For most humans, foods that have been cooked or otherwise processed are a part of everyday life. But what happens on a molecular level when you chop, mash, and sautee your meal? How has cooking given humans an evolutionary edge? And how is new research on the human microbiome challenging information listed on nutrition labels? Harvard University’s Rachel Carmody tackles these questions by studying the past, present, and potential future of how, and why, humans eat the way they do.
Podcast: Download | RSS | iTunes (47:42, 46.2 MB)
RACHEL CARMODY (ASSISTANT PROFESSOR OF HUMAN EVOLUTIONARY BIOLOGY AT HARVARD UNIVERSITY): It is such a pleasure for me to be here tonight. The American Museum of Natural History has got to be one of my favorite places on Earth. To me, it’s one of those places where both human knowledge as well as what we don’t know seem limitless. And I think for someone like me who’s a scientist, it just doesn’t get better than that. So I’m delighted to be here.
So, on the whole what we can say is, as a group, we eat a highly processed diet on average. And we’re really not alone in this at all. Cultures all around the world—in fact, every human civilization known actually processes its diet extensively, both by cooking as well as by a variety of non-thermal processing methods. And this includes cultures like the Inuit who famously do consume a portion of their hunted and fished materials raw. But even among the Inuit, the standard evening meal is both cooked and highly processed.
And so tonight I wanted to ask two basic questions about this unique and universal human behavior. And the first is: Why is food processing so universal? And the second is how might it threaten our health today?
And it’s worth noting that a lot of the ideas that I’m going to talk about I generated together in collaboration with my graduate mentor, Richard Wrangham. And I think he was really the first who dared to look behind and beyond the blinding, everydayness of cooking and really say, why do we do this in the first place? And for those of you who get interested in this topic and want to learn more than what I can really tell you in half an hour, I suggest looking up his book, Catching Fire. It’s a really great read.
So, first, let’s think about why is food processing so universal? And anytime we think about diet and why an organism consumes the diet it does, it’s really important to start thinking about the dietary adaptations of that particular species, because that really sets the playing field for what’s possible.
So if we think about what some of the human dietary adaptations are in comparison to our closest relative, the chimpanzee, we can see that humans have evolved a suite of features that necessitate a high daily energy budget. By which I mean we have to take in a lot of calories. And one bit of evidence for this is our increased body mass.
So, on the average, a human male, even in traditional populations that do not tend to be obese, human males tend to be 34% larger than male chimpanzees. Human females tend to be 56% larger than female chimpanzees. [And all else 10:37] equal a larger body does require more calories to sustain itself.
We also have a dramatically expanded brain size compared to chimpanzees. So, on average, our brains are about 1,350 cubic centimeters whereas chimpanzees are about 400 cubic centimeters. And the reason that this is relevant to a high energy budget is that the brain is an extraordinarily expensive tissue. It weighs about 2% of our body mass, but in a body at rest it consumes somewhere between 15 and 25 percent of our energy.
We also, especially in traditional societies, tend to be much more active than chimpanzees are. So, in wild chimpanzees, the average is they travel distance per day about 3 to 5 kilometers, whereas in traditional foraging populations, 10 to 20 kilometers per day of travel is the norm. And, obviously, travel does cost quite a bit.
So, when you do the math, this works out to humans having much higher rates of energy expenditure, even after you control for the differences in body size. So based on equations that were suggested by Leonard and Robertson several years ago, we calculate that human males expend about 44% more energy than do chimpanzee males, adjusted for body mass. And human females expend 17% more than do female chimpanzees, again adjusted for body mass. And that number is probably a dramatic underestimate because these equations didn’t account for the cost of gestation and lactation which are expected to be higher in humans due to our shorter interbirth intervals.
And so, in order to fuel our large bodies and our large brains and our pretty active lifestyles, humans require a lot of calories compared to chimpanzees. But we don’t just eat more of the same food. We eat fundamentally different food and we know this because humans also show adaptations that signal a loss of digestive capacity and I’ll explain what I mean.
So, what you’re looking at here are chimpanzee and human skulls shown from the underside. So you can see the … do I have a pointer here? You can see the holes at the bottom, that’s the foramen magnum, that’s where the spinal cord passes through and inserts. So you’re looking at these skulls from the underside.
And what you can visualize here is that we have smaller mouths, shown here on the right, in blue, compared to chimpanzees when the two crania are scaled to the same size. We also have smaller muscles for chewing, including the temporalis, which attaches through to the side of the cranium and passes through the zygomatic arch, which is here shown kind of in cross-section in red. And then it attaches down to the mandible and allows us to chew. We’ve got these sort of small, little muscles that we can deduce by the size of the zygomatic arch.
We also have a mutation in the myosin component of the jaw muscle that effectively limits bite force. So for humans, chewing and chewing power is reduced compared to chimpanzees.
We also have relatively smaller guts than chimpanzees do. About 60% of the size that you would expect for a primate of our body size. And one way of visualizing this is by looking at the shape of the thoracic cavity, the ribcage. So, in chimpanzees, which you can see here in blue, chimps have got sort of an inverted, funnel-shaped ribcage that sits atop a relatively wide pelvis. And what this essentially creates is a large space for intestines and so chimps have a stocky body full of intestines. Whereas in human, we’ve got a more tapered, barrel-shaped ribcage that sits atop a narrower pelvis. And this actually creates the human waist which obviously creates less space to hold a bunch of intestines for digestion.
