FOR 80 years we have misunderstood the feeding habits of cancer. It鈥檚 a controversial suggestion that, if correct, could open up a host of alternative ways to fight the killer disease, and may even mean that in some circumstances chemotherapy drugs promote tumour growth rather than inhibit it.
In the 1930s, Otto Warburg suggested that cancer cells produce the bulk of their energy by breaking down glucose in the absence of oxygen, a process called glycolysis. The Warburg effect, as it is called, is now widely accepted in cancer research. It is also incorrect, according to Michael Lisanti at the Kimmel Cancer Center in Philadelphia, Pennsylvania.
Lisanti thinks that when a cell turns cancerous it begins to spew out hydrogen peroxide. The free radicals this generates cause oxidative damage that prompts support cells in the connective tissue around the cancer cells to begin digesting themselves (see diagram). Once these support cells, called fibroblasts, have consumed the mitochondria that normally provide their energy, they switch to glycolysis. The cancer cells then feed off the nutrients glycolysis generates.
Advertisement
鈥淐ancer sends a signal that prompts the cells around the tumour to begin digesting themselves鈥�
鈥淚t鈥檚 the Warburg effect, but in the wrong place,鈥� says Lisanti, who presented the idea earlier this month at the meeting in Cambridge, UK. 鈥淐ancer cells can feed off normal cells as a parasite.鈥� In fact, he says, cells infected with malaria behave in much the same way. 鈥淭he malaria parasite enters cells, induces oxidative stress, and gets free food鈥� by feeding off the structures inside host cells that self-digest as a result of the stress.
鈥淭he importance of the micro-environment is something that has been gaining recognition over the last few years,鈥� says of the Paterson Institute in Manchester, and chief scientist for Cancer Research UK. 鈥淭his adds a very important and exciting twist, where the communication between the cancer cell and the fibroblast fuels the development of the tumour.鈥�
This form of 鈥渕etabolic coupling鈥� also mirrors the way in which the epithelial cells that make up the skin and the surface of the body鈥檚 organs produce hydrogen peroxide during wound healing. In doing so they rally immune cells to repair the damage 鈥� but in cancer the signal is never turned off. 鈥淐ancer is a wound that doesn鈥檛 heal, because it keeps on producing hydrogen peroxide,鈥� says Lisanti.
He has experimental data to support his radical idea. When his team cultured breast cancer cells alongside fibroblasts for five days, they spotted the cancer cells releasing hydrogen peroxide on day two. By day five, most free radicals generated by the hydrogen peroxide were found inside the fibroblasts (Cell Cycle, ). The team also found a reduction in mitochondrial activity in fibroblasts, consistent with the cells self-destructing. There was also an increase in glucose uptake by the fibroblasts 鈥� a sign of glycolysis (Cell Cycle, ).
Lisanti is now gathering evidence to find out whether his ideas can be applied to many cancers or just a few. He has discovered a 鈥渕arker鈥� to identify patients in whom the metabolic coupling is occurring: as the fibroblasts are destroyed they stop producing a protein called caveolin-1. Lisanti has recorded a drop in caveolin-1 levels in 40 to 50 per cent of patients with breast cancer, and loss of the protein correlates with early tumour recurrence, metastasis, and resistance to the drug, tamoxifen (Breast Cancer Research, ). He also has evidence for caveolin-1 loss in prostate cancer. Those results suggest that new cancer therapies based around Lisanti鈥檚 ideas might be possible (see 鈥淭he cells that die so cancer can live鈥�).
Lisanti believes the reason Warburg got it wrong is because he looked at cancer cells in isolation, rather than in co-culture with fibroblasts.
鈥淭he provocative use of the term 鈥榬everse Warburg鈥� is certainly catchy,鈥� says of Johns Hopkins University in Baltimore, Maryland 鈥� but it ignores some important observations. For example, many previous studies have found increased glycolysis in cancer cells.
Lisanti鈥檚 model also runs into problems when taking the long view. 鈥淚f these [fibroblasts] are sacrificing themselves so that the cancer can eat, sooner or later they are going to be completely depleted. And that doesn鈥檛 happen,鈥� says of Barts Cancer Institute in London, UK. It is possible that tumours recruit stem cells from the bone marrow to replace the fibroblasts, but Hart says more evidence is needed to confirm this.
However, if Lisanti is correct, his ideas could also explain why people become more susceptible to cancer as they age. More than 100 years ago, Steven Paget proposed that cancer cells are seeds that need the correct micro-environment in which to grow. 鈥淲hat we鈥檙e now saying is that the hydrogen peroxide is the fertiliser,鈥� says Lisanti. 鈥淎s you age, your body produces more hydrogen peroxide and free radicals and becomes a fertile ground for cancer.鈥�
The cells that die so cancer can live
For decades, cancer therapies have focused on destroying cancer cells and ignored the healthy cells tumours also contain. The discovery that cancer cells form a parasitic relationship with the 鈥渘est鈥� of fibroblasts or support tissue that surrounds the tumour may therefore open up other opportunities for treating the disease.
鈥淪o far, all mainstream cancer therapies are aimed at [removing] these transformed cells,鈥� says Ian Hart of Barts Cancer Institute in London, UK. 鈥淩ather than killing every last tumour cell, let鈥檚 modify the [fibroblasts].鈥�
In his model, Michael Lisanti at the Kimmel Center in Philadelphia, Pennsylvania, proposes that cancer cells use hydrogen peroxide to strike up their metabolic relationship with the fibroblasts. The chemical generates free radicals in the fibroblasts, kick-starting a self-digestion process which frees up nutrients to fuel cancer growth. His team found that treating cancer cells with catalase, an enzyme that destroys hydrogen peroxide, triggered a five0fold increase in cancer cell death, possibly by cutting off the cells鈥� fuel supply.
This raises the prospect of treating cancer with antioxidants, which mop up free radicals. However, although some studies hint that antioxidants may be beneficial, particularly for cancer prevention, the results have often been disappointing, says Hart.
Killer free radicals
Lisanti thinks that鈥檚 because most chemotherapies work by generating lethal doses of free radicals to kill the cancer cells, which would cancel out the effects of any antioxidant treatments. He believes we need trials of antioxidants alone, rather than in combination with chemotherapy.
If he is correct it is also possible that in some situations, chemotherapy might help cancer spread by making more fuel available to the cancer cells.
鈥淐onventional chemotherapy saved my father from colon cancer, but when it does not work, you get recurrence and metastasis,鈥� says Lisanti. 鈥淭here is a lot of luck involved here, ensuring that you got just the right dose.鈥�
Hart believes a more promising approach might be to target specific molecules that enable cross-talk between cancer cells and fibroblasts.
One possibility is using drugs that block 鈥渁utophagy鈥�, the process by which the fibroblasts self-digest and release nutrients that then fuel cancer growth. The malaria drug, chloroquine, works in this way, so could also be tested against cancer, says Lisanti.
Drugs that inhibit the ability of mitochondria to burn lactate and other products of glycolysis may also serve to cut off the tumour鈥檚 food supply. One such drug is metformin, widely prescribed to treat diabetes. Indeed, several recent studies have suggested that people taking metformin have a reduced risk of developing cancer (Gastroenterology, ).
