An energy-deficit diet is the cornerstone of therapy for obesity. However, the most appropriate macronutrient composition of diet therapy needed to improve metabolic health remains controversial. In the present study, we carefully evaluated the longitudinal metabolic effects of short-term (48 h; 2% weight loss) and longer-term (11 wks; 7% weight loss) calorie restriction (1000 kcal/d energy deficit) with either a high- or low- carbohydrate diet in obese, insulin-resistant but non-diabetic adults. Our data demonstrate that short-term CR caused a rapid decrease in IHTG content, increase in hepatic insulin sensitivity, and decrease in endogenous glucose production rate, whereas longer-term CR and moderate 7% weight loss improved skeletal muscle insulin sensitivity, in conjunction with an increase in cellular insulin signaling. In addition, short-term CR with a low-carbohydrate diet caused a greater change in liver fat content and metabolic function than short-term CR with a high-carbohydrate diet. These data underscore the complexity of the metabolic effects of CR with diets that differ in macronutrient composition, and demonstrate temporal differences among organ systems in the adaptive response to CR itself and subsequent weight loss.
Our results refute our original hypothesis that a LC diet will cause insulin resistance because of increased adipose tissue lipolytic rates and excessive FFA release into the bloodstream. In fact, we found that LC intake rapidly caused a greater reduction in IHTG content, improvement in hepatic insulin sensitivity, and decrease in endogenous glucose production rate than consumption of an isocaloric low-fat diet. The mechanism responsible for the early beneficial effects on liver metabolism is not known, but is probably related to the greater decrease in plasma insulin concentrations in subjects consuming the low-carbohydrate diet. The decline in circulating insulin likely decreased IHTG because of enhanced lipolysis of IHTG and hepatic fatty acid oxidation14
, and decreased hepatic glucose production because of hepatic glycogen depletion17
and decreased glycogenolysis4, 34
. These metabolic alterations are similar to the physiologic adaptations that occur during the early response to starvation, which are also triggered by a reduction in carbohydrate intake16
. However, in contrast with data obtained from studies evaluating the metabolic effects of brief fasting12–14
, we did not detect a significant decline in skeletal muscle insulin sensitivity after 48 h of CR with a low-carbohydrate diet.
Weight loss, but not short-term CR, was necessary to increase skeletal muscle insulin-mediated glucose disposal. The improvement in muscle insulin sensitivity we observed in vivo
is explained by enhanced cellular insulin signaling (increased insulin stimulated IRS-1 tyrosine and Akt/PKB serine phosphorylation) detected after 7% weight loss but not after 48 h of CR. These results are consistent with data from a study conducted in subjects with type 2 diabetes that found insulin-stimulated Akt/PKB did not change after 2 days of CR35
. In addition, our data suggest that the mechanism responsible for the increase in insulin signaling involves down-regulation of JNK, which inhibits IRS-1 serine phosphorylation and the proximal component of the insulin signaling cascade36
. Therefore, these findings demonstrate that the increase in JNK associated with obesity and type 2 diabetes is responsive to nutritional manipulation and can be normalized by weight loss.
Nonalcoholic fatty liver disease is associated with insulin resistance37, 38
and is an important risk factor for diabetes39
. We previously found a linear inverse correlation between IHTG content and insulin sensitivity in both liver and skeletal muscle38
. In the present study, dietary manipulation of IHTG content allowed us to dissociate the interrelationships among IHTG and insulin sensitivity in liver and skeletal muscle. After 48 h of CR, IHTG content decreased by ~20%, which was associated with a decrease in basal glucose production rate and an increase in hepatic insulin sensitivity, whereas skeletal muscle insulin sensitivity did not change. Continued CR until subjects lost 7% of initial body weight caused a further decrease in IHTG content, without a further decrease in basal glucose production or improvement in hepatic insulin sensitivity. However, 7% weight loss up-regulated skeletal muscle insulin signaling and increased muscle insulin sensitivity. These data support the notion of a causal link between steatosis and hepatic insulin resistance. The mechanism responsible for the link between IHTG content and hepatic insulin sensitivity is unknown, but could be related to an accumulation of intracellular fatty acid metabolites, which can antagonize the effects of insulin signaling on endogenous glucose production40
Our data provide new insights into the potential mechanism responsible for the marked improvement in glycemic control observed within days after Roux-en-y gastric bypass (RYGP) surgery in obese patients with type 2 diabetes41
. For example, in one study, 90% of patients were able to discontinue all diabetes medications and maintain normal glycemia at discharge from the hospital 6 days after RYGP surgery, before much weight loss occurred42
. These observations have led to the hypothesis that diversion of ingested nutrients from the upper gastrointestinal tract has beneficial effects on glucose homeostasis, possibly because of an altered incretin response to meals43
. However, our results suggest that the rapid decrease in liver fat and improvement in hepatic insulin sensitivity that occur after brief CR can completely explain the early improvement in glucose homeostasis observed after bariatric surgery. Food intake is limited after RYGP surgery, and patients usually consume less than 250 kcal/d for several days after the operation41
. Therefore, the marked postoperative reduction in calorie intake, itself, likely has profound effects on hepatic fat content and metabolism40
. Moreover, the decrease in calorie intake makes it is unlikely that diversion of ingested nutrients from the upper gastrointestinal tract has an important effect on glucose metabolism.
In summary, the data from this study demonstrate that the effect of moderate calorie restriction in obese subjects with either a low-fat or low-carbohydrate diet on metabolic function is a continuum, with differential effects on specific organ systems. Brief (48 h) CR and minimal weight loss (~2% of initial body weight) primarily affects the liver, manifested by a decrease in IHTG content, an increase in hepatic insulin sensitivity, and a decrease in endogenous glucose production, whereas longer (~11 wks) CR and moderate weight loss (~7% of initial body weight) primarily affects skeletal muscle, manifested by an increase in muscle insulin-mediated glucose uptake and enhanced cellular insulin signaling. These findings help explain the rapid improvement in glucose homeostasis observed after low-calorie diet therapy and bariatric surgery.