But it’s not the size that matters. It’s also the proportions of the intestines that we can compare between human and chimpanzees that matter. And so if you take the chimpanzee gut and the human gut and you scale them to the same size, you can get a sense for what I’m talking about here. And what you can see is that, compared to the chimpanzee gut, the human gut actually has very long, small intestines which contain all the nutritive absorptive surfaces. And we actually have a relatively short and relatively smooth colon which suggests that we don’t have a lot of capacity for fermenting foods that aren’t easily absorbed. We just can’t fit it in there. And so jointly this kind of anatomy suggests that humans are specialized for foods that are relatively easy to digest.
So, jointly, we’ve got this high daily energy budget and we’ve got low digestive capacity and this actually suggests two things about the human dietary niche that are rooted in our ancestry. The first is that we require foods that are rich in calories, a lot of calories in a limited amount of space. And we require foods that are easy to digest. And I’m going to argue here that by externalizing part of the digestive process, food processing allows foods to meet these needs.
So let’s consider for a second how the average American meets their caloric needs. We get on average about 50% of our calories from carbohydrates, primarily in the form of starch. We get about 12% of our calories from protein and the remaining 38% of our calories from fat. Of course, if you drink alcohol, there’s additional calories that are added into this mix, but on average, that’s what it looks like.
Food processing has differential effects on each of these macronutrients that I thought it’d be useful to just briefly review to give you an understanding of how food processing is shaping these different macronutrients.
So, first, as I mentioned, the majority of our carbohydrates come in in the form of starch. And heat has a very well-known effect on starch. Essentially, starch consists of long chains of glucose all strung together. And we’ve got enzymes called amylases that are specific to starch that come in and cleave off di- and trisaccharides. So little two glucose units and little three glucose units that our bodies can then break down with other enzymes and absorb.
But the problem is that amylases can’t always get to the starch string in the places they need to cleave. And that’s because, when foods are served raw, what you can see here on the upper right, starch exists in these tightly packed granules. And in this form, amylases cannot penetrate the starch granule in order to get at the cleavage sites that can start cleaving off those sugars that we can absorb.
But when starch has been cooked, that granule essentially wiggles around and it swells and it takes up water and what you can see is this exploded mass down here in the lower right. And this is what’s called, this is a process called gelatinization. And in this form it’s very easy for amylases to come in and cleave off those di- and trisaccharides. And that’s one of the reasons that, if you’ve ever tasted a raw potato next to a cooked potato, the cooked potato tastes relatively sweet in your mouth and that’s literally because salivary amylases are starting to create sugar while the potato is in your mouth, but that doesn’t happen when the potato is raw because it’s still in granule form.
So, when we think about protein, a lot of us will get the majority of our protein through animal products, particularly meat. If we think about the structure of meat, meat essentially is muscle-fiber bundles that are surrounded by a collagen matrix. And what heat basically does is it causes the muscle fibers to toughen and shrink along the grain. But at the same time, it gelatinizes the collagen matrix. And this sort of toughening of the muscle-fiber bundles but the gelatinization of the collagen means that it’s easier for cracks to propagate through the tissue. That’s why it’s easier to chew through cooked meat than it is to chew through raw meat. It also makes mechanical digestion in the stomach more efficient when you’ve got easy cracks propagating through the tissue.
The other thing that goes on is that protein in nature exists in these tightly wound bundles. It’s almost like a massive of yarn, tangled yarn. And what heat does is it causes that ball of yarn that is the protein to unwind. And that allows protein-degrading enzymes called proteases to come in and get access to the cleavage points that again create the peptides that our bodies can absorb.
Now, for a really long time, it was thought that heat didn’t have any impact on fact and this is because by and large, very little fat appears in feces, so the assumption was always that, well, we must be absorbing 100% of it so it doesn’t really matter it’s form, it’s all going to use by the body.
But we now know that that’s just not the case. And a key reason is that fat, when it comes in in a food, is often encased in other structures that depend on cooking. And so, for example, what you can see here—this is an example of peanuts with raw peanuts and cooked peanuts in the lower panel. And the fat globules have been stained in red. And, in the raw form, what you can see are these little fat globules that are encased within intact cell walls. Whereas in the cooked you can see the cell walls have broken open because the polysaccharides that keep those cells nice and intact have broken open. And the other thing that’s happened is that the fat globules themselves, the reason they’re round, is you’ve got fat that’s surrounded by this layer of protein called the oleosin layer. And, as we just talked about, heat degrades protein, causing them to unwind. And, for the same reason, when you heat this oleosin layer, it unwinds and that allows fat-degrading enzymes like lipases to come in and to be able to metabolize that fat.
And so jointly, these effects of heat on starch and protein and fat can be expected to lead to important increases in nutrition digestibility. But what about non-thermal processing, right? What I’ve shown you so far is all about heat.
Now, non-thermal processing can only modify the physical structure. It can’t change these compounds chemically, so it can’t, for example, gelatinize starch or it can’t denature protein. It can essentially deal, though, with particle size and we’ll consider this for just a second.
So, for example, this is an example of starch digestibility given cooking versus milling. And what you can see in the white bars is that the fraction of starch that is digested quickly rises to close to 100% when that starch is cooked. So, the white bars are cooked, the green are raw. And that’s true whether that starch has been coarse-milled or fine-milled.
