Show me the (BCAA) studies

You asked for it. Literally.

Tom Cruise in the blockbuster Isoleucine Restriction: Maverick

Protein restriction, or specifically BCAA restriction, or maybe even specifically isoleucine restriction, is all the rage right now.

Brad at Fire in a Bottle just lost 14.5lbs in 28 days doing it. He also lowered his fasting glucose by 20mg/dL.

My own diet, ex150, is also infamously low in protein. In fact, I called it “low-protein keto” in my own notes before I came up with the more neutral name “ex150.”

But a lot of people don’t think these anecdotes are enough. Show me the studies?!

So here we go. I recently started collecting all the interesting nutrition studies I’d been reading in a big file, and these are all the ones I’ve read on protein/BCAA restriction. A lot of the work is very recent, as in “came out last week” for one of them. And a lot was done by the Lamming Lab in Madison, WI. Very cool stuff.

A lot of the studies are in mice. I am aware that mice are not exactly identical to humans, but they’re fine for generating hypotheses. Especially if we already have anecdotes and an n=10 human trial showing protein restriction works.

So let me introduce you to the science on protein restriction.

Dietary Protein Restriction Improves Metabolic Dysfunction in Patients with Metabolic Syndrome in a Randomized, Controlled Trial

https://www.mdpi.com/2072-6643/14/13/2670

21 (human) patients with metabolic syndrome were either assigned to a Calorie Restriction (CR) group (by 25%) or an isocaloric (=they ate the same calories as before) Protein Restriction (PR) group. The trial ran for 27 days.

The CR group ate 25% of their calories from protein, the PR group only 10%. So not even a crazy protein restriction. Here are the results:

The CR group reduced weight by an average of 8%, the PR group by 6.6%. So the Protein Restriction group lost almost as much weight as a group restricting energy intake by 25%! Pretty nuts.

But here’s the kicker: they also tested energy expenditure before and after the trial was over.

The CR group had a markedly reduced metabolic rate after just 27 days of caloric restriction. The PR group did not!

What about blood markers? These are commonly used to identify metabolic syndrome.

Marker: PR / CR reduced by

Fasting glucose: 52.7% / 56.9%
HbA1c: 9.2mmol/mol / 17.5mmol/mol
Total Cholesterol: 35.4% / 34.4%
LDL: 38.6% / 36.9%
Triglycerides: 39.5% / 49.9%
C-reactive protein (an inflammation marker): 69.4% / 57%

Everything that you’d expect a “healthy weight loss diet to do” by restricting calories is also achieved by nearly the same amount without any restriction in calories whatsoever - merely by reducing protein intake from 20% to 10%.

Now 10% calories from protein isn’t even that low - for an adult man with 135lbs of lean mass, eating 3,000kcal/day, that’d be about 75g/day of protein, or probably just over 1g/kg of body weight at 20% body fat. This is well above the RDA of 0.8g/kg of body weight.

It’s just not insanely high protein, which is what we’re used to in the West.

And also notice these people weren’t eating a ketogenic or potato-only style diet: 30% calories from fat in both groups, and 50/60% respectively from carbs. They were straight up in the metabolic swamp.

Decreased Consumption of Branched-Chain Amino Acids Improves Metabolic Health

https://www.cell.com/cell-reports/fulltext/S2211-1247(16)30733-1

This 2016 study by the aforementioned Lamming Lab begins investigating what parts of protein restriction are mediating these benefits.

Confusingly, this study has both a human and a mouse component.

In the humans, they tested protein restriction and found it effective to lower weight and body fat, improve glucose control, and up-regulate FGF21. FGF21 is a hormone excreted by the liver that seems to be involved in satiety, energy metabolism, and tissue sugar intake.

In the mouse component of the study, they tested two branches: one was reduced in all amino acids (a general protein restriction) whereas the other group only restricted Branched Chain Amino Acids (BCAAs).

The 3 BCAAs (leucine, isoleucine, and valine) are essential amino acids that seem to have certain special properties which allow them to be used directly in energy metabolism, unlike other amino acids.

Both groups of mice experienced weight & fat loss as well as improved glucose control, but only the total protein restriction group saw improvements in FGF21.

Still, pretty impressive - they nailed down that BCAAs seem to independently mediate the weight and glucose control benefits previously seen in total protein restriction.

I think they do a lot of these studies in mice because it is difficult to get such fine-grained control over human diets. How would you feed humans a diet reduced only in BCAAs? You couldn’t feed them any natural foods, as those all have relatively similar amounts. Certainly most muscle meats have similar BCAA percentages, generally about 15% of protein. Dairy is usually at 20%.

