eLife Episode 7: Sedatives, Maths and Evolution
In this episode of the eLife Podcast, the growing problem of drug resistance, severe brain damage, sugar versus sweetener, public dilemmas, and the evolution of translation...
In this episode
00:40 - Antibiotic-resistant bugs grow better
Antibiotic-resistant bugs grow better
with Stephen Baker, Oxford University Clinical Research Unit, Ho Chi Minh City, Vietnam
Contrary to prevailing wisdom, antibiotic-resitant typhoid bugs grow better than their unmutated counterparts...
Chris - Now so the dogma goes, when bugs mutate to out-manoeuvre an antibiotic, they're forced to surrender some of their reproductive fitness. So, they reproduce less well than non-resistant or wild type microbes. So, when Steven Baker found that his antibiotic resistant mutant bugs in fact grew much better than their parents, he originally thought the results must be wrong.
Steven - So, I've had a long term interest in typhoid fever, a disease that's quite common in some developing countries caused by bacteria, Salmonella typhi. We've been interested in working on how the organism becomes and sustains resistance to a number of different drugs that are used to treat it.
Chris - How widespread is resistance to the first line treatments?
Steven - First line treatments, we no longer really use. The majority of places would have widespread resistance to the common treatments that were being used 10-15 years ago. And there's been more common use of a group of drugs called fluoroquinolones, but resistance to that group of drugs has been spreading quite rapidly where typhoid is common.
Chris - So, how did you do this study?
Steven - What we wanted to do is calculate whether we could explain why the organism seemed to spread very rapidly and whether it was merely dependent on the use of fluoroquinolones. So, we made organisms that were resistant to the drugs that we're interested in by introducing single mutations that we know are associated with the change in resistance to fluoroquinolones, and then we basically competed them with one another. The best way to describe it would be to simply to do a drag race with bacteria over a 150 generations to see which ones are the winners.
Chris - I like the analogy. So, the idea here is that you're comparing whether having these mutations, which on the one hand give the bacteria the ability to resist the antimicrobial drug and whether that encumbers them or impairs them when they're growing against other microbes which don't have that disabling mutation.
Steven - Absolutely, yeah. The scientific dogma really is that organisms that have these mutations or resistant to antimicrobials are less fit because organisms are adapting to work in a particular environment. So, that's what we really thought when we started the experiments, and what we were trying to do is calculate that negative effect.
Chris - So, you expected when you did this that the ones you are mutating were going to grow less well, and you just wanted to know by how much.
Steven - That was our original working hypothesis, yeah. So, we have gathered quite a lot of data on the circulation of these strains in different countries, and there is one particular strain that is dominant, and we thought that that would have a selective disadvantage.
Chris - And is that what you saw?
Steven - No. In fact, we saw quite the opposite. And at first, we were quite concerned about our results and I had several long conversations with a student of mine about the way the experiments have been done. We repeated them to the point of insanity really, to make sure that what we were seeing was correct, and what we found is, some of the organisms actually outgrew or out-competed the parent strain.
Chris - So, this flies in the face of perceived wisdom by becoming resistant to the antimicrobial agent, these bugs are fitter, they grow better.
Steven - That's exactly what data showed and in fact, to emphasise that a little bit more, what we found is that we induced strains with single mutations, double mutations, and also triple mutations. So, one, two, or three different mutations in sites that drugs actually target. And we found if you have two mutations in a particular gene, that actually exaggerates the effect.
Chris - So, why do you think that bugs haven't naturally evolved to have those changes in their DNA?
Steven - So, that's a very good question and it's something that we don't really have a great answer for. The one limitation of our experiments is that we did it in a laboratory. What we don't really know is how that relates to what's going on in the natural environment and there may be other ecological niches that we can't replicate in a laboratory or we don't quite understand that maybe those organisms may behave differently.
Chris - Do you think that in some way, you need the antibiotic agent there in order to maintain the stability of that change? Is it that it's good in the short-term but in the long-term, it does have some disbenefit and you just didn't do the experiments for long enough to see it?
