Welcome all seeking refuge from low carb dogma!

Monday, January 28, 2013

Lessons from LIRKO

The LIRKO mouse has no insulin receptors in its liver.  It's one pretty sickly mouse.  In most studies it becomes hyperglycemic and hyperinsulinemic very early in life, but by six months or so of age is either normoglycemic or hypoglycemic.  Despite the concurrent hyperglycemia and hyperinsulinemia, the LIRKO remains normal weight, it is even a bit underweight if anything.  This is despite the fact that  circulating free fatty acids (FFA, or my preferred acronym, NEFA)  are suppressed by 40%.  {Here are the two papers I've discussed in previous blog posts: Loss of Insulin Signaling in Hepatocytes Leads to Severe Insulin Resistance and Progressive Hepatic Dysfunction , High Circulating Leptin Receptors with Normal Leptin Sensitivity in Liver-specific Insulin Receptor Knock-out (LIRKO) Mice both links are to free full texts}

LIRKO presents two problems for the TWICHOOB:
1.  Chronic hyperinsulinemia does not cause rampant fat accumulation 
2.  Insulin "locking away" fats does not lead to hyperphagia (overeating)

The real gossip mongers on the outskirts of the IHC have been buzzing about various of us bloggers having been "schooled" by Dr. Jim Johnson, whom I will call Jim because that is the way he signed his comments here on this blog.  I hope that Jim will continue to engage this discussion, but for my part I have to beg off the Twitter part of it.  Twitter, with its character limits and difficulty following anything more than a short back and forth, just is not the appropriate venue for productive conversation on these topics.   I have offered in email to do a Todd Becker style thing (search on Insulin Wars and Todd Becker here to see what I mean) if that might work better and Jim has mine so if he's reading this and comments are too cumbersome b/c of length and formatting limits, just email me and I'll post it up!

In any case, as I understand it, Jim has written off the *IRKO mice as complicated beings that are not representative of realistic physiological scenarios that occur in nature.  I would remind that his "half insulin" KO mouse is not either ... especially in humans as the insulin deficient invariably exhibit some manner of disturbed glucose and lipid metabolism.  The other argument Jim has put forth is that the LIRKO is irrelevant b/c it wasn't created to study obesity.  I find that an odd argument for any scientist to make given the prevalence of great scientific discoveries that were the product of experiments or even accidents while investigating often wholly unrelated phenomena.  But I find it additionally troubling in light of the fact that Jim's own mouse and experiments led to the "surprising" discovery of obesity prevention.  According to their own press release, this mouse was not designed for obesity research either!  {see Does Hyperinsulinemia Drive Diet-Induced Obesity?}

To conclude that hyperinsulinemia "drives" obesity from just his work, requires ignoring considerable data -- both observational and experimental -- that is inconsistent with that hypothesis.  It is not sufficient to me to disregard this data, in a manner that appears flippant but I think that's the Twitter-factor ... or at least I hope so.

If hyperinsulinemia drives diet induced obesity, then a complete hypothesis must include at least a hypothesis as to how.  A mechanism.  In the low carb community, that mechanism has always been an imbalance of fatty acid sequestration (esterification/acylation) as compared to mobilization (lipolysis).  As Taubes put it, anything that works to put fat into adipocytes (e.g. insulin, but not ASP oddly enough) is "fattening" and anything that works to mobilize fats (e.g. less insulin, but more oddly enough not exercise) is "unfattening".

So with the LIRKO mouse the hyperinsulinemia should "trap" fatty acids in the adipocytes and this mouse should be fat.  But it is not.  The next thing that must be looked at then, is if this "violates" the mechanism of the hypothesis.  IOW, is there something about hepatic insulin signaling that interferes with the adipocyte signalling?  Not according to the studies.  Insulin is doing its job and NEFA (circulating free fatty acids) are considerably depressed (40%) in the LIRKO mouse.  So why is it not fat?  A reasonable explanation has not been provided.

But leaving obesity aside for another day (feel free to discuss now), LIRKO holds a few more "lessons" for us.

Pancreatic Exhaustion is countered by LIRKO:  This is a common meme.  Eat a lot of carbs, use your pancreas, which leads to abusing your pancreas as it needs to secrete more and more as your cells become deaf to the signaling.  Eventually you wear your poor pancreas out and you become diabetic.  If this were the case, LIRKO should eventually become hypoinsulinemic with evermore out of control hyperglycemia.  But exactly the opposite occurs as LIRKO ages.  The LIRKO at 6 months is still pumping out crazy amounts of insulin (9X fasting, almost 25X fed) compared to controls, while it's fasting hyperglycemia has turned to hypoglycemia and postprandial hyperglycemia is significantly muted compared to the young LIRKO (though still hyperglycemic vs. controls).   The livers of these mice are giving out, no longer able to keep up with the unchecked gluconeogenesis the liver thinks it's supposed to generate.

LIRKO counters Hyperinsulinemia causing Insulin Resistance:  If we define IR as skeletal muscle IR measured or ascertained by postprandial glucose clearance, this, too is countered by LIRKO.  In the glucose tolerance test administered to the LIRKO, the same amount is administered.  If chronic hyperinsulinemia led to diminishing insulin receptor action, reduction in glucose transporters or anything of the sort that reduced glucose "clearance", we would expect worsening postprandial hyperglycemia over time.  But we don't.  This is not to say that HI cannot precede and/or cause IR, but not by this mechanism analogous to building up a tolerance.

LIRKO counters Glucotoxicity Killing Beta-Cells:   One of the mechanisms by which hyperglycemia is thought to accelerate the progression of diabetes is that of glucotoxicity damaging beta cells.  Although LIRKO's hyperglycemia subsides with age, it certainly does not appear to damage the pancreas.

