 I'd like to welcome everybody this evening. Again, I'm Van Hubbard from the Division of Digestive Diseases and Nutrition within NIDDK here at NIH. And I welcome everybody to another one of our series of lectures on clinical nutrition and obesity. This evening, we are pleased to have Marcy Greenwood, who is currently Dean of Graduate Studies at University of California at Davis campus and professor of nutrition there. She originally got her degree at Rockefeller University in New York and was a part of their obesity research center and is eventually establishing an animal models corps while associated with that center. She moved to Vassar where she rose to be chairman of the Department of Biology and then in 1989 moved out to California. She has been instrumental in numerous committees throughout the US. She currently is chair of the Food and Nutrition Board at the National Research Council, National Academy of Sciences. She has been a member of that board for several years in addition. She's been a member of the Diet and Health Committee that was responsible for the publication Diet and Health, also out of the National Research Council, National Academy of Sciences. She's been involved with numerous reviews of scientific grants for the NIH, a past president of the North American Association for the Study of Obesity or among her past honors. And recently, she has been elected to membership in the Institute of Medicine at the National Academy of Sciences. We are pleased that she can join us this evening for her presentation on gender, genetics, and obesity. Marcy. Thank you very much, Dr. Hubbard. And I'm very pleased to be here this evening and to talk with you all about our current understandings about the impact of gender and genetics on the problem of human obesity. And what I'd like to try to do this evening is to talk with you a bit about what we understand with respect to the human elements associated with obesity. Introduce some of the research that our laboratories and others have done using models of animal obesity, primarily the animals who inherit obesity as a genetic disorder. And then to try to relate this back to human behaviors and perhaps put some perspective on why it is we're finding it so difficult to treat obesity and why it is that even in spite of the fact that many very talented professionals and a great deal of hard scientific work has gone into understanding the problem of human obesity that we still have such a great deal to learn and have such a great need to develop targeted innovative and integrated treatments. If I could have the first slide, just to review briefly what I'm sure a number of you probably already know, let me point out that a number of important documents have come out in the past several years on the relationship of nutrition and health or diet and health. This is the cover of the Surgeon General's report on nutrition and health. And among the many issues that it addresses, one of the issues it does address is the issue of the preponderance of obesity in the United States, the growing dimensions of this problem and the fact that we are increasingly seeing obesity indicated even in fairly young groups of children. Next slide. The diet and health report with which I had the pleasure of being associated also addressed many of the issues associated with diet and chronic disease risk. And as no doubt many of you know, depending upon the disease that you're looking at, somewhere between 25 and about 45 to 60% of some of the chronic disease risks can be associated with variations in diet associated with long-term risk. But particularly today, I want to talk about the problems identified by both of these studies and a number of others as well in which there is a considerable consensus in the nutrition community, something we haven't seen as strongly expressed in a number of decades. I want to talk about the problem of human obesity and its prevalence in the United States today. And to begin, let's look at the National Center for Health Statistics Second National Health and Nutrition Examination Study, sometimes called Haines too. And I'm sure some of you have seen these data. They've been popularized a lot over the past few years, and not notably in some of the recent obesity meetings that have been held both by the NIH and other groups in the country in the past year or two. An important point to notice here is that among men in the United States, from about 35 years on, somewhere between 30 to 40% of the population or overweight, when classified by race and age, as you see on this slide, there's not a marked difference in the percent of males overweight who are black or who are white. Although in these middle years, there is a slight preponderance of overweight among black males. These data, however interesting and certainly convincingly demonstrating that a substantial segment of the male adult population has a problem with weight, do not help to distinguish for us between those individuals who are only modestly overweight and those who are manifestly overweight with some of the more significant complications, both psychological and physiological, associated with marked obesity. So while these data help us to understand the preponderance of individuals who are overweight, they really don't help us in separating out what is increasingly getting to be known as the obesityes. Now, among women, the pattern of the prevalence of obesity or overweight is somewhat different than we saw with men. First of all, notice that in both black and white women, the percentage of women who are overweight by age category increases through the decades so that in postmenopausal categories, 36 or in some cases up to 40% of women are overweight using the standard criteria of heights and weights. If one looks at black women, you'll notice that up to 60% or more of these women are classified as overweight using this methodology in the adult population, particularly in the perimenopausal and postmenopausal period. So from the national statistics, we can glean two potentially important observations. One is that there are gender differences in the prevalence of overweight between males and females. There are gender differences in the age distribution of the prevalence of overweight between males and females. And that there is a potential difference between at least black and white classifications in the prevalence of overweight in these categories. I might add, although not done in this particular survey, that the Hispanic Hain study would show also a significant prevalence, higher degree of prevalence of overweight among Hispanic females than among white, although it is increasingly higher in all perimenopausal and postmenopausal women. Now, this epidemic, if you will, of overweight in the United States has presented a bit of a public health problem over the years and one of the reasons for this is indicated on this particular slide. One of the prevalent popular misconceptions about the obese American public is that we are becoming a nation of gluttons. And as the Diet and Health