I suppose you could feed humans a diet entirely based on protein powder, and create a special powder mix restricted in BCAAs. But good luck finding people to sign up for that study.

Of Mice..

This study also sees the appearance of the infamous 7%-protein mouse chow diet. Most studies by the Lamming Lab seem to use this 7% figure. They do mention they tried an extra-low protein diet (called ExLow AA lol.. sound familiar?!) with 5% protein, but the mice on this one lost lean mass in addition to fat mass, so they stuck to the 7% one.

The control mice ate 21% protein.

I’ll leave you with this amazing quote:

Despite consuming more food, mice on the 7% PR diet gained less weight than mice on the 21% protein control diet over the course of 2 months.

Body composition analysis suggested that while consumption of a low-protein diet slowed the gain of lean mass, fat mass accumulation was almost entirely blocked.

Mice eating the 7% PR diet had no change in spontaneous activity, but exhibited increased respiration throughout a 24-hr cycle and had increased energy expenditure at night

So the protein-restricted mice:

1. Ate more

2. Gained practically no fat

3. Didn’t have reduced activity levels

4. Burned more energy

Protein leverage me that!

..and BCAAs..

Wanting to know what exactly about protein restriction was the cause of the benefits, they designed two more diets: a leucine-restricted diet (leucine is one of the 3 BCAAs, and is implicated, in humans, in muscle growth), and a diet restricted in all 3 BCAAs. In both of these, the leucine/BCAAs were restricted by 2/3 respectively (21% protein reduced by 2/3 is 7%).

Interestingly, they found that the LowAA and LowBCAA groups ate significantly more, yet barely gained any fat, compared to the control high-protein mice, who got obese.

What’s interesting, too: the LowLeu group didn’t actually seem to get the benefits. They ate slightly less than than the control mice, and gained even more fat!

Confusingly, the study has this to say about the LowLeu group:

[..] we observed that mice fed a Low Leu diet showed a trend toward increased adipose mass and decreased lean mass; further, the mice appeared fatty upon necropsy.

Yet in the chart above, it looks to me like the LowLeu mice gained nearly as much lean mass at 10 weeks on the diet as the control mice?

In any case, this was a cool experiment. They tried protein restriction, it worked. They tried BCAA restriction, it worked. They tried leucine restriction - it didn’t work! Huh.

So restricting leucine, at least in mice, seems like a dead end. So presumably it’s one of the other 2 BCAAs, isoleucine or valine, right?

..and Men

In the human portion of the trial, they restricted 19 volunteers to 7-9% calories from protein, whereas another 19 volunteers ate their habitual diet. On average, these experiments lasted 43 days.

Interestingly, these “normal” diets weren’t that much higher in protein, only on average about 50% higher, which would come out to 10.5-13.5% protein.

In comparison, the mouse diets were 7% vs. 21% - a tripling of protein.

The protein restricted humans lost on average 1.37kg (3lbs) over the (on average) 43 days, whereas the control humans lost on average 0.36kg (0.8lb).

3lbs in 43 days isn’t exactly stunning weight loss, but it’s actually better than pretty much all studied diets out there. And, of course, 7-9% protein isn’t particularly low. The aforementioned RDA of 0.8g/kg body weight (again at 20%) is 7.6%. So only a few of these people were even below the RDA.

They also tested blood markers, on which the PR group generally did better. They improved their glucose by 7mg/dL, for example, whereas the control group got worse by 2.67mg/dL on average.

This might be an indicator that restricting protein in humans to “just” RDA levels is not enough for significant beneficial effects, though it seems to be slightly beneficial.

Restoration of metabolic health by decreased consumption of branched‐chain amino acids

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5813603/

This 2018 study is also by Lamming Lab. Having previously found that mice fed a reduced-protein or reduced-BCAA diet didn’t gain weight even on the notorious “high-fat diet,” they decided to test the next step: what if you fed it to mice who already had metabolic syndrome?

BCAA restriction could prevent obesity and impaired glucose tolerance, but could it reverse them?

Spoiler alert: yes.

We find that specifically reducing dietary BCAAs rapidly reverses diet‐induced obesity and improves glucoregulatory control in diet‐induced obese mice. Most dramatically, mice eating an otherwise unhealthy high‐calorie, high‐sugar Western diet with reduced levels of BCAAs lost weight and fat mass rapidly until regaining a normal weight.

Importantly, this normalization of weight was mediated not by caloric restriction or increased activity, but by increased energy expenditure, and was accompanied by a transient induction of the energy balance regulating hormone FGF21 (fibroblast growth factor 21).

