In the Footsteps of Wilbur Olin Atwater: The Atwater Lecture for 2019

ABSTRACT

A central theme of Atwater's research was the development and application of methods to understand how human beings and animals adapt to the nutrients they ingest. The research described in this article also deals with adaptation to nutrition focusing on adaptation to overnutrition, adaptation to undernutrition, adaptation to dietary fat, adaptation to dietary protein, adaptation to micronutrients, and adaptation to sugar and high-fructose corn syrup (HFCS). Studies using overfeeding have shown several things. First, overfeeding did not change the thermic response to ingestion of food nor the coupling of oxidative phosphorylation in muscle to energy expended by muscles during work on a bicycle ergometer between 25 and 100 watts. Second, the response to overfeeding was significantly influenced by the quantity of protein in the diet. During carefully controlled studies of underfeeding of people with obesity, the macronutrient composition of the diet did not affect the magnitude of weight loss. However, baseline genetic and metabolic information could provide guidance for selecting among the lower or higher protein diets, and lower or higher fat diets. Adaptation to an increase in dietary fat from 35% to 50% is slow and variable in healthy sedentary men. Adaptation is more rapid and complete when these same men were physically active. This effect of muscular exercise was traced to changes in the metabolism of glucose in muscles where pathways inhibiting glucose metabolism were activated by exercise. Dietary patterns that increased the intake of calcium, magnesium, and potassium effectively lower blood pressure in individuals with high normal blood pressure. Finally, the intake of sugary beverages was related to the onset of the current epidemic of obesity.

Introduction

Wilbur Olin Atwater is one of the greatest American pioneers of nutritional research. A central theme of Atwater's research was the development and application of methods to understand how human beings and animals adapt to nutrients they ingest. He built the first American respiration chamber which was of such value that it has been adapted by many other investigators. Using a metabolic respiration chamber patterned on the instrument that Atwater developed provided the road map for many of the studies described below and it is in this sense that I follow in the footsteps of Atwater.

It was in 1893 that Atwater and the USDA cemented a permanent relationship in a way that led to the development of the Cooperative State Research Service, and subsequently the Agricultural Research Service. The contributions made by Atwater using the respiration calorimeter, which he and E.B. Rosa built (1), were monumental in the field of nutritional science. Construction and operation of this respiration calorimeter at Wesleyan University was “big science.” It was made possible by what is now derogatively called “pork-barrel science” (2).

Wilbur Olin Atwater was born in 1849, the son of a Methodist minister. He began his collegiate education at the University of Vermont but transferred to the Methodist-dominated Wesleyan University in Middletown, CT, USA, where he graduated in 1865. He earned his PhD 4 y later from the Sheffield Scientific School at Yale University. Following 2 y of study at German Universities in Leipzig and Berlin, Atwater returned to begin an academic career in the USA.

Among nutritional scientists, Atwater is remembered for his confirmation that the first law of thermodynamics applied to human beings as it does to all living things. He is also remembered for his development of tables of food composition and for the relative energy value of macronutrients oxidized in the body. He provided energy values for the major macronutrients that are often cited as 4 kcal/g for carbohydrate, 4 kcal/g for protein, 9 kcal/g for fat, and 7 kcal/g for alcohol (3).

Atwater's interest in energy expenditure had been reignited by a tour of Europe in 1882 and 1883. During the 19th century, Munich was a leading center for studies in nutritional science. Justus von Liebig, one of the giants of nutrition and agriculture in the first half of the 19th century, eventually became a professor at the University of Munich, which also counted Carl von Voit as professor of physiology and Max-Joseph Pettenkofer as professor of hygiene. As Kirkland says, “for the devout nutritionist, Munich became a sort of Mecca” (4).

It was in Munich that Atwater saw the respiration calorimeter invented by Rudolph Pettenkofer. With the availability in 1887 of additional funds from the state of Connecticut, Atwater, along with Rosa, a new professor of physics, started to build a respiration calorimeter in 1892 (1). The costs of construction and operation of this chamber were >$10,000 per year, which exceeded the salary of a Wesleyan professor at that time by >5-fold. The respiration calorimeter was a room-size instrument in which subjects could live for 1 or more days (1). Using this apparatus, Atwater and his colleagues demonstrated that the law of conservation of energy applied to human beings (1).