Report was trying to substantiate this strongly held opinion but scientifically difficult to prove conception, one of the observations that is consistently noted, and there are several of those in the Diet and Health Report and I've only chosen one to discuss with you this evening, but these are data looking at macronutrient availability and total caloric intake and body weight changes. And the macronutrient availability is basically shown by these lines at the bottom. And macronutrient availability, which is basically disappearance data, is frequently used as a surrogate for the population intake. And the important point to be seen in these data is that there's been very little change in the macronutrient availability, although its composition has changed some over the years from 1910 to 1985. And so you can't substantiate a marked overindulgence, if you will, of the population. Now this doesn't suggest that individuals who overeat don't become overweight or gain weight, only that you cannot invoke a national epidemic of gluttony as the prime culprit for the development, the developing problem of obesity in the United States. And this is particularly so if you look at a couple of the national surveys, for example, NHANES-1 and the Health Examination Survey, which took place in 1960, compares 60 to 70 data, in this case 65 to 77 data. And these are some of the best national databases we have in which body weight was actually measured and food intake was estimated using a variety of techniques. And I know that these bars are probably somewhat difficult to see, but the important point to notice here is that body weight in males, shown in orange and green for the females, is increasing while measurable caloric intake is actually decreasing in these particular groups. This then of course leads those of us who would be on committees that have to make national population public health recommendations into a significant dilemma. We have an increasing problem of overweight in the United States and yet significant and compelling data that increased caloric intake is not the culprit, leaving of course as the probable explanation in many cases for the population data, at least a partial component associated with either genetics or the environment. And one of the main components associated with the environment of course would be supposed changes in physical activity. Unfortunately, we don't have good longitudinal data on the physical activity or any reasonable surrogate in the United States and so we have to speculate that by default the changes in physical activity are as much a culprit as any purported change in food intake. And just to give us sort of an example that's sort of understandable to all of us, it doesn't have to be a substantial change in physical activity to get a substantial change in body weight. And one of the old examples that I remember from my earlier days studying exercise physiology is that if you're a secretary and you're using a manual typewriter, which of course hardly anyone does now so I really need a new example. But in any case, as you're using a manual typewriter, if you switch to an electric typewriter or to a computer and you don't do anything to compensate, for example, change your typing speed, you'll gain five pounds in a year. So it's a relatively small change in physical activity that can accommodate at least many of the sorts of differences that one can see in populations. And because of that, you'll see what some people consider fairly wimpy recommendations from both the Surgeon General's report and in this case from the Diet and Health Report where you find recommendations that say balance food intake and physical activity to maintain appropriate body weight, maintain a moderate level of physical activity and improve physical fitness. But no specific recommendation that caloric intake be reduced in the population. The other reason that such recommendations seem to be conservative and appropriate for population recommendations have to do with other observations which are particularly pertinent when one is looking at young and reproductive age males and females. And one of the consistent observation of a variety of studies, about eight studies now, is that females and males, of course, eat different total number of calories and females are eating about 1,500 and males eating about 2,500 calories a day. Now there's always a lot of dispute in the nutrition community of whether these measuring methods are accurate or whether they're underestimating or overestimating the actual caloric intake of males and females, but there is a type of internal consistency over a variety of studies that have suggested that these numbers seem to be, if not absolutely accurate, at least consistent with respect to their observations. So women are eating about 1,000 calories a day than males and they're eating a diet that has relatively the same nutrient density as you can see by looking at these nutrients represented here. They do eat a diet that is slightly higher in vitamin A and it is about the same otherwise. However, on this caloric intake of approximately 1,000 calories a day, different. If you look at the percent of the RDA, an estimate of, admittedly, a very crude estimate of the nutritional adequacy of diets, you'll note that for males on this nutrient intake, this dietary caloric intake, they are meeting the 100% or greater line for virtually all of these important nutrients, but most women, shown in green, are in fact substantially below this and particularly so in important micronutrients such as folicin, iron, and zinc, which are critically important to normal reproductive success. So once again, then, just based on these observations alone, one would be hesitant to ever make a population recommendation particularly since we know the young adolescent girls are so prone to dieting, even when they're not overweight, to make a recommendation to the general public that further reductions in food intake would be warranted. Once again, I wanna point out that I'm not suggesting that this means that individuals who have problems with their weight would not necessarily want to try to reduce their own caloric intake if indeed that was indicated. Another observation that I think becomes increasingly important as we try to understand the impact of obesity on health as opposed to physique or cosmetic areas of one's life is that it's not only important how overweight you are, how fat you are, what your age is, but it's also important where your fat is at, so to speak. Now this is the cartoon of a male with a typical distribution of fat that is characteristic of many men as they increase their body weight and it's called central obesity, android obesity and is characterized by subcutaneous deposition of fat or more importantly and more pathologically significant the increasing intra-abdominal deposition of fat. This pattern of fat distribution in males we now know has increasing pathophysiological salience and is increasingly important as we begin to understand which individuals who are