Steven - That may be the case. I also think that potentially, what's happened is that these strains may have been circulating for some time, even before the introduction of the drugs and then the introduction of the drugs formed a greater selective pressure for these organisms and then because of their selective advantage, they've become sustained in the population.
Chris - What do you think the implications are for medical usage of these sorts of agents and therefore, what we're dubbing these days - better antimicrobial stewardship - making sure that we minimise the chances of antimicrobial resistance occurring?
Steven - What it says for typhoid, as the resistance is increasing there's some question about how long we can continue to use these agents for. These organisms are so common that therefore if we stopped using fluoroquinolones now, then actually, we probably wouldn't remove these organisms to the extent that we could start using fluoroquinolones again at any time soon.
Chris - Sobering finding, isn't it?
06:17 - Why a sedative reawakens coma patients
Why a sedative reawakens coma patients
with Nicholas Schiff, Weill Cornell Medical College, New York
The sedating medication "zolpidem" has the paradoxical effect of waking up some patients with severe brain injuries...
Nicholas - We had the good fortune of finding three subjects, all of whom came to us with the history that they would respond paradoxically and improve their behaviour with zolpidem. And, what we did was we brought them in, carried out behavioural assessments, allowed them to take their medications as they were doing and collected continuous electrophysiology data, behavioural data, and cerebral-metabolic data before and after they received the medication zolpidem, to look at brain activity across different times of the day and to make comparisons within and across the patients as to what might be happening.
Chris - Tell us what these patients were like before you gave them any zolpidem, what was the baseline situation?
Nicholas - Each patient was different. One patient was in what's called the minimally conscious state, he showed very clear responses to his environment to indicate that he had some level of consciousness ,but he could not reliably follow commands to do things. He did not speak, he did not answer questions, and he could not organise movements to show basic goal-directed behaviours. On the drug, he reliably and consistently could speak in sentences, he could answer basic questions about himself, historical information about his life, he was able to demonstrate easily the use of common objects, he could also write, and he could ambulate, and he could carry out a wide range of relatively high level cognitive operations. And he would maintain this level of function on the drug for about two and a half hours and then it would start to fade.
Chris - And if you asked him to recall what it was like being off the drug next time he was on the drug, could he recollect that at all?
Nicholas - No. He was not oriented to his situation and he did not show a continuity of memory of his state when he was not on the drug.
Chris - So, we're in this bizarre situation where you've got patients like this and you give them what's otherwise, a sleeping medication, and they wake up. Why do you think this is happening?
Nicholas - Well, the most interesting thing was that when we looked at these very different patients, their brain injuries were very different. But, yet when we look at their electrical activity, we found that remarkably they had very similar frequencies of oscillations, at around 7.5 cycles per second, mostly over the front part of their brain. And we realised that the number of the oscillations, the 7.5 cycles per second might actually point to a very specific hypothesis which was that these oscillations could be generated by what's called intrinsic activity of the cell membranes, that the cells just oscillate on their own like a crystal. When they have a little bit of input but not enough to generate the typical kind of firing the nerve cells generate by getting lots and lots of inputs from other nerve cells, integrating them and then deciding to fire an informative signal.
Chris - So, why would giving zolpidem remedy that?
Nicholas - What we believe is that it would have a direct excitation effect at the cortical cells, but also a suppression of the cells in the globus pallidus, which is the structure in the centre of the brain that can inhibit the thalamus, and the thalamus is the critical structure for arousing or turning on activity across the cerebral cortex. So, the old model, which we call the the MISO circuit model, suggests that after structural injury, in general, the front half of the brain is much more down than the rest of the brain. And that the critical node here is the central thalamus, which plays a role in activating all these structures. And that zolpidem allowed the restoration this thalamic activity and that's why these cells which are just sitting in an idling position, start firing the way they normally would by responding to inputs selectively, and playing a role in the exchange of information and computations at the level of the cerebral cortex.
Chris - And do you think that by bringing people back to life almost in this way, that this might enable you to tap into more plasticity and therefore, their brain can recover better when they're in an on-state like this on the drug?