But when the starch is raw, it lags at every point and, in the coarse-milled case, even after 24 hours, the digestibility of that raw starch never approximates that of cooking. And in the fine-milled case, it really only approximates the digestibility of cooked starch at 24 hours and, if you think about passive rate of starch through the gut, it’s really only in the small intestine for 4 to 6 hours. So this is actually never achieved. These levels are never achieved in real life.
So, the effect of cooking and non-thermal processing on starch and protein and lipid allow us to make two predictions. And the first is that the consumption of foods processed by these methods should lead to increased energy gain. And the second is that the energy gain conferred by cooking should exceed the energy gain conferred by non-thermal processing.
So, those are pretty basic questions and you would think, gosh, we’ve all been eating forever. We’ve been cooking, we’ve been processing our food, everybody does it. Surely somebody would have researched this and this would be known. And this is like my nightmare that one day I’m going to just find the text that overrides all of my research. But we haven’t found it yet. Surprisingly, no one has actually studied the energy gain due to non-thermal and thermal food processing. And so we decided to set up a really simple experiment.
And in our experiment, we selected two foods for testing. We selected sweet potato and we selected lean beef. And the reason we selected these foods is partly because they have very different macronutrient profiles so that we could see how food processing was affecting all of these nutrients. But we also selected it because I’m a human evolutionary biologist and these are examples of two food classes that were though to provide the bulk of calories for ancestral humans, so it gave us perspective into how the adoption of food processing techniques may have enhanced energy gains for ancestral humans.
And we took these sweet potato and lean beef and we processed them in four ways. Either what we call NP—Not Processed. We pounded it up. We cooked it. Or we both cooked and pounded it. And we fed these treatments to mice for a period of just four days to see what would happen with the energy metabolism of those mice. And I can attest that these mice actually were really happy about the experiment and they enjoyed eating both the sweet potato and the beef compared to their normal chow. Which was nice.
And what you’re looking at here are the changes in body mass that these mice experienced over just the period of four days, and this is controlled for differences in activity as well as slight differences in food intake across these treatments. And what you can see is that, on the sweet potato diet, when mice were eating the raw, not processed sweet potato, they lost about 4 grams. When they ate the pounded sweet potato, they lost about 3 grams. But when they ate either of the cooked sweet potato, whether it was whole or whether it was pounded, they were able to maintain their body masses just fine, no problem at all. And they actually did so despite eating less overall sweet potato.
Now, the story was similar in meat. In the case of meat, when mice were placed on the meat diets, they actually all lost weight and this was to be expected because mammalian omnivores actually don’t deal with lean meat very efficiently in terms of our metabolism. But what’s important is that, on the raw treatments, mice lost more weight than they did on the cooked and there was really no difference between whether those were served whole or pounded.
A recent experiment by one of my students, Emily [Grupman], showed that this result also holds for fat-rich foods. So, what Emily did is to feed mice unprocessed, blended, cooked or blended and cooked peanuts. And they were also happy about this experiment.
The mice that received the cooked peanuts, on average, gained more weight. This is an illustration here, the mice didn’t actually gain this much weight over the period of time. This was to visually help you understand what we did. But the mice who ate the cooked peanuts did in fact gain more weight per gram of food eaten. And this was primarily due to differences in the absorption of fat. So, when mice were eating raw peanuts, regardless of whether they were whole or blended, about 11% of the fat in the peanut was coming out the other end.
But when they were eating the cooked peanut, the amount of fat coming out the other was about 8%, rather than 11. And this difference in digestibility, the fact that the mice were able to retain more on the cooked diet really led to these long-term differences in net energy gain and changes in body mass.
So we see across and diverse foods—right, sweet potato and lean beef and peanuts—processing increases energy gain and that cooking does to a greater extent than non-thermal processing.
And so we can say that processed foods meet our requirement for a diet rich in calories—right? But what about a diet that’s easy to digest? What you’re looking at here is a phenomenon called diet-induced thermogenesis. And I’m sure you’ve experienced it yourself. If you eat meat and if you’ve ever eaten a large steak or a large burger and after you’ve eaten you feel really tired and you feel really warm and maybe you sweat, you are not making it up. Your body’s working really, really hard to break down that food, to produce all the acids and enzymes that create the digestive juices that break down the food. You’re conducing peristalsis to squish that food down the gastrointestinal tract. You’re absorbing, you’re assimilating all of these nutrients. And that is costly.
The cost of diet-induced thermogenesis actually differs by different macronutrient. So, for carbohydrates, it’s about 5-to-10% of the caloric value of that carbohydrate is essentially used in its own digestion. If you eat a protein-rich food, that’s a whopping 20-to-30% of the caloric value essentially is spent in its own digestion. For fat, the costs are very low, zero to 3%. And the good news is, for those of you who are drinking tonight, the cost of digesting alcohol are also quite high, so actually you’re saving on calories purely by spending some energy digesting your drinks. So, you go.
Over the course of a day, this is essentially what diet-induced thermogenesis does to the metabolic rate. So what you’re looking at is metabolic rate here on the Y axis and time of day on the X. And what you can see is in the early morning hours, the metabolic rate actually drops below resting levels, because when you sleep, your metabolic rate’s very low. And then, with the first blue arrow—Oh, you have breakfast and your metabolic rate rises. And you start to digest it and assimilate those nutrients and the metabolic rate starts to come back down. And then you have lunch and your metabolic rate bounces up again. Digest. Have dinner. Bounces up again. It’s really only once you’ve digested your evening meal that your metabolic rate drops back down to resting levels.