overweight are at the greatest risk for associated pathologies such as hypertension and diabetes and thus perhaps those who most urgently need medical attention as well as dietary counseling. This three dimensional figure depicts the impact of increasing waist to hip ratio that is more of an accumulation of fat in the abdominal region with number three being the most centrally distributed individuals and also increasing adiposity indicated here as turtiles of body mass index with number three once again being the fattest and as you can see those individuals who are both the fattest and have the greatest amount of fat in their intra-abdominal region are those who are at the greatest risk for in this case having a stroke. However, one of the things we're learning is that the distribution of body fat can be critically important even if an individual is not markedly overweight. These are data taken also from the same rather well-known Swedish study now on all cause death and the remarkable thing here is that the individuals who seem to be at greatest risk in this particular surveillance study are those who had the greatest amount of fat distributed in increased waist to hip ratio but who were relatively thin. So that there's an indication that weight while weight is one component of the risk factors associated with obesity, the distribution of body fat may be increasingly and in some cases perhaps even more important than obesity per se. This is particularly important and is particularly striking when one looks at obesity in women because women tend to have at least two patterns of fat distribution. One is the central or android in honor of its contribution of androgenic hormones. One form is the android or central distribution of obesity and the other is the peripheral sometimes called femoral or guinoid distribution of obesity. Centrally distributed women have this increased fat deposition both subcutaneously and in some cases intra-abdominally. They also have as a group lower sex hormone binding globulin and are relatively androgenized relative to other women. This is not to say that they're here suit or markedly androgenized, they are just more androgenized than equivalently obese women with this pattern of fat distribution. The women with this particular pattern of fat distribution have a normal waist to hip ratio that is most women have smaller waist than they have hips. The ratio is normally about 0.8 and as long as they're developing their obesity in this pattern it is relatively less pathological than in this pattern. What do I mean by that? Well an example is shown on this slide. Now these are data looking at insulin and glucose levels in women of increasing adiposity up to 100 kilograms of body fat which is really quite a substantial degree of obesity about 200 pounds of body fat. So these individuals are 300 pounds or more. The open circles represent women who are increasingly obese but whose waist to hip ratio is normal. And the closed circles represent individuals who are increasingly obese but whose waist to hip ratio is of the android or the central distribution. And as you can see, those individuals with the normal distribution even as they become markedly obese show really substantially less of an impact on insulin or glucose than one sees with the individuals who have the increasing central distribution of obesity. And there are many studies now that have substantiated and expanded upon the impact of fat distribution on associated pathologies particularly diabetes and hypertension. And in fact these data such as these that have helped to shed some light on a particularly puzzling public health problem that continues to be of interest to researchers in the area. And that is that while there are substantially more overweight women than there are men in the United States they do seem to be at somewhat less of a risk for some of the complications such as cardiovascular disease, diabetes and hypertension. Now no doubt some of that is related to hormonal status but a part of it may be related to the fat distribution which may also be at least partially under sex hormone regulation. Now I'd like to take some time to talk about the impact of genetics as well as gender differences and fat distribution on the problem of obesity. And genetics can be conceptualized in a variety of ways from isolating genes and looking at the molecular effects of known genes or in fact identifying genes that cause obesity to the impact of genetics on cellular effects specific hormone nutrient and nutrient genomic interactions. Sex effects of course themselves are genetically based and fat distribution has a strong component associated with genetics. So let's review for a moment what we currently understand about the status of genetics as a potential explanation or partial explanation for obesity in humans. These are data taken from a study done by Garn and others at the University of Michigan some years ago in which they did a fairly simple study which was to examine the fatness of children and compare them to the fatness of their parents. And what these data show in a nutshell is that lean parents that is lean-lean tend to have lean children. They do occasionally have obese children but by and large they have lean children. And obese parents by and large have obese children. Very few of them have lean children. And there's something of a distribution in between if you have a lean parent with a mediumly overweight person this is the distribution. If you have a lean one, if you have one lean and one obese this is some of the distribution. And the distribution is reminiscent of although not completely concurrent with the distribution of a putative gene for the development of obesity. But Garn and others are frequently reminding us that obese families have obese pets and it is certainly the case that they have not conveyed any of their genes to their obese pets. And so while these are compelling data to suggest familial association they can't be used to prove in any scientific way that there's a genetic component of obesity. But indeed many lines of data have begun to converge in very convincing ways to demonstrate that substantial components of what we now understand to be the problem of body weight regulation does in fact travel in a genetic fashion. One of the important observations one of the important groups that we rely on in humans for understanding genetic effects is of course effects in twins. And these are dizygotic or fraternal non-identical twins. And what you see here is that in some pairs of these twins both the degree of overweight and the distribution of the fat is really quite concordant. On the other hand there are pairs where although the distribution of fat is reminiscent the degree of obesity is markedly different between the pairs as one would expect in related siblings but non-identical twins. On the other hand when one looks at identical twins you'll notice that not only is the body fat distribution extremely concordant but the distribution of the body fat is remarkably similar as