eLife Episode 34: Man's First Footsteps

Clinical trials are usually based on findings from animal studies. But, it turns out some of those studies are based on dubious methodology, so a comprehensive review of the background literature is absolutely critical, as clinical investigator Manoj Lalu explains to Chris Smith...

Manoj - Many investigators in the past have looked at individual studies that have been performed in the lab and said, “Hey, this looks great. Let’s try this out in patients now.” However, that’s been met with largely, unfortunately, negative results. Meaning, most things don’t work that way.

Chris - Is that not what we call evidence-based medicine?

Manoj - It is in some ways, but the problem with what we called preclinical work or basic science work, the evidence base is usually pretty small, and you're talking about experiments with 10 to 20 animals. However, clinical medicine usually, we’re talking about evidence that covers thousands of patients.

Chris - So you're saying there's a fundamental problem here. people do research on a small group of animals and then call those men not mice.

Manoj - Unfortunately yes, that’s usually what occurs and what we’re trying to do is provide an alternative pathway that perhaps people can explore to get things from the bench to the bedside. Basically, we’re looking at using stem cells to treat septic shocks. So that’s when you have a life-threatening infection – when a person has this, their blood pressure gets very low. They get admitted to an intensive care unit. Right now, what we do for these patients is we give them antibiotics. We give them lots of fluids and then we give them other supportive treatment. We treat the underlying infection if we can find one as well.

Chris - What would be the rationale then for attempting to use stem cells in that sort of setting? How might stem cells assist a doctor making someone with septic shock get better?

Manoj - We think they might work in several ways. So, often people think stem cells go in and they become part of your body. However, in the setting of septic shock, we think these cells go in and they have anti-inflammatory effects. So, when you have septic shock, you get really a huge amount of inflammation that occurs in your body. We think these cells decrease the inflammation and allows the body to recover a little bit better that way.

Chris - What did you find when you began to survey the landscape of this subject a bit?

Manoj - What we did which was different from what other people have done in the past when they're trying to move from bench to bedside, we looked at the complete literature. So, we performed what's called a systematic search to try to find every single study where they had given animals these stem cells. What we found was that overall, that these stem cells, when you looked at the broad literature, reduced death in these animal models of septic shock.

Chris - Were you happy that those studies that you were looking at had been conducted in a rigorous way, in a way that could be translated to the bedside in an appropriate way rather than in a speculative one?

Manoj - First of all, I guess on the surface, yes, we’re happy that many different labs and many different groups have almost replicated the same sort of results, meaning that they had these animal models of septic shock and overall, it appears that these cells are protective. On the flip side which is a little bit concerning, the rigour that you were speaking about there, that’s where I think we can improve ourselves as basic scientists in terms of looking at ways that we can reduce what we call bias in our studies.

Chris - Is that not something though that basic science should be across anyway, somehow who is training to become a professional scientist should understand from the get-go, how you do an unbiased balanced, and rigorous study? Otherwise, it’s not valid science.

Manoj - That's a very good point and I think it speaks to something that we’re only recognising in the last few years, and that’s that we don’t get this sort of training. Really, when you're looking at these sort of methods to reduce bias to improve rigor, they're not really taught in a systematic way to trainees.

Chris - Would not a cynic put it to you though Manoj, that actually, what you’ve done is what you should be doing anyway before going anyway near a patient with any piece of research. We ought to make sure that we have thoroughly appraised the literature so that we make completely evidenced based and informed decisions about anything we do.

Manoj - I would agree with the cynic and say, yes, this is what we should be doing. However, unfortunately, that’s not what's been done in the past. There's a couple of very good examples of this where clinical trials have been designed and the push to move from animal work to patients was around a couple of positive papers. So, one was on endothelium receptor antagonist for heart failure and second was on an antioxidant agent for stroke. There was what we thought were promising preclinical results. However, the clinical trials ultimately failed and then retrospectively, scientists went back and looked at the complete literature. Surprisingly, what they found was the evidence really wasn’t that great when they conserved the complete body of literature.

Let's say you're an ant and you've just found a juicy treat; but it's too big for you to manage alone, so you retrace your steps to your nest and invite your sisters to come and help you. But the route you took to the treat involved threading your way through several ant-sized tight spots. The food won't fit through there, so even with your nest-mates to help you, you still won't get it home. So what do you do?

