eLife Episode 48: Bugs and Drugs, and Chocolate Cake

Conserving trees, gut bugs and drugs, and why there's always room for chocolate cake...
03 July 2018
Presented by Chris Smith
Production by Chris Smith.


Eating Cake


In this episode of the eLife Podcast, signs that trees exchange genes over hundreds of kilometres, how our gut bacteria protect us from plant toxins, and new insights into the placebo effect...


In this episode

Eating Cake

00:34 - Just one more slice

The brain region that encourages you to binge, even when you're full...

Just one more slice
with Saleem Nicola, Albert Einstein College of Medicine

Picture the scene: you’ve just finished a massive Sunday lunch and you feel like you can barely move. In fact, you’ve just slackened your belt a notch because your trousers feel uncomfortably tight. You honestly couldn’t manage another thing. But then someone says, “who would like a slice of this lovely chocolate cake?” And, miraculously, your stomach forgets how full it is and finds a convenient cake-sized corner to enable you to accommodate a generous portion, despite the fact that you definitely don’t need to eat it and just now you felt fit to burst… Speaking with Chris Smith, Saleem Nicola explains how he's uncovered the brain circuitry responsible for what we think is an evolutionary mechanism that once prevented us from starving but is rapidly becoming our downfall…

Saleem - There has been an epidemic of obesity recently, where many people in the modern world gaining much more weight than they did even 50 or 100 years ago. And of course, one reason for this is because many of us don't exercise enough. But it's also the case that we eat too much and typically we tend to eat calorie dense high fat, high sugar foods and we were interested in understanding what the mechanisms of that might be. So that's kind of the broad outline of what motivated this study.

Chris - I guess where you're coming from is you've got people who are already well-fed - clearly, in fact, overfed - yet we overeat despite being well nourished. And the question is why we would want to do that?

Saleem - That's exactly right. And so the way that we normally study animal models of eating has a couple of problems that we tried to solve. One of those problems is that what most people do, most scientists, neuroscientists do when they're trying to figure out the neural mechanisms of caloric intake is they just give the animal food and they say "well how much are you going to eat?" And they do some manipulation and maybe they change it and they say "a-ha, I figured out how this aspect of regulation of calorie intake works." And the problem is that, you know, you might tell your children it's time to eat, and they say I'm not very hungry, but you say but no you don't understand, I've got a chocolate chip cookie. And then they'll say "oh great yes, now I'm hungry," right? So the stimulus makes all the difference in the world because the stimulus predicts the positive outcome. And so we respond to those stimuli and that facilitates and increases the amount that we eat. That hasn't been studied as carefully partly because when people study those mechanisms they tend to study them in animals that are food restricted. But that's a totally different situation than most of us where we really do have enough to eat.

Chris - So how did you do the experiment, the rat equivalent of saying "I've got a cookie for you?"

Saleem - Right, of course we can't tell rats what to do but we can train them. And so we have animals, rats, in what we call an operant chamber, where we can present them with sounds, and lights, and things like that. And also a little well where we can deliver liquid reward, we used a cream reward which is high in fat. And so we had an auditory stimulus that just came on occasionally and told the animal that if it goes well there will be cream available for him. You know, within a week or two you can train animals that respond reliably to these stimuli and they'll do it even when they're not restricted. Even if you give them as much of an ordinary sort of yucky but very nutritious rat chow to eat as they want.

Chris - So you get animals here, which you're allowing them to eat as much they would want to normally, but then you can present the stimulus and they'll still go and overeat because there's a reward on tap which they love?

Saleem - That's correct. Rats, as people do love to drink cream and so they respond to the stimulus even though they're not restricted on food. So we actually had two groups of animals, we had one group of animals that was restricted and one that wasn't. And the study was to ask whether in fact the mechanisms that control responding to the stimulus that predicts cream, are different in the restricted versus non restricted group.

Chris - So how did you study the brain angle of this? What's going on in their nervous system to make them have the "I want to eat the cream, even when I'm full" behaviour?

