Metastatic disease is a major cause of morbidity and mortality. However, it is not clear if all metastases originate from the primary tumour or whether metastases themselves have the capacity to metastasize. In this review, the evidence for the latter phenomenon, and its biological and clinical implications, is discussed. Copyright © 2004 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.
The major cause of death from cancer is metastatic disease 1. In recent years, the understanding of how primary cancers metastasize has increased. This in turn has led to research into new treatment modalities to target specific steps in the metastatic cascade. However, do metastases metastasize? Or do all metastases originate from the primary lesion with early dissemination of tumour cells and micrometastases, which grow after a variable period of dormancy? These questions have been explored previously but remain unanswered. The clinical implications of this are important. Should all metastases be treated early, before they lead to further metastases? Is this why adjuvant therapy is more successful in terms of improving survival compared with treatment at recurrence? If a patient with a single pulmonary metastasis develops further metastases, did these develop from the first metastasis, and therefore should the ‘single’ metastasis be treated aggressively?
Pathogenesis of metastasis
The three main routes for tumour metastasis are direct spread, via the lymphatic system, and blood borne. The process of metastasis is inefficient. It consists of a series of sequential steps, which many cells fail to complete successfully. Initially, cells must invade local vascular or lymphatic channels 2. They are then transported in the circulation and must survive physical destruction by haemodynamic forces and host immune defences 1. Prior to becoming an established metastasis, cells must extravasate, proliferate, and migrate to a suitable location, where they initially form a micrometastasis 3. Once established, they must secrete factors to promote angiogenesis if they are to grow beyond 1–2 mm in diameter 4. The primary tumour is regarded as being heterogeneous. It consists of some cells that are capable of metastasis and some that are not 5. It is the highly metastatic cells that are successful in forming distant metastases and it is believed that all metastases arise from these cells 5. But do these cells then remain at this distant site once established, or do they retain their superior metastatic properties and re-metastasize?
Metastasis of metastases
There is little in the literature addressing the issue ‘do metastases metastasize?’ August et al presented a series of case studies, which they proposed supported metastasis of metastases. They reported seven patients who had developed hepatic metastases from colorectal primary tumours. At laparotomy for potential hepatic resection, metastatic disease was found in lymphatics draining the liver. There was no evidence of other lymph node involvement, no local recurrence, and no intra-peritoneal extra-hepatic tumour. It was suggested that the lymphatic disease was due to re-metastasis of the hepatic metastasis from the primary colorectal tumours 6. However, this is an anecdotal report of a small number of patients.
Experimental models have been inconclusive. Hoover and Ketcham induced primary tumours in mice and then amputated the primary tumour and left the mouse for several days to allow pulmonary metastases to develop. The host from which the primary tumour had been amputated was then joined at its peritoneal cavity with an immune syngeneic mouse for 3–4 weeks and they were then separated. This allowed the two mice to be considered as one animal with two sets of organs but only one circulation. The host mice and the guest mice were sacrificed and analysed for metastases, which were present in 80% of the guest mice 7. They thus concluded that the metastases in the host mouse had re-metastasized to the guest mouse. However, dormant disseminated tumour cells in the circulation of the amputee could explain these findings. Conversely, Sugarbarker et al transplanted pieces of healthy lung into mice with pulmonary metastases. No secondary metastases occurred 8. They thus concluded that there was no evidence that metastases metastasized.
Overall, there is a paucity of inconclusive experimental evidence regarding whether metastases can metastasize.
Disseminated tumour cells and micrometastases
The most widely accepted theory regarding the development of metastases is that a proportion of cells from a primary cancer become disseminated tumour cells, which then form micrometastases, some of which then progress to become macrometastases. Disseminated tumour cells are distinct from micrometastases as they have not yet extravasated and become implanted in the end organ. Micrometastases are defined as a metastasis 2 mm or less in diameter 3. Only a small proportion of disseminated tumour cells become micrometastases and only a fraction of micrometastases progress to become macrometastases 4. Is there a selection process that determines which tumour cells metastasize, or can all cancer cells metastasize once they have invaded sufficiently?
In-vivo video microscopy studies confirmed the above and demonstrated the inefficient nature of the process. Eighty per cent of melanoma cells injected into mouse liver successfully extravasated, but only 1 in 40 of the extravasated cells formed a micrometastasis and only 1% of those formed macroscopic tumours. However, 36% of the injected cells were still present at day 13, but 95% of these were dormant 9. These results have been reproduced by others 10.
