The Skinny Fat
From an outsider's perspective, obesity seems like a simple problem to solve: Eat less, exercise more. But, the body regulates food intake and feelings of satiety as part of a tightly regulated homeostatic process. Once a person becomes obese, it's these same regulatory feedback loops that also defend the obese state as the new "normal." Losing weight is difficult, in part, because of the starvation signals that your body sends in order to keep your weight constant.
Today a full third of all Americans are obese, more than 50% are overweight, and 300,000 die annually from obesity-related metabolic diseases such as diabetes, hypertension, cardiovascular disease and cancer....
Most therapies are geared at blocking fat absorption or curbing appetite, yet the precise contribution of overeating to obesity is unclear. Studying diet in obese patients is confounded by the fact that these patients tend to underreport their food intake by as much as 30%. Overeating can be gauged only in relation to that individual's energy expenditure.
We're missing a big opportunity by focusing so much on the molecular control of food consumption. If obesity comes from an imbalance in the "energy consumed-energy burned" equation, we should also be focusing on the molecular basis of energy expenditure. That's exactly where I've devoted much of the last 10 years of my research.
I haven't always been interested in issues of obesity and metabolic disease. As a biochemist and cell biologist back in the early 1980s, I was fascinated by the coordination of cellular systems in development. We used adipocytes as a model to study developmental signaling. It was a lucky choice, as the United States was beginning to notice the growing numbers of obese people. That the study of fat cell differentiation and the study of metabolic disease would converge was, with the benefit of hindsight, inevitable. Tracing the connection between fat cell development and the prevention of obesity has been a long and eventful journey.
When I started my own lab at Dana Farber Cancer Institute and Harvard Medical School in 1982, I was committed to researching fat-cell regulation. More than a decade of study led me, in 1998, to the metabolically hyperactive member of the fat-cell family that is abundant in infants but almost absent in adults. Our search for a trigger that could be important in obesity also revealed other mechanisms involved in diseases unrelated to obesity.
The number of fat cells does not increase in adults with morbid obesity. This suggests more fat cell differentiation in the development of morbid obesity. My lab's major goal starting in 1982 was to identify the molecular switch that triggered fat cell differentiation.
I was interested in the fat cell differentiation pathway as a model for understanding a basic process that might be applicable in other differentiation systems and to the loss of cell differentiation in cancer. We first identified genes that were expressed specifically in fat cells compared to preadipocytes, fat cell precursors. Then we looked upstream for the enhancers and transcription factors that turned on the genes.
We had a catalog of hundreds of genes that were turned on during fat cell differentiation, but among them the aP2 gene stood out because it was abundant and seemed an appropriate model system. So we set out to find out what controlled expression of aP2, and that was how we first came across the peroxisome proliferation-activated receptor gamma (PPARγ) in 1994. Researchers now recognize it as the "master gene" of fat cell development as well as the receptor for a few antidiabetic drugs. Reed Graves and Peter Tontonoz spent years dissecting the aP2 enhancer and found that one DNA binding complex, which we called ARF6, seemed to have all the properties suggesting that it could be the "key" factor. Peter and Erding Hu cloned ARF 6, and showed it was PPARγ in complex with RXR. PPARγ had been just described as a new member of the nuclear receptor family, but nothing was known about its function. When Peter transduced the gene for PPARγ into fibroblasts using viral vectors and provided a ligand, they differentiated into fat cells.
We were ecstatic, because this suggested that we might have found the switch for fat cell formation. I think that this was one of the most important findings that's come out of my laboratory to date.
It took quite a few years of work and the combined effort of several labs to work out the pathway by which PPARγ functioned. We now had a good idea of how to turn on the production of new white fat cells. Other labs had shown that PPARγ also played a key role in brown fat differentiation, which was a little surprising, given how functionally different the two cell types are.
Brown fat is commonly found in infants, where the cell type helps generate enough heat for the child's survival. Brown fat cells have no reason for existing except to generate heat. They do this by leaking hydrogen ions across the inner membrane of the mitochondria, generating heat, instead of converting it into ATP for other metabolic processes (see Energy-burning baby fat).
While it's abundant in infants, brown fat is almost absent (or at least difficult to find) in adults. Both cell types can originate from the same precursor cells, so we thought (and still think) that if we could find the specific trigger of brown fat, we might be able to increase the number of brown fat cells in obese adults. If we could find the right switch, we'd have a way for obese individuals to siphon excess energy into heat via brown fat cells, rather than into the storage chambers of white fat cells.