And, on the whole, this phenomenon of diet-induced thermogenesis accounts for about 10 to 15 percent of your daily energy expenditure. It’s an amount similar to locomotion or physical activity. All the moving about that you do every day is matched in the calorie expenditure by just you sitting there digesting your food.
And you would think that because food processing externalizes part of the digestive process, it should actually make your foods easier to digest. Right? It’s not rocket science. But nobody had actually shown this before. So we decided to do it. And we chose to do it in pythons—admittedly, a weird choice, but they’re actually perfect animal models for this system. They don’t move, so you don’t have to worry about energy expenditure due to physical activity. You can control their body temperature because they’re cold-blooded. They basically only eat once a month so you can be sure that you have a baseline where they’re not still digesting their prior meal. And, when they do eat, they eat an enormous meal, which gives you a nice huge peek that you can measure. That’s why we chose pythons.
And we fed them lean beef that was either not processed, ground, cooked, or both ground and cooked. And we compared the diet-induced thermogenesis. And what we found is that ground meat led to 12% less cost of digestion, cooked meat led to 13% less cost of digestion, and both ground and cooked meat, there was almost an additive effect, where it was 23% less costly to digest compared to the unprocessed meat treatment.
Other researchers have subsequently shown something similar, even for degrees of highly processed food. So, eating a white bread and processed cheese sandwich resulted in about a 40% less cost of digestion compared to eating a whole wheat bread and I think it was a maybe cheddar cheese…? It was some sort of artisanal cheese. And they nevertheless were able to measure a difference in diet-induced thermogenesis.
So, the evidence does suggest that processing makes foods both richer in calories and easier to digest. And so one simple reason why food processing is so universal among human cultures is that it renders food suitable for human consumption. And, for most of human evolution, maximizing energy gain from food and minimizing the cost that went into digesting that food would have been advantageous, since suitable foods were so hard to come by. And so processing food would have actually given ancestral humans a competitive advantage, both in terms of survival but then also in terms of reproductive success.
But of course today we live in a really different environment and we’ve got, you know, I don’t know if you have it here in New York, but we’ve got Prime Now where you can get all of your Whole Foods groceries delivered to your door, there’s no such thing as foraging, really, anymore. We don’t really spend a lot of energy to go collect these foods. We don’t really have meaningful seasonality in the food supply. And by and large, a lot of our foods are processed to such an extent that they almost rarely resemble the ingredients from which they’re actually prepared.
And this means that some of the energetic advantages that ancestral humans gained by processing are today likely to be disadvantages.
And so how might it threaten our health today? I don’t have to be depressing and sobering, but it is worth reminding that we have an obesity epidemic in this country. I’ve just put up data for the last 30 years, comparing 1990, 2000, 2010. There’ll be a new survey coming out next year, which I’m sure will be even worse. But if we just look at this kind of 20-year time span, what we can see is that, in 1990, there was no state in the country where the rate of obesity reached 15%. So, every state had a rate of obesity less than 15%. And within a 20-year span, we reached the state where there was no state where the rate of obesity was less than 20% over a 20-year span. And today we’ve got 2 of 3 U.S. adults being overweight. We’ve got about 35% of adults and 15% of children being clinically obese. And this is a diagnosis that contributes to at least 400,000 deaths per year and something like $100 billion drain on the healthcare system. So, obviously, this is a problem.
And, ultimately, obesity is a problem of too many calories and not enough calories out. We all know that. But rather than me standing here and tell you what you should eat or how you should exercise, I wanted to highlight something that is less well-publicized, and that’s the main tool that we all have for managing caloric intake. The food label. It’s actually not a very good tool. It’s woefully inadequate in reporting the number of calories that our bodies are actually gaining from our diet. And I’ll illustrate what I mean here.
So, if we take the three foods for which I’ve already presented data showing that there is a net energy gain associated with processing—we’ve got lean beef, sweet potato and peanuts. And if we look up the food labels for these items served raw and cooked, we get this.
So, 100 grams of lean beef served raw gets you 173 calories, 100 grams served cooked gives you 212. For sweet potato, raw, 100 grams gives you 86 calories, cooked gives you 90. And for peanuts, raw gives you 567 and cooked it gives you 587.
So you’re saying to yourself, what is she talking about? We’re reporting that the cooked item has more calories, isn’t this what the whole point was about? She’s crazy.
But actually, no. Well, yes, maybe, but not at least for this point. And the reason is that 100 grams, so much of food is water and 100 grams of items that have been cooked actually have lost a lot of that water so there’s more food in 100 grams of cooked food than there is in 100 grams of raw food. And so if we actually scale these values on a dry matter basis where we just look at how are these labels actually capturing the caloric gains for the stuff that’s not water? Here’s what we see instead.
That, for example, with meat 100 grams of dry matter gives you 550 calories if raw, 549 if cooked. Sweet potato, 379 if raw, 372 if cooked. Peanuts, 606 if raw, 598 if cooked. So, somehow these food labels are not capturing the caloric benefits that we know come with cooking.