LIRKO counters Chronic Hyperinsulinemia interfering with acute GSIS by depleting insulin stores:   This isn't something we hear often, but it's a corollary of the exhausted pancreas theory.  I blogged on a study about exactly this Chronic Exposure to Free Fatty Acid Reduces Pancreatic β-Cell Insulin Content by Increasing Basal Insulin Secretion That Is Not Compensated For by a Corresponding Increase in Proinsulin Biosynthesis Translation almost two years ago.  That study, on which renowned diabetologist JD McGarry was a contributing author, boils down to chronic insulin production by stimulation by elevated NEFA doesn't stimulate replenishing of the stores so that glucose stimulated insulin secretion, GSIS, is impaired.  Given the chronic hyperinsulinemia, merely producing a lot of insulin chronically can't be the cause of impaired GSIS.  LIRKO still mounts an impressive GSIS despite pumping out a lot of basal insulin.  What is different here?  Well,

LIRKO supports "It's all about the NEFA"  The LIRKO mouse is a sick mouse alright and doesn't fare well as it ages.  But the one thing LIRKO is not lacking is pancreatic function.  They have super pancreata!!  They produce both basal and glucose stimulated insulin admirably, something Type 2's don't do from earlier on than is often recognized.  The basal hyperinsulinemia in T2 is often accompanied by impaired GSIS.  Why?  Well, the study in the previous paragraph holds the clue.  It is the elevated NEFA that leads to interfering with the insulin production machinery.   This may be due to excessive buildup of fatty acids in pancreatic cells as well (recall the "crash diet" study? Reversal of type 2 diabetes: normalisation of beta cell function in association with decreased pancreas and liver triacylglycerol )   But LIRKO has no such problem because if anything, NEFA are suppressed and properly tucked away where they are supposed to be stored when insulin is functioning properly at the adipocyte level.  


Gabriella Kadar said...

So, in other words, it's all about the liver?

Has anyone studied the glucose blood levels of those overfed geese used to make fois gras? They kill them too early, I guess. They are not mammals but still.

Sanjeev said...

I've read these studies and the reporting on them, trying to draw out the interactions as best I can - I'm curious about the technology now.

are there catalogs (in both meanings - compendiums and lists of items for sale/purchase) and taxonomies/hierarchies of these knockout models that have been made and studied to date?

If another lab wants to study LIRKO or derive another model would they need to generate their own or can they buy them - or is this standardized enough now if they make their own they can guarantee they got a duplicate?

Sanjeev said...

> derive another model

derivative of LIRKO, not "something similar to LIRKO"

Diana said...

I'm late to the Johnson/Mehran study, with the LC crowd seemed to regard as the nailing of the 95 theses. (Didn't that already happen with the publication of GCBC?) My takeaway from that was, "don't eat a high fat diet," esp. if you are a house. It was a great study, but the LCers only saw in it what they wanted to see. They didn't see that the process began with a high fat diet.

markgillespie said...

2. Insulin "locking away" fats does not lead to hyperphasia (overeating)

Should read 'hyperphagia'

OnePointFive said...

Well I thought I might go and ask my neighbours if I could test their duck's blood. (they all gavage ducks for foie gras) but google seemed easier! There are a huge number of research papers on foie gras production, it's a big industry
Most studies are on ducks since most foie gras comes from ducks nowadays,
There is an increase in glucose levels during the overfeeding phase. Two studies I found put it at about 12%. Plasma insulin was checked in one study and this was 2.23 times higher in the over fed duck(6.48µU/ml before and 14.45µU/ml after )
Obviously in this case the overfeeding precedes the hyperinsulimia

The overfeeding period is relatively short ie 10-14 days and the effects on the bird's liver are apparently reversible if overfeeding stops. ( studies like that by Roy Taylor are now suggesting the same occurs in humans and might also reverse early T2 diabetes)
As you say they aren't mammals and these birds are bred from varieties that naturally overfeed to store fat for winter. The de novo lipogenesis pathway is more important in them than humans.

Evelyn aka CarbSane said...

It seems in most cases these mice (all) are bred as needed. I don't think you can call up Charles River Labs and order up a bunch of LIRKO.

Evelyn aka CarbSane said...

Ugh, good catch. Will fix, thanks!

Diana said...

@Gabriella, If you want to study the metabolic effects of force-feeding on humans, go to Mauritania.

Horrible - but it would be an interesting study, wouldn't it? What happens to these girls every step of the way? What is going on before, during and after? If it is going on, we should learn something about it.

Gabriella Kadar said...

Thanks Diana. I've read about the force feeding of women in Africa but didn't know where. There are several areas in Africa where women are provided super nutrition in order that they breed well. I'm assuming that the high rate of pregnancy 'eats' up the excess. But probably the longterm adverse effects have not been studied.

Jane said...

Yes, I think the crucial thing is that as you say, the LIRKO pancreas is in very good shape. It will be releasing insulin in regular 5-10 minute pulses. One of the big shots in this field is Steve O'Rahilly, who told me many years ago he could get insulin resistance in his animals any time he liked if he abolished pulsatile insulin secretion. It may be that the total amount of insulin is much less important than the pattern of release.

Adipocytes also oscillate, at around the same frequency, apparently between lipogenesis and lipolysis. LIRKO adipocytes will be oscillating normally, unlike (I imagine) obese human ones. So the whole thing may be working just as it's supposed to work despite vast amounts of insulin.

Remember Jim telling us that if the animals are fed in a small time window they don't get obese? It means, periodic absence of insulin is necessary for things to work properly. Just what Steve O'Rahilly says.

Nigel Kinbrum said...

...or even a mouse ;-D

The Johnson/Mehran study was a great study and used a specific type of high fat diet (D12492?) with a specific type of mouse (C57BL6/J). You can't extrapolate the results to other types of high fat diets or to other species.

Astronautjane said...

A couple of quick comments:
Part, perhaps most, of LIRKOs hyperinsulinemia is due to reduced insulin clearance.

LIRKO also has hyperleptinemia without leptin resistance--might explain lack of hyperphagia.

Nigel Kinbrum said...

In Pancreatic Exhaustion is countered by LIRKO:
"with evermore out of control hyperinsulinemia." should read
"with evermore out of control hyperglycemia."

Atif Ahmed said...
This comment has been removed by a blog administrator.
Evelyn Lee Barney said...

Please excuse the OT - but I couldn't find your email address Evelyn - sorry!

This is an interesting article:

One thing in particular jumped out at me: “The poster child for quick evolution in humans is dairy tolerance,” Okay, makes sense. BUT, dairy INtolerance is common, and tends to increase with age. So maybe we are still evolving vis a vi dairy.

