Chemists take on immunology

Like Comment

Day two of my conference in sunny Paris, and the talks moved on this afternoon to biological chemistry. Long-term readers of this blog will know that I’m not a chemical biologist, so this will be a relatively brief overview of the topics discussed – but what topics!

Kim Janda largely focused on his work discovering antibodies for various therapeutic applications. I never knew that it might be possible to develop a vaccine for cocaine addiction, but Kim seems to be well on his way, with an antibody treatment that – in animal studies – reduces the amount of cocaine that passes through the blood-brain barrier, and which offers some protection against overdoses with the drug.

Carlos Barbas’s talk also strongly featured antibodies. I was particularly taken with his idea of using bifunctional antibodies in combination with prodrugs for anticancer therapy. The antibodies bind to tumours, but they also unmask anticancer agents from the prodrugs; this ensures that cytotoxic anticancer agents are targeted exclusively to tumours (which is good, because such agents are generally also toxic to healthy cells). He showed that a prodrug of the anticancer drug etoposide, when given to mice with appropriate antibodies, elicits a 75% reduction in tumour growth compared with controls, and with no signs of toxicity.

Thomas Carell gave a brilliant talk about his studies of DNA repair mechanisms, having isolated crystal structures of some of the enzymes concerned when in complex with damaged DNA. And here’s one of those mind-boggling statistics: apparently, in humans, every cell sustains 100,000 lesions to its DNA every day. So be thankful for your photolyases and Y-polymerases, which deal with the problem.

And finally, what was the dispute between George Whitesides and Jean-Marie Lehn yesterday? Well, George was discussing the problems of designing ligands for proteins – his conclusion is that our models for the hydrophobic effect and for how water molecules behave in proteins are basically wrong. But he also declared that the ‘lock-and-key’ model of enzyme-ligand interactions is bogus too: you don’t need a precise, tight fit, as implied by the model, but rather a ‘sloppy’ fit. Jean-Marie, however, defended the lock-and-key model, saying that, at the very least, it is the best way of explaining protein-ligand interactions for non-specialists. And then he went on to use the lock-and-key analogy in his slides.

More from Paris tomorrow…


Andrew Mitchinson (Senior Editor, Nature)

Andy Mitchinson

Chief News & Views Editor, Nature