And why is this? Helps to understand a little bit about digestion. I promise not to quiz you on this. But essentially if we just consider what goes on—let’s say, imagine you eat this cashew. So you eat a cashew. You chew it in your mouth. You’ve got all this saliva being produced and salivary amylase that begins that process of starch digestion. At some point this all comes together and forms what we call a bolus. That bolus is swallowed. It passes down your esophagus aided by the muscular contractions that we call peristalsis. Goes into the stomach. In the stomach you’ve got squeezing going on that’s mechanically breaking down your food. You’ve got gastric acids and enzymes that are basically starting the process of protein digestion. It basically turns your food into a slurry called [kimes]—really gross, really mushy. That passes out of the stomach and into the small intestine where secretions from the pancreas and liver start to break down carbohydrates and emulsify fats and continue to process the protein digestion. And all the stuff that you can possibly absorb gets absorbed.
But some of that cannot be absorbed by the end of the small intestine. And what can’t passes into the colon where you can’t really do much with it anymore, but your microbial community takes over. It ferments that food. It can produce short-chain fatty acids, some of which we use as energetic substrates. And then you are sucking out the water, you’re sucking out additional minerals and, finally, you’re getting rid of what’s left over. That’s a lot of stuff that happens between Point A and Point B.
But the food labels that we’ve got today only look at what goes in and what comes out. It actually ignores everything that happens in-between. And that’s kind of a problem. And I’ll explain why.
So, what it measures is what we call total tract digestibility. What goes in minus what comes out. We assume that everything that has disappeared between Point A and Point B has gone to us. What we know it doesn’t include are diet-induced thermogenesis, which we just talked about. And, as I’ve already shown you, diet-induced thermogenesis can be really different based on different macronutrients. So it’s not just that, oh, we haven’t accounted for diet-induced thermogenesis, we know that all the calories are wrong by 10%. No, it’s variably wrong between different foods depending on that macronutrient composition, but also depending on whether that food has been processed.
We also know that food labels don’t take into account a property we called ileal digestibility and this just means we need to know where in the gastrointestinal tract things were absorbed. And I’ll give you an example.
So, this is my poor drawing of the gut. It’s a simplified version of the small intestine followed by the colon. And these little red, blue and green dots—those are nutrients coming in and some of them get absorbed and there’s lingering amounts that then pass into the colon.
Now, if you absorb nutrients in the small intestine, you get the kilo-calorie per gram values that probably all of you are familiar with. Nine kilo-calories per gram for fat, 4 kilo-calories per gram for either carbohydrate or protein.
But, here’s what. So, the nutrients that are not absorbed in the small intestine enter the colon where, as I’ve mentioned before, microbes ferment these nutrients and they produce short-chain fatty acids, some of which we can absorb and utilize for energy gain. But these short-chain fatty acids, if we can use them for energy gain, are only worth 2 kilo-calories per gram to us. And so a gram of carbohydrate absorbed in the small intestine is worth 4, but if that gram of carbohydrate is actually metabolized by microbes, it’s worth 2. Which is a pretty big difference. And this happens all the time. Our microbial communities are very active digesting our food and food labels are not accounting for this differential.
It’s worth thinking about a good example of a high digestibility food, or one that we think of as perhaps perfectly digestible, and that’s egg. You’ve got all these weight lifters or whatever downing the raw egg because they think that the protein is highly bio-available. And that’s because food labels suggest that it is. Because when you eat a raw egg or you eat a cooked egg, almost none of that protein comes out the other end, Point B. But, we haven’t accounted for where that protein has been digested. So, there was a study done by a team of Belgian scientists led by Evenepoel, that actually developed an ingenious technique where they isotopically labeled eggs by feeding chickens an isotopically-labeled diet and they were therefore able to track how proteins were being dispensed along the course of the gastrointestinal tract.
And what they found is that cooked egg, most of it was absorbed and only 6 to 9 percent came out the other end of the small intestine undigested. But if that egg was served raw, about 39 to 45 percent of that protein exited the small intestine and entered the colon. So, in other words, even though you basically get a full disappearance of egg protein from Point A to Point B along the gastrointestinal tract, if that protein is consumed raw, it’s your microbes that are benefiting from the energy, not you. And so cooking increased ileal digestibility, what we consider, by 45 to 78 percent, which is a pretty big difference.
The other thing that food labels do not account for is something we haven’t actually quantified very well so far, but it’s something my lab is working on. And these are host-microbial interactions in energy gain. So I assume that many of you, being interested in science, know at this point that the human body is an ecosystem. There’s about 150-to-one ratio of bacterial to human genes in the human body, so from a genetic perspective, we are far in the minority. There’s about a 1-to-1 relationship of bacterial cells to human cells in the human body. So we’re about half-human if you count us by cell number.
And there are different communities all over the body. Basically, every human surface that contacts the environment, including things like the lungs or the vaginal tract—anything that has exposure to the environment is fully colonized by microbes. But by far the most dense and rich and complex community lives in the human gut, where there’s an estimated 100 trillion organisms. And I knew we were going to be in the Hall of the Universe, so I had to come up with some sort of astronomical analogy. And, basically, this is a thousand-fold larger than the number of stars in our galaxy. That’s a lot of bugs that live in each of our guts.
And we know that this gut microbial community is linked to so many different aspects of human physiology, from digestion to immune function—now there are even links to social behavior. But one of the best worked out systems is how this microbial community interacts with energy metabolism. And we now have very strong data linking changes in the composition and function of the gut microbial community to basically all of the major metabolic disorders—arthrosclerosis, Type II diabetes, non-alcoholic fatty liver disease, low-grade inflammation, and obesity.