Consumption of a Western diet reduced in BCAAs was also accompanied by a dramatic improvement in glucose tolerance and insulin resistance.

For the setup, they fed the mice the evil Western Diet for 12 weeks. There was also a control chow group.

Then, they switched them over to a series of diets: ControlAA, ExLowAA, ExLowBCAA, Western Diet, Western Diet + BCAA. Isn’t it hilarious that this lab came up with the same diet naming scheme as me? I swear I didn’t know about these.

Let’s look at the results. As usual, Lamming Lab has great graphs.

All groups except the control-chow group start obese. The mice that stay on the WD, and those that eat the WD + more BCAAs, get even fatter. The extra BCAAs slightly increased the rate of weight gain.

Mice started on the control chow - who were not obese to start with - and fat WD mice switched to control chow, pretty much converge on the exact same weight at the end of the study. Note that this is not much lower than the “obese” level all WD mice were at the beginning - going on the control chow caused the mice to go from 41g to 35g.

The mice on the BCAA and AA restricted diets, on the other hand, lost a ton of weight VERY rapidly, and at pretty much the exact same rate.

After only 2 weeks on these restricted diets, they’re already at 30g - lower than the control chow mice will end up at the end.

They even have a breakdown of fat vs. lean mass:

We can see that all mice (bar the control mice) started out around 15g fat mass, which, at 41g starting weight, gives them about a 36.5% body fat percentage.

The mice continuing on the WD and WD+BCAA diets will reach about 41% and 44% body fat respectively.

The mice starting on/switched to control chow all end up around a respectable 20% body fat - certainly an impressive loss.

Yet the protein/BCAA restriction yields around 15-16% body fat. The BCAA-restricted mice are slightly leaner than the protein-restricted mice, but it’s quite close.

We can also see that both the low-protein and low-BCAA groups lost some lean mass, wheres the control-chow and WD mice all pretty much had the same lean mass. Again, the BCAA-restricted mice seem to have lost slightly less than the protein-restricted mice.

Mice can break the laws of thermodynamics

But how did these low-protein and low-BCAA mice get so shredded? Surely they exercised more and ate less, right?

Of course not.

The researchers measured the mice’s food intake over 4 day periods, once at 2 weeks and once at 10 weeks. At 2 weeks, the ExLowBCAA mice already ate the most, followed by the ExLowAA mice and then the control chow mice.

At 10 weeks, the difference is stark. The ExLowAA mice are now in the lead, eating a whopping ~80% more (!) food than the WD and WD+BCAA mice. The ExLowBCAA mice are not far behind. What’s interesting here is that the mice switched from the WD to the control chow have caught up - they’re now on par with the mice that were on control the whole time.

This seems to indicate that food intake is regulated by the type of diet. Who knew!

When it comes to energy expenditure, mice, who are nocturnal, have very different movement patterns in the day vs at night. But the relations stayed the same: both control chow groups moved the most. The WD mice moved the least.

Another interesting finding here: the WD+BCAA mice actually move way more than the pure WD mice, closer to the protein/BCAA-restricted groups.

But if we look at energy expenditure, not “working out” by moving around, the picture is different. The ExLowAA, ExLowBCAA, and control chow mice all had very similar energy expenditure. But both the WD and WD+BCAA groups had much lower energy expenditure. The difference is quite drastic: the highest expenditure group (ExLowAA) had 50-70% higher energy expenditure than the lowest (WD+BCAA).

I think this is pretty nuts, but it matches my own anecdotal experience when talking with people about their TDEE (total daily energy expenditure).

A ton of gym bros run around on Twitter talking about how “nobody needs more than 2,000kcal” and how they, adult men exercising 3-5x/wk, get by fine on that amount.

Yet I can lose weight eating 2,300kcal, and stay weight stable at nearly 4,000kcal/day. During the ex150 trial, a 38 year old female participant lost 11lbs in 30 days while eating 2,700kcal/day.

How can we reconcile that? The answer is, of course, that your TDEE is not an inherent, set-in-stone number. It’s a function of many factors, including your diet. Nobody knew.

To create an energy deficit, eat 80% more?

The protein/BCAA restricted mice ate 80% more food, and their TDEE was 50-70% higher, although they did not move more than the control mice - and they lost a ton of fat, while the WD mice continued gaining fat.

Clearly, these mice did not get the “eat less and move more” memo.

The WD seems to make mice lethargic - they eat way less, they don’t move much, and their TDEEs tank. The protein/BCAA restricted mice are the opposite - they eat a ton, and they burn even more.

If only we knew of some food ingredient that reliably put mammals into some sort of “torpor.”

