The oncogene c-Myc controls the fine line between life and death, as it can induce both cell proliferation and apoptosis. But whether c-Myc-induced cell death can actually restrain tumour growth has remained undetermined. Stella Pelengaris et al. have now shown that c-Myc-induced apoptosis can indeed prevent tumour formation, and that switching off apoptosis allows the tumorigenic capability of c-Myc to proceed unchecked.

A switchable c-Myc — in which the gene is fused to the 4-hydroxytamoxifen (4-OHT)-responsive oestrogen receptor (c-MycERTAM), so the protein is activated following the intraperitoneal administration of 4-OHT — was specifically targeted to pancreatic β-cells in mice using the pIns insulin promoter. Induction of c-Myc initially resulted in cell proliferation, but this was accompanied by a faster rate of apoptosis, such that the net effect was β-cell ablation and hyperglycaemia caused by loss of insulin-producing cells. Interestingly, when 4-OHT was withdrawn, which switched off c-Myc, pancreatic islets rapidly regenerated and blood glucose levels returned to normal.

If apoptosis overpowers cell proliferation, inhibiting apoptosis should allow proliferation to proceed unchecked. Expressing the apoptosis inhibitor Bcl-x L under the control of the rat insulin promoter (RIP7) allowed this hypothesis to be tested. pIns-c-MycERTAM/RIP-Bcl-xL transgenic mice had normal pancreatic islet formation until 4-OHT was added, at which point proliferation was induced throughout the pancreatic β-cells. As apoptosis was inhibited by Bcl-xL, this resulted in hyperplasia within 7 days.

But can deregulated expression of c-Myc induce tumour formation, which is thought to require the cumulative effect of multiple mutations? Pancreatic β-cells in pIns-c-MycERTAM/RIP-Bcl-xL mice not only hyperproliferated, but also underwent de-differentiation — as seen by the reduced production of insulin — and extensive angiogenesis. Expression of the intercellular adhesion molecule E-cadherin was lost as well, which is a prerequisite for loss of cell–cell contacts and invasion. c-Myc therefore seems to be able to directly induce several of the hallmarks of cancer. Two weeks after induction of c-Myc expression, pIns-c-MycERTAM/RIP-Bcl-xL mice had developed pancreatic tumours, and by 8 weeks the tumours were large and vascularized, with sites of local invasion in local blood vessels and draining lymph nodes.

So, expression of a single oncogene — c-Myc — is sufficient to induce several steps of carcinogenesis, as long as its innate apoptotic activity is curtailed; but is it also required to maintain the tumours once they have formed? Switching off c-Myc, 14 days after its induction, resulted in a reversal of the tumorigenic process: β-cells exited the cell cycle, E-cadherin was re-expressed and cells re-established cell–cell contacts, and endothelial cells and β-cells apoptosed. Even mice that had expressed c-Myc for 8 weeks, with extensive tumours that had invaded into lymph nodes, made a full recovery following c-Myc deactivation.

These results challenge the paradigm that carcinogenesis is a multistep process that requires many mutations, and indicate that, instead, it can be driven by deregulated expression of a single growth-deregulating oncogene, provided apoptosis is suppressed. If this is found to be true for other commonly mutated oncogenes, new cancer therapeutics should aim to inhibit these few crucial molecular targets.