Nicholas - Well in fact, of the three subjects, there's a little bit of a hint that something like that may be the case. The second subject, unlike the first two, regularly took the drug for 5 years, 3 times a day, every day, and over that 5-year period, the patient changed from being in minimally conscious state when off the drug, to recovering to a level above minimally conscious state, even at base line. So that suggests that there is an increase in background activity and some element of plasticity arising through the continued use of the drug. So, I think that is likely to be true.
11:46 - MCH Neurones give the brain sweet tooth
MCH Neurones give the brain sweet tooth
with Jeffrey Friedman, Rockefeller University, New York
Why does the brain prefer sucrose over sweeteners? The neuronal cell population that gives sugar its appealing qualities has been discovered...
Jeffrey - Sugar is a very attractive nutrient for two reasons. One, it tastes sweet, but is also attractive because it has nutrient value and it's actually the nutrient value of sugar, meaning its ability to provide calories, that explains a lot about why sugar or sucrose in particular is preferred to artificial sweeteners which don't have any nutrient value.
Chris - But I'll put it to you that lots of people do drink diet drinks or they do add sweeteners. So, they obviously do think that the flavour is quite nice. So, when you do subvert your sense systems into experiencing a sweet taste with no calories in it versus glucose, which is obviously more calorie-rich, how does the brain respond to that?
Jeffrey - If animals are given the choice between a natural sugar like sucrose or an artificial sweetener, they invariably prefer the natural sugar. The same is probably true in humans, because all the statistics would seem to indicate sweetened sodas outsell diet sodas by a great margin. Now, it's true that many people use artificial sweeteners and in the short-term of course, they'll find that satisfying but what that doesn't tell you is whether they might go back and have sucrose in some other form later on in the form of a candy bar or a candy. It really seems as if a lot of the intake of sugar and nutrient in general, is unconscious and is driven by signals of the nutrient value that is present in our food.
Chris - Now what actually detects that nutrient value then? So, when I put something sweet into my mouth, is it just the sweet taste and there's therefore a neurological inference of the calorie value of what I'm eating or is there integration of multiple analyses of the food, the calories coming in, plus the sweet taste, or is the brain just looking at blood glucose and going, "Ah! I ate that Mars bar, therefore the sugar's gone up, therefore it must be good for me!"
Jeffrey - The answer appears to be both. So, certainly a sweet taste such as either a sucrose or an artificial sugar, is attractive. But it's attractive because it predicts that that nutrient or that compound you just ate will have calories or nutrient value in it. But if that taste is empty of calories, it won't be reinforced. What reinforced the desire to eat sweet things is ultimately the presence on top of it, of calories which are to some extent thought to be sensed by increase in the level of glucose in the blood. Unknown until recently, however, was where and how that glucose increase in the blood is sensed.
Chris - So, how did you probe that?
Jeffrey - A very talented colleague of mine, Ana Domingos, hypothesised that a particular nerve cell population in the brain referred to as MCH - Melanin Concentrating Hormone - might have something to do with the sensing of sucrose and why it is preferred to sucralose. The reason that Ana hypothesised these neurons might play a role in sugar sensing is because, firstly, the neurons can be shown to fire more rapidly when exposed to sugar, and secondly, because if you kill these neurons, animals lose weight and weigh less. And those two published findings that are to hypothesise that perhaps these neurons played some role in the sensing of sucrose and establishing a preference versus an artificial sweetener.
Chris - Do you think that sensing goes on when the glucose, which just come from the sucrose, goes into the brain or do you think that there are sensors for sweetness in the tongue, perhaps even in the intestine, and they're telling those MCH cells in the brain, "Hey! I'm being stimulated by something sweet"?
Jeffrey - So, the experiment Ana did was, in the first case, kill all the MCH neurons and if you did that, animals could no longer distinguish between sucrose and sucralose. They now had an equivalent preference. Conversely, if she activated those neurons using some new experimental tools, she could switch an animal's preference to the artificial sweetener if it was linked to the activation of those neurons. This tells us pretty definitively that these neurons play some role in the sensing of sucrose but it doesn't actually tell us whether or not the glucose itself or the sucrose itself is actually being sensed by these neurons or by some other group of neurons, such as sweet sensors elsewhere in the body.