Saleem - We focused on a part of the brain called the nucleus accumbens, because it's been shown before that a particular neurotransmitter system in the nucleus accumbens can promote animals eating particularly of calorie dense rewards, and particularly of high fat food. This system is an opioid system. there's a kind of opioid receptor called the mu opioid receptors, of which there are a lot in the nucleus accumbens. A neurotransmitter that activates these receptor peptides called opioid peptides, and there are a lot of those in the accumbens as well. And so we reason that maybe this opioid system is involved in responding to that kind of stimulus. And one of the first things we did is we blocked those opioid receptors and asked whether manipulation had different effects in the restricted versus the non restricted animals. And lo and behold we found that there was actually no effect on the restricted animals but it severely impaired responding to the cream predicted cue, in the animals that had as much chow to eat as they like.

Chris - Can I speculate then by saying that are the animals that have eaten normally, once you're sated that, sort of, sets up their system and primes it to be responsive through these mu opioid receptors to the stimulus saying well now you can find room for a bit more a few more highly palatable calories and is that why the animals are susceptible in that context?

Saleem - Right, so we don't know exactly how it works. And one idea is that maybe the opioid peptides that bind to those receptors and activate them, maybe those are released only in situations where: A) the animal or maybe even the person is relatively sated and B) when there are high calorie rewards available. But exactly how that works is really a mystery and that's partly because it's hard to study when these peptides are released because it's released in such tiny quantities that they're very very hard to measure. But what we did do, is we can figure out how these peptides work because we can block that with an antagonist, we can ask what that actually does to the neurons in the nucleus accumbens

Chris - And what are you going to do now to follow this up? What's the outstanding question and the implications of this?

Saleem - There are a couple of questions. We kind of had an idea what happens, when opioid receptors, when they're bound, and then it does this to the firing of these accumbens neurons but what controls the release of those opioid peptides that bind to the receptors. And the other question is we don't know exactly how this works, so we know that this is reduction in the firing of these accumbens neurons. We don't know why and there are a number of different possibilities that we can investigate. We know that the neurotransmitter dopamine is very important. The levels are much lower in animals that are not hungry and it could be that what's happening is that opioid peptide, by activating this receptor is somehow increasing the dopamine levels so that there is sufficient to actually drive this firing.

Eucalyptus melliodora

07:04 - Seed collecting for tree conservation

From where should scientists collect seeds to conserve tree species?

Seed collecting for tree conservation
with Megan Supple, University of California Santa Cruz

When conservationists are seeking to restore woodland and regenerate populations of endangered tree species, where should they go to collect seeds? One school of thought says it’s best to source them locally, since locally-growing specimens may be carrying genes that confer advantages for that specific environment. But, actually, we didn’t know over what ranges trees could exchange genes, or how climate change might feature in the equation. Now, thanks to Megan Supple’s study on one strain of Australia Eucalypt, we have a much better idea, and Chris Smith hears what it is…

Megan - We were interested in asking whether we can build more resilient plant communities through restoration, in particular by incorporating climate change models into seed sourcing decisions.

Chris - And what strategies are there at the moment for doing that?

Megan - So we focused really on seed sourcing, so in terms of how you source your seeds for restoration, it's often quite local. You select seeds that are from near the restoration site with the hope that they are best adapted to the local conditions.

Chris - Right, so what's wrong with that strategy or why shouldn't that be a sound approach?

Megan - Well, it can go awry when you're planting a long-lived species such as a tree in a changing environment. So they're locally adapted to the current conditions but the conditions are changing and so they may not be adapted to future conditions.

Chris - But aren't you sort of asking for the arboreal equivalent of a crystal ball, because no one knows what climate change is going to do to a different geography and no one knows how the species or any species is going to respond to that?

Megan - That's true and so there's climate models out there. They're not perfect, but we can make a best guess as to what we think will be the best seed sourcing strategy. And in particular one of the issues with going quite locally is you have a reduced gene pool to select from. And so when you begin to source more broadly you can incorporate more diversity and then you have that genetic variation there that you need for these unknown things in the future.

