Some Suggested Reading ...
In light of other recent events, I present some suggested posts from this blog on the questionable accuracy of the blogging of one Peter Dobromylskyj aka Hyperlipid.
Summary: The third post in a series (links to I & II within) outlining how Peter deliberately fudged the data from the "infamous" Grey & Kipnis study. He even told his readership he was doing it and tips his hat to Ancel Keys. I feel to this day that if anyone can read this series and take Peter seriously after that. It was done in all seriousness on his part.
Summary: Peter blogged on the LIRKO mouse with a "cute" quiz about making jam from the urine of this mouse. One problem, LIRKO doesn't get glucosuria. Other claims made are also easily dispelled including that this mouse's liver (and other cells) can't process glucose so they burn through fatty acids like mad and that's why they don't get fat, and that this mouse's muscle cells "internalise their insulin receptors".
Summary: Peter compares putting a mouse on a ketogenic diet to a FIRKO mouse while there is no evidence whatsoever that KD's alter insulin receptors on fat cells, "rendering them insulin resistant" in any way meaningfully comparable to knocking them out entirely.
Summary: Highlights two definitive statements made by Peter: "Hyperglycaemia causes insulin resistance. This is not controversial, as far as I am aware", and "Excess insulin causes insulin resistance".
Comments
"LIRKO mice are hypoglycemic in most papers. That is why they don't gain weight - way TOO much insulin. Well outside the physiological range."
I don't want to ask on Hyperlipid as that thread is a car crash!
I see that Evelyn has queried the first part of the comment on Twitter.
My question is this:- If too much insulin causes weight/fat gain, how can way too much insulin have the opposite effect?
http://www.dietdoctor.com/paleo-and-calories-2
"Based on a decade of experience treating obese patients, reading studies, going to obesity conferences and discussing this topic with the world’s biggest experts, here’s my best advice for maximizing your weight loss."
http://www.dietdoctor.com/how-to-lose-weight
I think the the LIRKO is a red herring, as are many of the other IRKO models. Each only examines the effect of tissue-specific insulin resistance. I don't think there is anyone that would argue that diet induces liver-specific insulin resistance. Even the MIRKO data are tricky, and that is the tissue most people think of being 'resistant' first and foremost.
Basically, I would argue, as I did in my paper, that reducing insulin (in the way that human genetics, environmental factors and drugs can do) is the best way to understand the integrated physiology of this hormone...
Also, Nige, I posted some thoughts on your blog about the effects of insulin on energy expenditure, or at least I tried to... they are below...
"Hi Nige...
I wrote the paper you are commenting on and I have been following it around the internet as it gets discussed/interpreted/misinterpreted. Insulin had multiple effects on the adipose tissue, including regulation of lipolytic genes. I think the evidence that insulin is important in obesity is strong. I don't see any reason why insulin's effects can't include effects on expenditure as well.
As for humans, no one has any idea what CHRONIC (life long) reduction in insulin hypersecretion would do the Ucp1 in WAT. The ACUTE (minute to minute) effects of insulin on EE are less important here.
I think a lot of people are missing the fact that the experiment represents 1/3 of the mouse lifetime and starts just after weening."
I think I'm going to avoid Twitter convos as the character limit really makes it difficult to have constructive conversations. For example, @ProfDocHealth made a comment about the liver's pivotal role in energy homeostasis.
On the one hand you seem to be saying that the results of your experiment show that insulin drives DIO, because if you produce an insulin deficient state (for which there is no equivalent in human physiology that I am aware) you can prevent DIO. This was translated into "hyperinsulinemia drives DIO". But as you can see I'm not the only one left wanting by your explanation of how LIRKO doesn't get fat. On Peter's site you said hypoglycemia coupled with extreme levels of insulin. I see that the mice in the paper linked last evening were mildly hypoglycemic, but that appears to be the exception to the rule for LIRKO which is almost universally described as hyperglycemic. In any case, you have a mouse that has 10X insulin but nominal suppression of glucose (112 to 88, stat sig?, http://www.ncbi.nlm.nih.gov/pmc/articles/PMC151923/table/T1/) It seems to be that you are saying hypoglycemia is incompatible with the development of obesity?
What is the mechanism by which you propose hyperinsulinemia causes DIO? How would hypoglycemia (if seen) interfere with this mechanism? It doesn't seem sufficient to just dismiss LIRKO as effed up, but say I'm being silly when I suggest that your mouse doesn't have an equivalent in the human condition.
If LIRKO's adipose was messed up, that would be an explanation, but this does not appear to be so. Forget obesity, since you study lipotoxicity, LIRKO appears to be pretty solid evidence that beta cell dysfunction -- e.g. deficiency -- is a function of NEFA. The pancreata of these mice seem capable of pumping out ever more insulin, GSIS and basal, to at least attempt to control what's being thrown at them. In IR and T2, the basal insulin is often high but GSIS is depressed. LIRKO doesn't suffer lipotoxic effects because fatty acid metabolism in adipose seems to be intact.