Chris - Where are these MCH neurons, and what, more importantly, perhaps do they connect to?
Jeffery - MCH neurons are present in a brain region known as the hypothalamus. In a particular region of the hypothalamus, known as the lateral hypothalamus, and this nucleus is generally thought to activate hunger. This particular hypothalamic region is intimately connected to reward circuits that play a role in changing the motivation for certain behaviours or certain preferences. And so, they're well positioned, both to sense glucose signals directly or indirectly and then connect them to reward circuits.
17:21 - Modelling microbial public good
Modelling microbial public good
with Benjamin Allen, Emmanuel College, Boston
Can mathematics model the evolution of altruistic microbes that make molecules benefiting other bugs? Some microorganisms secrete factors that benefit, not just themselves, but their neighbours too. And mathematician, Benjamin Allen has been developing a way to model how this comes about...
Benjamin - So, we've been looking at cooperation in microbes, and this happens a lot in nature. One of the main ways that microbes cooperate is by releasing chemicals that are useful to others. In a colony of cells, one cell or more has to produce some useful chemical and then that benefit is shared by other cells. So, one example of this is yeast cells, when they are out of their preferred nutrient, glucose. They can produce an enzyme that breaks down more complex sugars into glucose. These digested sugars then, they don't go the cell that makes the enzyme, but they're shared kind of throughout the colony.
Chris - So in other words, that one cell that makes, I think it's called invitase isn't it, the enzyme. That goes into the media that all the cells are growing in. So, it's sort of, one cell pays to make that molecular machine, all of the cells benefit from the fact that it's chopping up other sugars to make things available for everybody.
Benjamin - Yes, that's exactly right.
Chris - So, what's the big unknown then?
Benjamin - Well, so this is a form of costly cooperation and the question is how does this behaviour evolve? So, if we think of survival of the fittest, you would think that the cell that produces this enzyme is at a disadvantage because it is doing all the work to produce this and then other cells get to reap the benefit.
Chris - Doesn't that sort of assume that the other cells are going to cheat on the equation in the sense that if the cell is producing some invitase secreted into the local environment, won't its neighbours do the same?
Benjamin - You have to think that behaviour evolved somehow, so at some point there must have been competition between cells that produce this enzyme and cells that don't. And somehow, this production behaviour survived through evolution. So the question is what gives this production behaviour an evolutionary advantage over the cheating behaviour? There's two answers to that. One is that the cell actually does get to keep some of the benefits from its own production of whatever useful resource it makes. The other thing is that if the cells are arranged in some spatial configuration and they reproduce locally, so that they're daughter cells remain near the parent, then they're sharing this benefit with other cells who are also likely to be producers of this public bit.
Chris - Simple to say, I suspect rather tricky to model mathematically.
Benjamin - Right, it is. Our test was to simplify nature enough to be able to do exact mathematics on it, but not so much that we'll lose the essential features of what's going on. For example, in colonies of microbes, cells can be arranged in all kinds of chaotic, interesting ways, but there's a certain regularity to it. E. coli for example, are long, thin cells. So, when they're arranged in a colony, there's some chaos, but there's also spaces where the cells are kind of arranged in a predictable end to end fashion. And so, we took that regularity and said, well let's assume that all the cells are arranged in some very predictable, regular fashion. And so then we can say, well, what if one of these cells produces this chemical, how does that become shared with the various neighbours?
Chris - And what do you find if you do that?
Benjamin - We find that the cooperation behaviour has an evolutionary advantage if the benefits are high enough and if they're shared locally enough. We call this a diffusible of public good, because some of it is kept by their producer, some of it is kept by the neighbours, some of it is kept by the neighbours of neighbours, and that rate of sharing has to be small enough so that most of it really goes to the cell and its immediate neighbours.
Chris - What about the concept of things cheating because it could be that an external organism comes in and steals, it takes away but doesn't pay back into the pot? And equally, it could be that the organism itself evolves to do that and becomes parasitic on its parents, if you'd like, rather like modern day teenagers.