Adaptation alcohol is one of the nutrients that I did not study, in part because of the pioneering work of Atwater (3) with later additions from other scientists (5). Atwater and colleagues drew a number of conclusions from their study of the metabolism of alcohol. First, >98% of the ingested alcohol was absorbed and oxidized in the body. The presence of alcohol in moderate amounts tended to very slightly increase the availability of other nutrients in the diet, especially proteins. Atwater also found that alcohol was more completely absorbed than other nutrients in the ordinary mixed diet. The laws of conservation of energy applied to the metabolism of alcohol in the diet as they did for other components of an ordinary diet. Finally, alcohol appeared to protect body fat by serving as a substrate for oxidation, but its effects in sparing body protein were somewhat more variable.

Adaptation to overnutrition

The respiration calorimeter invented by Atwater, played an important role in my studies of overnutrition by overfeeding described below. I was fortunate to be working in the laboratory of Edwin B. Astwood in the 1960s when overfeeding investigations from the Vermont study were first presented by Ethan Allen Sims (6, 7). Sims graciously let me work during his group beginning with my fellowship and continuing after I became a faculty member at Tufts New England Medical Center, and even when I moved from Boston to Harbor-University of California, Los Angeles (UCLA) Medical Center in Torrance, CA, USA. The collective studies with Sims, my medical school classmate Edward Horton, and Lester Salans who subsequently became Director of the National Institute of Diabetes and Metabolism [precursor of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK)] focused on the metabolism of adipose tissue (8, 9). In 1 of these studies we found that the altered basal glucose metabolism and response to insulin in adipose tissue from patients with spontaneous obesity were reproduced by weight gain resulting from overfeeding. Moreover, they were reversible by weight loss (8). In the second study comparing lean individuals who gained weight by overfeeding with individuals with spontaneous obesity found that differences in the metabolism of adipose tissue between obese and lean subjects persisted when differences in the size of fat cells and caloric intake were controlled, suggesting that spontaneous obesity may have metabolic differences from obesity due to overfeeding (9).

One of the take-away messages from the Vermont study was that men seemed to require significantly more energy per unit of body surface area to maintain their elevated weight than they had required when they were at normal weight. This observation reflected on the concept of “luxuskonsumption,” the idea that you could maintain a stable weight over a wide range of energy intakes (10, 11). It obviously needs to be tested in more detail.

Once established at the Harbor-UCLA Medical Center I began to plan an overfeeding study to answer 2 questions from the Vermont studies. Could this apparent “luxuskonsumption” be the result of an enhanced thermic effect of food, i.e. the increased oxygen consumption during food ingestion, or could it be the result of inefficiency of coupling oxidative metabolism in muscle to muscular contraction thus producing a kind of metabolic inefficiency?

Before recruiting other volunteers, I decided it would be appropriate for me to gain weight first (12). Beginning in January 1972 at a weight of 72.3 kg, I began by trying to double my usual food intake. This didn't work because I could not ingest all of the food, so I switched to a more energy-dense diet with higher fat and sugar content and less water. Gradually, over the next 10 wk my weight increased just over 10 kg to reach 82.7 kg at which time I repeated my study of thermic effects of food and the efficiency of energy expenditure on a bicycle ergometer. My thermic effect of food, like that of the other 4 men who were overfed, did not change. Neither did the efficiency of oxidation to muscular activity which was the same for each subject before and after gaining weight and ranged between 27% and 30% (13, 14). When my overfeeding ended, my weight rapidly returned to normal where it has remained for the last 45 y (12).