Chris - So why doesn't everyone just say "Well the simple solution is then to just source really broadly because then I'll have loads of diversity and I've safeguarded the future?"

Megan - Because they're quite worried about this idea of local adaptation, that over long periods of time these plants have become locally adapted to their environment where they live. And so that potentially seeds from further out might not do well in that environment so there's a real concern about local adaptation.

Chris - So how did you pursue it? What did you do to try and nail which is the more dominant or more important determinant?

Megan - What we did is through a citizen science program, we obtain leaf samples from trees across this species distribution, and then we genotype those at about 10000 genetic markers. Using those markers we generated a model explaining how those markers are associated with geographic location and the climate variables at those sites. Then using that model we can project the model on to predict future climates to see how we expect that different genotypes will fare in different climates.

Chris - And what emerged?

Megan - The first big finding was, we looked just at the patterns of gene flow across the landscape. And so now remember these are 200 year old trees. And so we're really looking at a historical baseline of natural gene flow prior to a lot of the habitat fragmentation that we see now. And we found the gene flow extends really far. We estimated it extends beyond 500 kilometers. That means then when sourcing seeds we can source quite broadly across the landscape and mimic that and have that historical gene flow.

Chris - So when you say gene flow this means that a tree in position A, there's evidence from your data that that's been exchanging genes with trees up to 500 kilometres away?

Megan - Yes.

Chris - So that's a big distribution isn't it?

Megan - Yes it's quite broad. So in some of these species historically they would source seeds from within 15 kilometres of the restoration site.

Chris - But can there not be superimposed on some of these gene flows from further afield, more local effects so you could actually have both things going on so you could be still robbing Peter to pay Paul if you go too far afield?

Megan - So we did not specifically look at those local adaptation and very specific genes, we took a genome wide approach to look at the overall genome signal. But one thing we can do in a future study is use the data that we've collected here and begin to look at that.

Chris - You also looked at one particular eucalypt in Australia didn't you? So what extent can I say right okay that applies to eucalypts in general or that applies to an oak tree? Are they likely to be quite different or quite similar?

Megan - I think there will be similarities in other tree species that have similar pollen flow. And so while this analysis is particular to this specific species, I would think that in species with similar life histories you would expect to see a similar pattern.

Chris - Did you look at epigenetic effects as well because that's become apparent in recent years, that we can see this having a heritable effect so that species can react to the exigencies or particular opportunities in a particular environment? Now you wouldn't see that if you just looked at the genetic code would you?

Megan -  No and we would not see that with this and we did not look for epigenetic effects, and it's hard because these are really complex systems that have a lot of things going on and so we focused on these two particular things of gene flow and climate change.

Chris - So pulling it all together then, if you were now writing the guidebook on how one should collect seed to restore a species or ensure and safeguard the future of a species, what are your bullet point take home messages that this study suggests we should adopt?

Megan - So the first thing I think for this species is again going more broadly across the landscape, and so incorporating seeds from a wider distribution to incorporate more of that genetic diversity that naturally occurs in this species. And the second thing would be; we generated a model that allows us to predict how well different genotypes will perform at a different restoration site. And so from that we can generate a map that suggests where are the best collection sites for a given restoration site and so that's a useful tool because you can incorporate that with other knowledge that you have to figure out where your best likelihood of success is.

Digitalis - foxglove

13:32 - Microbiome plant toxin and drug digestion

Does the microbiome hang on to genes to degrade plant toxins, and drugs, just for our benefit?

Microbiome plant toxin and drug digestion
with Peter Turnbaugh, University of California San Francisco

A few years ago it was discovered that intestinal microbes can detoxify certain drugs and affect the dose that a patient effectively receives. Eggerthella lenta, for instance, can degrade the cardiac drug digoxin. What scientists didn’t know was why they do this, and what role this might serve for the microbe. Now a broad sweep taking in a larger range of E. lenta strains has revealed something very intriguing: the genes that do this seem to serve no direct purpose for the bacterium, and not all strains carry them. So, could the bacteria be making these gene products, Peter Turnbaugh suggests to Chris Smith, just for our benefit, to keep their hosts happy...