Benjamin - Right. So, that's then the evolutionary question. If you have a colony that's already cooperating, and then a cheater arises in that colony, you have to ask, well, what are the odds that that cheating behaviour will spread throughout the colony? What we find here in our model is that even though this cheating behaviour might initially have an advantage, it doesn't have enough of an advantage to take over because say, the cheater does well initially and produces some offspring, some daughter cells then - those daughter cells - will also be cheaters. And so, the cheater will then be surrounded by other cheaters and thus, in the long term, it won't be actually getting any of the useful public goods that it needs.
22:55 - Yeast evolution reveals cancer strategy
Yeast evolution reveals cancer strategy
with Yitshak Pilpel, Weizmann Institute, Israel
What happens if you delete the gene encoding an essential transfer RNA in yeast?
Yitzhak - This particular study, aimed at understanding how the protein translation machinery evolves. In between a protein and a DNA, is a process called translation, carried out by a factor we call a ribosome that looks at RNA and it converts it, or as we call it, translate it into protein. Its a very sophisticated process that has evolved in living species for billions of years. And we wanted to understand how it evolved.
Chris - How can you probe what's taken billions of years of evolution in a test tube in the laboratory within the scope of a grant?
Yitzhak - So, the way we probed that is we mutate the system and then we ask the organism to relax back to a normal state in which it has to mutate so as to restore all balances and checks, and so on, within it. So we went in particular ,to one type of molecule that is at the heart of the translation process. This molecule is called tRNA, a transfer RNA, that helps the ribosome decode information that is stored in the RNA and converts it into a protein. And we decided to delete them from the genome of the yeast cell. If you remove a particular tRNA, you just eliminate the ability of the cell to incorporate one of the 20 amino acids into its proteins and then you ask it to solve the challenge. And we were amazed to see that although this a major assault on the cell, after just about a month or two of evolution, it overcomes the challenge.
Chris - How does it actually do that? Does it take one of the other tRNA molecules that would code for the incorporation of different amino acid, and subvert that into using a pathway that would have been subserved by the now missing one, or does it do something else?
Yitzhak - So, actually it turns out be exactly the solution. The cells convert another tRNA into effectively, being similar to the tRNA that we have deleted. So, if different tRNA serves as almost like synonymous words to one another, you take one synonymous word and you convert it by a mutation process to become identical to the one that was lost and by that, restores the balance.
Chris - Do you not end up robbing Peter to pay Paul, whereby you take one of these tRNAs to replace the missing one? Well, what happens to the job it used to do?
Yitzhak - It turns out that the pool from which it takes, contains many redundant copies, 11 copies. And the one that we have deleted appears in the single copy. So, it makes a lot of sense in fact, for the cell to sacrifice one of the 11 copies to solve the problem of a singleton gene of another type.
Chris - Seems to be a day for really bad analogies and metaphors, but chicken or egg? Why are there 11 copies of some of these tRNAs? Is that with this very process in mind?
Yitzhak - Yeah. That's a very interesting idea, that the cell is generating a reservoir of backups providing molecules, such that mutations can shift one into another. It's a very interesting concept, although I believe that it's the mere role of the 11 copies situation, was just to backup the cell in case of a mutation. In some other tRNA, it would not have been stably maintained in the genome as such. So, we quite pretty much believe that the 11 copies serve for a good purpose in the cell. But when you have zero copy of another, it's an emergency situation and one of the 11 copies needs to be sacrificed.
Chris - What do you the implications of what you found for biology and evolution are?
Yitzhak - The main value of lab evolution exercises, is that it allows us to ask what happens in nature. So, in fact we were able to get a clue from what we found in the lab, and then examined all the genomes of all the species that the genome project has explored so far. And we discovered that its mechanism was evolving the translation apparatus in which one tRNA mutates into another is very prevalent in nature. And natural history of species has utilised this mode of evolution to a very common solution to challenges that must have occurred on Earth. And in addition, we have a good reason to believe that something along these lines occurs also in cancer. Cancerous cells are cells that need to evolve a capacity to grow much faster than the normal cells in the body. We think today that cancerous cells may have accumulated mutations in their translation apparatus that would help them to convert between different synonymous tRNAs to serve the needs of cancerous cells.