well. And a number of recent studies have been published that have demonstrated that identical twins even raised apart in substantially different environments over very long periods of time tend to have a concordance for body weight within a few percentage points of each other. And this is remarkably true across ethnic and gender groups as well. An important set of observations also amplified upon and recently there have been quite a spate of papers coming out in this area has to do with some work done by Stunkard and his associates examining the Danish adoption register. And one of the great advantages of the Danish adoption register which is rather famous for a variety of studies is that it keeps information on both the biological parents and the adoptive parents so that it's possible to look at the adopted child's weight class at a certain age and compare it to the body mass index or the adiposity index of both the biological parents and the adoptive parents. And when that's done in this particular set of data you'll notice there's a good correlation between the weight class of the child and the body mass index of either the biological father or the mother. On the other hand, when one looks at the same child and compares it to its adoptive parents there's virtually no correlation at all between the weight of the child, the adopted child and the weight of the adoptive father or mother. In addition to this, Arlen Price, a human geneticist has recently published a number of studies and one that I'm particularly impressed with in human genetics fairly recently looking at a substantial number of kindreds and doing a careful genetic analysis of these data. And he estimates that the body weight regulation component or the obesity component in this substantial study he's done that about 25% of the variance in the body weights can be attributed to large effects of rare genes and about 45% of it or so can be attributed to cumulative small effects of multiple genes that affect body weight regulation and obesity. So there's converging evidence that there are strong genetic and biological components to this very vexing and difficult problem of obesity that so many individuals in our society are currently facing. And I think Claude Bouchard has said this particularly well. One of our colleagues from Canada who has been, who is currently conducting a series of very elegant studies on overfeeding and exercising in identical twins. And what he had to say was that nutrient partitioning that is the distribution of nutrients between muscle and fat is the single most important factor to explain the individuality in body mass gain. A significant proportion of those prone or resistant to obesity find themselves in this vulnerable or desirable position because of inherited or acquired differences in nutrient partitioning mechanisms. Which leads me to the part of my talk which focuses on cases where we know we are dealing with a mutant gene. And there are several strains of rodents, mice and rats who carry genes which are known to cause obesity. The OBOB gene and the DBDB gene, sometimes called the obesity and the diabetes gene in the mouse are mutant recessive genes. The condition requires homozygosity. That is, you must inherit one gene from each parent. The AY, which also causes obesity of all the way more moderate versions. Sometimes this is called the yellow mouse. This is a dominant gene. And the fatty gene, which is a rat gene, is now known to cause hypercellular, hypertrophic obesity in rodents as well. And in each case, the chromosomal location of the gene is now known. And you'll notice that even among the mouse strains, the gene is located on different chromosomes. And we now have its location thanks to our colleagues at the Rockefeller University on chromosome five in the rat. This fatty rat or so-called Fatha rat has been the focus of a great deal of attention in our laboratory over the past two decades. This is the genetically obese Zocrophati rat. And here is its lean littermate. Nothing has been done to these animals, except to allow them to eat ad libitum, and they are about 30 weeks of age now. And even my freshman, when I was teaching, had no difficulty telling me which animal was carrying the mutant genes. So there's quite a striking difference between these. This could be either the homozygous or the heterozygous lean animal. That is, it could be carrying one dose of the gene and would still be lean. The Zocrophati rat, like many of the cases that one sees in developing obesity in young children, cannot be told from a lean animal early on in development. That is, while they're suckling on their moms, while they're in their nursing stage, you can't tell who's going to become obese or who's going to become lean just by looking at their body weight. And we also know from a variety of studies that our laboratory and other laboratories have done as well that during the suckling period, the animals do not eat differently. One group, the ones carrying the gene for obesity are not overeating during the suckling period. So during this period of time when their body weights are essentially identical between the obese and the lean, their body compositions, however, are markedly different. So even though they weigh the same, the animal who's carrying the obesity gene is already at 15 days, almost twice as fat as its lean littermate. And this is exacerbated by day 22, which is approximately when these animals are weaned away from their mothers and onto solid chow. Probably more remarkable, I mentioned to you earlier that they don't overeat in the litter. And so one of the things that we were interested in was whether or not if you prevented these animals from overeating for most of their life, I mean the sort of simplistic notion that we often apply to the prevention of obesity in some of these terribly difficult pediatric cases, if we prevented these animals from overeating all of their life, would they in fact be lean? Here is the effect of chronic food restriction on body weight at 33 weeks of age. These are ad libitum-fed lean animals. That is, they've been allowed to eat whatever they want. And here's the ad libitum-fed fatty. And as you can see when they're allowed to eat what they want, the fatty weighs a good deal more than the lean. This animal, which is called the restricted fatty, has eaten the same number of calories throughout its life as the ad libitum-fed lean. But you'll notice that it is significantly heavier than the lean, and although it is somewhat less obese than the freely-feeding fatty, it is still notably overweight, even on the same caloric intake. And perhaps most depressing or most exciting, depending on whether you're obesity researcher or fat person, are the data shown on this slide. And I happen to be both. And that's, these are the data on body fat composition. You have an ad libitum-lean rat here with about 19% of its body weight as fat. And the ad libitum-fed fatty has about 52% of its body weight as fat. The restricted fatty, whom you will remember, weighed more than the lean