¿ìè¶ÌÊÓƵ

CANCER is not a single disease, but a group of diseases with one thing in common: somewhere in the body there is a rapidly growing mass or lump of disorganised tissue called a tumour. This has been recognised since ancient civilisations first practised medicine, and for most of this time the tumour has dominated the attention of doctors treating the patient. More recently, it has become clear that cancers also have general, or systemic, effects on the whole person that are often even more debilitating than the original tumour. The main goal of the research in many laboratories in this field is to understand these systemic effects and provide information which may open new pathways for controlling cancer.

Cancer is often difficult to treat and frequently fatal because the proliferating cells of which it is composed tend to invade and destroy tissue near the site of the tumour and spread via the blood or the lymph to other organs, in which they form secondary tumours. The widely scattered tumour cells may also lie dormant for many years before producing fresh crops of tumours. These are major problems for those trying to treat people with cancer because the number of tumours increases (often dramatically), so that the removal of the original tumour cannot cure the disease. Moreover, attempts to destroy these secondary tumours often cause widespread damage to normal tissue in many organs.

The task is rather like fighting a guerrilla army. The defences within the body find it hard to distinguish normal cells from tumour cells, especially as these are dispersed as small units in unknown numbers over a large terrain, which is difficult to police effectively. The defences that the body mobilises are usually ineffective, because as soon as one small unit has been liquidated, another can emerge elsewhere.

Most people who are dying from cancer have secondary tumours in many sites. Clinical studies show that these tumours, or the ‘seeds’ that started them, had already spread by the time people first took their symptoms to a doctor. Therefore, the successful treatment of cancer must include eradicating already established multiple secondary tumours, or arresting their growth, in addition to preventing further dissemination. Currently, the combined resources of surgery, radiotherapy, drug therapy, hormonal therapy and manipulation of the immune system are successful in only a small proportion of people with advanced cancer. It is therefore necessary to know more about how tumours spread.

The spread of tumours in the body was first recognised in 1829 by Jean-Claude Recamier, a doctor who had served briefly in the Napoleonic Wars, before working at the Hotel Dieu in Paris. He described a patient with breast cancer who also had a tumour in the brain and concluded that the brain tumour was a distant colony of the main one in the breast. He called the process ‘metastasis’, a hybrid term of Greek origin meaning to transfer from one place to another. So the secondary tumours are called metastases. Each secondary tumour can shed more cancer cells that may go on to seed new sites. Consequently, the threat to the body increases exponentially rather than arithmetically.

Tumours are not all capable of metastasis. They have been divided into two main groups described as benign and malignant. Those that are benign remain well defined and stay where they originated. Other tumours invade the surrounding tissues but rarely spread to other organs. This can make them hard to treat, because it is difficult to remove their finger-like extensions completely, and so they often recur. These locally invasive tumours, together with the many different types that spread around the body, are malignant. The word cancer is a general term applying more to malignant tumours, or ones known to be capable of becoming malignant, than to benign ones.

In the years that followed Recamier’s insight, scientists debated about exactly what was being transferred from the original tumour to the secondary ones. Two theories were popular: the possibility that portions of the tumour broke away and seeded elsewhere, or that the primary tumour released chemical or infective agents that caused new tumours to form elsewhere. Eventually, the development of good microscopes settled the issue. Now, no one seriously doubts that cells from the original tumour are carried along with the blood or the lymph to new sites, where they become trapped and either die, lie dormant, or grow to form secondary tumours.

Because metastasis is a step-by-step process in a living organism, involving the transfer of tumour cells from one site to others through channels such as blood vessels, it cannot satisfactorily be modelled outside a living body. The malignant tumour cells must be able to perform every step in the sequence in order to establish a metastatic colony. The first event is the escape of the cells from the primary tumour. The cells must then be able to enter a transport system such as the blood or the lymph. Once there, they must survive the challenges in the new environment, including high-speed collisions with other cells and vessel walls within the bloodstream and elude destruction by the immune system (see Figure 1).FIG-mg17263901.GIF

A suitable analogy is to imagine being carried along with numerous other passengers, down a tunnel in the London underground by a tidal wave, with the added refinement of the tunnel ending in several branches of small diameter. Clearly there is considerable risk of injury from getting stuck in narrow channels and from colliding with the walls and with other passengers, especially in zones of high turbulence near branches. Those cells robust enough to survive the experience subsequently have the opportunity to insinuate their way through the vessel wall into the surrounding tissue. Alternatively, they may jam within the cavity, or lumen of the small vessels, start to grow, and eventually burst through the wall. Some pass through wider channels in the first organ to repeat the experience in downstream organs.