There's lots more to say on this article - but I'll leave it to you, Evelyn, to address it or not.

Charles Grashow said...

"there is no difference in the following hours and - what's even more important - the total energy expenditure was identical - 2594kcal/day vs. 2589kcal/day in the before and after breakfast trial, respectively. In addition to that, a detailed analysis of the unpublished study by Shimada et al. the above data was taken from, does also show that
working out before breakfast reduces the energy expenditure in the time before lunch -- over the whole period the subjects burned about 500 kcal less, when exercise was performed before breakfast
working out before breakfast burns more glycogen and increases non-oxidative carbohydrate storage during / after breakfast -- with the carbohydrate content of the breakfast being used for glycogen repletion, this does in fact lead to another increase in fatty acid oxidation, simply because the alternative fuel, namely the carbs are not oxidized, but stored
So, 2x thumbs up for cardio before breakfast - at least in the short run and when your goal is to maximize fatty acid oxidation, but ...

Is maximal fatty acid oxidation even what you should be aiming for during a workout?

I guess you will already have read between the previous lines that my answer to this question is a definitive "no". Moreover, most of you are so clever and have been following the SuppVersity posts for so long that they could come up with their own arguments against an overemphasis of intra-, post and total 24h fatty acid oxidation, when getting lean and healthy is your goal. And probably, some of them are even identical to mine:

Firstly, and most importantly, burning fatty acids for fuel does not equate fat loss. If you follow a real ketogenic diet (not one with tons of protein in it), you'll burn (almost) exclusively fat, but even under these "extreme" conditions most of the fat will come from the fat you eat, while the small amount that's actually taken from your hips, buttocks and whatever, will be restored unless you are in a caloric deficit, when your fatty acid oxidation will increase anyways.

Secondly, many of the metabolic benefits of exercise are closely related to the act of glycogen depletion. This is particularly true for the increase in GLUT-4 expression and consequent improvements in muscular glucose uptake, burning only fat for fuel during a workout is thus a questionable ideal.

Thirdly, working out "in the zone" may burn the most fat but won't have the conditioning effects high(er) intensity workouts have. While obese individuals and people who have been sitting around their whole lives will see improvements in their VO2max (and in the long run their heart-health), anyone who is not totally unconditioned misses out on the structural changes in the musculature, and as you've learned on day one of the SuppVersity Exercise Science Week adipose tissue, as well.

In short, the importance of burning fat for fuel is so overrated that exercise prescriptions that are based on the paradigm of maximal fatty acid oxidation are at least suboptimal for health, fitness and physique purposes. Some people, I guess, would probably even go so far to say that they do more harm than good. I for my part leave it up do you to decide whether you join sides with my carefully worded or the more extreme version of this conclusion, or - and this would be your good right - to wholeheartedly disagree with both of them."

Diana said...

Gabriella, I didn't know that it was done in several areas of Africa, I just thought it was in Mauritania, but the Wikipedia article says that it started as a Berber culture, so it probably is practiced in a variety of places. I wonder if it started as simply a sensible way to help fertility in places where food is scarce, and then turned into a (disgusting) fetish.

Whatever, it's clear that forcing a girl to gorge on high carb/high fat, repeatedly, makes them fat. I've seen pictures. They don't start out fat at all - they are pretty lean folk. The women turn into real camels.

I guess this also explains Gary Taubes' conundrum: "Obese mothers and starving children." It is a favorite meme of his, the subject of his famous confrontation with Stephan. I can answer that: when you are in a refugee camp, and all you do is gorge on crappy carb/fat rations, you get fat. And your kids get malnourished.

Diana said...

"esp. if you are a house. " "or even a mouse."

LOL, Nigel. I am making the strangest misteaks in my typing nowadaze.

Charles Grashow said...

"...around 10,000 B.C., this began to change. A genetic mutation appeared, somewhere near modern-day Turkey, that jammed the lactase-production gene permanently in the “on” position. The original mutant was probably a male who passed the gene on to his children. People carrying the mutation could drink milk their entire lives. Genomic analyses have shown that within a few thousand years, at a rate that evolutionary biologists had thought impossibly rapid, this mutation spread throughout Eurasia, to Great Britain, Scandinavia, the Mediterranean, India and all points in between, stopping only at the Himalayas. Independently, other mutations for lactose tolerance arose in Africa and the Middle East, though not in the Americas, Australia, or the Far East.

In an evolutionary eye-blink, 80 percent of Europeans became milk-drinkers; in some populations, the proportion is close to 100 percent. (Though globally, lactose intolerance is the norm; around two-thirds of humans cannot drink milk in adulthood.) The speed of this transformation is one of the weirder mysteries in the story of human evolution, more so because it's not clear why anybody needed the mutation to begin with. Through their cleverness, our lactose-intolerant forebears had already found a way to consume dairy without getting sick, irrespective of genetics."

Sanjeev said...

> poster child for quick evolution in humans

For dietary adaptation, maybe


the most dramatic recent case of fast evolution IMHO is the human/ape changes in uric acid handling (loss of the enzyme uricase) and the apparent simultaneous loss of the ability to make vitamin C - the extent of the changed internal chemistry, the massive departure from most other life on earth, and so on again IMHO are more far-reaching than the dairy adaptations.

And that's just the chemistry of course - other evolutionary changes can be argued to have been a lot more drastic - what about bipedalism and language?

Massive evolutionary change can happen VERY fast.

Gabriella Kadar said...

It is worth considering that most of the dairy was fermented. Fresh milk will ferment in a very short period of time at warmer temperatures. Probably humans were consuming fermented dairy for a long time before some freak 'guy' grew up and could drink it fresh from the cow without getting diarrhoea.

The global distribution of multiple sclerosis coincides with populations that are lactose tolerant. (Except the Masaai)

If you look at the multiple sclerosis incidence in Finland, for example, the people on the west and south coast have a high incidence (they are people who migrated to the area from what is today Sweden and Denmark). The people who live in the eastern part of Finland, settled later by people from the east, are not lactose tolerant and also have a much lower incidence of MS. (Possibly the people in eastern Finland are migrants from the coast.) As an aside there are families in eastern Finland whose members are lactose intolerant from birth and hence the babies are given fermented milk.