Peter - Our lab is really interested in ways in which bacteria that are found within our gut metabolise the drugs that we take. We've known that microbes can impact drug metabolism for many years, but we know very little about the details, and how important they are for the outcome of a given medication.

Chris - Now when you say they metabolise the drugs that we swallow, they change them in some way. Is that because they just naturally happen to have metabolic pathways, enzymes, that can do that? Or is there more to it than that?

Peter - Yes that's one of them questions that we tackle here. There's a general question of why bacteria would choose to metabolise a drug. Is that because they're intentionally targeting the drug with enzymes that are specifically evolved to modify those types of compounds? Or is it because they're intending to metabolise something else that's normally in the body and they have promiscuous enzyme activities. So their enzymes may actually be able to accept, not just normal compounds that are found in the body but also drugs.

Chris - So how did you look into this?

Peter - We, a few years ago, had identified two genes that were found in this prevalent member of our gut microbiome named Eggerthella lenta by using something called RNA sequencing. So we have looked for genes that were turned on in the presence of the drug. And that really allowed us to zoom in and focus on these two particular genes. But one of the challenges was that, at the time, we didn't know whether or not these genes were the correct ones responsible for the reaction, and we knew very little about their biochemistry and how exactly they worked. To get at that, we started comparing a broad collection of different strains. Originally, we had discovered that not all strains of this particular type of bacterium are the same. So, by expanding that now to 25 different strains, we were able to look across the entire genome of this bacterium and find that there is actually just one set of genes that's found in all of the strains that can metabolise the drug digoxin, and missing in strains that can't metabolize the drug.

Chris - Now it's interesting isn't it, that the bacterium has this ability? When you look at that gene and you look at the biochemistry that's making the degradation of the drug possible, what does it tell you about the possible role for that gene in this group of bacteria?

Peter - Well what's really surprising is that we found that this enzyme doesn't work very well against compounds that are normally in the body. It seems to only be capable of metabolising digoxin, this cardiac drug, as well as other closely related compounds that represent other toxic compounds found in plants. And so, you know, unlike what we had expected it doesn't look like this enzyme is acting on compounds that are normally found in the body, it seems to be really specifically tailored for plant toxins.

Chris - So what's it doing there?

Peter - Yeah, I mean, it's still really a mystery why bacteria would hang onto a gene and an enzyme that doesn't normally have an important role in the body. Could be that we just haven't found what it's normally metabolising in our body. Or alternatively it could be that the bacteria are just maintaining these really specialised enzymes in the off chance that we may ingest a toxin.

Chris - Well people do say we should regard the microbiome as an additional organ in its own right, because of the huge number of molecular "knives and forks" that they bring to the table. So this, sort of, could fit that couldn't it, in the sense that we could be providing a home for these bugs and the payback is that they detoxify things from plants that we eat?

Peter - Yeah I think that's a really fascinating way to think about it. What's surprising to me is that this may be a case where the particular knife or fork these bacteria is using to deal with these toxins may not benefit the bacterium itself but may only serve to benefit the bacteria through this indirect effect on the host.

Chris - Now given that you've discovered this and that it may be potentially generalisable to many more organisms and therefore many other compounds. Something like a third of the top 10 agents in a doctor's medicine bag have direct origins in nature. So does this mean then that we actually probably, should be a lot more careful about how we trial agents and how we test agents that are already active in order to take into account what you have found here.

Peter -  Yeah, it's something we think about a lot in our lab, how this information could be used in drug development or even after drugs are on the market. Can we gather data that would help us determine whether or not the microbiome is having an effect on a particular drug.

Periaqueductal grey matter

19:24 - Manipulating the placebo effect

The size of the placebo effect depends on how precise our expectations about a treatment are.