It is difficult to get a read on where you stand from Twitter bites. This cannot be simultaneously simple and complex. If demonstrating loss-of-function enables the conclusion that HI drives DIO, then your hypothesis should be able to explain cases like LIRKO where HI is accompanied by normal or even slightly reduced body weight. You have also made some sweeping statements about the translatability of mouse physiology to human physiology yet seem to dismiss when human conditions we actually observe are brought up in discussion.
In the beginning you were careful to point out that your paper referred to basal insulin secretion. And yet a goodly proportion of obese humans have normal basal insulin. So they got that way with elevated basal insulin and what? It went away?
Normal insulin secretion and signaling is almost certainly required for obesity to develop. This does not logically translate to excessive insulin causes obesity.
Eff these people. One big lesson I learned on my way to mastering my eating issues is, if you feel in your gut it's wrong, it's wrong.
Listen to your body!!!! (Remember that one, Evelyn?)
LIRKO mice have insulin signaling problems in the SREBC-1c pathway in the liver. From my understanding, this accounts for the low circulating NEFA in LIRKO. The studies/reviews I've read maintain the LIRKO are muscle/adipose IR.
Hi Dr. Johnson,
Thanks for joining in the comments. I'm curious, did you expect when designing your experiment for your knockout mice to become diabetic on the high fat diet?
Thanks,
Sara
If LIRKO are also adipose IR, this would reduce adipose storage & increase adipose energy expenditure which would explain lack of obesity. Got links?
I'm not aware that the primary mechanism by which insulin is purportedly causing obesity has changed: trapping fatty acids in adipose tissue. There is no indication that I've seen that LIRKO have impaired triglyceride storage capabilities like FIRKO, C3KO and C2L5KO (latter two are ASP deficient and receptor KO respectively).
I meant to suggest that since LIRKO mice have low circulating NEFA, they have low NEFA uptake into adipose tissue. Does this make sense? Granted these mice appear to have normal adipose tissue signaling, they still fail glucose tolerance tests, so their adipose tissue are not a signicant source of glucose disposal. The reason the LIRKO mice are not fat despite normal storage capabilities is that perhaps there is not much (i.e., low NEFA)to store. Does this seem plausible? I am just trying to understand the physiology, not trying to support a particular diet.
Thanks,
Sara
These mice eat about the same as normal mice. LIRKO does not appear to alter metabolic rate, fatty acid oxidation, etc. Thus by good old CICO, LIRKO is about the same size as normal mice. This explanation is consistent with what the authors of that study looked at, reported and concluded.
I have been on the road (still am) so twitter has been my only option most of the time. I only have a little time at the moment as well but I will try to address a few things.
1. LIRKO - the specific loss of liver insulin sensitive is not found in man or DIO mouse models, as you know. I agree that the lipotoxicity is interesting. I still question it's relevance and I think any one really understands all aspects of its phenotype. There are many interesting models (made by Colin Nicols at Wash U) of PHHI that have hypo turning to hyperglycemia. I think the best analogy related to massive hyperinsulinemia are patients with insulinoma, many of whom do not gain weight. I don't think anyone knows why. Perhaps very very high levels of insulin do have satiety effects that are less apparent at more physiological doses.
2. With regards to relevance. Humans have significant variation in insulin gene expression, the feature modeled in our mice, so they are surely much more relevant than the LIRKOs. Also, as many have noted, diet can affect insulin levels, both acutely and long-term. You will note that throughout my internet discussion I have avoided discussing specific diets. There is a reason for this.
I do think scientific problems can be simple and complex at the same time. The 'genetic logic' of our experiment is clear. The LIRKO and integrating all of the data on insulin is complex.
The sweeping statements are about mammalian physiology and insulin, for which we have the best available model. Better models will permit us to modify these statements.
What I dismiss (mostly) are the correlative/cross-sectional human studies. Finding some percentage of people who are obese with low insulin doesn't mean that much to me if we don't know what their insulin levels were when they were getting obese or even when they were developing and growing. It is like finding a lung cancer patient who doesn't smoke at a specific point in time, but may have smoked for 30 years and quit years ago, and concluding that smoking is unrelated. There is a strong correlation between hyperinsulinemia and obesity, but I don't much care for correlations. I worry even less about seeming exceptions to those correlations.
We have lots to still understand. I think the discussion is constructive, but overly focused on a few tidbits of information. I disagree that every mouse model have to elucidate the details of every other mouse model and every patient. There are many types of human obesity, and obviously to me, many causes. I see hyperinsulinemia as a requirement, but I don't see it as sufficient, especially at all ranges of insulin.
I have you book marked to look at new stuff you post…
My weblog ... teichpumpe 12v solar - -
Post a Comment
Comment Moderation is ON ... I will NOT be routinely reviewing or publishing comments at this time..