The next foray into overfeeding asked whether protein in the diet affected efficiency of weight gain in studies supported by the USDA. Earlier studies by Miller and Mumford (15) had shown that overeating a low-protein diet produced less weight gain than eating the same energy excess with a normal protein diet. In an analysis of studies examining the effects of overfeeding, Michael Stock (16) had shown that there was a “V” shaped relation between the efficiency of food storage and dietary protein. At both high and low levels of protein the amount of food energy stored seemed to decrease suggesting that protein could induce a metabolic inefficiency and account for “luxuskonsumption.” To test this hypothesis, we used a metabolic respiration chamber that was a descendent of the one developed by Atwater and Rosa (1). With it, we overfed men and women by 40% above their energy needs with 1 of 3 diets: a 5% protein diet, a 15% protein diet, or a 25% protein diet for a period of 8 wk while the individuals lived in the Pennington Biomedical Center research facility (17–23). We did indeed find that weight gain was significantly lower with the 5% protein diet than with either of the other 2 diets (17). However, this difference was due to changes in body composition. With the low-protein diet lean body mass did not increase, and actually showed a statistically significant decrease, whereas lean body mass increased in the other 2 groups. The quantity of body fat stored was nearly the same in each group, although there was a significant inverse relation between fat storage and protein intake. The high-protein diet was also associated with a higher metabolic rate, both during the day and at night. Thus, specifying the amount of protein is important in any metabolic study (17).

Adaptation to dietary fat

Animal studies

In an “obesogenic” environment with large quantities of high-fat high-sugar foods readily available to everyone, the important question is “why isn't everyone fat?” We have addressed this question from 2 perspectives. The first began as a comparison of 2 strains of rats, one which readily becomes obese eating a high-fat diet, the Osborne-Mendel (OM) rat, and the other strain called the S5B/Pl rat, which resists the temptation to overeat a high-fat diet, as well as cafeteria or supermarket diets and maintains body weight and body fat (24). These studies were initiated in collaboration with Rachel Schemmel who had selected these 2 animals by choosing animals which fattened and didn't fatten in a project with Olaf Mikkelsen at Michigan State University (25), one of the leaders in the field of nutrition.

One of the key observations about these animals was that they differed in the way they responded to ingested fatty acids. In 1 set of experiments conducted with Tim Gilbertson we found that taste receptors isolated from the sensitive rat and resistant rat had very different responses to linoleic acid (26). When the taste buds of the resistant rats were bathed with linoleic acid, electric activity was suppressed by a larger amount and for a longer period of time than the taste buds of sensitive rats which only had a transient short-lived response. A similar difference in response was observed when linoleic acid was infused into the intestine in studies conducted with Danielle Greenberg (27). For these experiments, the rats had cannulas in their stomachs from which food could be drained when the cannulas were opened. Putting linoleic acid into the intestine of the S5B/P1 rats that are sensitive to a high-fat diet, produced a more complete and much longer lasting suppression of sham-feeding than when the same amount of linoleic acid was infused into the duodenum of the OM rats which are sensitive to becoming fat eating a high-fat diet (27).

These differences in response of food intake to fatty acids suggested that there must be a receptor system in the brain that behaved differently in these 2 strains of rats. Collaborative studies with Michaelides (28) have confirmed this idea. Positron-emission tomography with [¹⁸F]fluoro-2-deoxyglucose as a marker showed that rats sensitive to a high-fat diet had significantly lower uptake of glucose in several brain regions, including the striatum-putamen, hippocampus, cerebellum (both the reticular and tegmental nuclei), than in rats that were resistant to obesity. The striatum had the biggest difference particularly in the brain of the resistant rats. Since the striatum is involved in modulation of food intake, the next step was profiling gene expression. Compared with the rats that were resistant to the high-fat-induced obesity, those that are sensitive to a high-fat diet showed significantly higher levels of mRNA expression for 5 genes (Rgs4, Drd1a, Drd2l, Grm5, and Cnr1). Several receptors, including dopamine D1 and D2, metabotropic glutamate receptor 5, and cannabinoid receptor 1 were next examined and were not different between the strains of rats. In contrast, the gene, regulator of G-protein signaling 4 (Rgs4) was significantly higher in the rats that get fat on a high-fat diet (OM) compared with those that do not (S5B). The difference in genetic expression was related to repressive epigenetic mechanisms associated with the differential regulation of the Rgs4 gene. Rgs4 acts to terminate G-protein activity. The striatum is composed primarily of 2 cells types. One of these types of cells, the striatopallidal medium spiny neurons, has been implicated in both human and rodent obesity. Rgs4 is enriched in the neurons of fat-sensitive rats (OM), but not in the fat-resistant (S5B/P1) rats. Suppression of gene expression for Rgs4 in the fat-sensitive rats using small interfering RNA (siRNA) injections reduced food intake significantly more than in the resistant rats. These rodent studies led to important clinical correlations. In human brains from postmortem examinations, overweight individuals with higher BMI levels had significantly more Rgs4 protein compared with normal weight controls (28). The Rgs4gene may thus provide a new focal point for developing agents to treat individuals with obesity.