And in most cases, we’ve moved beyond simple associations by saying, okay, the microbial communities of lean and obese individuals look different to actually doing work showing that the differences in the microbial community are contributing to the differences in the host phenotype. Meaning, contributing to that leanness or contributing to that obesity. And we can do this through fecal transplant experiments where we take microbial communities, say from a lean and obese individual, put them into germ-free animals and we recapitulate the phenotype of the donor purely by transferring the microbial community. And this—the first time I saw this, it blew my mind. We do this in the lab all the time now, it still blows my mind.
And since the composition we know now of the microbial community impacts energy metabolism, it is worth considering how food processing is going to be impacting the composition of that microbial community, as well as its function.
And we know just from a pure ecological theory basis that the gut microbiome should be very sensitive to digestibility. And if just think from the perspective of a microbe for a second—you know, you’re sort of swimming in the colon. I mean, the microbial community actually goes all along the GI tract, but the majority is in the colon.
And if you just think, like any other ecosystem, when nutrients flood into a system, they’re going to favor the species that can actually make use of those nutrients and they’re going to be biased against those species that cannot make use of that influx of nutrients. This is true in any ecosystem.
So if you imagine within the colon you have changes in digestibility, some microbes will benefit, other microbes will benefit less. This will actually change the balance of the different microbial taxa and how they compete with one another.
And so the key concepts here are that the nutrients reaching the colon should select for microbes that are capable of the metabolism of those nutrients. And, if that’s the case, if it’s now selected for a microbial community that is good at metabolizing these nutrients, that should actually mean that the microbial community returns a greater fraction of the energy it [sees 45:16] back to the host because it’ll be able to utilize these nutrients to a greater extent.
Excuse me. Yay. Okay.
So, what we wanted to do is we set up a study that tested whether digestibility impacts the gut microbial community and to this we designed custom diets that contained 50% starch, but the starch was in one of two forms. It was either highly digestible in the small intestine or was fairly resistant to digestion in the small intestine. But the actual macronutrient composition of the two diets was exactly the same. We fed these to mice and we measured their microbial communities.
What I’m going to show you here is what we call a principal coordinate plot. All it is, and the only thing you have to understand is that dots that are closer together mean microbial communities that are more similar to one another. Dots that are further apart mean microbial communities that show more distance or differences to one another.
And here’s what we observed. So, among the mice fed the high digestibility diet shown in green versus the low digestibility diet shown in yellow, we see basically a perfect separation of those microbial communities based on the diet fed to the mice. And these open circles reflect samples collected after 24 hours. So these changes in the microbial community happen immediately and, within a single day, you already see separation on the basis of diet.
And just to be sure that we didn’t have some bias between the mice that got assigned to one diet versus the mice that got assigned to another diet, we repeated the entire experiment with germ-free mice that were all colonized from a pool of two donors. So we mixed together two microbial communities and we gave that microbial community to all the mice and then we fed them these two different diets and we basically saw exactly the same thing.
So, what you can see here in the [grey of 47:11] dots are the donor communities as well as the inoculum, and within the first 24 hours that you can deduce from the open circles, you see that the microbial communities within the first day largely resemble those of the donors and then they separated out again according to diet.
We then wanted to ask, okay, well, we showed this with these custom trials that had kind of engineered starch. Does this actually happen in a real food? Where, if we cook the food, we know we’re changing the digestibility of that starch, the allele digestibility of that starch because of starch gelatinization. Can we pick up the same effects by feeding raw versus cooked starch-rich sweet potatoes, for example?
So what we did is we prepared raw sweet potatoes, cooked sweet potatoes, or cooked sweet potatoes that we served in a restricted ration and we did this because, as you recall, on a cooked diet, mice can maintain their body mass. On a raw diet, they lose body mass. So our restricted portion allowed us to control and separate out the effects of weight loss versus those of eating a raw versus a cooked diet.
So we fed these diets to mice, measured their microbial communities, and looked at both the composition of those communities and how the functions change. And, again, this is the principal coordinate plot summarizing how composition looked of mice that were fed raw versus cooked sweet potato.
And what you can see is perfect separation in the microbial communities depending on whether those diets—the sweet potato—was fed raw or cooked. And you can see that, in the dark green versus light green circles, which represent the cooked ad libitum versus the cooked restricted ration, this was not a story about weight loss because the mice getting the restricted diet were actually losing just the same amount of weight as those on the raw diet. This was about whether the food was served raw versus cooked.
We also looked at microbial transcription. So this is gene expression in the microbial community itself and we wanted to know, does that look different, as well? So is it not just the different species that we’re affecting, but are we actually changing the behavior of those microbes, as well? And I’m just showing you a little tree plot that shows that, in terms of the microbial function, there was perfect separation, again, between microbial communities that were seeing a raw diet versus those that were seeing a cooked diet.
And, interestingly, one of the key distinguishing features on the raw diet is that we saw more expression of microbial beta amylase. So this is the same starch-digesting enzyme that we have, it’s just the microbial version. And it’s consistent with the idea that, on a raw diet, more starch is reaching the colon because we can’t absorb it as well. And that this would then select for behaviors in microbes that process that starch.
