New insights into how cancers recur have been revealed by scientists in the US.
Tests suggest that up to 30% of women with early-stage breast cancer have "silent" metastases (cancer spread) in the bone marrow at the time of their diagnosis.
This means that recurrence of the disease, even years after apparently curative therapy, remains a risk for many cancer patients.
Now a team from Duke University have some good news, because researcher Trevor Price and his colleagues have uncovered how rogue cancer cells become dormant at remote sites around the body.
They have also found how to block the process and even how to flush hibernating cancer cells back out into the blood stream where they might become susceptible to anti-tumour drugs.
Writing in Science Translational Medicine, the team have been studying breast cancer, which has a known predilection for spreading to bone.
Working with experimental mice, the researchers injected into the bloodstream cancer cells that had been labelled with a coloured probe.
They then used a microscope capable of seeing through the intact skulls of the mice to watch as the cancer cells floated through - and then lodged inside - the blood vessels supplying the bone itself.
Critically, the tumour cells regularly settled at specific regions in the bone marrow known as the perisinusoidal space.
These areas, the team have found, contain large amounts of a sticky molecule called E-selectin, to which cancer cells naturally bind.
Another molecule, called stromal cell derived factor 1 or SDF-1, a signal for which cancer cells express a receptor on their surfaces, is also enriched in this region of the bone.
Reasoning that these molecules might be leading their breast cancer cells to preferentially spread to these sites, the Duke University team set about using drugs to block these signals to explore whether this could alter the ability of the cancer cells to invade.
Administered into the bloodstream, the agent GMI-1271, which blocks E-selectin, successfully abolished the spread of cancer cells into the bones of the mice.
Another drug, called AMD3100, which blocks the SDF-1 signal, had no effect on the spread though.
But when the researchers administered AMD3100 to mice that already had tumours that had spread into their bones, large numbers of cancer cells were flushed out into the bloodstream.
This, the team speculate, suggests that E-selectin probably acts as a signal to attract cancer cells out of the blood and into bone in the first place.
The SDF-1 signal, meanwhile, then applies a kind of cellular handbrake, holding the cells in-situ.
The discovery is important because it sheds light on why many cancers naturally invade bones and how they probably lurk in these sites, undetected and in a dormant state, for years before triggering cancer recurrence.
The researchers think that the cancer cells are mimicking bone marrow stem cells that often naturally move into these bony "niches" where they receive pro-dormancy signals that temporarily power them down until they are needed.
The ability to reawaken and flush out cancer cells using the AMD3100 agent, while it sounds like a bad idea, might ultimately hold the key to ridding the body of malignant cells by forcing them out of the protective environment of the bone marrow and exposing them in the bloodstream to anti-cancer medications.