One more thing

The study also mentions a ton of blood work, but I just want to show you 2 of those: glucose and insulin.

This shows a glucose tolerance test (GTT). The WD mice are diabetic. Their glucose spikes 3x from baseline to 300mg/dL, stays up there for an entire hour, and doesn’t come back to baseline over the course of 2h, staying at 200mg/dL. The WD+BCAA mice are even worse!

On the contrary, the 2 control chow groups did relatively well. But even they don’t come back to baseline by the 2h mark. I’d call them pre-diabetic by Kraft test standards.

But, again, the BCAA and protein-restricted groups did significantly better. Their peak is below 200mg/dL, and it comes back way quicker than even the chow mice.

Now let’s look at their fasted insulin:

Mice who’d always been on control, ExLowAA, and ExLowBCAA mice all had roughly the same insulin response. The WD→control mice fared a tiny bit worse, but they’d only been switched over for 5 weeks at this point - maybe they would’ve caught up completely by the end.

The WD mice had incredibly high fasting insulin, more than double that of the healthy mice. And the WD+BCAA mice did even worse at about triple (!) the healthy insulin level.

Conclusion

This was actually only the first half of the study. In the second part, they KEPT the mice on the evil WD diet while restricting protein/BCAAs. The results were even more insane, but there are so many other studies to get to.. read it yourself if you’re interested.

The adverse metabolic effects of branched-chain amino acids are mediated by isoleucine and valine

https://www.sciencedirect.com/science/article/pii/S1550413121001662

This study is from 2021. It’s all in the title. Lamming Lab doesn’t mess around. These are many of the same researchers from the previous studies. Having tried leucine restriction and finding it wanting, they went for isoleucine(Ile)/valine(Val) this time, the other 2 BCAAs.

Again they did the weird human/mouse hybrid study. They tried several different diets in the mice to drill down on what mediates the metabolic benefits of protein restriction.

The human portion was just an analysis and correlation of blood isoleucine levels with BMI. I suppose that after their mouse isoleucine results came in, they just wanted to get a sanity check on humans and so they included it. Scientists! How do they work?

Step 1

First, they ran 4 diets on the mice: ControlAA (21% protein), LowAA, LowBCAA, LowNon-BCAA (restricting the other essential amino acids, but not the BCAAs). The “low” diets restrict the respective AAs by 2/3. All diets were isocaloric and isofat.

They found that LowAA and LowBCAA mice lost weight and had improved glucose control, but LowNon-BCAA did not help much. I think they were just trying to confirm that it was actually BCAAs here, and eliminate the possibility of the other EAAs playing a role.

Step 2

Now confident that they were honing in on the target, our brave researchers designed several more mouse diets to narrow down the root cause: LowLeu, LowIle, LowVal, one for each of the 3 BCAAs. Again, all reducing the respective ingredient by 2/3.

The isoleucine-restricted LowIle diet was the clear winner. The mice eating it were by far the leanest. Again, the LowLeu mice actually gained more weight than the control mice. LowVal performed just about as good as LowBCAA. The effects on blood glucose were similar, with LowIle outperforming everyone.

Surprised by their finding that LowLeu was not working, they even designed a diet that restricts leucine by 87%. This still didn’t bring any benefits. Leucine did nothing wrong.

Step 3

It was clearly isoleucine, right?

But no, our investigators wanted to find out even more. They designed a new series of diets called “addback” diets. In these, they took a baseline protein-restricted diet and then added back certain ingredients back to “normal” protein levels, namely: Leucine, isoleucine, valine, or all 3 BCAAs. These were named, respectively: LowAA+Leu, LowAA+Ile, LowAA+Val, LowAA+BCAA.

The addback diets confirmed what they’d found: adding back all BCAAs did nearly as bad as the control high-protein diet, adding back isoleucine was the next worst. Adding back valine did basically nothing, that performed the same as just having a generic protein-restricted (LowAA) diet. Adding back leucine was fine - because both isoleucine and valine were still restricted.

The researchers did a few more things that I’ll gloss over, like test mTOR activity caused by various BCAAs and testing for changes in gene expressions.

Coup de grace

Because apparently they hadn’t found out enough, the scientists decided to test the effects of their restrictive series of diets on mice eating the infamous Western Diet (spooky music!) or WD. This diet is known for making mice extremely obese very quickly. They started all mice off on this diet, i.e. the mice were all obese to begin with.

Holy canneloni. Despite continuing on the crazy Western Diet, restriction of all BCAAs, valine, and especially our favorite, isoleucine, caused the mice to drastically lose fat mass. Within 3 weeks, all of those 3 were leaner than the control-chow mice, who had never been obese to begin with.