Pere Puigserver and I decided to try to find the fat cell-specific triggers of PPARγ. We screened brown fat gene libraries for molecules that interacted with PPARγ, and in 2004 we published our findings on a molecule that bound to PPARγ and activated genes important to brown fat-specific differentiation. With great originality, we called this protein PGC-1, for PPARγ coactivator-1. When the gene for PGC-1 was expressed in white fat cells, it increased mitochondrial respiration and mitochondrial biogenesis and induced the expression of an uncoupling protein (UCP) that makes mitochondria energetically leaky. Instead of retaining the energy that mitochondria produces, the UPC causes its release as heat. It's the reason why brown fat is so much more abundant in hibernating animals and babies who cannot rely solely on shivering to provide sufficient heat.
PGC-1 alpha has turned out to be a key regulator of those brown fat genes linked to mitochondrial biogenesis and thermogenesis. It is also a key regulator of mitochondrial gene expression and respiration in many if not most tissues. In fact, PGC-1 alpha is turned on in many tissues by external stimuli such as exercise (in muscle) or cold (in brown fat), and it mediates the effects of those stimuli on mitochodrial number and function. Moreover, PGC-1 alpha also causes muscles to switch their fiber-type to more oxidative fibers - something that consistent exercise is known to do.
Most recently we realized that although PGC-1 alpha controls the mitochondrial biology and thermogenesis of brown fat, it does not control all the genes characteristic of brown fat. Patrick Seale and I reasoned that if PGC-1 turned on multiple, but not all, brown fat functions, there had to be another regulator that was located upstream genetically that was acting as the "master regulator" of brown fat differentiation.
We assayed a published database of all the transcriptional components of the mouse genome, around 2,000 of them, for components that were specific to brown fat. We then screened the subset of the 20 that we found and discovered that only three genes were selective for expression in pure brown cells versus pure white cells. Of those three, PRDM16 was the only one that fit the bill. It was almost nonexistent in white fat. Moreover, PRDM16 not only activated the PGC-1 gene, but it also ramped up expression of nine other brown fat-specific genes that we had identified when we expressed it in fibroblasts, destined to otherwise become white fat cells. When PRDM16 expression was knocked down with interfering RNA, brown fat cells lost their characteristic phenotype and looked more like white fat cells, from a genetic perspective. We published our results in July 2007.
It wasn't too far-fetched to think that if we found a way to activate PRDM16 in human white fat, we could tip the energetic balance in the direction of passive energy expenditure. However, we could get the brown fat phenotype only when we turned on PRDM16 in undifferentiated adipose cells in culture, but not those that had already become white fat cells. While this result may seem discouraging in terms of applying it to humans, it's not as bad as it seems. Although preadipocyte cells are not abundant in adults, a change of 1-2% might be enough to create a significant change in obesity over time.
I'm betting that there will be an effect on obesity in humans. We've recently started working with the Broad Institute in Boston to screen every FDA approved drug for possible effects on PRDM16. Just because a drug is being used for another purpose doesn't mean it won't have other effects that we haven't yet discovered in the body. Also, we're looking for a small change over time that might not be picked up by clinicians. If we find an approved compound or combination of compounds that work, they would already have safety and pharmacokinetic information associated with them - immensely speeding the process of bringing the drug to patients.
I had set out with the goal of understanding differentiation pathways. It's fortunate that the model system I used happened to reveal pathways that are so important in basic physiology.
While looking for the genetic switch for fat cell differentiation (brown and white), we realized that these brown-fat molecules, PRDM16 and PGC-1, were important in many other pathways. When we knocked out the PGC-1 in mice, the first and most obvious effects we saw were the changes in the brain. Although we didn't see changes in body weight in this model, the mouse was very sensitive to cold, showing an inability to generate heat through uncoupled mitochondrial respiration.
How can brain functions share the same pathways as fat cells? The whole point of brown fat is expanded and accelerated mitochondrial biology. Brown fat produces energy through a proton leak process that releases heat instead of converting the energy into ATP. The brain, like the heart, is extremely expensive energetically, and these are usually the first organs to show the effects of malfunctioning mitochondria. Our PGC-1 knockouts were much more sensitive to oxidative damage in the brain from reactive oxygen species, which the mitochondria produce. In 2006 we published our findings showing that when the gene for PGC-1 alpha is expressed in cells, it increases mitochondrial respiration while at the same time increasing the enzymes responsible for detoxification of the reactive species.
We're very interested in the possible role these two molecules have in the brain, and there are suggestions that PRDM16 may be important in Huntington's disease. It may seem dangerous to try fiddling with genes that are so important in regulating mitochondria - the cell's powerhouse. But, to be useful in obesity we're looking for a drug that might change the balance by 1-2%. Over time and with other management methods, that could make an important difference for an obese individual.
The more we learn about them, the more PRDM16 and PGC-1 alpha appear to be crucial for many functions. Indeed, researchers studying different tissues in other fields could have discovered them first. But because brown fat so emphasizes these pathways, they've been the best tool for finding the "master" mitochondrial regulators. If we ever had had a really good idea, the study of this cell type was probably it.
Bruce Spiegelman is a professor of cell biology at Harvard Medical School and the Dana-Farber Cancer Institute.