We next wanted to know whether the changes in the gut microbial community composition and function that we saw actually had any energetic consequences for the host. So, to do this, we took mice and we fed them either raw versus cooked sweet potato. And we took their microbial communities out and we transplanted them into germ-free mice in one of two forms. We either took that microbial community and we put it in live into the recipient or we autoclaved it—we basically fried the microbial community and then we put it into the recipient to create a recipient that was still germ-free but received all the same stuff that was in that inoculae. And so obviously the mice that received the live treatments from the raw-fed donor and the cooked-fed donor were then colonized. And the ones that received the killed microbes remained germ-free. And we fed all of these recipients a chow diet.
So what this design allowed us to do was to separate out the effects of preconditioning a microbial community on the host phenotype in the recipients that received these microbial communities. And recall that these recipients never saw a different diet, they were all just eating chow the entire time.
And yet what we observed were changes in body mass and body fat that suggested there were these really interesting host-microbial interactions. So what you can see here is, in the gray bars, these are the mice that remained germ-free. Changes in body mass on the top panel. Changes in body fat or just body fat in the lower panel. And both of the donor groups that received live inoculum gained more weight and gained more body fat compared to those that remained germ-free because now they’ve got a functioning microbial community that’s helping them rescue a little bit of the food that they couldn’t digest and produce those sort-chain fatty acids that we talked about.
But, interestingly, recipients of the raw-conditioned gut microbial community gained a lot more body mass and a lot more body fat compared to either those that remained germ-free or those that got the cooked-condition community.
And so the model that is emerging is this. On a raw diet, you’ve got less energy gained directly by the host. Right, we talked about the reasons that might be. That means that more nutrients are available for fermentation, because if you’re not absorbing it in the small intestine, it’s making its way into the colon. This ecological shift is altering gut microbial structure and function. And leading to a greater fraction of those indigestible nutrients being returned by the microbiome to the host. And this creates this really interesting partnership between humans and our microbial communities where, when we can absorb less of our diet, we’re actually selecting for a microbial community that returns a greater fraction of the energy it sees back to us. And so there’s this very interesting dynamic that, when we think about the average food labels, we’re not capturing at all.
And so, on the whole, by failing to capture diet-induced thermogenesis, the distinction between digestion in the small intestine and the colon, and host-microbial interactions and energy gain, all of which we’ve now seen are likely to be influenced by food processing, our standard energy assays fall pretty short. And, while we may be some ways away from being able to really develop rules around how we predict host-microbial interactions, we know enough at this point about diet-induced thermogenesis and the effects on ileal digestibility to improve these models.
And so, without tools that capture the energetic effect of food processing in a world full of processed foods, it’s going to be difficult indeed for consumers to manage their caloric intake, even if they really try.
And I just wanted to point out, as my final slide, that March happens to be National Nutrition Month. And I hope that our journey tonight just gave you a little bit of a new perspective on why we eat the way we do. And, to sum up, I think it’s quite simple—we eat processed foods because we can. And because for most of our history, the increased energy from these processed foods gave us an advantage in terms of survival and in terms of reproduction. And, although we may not appreciate or benefit from these energetic advantages today in the modern world, I think we can take the reins of our energetic legacy just by keeping in mind that it’s not just the food that matters, but it’s also the form of the food that matters.
And, with that, I just want to acknowledge, again, the American Museum of Natural History and the Leakey Foundation for sponsoring this. I want to thank my two main mentors, Richard Wrangham and Peter Turnbough for really inspiring a lot of this work, as well as some of the other collaborators that have directly contributed to the data that I showed you tonight, my lab and, of course, other funding agencies for their support. And thank you all for your attention and I’m happy to open the floor for discussion if there’s time.
[applause]
MODERATOR: Thank you, Rachel. I wanted to start with a Twitter question and then there’ll be two roving mics. The first question comes from Twitter asking if it’s better to eat a 3-ounce hamburger with a glass of beer than it is to eat a 3 ounce steak with a glass of milk? Sorry, better in the terms of getting more calories out of…
RACHEL CARMODY: So, better in terms of getting more calories out of it as opposed to getting fewer? Again, we have to always keep in mind that what we consider to be a benefit can be interpreted quite differently if we’re thinking about ancestral humans or even modern humans that are living in traditional societies versus industrialized societies, where often our problem is that we get too many calories.
But it’s definitely the case that, if your choice is between a steak and a burger, you’re going to get more calories, net calories out of the burger, in part because a greater fraction of that burger will be digested and absorbed and in part because you’re going to expend less energy to break down the burger than you will the steak.
In terms of the beer versus—what was the other option? Was milk?
Yeah, well, it’s really interesting, because they both have high cost of digestion and probably similar caloric content. And, as I mentioned before, protein and alcohol are actually quite similar in their metabolic costs of digestion. I think we probably have to test that. I think it’s almost a wash on the drink, but there’s no question that you get more calories out of the burger.
MODERATOR: All right, we have another question over here.
AUDIENCE QUESTION: Actually, I have like two questions in one. So, if somebody were to only eat a raw, unprocessed diet, they could theoretically die from starvation because they’re not absorbing enough calories? But also, when you cook foods, some of the nutrients are—they disintegrate from the heat, like Vitamin C. So there’s this like damned if you do and damned if you don’t?
RACHEL CARMODY: No, that’s a really good question, thank you for asking. So I’ll answer the first part first and actually it’s surprisingly, I have … I had a whole bunch of stuff to talk to you about that I didn’t talk to you about. But I can show you—this is the chart that answers the raw food question.