And check that out: the restriction of BCAAs/Ile/Val also caused the mice to eat more - way more! The isoleucine-restricted mice ate nearly 2x as much as the control or Western Diet mice. Talk about a deficit - eat 2x as much, lose 2/3 of the weight!

Leucine restriction was once again useless - WD mice restricted in leucine never lost any weight.

Conclusion

Lamming Lab really hammered this one home. They drilled down and drilled down and then tried invalidating their candidate (isoleucine) from all sides.

Really cool to follow the researchers down this rabbit hole.

This study alone proves that you can apparently “break the laws of thermodynamics” and double your food intake, while dramatically losing weight, just by doing this One Weird Trick: restricting isoleucine.

A Novel Dietary Intervention Reduces Circulatory Branched-Chain Amino Acids by 50%: A Pilot Study of Relevance for Obesity and Diabetes

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7824725/

This one’s much shorter: in this 2021 study, the authors were interested in finding out if restricting the food intake of BCAAs actually reduces circulating BCAA levels in the blood of huumans.

Yes, it does.

Fun quote from the abstract:

Elevated circulating branched-chain amino acids (BCAAs; isoleucine, leucine, and valine) are associated with obesity and type 2 diabetes (T2D).

How come everybody in science seems to know this? This is pretty much the opposite of popular opinion on protein and obesity/health.

A Branched-Chain Amino Acid-Related Metabolic Signature that Differentiates Obese and Lean Humans and Contributes to Insulin Resistance

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3640280/

This study is an attempt in “metabolomic profiling” of obese versus lean people. Essentially, they were trying to broadly compare the metabolic profiles of obese and lean people, to glean an understanding of what makes people obese. They just tried to measure as many hormones, metabolites, and other values as possible.

They selected 74 obese (but disease-free, e.g. not diabetic) and 67 lean subjects.

What they found was that, not surprisingly, the obese people were more insulin resistant. But they also had a unique “metabolic signature” of branched-chain amino acids (BCAAs).

They discovered via food questionnaire (questionable of course as these are from the subjects’ memory, not objective measures) that the obese people ate slightly more fat, slightly less carbohydrates, and more protein.

After taking the subjects’ blood, they used statistical software to find the blood markers that showed the biggest differences between the obese and lean group. These included the BCAAs. The BCAAs also strongly correlated with insulin resistance, in both obese and lean people.

Testing the theory in rats

To find out if they were onto something, the researchers ran an experiment on rats. They use 3 diets: HF (high-fat), HF/BCAA (high-fat with added BCAAs), and SC (standard chow) as control.

The yellow line in the graph is a bit hard to see, but the HF rats gained the most weight, followed by the HF/BCAA and then the SC rats. Interestingly, this was apparently because, while the HF and SC rats at the same amount of food, the HF/BCAA rats ate significantly less.

The BCAA supplemented rats also had significantly higher levels of BCAAs in their blood, as expected per the “metabolomic signature” found in humans.

In panel C, we can see that although the HF/BCAA rats did not become obese, they still became insulin resistant and are basically diabetic.

These experiments demonstrate a contribution of BCAA to development of insulin resistance that is independent of body weight.

They repeated the study by supplementing the SC group with BCAAs, but the SC rats did not gain more body weight or become any more insulin resistant than their non-BCAA supplemented brethren. BCAAs seem to only help in the context of the “high-fat” diet. (Ominous foreshadowing!)

How do BCAAs mess with insulin?

The researchers suspected that mTOR (and a few other things) activation mediates the interference with insulin signaling. They found significant increase in mTOR for the BCAA-supplemented group vs. the other two groups.

The diets in detail

I always try to look at the specifics of the diets used. The studies often just say “high-fat diet” and somehow it’s never butter. And, of course, we see the same here.

HF (high-fat)

45% kcal from fat (12.3% soybean oil, 87.7% lard)
19% kcal from protein (53% casein, 47% bulk amino acids)
35% kcal from carbs (21% corn starch, 29% maltodextrin, 50% sucrose)

Dear Lord. This is just about the most nightmarish rat chow you can imagine. Soybean oil is 60% PUFA, lard is 20-30% PUFA for a total of about 11.2% kcals from PUFAs in this “high-fat” diet.

Casein is 20% BCAA and somehow I suspect it’s messed up in other ways, at least in rodent studies. Maybe rodents just really shouldn’t eat casein? Or maybe it’s cause they’re always eating soybean oil and lard.

Why do they never feed the rats butter or beef tallow and beef protein? Assuming that soybean oil and lard represent all “fat” and casein represents all “protein” seems like a massive leap.

