Leptin Improved Belly Fat/Dyslipidemia - The Effects of Recombinant Human Leptin on Visceral Fat, Dyslipidemia, and Insulin Resistance in Patients with Human Immunodeficiency Virus-Associated Lipoatrophy and Hypoleptinemia - Pilot Study
"Leptin treatment was associated with marked improvement in dyslipidemia. Hepatic insulin sensitivity improved and lipolysis decreased. Visceral fat decreased with no exacerbation of peripheral lipoatrophy. Results from this pilot study suggest that leptin warrants further study in patients with HIV-associated lipoatrophy."
"The mean decrease of 32% in VAT induced by leptin administration in this study exceeds decreases reported with other therapies. For example, administration of pharmacologic doses of GH decreased VAT by 21-24% in two randomized trials in subjects with abdominal obesity"
"A recent randomized trial of a synthetic GHRH, TH9507, reported a more modest reduction in VAT (15% in subjects who received active therapy) with no significant effect on glucose metabolism (29). In contrast, leptin improved glucose metabolism."
"Total, LDL, and non-HDL cholesterol decreased by 15-20%, whereas HDL-C increased by a similar amount. In addition, triglyceride levels, measured under fasting, hyperinsulinemic, and postprandial conditions, decreased by about 30%."
"In conclusion, results from this open-label pilot study demonstrate that treatment with leptin was well tolerated and led to significant improvements in dyslipidemia, lipolysis, and hepatic insulin sensitivity that were accompanied by a marked decrease in VAT. These improvements occurred without any apparent exacerbation of peripheral lipoatrophy. These findings suggest that leptin may have therapeutic potential in HIV-infected patients with metabolic and morphologic alterations in association with lipoatrophy and warrant a larger, placebo-controlled trial in this population."
"In transgenic mice engineered to have virtually no white fat, a severe metabolic syndrome develops, characterized by insulin resistance, hypertriglyceridemia, hepatic steatosis, and decreased levels of adipocyte-derived hormones such as leptin and adiponectin....Treatment with recombinant methionyl human leptin ameliorated these metabolic abnormalities in both fatless mice (8, 9) and humans with lipoatrophy (10, 11, 12). Prompted by these findings, we undertook an open-label, proof-of-principle, dose-ranging study to comprehensively assess the metabolic and morphologic effects of recombinant human leptin treatment of HIV-infected subjects with lipoatrophy and hypoleptinemia."
"Results of the current study will need to be confirmed in larger, placebo-controlled studies to determine whether leptin offers advantages over other potential treatments for HIV-associated abdominal fat accumulation."
The Journal of Clinical Endocrinology & Metabolism April 2009 Vol. 94, No. 4 1137-1144
Kathleen Mulligan1, Hootan Khatami1, Jean-Marc Schwarz, Giorgos K. Sakkas, Alex M. DePaoli, Viva W. Tai, Michael J. Wen, Grace A. Lee, Carl Grunfeld and Morris Schambelan
Department of Medicine, University of California, San Francisco (K.M., H.K., J.-M.S., G.K.S., V.W.T., M.J.W., G.A.L., C.G., M.S.), San Francisco, California 94143; San Francisco General Hospital (K.M., H.K., J.-M.S., G.K.S., V.W.T., M.J.W., M.S.), San Francisco, California 94110; Department of Basic Sciences, Touro University (J.-M.S.), Vallejo, California 94592; Amgen Inc. (A.M.D.), Thousand Oaks, California 91320; and San Francisco Veterans Affairs Medical Center (G.A.L., C.G.) , San Francisco, California 94121
Context: Leptin deficiency is associated with dyslipidemia and insulin resistance in animals and humans with lipoatrophy; leptin replacement ameliorates these abnormalities.
Objective: The objective of the study was to evaluate the effects of leptin therapy in lipoatrophic HIV-infected patients with dyslipidemia and hypoleptinemia.
Design: This was a 6-month, open-label, proof-of-principle pilot study.
Setting: Metabolic ward studies were performed before and 3 and 6 months after leptin treatment.