Finally, tumour cells entering the tissues of a new organ must be able to multiply in that environment and induce the growth of local blood vessels. If they cannot do this, the size of the colony will be limited to a few millimetres, because the only other way to transport nutrients and waste materials to and from cells, namely diffusion, is too inefficient over larger distances to support their survival and growth.

If all of these conditions are satisfied, secondary tumours can grow very large, sometimes bigger than the primary tumour. But the process is not always successful. Many cells are eliminated at each step in the sequence, either because they are unable to proceed to the next, or because of mechanical damage during transport, or because the host’s defences destroy them, or because they cannot adapt to the new conditions around them.

¿ìè¶ÌÊÓƵs once believed that all cancer cells in malignant tumours could form metastases and that once they are shed into the blood, or lymph, a secondary tumour would inevitably form downstream. But modern work has shown that this is not so: individual tumours and even different cells within each tumour have different capacities to form metastases.

It follows that the defect that initiates and drives the process lies primarily within the spreading tumour cells themselves rather than in the normal tissues. The abnormality is not transient and can be passed from cells to their descendants. Some cell populations continue to be metastatic over many cell generations. This implicates the genes of the tumour cells as the primary cause of the problem and in turn raises the questions of which genes, and of how they are regulated.

Genes drive the spread of tumours

One clue is that any property possessed by a cancer cell can also be found in normal cells somewhere in the body. This suggests that the genes that drive metastasis are also normally present in non-cancerous cells. So a cancer probably spreads because genes that are normally ‘silent’ in most cells become activated inappropriately, although they may be essential for some process in some other cell, or perhaps even at some other time in life. Genes that may be involved include those required for certain mobile blood cells (for example, lymphocytes and granulocytes) to move through the tissues of the body.

One way to identify and study the genes which may be responsible for metastasis is to try to transfer this behaviour to cells of benign tumours by transferring DNA from metastatic cells. In our laboratory, we culture the non-metastatic cells in a solution containing DNA from the metastatic cells, under conditions that help cells to take up and use this DNA; this process is known as transfection. The donor DNA is obtained from metastatic human cancer cells and transfected into non-metastatic mouse cells so that the recovery and purification and cloning of any donor DNA in the recipient cells can be facilitated by searching for human specific sequences.

The DNA is first mixed with a marker gene which can make cells resistant to a specific toxic drug (such as the antibiotic neomycin). This makes it possible to select those cells that take in the DNA and express the genes concerned: culturing the cells in a medium containing the relevant toxic drugs kills off any cells that have not incorporated the DNA supplied.

Our results indicate that transferring ‘metastatic’ DNA into non-metastatic cells can make them able to spread to other organs. This does not happen in every transfection experiment, but the observation provides more evidence that metastatic behaviour, once initiated, can be inherited by cells from their predecessors. So it should be possible to identify the specific genes responsible and study why they function abnormally. Identifying and cloning these genes may also be useful in cancer treatment. The use of cloned genes as probes can reveal whether individual cells contain corresponding messenger RNA. Eventually, such methods may show whether some cells in a given person’s tumour are transcribing genes relevant to metastasis. This could help to tailor treatment to the likely behaviour in a specific cancer in a patient rather than to that type of cancer in general.

The evidence summarised above supports the conclusion that metastasis is initiated by the release of a drive to spread and colonise. However, this working interpretation does not, on its own, fully explain other important aspects of the metastatic process. Doctors have known for a long time that different types of cancer tend to form secondary tumours in certain organs and not in others. For instance, as early as 1889, Stephen Paget noticed that cancer of the breast tends to spread to the bones, liver and lungs, but rarely involves other organs, such as the intestines, thyroid gland or muscles. Subsequent studies have confirmed this and described different patterns of spread favoured by cancers of other sites.

Paget likened tumour cells that could spread to other sites to seeds blown in the wind: they could grow only if they landed in congenial soil. Many recent observations support this interpretation, but some investigators place the emphasis on the ‘wind’, and argue that the sites of growth of the ‘seeds’ depends entirely on which way the ‘wind’ is blowing. According to this school of thought, a tumour cell that enters a blood or lymph vessel is carried passively in the direction of the flow and will grow to form a secondary tumour in any organ where it becomes trapped by the network of fine capillaries.