Also, 45% kcal from fat isn’t exactly high-fat in my book. That’s not even a low-carb diet.

To summarize, the HF diet is 1. super high in PUFAs, 2. right in the middle of the swamp at 45% kcals from fat, and 35% from carbs, 3. already very high in BCAAs.

The HF/BCAA diet is almost exactly the same, except it’s 2% lower in fat and 1% lower in carbs, and in return 4% higher in protein. I don’t know why it adds up to 100% while the HF diet only adds up to 99%.

Now let’s control these 2 to the control diet:

SC (standard chow)

10% kcal from fat (55.5% soybean oil, 44.5% lard)
19% of kcal from protein (53% casein, 47% bulk amino acids)
71% of kcal from carbs (45% corn starch, 5% maltodextrin, 50% sucrose)

This diet sure seems a lot more low-fat than the other diet was low-carb. In addition, not sure why this one had 5% maltrodextrin vs. 29% in the previous one. Maybe they kept maltodextrin the same in absolute terms and added all the extra carbs via corn starch? I’m not sure if corn starch vs. maltodextrin makes a difference, but I wouldn’t be surprised.

In percent of fat, this diet is even higher in PUFAs than the HF ones, but of course it only has 10% total fat, so assuming 60% PUFA in the soybean oil and 20% in the lard, it comes out to about 4.2% kcals from PUFAs.

This alone could explain the difference. In seed oil circles it’s generally assumed that 2% kcal from linoleic acid is “ancestrally safe” and 8% is definitely too much. Not sure these numbers would be exactly the same in rats, but if they were, 11% would definitely be way above “bad” and 4% might just be close.

The difference is even a bit starker if we look at linoleic acid specifically, not overall PUFAs. Soybean oil is about 50% linoleic acid, whereas the lard in lab rat “high-fat” diets has actually been measured to be 14%, much higher than assumed.

This would give us 3.4% and 8.3% linoleic acid of calories, respectively. So the SC is pretty close to “known safe” and the HF diet is just over “known unsafe.”

I really just wish someone would reproduce these rodent studies with butter or beef tallow for fat, and maybe replace the casein, too.

Restricting Branched-Chain Amino Acids within a High-Fat Diet Prevents Obesity

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9030079/

This 2022 study builds on the “metabolic signature” of amino acids above, if done by different researchers.

The goal was to find out if the “metabolic BCAA signature” had any clinical potential, i.e. can we help people reverse obesity or metabolic syndrome by pulling on the BCAA intake lever?

First, the researchers performed a meta-analysis of nearly 3,400 people from six cohorts in Canada, Australia, and the U.S. and compared their “metabolomic signature” (sorry, I can’t type “metabolomic” without using quotes lol, what a word!) to their BMIs.

They then took the most promising “metabolomic factors,” which included BCAAs, phenylalanine, and tryptophan (all amino acids) and performed a mouse trial with 8 different diet interventions, 5 mice per group.

So meta

Interestingly, almost all the amino acids that were highly correlated with obesity and metabolic syndrome in the human meta-analysis are part of the BCAA metabolic pathway, part of the TCA/Krebs cycle, which basically runs energy production in each and every one of our cells.

This pretty awesome graphic from the study shows some of the implicated factors and where they fit in the Krebs cycle:

Bold means “associated with obesity,” red are BCAAs, green is “positively associated with obesity, “ and blue is “negatively associated with obesity.”

The two BCAAs isoleucine and valine are associated with obesity, as are a few others. Glycine is negatively associated with obesity.

If you’re familiar with Fire in a Bottle, you’ll notice that glycine leads through alpha-ketoglutarate (AKG), which Brad talks about a lot and sells on his website. Isoleucine and valine seem to use this up, and it could be a rate-limiting factor in the TCA cycle. An increased glycine/ile+val ratio might help increase this if rate-limited, either by increasing glycine intake relative to ile+val intake, or by directly taking AKG.

Science is pretty crazy, huh?! I love that they did this, makes it super visual and intuitive to follow the pathways and speculate what could be happening.

And after you speculate, you test.

Mice HATE this study!

Let’s look at the results. First, body weight:

The HFD (high-fat diet) group was the heaviest, as usual. The SCD (standard chow diet) mice were much lighter.

Interestingly, the HFD-BCAA (BCAA-restricted) mice were even leaner than the control mice! I think this is because standard lab rodent chow is actually already slightly obesogenic. Remember how, in the last study, even the control chow was 3-4% linoleic acid/PUFA? That’s probably better than 8%, but not optimal.