In Sardinia, the people who live in the north and west of the island, settled by people from Europe, the incidence of MS is both higher and the lactose tolerance also higher than in the other parts of Sardinia.

What is interesting is that the lactase gene is on chromosome 2. This is a weird chromosome having been formed through a Robertsonian translocation. Is it possible that the genetics for MS are associated with a gene that is close to the lactase gene on the number 2 chromosome? Or is that too simple?

My comment is an 'aside' to your comment about lactose tolerance.

Gabriella Kadar said...

Diana, what about steatopygia (sp?)? African men prefer the contours of the gluteal region as opposed to American men who are attracted to the breasts. Is it about fertility or what?

Diana said...

This whole discussion encapsulates my frustration with Paleo in general. And I was a big fan of Paleo at one time, due to my reading of one of its originators, Melvin Konner. The Paleolithic era is 2.6M years long, and took place over vast areas, so to me it is meaningless.

The most important thing IMO for optimal diets (whatever that is) is to look at your own family, if you can. That will tell you much more than theorizing what a group of hunter gatherers did 200K years ago.

I used to be a huge cholesterol denialist, for the simple reason that on my maternal side many of my aunts and uncles lived to ripe old ages with high cholesterol. But I've had my mind opened. The conventional cholesterol theory has something to say to other people. If your older relatives died of CVD in their mid-60s, with high cholesterol, then pay attention to it. If they lived to 90-100 with high cholesterol, then be a skeptic.

@Gabriella, nothing that I said should be taken as saying that your info about Finland and lactase intolerance wasn't fascinating. It was!

Diana said...

@Gabriella, you should be copy-editing the NY Times: " evolutionary eye-blink, 80 percent of Europeans became yoghurt-eaters, kfir drinkers, and milk-fermenters" some areas the diet is indeed "milk-sodden."

Diana said...

"It is about fertility or what?" I would imagine.

Maybe since that is where sub-Saharan African women tend to put the fat on, guys become attracted to that. If women put the fat on in the boob area, they go for that.

I imagine some Ph.D. somewhere has written a paper on this with some fancy name on it. I would call it the "Maxim" theory of evolution.

Gabriella Kadar said...

Diana, could I earn a living copy editing? LOL! But I appreciate the compliment.

rodeo said...

Could you eleborate? I found your comment interesting and would want to learn more.

Jane said...

Well, this comment thread is full of absolutely fascinating stuff. LIRKO takes a back seat while we ponder what Evelyn has said. I should have mentioned that her analysis of LIRKO is the best I've seen. I don't think Jim Johnson could do better, and in some ways he might do worse, having a vested interest so to speak. It isn't good enough to just say 'LIRKO is hypoglycemic'. I expect he regrets that.

OnePointFive said...

'It is about fertility or what?'
Interesting chapter about the meanings of fatness, the writings of a researcher(participant observer) in rural Jamaica during the late 1980s.
The Sweetness of Fat E Sobo in
Many Mirrors: Body Image and Social Relations (can read the chapter on google books).

MM said...


I've done a journal search on lactose intolerance and multiple sclerosis. I've mostly come up empty. The only article I've found is this one: "The Molecular Basis of Lactose Intolerance"

It states, "The symptoms of lactose intolerance are caused by gases and toxins produced by anaerobic bacteria in the large intestine. Bacterial toxins may play a key role in several other diseases, such as diabetes, rheumatoid arthritis, multiple sclerosis and some cancers." It seems to come to the opposite conclusion.

Unfortunately, I don't have access to the full text on that one, but I'm having a hard time coming up with any articles that link the two at all. I'm not questioning your conclusions, but do you have links/references? Thanks.

Evelyn aka CarbSane said...

@Gabriella: I think the "it's all about the liver" or "it starts in the liver" was the first interpretation of the various IRKO mice. But there really is no pathology that presents as isolated severe liver IR. The IRKO mice are helpful in ferreting out the roles of various organs, however.

Charles Grashow said...
This comment has been removed by a blog administrator.
Evelyn aka CarbSane said...

Misteaks? Is that like a Mistersteak? ;-)

To me, Johnson/Mehran showed that HFD was responsible for elevated basal insulin ... yet another piece of evidence that postprandial insulin isn't the bad guy it's made out to be.

The other take home is that adequate insulin production and signaling at the pancreatic level may well be required for DIO to occur. The problem is, in this mouse, the excess fatty acids were burnt off in the fat cells by upreg of UCP1 activity. Does this happen in humans? I don't think so. All insulin deficient states are accompanied by "diabetes" (hyperglycemia).

Evelyn aka CarbSane said...

Hi Jane, Jim sent me that paper but I haven't had a chance to look at it with the undivided attention it needs. I would point out, however, that in that study they also compared 24hr ad libitum vs. restricted feeding window using a standard mouse chow and these mice were essentially the same. So, sorry Jim, but I think his extensions of his data to "keeping insulin low = no obesity" are premature, and they are, of course, being taken totally out of context by the low carbers. Low carbers who, I might add, generally ignore the action of fatty acids on insulin production.

I think leptin and fatty acid ox rates are likely at play here.

Evelyn aka CarbSane said...

Hi AstronautJane! Yes, LIRKO does exhibit reduced insulin clearance, but there is another study I plan to blog on in which both pancreatic and liver insulin receptors are KO'd and in those mice, the beta cell expansion that occurs in plain LIRKO doesn't happen. If the insulin-locks-fat-away hypotheses hold, however, it should not matter how the insulin remains high, only that it does!

Diana said...

Misteaks and Mistersteaks, whatever, they are all Low-Carb and all certified Kosher Paleo!

Regarding the diffs between mice and human fattening, this is another reason I'd love to study those fattened up girls in Mauritania, as unethical and inhuman as it sounds. What the hell is going on with their insulin levels, leptin, etc., with a perfectly normal group of girls, forced to become fat?

Of course I'd like to stop the practice altogether.

Evelyn aka CarbSane said...

Thanks! Fixed.

Diana said...


