Insulin Action on Fat Cells Over the Course of Fat Gain
The idea that insulin acts on fat cells to promote obesity requires that insulin suppress fat release in people with more fat (or people who are gaining fat) to a greater extent than in lean people. As I have written before, this is not the case, and in fact the reverse is true. The fat tissue of obese people fails to normally suppress fatty acid release in response to an increase in insulin caused by a meal or an insulin injection, indicating that insulin's ability to suppress fat release is impaired in obesity (1, 2, 3). The reason for that is simple: the fat tissue of obese people is insulin resistant.
There has been some question around the blogosphere about when insulin resistance in fat tissue occurs. Is it only observed in obese people, or does it occur to a lesser extent in people who carry less excess fat mass and are perhaps on a trajectory of fat gain? To answer this question, let's turn the clocks back to 1968, a year before Neil Armstrong first set foot on the moon.
The question was first investigated by Dr. Jules Hirsch's group (4). They took fat biopsies from people with a range of different fat masses, exposed them to insulin, and determined the degree of insulin sensitivity of the biopsies. They found that insulin sensitivity of fat tissue declines as the size of fat cells increases. This was true across all cell sizes, not only the largest ones. As body fat gain mostly involves an increase in fat cell size rather than number, this suggests that fat tissue insulin sensitivity progressively declines as fat mass increases.
But they went further. They caused weight loss in their obese subjects using a calorie-restricted diet (15:45:40 protein:carb:fat), which shrunk their fat cell size. Following this intervention, insulin sensitivity in fat tissue increased, and both the blood glucose and insulin response to an oral glucose load improved considerably. They concluded:
Weight loss and reduction in adipose cell size restored plasma insulin concentration to normal, concomitant with the return of normal tissue insulin sensitivity.These data are consistent with the rest of the literature suggesting that elevated insulin and insulin resistance are the result of obesity. They suggest that excess fat mass, particularly enlarged fat cells, is the ultimate cause of insulin resistance. This hypothesis has been buttressed further since 1968.
The Relationship Between Fasting Insulin and Future Weight Gain
As a further data point, consider the review paper published in 2007 by Hivert and colleagues (5). They reviewed all the studies that examined the relationship between fasting and/or post-meal insulin level and future weight gain (there are a number of them). Here's what they found:
The majority of prospective studies that included non-obese adults failed to show an association between insulin level at baseline and future weight gain.
On the other hand, other large cohort studies have shown that insulin resistance, which is usually associated with high plasma insulin levels, could be protective against weight gain.High insulin does not generally predict future weight gain, and sometimes it even predicts lower future weight gain. This could be because of insulin's anti-obesity action in the brain, although that isn't clear because we don't know how hypothalamic insulin sensitivity tracks with hyperinsulinemia.
The Case of Diazoxide
Much fuss has been made about a study showing that the potassium channel activating drug diazoxide accelerates weight loss in obese people (6). One of the effects of this drug is a substantial reduction in insulin secretion, which is why it's used to treat hypoglycemia.
There are a number of problems with using this study to support the insulin hypothesis of obesity. One problem is that the finding has not always been replicated by other investigators (7). Also, this drug is already approved by the FDA for the treatment of hypoglycemia and hypertension. If it's so effective for fat loss, why isn't it being used as a fat loss drug?
A second problem appears as you sort through the first study's results. Basal metabolic rate and the proportion of carbohydrate and fat being used for fuel remained unchanged by diazoxide, suggesting that even if more fatty acids were being released by fat cells, they were not being burned at a faster rate, and thus they were also being re-incorporated into fat cells at an equally high rate. Think about this for a moment. Diazoxide decreased fasting insulin by 36 percent, and this had no effect whatsoever on fat burning or resting energy expenditure. This study confirms, ironically, that insulin does not regulate the net fatty acid flux of fat cells. Even if reducing insulin increases fat release from fat cells, if the fat is not burnt, it just does a loop through the circulation and ends up right back where it started. This is partially because insulin is not the main factor controlling fat re-incorporation into fat cells-- that job seems to be held by acylation-stimulating protein (ASP).
The fat loss coupled with unchanged basal metabolic rate means that either a) diazoxide made them start exercising a lot, and/or b) they ate less. Since I've never heard of a drug that causes obese people to run three miles a day, it was almost certainly (b). So did reduced insulin action on their fat cells make them eat less? Given that fat cells don't regulate food intake (except indirectly via their production of leptin, which acts in the brain), and the brain does, perhaps we should shift our focus to the brain for a moment.
But first, what is diazoxide? It activates ATP-dependent potassium channels, which are required for glucose sensing by the insulin-secreting pancreatic beta cells. But as the biologists in the crowd may know, these channels appear in a lot of places in the body. One of the places they appear is in the smooth muscle tissue that lines the arteries, which may be why diazoxide is used to treat hypertension. Another place they appear is in the brain, where they regulate the electrical activity that is intrinsic to neuron function.
As the hypothalamus is a critical area regulating food intake, it makes sense to see if diazoxide can influence the activity of neurons there. It turns out, diazoxide influences the activity of POMC neurons, one of the critical cell types that regulates food intake in mammals (8). As these cells are contained in the hypothalamus, a region that has a high blood-brain-barrier permeability, it is plausible that diazoxide actually exerts its effect there. The larger point is that diazoxide is not a specific drug-- it has effects on many parts of the body, so one cannot assume that its effect on body weight in some studies is due to a reduction in circulating insulin. The evidence on diazoxide does not support the idea that it causes fat loss by reducing insulin action on fat cells. Other mechanisms are more plausible at this point.
Drugs that influence food intake and/or body fatness usually do so via an effect on the brain. For example, rosiglitazone is an anti-diabetic drug that increases insulin sensitivity. One side effect is fat gain. Originally, it was postulated that fat gain was due to the effects of rosiglitazone on fat cells. Recently, it was shown by Dr. Jerrold Olefsky, in collaboration with my colleagues at UW, that rosiglitazone exerts its obesity-promoting effect mostly via the brain (9). This same story has repeated itself many times in the scientific literature, therefore whenever an intervention causes a change in food intake or fat mass, the first thing I think is "what's happening in the brain?"
High circulating insulin is probably an adaptive response to insulin resistance in the body, which develops as fat cells enlarge and become less effective at trapping fatty acids and keeping them where they should be (there may also be a contribution from inflammation that may or may not be independent of the changes in fatty acid handling). Elevated insulin is probably the body's way of trying to compensate for this defect and keep fat in fat cells, but it does not fully compensate for the insulin resistance in fat tissue that progressively develops as fat cells enlarge. Evelyn Kocur has written about this quite a bit. This defect can be largely reversed by fat loss, as demonstrated by the fact that a number of fat loss diets, including low-carbohydrate, low-fat and calorie restriction diets lead to improved insulin action as long as sufficient fat is lost.
I have pointed out the reasons why the carbohydrate-insulin-fat hypothesis is not generally considered viable by the scientific community. I feel I have convinced those who are able to be convinced. I can't convince everyone, and that's all right. It's time for me to move on from this topic, and on to more useful things!