but less than the fatty than the ad libitum-fed fatty has got actually statistically significantly more fat as a consequence of being prevented from overeating. So this particular gene, which is that we are studying in this rat, is extremely resistant to manipulation. Dietary manipulation, surgical manipulations, exercise do not return the LPL activity, oh, excuse me, do not return the body fat to the lean. One of the mechanisms by which this fatness accumulates has to do with the enzyme lipoprotein lipase. This is, as I'll show you on the next slide, is probably the most important enzyme in regulating the fat size and the fat accumulation of adipocytes. In lean animals, this is the activity level of the LPL. In the ad libitum-fed fatties here, you'll see this is the activity several fold greater. In restricted fatties, it's at least as high as ad libitum-fed and in many cases, it's statistically significantly higher in animals that have been food restricted. Unfortunately, it's also higher in animals that have been exercised. And that's been repeated not only by our laboratory but by a variety of other laboratories that most recently eckles. The LPL hypothesis that we developed earlier basically then stated that early developmental changes in adipose tissue LPL lead to increased deposition of triglyceride into fat in the absence of overeating and hyperphasia. Then as a result of the disproportionate deposition of the nutrients into adipose tissue, other growing tissues are relatively deprived. That, and that is manifestly true if you remember the body composition data I showed you a few slides ago in which while the animals are at the same body weight, one group is fatter than the other. Under those conditions, other organs are clearly decreased, most notably skeletal muscle. Consequently then, when the animals are allowed to behave, that is to eat as they want, they do increase their food intake. And the impact of this increased hyperphasia is an appropriate adaptive response to this nutrient flux caused by LPL. And they will become hyperphasia until their muscle mass reaches the size of a normal lean rat at which point the hyperphasia is markedly reduced. So it would appear that the overeating is caused by the metabolic nutrient changes rather than that any invocation of overeating can be implied as a cause of the metabolic change at least in this particular model. And most recently, the LPL story has become even more exciting because as we begin to understand the phagean story, and we don't know yet what it is that the fatty gene is producing. I have called it element Y here for lack of a better name, although we do believe it's a transacting factor that turns on a variety of lipogenic genes early on in the young rat or perhaps in the young human or other species life. LPL, however, is an enzyme that is made by the fat cell and which is exported from the fat cell to the capillary endothelium where it acts on the very low-density lipoproteins. So although some of these other lipogenic enzymes are elevated, LPL is the only one that's able to leave the fat cell and move to another site distilled to the fat cell where it is able to modify the nutrient environment of the fat cell by providing excess-free fatty acids, which are then reesterified stored as triglyceride. The other enzymes, which may also be coordinated, are very important in providing alpha-glycerol phosphate for the reesterification and the development of triglyceride. We have recently described what we think is a very interesting and exciting observation, and that is that fat cells have LPL binding sites. This was not known before. And the function of these binding sites and the regulation of these binding sites may be very important in the post-cellular regulation of LPL, especially since we know that the increased LPL is not a simple change in transcription, but is rather more markedly attributable to post-translational, and for reasons I don't have time to go into in detail this evening, probably post-cellular mechanisms. So the observation that there is a pool of LPL that can be held on the surface of the fat cell released rapidly and regulated, and which may also regulate this whole cycle of providing free fatty acids may indeed be an important new observation that will allow us to understand how LPL relates to obesity, especially in the early developmental stages. Another observation that is important is that free fatty acids are becoming increasingly important as signal systems in differentiation. And one of the consistent observations, particularly in marked obesity, where you see both hyperplasia and hypertrophy, is that hypertrophy almost always precedes hyperplasia, and it may very well be in a situation such as this, that the free fatty acids are first taken up and reesterified and deposited in pre-existing preadipocytes, but that the free fatty acids are also acting on precursor fat cells, causing them to differentiate and for your organism to become hypercellular. That would be very consistent with the pattern of development of the obesity that one sees both in young developing obesity and also in adult cases, where we do see hypercellular obesity, first the hypertrophy occurs, then the hyperplasia occurs, and so we've become very interested in examining the impact of various free fatty acids on the rate of differentiation of pre-adipocytes from obese and lean cells in culture. Now, one of the observations that has been made recently is that the fatty gene in the zucca rat may be homologous to the DB or the diabetes gene in the mouse. This is a particularly interesting observation because the DB gene in the mouse when expressed on different genetic backgrounds has varying degrees of diabetes associated with it. And we've recently become interested in a model which we think shed some light on the interesting issue of whether or not the same gene in different genetic backgrounds produces different and associated pathologies. This is an important issue in human obesity where we have many obese individuals who do develop diabetes or hypertension, but others who are equivalently obese who don't develop the obesity or the diabetes. And this is the Wistar fatty rat, has the same fatty gene, but it's a different rat. It's an albino strain of rats and it was produced by our colleagues in Japan to cater industries. What's interesting about this obese rat is that much like the zucca rat, both the males and the females get overweight if you just allow them to overeat compared to the lean males and females. But if you feed all of these animals, obese and lean, a high sucrose diet, you'll notice that the fatty male becomes markedly hyperglycemic. The lean males and females of this strain do not respond with hyperglycemia to a high sucrose diet fed over as long as 22 weeks of age, 22 weeks of experimental period. The females show some peaking of hyperglycemia, which then comes back to normal and a number of experiments never show this peak either. But so there's a striking sexual dimorphism in this expression of the diabetes, but