Tuesday, August 5, 2014

Can Hypothalamic Inflammation and Leptin Resistance be Reversed?

A new study by yours truly begins to address the key question: can hypothalamic inflammation and leptin resistance be reversed?

Leptin is the primary hormonal regulator of body fatness in the human body (1).  Secreted by fat tissue, it acts in many places in the body, but its most important effects on body weight occur via the brain, and particularly a brain region called the hypothalamus.  The hypothalamus is responsible for keeping certain physiological variables within the optimal range, including blood pressure, body temperature, and body fatness.

In obesity, the brain loses its sensitivity to leptin, and this causes the body to begin 'defending' a higher level of body fatness, analogous to how a person with a fever 'defends' a higher body temperature (1).  Once a person has become obese, it's difficult to return to true leanness because this system vigorously opposes major fat loss.  Leptin resistance makes fat loss more difficult.

In rodent models, leptin resistance is caused at least in part by inflammatory signaling in the hypothalamus.  We can observe this in multiple ways, but one common way is to look at the appearance of specific cells in the brain that change number, size, and shape when inflammation is present (2).  These cells are called microglia and astrocytes.  In addition to the work in rodents, we've published preliminary evidence that these same inflammatory changes occur in the hypothalamus of obese humans (2).

A key question is whether or not these inflammatory changes can be reversed.  Is a person with leptin resistance doomed to have it forever, undermining fat loss efforts for the rest of his or her life?  Or can it be corrected, possibly allowing easier and more sustainable fat loss?  We just published a study in Endocrinology that begins to answer this question, using a mouse model of dietary obesity (3).  I'm co-first author of this study along with my colleague Kathryn Berkseth, MD.  My former mentor Mike Schwartz, MD is senior author.

The Study

We divided the mice into three groups:

  1. CHOW.  This group received standard unrefined rodent food for 5 months.  This is a diet composed primarily of unrefined corn and soybeans, with a little bit of added fat, meat, and micronutrients.  Rodents do well on this diet, particularly if they're encouraged to exercise (which they weren't in our study).
  2. HFD.  This group received a refined high-fat diet for 5 months.  This is the same diet many research groups use to produce dietary obesity.  Most of its calories come from lard.
  3. HFD-CHOW.  This group received the refined high-fat diet for 4 months, then was switched to unrefined chow for an additional month.

Each mouse had unlimited access to its assigned diet at all times.  We periodically weighed each mouse and measured its fat mass and lean mass.  At the end of the 5-month period, we collected brains and looked for evidence of the cellular changes that indicate inflammation.  We also used MRI to attempt to measure signs of brain inflammation non-invasively.


What we found is quite remarkable.  The HFD group rapidly became obese, achieving a five-fold higher fat mass than the CHOW group.  However, when we switched the HFD-CHOW group back to the unrefined chow diet for one month, the mice lost nearly all of their excess fat mass!  It was a remarkable physical transformation in just four short weeks.  We were expecting fat loss, but not to that extent.

When we examined their brains, we found that the inflammatory changes in the hypothalamus were essentially gone in the HFD-CHOW group.  Microglia and astrocytes had returned to their resting state.  Although we didn't directly test leptin sensitivity (for practical reasons), it was likely restored since body fatness normalized and hypothalamic inflammation disappeared.

Unfortunately, the MRI arm of the study didn't work out as well as we had hoped.  MRI was able to detect an increase in inflammatory changes in the HFD group, as previously published, but not reversal in the HFD-CHOW group.  The sensitivity of this technique is currently much lower than looking at the hypothalamus under a microscope.  Our collaborators Ellen Schur, MD, and Joshua Thaler, MD, PhD, are continuing to work on improving it.


Our study shows it's possible to reverse hypothalamic inflammation in mice and cause a nearly complete reversal of obesity, using diet alone and without imposing calorie restriction*.  This raises the possibility that the same could be true in humans.  Since our study was conducted in mice, and under highly controlled conditions, it remains unclear how relevant our findings will be to humans.  However, it provides a rationale for further study.

Dietary fat loss strategies typically have limited effectiveness in humans.  We think this is due primarily to two factors: 1) limited adherence to the weight loss diet, and 2) opposition to fat loss by the brain/body, particularly the leptin system.

The cool thing about rodent studies is that we can exert complete control over what the animals eat.  This is a major difference between animal weight loss studies and most human weight loss studies.  Humans like the food we like, and it's tough to change habits.  In our experiment, we were able to completely restrict the HFD-CHOW group to an unrefined healthy diet for the last month of the experiment.  This is impossible in free-living human studies, and difficult to achieve under tightly controlled conditions due to high costs and the difficulty of recruiting volunteers.  It allowed us to eliminate the diet adherence factor, focusing on the diet's effects on the system that opposes fat loss.

Our findings raise the possibility that leptin resistance could be reversible by diet in humans, under the right conditions.  I hope future research will pursue this important possibility.