Manipulating the placebo effect
with Arvina Grahl, University Medical Centre Hamburg-Eppendorf

The word placebo means “I will please” in Latin, and it’s the phenomenon that, if someone thinks something is going to help them then they get benefit even if the thing does absolutely nothing. But how much benefit the person gets depends heavily on their expectations, and those, it turns out are coloured by the reliability and reproducibility of their past experiences. Arvina Grahl subjected volunteers to mild heat pain which she relieved with a fake “TENS” device. For some of the participants she tweaked the pain levels to make it look like the TENS device delivered less reliable relief. For others it worked well every time. As she explained to Chris Smith, the question was how would this affect the performance of the placebo effect, and how would it be reflected in brain imaging measurements…

Arvina - So what we actually wanted to address here was to create two different groups and present a painful stimulus which we did with heat pain. One group with very precise expectations about the efficacy of the treatment. And one group with more unreliable expectations concerning this treatment, and the idea was to see that the group that had much more precise expectations concerning this treatment would benefit from this precision because that information were much more reliable concerning the efficacy of that treatment compared to the other group. This would mean that the placebo effect would be larger for the more precise group compared to this group with more unreliable information.

Chris - What was the nature of the treatment that you gave them for the thermal pain?

Arvina - So the treatment we told our participants that it is an electrical stimulation which is called TENS it is well know. We used this to tell them this would help to give you some pain relief during this thermal stimulation but actually we never applied this.

Chris - And how did you then factor in the precise information versus the imprecise information situation? So there was that variation between the two groups, how did you do that?

Arvina - A classic placebo study always works that you present your participants with two conditions; a treatment condition and a non treated controlled condition, and then the second phase would be that you present the same thing again but in the first phase we don't tell our participants that we present them two different intensities are more painful for the control and a less painful for the treated to let them really experience this treatment effect. And in the group with the high precision, we actually presented as this lower intensity always the same intensities of pain, and then the other group where we wanted to create some unreliable information some more variable information we actually varied treatment effects or sometimes that was more pain relieving and sometimes there was less pain relieving.

Chris - So in other words they dont realise it but you're making the stimulus change so that its more or less painful and they think its because the pain relief is better. Its not. Its because youre actually subjecting them to more discomfort and then youre using that to make them think they're getting relief.

Arvina - Thats exactly what we do. We will always tell them for the whole experiment, it is always the same pain intensity that they would get and due to the treatment they will feel a certain relief.

Chris - And what happens?

Arvina - As we did this experiment also in the fMRI scan where we have a look at the brain responses as well, and wanted to see if the brain processes the pain differently when we induce different expectations. We first see in the behaviour that the placebo effect was actually nearly none present in this more unreliable information group, and much larger in this higher precision group. And we also could see that this was modulated in the brain in a very well-known area called the P G.

Chris - This is the Periaqueductal gray matter isnt it? It's where theres a centre of endogenous opioid activity in the nervous system.

Arvina - Exactly. Its a very exciting area actually because it has strong modulating influences on the pain perception, also in pain that incoming and also pain that was processed in a brain and this feedback back to the skin for example whether painful stimulation was presented. Butt what we actually see which is very interesting is we see a higher activation for those participants that had the more unreliable expectations prior to this test phase. So we see that there is more modulation needed if you have more unreliable information concerning the treatment.

Chris - That sounds a bit counterintuitive one would expect that you'd see more activity if you were seeing more reliable pain relief. So how do you account for that?

Arvina - We actually referred that to two different resource demanding processes. So when we know that expectations are very reliable, when we see that there is less activation needed to process this information because there are already similar to those that I expect then you need less resources in this modulatory area.

Chris - And can we utilise this information now to try and work out a way of exploiting this mechanism to a better effect?

Arvina - I would say it is important that a physician actually asks for expectations of a treatment and assesses was the patient actually experienced in other treatments to increase the expectation effect of the treatment in combination with the actual treatment.

An in vitro toxoplasma tissue cyst, 20 days post infection.