That's because most of us don't have the science chops to discuss it. OK, I speak only for myself. I don't have the science chops. I can only gape from the sidelines. (Simon Carter, where are you?)

What I wish is that Taubes would come here and discuss LIRKO, but he won't, because he can't!!

Evelyn aka CarbSane said...

@Evelyn, It's carbsane at gmail dot com for future reference. I was really hoping to migrate over my break so didn't bother updating. I think I'll have to do that now and add a contact button. Will take a look!

Evelyn aka CarbSane said...

Thanks Jane! Yes, Jim's dismissal of LIRKO somehow having supraphysiological insulin and thus hypoglycemia preventing obesity is an insufficient one. It needs to be explained better if it's a viable explanation. I and others who have commented on this blog after that exchange on Peter/Hyperlipid's never really got a fleshed out explanation. I would note that supraphysiological is a term that often refers to in vitro and in vivo infusion studies where levels exceed those possible or observed. He may have a point on the latter, that such levels are not observed, but no leg to stand on for the former, because clearly it's physiologically possible for the mouse pancreas to sustain these elevated insulin levels for 6 months (or more). I would note that the hypoglycemia when it occurs is mild fasting (as in a study he cited on Twitter) and the study the table in this post came from. But note on my table that LIRKO remains postprandial hyperglycemic (just not as much) at 6 months, which in a mouse should be fattening if it's all about the insulin!

Evelyn aka CarbSane said...

Oh yeah, Diana! Where is Simon?! LOL

I personally love when conversations like the one Gabriella sparked that led to the overfeeding of women you describe, etc.etc. develop. I only wish they didn't always seem to happen when my time is occupied elsewhere so I can't contribute. Carry on gang! I'm reading with interest.

Evelyn aka CarbSane said...

I moved this to the Konfederacy of Dunces post where it is more on topic.

Diana said...

"I personally love when conversations like the one Gabriella sparked that led to the overfeeding of women you describe, etc.etc. develop."

Good. I worry about wandering O/T. While I think the "Maxim" aspect *is* OT, I don't think that measuring what's going on physiologically with girls who are actually being experimented on in a real life fattening experiment is technically O/T. Aren't they better subjects for what happens in a fattening experiment than mice?

Although I admit it is unethical.

Evelyn aka CarbSane said...

Nope perfectly fine to go OT or even semi-onT (I liked the Maxim comment). I only moved Charles' comment below to another post b/c it was more relevant to that conversation.

Nigel Kinbrum said...

Oh, Diana. I think you need more mussels!

Diana said...

Nigel, I am a hard mussels-sprouter.

Gabriella Kadar said...

Maybe this type 2 diabetes has something to do with ambient testosterone levels. Maybe diet just brings it on. I have no idea in re: mechanism of action. Maybe mice aren't 'apples' or 'pears'?

Gabriella Kadar said...

MM, do you remember about the Italian doctor who has the theory that venous drainage from the brain is the problem with MS?

There are no research articles correlating lactose tolerance with MS. But after I read about his theory that some sort of developmental aberration of the venous drainage is responsible for the MS based on his studies in Sardinia, I started to look around.

I decided to spend some time entertaining myself. Hypothesis: Venous drainage abnormality is responsible for MS. Let's say it's true. Who knows but as a hypothesis, it is possible to just check around the problem.

MS global distribution coincides in large part to ability to digest lactose.

Next step: where is the gene. Chromosome 2. (21P or something like that, I don't remember and the article files are in the living room.) Chromosome 2 is a weird one because of the Robertsonian translocation. The gene for lactase is very close to a homeobox gene. Homeobox genes are responsible for the development of body parts.

Is it possible that during meiosis, these two genes sort together more often than apart?

Of course Dr. Zamboni's hypothesis is probably total shit, but I spent some lovely hours correlating the incidence of lactose tolerance and MS. And it may have nothing to do with a homeobox gene, but it doesn't mean that there's some other gene close to the lactase gene on chromosome 2 that may be a factor in the development of MS. Truly, everything just fell together. I wasn't even trying to make things fit.

JJ said...


I’ve been swamped with teaching and Feb 1 grant submissions, so I have not had time to look up discussion of the paper on the internet for a while.

First, thanks for your post. It seems to have spawned some interesting discussion. By the way, I’m not really trying to ‘school’ anyone… just providing some expert perspective. I hope no one takes offense. That’s not my intention.

OK. Back to the post. I stand by my interpretation that the IRKOs are not great for looking at insulin alone because (tissue-specific) insulin resistance and cross-tissue compensation are always confounding factors.

As for the relevance of reducing insulin, I would say it is much higher. There are humans that make 30% less insulin than normal due to relatively common variation in the VNTR region of the insulin gene. There are also humans that make more or less insulin due to short and long-term changes in the diet. Our mouse models these changes in isolation, so that we can test hypotheses related to insulin, independently of diet.

“To conclude that hyperinsulinemia "drives" obesity from just his work, requires ignoring considerable data -- both observational and experimental -- that is inconsistent with that hypothesis.”

In our study it did, that is just the logic of the experiment. Also, the paper cites many other papers indicating the same thing, including the FIRKO and the human diazoxide and octride papers. It also cites quite a bit of data indicating that hyperinsulinemia precedes obesity and insulin resistance in humans and models.

JJ said...

Below I will try to address some of the comments that come from the comments….

@Gabriella – the liver is super important for glucose homeostasis and insulin clearance. Not sure it is so critical for weight.

@Evelyn “But there really is no pathology that presents as isolated severe liver IR. The IRKO mice are helpful in ferreting out the roles of various organs, however.” Exactly. The IRKOs are good (not perfect) for understanding the roles of insulin signalling in all these different tissues, but the major drawback is that every other tissue goes into an unusual compensatory state. Not many of these approximate any human disease (but maybe some do).

@ Sanjeev @Evelyn – Actually scientists can buy many of the genetically engineered mice (Jax or CR), but it still takes around a year to do the crossing to get all the genes you want onto one mouse. The LIRKO for example is a cross between the floxed insulin receptor mouse and the liver-specific Cre mouse. Some of the mice are not made readily available, so it prevent other scientists from confirming the work. The IRKOs were not made widely available until recently.