There also appears to be a relationship between the primary tumour and distant spread, the nature of which depends on the type of primary tumour. On occasion, removal of the primary tumour results in regression of spontaneous micrometastases, suggesting that the presence of the primary tumour is required for maintenance of micrometastases and that the micrometastases result from continued seeding, or other forms of support, from the primary tumour 11-13. However, Folkman and co-workers previously demonstrated explosive growth of previously dormant lung metastases following removal of a primary tumour in mice. This led to the discovery of angiostatin, an angiogenesis inhibitor produced by the primary tumour that inhibits the spread of metastases 14, 15.
Significance of disseminated tumour cells and micrometastases
Disseminated or isolated tumour cells found in bone marrow specimens from patients with malignancy are believed to be an indicator of poor prognosis. Approximately 30% of patients with breast cancer without evidence of lymphatic spread or distant metastases will have disseminated tumour cells in the bone marrow 16. Micrometastatic cells in the bone marrow are viable, capable of proliferation, resistant to immune attack, and insensitive to some chemotherapeutic agents 12. Thus, patients with carcinomas who have micrometastases may have a prognosis similar to that of patients with documented macrometastases 17 and patients with early breast cancer who have micrometastases in their bone marrow have an increased risk of disease recurrence 18.
Tumour dormancy explains the latent period between detection of a primary cancer and evidence of metastases. The tumour dormancy hypothesis implies that micrometastases become established but then do not grow for a given time period, depending on both tumour and host factors 19. Dormant cells are cells that are neither proliferating nor undergoing apoptosis at a given time. As they are quiescent, they are often resistant to therapies targeted at dividing cells such as chemotherapy 9. Both clinical and experimental studies have shown that tumour cells and micrometastases undergo a period of dormancy 3, 20. An alternative theory to tumour dormancy is that metastases develop due to continuous tumour growth from tumour seeding until clinical recurrence is documented; this has not withstood recent analysis 21.
Most cancers spread via the lymphatic system. It is well documented that lymphatic metastases correlate with a worse clinical outcome. In pathological staging of patients with carcinoma, emphasis is placed on apical nodes to determine further treatment and prognosis. But how do metastases spread through the lymphatic system? It is not clear whether metastases spread in a stepwise progression from node to node moving more apically or whether the lower nodes initially become congested with tumour so that subsequent metastases bypass them and move higher up the lymphatic chain. Lymph nodes are efficient filters, mainly due to adhesive interactions, which are normally involved in the T-cell immune response. Adhesive interactions provide a suitable environment for tumour cells to invade and proliferate 22. The lymphatic system is also amenable to the entry of metastasizing tumour cells 23. Sleeman argues that primary tumours are heterogeneous and only a small proportion of cells within the primary lesion are capable of forming distant metastases. These cells initially metastasize to the lymphatics and Sleeman states that they then re-metastasize, as these cells possess the necessary properties, and have already successfully managed to do so 23. However, this theory has not been tested experimentally.
Metastatic progression of a primary tumour is both complex and inefficient. A tumour consists of a heterogeneous population of cells with varying degrees of metastatic potential. The most likely mechanism of metastasis formation is a stepwise progression as follows: cells with metastatic potential disseminate from primary tumour and a small proportion of them will begin to spread. An even smaller sub-population will have the ability to form micrometastases, of which even fewer will progress to form macro-, and hence detectable, metastases. To confound our understanding, disseminated tumour cells and micrometastases may be held dormant for prolonged periods before an as yet unidentified trigger promotes their further development.
Evidence to date suggests that, as yet, there is no unifying mechanism by which tumours progress to metastases and that both the route of spread and the subsequent development to detectable metastases are dependent on both the type and the grade of the primary cancer. Few in vivo studies address these issues and, at present, published results serve only to confuse the area further. That leaves us with one of two scenarios: the mechanism of tumour cell dissemination and metastasis is individual to each type of cancer and molecular profiling of each patient will be needed to ensure that patients receive the best treatment regime. Alternatively, there is a generic mechanism for this process that we do not as yet understand, due largely to the lack of experimental approaches. It is not unprecedented for seemingly unrelated disparate pathways to be explained by a single mechanism once the underlying processes have been delineated and understood.
