快猫短视频

IT DID not make sense. Back in the 1930s, biologist John Bittner was studying a strain of mice in which breast cancer ran in the family. This suggested the cause was genetic 鈥� and yet if the pups were taken away from their mothers at birth and fostered by other females, they did not develop the disease.

The answer, Bittner soon realised, was that something in the milk of the cancer-prone mice must be causing the tumours. Over the following years it was shown to be a virus, called the mouse mammary tumour virus or MMTV. Its discovery raised an obvious question: might a similar virus cause breast cancer in people, too?

The idea was plausible, yet no one found any convincing evidence to support it. There was no hint of any link between breastfeeding and cancer, for instance: at the time, fewer and fewer women were breastfeeding yet breast cancer rates were rising. By the 1970s, the idea had fallen out of favour.

Many decades later, though, we still have very little understanding of what causes 90 per cent of breast cancers 鈥� all the ones, that is, that are sporadic rather than due to inherited mutations. Recently, a few researchers have revived the idea that a virus similar to MMTV might be to blame for at least some sporadic cases. It would be good news if a virus is to blame: it is possible to prevent breast tumours in mice by vaccinating them against MMTV, suggesting a human vaccine could save many lives.

Supporters of the idea can point to a growing number of studies over the past decade that implicate a virus similar to MMTV in up to a third of breast tumours. But there are many contradictory results, too. So are we getting any closer to resolving this long-standing mystery?

Despite the waning of interest in the human connection in the 1970s, researchers continued to study MMTV. It has turned out to be a sophisticated retrovirus that exploits the immune system of its host. In mice, MMTV initially infects local immune cells in the gut of newborns. It produces a 鈥渟uperantigen鈥� that encourages the proliferation of other immune cells: rather than being destroyed by these cells, though, the virus infects them and is carried through the bloodstream to the developing mammary glands.

Being a retrovirus, MMTV inserts its genes into the genome of its host cell. That means it can lie dormant inside mammary cells for many years. When female mice enter puberty or become pregnant, MMTV springs into action again, infecting cell after cell. Every time it infects a cell, it inserts its five genes into the cell鈥檚 genome, along with a 鈥渟witch鈥� or promoter for activating these genes, which can also turn on nearby mouse genes. The more cells MMTV infects, the more likely it is for its genes to land near genes crucial for controlling the growth of breast cells, such as Wnt-1, which can kick-start the chain of events that leads to a tumour.

While most researchers focused on working out what MMTV does to mice, a few wondered what MMTV does to researchers. In the 1980s, Arnold Dion of the Garden State Cancer Center in New Jersey found that some lab workers showed signs of an immune response to MMTV, suggesting they had been infected by the virus. One women developed breast cancer not long after testing positive.

What really revived interest in the human connection, however, was a 1995 study by Beatriz Pogo at Mount Sinai School of Medicine, New York, and colleagues. The team used PCR, a technique that can be used to detect tiny amounts of specific DNA sequences, to look for sequences resembling part of an MMTV gene called env in tissue samples. The researchers found matching sequences in 38 per cent of samples from breast tumours, but not in normal breast tissue or in other tumours.

Over the following years, several other groups reported similar findings. In 2000, Polly Etkind at New York Medical College not only found MMTV sequences in human breast cancer tissue, but also found that some samples harboured more than one strain of MMTV, suggesting multiple recent infections. In 2003, Caroline Ford, now at Lund University in Malm枚, Sweden, detected MMTV sequences in 42 per cent of breast tumour samples from Australian women, but in just 1.8 per cent (2 out of 111) of samples of normal breast tissue.

Ford鈥檚 team also found MMTV sequences in six of nine male breast tumours examined. And when Ford investigated the two samples of normal breast tissue that tested positive, she found one came from a woman who had had a tumour in the other breast, while the other came from a woman who developed a tumour after the sample was taken.

Yet if MMTV or a closely related virus really can cause breast cancer, how are people being infected? It seems clear that women do not pass on the virus in breast milk. A 1998 study of more than 8000 women by Linda Titus-Ernstoff at Dartmouth Medical School in Hanover, New Hampshire, found no evidence that being breastfed increases the risk of developing breast cancer, even when the mother herself later develops breast cancer.

Source of infection

If it鈥檚 not from milk, could mice be infecting people directly? In theory, it鈥檚 possible that MMTV could be passed from mice to humans by insects such as fleas or mosquitoes, or even directly via mouse droppings in food. Most countries鈥� regulations tolerate a certain level of rodent faeces in grain.

To explore this connection, a 1999 study by Thomas Stewart of the University of Ottawa in Ontario, Canada, compared the distributions of mouse species with human breast cancer rates. There are huge differences in the incidence of breast cancer around the world: some western countries have rates five times as high as those in some Far Eastern countries. This cannot be genetic, as the children of immigrants to the US suffer the same high rate of breast cancer as everyone else, and yet so far no especially strong link has been found between breast cancer and any single environmental factor such as diet.

Stewart found that around the world, the incidence of breast cancer tends to be higher in regions where a particular subspecies of house mouse, Mus musculus domesticus, is found (see Map). This subspecies is native to western Europe but has spread with colonists to many other areas, including North America, Australia and New Zealand, and there is some evidence that it is more prone to MMTV infections than other mouse species.