@ Diana – there are no nails in coffins or theses in science, just bits of information and hypotheses that are tested. Gary’s book, which he sent me to read, is simply a review of some early research on nutrition. It doesn’t ‘test’ any hypothesis on its own.

The key thing with our experiment is that we were able to manipulate insulin secretion, rather than ‘insulin signalling’. As you will note, I have not used the information from our study to advocate for one diet or another, just as evidence that a reduction in insulin secretion (in this case genetic) can prevent obesity. While we have not completed the studies, we are looking at other diets. I suspect that hyperinsulinemia will play a causal role in the obesity generated from any diet that leads to chronic over-secretion of insulin.

@nigel – True, we are the first study to look at reducing insulin itself in any mouse type. However, since insulin and insulin signalling are required for weight gain in flies, worms, our mice and probably humans, I am willing to bet that the finding will be broadly applicable. We have quite a few ongoing studies in other strains of mice that are supportive. Also, work on the FIRKO mice agrees with our observations, as well as other studies, including S. Panda’s timed feeding work.

JJ said...

“To me, Johnson/Mehran showed that HFD was responsible for elevated basal insulin ... yet another piece of evidence that postprandial insulin isn't the bad guy it's made out to be.”
I don’t know the answer to this. Our study certainly doesn’t address it. The chronic hyperinsulinemia is almost certainly related to the expansion of beta-cell mass (and causing it).

“The other take home is that adequate insulin production and signaling at the pancreatic level may well be required for DIO to occur. The problem is, in this mouse, the excess fatty acids were burnt off in the fat cells by upreg of UCP1 activity. Does this happen in humans? I don't think so. All insulin deficient states are accompanied by "diabetes" (hyperglycemia).”

It depends on how insulin deficient the state it. Although estimations vary, most scientist and clinicians I know believe that humans are normoglycemic until they loose about 75-80% of functional beta-cell mass. I think mice and dogs are exactly the same. Remember, most of our mice were close to diabetes, but without crossing over into hyperglycemia.

Regarding UCP1 in humans, how do you know? These studies are only just beginning and there is some very exciting work coming down the pipe.

@Jane – you (and Dr. O’Rahilly) are totally correct regarding pulsatile insulin secretion. This is a very important area that is almost always overlooked. I’ve done quite a bit of work on the mechanisms of insulin secretion. I’m not sure if the LIRKOs are normal in this regard. We didn’t test this in our mice either (it is actually super hard to do in vivo in the mouse, or even in humans). Regarding Evelyn’s comment, Satchin’s paper, just like mine, shows difference in high fat, but normal chow. The data are very similar, actually, so it is likely that there is a similar underlying mechanism. We are actually planning to test this hypothesis directly and we will have data in a few years. Science is slow!

Regarding your comments to Astronautjane. I must disagree with the comment that it doesn’t matter why insulin is high. The context of the high insulin (or any other physiological adaptation) is really important. Unfortunately, we really don’t know what else is going on in the LIRKO.

JJ said...

“I don't think Jim Johnson could do better, and in some ways he might do worse, having a vested interest so to speak. It isn't good enough to just say 'LIRKO is hypoglycemic'. I expect he regrets that.”

Ouch! ☺ Well, I don't regret dismissing the LIRKO as a good model for studying the role of insulin (in relative isolation) in obesity, because insulin is one of a whole group of parameters that change. Glucose is altered, both up and down, and the level of hyperinsulinemia is extreme. There are really no correlates elsewhere in physiology. On the other hand, there are many studies that have addressed insulin and insulin signalling in fat, and they all seem to point the same direction to me. I think to focus on the LIRKO, and dismiss all the other IRKOs is a bit premature.

Since I was asked about LIRKO and people blogged about LIRKO in relation to our paper, I have answered as many questions as possible. Unfortunately, they are scattered all of the web and twitter… so, I will try to put them here in one place in a short form.

1. LIRKO is a model of liver-specific insulin receptor ablation, not a model where insulin levels are modulated specifically. There are >10 papers on LIRKO and similar models and they have some key differences. My interpretation is that the LIRKO shows that insulin receptors on liver are very important and all hell breaks loose when they are gone. I think it is a mistake (missteak) to over-interpret these data, because….

2. There are cases when pathologically high levels of insulin don’t lead to weight gain. For example, many patients with insulin-secreting tumours have normal weight . Also, you can’t just give mice large amounts of insulin in pumps and expect them to gain weight because very high insulin leads to hypoglycemia. It is clear to me that modest hyperinsulinemia (in the range that can be caused by changes in diet) is a very different animal than extreme hyperinsulinemia, which probably results in other compensatory changes in other tissues. Our study only addresses the former. I don’t think that the extreme insulin case invalidates the effects of modest, long-term insulin hypersecretion. In fact, we and others have spent a lot of time studying the complex relationship between insulin doses and biological effects. It is very non-linear, with physiological concentrations (picomolar) having entirely different signalling mechanisms when compared to so called pharmacological concentrations (nanomolar).

Anyway, that is my ‘dismissal’ ☺ insufficient or not. To ask questions about the role of insulin, I would rather focus on models where insulin, not insulin receptors, are the manipulation.

About “physiological”. Not everything that can happen in an animal or person is physiological. Things outside the normal homeostatic range are called super-physiological or pathophysiological. That is just jargon.

I don’t think it is wise for anyone to say it is “all about the insulin”. I have mentioned this to others on their blogs who were over-interpreting our study. Clearly, the increase in insulin is ‘required’ but it is not alone ‘sufficient’. This is a critical distinction. There needs to be other factors.

JJ said...

And Finally! I have no vested interested in the outcome of these studies or any of the other ones. We started the experiment to look at the role of hyperinsulinemia in beta-cell mass expansion (something I have worked on for over a decade). When we noticed that the mice could not gain weight, we did the necessary studies to figure out why. Simple as that. Now we have a lot of new questions, some of them coming from interested parties on the internet, and we can address these in future studies. But trust me, I really don't have an attachment to the outcome. I'm a tenured professor and I get my (not so gigantic) salary regardless of whether high levels of insulin plays a role in obesity, cancer, Alzheimer's, or heart disease. If we do a better study that leads us to alter our conclusion, that is exactly what we will do. In science you have to stay humble, because we are have really simple brains and we are trying to understand really complex systems.

