Clinical observations and studies that found that taking common non-steroidal anti-inflammatory drugs (NSAIDs) was linked to a lower risk of certain cancers was good news for many companies, as it bucked the trend at that time of creating anticancer therapies by rational design. And when a next-generation NSAID, the cyclooxygenase-2 (COX2) inhibitor Celebrex (celecoxib), was approved for familial adenomatous polyposis — an inherited predisposition to colorectal cancer — in 1999, there was hope that other COX2 inhibitors would also prove to be safe and powerful anticancer treatments.

However, further studies showed that not all COX2 inhibitors are created equal. The antitumour effects — thought to occur by sensitizing tumour cells to apoptosis — of other equally powerful COX2 inhibitors, such as Vioxx (rofecoxib), were much lower than celecoxib, which implied that the antitumour effects of these drugs are distinct from their effects on COX2 inhibition.

So, Chen and colleagues went back to basics and looked for structural differences between celecoxib and rofecoxib that could explain this discrepency. They used a systematic chemical approach to modify the structures of both celecoxib and rofecoxib to produce 50 compounds, and then tested them for their ability to induce apoptosis in human prostate cancer cells. The pathways through which apoptosis acts was monitored and molecular models were used to identify the key structural elements involved in COX2-mediated apoptosis.

The results, published in the Journal of the National Cancer Institute, confirm that the structural requirements for the induction of apoptosis are distinct from those that mediate COX2 inhibition. The induction of apoptosis required a bulky terminal ring, a heterocyclic system with negative electrostatic potential and a benzenesulphonamide or benzenecarbonamide moiety. To prove their observations, Chen and colleagues modified the structure of rofecoxib to create four compounds that mimicked the surface electrostatic potential of celecoxib — one of which showed a substantial increase in apoptotic activity.

The researchers found that apoptosis was mediated by downregulating the production of AKT and ERK2, which are essential to cancer cell survival. Interestingly, the crucial role that AKT and ERK2 also have in angiogenesis, taken together with previous observations that celecoxib can inhibit angiogenesis, indicates that these anti-angiogenic effects could be mediated by a similar mechanism to that which induces apoptosis. To investigate this further, Chen and colleagues are now looking at the relative contributions of the apoptotic and anti-angiogenic mechanisms to the in vivo effects of celecoxib and its derivatives on tumour growth, and are also assessing the pharmacokinetic, pharmacodynamic and toxicity profiles of the apoptosis-inducing agents.