* To clarify, calorie intake did decrease temporarily, but not because we restricted the amount of food available to the mice.  They chose to eat less.


Howard said...

I would be very interested in the exact makeup of the HFD. You mentioned that the majority of the calories came from lard (unlike the usual Atkins-bashing Purina 5TJP, which features partially-hydrogenated cottonseed oil as its main fat, along with soybean and corn oil) -- but not what the other ingredients were.

From the increase in inflammation reported -- and the weight gain, I would bet money that there was a substantial amount of sugar in the

g-mo said...

The voluntary reduction in calories of the mice reminds me of the "bland liquid diet" study that you've mentioned before. The results don't seem to contradict your food reward model.

Stephan Guyenet said...

Hi Howard,

The diet is Research Diets D12492. It's 60% fat by calories, and 7% sugar by calories. There is no hydrogenated fat, but there is a small amount of soybean oil for n3 fatty acids. The exact composition can be found here:


That is substantially less sugar than what most Americans eat and it can be considered a moderate-to-low sugar diet. Almost any type of fat tends to be fattening in rodents, however, the added sugar does increase the fattening effect of the fat.

This study wasn't designed to isolate what properties of the diet are responsible for the fattening effect. We make no specific claims that the differences in fat content, carb content, calorie density, palatability, etc. are responsible for the effect. The high-fat diet is a refined food and is probably unhealthy/fattening for multiple reasons. We use this diet because it makes them fat so we can study the process of obesity development, that's all.

Hi Garrett,

It's strikingly similar, isn't it? In both cases, the diet shift was extreme and adherence wasn't an issue.

Stephan Guyenet said...

Garrett, to expand on that a bit-- the mice LOVE the high-fat diet. It's softer, fattier, and sweeter than their regular dry pellets. Once they've tasted it a few times, they'll drop everything to gorge on it at any time of day, even while they should be sleeping.

Ned said...

I'm always grateful when you make this stuff fairly understandable for non-science folks like me.
I wonder what would be a comparable human "chow" diet? I guess it should be kind of boring, include essential nutrients, and be non-inflammatory. I'm desperate enough to be an N=1 if someone comes up with Bland Chow.

valerie said...


you mention in a previous comment that all fats tend to be fattening to rodents.

Do all those fattening diets also bring hypothalamic inflammation and/or leptin resistance?

Also, what happens when rodents are fed a HFD restricted in calories? Do they become "starved obsese" rodents or do they stay normal?


rpforensic said...

Any similar studies being done on human brains examined at autopsy? Although the diets of the victims of sudden, unexpected death cannot be controlled, the hypothalamus could easily be collected and examined.

Vladimir Heiskanen (Valtsu) said...

Nice results.

I have seen something quite a similar reversal of metabolic syndrome in another study, in which they used glycine.

They added 30% sucrose to the drinking water of rats, which led to significant increase in fat cell size, and hypertension/dyslipidemia/other stuff.

But addition of 1% glycine to drinking water reversed most of the harms in 4 weeks.

http://www.ncbi.nlm.nih.gov/pubmed/15331379 Glycine intake decreases plasma free fatty acids, adipose cell size, and blood pressure in sucrose-fed rats.

Other similar papers:


I think the effects might be caused by the hyperpolarizing effect of glycine on immune cells. That decreases pro-inflammatory cytokines, and probably prevents many of the harmful effects of inflammation in liver, abdominal tissue and hypothalamus.

I have recently written an article on endotoxemia, inflammation and metabolic syndrome. The reference list has a lot of papers you might find useful (some on hypothalamic inflammation): http://180degreehealth.com/thyroid-inflammation-metabolic-syndrome-suprising-power-context/

Raza said...

This is great research Stephan. You mentioned that the HFD-CHOW group lost nearly all of their excess fat mass. Do you know if their was any lean tissue loss in addition to the fat loss?

Also, do you have any insight on how different time periods affect the reversal of the inflammation/leptin resistance? In this case the mice ate HFD for 4 months, then CHOW for 1 month. Might you see different results if the mice were fed HFD for 1 year then CHOW afterwards?


aluchko said...


So that means the amount of food available to the mice was unlimited?

Are there plans for longer term mouse studies? I'd be curious to see if the mice become more resistant to weight loss the longer they've been obese. Outside of this guy there's not a lot of humans who are only obese for ~6 months.

Travis Jensen said...

Is it really surprising that humans might be able to recover from obesity without calorie restriction just by changing diet? I’ve done it, and I interact regularly with several hundred other people who have, too.

Of course, doing so requires things like seriously reducing carbohydrates, increasing fat, and eating real foods (not saying those are the only ways, but they do seem to work pretty consistently), something most nutrition researchers seem to think is impossible (and that is saying nothing about the ludicrous, unsubstrantiated fear of saturated fat so many have). You’d almost think most nutrition researchers are being paid by the industry they are ostensibly researching.

Dale said...