So if you restrict their BCAAs, you not only help them get rid of the difference between SCD and HFD, but also the 5-10% of extra weight that the SCD made them put on.

The diets that restricted phenylalanine or phenylalanine + tryptophan on top of BCAAs did not produce any additional weight difference. The researchers concluded that these two amino acids are not implicated in obesity, at least not independently of BCAAs.

Looking at body composition, the HFD mice were the fattest, the BCAA-restricted mice were the leanest, and the SCD, BCAA-Phe, and BCAA-Phe-Trp mice were in between. Pretty much what you’d expect.

In the glucose tolerance test, the pattern is exactly the same. HFD mice are diabetic as heck, BCAA-restricted mice have the best glucose response, and the other mice are just a smidge above the BCAA-restricted ones.

I really like this study because it uses 2 parts of science really well: it does a meta-analysis of epidemiology to generate a bunch of hypotheses, and then it constructs a mouse trial to test these hypotheses. Some pan out, others don’t.

The Diets

Similar to the Lamming studies, “restriction” here meant by 2/3. E.g. the HFD-BCAA diet had its BCAAs restricted by 2/3 vs. the normal HFD. Protein was kept constant by increasing the amounts of the other amino acids.

All high-fat were 61% kcals from fat, and 21% kcals from carbs. All standard chow diets were 10% kcals from fat, and 72% kcals from carbs. All diets were 18% kcals from protein.

The vast majority of carbs in the SCD came from corn starch, while all diets contained equal amounts of maltodextrin and sucrose.

All diets contained the same amount of soybean oil, and the high-fat diets contained the rest of their fat from lard.

Once again I am asking researchers to not use soybean oil and lard for everything.

The presence of the soybean oil baseline makes my above theory reasonable, I think: even the SCD mice are slightly insulin resistant and metabolically unhealthy, so the BCAA mice outperform even them.

10% kcals from soybean oil means 6% kcals from PUFAs and 5% from linoleic acid alone. That’s quite high, I’d say.

I think the reason they always use soybean oil is that it contains the essential fatty acids, so this ensures none of the mice become deficient in them. But it comes at the cost of confounding every single rodent study.

Resistance exercise protects mice from protein-induced metabolic dysfunction

https://www.biorxiv.org/content/10.1101/2022.10.18.512689v3

This study came out last less than a month ago, in October 2023. It is another excellent Lamming Lab study.

The method of the study is kind of hilarious. Apparently, someone recently invented a way to make mice do resistance training. Before that, it seems, lab mice were only ever doing endurance exercise by running in a wheel.

How they do it is they make the mouse drag a cart loaded with heavy weights. This study made the mice do this strength training 3x/wk. They also had a control group where the cart was unloaded, so there was practically no resistance. I suppose this was for placebo reasons or something, or were they trying to double-blind a mouse study?!

The goal was to find out if the negative effects of BCAAs on a high-fat mouse diet could be mitigated by strength training.

In short, yes, to a degree. But not as completely as the 2/3 BCAA restriction they’d tested previously.

Study design

They ran a 2x2 design. Half the mice were doing the Weighted Pulls (WP), half did the “sham” cart. Half the mice were doing Low Protein (LP) containing 7% of kcals from protein and half were doing High Protein (HP) with 36% of kcals from protein. So you had LP/WP, LP/Sham, HP/WP, and HP/Sham.

Strength training protects high-protein mice from obesity, but not glucose issues

First up, the low-protein mice ate significantly more than the high-protein mice. Protein leverage hypothesis, amirite? This is true for both the WP and the sham training mice.

But what’s this? The LP mice, despite eating more total food and more food per body weight, stayed way lighter!

This is where it gets interesting with the strength training. As expected, the high-protein sham mice were the heaviest, and the fattest. The LP mice, both WP and sham, were the lightest and had the least fat mass.

But the HP strength trained mice were significantly lighter, and had almost as little fat mass as the LP mice! They also had significantly more lean mass than the LP mice.

Interestingly, the HP sham mice had almost the same lean mass as the HP strength training mice. I’m not sure what this means, maybe even pulling an empty cart is quite a workout for a mouse? Or high protein intake alone stimulates mTOR enough?

You can see that the strength training didn’t entirely protect the HP mice from fat gain, they were still fatter than either of the LP groups. But they were much closer to the LP level of fat mass than to the HP sham group.

So the strength training definitely looks like a win.

In the glucose tolerance test, the strength-trained HP mice did almost as badly as the sham HP mice. Interestingly, the strength-trained LP mice did slightly better than the sham LP mice.

So strength training does seem to have an effect on glucose and insulin sensitivity, but it’s minimal, at least in these mice.

