Local anesthetics - and for that matter, general anesthetics, too - do not specifically target pain-sensing neurons. Nor do they bind to extracellular receptors. Instead, they diffuse across cell membranes and block voltage-gated sodium channels from the inside, binding to a pocket in the narrowest part of the channel pore.

The result is a loss of pain sensation, but at the price of losing other types of sensation, too. Anesthesia comes at the price of temporary numbness and paralysis.

In the Oct 4, 2007, issue of Nature, scientists from Harvard Medical School showed that it is possible to smuggle a normally ineffective local anesthetic specifically into pain-sensing neurons by mixing it with another compound that opens a different channel, enabling the anesthesia to pass through.

Alexander Binshtok, Bruce Bean and Clifford Woolf wanted to test whether it would be possible to selectively target pain sensation by taking advantage of the fact that pain-sensitive neurons do double duty as detectors of so-called noxious heat. For that job, the neurons have a receptor that is sensitive to capsaicin - the ingredient that gives chili peppers their burn.

That receptor, the TRPV1 receptor, also goes by the name of vanilloid-1 or VR-1 receptor - which might seem like an odd name for something that is probably the polar opposite of vanilla on the flavor scale, but Bean explained that "capsaicin is a 'vanilloid' chemically, in that it has a backbone similar to vanillin. (He also noted that "the odorant receptors acted on by vanillin in the nose are almost certainly not TRPV1 receptors.")

The researchers combined capsaicin with QX-314, a relative of the local anesthetic lidocaine that cannot cross cell membranes directly because it is positively charged. The combination allowed QX314 to enter pain sensing cells via the TRPV1-gated channel, and then shut them down by binding to the voltage-gated sodium channel.

In electrophysiological experiments on isolated cells, the cocktail had no effect on the activity of large sensory neurons, which do not possess TRPV1 channels, but almost completely blocked the activity of pain-sensing neurons.

When the scientists injected rats with the mixture into either their hind paws or near their sciatic nerve, they found the animals had increased pain thresholds for about two hours. The decreased sensitivity to pain was not accompanied by motor or sensory deficits.

Harvard University and Massachusetts General Hospital have filed patents on the technology and co-author Clifford Woolf is co-founder of Boston-based Solace Pharmaceuticals, which received Series A funding for the development of pain drugs earlier this year. (See BioWorld Today, May 2, 2007.)

But Bean said there are "no concrete plans at this point" to commercialize the work reported in Nature for the time being, although the ultimate plan is, of course, to bring the approach to the clinic. "The application that seems clearest is for local . . . or regional anesthesia," though there is also the "possibility" of treating those types of chronic pain that are due to hyperactive pain-sensing neurons themselves, as opposed to chronic pain that is generated by the central nervous system.

For starters, Bean and his colleagues are working on "developing different agents for the same idea," Bean said.

QX314 actually is not a particularly effective painkiller, and capsaicin, of course, produces a burning sensation that is less desirable in analgesics than at the dinner table. So although they were easy to obtain and provided effective proof of principle, the two agents are "not necessarily the ones that would be optimal for extending this to humans."