The majority of people reading this sentence will, at some point in their lives, undergo a medical treatment that requires general anesthesia. Doctors will inject them with a drug, or have them breathe it in. For several hours, they will be unconscious. And almost all of them will wake up happy and healthy.
We know that the general anesthetics we use today are safe. But we know that because they've proven themselves to be safe, not because we understand the mechanisms behind how they work. The truth is, at that level, anesthetics are a big, fat question mark. And that leaves room for a lot of unknowns. What if, in the long term, our anesthetics aren't as safe for everyone as we think they are?
The only way to know for sure is to figure why anesthetics cause unconsciousness, and how one drug differs from another. Roderic G. Eckenhoff, MD, is a professor at the University of Pennsylvania's Perelman School of Medicine. He's one of the people trying to figure out what general anesthetics really do inside the human body, and how we can use that information to discover even safer drugs than the ones we already rely on today. How does he study that? By drugging tadpoles.
This week, Chemical and Engineering News published a profile of Eckenhoff and his work, written by journalist Carmen Drahl. That piece inspired me to call up Eckenhoff and find out more about what we think we know about anesthetics, why it's taking medical scientists so long to understand such a commonly used class of drugs, and why tadpoles make an ideal model animal.
Maggie Koerth-Baker: Describe for me, in your own words, the current basic theory of how anesthetics work. We're talking about chemicals binding to protein receptors, correct? But what does that mean? Why do proteins matter?
Roderic G. Eckenhoff, MD: The real simple answer is that we don't know. We don't even know what class of macromolecule, for certain, underlies the effects of general anesthetics. And anesthetics don't just do one thing. They produce a myriad of effects ranging from hypnosis, to amnesia, to pain relief and a range of other effects that are much less desirable, like hypothermia, nausea, and vomiting.
But most of us think about the primary effect, which would be unconsciousness, and the answer is still we don't know for sure. But there has been a gradual shift in the field to thinking that protein targets are the likely candidates for this interaction. The main reason is selectivity. Even though "unconsciousness is unconsciousness" the whole spectrum of what the unconsciousness looks like aren't always the same from drug to drug. Some patients are more dysphoric afterwards, for instance. There are components of the electroencephalogram that look different from one drug to the next. That leads us to believe that there's some selectivity.
That's a bit of a surprise, actually, that the drugs aren't all working the same. In the last 5 years or so that's come more to the forefront. A very, very small molecule like halothane, which isn’t used in the United States anymore, might act differently than a drug that's more popular today called isoflurane. There's surprising selectivity to these drugs and we're only now starting to appreciate. And when you talk about selectivity, you're talking about proteins because they have the most diversity in terms of structure.
Ion channels [A type of protein—MKB] are also important because they transduce most communication and signaling in the central nervous system. If we think the drugs affect synaptic transmission, for example, then there's a host of ion channels that could be candidates. They are prime targets simply because of what they do. The evidence to date looking at ion channels in vitro strongly supports that notion.
MKB: In the Chemical and Engineering News article, writer Carmen Drahl talks about some of the major discoveries behind how anesthetics work, and we find out that these discoveries happened in the 1980s. What took so long?
RE: That gives you some insight into how difficult the problem will be. A large group of people has been working on this for a long time and we barely know what class of macromolecule underlies the effects. That's remarkable at this stage, given the millions of dollars that have been poured into the problem. It’s taken so long because we've been searching for the single target or just a few ... but it's probably a bigger problem than that.
My bias, which is somewhat speculative, but evidence supported, is that we're talking about interactions with as many as 10, 20, or even 50 different protein targets. That constellation of small effects disrupts the extraordinarily well-timed signaling in the central nervous system to produce the final common pathway of unconsciousness.
What we’ve seen is that people have their favorite targets that they work on in vitro and when they work their way back up to an intact animal they find that this target has only a very small contribution to the overall effect, in contrast to in vitro work. That's been reproduced time and time again. The model that seems to work best is a small-effects-at-multiple-targets model. How does one achieve selectivity then? What you’re probably seeing is each drug affecting different but overlapping mix of targets.
MKB: We've been using anesthetics for more than a century without really understanding how they work. What does that mean for safety? Are there cases where we know, in retrospect, that an anesthetic was being used improperly because we thought it worked differently than it really did?