Clinical Studies

The same question of why some people respond to dietary fat differently from others was also evaluated in clinical studies supported by the USDA. In 1 study, the amount of fat and carbohydrate were measured when men living in a respiration chamber, like the one of Atwater, switched from eating a 35% fat diet to eating a 50% fat diet. The respiratory quotient (RQ) that we measured varies between 0.7 (fat oxidation) and 1.0 (carbohydrate oxidation), although it can be transiently >1 if carbohydrate is being converted to fat. When 6 healthy young men each spent 4 d in an Atwater-type respiration chamber at the Pennington Center, they showed a wide range of shift in RQ over the 3 d after fat in the diet was increased from 35% to 50%. In 1 man, the adaptation was complete in 4 d, and at the other extreme 1 man showed no adaptation to the higher fat diet as reflected in fat oxidation. The other men were between these extremes. When the same individuals underwent the same experiment but had compulsory exercise each day during their time in the chamber, all 6 men increased their oxidation of fat day by day and all of them had adapted completely by day 4. Thus, exercise can enhance fat oxidation, suggesting that muscle adapts to the higher fat diet (29, 30).

To pursue this further, we examined the metabolism of glucose and fatty acids in muscle biopsies obtained before and after eating the 50% fat diet. One way to enhance fat oxidation is to inhibit glucose oxidation. There are ≥2 key steps in the metabolism of glucose where this can occur. One is in the conversion of fructose-6-phosphate to fructose 1,6-diphosphate, and the other is the conversion of pyruvate to acetate by pyruvate dehydrogenase. In both cases there are activators and inhibitors of the phosphorylation of these 2 enzymes. When we examined the muscle biopsies before and after the transition to a high-fat diet, we found that pyruvate dehydrogenase kinase 4 (PDK-4), which inhibits pyruvate dehydrogenase, and 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (PFKFB3), which inhibits the activation of fructose 1,6-diphosphate, were both higher with the high-fat diet (31).

Adaptation to underfeeding

Diabetes Prevention Program

The Diabetes Prevention Program and Outcomes Study was designed to ask whether there were strategies that could reduce the conversion rate of prediabetes to diabetes. The initial trial included 3 different experimental strategies: an intensive lifestyle program designed to produce weight loss, metformin which is a drug widely used to treat diabetes, and troglitazone, an antidiabetic drug that works by activating peroxisome proliferator-activated receptors (PPARs). Reports of liver toxicity led to termination of the troglitazone arm of this trial. The trial was conducted in 22 centers and included over 3000 people who were almost all overweight or obese. After 3 y of follow-up the study reported that those in the lifestyle group who lost 5.5 kg on average reduced their incidence of diabetes by 58% compared with the standard treatment arm. Metformin reduced the risk by 31% (31). With continuing follow-up to 10 y the lifestyle group still showed a 34% reduction in conversion rate and metformin an 18% reduction indicating the durability of these treatments (32, 33). The impact of the lifestyle program was felt in many aspects of the participant's lives, and the concepts of this program have been widely replicated around the country in an effort to lower the incidence of diabetes.

Look Action for Health in Diabetes study

The Look Action for Health in Diabetes (AHEAD) study is a second multicenter trial involving 16 centers that focused on the question of whether an intensive lifestyle program, similar to the one in the Diabetes Prevention Program could reduce the incidence of a combined end-point for cardiovascular disease. There had already been data from the surgical literature showing that weight loss in the range of 10% or more could reduce overall mortality and cardiovascular disease (34).