Tumour cells entering veins are carried via successively bigger veins back to the heart from where they are pumped into the pulmonary artery. The first capillary bed they encounter is that in the lungs. Many tumours, including cancers of the mammary glands in mice, tend to spread to the lungs; this suggests that the anatomy of the blood circulation affects where secondary tumours are formed.

Other studies, however, indicate that this is not the only mechanism controlling where metastases develop. For instance, lung cancers in men and women tend to spread to the adrenal glands, while women with breast cancers often develop secondary tumours in their bones. The anatomy of the blood circulation alone cannot explain these tendencies because to reach these sites, the cancer cells must have passed through the capillaries of the lungs to get into the arterial blood and have been distributed rapidly to all organs, yet metastases are often not present elsewhere. So it seems that the ‘seed’ also needs the right ‘soil’ if it is to grow.

In support of this view, several studies have shown that tumour cells liberated into the blood arrive rapidly in all organs in the body. In our laboratory we recorded the distribution patterns of dispersed mammary tumour cells, labelled, before injection into the blood stream, with a radioactive compound or with a fluorescent dye. These experiments showed that tumour cells released into the circulation, either via the veins or via the arteries, reached every organ within 15 minutes, and that the dose of cells delivered to each organ is not the main determinant of the pattern of metastases.

Despite the fact that cells were being distributed everywhere, secondary tumours formed most often and most extensively in the lungs. This still happened even when we injected tumour cells into the main artery leaving the heart (the aorta). In such injections, the cells bypass the lungs so that the first capillary ‘sieve’ is that of organs such as the brain, liver, kidneys and muscle. Contrary to the predictions of the ‘wind’ hypothesis, the tumour cells did not colonise all these organs indiscriminately, and the lungs were still the commonest site for metastases even though the cells making the colonies must have passed through, and thus had the chance to colonise, other organs first.

Our conclusion, like Paget’s, was that there must be some characteristic of the ‘soil’ of the lungs that encouraged the growth of tumour cells. Conversely, it seemed that the tumour cells could not thrive in other sites and this interested us because it indicated that cancer cells are not totally self-sufficient and able to grow anywhere. To investigate this further, we devised the following experiments. First, we incubated fragments of various organs in tissue culture fluid for 24 hours. Once the fluid medium had been ‘conditioned’ by the organ in this way, we used a centrifuge to remove all cells and debris. Then we cultured suspensions of mouse mammary tumour cells for 24 hours in these various ‘organ-conditioned’ media.

At the end of the experiment, the cultures enriched with lung-conditioned medium always contained far more tumour cells than those supplemented with other organ-specific media. These findings correlate well with the tendency of these tumours to form secondary tumours preferentially in the lungs. Ovary and kidney-conditioned media varied in their effect: cells from some mammary tumours grew in them, but the cells of others did not. This fits well with our observations that the cells from a few tumours grew in the ovary or kidneys after inoculation via the aorta, while the cells from many others did not. Media conditioned by liver or thyroid killed all tumour cells, and mouse mammary tumours never spread to these organs. Our experiments therefore suggest that the organ fragments released some soluble factors that affected the ability of tumour cells to survive in culture, and might also do so in the body. Such factors could be one of the mechanisms by which the ‘soil’ of various organs affects whether cancer cells succeed or fail in forming secondary colonies.

The above observations are based on the study, in many laboratories, of cancer in animals. In recent years we have also had the opportunity to confirm some of these conclusions by studying the spread of cancer in people. Certain types of cancer cause large amounts of fluid to accumulate in the abdomen, making it distended and painful. To relieve this distressing condition, known as malignant ascites, doctors traditionally drained this fluid via a needle inserted through the abdominal wall. The fluid contains proteins and ions such as sodium, potassium and chloride, as well as living malignant cells, so when it is repeatedly drained, people suffer severe metabolic imbalance which makes them feel very unwell. An alternative form of treatment which conserves these important metabolites was therefore proposed: the idea is to transfer the fluid back into the bloodstream, via a plastic tube known as a ‘shunt’. At first, the use of this technique was controversial because some doctors feared that the procedure would spread the malignancy. The treatment is now generally accepted, however, because there is no evidence of deterioration in the condition of the patients receiving it.