RE: The first question is about safety. We started with two principle drugs in 1850, chloroform and diethylether. Those two grew up together and the latter is a very safe anesthetic, but it’s explosive and flammable. So it doesn't mix well with today's electronics. Chloroform is unsafe, in the sense that it's metabolized into reactive products in the liver and causes liver toxicity. It also has bad cardiovascular effects. So both drugs have gone by the wayside and the safety profile of the drugs we do use has continuously improved. But it’s not because of us knowing how they work. It's all been empirical, trial and error stuff.
Today, we have drugs that are safe in the short term, but we’re worried about long lasting effects, especially cognitive effects in the vulnerable brain, for example the elderly and children. Effects could last well beyond duration of administration. That's gotten people worried. That's why we're trying to come up with other chemotypes that don't do these things.
MKB: I think this is one of those things that would sound quite scary to a lot of people — that their anesthesiologist doesn't really know exactly how anesthetics work. But we use these things every day, so it must be safer than it sounds. Why is that not actually as big of a problem as people might think?
RE: Because we get away with it. Worldwide, it's estimated that over 200 million general anesthetics are given each year. In this country alone it’s something like 40 or 50 million per year. Really, only 30-40% of people make it through life without experiencing a general anesthetic. Based on the safety profiles, the bad things that happen and are directly attributable to anesthetic are very rare. But that's just the tip of the iceberg. We don't know what else we're doing long term. We're just not set up to know that yet.
MKB: Besides the fluorescence which makes it easier to track through a tadpole's body, what makes 1-aminoanthracene, the anesthetic you're working with, a better drug? What could it do for humans that existing anesthetics don't already do?
RE: I wouldn't give that to any human. Aminoanthracene is strictly a lab anesthetic that helps us to understand what the microscopic and molecular targets might be. It's only advantage is that it’s fluorescent. If you do reading on 1-aminoanthracene you know they aren't good molecules to have in you for any length of time. Probably carcinogenic.
We have two arms to our research, finding the targets and trying out new drugs. Aminoanthracene has helped in both arms. One arm of the project is discovering new drugs — we've done a large screen of half a million compounds and are now sorting through the hits to find a new class of general anesthetic. The other arm tries to identify what the molecular targets might be. Finding the targets helps us to direct drug development a bit more.
MKB: How is this different from localized anesthetics? Do we understand those better?
RE: The mechanism of local anesthetics for things like epidurals, spinals, local anesthesia, we think we understand that a lot better. They're bigger molecules and there’s a good relationship between selectivity of the molecule and its size — local anesthetics are more selective about what they affect. And part of the safety also comes from the fact that we only give a little bit in a very selective place, to begin with. By the time it disseminates into the rest of the body the concentrations are so low that it does nothing. Dose matters. For example, if you give enough local anesthetic intravenously, they can cause seizures and cardiovascular collapse. But in small doses it’s safe.
MKB: I think a lot of people will be interested in the fact that you work with tadpoles, and not a model that's more familiar to the general public, like mice. Carmen Drahl writes that this is because tadpoles are cheap, and that they are an excellent mimic for human responses to anesthetic. When we say "excellent mimic" what are we really talking about? How does a tadpole on anesthesia resemble a human?
RE: Basically, this sounds kind of primitive, but the basic endpoint used in anesthesia is that when a surgeon cuts the patient, they don't move. I'm serious. It's very, very crude, but it's the coin of the realm. The bottom line is that when you do something that ought to hurt the animal, it doesn't respond. In a tadpole that means trying to elicit a startle reflex by tapping their dish, or tapping the tadpole itself. If it doesn’t do anything, it’s considered anesthetized.
That behavior, loss of movement, we see in animals going all the way down to the fruit fly or the nematode. Any animal that can move can be an anesthetic model. But what I really mean by “mimic” is that concentrations required to produce that endpoint are almost the same, within 10 or 20% or so, of those required to achieve the endpoint in humans. And that’s right across a large number of common anesthetics.
The ability to be anesthetized is a very conserved response. I wrote a paper a few years back on “Why can all of biology be anesthetized?” The response even extends to plants!
MKB: So why can all of biology be anesthetized?
RE: I have a theoretical, protein-based argument — that proteins have small hydrophobic cavities that are essential to their movements and function. If you fill those holes with a small hydrophobic molecule, like anesthetics, you're going to inhibit or change the function of the protein in some way. It may be such a small change that it doesn't matter, or it can matter a lot. But all proteins have these cavities, so all of biology should be affected.
Thanks to Aaron Rowe!