Participants: Participants included eight HIV-infected men with lipoatrophy, fasting triglycerides greater than 300 mg/dl, and serum leptin less than 3 ng/ml.
Intervention: Recombinant human leptin was given by sc injection (0.01 mg/kg and 0.03 mg/kg twice daily for successive 3 month periods).
Outcome Measures: Measures included fat distribution by magnetic resonance imaging and dual-energy X-ray absorptiometry; fasting lipids; insulin sensitivity by euglycemic hyperinsulinemic clamp; endogenous glucose production, gluconeogenesis, glycogenolysis, and whole-body lipolysis by stable isotope tracer studies; oral glucose tolerance testing; liver fat by proton magnetic resonance spectroscopy; and safety.
Results: Visceral fat decreased by 32% (P = 0.001) with no changes in peripheral fat. There were significant decreases in fasting total (15%, P = 0.012), direct low-density lipoprotein (20%, P = 0.002), and non-high-density lipoprotein (19%, P = 0.005) cholesterol. High-density lipoprotein cholesterol increased. Triglycerides, whole-body lipolysis, and free fatty acids decreased during fasting and hyperinsulinemia. Fasting insulin decreased. Endogenous glucose production decreased during fasting and hyperinsulinemia, providing evidence of improved hepatic insulin sensitivity.. Leptin was well tolerated but decreased lean mass.
Conclusions: Leptin treatment was associated with marked improvement in dyslipidemia. Hepatic insulin sensitivity improved and lipolysis decreased. Visceral fat decreased with no exacerbation of peripheral lipoatrophy. Results from this pilot study suggest that leptin warrants further study in patients with HIV-associated lipoatrophy.
HIV-associated metabolic and morphologic alterations, including dyslipidemia, insulin resistance, lipoatrophy, and central fat accumulation, are recognized increasingly as a major complication of otherwise effective highly active antiretroviral therapy. Because this cluster of metabolic abnormalities, commonly referred to as HIV lipodystrophy, is associated with increased risk of cardiovascular disease in non HIV-infected populations, it is now feared that lipodystrophy may impact the long-term prognosis in HIV-infected patients whose life expectancies have been significantly extended due to effective viral suppression by highly active antiretroviral therapy.
Whereas the association of dyslipidemia and insulin resistance with obesity is well recognized, more recent studies have demonstrated that fat loss (lipoatrophy) can result in similar metabolic abnormalities. In transgenic mice engineered to have virtually no white fat, a severe metabolic syndrome develops, characterized by insulin resistance, hypertriglyceridemia, hepatic steatosis, and decreased levels of adipocyte-derived hormones such as leptin and adiponectin (1, 2). Evidence that the lack of fat is responsible for the phenotype is provided by experiments in which transplantation of fat from wild-type mice ameliorated these metabolic abnormalities (3). Both HIV-infected (4, 5) and uninfected (6, 7) humans with lipoatrophy manifest a similar metabolic profile, including hypoleptinemia, hypertriglyceridemia, and hyperinsulinemia.
Treatment with recombinant methionyl human leptin ameliorated these metabolic abnormalities in both fatless mice (8, 9) and humans with lipoatrophy (10, 11, 12). Prompted by these findings, we undertook an open-label, proof-of-principle, dose-ranging study to comprehensively assess the metabolic and morphologic effects of recombinant human leptin treatment of HIV-infected subjects with lipoatrophy and hypoleptinemia.
