A couple of colleagues turned me on the other morning to a press release by researchers at the University of Warwick who recently published in PNAS that their data apparently overturns the Meyer-Overton Rule regarding solubility of a compound in olive oil and its propensity for crossing biological membranes. I'm having trouble understanding exactly why their conclusions are earth-shattering.
At the turn of the last century, Meyer (1899) and Overton (1901) independently conducted experiments to demonstrate that the longer the carbon chain of a molecule, the better it partitioned into olive oil relative to water, and the more potent it would be as a general anesthetic (I'm almost certain that Meyer did his anesthetic work in tadpoles). This is back when we thought that general anesthetics worked primarily by disrupting ion channel function by altering cell membrane structure (it's actually more complicated than that, involving anesthetic molecules directly interacting with hydrophobic surfaces on ion channels, but the bottom line is that the better a molecule partitions into the lipid portion of the cell membrane, the greater its anesthetic potency. This point only holds true for inhaled gases from nitrous oxide to halogenated hydrocarbons like isoflurane and is unrelated to sedative/hypnotics used in anesthesia that have discrete molecular actions such as benzodiazepines like diazepam or opiates like fentanyl).
The press release from the University of Warwick describes what appear to be really cool electrochemical experiments with ultramicroelectrodes and confocal microscopy that are to be published in the 26 August 2008 issue of the Proceedings of the National Academy of Sciences (PNAS). The research team apparently provides direct evidence that increasing carbon chain length of carboxylic acids (acetic, butanoic, valeric, and hexanoic) cause them to pass through membranes progressively more slowly.
But instead of being at odds with the Meyer-Overton correlation, this is exactly what one would expect from the Meyer-Overton experiments.
The point is that membrane-disrupting anesthetics act primarily by staying in the membrane. (I might add that the press release is vexing to me because it fails to refer to the work of Meyer, calling it the Overton rule, and uses the term "cell wall" instead of "cell membrane.")
Perhaps I am wrong in my perception of the Meyer-Overton correlation. My German isn't good enough to review their original papers but I don't believe they made any mention of how fast a drug passes through membranes but rather how easily they partition to lipid membranes.
However, text in Chapter 2 of my 1941 first edition of Goodman and Gilman's The Pharmacological Basis of Therapeutics states that the "law" already was known to have exceptions more than half a century ago:
The lipoid theory was advanced by Meyer (1899, 1901). In its simplest terms, it states that there is a direct parallelism between affinity of an anesthetic for lipid and its depressant action. In other words, the more soluble an agent is in oil and the less soluble in water, the more anesthetic it will prove to be. . .
. . .Evidence for the lipoid theory rests mainly on the fact that the arrangement of a series of chemical agents according to their solubility coefficients agrees fairly well with the order obtained by grouping the sames substances according to their anesthetic potency. The many arguments against the theory cannot be given at length but mainly center around the facts that most of the solubilities have been studied in vitro with vegetable oils and water rather than with brain lipids and bodily fluids; that a number of anesthetics fall out of line, notably chloral hydrate; that alkaloids and inorganic ions do not comply with the theory; and that many fat-soluble substances closely related to anesthetics have no depressant action on the nervous system. [emphasis mine]
I believe I even learned that in 1982, and was certainly teaching in 1992, that the Meyer-Overton correlation was a great initial hypothesis in its day but still had its limitations and went out the window after a carbon chain length of about 10. In the late 1980s, we recognized that lipophilicity was necessary, but not sufficient, for anesthetic action. Little did I know that this rule had already been questioned, and already had accumulated significant exceptions in 1941.
(For those wishing to read a nice description and graphic representation of the Meyer-Overton law in the context of the history of anesthesiology, see this discussion of the University of Pittsburgh's Dept of Anesthesiology)
As an aside I should also make a brief statement about the many, many other drugs that act either by binding to cell receptors or acting on effectors inside the cell or on enzymes in plasma or other extracellular fluids.
The Meyer-Overton law bears absolutely no relation to non-anesthetics. In fact, this optimal lipophilicity for drugs other than anesthetics is depicted more appropriately as one of a balance. Everyone in drug discovery and development is already aware that a combination of lipophilicity and hydrophilicity are required for "druggable" molecules (the 1997 Rule-of-Five proposed by Lipinski and colleagues from Pfizer comes to mind; oft-debated but well-understood that the octanol/water partition coefficent vs. membrane permeability is an inverted U relationship) - remember that there are those dang negative phosphate groups in the cell membrane that must be contended with before one gets to the acyl tails. Simply put, stuff has to have some degree of aqueous solubility to get to the point where it has a chance to pass into biological membranes, some lipophilicity to negotiate the rich, fatty center, and a bit more hydrophilicity to escape the other charged side.
But back to the University of Warwick press release.
The lead researcher on the study from the University of Warwick, Professor Patrick Unwin, said:
"This was a surprising and exciting finding. Our direct observations appear to totally undermine a key rule that has withstood the test of time for over a century. We will now make observations with a range of other chemicals, and with other techniques, to further elucidate the molecular basis for our observations. Text books will have to be rewritten to revise a rule that has been relied on for over a century. Advanced techniques, such as the one we have developed, should give much clearer insight into the action of a wide range of drug molecules, which will be of significant interest to drug developers."
No offense, Prof Unwin, because I suspect the techniques of your team are pretty fabulous.
However, textbooks needn't be rewritten.
They've already been noting the limitations of the Meyer-Overton rule since 1941.
It gets worse, to be honest - they've missed the point that just because an organic acid is swapping sides in a membrane very fast, it doesn't necessarily follow that it'll deprotonate when it does so: the pKa of an organic acid in solution is ~4.5, whereas in a membrane it's nearer 8. There's a 3-order of magnitude preference for the longer-chain models they've used to remain membrane bound that for the smaller ones, and they haven't bothered to correct for it.
Oh, and their error bars are >+/-20%. You can tell it was a PNAS Direct Access job - most reviewers would have torn lumps out of it. If I was aiming to send my paper to all the world's media outlets, I probably would have checked the damn maths first...
Very interesting discussion, thanks. I think Orac had a recent discussion inre uncritical acceptance of articles in PNAS; sorry, the URL is not at hand.
I had a professor from Scotland who liked to say (imagine the accent) "We knew that thairty years ago."
Perhaps the bottom line is- we must not underestimate the value of the press-release.
Good job, bro. And thanks for the ack.
Interestingly, the paper seems to have vanished in the last couple of days. It's no longer searchable via the title or any of the authors, and the DOI doesn't lead anywhere. Strangely, this change doesn't seem to have merited a press release...