In this randomized clinical trial, patients with type 2 diabetes were randomly assigned to either an intensive lifestyle program or a program of diabetes support, but no intensive lifestyle activities. Weight loss with the lifestyle program was among the best ever reported with an average loss of 8.6% at 1 y and maintenance of >5% out to 10 y (35). However, even without an intensive intervention, there was weight loss in the control group and by the end of 10 y the control group had lost an average of nearly 4%. At the end of 10 y using an intent-to-treat analysis there was no significant reduction in the combined cardiovascular end-point (36). However, there was a wide range of weight loss among those in the lifestyle program. At the end of 1 y, 25% of the patients had lost <5% and 10% lost essentially no weight or even gained weight, despite the intensive lifestyle intervention (37). Since weight loss is the intermediary for the expected effect on cardiovascular disease, it would be irrational to expect people who didn't lose weight to get this benefit. When the people who actually lost 10% or more were compared with those who didn't lose weight, there was a significant 20% reduction in the composite cardiovascular events (38).

Prevention of Overweight Using Novel Dietary Strategies Lost trial

The question of whether there is 1 diet that is better than another diet to facilitate weight loss is a recurring theme. There are arguments that favor low-carbohydrate diets, whereas others favor low-fat diets, or balanced deficit diets, or liquid formula diets, or 1 of many other diets, and this literature is evaluated in several references (39–42).

It was this question that led to the design of the Prevention of Overweight Using Dietary Strategies (POUNDS) Lost trial which was conducted at the Harvard School of Public Health in Boston and the Pennington Biomedical Research Center of LSU in Baton Rouge. The study was designed to recruit 400 participants from each site. A final total of 811 participants who were overweight or obese and recruited and randomized in a 2×2   factorial design to 1 of 4 diets with target intakes of 15% (average protein = AP) or 25% protein (high protein = HP) and 20% fat (low fat = LF) or 40% fat (high fat = HF). Men and women with a BMI ranging from 25 to 40 kg/m², and aged between 30 to 70 y, and who were otherwise healthy and able to participate in a 2-y clinical trial were randomized with the first subject enrolled in October 2004 and the trial finalized in December 2007. All 4 groups of participants received instruction in their calorie-restricted diets. Energy intake was reduced by 750 kcal/d below the resting metabolic rate determined by using a metabolic cart and then multiplied by an activity factor with a minimum caloric intake of 1200 kcal/d. All diets were low in saturated fat (43). The foods used were similar in all diets, but with differing amounts. Dietary adherence was based on 24-h telephone interviews done at 6 mo in a 50% subset of participants at each site. Reported data were compared with assigned goals and those within ± 5% were labeled as “adherent.” A behavioral program of similar programmatic content and intensity was uniformly implemented across the 4 dietary groups using personnel from each site who had been specifically trained in these techniques (43). An activity level of 30 min 5 d/wk was encouraged and nearly a quarter of the participants (N = 241) received and wore pedometers that were attached at the belt line. Data for the number of steps was recorded at baseline and 6 mo in this subsample from each of the 4 dietary groups. Investigators and staff strived for equipoise at each site by teaching participants that each diet had an equal chance of success and by emphasizing the importance of macronutrient goals. The staff members responsible for outcome measurements were unaware of treatment assignment.

Mean weight losses were similar across all 4 diet groups from baseline to 6 mo and to 24 mo. Weight loss in the 15% and 25% protein groups was, respectively, 3.0 compared with 3.6 kg at 2 y (P = 0.2). Weight loss was 3.3 kg for both the 20% fat and 40% fat diet group. The 4 diets in this trial provided a graded intake of carbohydrate ranging from 35% to 65%. Weight loss was not significantly different when comparing the highest carbohydrate and lowest carbohydrate diet (2.9 compared with 3.4 kg, respectively). Similarly, there were no macronutrient effects on weight loss or weight maintenance at 2 y in the intent-to-treat analysis. However, POUNDS Lost did show the benefits of behavioral adherence as reflected in attendance at group sessions (43, 44) and adherence to diet assignment, particularly protein, was associated with more weight loss.