So, based on what I showed you, it is the case that consuming a raw diet you’re going to extract lesser amount of calories and you’re going to spend more energy to digest that food, so the net gain is lower. And it is the case among humans who, for a variety of philosophical and … really, philosophical reasons, decide to eat raw and go back to what they view as an ancestral diet, they do adopt a raw lifestyle. And what you find among raw foodists who eat raw food over long periods of time is they do lose body mass. And so what you can see here in the red is age-adjusted body mass index as a fraction of the percent of the diet eaten raw. So, if you’re eating, say, 70 to 80 percent of your diet raw, you’ve got a body mass index of around 21. If you’re eating 100% of that diet raw, you’ve got a body mass index that’s a couple of points lower and there’s this kind of very dose-response effect in the fraction of the diet eaten raw versus body mass index.
But what the blue line is telling you is—herein lies the rub. I didn’t present a whole bunch of data showing you that we are also adapted to this processed diet at this point. We’ve got high energy needs. We’ve got tiny digestive structures. And when we try to eat a diet that doesn’t match those biological parameters, we don’t fare very well. So what the blue line is showing you is the incidence of amenorrhea. So this is basically the cessation of ovarian cycling, the shutdown of the reproductive system in women who consume raw diets. And it’s widely known that human reproduction, particularly in females, is very, very sensitive to energy balance, which is why you have reproductive issues in, say, elite marathoners and people who are either very lean or expending lots of energy.
And what we see here is that, as the percentage of the diet that’s eaten raw increases, the rate of amenorrhea increases to the point that, if you’re on 100% raw diet, 50% of women of reproductive age are not reproductive. And, from a evolutionary perspective, this is a disaster. Right? This is not a sustainable strategy.
And this is one of the reasons that we actually think that humans at this point are committed to a diet that includes processed food. And not just processed food—cooked food. Because these raw foodists are—believe me, they’re processing their diet in every non-thermal way you could imagine. And they’re typically living in industrial societies where they’re not subject to seasonality, they are not expending lots of energy to get their food, they’re not really physically active to the same level that hunter-gatherers are. So this is [indiscernible 61:09] the best possible energetic scenario and even then we become non-reproductive. So we actually think that humans have adapted to this processed diet to the point where we’re committed at this point.
With regard to your other question about micronutrients and, say, vitamins, you’re absolutely right. So I focused my comments in the talk on energy gain, thinking purely about calories. But it is the case also that the human body depends on various micronutrients as well as other components like some of the omega fatty acids that are important for brain growth. And we still need to learn more about the net effects for the essential macronutrients in terms of how they fit into this picture. But certainly it’s the case for micronutrients that cooking, heating, can actually inactivate some of those and make them non-bioavailable. And that tends to be worse for cooking methods that, say, can leach these micronutrients out so, like, boiling, where vitamins can not only be lost through to degradation, but they can be lost through to the cooking medium. Thank you.
MODERATOR: On your left is the next question.
AUDIENCE QUESTION: There’s an awful lot of information to digest that you’ve given us. What combination of all the independent variables you have—raw versus cooked, fat, protein, carbos, yadda, yadda, yadda—what combination of all those independent variables gives, especially as you get older, like I am, gives you a lean figure? As opposed to being a fat old slob, like me?
RACHEL CARMODY: That’s a very good and very complex question and part of the reason that it’s complex is what we’re learning about these gut microbial communities and their sensitivity to diet.
Now, every single person in this room actually is going to have a different microbial community. It’s amazing how variable we all can be. And we can actually respond differently to the same diet because of those host-microbial interactions. So it’s a complicated problem for that reason.
But, if we kind of break down some of these independent variables, right? Generally speaking, if the goal—and this is the big “if”—so in the industrialized world, if the goal is to remain lean, which of course to a hunter-gatherer is not the goal, if the goal is to remain lean, I think we’ve learned that minimal processing is actually beneficial to a point. If you are, as you phrased the question, if you’re an older individual, if you’re post-reproductive, it may not matter to you if you’re a woman that you become amenorrhoeic, because you’re amenorrhoeic anyway. And so it is one known way of keeping a lean body mass.
We also, if we think about it from a diet-induced thermogenesis picture, the greater fraction of daily calories that come in from protein are going to cost more to digest than calories that come in from fat and carbohydrate. We have to be careful with the balance of those different macronutrients. So, as I mentioned, the with the data on mice, mice lost weight on these meat diets. We do the same thing. Our bodies are not designed to eat lean protein as its primary source of calories.
But if you can push the lean protein fraction to close to where the physiology becomes unsustainable, which is actually the basis of the ketogenic diet, you could in fact foster a lean phenotype that way.
So, thinking about processing aspect, less processing, particularly incorporating more raw items when it’s safe to do so—which of course is not the case with a lot of protein-rich items—would be beneficial. Eating foods that have larger particle sizes. So, not grinding everything down to where it’s unrecognizable but leaving foods in larger chunks to force your body to work harder at breaking it down would actually be beneficial. And then of course weighting across the different macronutrients to macronutrients that are more costly to digest, keeping in mind that humans cannot really sustainably get more than 50% of their calories from lean protein.
MODERATOR: That’s all the time we’ve got for tonight. So, please join me in thanking Dr. Rachel Carmody.
RACHEL CARMODY: Thank you.
A video version of this SciCafe will be available on the Museum’s YouTube channel on Saturday, March 23.
Subscribe to the [email protected] Podcast on iTunes, Soundcloud, or wherever you get your podcasts.