Affinity help

Matt Jones jonesmat at
Mon Apr 12 13:20:53 EST 2004

BilZ0r <BilZ0r at> wrote in message news:<Xns94C77065B5DCEBilZ0rhotmailcom at>...
> "John H." <johnh at faraway.> wrote in news:4076a1bb at
> > Since you have earned my pity, give me a couple of examples, I have a
> > knack for extracting information from databases and I'll see if I can
> > establish some suitable generic search strings. I have registration at
> > this site so may be able to help.
> Well you could give me a 2nd opinion. One is preety common, but it still 
> took me a while to find:
> NMDA for the NMDA receptor
> SKF81297 a D1 selective agonist, for the D1 receptor.


You've earned my pity too.  I can't help much, because I've always had
to go through the same process you're struggling with now, and have
never found a shortcut other than to carefully review as much of the
relevant experimental literature as possible.

However, as if your task weren't difficult enough already, I feel
compelled to raise one more issue:

"affiinity' is not a very well-defined term in the first place. 

One needs to specify exactly which "affinity" one is after. For
example, for an agonist (like NMDA) the affinity revealed in a
radioligand binding experiment is going to be much different (e.g., up
to several orders of magnitude) than the affinity revealed in an
electrophysiological experiment. Further, the affinity revealed in an
electrophysiological expt will depend on how the response is measured.
A dose-response curve for peak current will yield a lower affinity
(e.g., up to a couple orders of magnitude) than a dose-response curve
for steady-state current.

These problems are generally less severe for antagonists, but they may
still exist.

By the way, the dependence of "affinity" on what experiment is used to
measure it is not really a matter of methodology, it's a bit deeper
than that. One can reproduce this dependence using mathematical
models, so the phenomenon is a result of the math underlying how
receptors work.

Also note that classical pharmacology usually insists on steady-state
measurements, but these will yield affinities that, in general, are
thoroughly unhelpful in predicting physiological responses to
"natural" (i.e., highly non-steady-state) stimuli.

Sorry to add to your headaches, but it sounds like you're getting
ready to do some sort of modeling of affinity, or comparison between
different receptors. Maybe knowing about this issue can save you some
problems later on down the road.



More information about the Neur-sci mailing list