One of the most striking features of this study, and most others dealing with weight loss diets, is the wide range of weight losses. Although the mean weight loss did not differ between groups, there was considerable variation of weight loss within each of the 4 dietary groups (45). Indeed, the patterns of weight loss in those who completed 6 mo in each diet look remarkably similar; some people gained a small amount and others lost a good deal of weight. Because each of the diets had a similar pattern of weight loss, with no mean differences, it allowed us to pool the data and examine the effect of genetic factors at baseline and whether the differences in weight loss interacted with diet assignment. A total of 19 genes have been examined so far (46) and of these, 17 interacted with diet. Eleven interacted with dietary fat, either high or low. Three other genes interacted with dietary protein and 3 genes with dietary carbohydrate (65% compared with 35%). For example, the amylase gene (AMY1/AMY2)(rs 11,185,098) was the only 1 to predict weight loss independently of diet. The AA genotype lost significantly more weight than the GG genotype (47). The protein phosphatase 1 K mitochondrial gene (PPM1K)(rs1440581) produces a protein associated with modulating the ratio of BCAA/AAA in the serum and is a marker for liver disease. In POUNDS Lost, individuals with the TT genotype eating the high-fat diet lost nearly 8 kg compared with 4 kg for those with the CC genotype (48). A final example is the FTO gene (Fat Mass and Obesity Associated Gene superfamily of hydroxy-α-ketoglutarate-dependent hydroxylase)(rs1558902) where the AA genotype lost 6 kg eating the high-protein diet compared with 3 kg for the TT genotype eating the same diet (49). Other examples of dietary–gene interactions that influence weight loss in the POUNDS Lost trial can be found elsewhere (46).

Adaptation to micronutrients

The DASH (Dietary Approaches to Stop Hypertension) Diet Study

A number of dietary factors influence blood pressure. These include high levels of fiber and minerals such as potassium and magnesium as well as a lower fat content. Observational studies of blood pressure have noted significant inverse associations of magnesium, potassium, and calcium intake as well as levels of dietary fiber and protein. Trials that have tested these nutrients as dietary supplements, however, have produced disappointing and inconsistent reductions in blood pressure (46). Against this background, the National Heart Institute issued a Request for Applications (RFA) to examine the effects of dietary patterns, as opposed to single dietary components, on changes in blood pressure. In response to this RFA, 4 field centers, 1 from Harvard, 1 from Johns Hopkins, 1 from Duke, and 1 at the Pennington Center, and a co-ordinating center were selected to design and execute this trial. Diets were planned by using data on the intake of magnesium, potassium, and calcium from the National Center of Health Statistics surveys (50). The comparison diet had micronutrient levels of calcium, magnesium, and potassium near the 25th percentile of average intake. The 2 intervention diets, a fruits and vegetables diet and a fruits, vegetables, and low-fat dairy products diet were designed to provide these same micronutrients at the 75th percentile of average intake. The trial enrolled 459 adults with systolic blood pressure <160 mmHg and diastolic blood pressure between 80 to 95 mmHg. The participants were enrolled in an 8-wk efficacy study in which the planned diets were provided at each of the centers.

The results of this study were clear (51). The individuals assigned to the fruits and vegetables diet lowered their blood pressure significantly more than the comparison diet. The combined diet with fruits and vegetables, and low-fat dairy products, named the Dietary Approaches to Stop Hypertension (DASH) Diet, lowered blood pressure more than the fruits and vegetables diet and more than the comparison diet. This study launched the DASH Diet which gradually became accepted as 1 of the best dietary patterns, and for 8 y was ranked number 1 among the diets reviewed by US News and World Report in their annual survey of diets. In a recent systematic review and network meta-analysis, the DASH Diet was ranked as the most effective dietary approach in reducing both systolic and diastolic blood pressure, and the credibility of evidence for this trial was rated high in contrast to the other studies in this review (52).

Following publication of the DASH Diet, we conducted a second study examining the interaction of the DASH Diet with dietary sodium (53). In this study, a total of 412 individuals were randomly assigned to eat either a control diet typical of what Americans usually eat, or the DASH Diet each at 3 different levels of sodium intake. Using a crossover design, participants ate foods with high, intermediate, and low levels of sodium for 30 consecutive days each, in random, counterbalanced order. At each level of sodium intake, the DASH Diet lowered blood pressure substantially, with the greater effects at higher levels of sodium intake (54).