There are even fewer data to support or negate the suggestion that metastases might themselves metastasize. It is tempting to presume that once a tumour cell has acquired the ability to metastasize, it will retain that ability. We must, however, bear in mind that the cell population within a tumour is continually evolving, the process by which the more metastatic cells originally occurred. It is entirely plausible that these cells retain this characteristic. However, it is also plausible that these cells acquire further alteration which may cause them to lose their metastatic potential. This again may be dependent on both tumour type and even the length of dormancy prior to progression to macrometastasis. Without convincing experimental data, it is difficult to draw definitive conclusions.
From a therapeutic point of view, it is important to know if metastases metastasize. If this is a phenomenon specific to certain tumour types, then those metastases can be identified and treated aggressively to prevent further spread. If this process turns out to be a general mechanism of tumour progression, then treatment regimes need to be carefully monitored and reviewed in order to maximize efficacy. With the currently available therapeutic modalities, there is no proven survival benefit for aggressive treatment of metastases. However, this issue will become increasingly relevant as newer, more effective therapies are trialled.
Thanks are due to the Leeds General Infirmary Breast Cancer Research Action Group for their continued support.
- 1. Critical determinants of metastases. Semin Cancer Biol 2002; 12: 89–96.
- 2, , . Evidence for the clonal origin of spontaneous metastases. Science 1982; 217: 361–363.
- 3. Disseminated tumour cells versus micrometastasis: definitions and problems. Anticancer Res 1999; 19: 2771–2774.
- 4, , . Cancer spread and micrometastasis development: quantitative approaches for in vivo models. Bioessays 2002; 24: 885–893.
- 5, . Metastasis results from pre-existing variant cells within a malignant tumour. Science 1977; 197: 893–895.
- 6, , . Lymphatic dissemination of hepatic metastases. Cancer 1985; 55: 1490–1494.
- 7, . Metastasis of metastases. Am J Surg 1975; 130: 405–410.
- 8, , . Do metastases metastasize? Ann Surg 1971; 174: 161–166.
- 9, , , . Multistep nature of metastatic inefficiency. Dormancy of solitary cells after successful extravasation and limited survival of early micrometastases. Am J Pathol 1998; 153: 865–873.
- 10, , . Molecular biology of breast cancer metastasis. Clinical implications of experimental studies on metastatic inefficiency. Breast Cancer Res 2000; 2: 400–407.
- 11, , , . Overexpression of fibroblast growth factor 1 in MCF-7 breast cancer cells facilitates tumour cell dissemination but does not support the development of macrometastases in the lungs or lymph nodes. Cancer Res 1999; 59: 5023–5029.
- 12, , , , . Bone marrow micrometastases and gastrointestinal cancer detection and significance. Am J Gastroenterol 2000; 95: 1644–1651.
- 13, , . Micrometastases: marker of metastatic potential or evidence of residual disease? Gut 1997; 40: 512–515.
- 14, , , et al. Angiostatin: a novel angiogenesis inhibitor that mediates the suppression of metastases by a Lewis lung carcinoma. Cell 1994; 79: 315–328.
- 15, , , , , . Radiation therapy to a primary tumour accelerates metastatic growth in mice. Cancer Res 2001; 61: 2207–2211.
- 16, , , et al. Differential expression of proliferation associated molecules in individual micrometastatic carcinoma cells. J Natl Cancer Inst 1993; 85: 1419–1424.
- 17, . The significance of breast cancer lymph node micrometastases. Surg Oncol Clin N Am 1999; 8: 481–496.
- 18, , , , . Minimal residual disease in bone marrow and peripheral blood of patients with metastatic breast cancer. Recent Results Cancer Res 2003; 162: 135–140.
- 19. Tumour dormancy: findings and hypotheses from clinical research on breast cancer. Semin Cancer Biol 2001; 11: 297–305.
- 20, , . Dormancy of micrometastases: balanced proliferation and apoptosis in the presence of angiogenesis suppression. Nature Med 1995; 1: 149–153.
- 21, , . Proposal for a new model of breast cancer metastatic development. Ann Oncol 1997; 8: 1075–1080.
- 22, . What can we learn from the phenomenon of preferential lymph node metastases in carcinoma? J Surg Oncol 1999; 70: 199–204.
- 23. The lymph node as a bridgehead in the metastatic development of tumours. Recent Results Cancer Res 2000; 157: 55–81.