JJ said...

No vested interest

typos are my nemesis

Nigel Kinbrum said...

Jim, what was the composition of the high fat diet? I can't access the Supplemental Information. The Google Document said 58% fat, which is slightly different from D12492.

JJ said...

Below is the text from the Supplemental Methods. The control diet was a 'breeding' diet that already had a moderate amount of fat. Also, as you can see the diets are not well matched, even including the type of fat used. Thus, the comparisons are most valid within diet and between genotypes. We are now planning to repeat the experiments with a variety of better matched diets.

"Groups of mice were divided into two diet groups at weaning (3 weeks), one group was kept on the control diet (total calories = 4.68 kcal/g; 25.3% calories from fat, 19.8% calories from protein, 54.9% calories from carbohydrate; Catalog #5015 Lab Diets, Richmond, IN) and the other group was put on a high-fat diet (total calories = 5.56 kcal/g; 58.0% calories from fat, 16.4% calories from protein, 25.5% calories from carbohydrate; Catalog #D12330 Open Source Diets/Research Diets, New runswick, NJ)."

Evelyn aka CarbSane said...

Hi Jim! Nice of you to stop by. On the schooling thing, I don't see it that way -- it's a poke at those who like to call it that.

OK, let me ask a question when you get time. It's no rush as new comments pop right to the top of my reader and dashboard so I won't miss them. By what mechanism do you believe insulin drives obesity? The hypothesis from Gary Taubes (I see you've read his misrepresentations -- grin) is that when insulin is high it works to put fat into the fat cells and slow the release so it accumulates. So here is where LIRKO blows that hypothesis out of the water. Are you proposing a different mechanism by which hyperinsulinemia drives obesity?

This is what I mean by ignoring considerable data. LIRKO is but one hyperinsulinemic case where obesity does not develop. So while FIRKO supports the insulin hypothesis (I would argue that diazoxide and octreotide studies are far from convincing evidence, I hope to get back to some of Lustig's work there soon), a working hypothesis should be able to explain LIRKO. But this mouse stores fatty acids in adipose tissue just fine. The large variability in human insulin production irrespective of body weight would, as well, counter insulin being a driving factor.

The explanation you put forth on Peter's blog (and here) was that their insulin levels were too high and accompanied by hypoglycemia. Which is not the case in the study the table in this post comes from. There was only fasting hypo at 6 months but still fed hyper. I think several people are confused by this explanation you are putting forth. If insulin drives, more insulin should drive more.

I agree with "necessary but not sufficient". But, perhaps it's semantics, "driving" is not synonymous with this.

Kindke said...

But what gets the ball rolling for the beta cell expansion in the first place? Is it hyperglycemic spikes from high-GI + high-GL foods? or is it the fats we are eating?

Or is it even something else?

Evelyn aka CarbSane said...

Hi Jim, wrt UCP1, I just haven't seen in humans where improper fat trapping leads to "blowing off" the fatty acids by increasing oxidation in some way. Have you?

"most scientist and clinicians I know believe that humans are normoglycemic until they loose about 75-80% of functional beta-cell mass. "

How can that be if normoglycemia can be re-established within days to a few short weeks of (a) GBP surgery, (b) very low calorie "crash diet" and (c) early insulin treatment in large percentages of diabetics?

We'll have to agree to disagree on the relevance of LIRKO, but my point (going back to my last comment) is if the mechanism by which hyperinsulinemia makes us fat is favoring fat accumulation, then it really doesn't matter why the insulin levels are high, it is that they are. The evidence -- low NEFA -- seems pretty straightforward that LIRKO is able to store fat properly (unlike FIRKO and, incidentally, ASP deficient C3KO or ASP receptor knockouts C5L2KO). So high insulin + receptive/functioning adipocytes = obesity ... but not in LIRKO.

I haven't yet gotten around to the timed-feeding paper, but it was only the HFD that induced obesity. There were no differences for the normal diet between ad lib and restricted.

Evelyn aka CarbSane said...

OK... responded above before reading this one. So ... you are saying that modest hyperinsulinemia drives obesity, but excessive does not?

Also, others may ignore various IRKOs that are inconvenient, I do not. What is very interesting to me is that you can knock out EITHER insulin or ASP receptors on fat cells and protect against obesity and the adipose tissue takes on a similar distinctive pattern of large and small cells along with increased fatty acid oxidation rates and normoglycemia. FIRKO and MIRKO are interesting b/c they tend to counter the "backed up glucose disposal" model and LIRKO's hyperglycemia attributed to unchecked gluconeogenesis favors this as being at the root of hyperglycemia.

It seems to me that the beta cell is highly susceptible to lipotoxicity which was more to the point of this post -- the other stuff LIRKO demonstrates. There's a goldmine in Taubes' cited scientists that he never bothered to read the works of (well, he did skim the papers he paid students to gather for him ...) BTW, I'd love to take a look at this of yours:

Nigel Kinbrum said...

Thanks for that. Some people believe that your study shows that "high fat diets cause obesity".

I have a gut (!) feeling that a high fat diet based on natural unrefined foods (fatty meat + veggies, say) with the same P/C/F ratios as D12330 would produce a different result.

Evelyn aka CarbSane said...

In the case of the HFD, I'm putting my money on elevated NEFA. There's recent work showing that HFD acts on hypothalamus in days leading to inappropriate lipolysis from adipocytes.

Jane said...

Sorry. My comment about vested interest was totally uncalled for and I think you were very nice about it.

May I just mention one other factor. The high fat diet might have an effect on gut bacteria, which might in turn have an effect on mineral absorption. Fermentable fibre improves magnesium absorption, and high fat diets are expected to discourage the bacteria that do this. This might explain why you need a high fat diet for insulin to cause obesity.

Magnesium deficiency has been found in obese humans, and according to one report, their adipocytes have calcium overload. You will be familiar with Zemel's work showing that calcium promotes lipogenesis and inhibits lipolysis. It means keeping calcium low is very important, and calcium pumps that keep it low are of course dependent on magnesium-ATP. I think we need to know whether insulin is capable of causing obesity in the absence of magnesium deficiency.