This is interesting: the LP mice ate more than the HP mice, and the LP sham mice were way more “spontaneously active.” I think this is just running around the cage and stuff.

The LP + WP mice were actually slightly less active than their HP + WP counterparts. Both WP groups were less active than either of the sham groups.

This seems to support the idea, I think promoted by Herman Pontzer, that energy expenditure is regulatede, and if you are highly active, your body will make up for it by down-regulating spontaneous activity and other non-conscious energy expenditure, like body heat.

This graph shows total energy expenditure in a metabolic chamber. The LP sham mouse were clearly expending the most energy (they also ate the most). The LP + WP mice followed, and then the HP sham mice. The HP + WP mice had the lowest energy expenditure. Since they also ate very little, that makes sense.

Once again, the best method to create a caloric deficit seems to be:

1. Eat more food than anyone else

2. Eat low protein

3. Spontaneously have your body expend all that energy without exercising

Now I do like this study because exercise is usually brought up for weight loss in the dieting context. It seems like, at least in mice, it can definitely help protect against the obesogenic effects of protein.

It’s unclear how exactly this translates to humans, especially since the HP sham mice gained nearly the same amount of muscle as the HP + WP mice. This seems to be not the case in humans… I think?

And to how much strength training do 3x weekly weighted pulls in a mouse translate in humans? Going to the gym 3x a week, doing what exactly?

But it does show us that strength training could be a potential tool for mitigating the obesogenic effect of BCAAs in the high-fat diet context.

And by high-fat I mean, of course, seed oils. Both the LP and HP diet contained 18.5% and 18.9% of kcals respectively from fat, and the fat was corn oil and olive oil.

Corn oil is 53% linoleic acid, olive oil about 10%. So these mice were all eating about 7% of kcals from linoleic acid, and we think 8% is the “definitely bad” number.

Once again, all of this takes place in a high-PUFA context.

Dietary restriction of isoleucine increases healthspan and lifespan of genetically heterogeneous mice

https://www.sciencedirect.com/science/article/pii/S1550413123003741?dgcid=author

Quickly closing in on the current day, this November 2023 study (note: it is currently still November 2023 as I am writing this) from Lamming Lab continues the excellent track record.

Having nailed down the causal effect of protein restriction to mostly isoleucine, they now sought to remedy one of the restrictions of their previous studies: the studies were mostly done on a certain type of mouse, usually all mice were the same age, and they were all male. I didn’t tell you this, but this is listed as a limitation of most of the studies we’ve talked about so far.

This time, they focused on expanding the mouse cohorts to a more heterogeneous population. They tested both young and old mice, male and female. They also used a stock of mice called UM-HET3, which is apparently specifically bred to be genetically heterogeneous. I suppose exactly for studies like this.

Unfortunately, I don’t have full access to the study, but it seems to have worked:

The graphs show that the isoleucine-restricted mice were much lighter and had better glucose control than both the total protein-restricted and the control mice. They also lived 33% longer.

Altered branched chain amino acid metabolism: toward a unifying cardiometabolic hypothesis

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6260824/

This 2019 paper discusses the current (at the time) research on BCAAs, impaired BCAA metabolism, and the consequences of that, as well as potential directions for future research.

Overall, BCAAs represent a promising emerging target for risk stratification and possible intervention, to support efforts to mitigate the burden of cardiometabolic disease in the population.

It highlights how impaired BCAA metabolism is associated with Type 2 Diabetes, heart disease, obesity, and more.

Overall, a great summary of the topic of BCAAs.

Conclusion: BCAA restriction works

Restricting protein, specifically BCAAs, specifically isoleucine, prevents pretty much all the negative effects of the “high-fat obesogenic Western” (=lard/soybean oil) rodent diet.

Elevated BCAAs have also been associated with obesity in humans, and protein restriction seems to bring them down.

Anecdotally, protein restriction, and maybe BCAA restriction in particular, seems to make obese people lose weight pretty effortlessly. Both on keto (ex150) and on a high-carb, low-fat diet, as in Brad’s case at Fire in a Bottle. In his case, it also vastly improved his blood glucose (never an issue for me on keto, so can’t say if the protein restriction helped).

There is the question if this is just some genetically “special” people who can’t handle protein, or maybe (as Brad believes) this is only due to the metabolic context of “torpor” aka excess PUFAs.

I’m hopeful we’ll see more of these studies in the near future, especially on humans. It’s difficult to put humans on diets as restrictive as those we give to lab rats, but there seems to be a lot of momentum behind the protein/BCAA restriction idea.

Just see how many of these studies are <3 years old, some even less than 3 weeks!