We studied a group of 23 terminally ill patients, with various cancers, who had shunts inserted to relieve their symptoms and obtained surprising results which prompted us to reconsider some of our assumptions about tumour metastasis. Although each person’s shunt delivered about a billion living cancer cells into the blood every week for months or years, only about half of the group had any blood-borne secondary tumours when they were examined during subsequent autopsies. The rest had no evidence of deposits in any organ, nor any indication that metastases had been formed and then destroyed by immunological or other defences. Even among those who did have metastasis, the deposits were small, caused no symptoms and did not contribute to death.

The most remarkable findings were seen in people who already had blood-borne metastasis in some organs before having a shunt installed. In these people, the number of deposits in affected organs increased, but other organs remained clear, providing compelling evidence that Paget’s conclusions were correct. The main implication is that tumour cells that form metastases are not invincible, and cannot grow just anywhere. This highlights an important aspect of metastasis: namely that scattered tumour cells must be able to multiply in their new site if they are to form secondary tumours. Cells that fail to multiply, or are prevented from doing so, no longer present a clinical problem, because they cannot produce a new seeding tumour capable of producing other tumours. This could perhaps be considered as a starting point for new approaches to treatment.

Recent studies on the causes of tumour-cell proliferation have indicated that this may result from loss of certain inhibitory genes which control or stop the normal tendency of cells to multiply. This work has its origins in the pioneering studies of Henry Harris at the University of Oxford, who first showed, many years ago, that the fusion of a cancer cell with a normal one resulted in a hybrid which was not cancerous. Harris reasoned that the normal cell contributed information, absent in its cancerous partner, which brought cell proliferation under control. This finding has since been confirmed and extended by Eric Stanbridge at the University of California at Irvine and others. Stanbridge recently described results indicating that replacement of a single normal chromosome 11 into malignant kidney tumour cells which have only one defective copy of this chromosome, suppresses the formation of tumours. Moreover, data recently published by Wen-Hwa Lee at the University of California at San Diego, and some findings from our own laboratory suggest that sometimes it is possible to stop the growth of tumours by transfecting them with single genes or with small bits of DNA from normal cells.

Lee’s research shows that putting the recently isolated RB1 gene into cells derived from cancers of the eye, breast and prostate can prevent them from developing tumours. In our laboratory we have been studying whether it might be possible, by transfection of DNA from normal (non-cancerous) human cells into mouse tumour cells, to identify and isolate genes which could control malignant behaviour in cancer cells whatever their origin. After many attempts we succeeded in suppressing the tumour-forming capability of a cell line that formerly produced tumours in all of the animals in about 20 days. The transfected line (which has been confirmed to contain human DNA) takes, on average, 130 days to produce tumours and often does not do so at all. Another tumour cell line, obtained from one of the tumours which eventually formed, does not contain any human DNA and again produces tumours about 20 days after inoculation. This suggests that tumour-forming capability has re-emerged after elimination of some regulatory or suppressor sequences in the human DNA.

The interest and significance of isolating the gene or genes that could inhibit the growth of tumours are obvious, but for the research to have practical clinical application, the problem of how to get the ‘growth control’ gene into the metastatic cells in a person suffering from cancer would have to be solved. Doctors would need to find ways of delivering the genes specifically to the malignant cells scattered throughout the body and of ensuring that they functioned properly once inserted in the DNA of the target cells. Specially constructed viruses might be suitable for ferrying the genes into the tumour cells because many viruses are known to have affinities for particular tissues (for example, the hepatitis B virus infects liver cells). Targeting based on the tissue of origin of the cells would be worth investigation.

Researchers at the University of Wisconsin and at the University of Edinburgh have used some similar techniques to correct specific gene defects in cells grown in the laboratory. They transferred the gene for the enzyme hypoxanthine guanine phosphoribosyl transferase into cells deficient in this enzyme, using a virus to carry the gene into the cell. The method they used ensured that the gene integrated itself properly into the host DNA and correct the deficiency previously inherent in the cell line. This is described as targeted gene replacement. So the idea of infecting a patient with a specifically tailored virus, which would bind to receptors specific to the tumour cell, enter its nucleus and control proliferation, does not seem too far-fetched.

Advances in the treatment of some special types of disseminated cancer have resulted in cures or in substantial improvements in survival. Unfortunately, the same cannot be said for many other types of cancer. Current treatments, involving radiation, drugs and surgery, although helpful in controlling the local growth or the recurrence of a cancer, can themselves cause damage to surrounding tissues and unpleasant side effects without significantly lengthening the individual’s survival. Doctors and scientists working to improve the treatment of people with metastatic cancer need to be open to unconventional possibilities. Recent advances in tumour genetics may offer promising new approaches to this old and difficult problem.