This was at odds with the notion proposed by the more famous Rudolf Virchow, father of modern pathology that cancers spread within an organism through some sort of liquid medium that somehow changes connective tissue cells at a secondary site into metastases. It was, therefore, a while before the notion that cells migrate from a primary tumour to form a secondary became established but, once it did, the obvious question arose: when cells metastasise, how do they know where to stick? Or, as Stephen Paget more elegantly phrased it in a landmark paper of 1889: ‘What is it that decides what organs shall suffer in a case of disseminated cancer?’
The simplest answer would be that it just depends on anatomy: that is, cells leave a tumour and then adhere to the first tissue to which the circulation carries them. But Paget had noticed that quite often this simply didn’t happen and in his paper, he described his own evidence and summarised the work of a number of other luminaries to show that ‘the distribution of secondary growths was not a matter of chance’. Paget’s speciality was breast cancer and in 735 fatal cases, he had found 241 with liver secondaries, 70 lung metastases, 30 kidney metastases and 17 metastases to the spleen. He also noticed that secondary tumours from the breast occur with marked frequency in the bones. Paget gives full credit to the contributions of others and particularly records Fuchs ’s prescient suggestion in 1882 that certain organs may be ‘predisposed’ for secondary cancer. All of which led Paget to a botanical analogy for tumour metastasis: ‘When a plant goes to seed, its seeds are carried in all directions; but they can only live and grow if they fall on congenial soil’. From this, then, emerged the ‘seed and soil’ theory as at least a step to explaining metastasis.
The great strength of the seed and soil view is that it conveys interplay between the tumour cell and normal cells and that their actions collectively determine the outcome. As we shall see shortly, that is at the centre of metastasis. Nevertheless, even at the beginning of the twenty-first century, the honest answer to the question ‘What controls the spread of tumours?’ is ‘We don’t know’. However, the facility with which genomes can now be sequenced is revealing the extent to which mutational evolution can continue in secondary tumours and hence that each is as unique as its primary precursor. Despite this continuous evolution, the presumption is that there are some common molecular strands to the dissemination process and these are gradually coming to light with the increasing sensitivity with which proteins and sub-sets of cells can be detected.
If we accept that the sites of metastases are not simply a consequence of anatomy, perhaps we should take a step backwards and ask how cells acquire metastatic capacity. Unfortunately, even that remains a pretty murky area but it seems probable that a subset of the mutations that are acquired early in tumorigenesis, giving a proliferative advantage, are also able to promote metastasis in later generations of tumour cells that have acquired further mutations. This idea is intuitively attractive because it implies that a primary tumour or at least a subset of the cells therein, gradually develops the capacity to make the critical step – dispatching cells to other locations – that will ultimately be fatal for the host. This scenario has been confirmed in pancreatic cancers by the application of whole genome sequencing to reveal that the primary tumours contain a mixture of sub-clones each localised to distinct areas within a tumour. Distant metastases in different tissues (liver, lung, and peritoneum) are derived from specific clones, each of which in turn developed from a single, parental, non-metastatic clone. Based on estimates of proliferation rates, at least ten years are required from the initiating mutation to the appearance of the parental, non-metastatic cell and a further minimum of five years for metastatic capacity to appear. On average patients live for two years after this event.
Notwithstanding the logicality of that model, as ever there are observations indicating that at least some cancers do it differently. Occasionally metastatic growths appear when no primary tumour can be detected. These are classified as ‘cancers of unknown primary’ and they are not that uncommon, falling in the top 10% of diagnoses. Ascertaining the primary tissue of origin for these tumours by conventional histology is often difficult, a problem that is being alleviated by the use of gene expression profiling to identify diagnostic patterns. Whatever the molecular events that promote these tumours, they clearly do not require the prolonged development of a primary. In mice with human breast cancer tumour cells implanted in a mammary gland, the human cells can be detected in bone marrow when the mammary glands show only increased growth of normal cells (‘atypical ductal hyperplasia’) – the earliest pathologically detectable stage of breast cancer. Similarly, in humans, equivalent numbers of tumour cells have been detected in the bone marrow of patients regardless of the development stage that their primary tumour has reached. That is, metastasis had occurred to much the same extent in patients at the earliest stage as at each of the later stages.
These results indicate that at least some primary tumours are capable of releasing cells that relocate to metastatic niches during the earliest phases of tumour development. This suggests that critical steps required to produce a fully malignant tumour can occur in metastases, not just in a primary tumour. This evidence also shows that for breast cancer at least, the poorer prognosis of patients with late-stage primaries is not simply because they have exported greater numbers of cells to metastatic sites. One possibility is that the larger primaries may be releasing factors that promote the growth of metastases.
Regardless of when, in the development of a primary tumour, cells become able to leave and form secondaries, it is now possible to look at the entire pattern of genes being expressed in a population of cells. Such analyses show that the patterns of gene expression in metastatic cells differ from those of their non-metastatic counterparts in a primary tumour. What appears to happen is that in response to signals they receive from nearby cells or other environmental triggers, cells start to make proteins that enable them to detach from a primary tumour and initiate the process of intravasation. Given the huge range of genetic disruption that occurs in cancer cells, there seems no reason why this shouldn’t happen very early in the life of some primaries, although one might suppose that the older the tumour the greater the chance that it will come up with a recipe for spreading.