25:48 - Parasite lifecycles

How the parasite Toxoplasma gondii adapts its growth rate to its environment

Parasite lifecycles
with Sarah Sokol, University of Pittsburgh

Toxoplasma is a major health burden. In some countries more than half of the population are infected, and if they become immunosuppressed, or pick up the parasite during pregnancy the consequences can be very serious. But toxo has a close relative called Hammondia that is much more limited in how it behaves. So could studying the two agents provide us with new insights into ways to turn off toxoplasma? Chris Smith speaks to the University of Pittsburgh's Sarah Sokol…

Sarah - The lifecycle of toxoplasma is very unique. It starts in the definitive host which is the cat. In this case the parasite will undergo sexual reproduction and the sexual reproduction will produce oocysts and these oocysts would then be excreted from the cat and cat faeces. And these are very environmentally stable. So whenever another host, so an intermediate host comes in contact and ingest these oocysts they will infect that intermediate host, rapidly replicate, and then they'll eventually differentiate into these tissues cysts that results in this chronic, or latent, infection. Those can then be passed back to the definitive host to initiate sexual reproduction or it can bypass sexual reproduction and those tissue cysts can be used to infect another intermediate host. and thus the parasite can be continuously propagated just in its intermediate hosts and its asexual form.

Chris - So let's say that there is some faeces in my garden, which cats seem to love doing. Along comes mouse, encounters the cat feces replete with the toxo, mouse catches toxo, cat could eat mouse completing the lifecycle or, say a rat could come and eat the mouse and become a rat with toxo?

Sarah - Yes, yes it could.

Chris - And how does that compare with the other parasite you've been studying? What is the other parasite you've been looking at?

Sarah - So the other parasite is Hammondia Hammondi. It is very similar to toxoplasma, so it has the same definitive host, it completes sexual reproduction and the cat produced oocysts which then when are consumed by an intermediate host will establish an infection and then eventually differentiate into tissue cysts. However those tissues cysts are only infectious to a cat. So if another animal came along and at an intermediate host maybe in this case mouse they would not contract Hammondia.

Chris - So how does the parasite, the Hammondia, know it's in a cat?

Sarah - That is a great question. I wish I knew the answer to that. Ideally, we would love to be able to understand those differences because I think understanding that would help us understand why toxoplasma doesn't need to be in a cat to continue to replicate 

Chris - Because Toxo is a massive global disease burden for humans isn't it? I mean estimates are that more than a billion people carry toxoplasma and it is a really serious issue for people who have immune problems or are pregnant. So if you can unpick what actually means it's able to bypass not being in a cat then actually you might have a way to intervene and turn off toxo?

Sarah -  Yeah ideally that would be the optimal situation, is if we knew exactly how to manipulate toxo so that it couldn't go and infect additional intermediate hosts.

Chris - Now given that they are so closely related and so similar, can you sort of line up the two genomes and ask well where do they differ. Because perhaps those differences in behaviours are underpinned by those differences in the genetics.

Sarah - Yes. So we've done this. What's surprising is that these genomes are greater than 97 percent synthetic. So that means they have the same genes in the same place within the genome and there's really not a lot of gene loss that would account for their developmental differences.

Chris - So that makes your life a lot easier than if you've got so few differences. Are there any in there, if you look at those differences that might point towards why there is this very dramatic difference in life cycle between the two parasites?

Sarah - Not really. So most of the differences that we see are likely due to assembly differences with the genome. So it's really hard to identify what those differences are. However, what we do know is that both of these parasites differentially express different genes during their developmental life stage and so that'swhat we're trying to use in order to identify differences.

Chris - Could it be some kind of epigenetic phenomenon, whereby the environment in which they find themselves sets a bunch of epigenetic tags signalling to the organism you are not in a cat or you are in a cat and that triggers different stages.

Sarah - Yes so that's one of our most recent hypotheses and where we see this work going in the future is thinking that there has to some type of epigenetic regulation resulting in this expression difference between both parasites.


Add a comment