The modified DASH Diet study

A principal concern with the DASH Diet was the significantly lower level of HDL cholesterol that it produced (55). Dietary carbohydrate and fat are known to influence the level of HDL cholesterol, as well as other lipids. We decided to test the effects of substituting regular dairy products, which have a higher amount of fat, for the low-fat dairy foods in the original DASH Diet. This was a 3-period randomized crossover trial in free-living healthy individuals who consumed 3 diets in random order: original DASH Diet with low-fat dairy products, a DASH Diet variant with regular dairy products, and a typical Western diet. Laboratory measurements included lipoprotein particle concentrations determined by ion mobility, which was done at the end of each experimental period. A total of 36 participants completed all 3 dietary periods.

Blood pressure was reduced similarly with the original DASH Diet with low-fat dairy products, and the regular dairy product DASH Diet, compared with the control diet. The regular dairy product DASH Diet significantly reduced triglycerides and large and medium VLDL particle concentrations and increased LDL peak particle diameter compared with the low-fat DASH Diet. The low-fat dairy product DASH Diet, but not the regular dairy product DASH Diet, significantly reduced HDL cholesterol, LDL cholesterol, intermediate density lipoprotein and large LDL particles, apoA-I, and LDL peak diameter compared with the control diet. From this study, we conclude that regular fat dairy products in the DASH Diet lowered blood pressure to the same extent as the original DASH Diet but had an improved lipoprotein profile compared with the original DASH Diet (55).

Adaptation to carbohydrate with a focus on sugar

The consumption of sugar-sweetened beverages increased 5-fold between 1950 and 2000 (56). It went from 10 gallons per person per year to 50 gallons per person per year. If the sugar content is 10% which approximates the value for most sugary beverages, then in 1 y 10 gallons would yield ∼3780 g (1 gallon = 3.785 liters) of sugar and at 4 kcal/g this is ∼15,120 kcal. The 5-fold increase in sugar intake from 1950 to 2000 provided an additional calorie load of ∼60,000 kcal or ∼165 kcal/d. Unless the individual performed enough exercise to offset these additional 165 kcal, or reduced their intake of other calories by ∼165 kcal/d to counterbalance the load, they would be expected to gain weight which could be >9 kg.

In the 1960s chemists learned how to convert starch into fructose by breaking down the starch in plants like corn to glucose and then enzymatically rearranging the glucose to produce fructose. By combining glucose and fructose to produce high-fructose corn syrup (HFCS), manufacturers were able to substitute sweet-tasting HFCS for sugar and reduce the cost of sugary beverages. The increased use of HFCS paralleled the increase in the prevalence of obesity (57). Although association is not causation, much subsequent data has shown the relation of sugary beverage consumption with the risk of developing diabetes (58), cardiovascular disease (59), obesity (59), nonalcoholic fatty liver disease (60), and most recently cancer (61).

Our observations relating the consumption of HFCS to the prevalence of obesity were published in 2014 (57). It is interesting to note that this was the inflection point near the peak of sugary beverage intake which has since gradually declined, although it is still nearly 4-fold higher than in 1950.

Randomized clinical trials show that replacing sugary beverages with artificially sweetened beverages will slow the weight trajectory of adolescents (62). Of particular interest is that the compensation for changed caloric intake with this replacement of sugary beverages by very low-calorie beverages was complete in the thinner adolescents, but not in the heavier adolescents. It was the adolescents with BMI values above the mean who did not detect the change in calories with the beverage substitution, and thus slowed their weight gain (63).

In summary, my studies in the field of adaptation to nutrition have relied heavily on the seminal work of Atwater >100 years ago and why I am lucky to be able to walk in his footsteps.

Notes

Author disclosure: The author reports no conflicts of interest.

Abbreviations used: AHEAD, Action for Health in Diabetes; DASH, Dietary Approaches to Stop Hypertension; HFCS, high-fructose corn syrup; OM, Osborne-Mendel strain of rats; Rgs4, regulator of G-protein signaling 4; RQ, respiratory quotient; UCLA, University of California, Los Angeles.