Kindke said...

According to the breakdown, D12330 is

12.2% Maltodextrin ,
13% Corn Starch,
4 % soybean oil,
54% Hydrogenated coconut oil.
16% Casein.

Sanjeev said...

What's the point of the bolding?

Since it's not 'partially hydrogenated' the 6% of the polyunsaturates and 2% monounsaturates in coconut oil get removed. Completely.

It's now likely all saturated - isn't that the "good", maybe even the BEST fat, according to some?

Nigel Kinbrum said...

According to Surwit Diets, the fat is >99% saturated so there is virtually zero trans-fat content. There is also zero fibre content.

Kindke said...

Right, so the obvious conclusion we are left with is coconut oil is fattening?

Nigel Kinbrum said...

The only conclusion that I come to is that a diet of D12330 is fattening to C57BL6/J mice.

Sanjeev said...
This comment has been removed by the author.
Sanjeev said...

high fat diet, gut bacteria, magnesium absorption, magnesium deficiency

> Zemel's work showing that calcium promotes lipogenesis and inhibits lipolysis. It means keeping calcium low is very important,

so what's the suggestion here, less calcium containing food?

(this is the deleted comment from below, it didn't go in the right place ...)

Jane said...

No, calcium in the diet actually has the opposite effect, because of vitamin D. A high calcium diet means low adipocyte calcium, and can cause weight loss. See 'Calcium modulation of adiposity', Zemel 2012.

See also 'Scientists link obesity to gut bacteria' in the Financial Times a few weeks ago.

and this paper
showing that in obese children, serum Mg and Mg intake were both inversely related to fasting insulin, suggesting that the hyperinsulinemia of obesity might be due to Mg deficiency.

'Obese children consumed a higher percentage of total calories from fat and a lower percentage of calories from carbohydrates. Fiber intake was lower in the obese group... Other potential causes of lower serum magnesium in obese youth include reduced magnesium absorption secondary to higher fat intake and lower fiber intake. ...several human studies have shown that fermentable oligo- or polysaccharides enhance magnesium absorption.'

Nigel Kinbrum said...

A high Ca diet also increases energy loss due to faecal fat excretion. See Effect of short-term high dietary calcium intake on 24-h energy expenditure, fat oxidation, and fecal fat excretion.

JJ said...

Hi Jane,

I think you might be the first person I have seen on the internet take something back... thanks.

With regards to diets, both hyperinsulinemic and non-hyperinsulinemic mice had the same diet, so the difference has to be something involving insulin. Interesting, we saw a decrease in adipose tissue inflammation with reduced insulin.

Basically, we can't say anything about diets with our study, just insulin. This is the key point that I think is being misunderstood and over-interpretated in the bloggosphere.

JJ said...

"OK... responded above before reading this one. So ... you are saying that modest hyperinsulinemia drives obesity, but excessive does not?"

Yes. This is my interpretation of all of the available data. As I mentioned before, this does not surprise me given our work on the dose-dependent difference in insulin signalling.

"Also, others may ignore various IRKOs that are inconvenient, I do not."

I don't ignore them, I just take them with a grain of salt given the caveats associated with tissue-specific insulin resistance. Our model did not have insulin resistance, so we can separate the effects of hyperinsulinemia from the effects of insulin resistance.

"It seems to me that the beta cell is highly susceptible to lipotoxicity which was more to the point of this post -- the other stuff LIRKO demonstrates."

Yes. We do a lot of work on lipotoxicity in beta-cells. There are better examples than the LIRKO. Insulin is anti-ER-stress.

"BTW, I'd love to take a look at this of yours:"

Sure. Send me an email and I'll send it to you.

JJ said...

This diet only fattened the male mice.

Again, I don't think it is a good ideal to make conclusions about diets from our study. The diet was simply a means to an end (induce predictable hyperinsulinemia). The conclusions should be limited to the role of insulin. It will be interesting to examine other diets.

Nigel Kinbrum said...

Do FIRKO mice support the Insulin Hypothesis? I thought that part of the IH was that excessive insulin results in excessive uptake of glucose & FA by fat cells and that low insulin results in low uptake of glucose & FA by fat cells.

FIRKO mice have WAT & BAT insensitive to insulin which, according to the IH, should mean that FIRKO mice have minimal uptake of glucose & FA by WAT & BAT, resulting in reduced fat mass with hyperglycaemia & other hyperaemia.

However, FIRKO mice have normal uptake of glucose & FA by WAT & BAT. They also have up-regulation of UCP1 activity in WAT & BAT (due to lack of inhibition by insulin) which "burns-off" the glucose & FA as heat, resulting in reduced fat mass without hyperglycaemia & other hyperaemia.

Jane said...

Yes it's interesting about the inflammation. I've just checked your paper and seen you looked at oxidative stress in liver but not in adipose tissue. I can't see how you'd get inflammation without some degree of oxidative stress, and that would mean calcium overload because oxidation activates calcium channels and inhibits calcium pumps in the ER. Insulin would be responsible because it activates Nox4.

JJ said...

Immune cells are thought to be attracted to 'swollen' adipocytes. The calcium dependent ER-stress is reflected in the mRNA levels of Ddit (aka CHOP).

Jane said...

Ddit is CHOP, is it? So you're getting ER stress, which would explain the inflammation. Swollen adipocytes means dysfunctional adipocytes, I imagine. Did you look at Nox4? There is a Nox4 knockout that isn't thin, it's fat, which says a lot about the complexity of these things. I suspect adipocyte oscillations = calcium oscillations = redox oscillations, and swollen adipocytes aren't oscillating properly. The calcium will go into their mitochondria instead of the ER. Then, no ATP to pump the calcium back where it belongs. Calcium = lipogenesis instead of lipolysis, so the adipocyte swells.

Simon Carter said...

Great discussion and I think it was wonderful that JJ (Dr. Jim Johnson) stopped by to voice his opinion and answer some questions.

Post a Comment

Moderation is currently on. Thanks in advance for your patience.

Found something helpful today? Please consider a small donation. Thank you!

Where to now?

align="middle" align="middle"

align="middle" align="middle" align="middle"