more importantly, it's the same gene on two different rat ethnic backgrounds, if you will, with markedly different expression of the diabetes associated with the obesity. There are many aspects of this particular rat model that I could discuss with you this evening, but time being relatively limited, let me just tell you that this rat may be a reasonable model of human non-insulin-dependent diabetes. It certainly has the obesity. It develops glucose intolerance and insulin resistance, dietary influences, such as the impact of high sucrose in the diet and the form of the sucrose as well influence the expression of the diabetes. There do seem to be differences between males and females. And when you compare the Wistar fatty in the Zooker strain, you get substantial differences, not just in the expression of the hyperglycemia, but in the response of these strains to manipulations such as adrenalectomy or other manipulations which are known to affect hyperglycemia and or obesity. So in a nutshell, what I'm saying is that when genetics plays a role, dieting alone is not enough to reduce an individual to a non-obese weight. And that raises the important issue which has been in the press a good deal lately on my dad. What happens when diets fail and repeated weight loss and regain occur? The answer to that question, which I should probably tell you straight out, is that we really don't know, although we are beginning to have both an increased interest and an increasing understanding of what weight fluctuations mean with respect to long-term risk. As an example, as an example, let me point out that repeated cycles of weight loss and regain are not unusual. This happens to be a weight history of a female patient in which the weight gain in this particular case has been associated with both reproductive events and some psychosocial events, depression and stress. But you could find many weight curves like this, some of which have children involved and some of which have nothing involved except a continual problem, trying to lose weight, often regaining it and frequently shooting above the pre-diet weight as a consequence of relapse. A common phenomena and one that the recent technology assessment, NIH technology assessment conference addressed in detail, there's no question that many individuals who try to lose weight do not succeed in treatment and even those who do lose the weight frequently gain it back. Some years ago, our laboratory and the laboratories of Kelly Brinnell and a number of other investigators became interested in a question of whether or not weight fluctuation independent of net weight gain or loss had an impact on long-term morbidity and mortality. And these data were taken from the famous Framingham study in which they have followed groups of males and females over a number of decades now. And without going into many of the details and let me tell you they are indeed laborious statistically, and this is the work by Lisner and Brownell. The important point to note here is whether you're one is looking at all-cause death, cardiovascular death, or cardiovascular risk factors, the group with the most fluctuation in body weight showed substantial increments over the weight stable group and the middle group was somewhere in between. These data are hardly conclusive but they do suggest that at least for some individuals weight fluctuation can be an independent risk factor. And this points perhaps most compellingly to the importance of developing weight stabilization programs and the problems of weight maintenance for most individuals. This is a particularly disturbing trend if true. I mean, if it's true that weight fluctuation has an independent risk factor, then one thing that one might be concerned about and certainly a number of us in the field are, and that is that dieting is a cultural phenomena in young pre-reproductive age women, women who are not overweight, who are constantly losing and regaining five or 10 pounds. And if in fact there is a separate independent risk of dieting behavior, which may set them up later for weight gain during their reproductive years or perimenopausal years, then this is something that it's important to understand the magnitude of because it does have significant educational and public health components associated with it. And it may be this group to which many of the diet, much of the dieting articles in popular magazines and papers are pitched to this particular group of young reproductive age women who don't really have to lose weight. To try to study this in some animals, we did a study a few years ago in which you looked at males and females who were, in fact, there were several additional groups in this study. I only pulled two out to try to keep it relatively simple. And we looked at groups of rats who were not cycled that is maintained on their own diet and then groups of rat who were simply fed a low, a chow diet, which is a low fat diet and then allowed to refeed on a diet of their choice. They could select protein, fat and carbohydrate as they chose. And we looked at them as they went through two diet cycles, both males and females. And there are two things to notice here. One, with the males, they lost weight and they regained right back to their pre-diet weight. With the females, one initially thinks that one is seeing that the female cyclers actually show a lower body weight at the end of this experiment, thinking that the dieting behavior may have actually been quite helpful. And I apologize for the fact that I left, I did, I left a slide out of the male-female comparison. But what I wanted to tell you was that even though the females weighed less than that study, they were equally as fat as the high fat cycling females. So that even though they weighed less, they were maintaining their nutrient distribution and were equally obese. If you look at just cycled and non-cycled individuals, the cycled individual rats had higher plasma insulins. And if you look at two of their inter-abdominal fat pads, the perimetrial and the retroperitneal, the cycled animals had more adipocyte, had more fat in these pads than did the control animals, even though you will remember that the control animals, in some cases, weighed less, for example, in the females. This is also an interesting comparison with the females and the males. This relates to the slide that I just showed you two slides ago. But in the females, the dieters were redistributing their fat so that they had more of the internal fat, or inter-abdominal, if you will, compared to the males. Even though, in fact, this group, the high-fat fatty animals, the black group, were fatter than the dieters. In addition, if you look at the dietary choices of repeatedly reduced and then regaining rats who are allowed to freely choose the composition of their diet, if you look at the speckled bar here, which are the cycled animals, it would appear, and we've now done this in several experiments, that as the animals regain, they choose diets that are higher in fat compared to the normal controls who are