The results of the studies looking for MMTV sequences fit in neatly with the idea that the virus can explain the regional differences in breast cancer incidence. In western countries, around 10 per cent of women develop breast cancer, and MMTV sequences have been found in about a third of breast tumours in the US, Australia and Italy.

鈥淏reast cancer incidence is five times as high in some countries, and it鈥檚 not down to genes鈥�

In Vietnam, just 1 per cent of women develop breast cancer, and when Ford looked at samples from Vietnamese women in her 2003 study, she found MMTV sequences in just 0.8 per cent of tumour samples and none in samples of normal breast tissue.

All these results seem to paint a plausible picture but most researchers have yet to be convinced. Robert Garry, a virologist at Tulane University in New Orleans, Louisiana, has found MMTV-like sequences not only in breast tumours, but also in 10 per cent of healthy men and women. To Garry, this suggests that the retrovirus has recently become integrated into the genomes of some people.

This is certainly plausible: there are several lines of breast-cancer-prone lab mice in which the retrovirus has become permanently integrated into the genome and is now passed down from parent to offspring. 鈥淲e haven鈥檛 eliminated the possibility of horizontal transmission [in humans], but I think it鈥檚 more likely to be an endogenous virus,鈥� says Garry. If he is right, some people will be stuck with MMTV.

Several other groups, using slightly different methods, have failed to find any MMTV sequences in breast tumour samples. In 2004, for instance, a team led by John Cason of the Guy鈥檚, King鈥檚 and St Thomas鈥� School of Medicine, UK, failed to find the env gene sequence in samples taken from breast cancer patients in London.

How could this be? One suggestion is that what Pogo and others are finding is the legacy of ancient retroviral infections rather than a recent infection. In our distant ancestors, some retroviruses managed to infect the cells that produce sperm and eggs. Every cell of any resulting offspring carried the retroviral genes. These genes soon became mutated and harmless, but their genetic remains, called human endogenous retroviruses (HERVs), make up 8 per cent of the human genome. Some HERV sequences are very similar to those of MMTV.

The most recent studies by Ford and others, however, have been based on looking for MMTV sequences in tumours that do not resemble any known HERVs. What鈥檚 more, these sequences have not been found in normal tissue from the same individuals, as would be expected if they were HERVs. 鈥淭his suggests a recent infection and integration of viral DNA,鈥� Ford says.

Other researchers are not convinced. No one has yet isolated the entire MMTV genome in one piece from a human cancer cell and shown that it yields viruses able to infect cells, points out Robert Weinberg of the Whitehead Institute for Biomedical Research in Cambridge, Massachusetts.

鈥淕iven the molecular techniques that have been available for far more than a decade, the definitive proofs of the presence of MMTV-like sequences, including complete proviruses, should have been forthcoming many years ago,鈥� he says. 鈥淭hat has not happened.鈥�

Indeed, there is a question mark over whether MMTV can infect human cells at all. The mouse protein to which MMTV binds when it infects cells is quite different from the human equivalent, suggesting MMTV should not be capable of entering human cells.

Infecting humans

Last year, however, a team led by Stanislav Indik at the University of Veterinary Medicine in Vienna, Austria, concluded that MMTV can infect several human cell types nearly as efficiently as mouse ones. The team carried out a series of steps designed to rule out alternative explanations for their findings.

Another criticism of the MMTV theory is that the risk of breast cancer does not appear to be raised by immunosuppression. 鈥淚f you look at other virally caused cancers, such as cervical cancer, they occur at an increased frequency in people with poorly functioning immune systems, who can鈥檛 kill off the virus,鈥� says Cason. This is what happens with Kaposi鈥檚 sarcoma, which is often seen in people with AIDS, as well as with liver cancer due to hepatitis B and C. 鈥淏reast cancer shows no such association with poor immunity,鈥� says Cason. 鈥淚 think that鈥檚 one of the strongest points against the hypothesis.鈥�

Defenders of the hypothesis, however, point out that, at least in mice, the initial spread of MMTV depends on the proliferation of immune cells induced by MMTV鈥檚 superantigen, and that immunosuppression during this phase might prevent the virus reaching the mammary glands.

Then there鈥檚 the issue of pregnancy. In mice, MMTV replication greatly increases during pregnancy. The more often female mice infected by the virus fall pregnant, the more likely they are to develop tumours. Yet as critics like Cason point out, women who get pregnant have a lower risk of breast cancer. If the human version of MMTV is as similar to the mouse strain as most studies suggest, it is hard to explain such differences.

None of the supporters of the MMTV hypothesis are claiming to have provided conclusive proof. 鈥淥ur results, and those of others, only indicate an association and point to some possible implications,鈥� says Pogo. 鈥淲e have not claimed causation in any of our papers.鈥� Ford agrees. 鈥淭his workis preliminary, and the topic as a whole is still really in its infancy,鈥� she says.

What would help convince some of the critics is the isolation of complete MMTV particles from tumour samples. That won鈥檛 be easy. 鈥淪earching for viruses in tissues has always been difficult and still is,鈥� says Garry.

Nevertheless, Pogo claims to have taken a step towards that elusive goal. Using an electron microscope, she says she has seen viral particles in cancer cells, as well as the budding of viral particles from the surface of cells. Both claims, however, have yet to go through the rigour of peer-reviewed publication. Given the long history of controversy over MMTV, you can bet her findings will not be the end of the story.

Update: suggests that all the epidemiological studies that failed to find a link between breast cancer and breast-milk-borne viruses could be fundamentally flawed.