Wow, these are really interesting results! I wonder what effect pushing the mice off the wagon once or twice during their chow month -- i.e. making HFD available to them a couple times -- would have. I have a notion it would block the return to normalcy almost completely.

Gretchen said...

What if you took people who were obese but couldn't lose weight in part because they couldn't afford decent food.

If you offered them free food for a couple of months, they might be willing to eat only the food you provided, with the understanding that if they cheated, they would be out of the study.

That would save the cost of an in-hospital study, and although 2 months wouldn't be as good as 4, it might suggest a trend toward weight loss and would be cheaper.

To match the mouse study, you could let them eat all they wanted of the food you provided, which again might motivate them to stick to the diet.

Ben Kennedy said...

How about putting the mice that lost the weight back on the high fat diet? Then you would compare the rate of the original weight gain against the second weight gaining episode. If there were any significant changes (for example, they regain the weight must faster) that would be evidence that they time they spent as obese mice somehow changed their internal body chemistry

Sanjeev said...

good stuff, I'm so glad these ideas are being fleshed out

> MRI was able to detect an increase in inflammatory changes ... but not reversal in the HFD-CHOW group

has it been ruled out that changes in other brain areas could compensate while this region stays inflamed?

If so how was that ruled out?

If this is a really dumb question only a neuro-newbie-layperson could ask, let me know that too ;)

Tobin said...

Very interesting work Stephan, congratulations.

I have a question about feeding studies in mice that I'm wondering if you would address. I am wondering about the applicability of findings about mice diets to human diets. (I realize your study was not advocating for a certain diet in humans, but it just made me think of this.)

Maybe this is too simplistic, but to me it seems like since mice are basically smaller than any other animals and are not predators or scavengers, they would not be evolutionarily adapted to eating meat of any kind, so it is not surprising that they do poorly on high fat diets.

For example, there is some published work about high fat diets causing gut inflammation in mice. But would that same adverse effect of high fat diet necessarily pertain to humans?

In your study, you showed that leptin resistance is reversible by diet in mice. But it could be that the diet that would achieve this reversal in humans would be different.

I understand the appeal of animal models for experimental purposes, but when it comes to diet, I just wonder if it makes sense to extrapolate from mouse to man.

Unknown said...

You can get meal replacement powders that taste bad. Also some that taste OK and dilute them to bland. Or mix and match to a bleh flavor.
Anything under 1,000 calories for any length of time, should be done under supervision.
Recently started the Newcastle diet. 800 cal which is for 2 months. (It is said to possibly reverse diabetes. I'm not but was approaching it.)
Assuming I make 8 weeks, I will up it to 1,000 cal.
Casein will help keep you full feeling for about 7 hours and you can get unflavored casein which is a step towards bland. But remember it is protein only you need other stuff for meal replacement.

Bariatric Advantage
has a meal replacement.
As does

There are vegetarian options but tend to be low in protein and taste bleh to begin with.
All products I've checked have something I don't like about them.
Assuming I can even get all the information I want.

Personally I am going with several different ones and mixing and matching.

Gabriella Kadar said...

Hypothalamic inflammation then would affect more than leptin. What about down the road to the pituitary? and thyroid? Maybe also PCOS all for the same reason. Except permanent? Isn't PCOS permanent?...'Eat yourself to infertility'.

Same with men and lowering sperm counts?

Laura said...

Very cool. Just posting this link for you; I'm pretty sure you have already seen it but just in case.

Richard Nikoley said...

It gets complicateder and complicateder:



Could a simpler answer be that when we do what we need to do instead of what we may want to do we obtain the results we want? It seems to work in business & real life! Trying to find a solution that defies Nature & our evolutionary traits seems to be a futile effort. The hard truth is if you do not follow simple steps that are in keeping with how we evolved we will be deemed to failure!

Chris said...

Hey Stephen,

Great work! Just wanted to get your take on skinny people who consider themselves ectomorphs. I am one of these people but most like me just eat very little but think we eat a lot. I listened to your AHS talk on leptin and wondered if the reverse is happening in mine and other cases.

If I don't track calories or what I eat I undereat. Could my leptin levels naturally just be very off?

Anecdotally I was eating for years only processed foods and snacking all the time and gained 40 lbs of fat. I changed to whole foods without tracking calories and lost 50 lbs within 6 months then had to start counting calories to ensure I was eating enough.

Wondering if any study has been done on the reverse side of leptin causing undereating.

GearĂ³id said...

I think you're wrong and here's why. I'm a retired Doctor and all my father's family were fat and his mother and one of his brothers had DM2. I've struggled with my weight until recently when hey presto, a low carb diet allowed me to lose 40lb and to keep it off, curing my leptin resistance and hypothalamic inflammation?
Not hardly Doctor. Lowering my insulin. Now insulin may increase Leptin resistance and all that, but I doubt that Leptin is crucial. All the indicators are that Taubes is right. But whether it is primarily Leptin or Insulin is moot, because LCHF really works.