also self-selecting their diet, but who have not lost weight. So there are some quite intriguing data, at least in animals, that suggest that dieting does have an impact on fat distribution, on degree of adiposity, and it may have an impact on dietary preference during the regain period. And weight cycling is a controversial topic, I'd be the first to admit that, but it may be associated with decreased effectiveness of weight loss methods. There are sex differences in the effects of weight cycling. It may redistribute body fat. I didn't have time to show you data, but there are several experiments which show that weight cycling in animals, and possibly in humans as well, increases blood pressure. It certainly increases insulin levels in rat studies, and increases fat preference, and I think I can show you some data in a moment that's suggestive in humans as well as in rats. Therefore, weight cycling effects are really an important area for future research and are of quite popular interest at the moment. Let me return for just a moment to the issue of lipoprotein lipase and weight loss and regain in humans. Normal individuals here, obese individuals here, but note that reduced obese individuals have notably marked an increased LPL activity. Regardless of whether their weight loss is only of a duration of several months or several weeks, I think this is actually, or several years. Along with this increase in LPL activity are some observations that have to do with fat preferences in humans. These data were collected at the University of Washington with my colleagues, John Brunzel, Adam Dronowski, and others. And what we did was we looked at the sugar and fat suite and fat preferences of obese and lean individuals and of reduced obese individuals. And I know these lines are a little busy, but it's okay because the only point to notice from them is it does not matter whether you are normal or obese or reduced obese, you are equally as accurate at estimating the increasing sugar content, fat content of the solutions that you're being asked to taste. And what you do in this particular psychophysical testing paradigm is to ask individuals to test, to taste test about 20 solutions and they taste them, spit, rinse, spit, rinse. So they don't ingest any nutrients in this case, they're just rating the intensity of the sweetness or the fatness and their preference for the sweetness or the fatness. We use two adjectives for fatness because psychophysical testing for fat was in its infancy at that time and we weren't sure which adjective would best track the fat composition of the solution. As it turned out, it didn't really matter. These are two-dimensional representations of the fat preferences. Normal individuals are indicated here. You probably can't interpret this too quickly, but trust me, this indicates that normal individuals like a sweet and fat solution, which is roughly the composition of whipping cream or good hog and oss ice cream. Now this may not be what they tell you that they are going to eat or their dietary records may suggest that they have, but when you're just in a testing paradigm, that's what they like. Obese individuals have a very mild sweet preference. It's been known for years from a variety of psychophysical experiments in a number of laboratories that sweet, that fat people don't have a sweet tooth. In fact, if anything, they do not prefer sweet as much as normal weight people do, but what does happen is that a very little bit of fat, I mean, a very little bit of sugar makes a lot of fat very palatable. They have a very high fat preference. You basically can't get a solution with a little bit of sweet fat enough for some overweight individuals, for obese individuals to prefer. Reduced obese individuals pick up both the normal sweet preference and maintain the obese fat preference so that during the reduced obese condition, while they're in the psychophysical testing procedure, where they're not ingesting anything, they're just rating solutions, they have a pronounced and enhanced sweet preference and fat preference. Whether this is metabolically based on something like lipoprotein lipase or some other component of weight loss is still to be determined, but it is of some considerable significance, I think, in terms of the types of treatment one might devise or the types of foods one might devise during the weight maintenance or potential regain phase. Now, just to return briefly to the weight cycling issue, I'll show you some data which are really, frankly, quite tentative and they're pilot data, but they're interesting and I thought you might enjoy seeing them this evening. These are some data collected in a collaborative study with Dr. Rodin at Yale in which we looked at fat preference in women who were going through pregnancy and the top solid line indicates individuals who have been identified as having a history of weight cycling and the dotted line with those individuals who have not shown a history of weight cycling and we use a weight cycling questionnaire to separate it out and we only take the bottom quartile and the top quartile and as you can see early on before these individuals become pregnant, the cyclers have a higher preference for high fat solutions than do the non-cyclers. It dips during the first trimester of pregnancy for reasons I'm sure a number of the women in the audience could explain and then it tends to go back up so that by the end of the pregnancy the fat preference is the same or higher than it was before the initiation of the preference. Looking at their dietary records, there was a statistically significant difference in the percent of calories ingested by these same individuals in their routine diets. It was not a huge difference but it was statistically significant in this population and this was correlated interestingly with very small but once again statistically consistent and significant differences in measures such as blood pressure and another one is blood glucose which I don't have with me this evening. So what does this all mean? Well, it certainly doesn't mean that I'm suggesting that individuals who are overweight shouldn't diet. What I am suggesting though is consistent with a good deal of what we know about the mortality ratio for overweight and if you look, this is taken and modified from the Diet and Health report and the Surgeon General's report and if you look at the mortality ratio of individuals with body mass indexes from about 19 up to about 27, this is this range here. You'll notice that the risk is normal to low and this can be a substantial weight range particularly among reproductive age women much as 20 to 40 pounds depending on how tall you are. What these data suggest is that in this range where individual body weight can fluctuate substantially there are a variety of health relevant risk factors which modulate overall risk and weight is only one of them and this is the area where factors such as physical activity versus dieting may be particularly salient in terms of long-term risk. On the other hand, when one gets to