Well, thanks to Ofer Feinerman, who has painstakingly videoed ants solving these problems, we now know that the answer is that you literally go off the beaten track; everyone chips in an idea regarding what direction you should all walk in and you pay a bit less attention than normal to the chemical signals that led you to the treat in the first place. Chris Smith hears how it works...

Ofer - We were watching ants carrying large things together. So when the ants need food items that they can't carry alone, they can just team up – 10, 20, 30 – even more ants, and just haul it to the nests together. We’re looking at this phenomenon which is very interesting because it gives you the connection between a single ant and the group. You're seeing single ants pulling whatever and the group, doing what it does, transporting the item.

Chris - If you watch what ants do individually, is it the same thing when they're working alone when they're carrying something small that a single ant can manage as when a big group of them do it or the different behaviours that go with those two different scenarios?

Ofer - Yeah. So there is a slightly different behaviour. When you have a big object, the first ant to find it, she will trace to the nest and she will recruit other ants and then these ants will come back with her to the item, and they can carry it together.

Chris - How do they find their way back to the nest? Do they follow a chemical trail laid down by the ants that first made that journey back and forth?

Ofer - Exactly. So, when we looked at this first ant, we saw like a very fast stop and go motion. She runs and stops, runs and stops, runs and stops, and actually, when we looked at this from the side, we saw that each time she stops, she puts down her gaster, touches it to the surface, and that’s when she lays a small scent mark like laying pebbles in Hansel and Gretel. When the ants come out of the nest, they can follow these scent marks back to the food item. What we actually discovered now is that the ants also use scent marks also in the stage after they start to move the object when they're navigating the large object home and then they treat them in a very different way.

Chris - Why do they behave differently? Why do they need new scent marks and how is their behaviour different on that occasion compared with the first occasion?

Ofer - So actually, we filmed the ants and just from the stop and go motion of the ants, we can deduce where the scent marks were and we start to pinpoint each and every scent mark that the ants were laying during this movement. They're actually guiding. They're telling the object where to go next. And then we saw something that’s very interesting that these ants that were trying to guide the object, the food with their scent marks, were often making mistakes. There is some conflict between the different scales of organisation in the group. Here, it’s very simple. The ants can go through any narrow passage and maybe destroy this way to the nest. They will mark through there, but the object cannot pass. What we found when we looked at the sum of all these scent marks, actually, what works here is this new ant trail that is just made of very short markings by the ants. They just mark the next step to go, not the whole distance to the nest. This is one aspect of it. The second aspect of it is that the ants that carry don’t follow it religiously. If they would follow it religiously, they just get stuck in these narrow corridors.

Chris - So, are you saying then that the ants which are carrying the load, they are contributing to the scent marks. Sometimes they get it right, sometimes they get it wrong, but equally, they don’t follow it religiously so they sometimes follow it wrong too, and this is going to introduce some sort of noise or error signal into the path they take and that’s in some way going to help them to get back with the big load to the nest more efficiently than if they were to just follow the original path laid down.

Ofer - Right. They know how to navigate. They do this by giving very small opinions. They don’t give a huge opinion, “this is the whole path you should take to the nest. This is the next step you should take.” The carrying ants, they listen to it, but partially, they often ignore it. These two together, they tune the way that the group listens to the individuals. It tunes just so that they can pass through these complex environments.

Chris - Obviously, you made the observations by watching in painstaking detail what these ants were doing, but you can speculate on that behaviour, but have you now taken the data and modelled it to say, “well, we think that’s what's going on so now, let’s do some simulations and see if this is genuinely what these ants have evolved to do.

Ofer - Right. That’s what we did actually. We collaborated with another lab in France, their lab that studies distributed computing. Actually, the easiest way to explain the model that we came up with is by an analogy. Imagine that you're driving in some foreign country and it was just hit yesterday by a hurricane. What this hurricane did was turn to the road signs, but not all of the road signs. But not all of the road signs, like 95 per cent of the road signs are correct and you want to get from one city to your destination. What would you do? How would you follow these road signs? Most of them are correct. What you have to do, you have to carry a dice with you. Each time you reach the junction, you have to throw the dice. If you get one to five, you follow the road sign as it is. But if you get a 6, you just choose a random direction. The theory told us that this is optimal. This is the best you can do and you will reach your destination in almost the fastest time possible. This is completely analogous to what the ants were doing.