body mass indices above about 27 or 28, you'll see that not only does the risk increase but it dramatically and logarithmically increases. So that it is probably helpful to distinguish among these groups. When one is talking about the potential risk of weight cycling or weight loss, it may be higher for this group than it is for this group. Something which is often missed in the press reports which are sometimes interpreted to mean that investigators interested in weight cycling are in fact suggesting that dieting is not good at this degree of obesity and that's really not what most people are suggesting. What they're saying is that in this risk group the risk of dieting alone may be greater than any weight change in physical activity, low-fat diet, smoking, not smoking, alcohol, may be as important or more important in terms of overall morbidity and mortality but in this category one can hardly deny that weight is a significant and independent risk factor for the development of morbidity and mortality. These are the desirable body mass indices. I won't go through them for you. The point I wanted to make though is that body mass index as opposed to weight height is the normal way of looking at adiposity right now. It has the great advantage of being gender neutral and also as you'll notice, generally speaking, there are differences, slight differences regarding the decades as one ages. However, please note that the top of the range for the youngest group is still within the range of the upper group so this doesn't suggest that every individual needs to be gaining weight as they age. So what to advise? Well, very importantly and certainly stressing diet composition over caloric restriction is a cogent and reasonable recommendation particularly as we are beginning to understand that those individuals who can maintain a low-fat diet often have slow but steady weight losses over a period of time and the low-fat diet and reduced alcohol intake, et cetera can be far more important in terms of long-term wellbeing. And a tried but true recommendation, sorry that I can't bring you the silver bullet this evening, is to stress physical fitness and permanent lifestyle change. These are recommendations that all of you I know are aware are extremely hard to affect in many populations. No amount of goodwill and hard work on the part of both clients and health practitioners has provided a program that we can with certain to say is going to be the success story for every individual. So on that note, let me stop and take questions and thank you very much for your attention. Thank you. Can you affect the body composition of the F-A diet-restricted rat by making you run on a treadmill? That's an interesting question. The question is can you affect the diet composition by making a rat run on a treadmill? Well, those experiments have been done. Judy Stern's done a lot of them. We've swum them four hours a day, six days a week, kind of studies you can do with enthusiastic students and the rats get so that they just love swimming and they jump off your hands into the thing. And interestingly enough, at the end of all of this, no, you do affect body composition, you improve muscle mass, but you don't, and it's a modest improvement, a five or 10% improvement, but you do not prevent the development of the obesity. We have run them in treadmills or in activity wheels or swum them from weaning on for as long as six to eight weeks, which in a rat's life is a long time. And although you improve their muscle mass and therefore shift the fat to muscle ratio somewhat, you do not reduce their body fatness markedly. I was specifically asking about the diet-restricted ones. Yes, oh, I'm sorry. Judy's done that, both on the treadmill and in swimming. And the thing that's really depressing about that is that the diet-restricted ones don't even get the special effective improved muscle mass. And they have increased LPL activity compared to all other groups. Now, I don't like to leave an audience with the feeling that this is just overwhelming. Please keep in mind that one of the reasons that we're so interested in the scientific value of some of these rodents is that they are effectively simplistic explanations. I mean, they are single known, homozygous genes in a model system that you can control the diet and the physical activity of. It may be, given Arlen Price's recent data, that single genes with large effects are going to be a partial explanation for the development of obesity in substantial numbers of humans. But probably it's more likely the cumulative effect of common genes with small effects that are subject to modifications by the composition of the diet or the degree of physical activity or the exposure or non-exposure to steroids that people take either as contraceptives or as a consequence of being pregnant or lactating that may have an impact. So I'm not prepared to say that this represents human obesity, but please do keep in mind that there are cases of human obesity where you see these 80 pound young children at the age of three. And I cannot think of any explanation or any set of animal experiments that's ever been done, even with substantial overfeeding from birth, where you can even come close to producing that degree of obesity at that age in a rat. Barbara Hanson's actually here and she might contest that over. I mean, you can certainly produce very fat monkeys by overfeeding them, but not in that timeframe. Barbara? Marcy, there's been a lot of publicity lately about the interpretation of the variability in weight data. Can you comment a little further about the relationship to mortality and possible causes? Since it seems to me none of these have really dissected out the cause of the weight variability from the cause of death. That's, I mean, Barbara, Dr. Hanson brings up a very important point here, which is that we're describing a phenomenology here and not a mechanism for understanding a scientific basis of our observations. And unfortunately, because of the nature of the data that we have on humans, it's hard to go too much beyond that because there haven't been any long-term trials in which there's been a purposeful decrease in body weight and regain, decrease in regain. What may be, I mean, I think probably the potentially most interesting type of explanation is that weight fluctuation could be shifting dietary preferences and could be shifting fat distribution. That would be consistent with an independent risk even in the absence of net weight gain or loss of some of the pathology that we see, insulin resistance in particular, hypertension and diabetes. But whether that's true or not remains to be shown. At this point, we don't even have data on substantial numbers of individuals who have weight cycled to even look at a simple question like are they, do they have more interabdominal fat at the same weight than individuals who haven't weight cycled? If they do, then we have a potential mechanistic explanation. If they don't, then I don't think we do. At least not one that I can think of this evening.