In this open-label pilot study in HIV-infected men with lipoatrophy and hypoleptinemia, administration of recombinant human leptin for 6 months was associated with a robust decrease in visceral fat, an increase in hepatic insulin sensitivity, and marked improvement of dyslipidemia. The decrease in VAT with leptin treatment was accompanied by a nonsignificant (14%; P = 0.645) decrease in abdominal SAT area, but neither peripheral (arm + leg) fat measured by DEXA or thigh SAT measured by MRI decreased. Although evidence of central fat accumulation was not an eligibility criterion in the current study, average values for VAT at baseline were comparable with those seen in studies of experimental treatments in HIV-infected patients selected for the presence of excess central fat. The mean decrease of 32% in VAT induced by leptin administration in this study exceeds decreases reported with other therapies. For example, administration of pharmacologic doses of GH decreased VAT by 21-24% in two randomized trials in subjects with abdominal obesity (25, 26). However, pharmacological GH treatment is associated with impairment of glucose metabolism and exacerbation of insulin resistance (25, 26, 27, 28) and loss of sc fat (25, 26) and thus is not an optimal therapy in a population already predisposed to insulin resistance and lipoatrophy. A recent randomized trial of a synthetic GHRH, TH9507, reported a more modest reduction in VAT (15% in subjects who received active therapy) with no significant effect on glucose metabolism (29). In contrast, leptin improved glucose metabolism. Randomized trials of rosiglitazone in patients with HIV infection improved insulin sensitivity but consistently failed to reduce VAT (30, 31, 32), whereas those of metformin have shown modest, if any, reductions in VAT (31, 32, 33, 34). Results of the current study will need to be confirmed in larger, placebo-controlled studies to determine whether leptin offers advantages over other potential treatments for HIV-associated abdominal fat accumulation.
The proportional loss of VAT relative to loss of fat or body weight in the current study is greater than has been observed in studies of individuals losing weight by simple caloric restriction, pharmacological weight loss therapy, or exercise (35). However, our subjects entered the study with relatively depleted stores of SAT, so we cannot determine whether their loss of VAT was selective or reflected its relative excess compared with SAT at baseline. Barzilai et al. (36) found a preferential loss of fat in intraabdominal depots in healthy rats that lost weight during leptin treatment when compared with pair-fed controls with the same rate of weight loss, and they speculated that the selective effect of leptin on VAT may be mediated through _-adrenergic stimulation. However, because loss of VAT in both the current study and that of Barzilai et al. occurred in the presence of weight loss, a role for leptin-associated reduced energy intake as a cofactor must also be considered. In addition to these central effects, leptin has also been shown to act peripherally, e.g. by activation of AMP kinase (37), up-regulation of peroxisome proliferator-activated receptor- coactivator-1 (38), and other effects that may contribute to enhanced lipid disposal. Our findings in humans therefore emphasize that further studies are required to determine the exact neuroendocrine and other mechanism(s) by which leptin reduces VAT.
Leptin administration was associated with marked improvements in dyslipidemia in a group of patients who, for the most part, were already treated with lipid-lowering agents. Total, LDL, and non-HDL cholesterol decreased by 15-20%, whereas HDL-C increased by a similar amount. In addition, triglyceride levels, measured under fasting, hyperinsulinemic, and postprandial conditions, decreased by about 30%. Because subjects who were on lipid-lowering therapies were required to have been on stable treatment for at least 3 months before enrolling, it is unlikely that the observed improvements were a result of the hypolipidemic agents, whose effects are generally seen within the first 4 wk of treatment. To our knowledge, no one has shown that weight loss per se can produce reductions in LDL-C of this magnitude. Our results are consistent with the results of studies in patients with non-HIV lipodystrophy, in which striking reductions in triglyceride levels were achieved during leptin treatment (10, 12). The magnitude of the changes in lipids observed in the present study is comparable with or greater than that seen in studies of other lipid-lowering agents in HIV-infected populations (39, 40, 41, 42, 43) and exceeds that achieved by experimental therapy with GH (25, 26, 28) or a synthetic GHRH (29). Given the difficulty in treating dyslipidemia in HIV-infected subjects, as manifest by the abnormal values in this cohort of subjects, nearly all of whom were on hypolipidemic therapy, the improvement in lipids profile raises the possibility that leptin therapy could potentially reduce the risk of cardiovascular disease.
Leptin administration was also associated with decreases in fasting insulin levels, as has been observed in other studies (10, 11, 12). There has been less consistency in results of studies that used dynamic measures of peripheral insulin sensitivity: Petersen et al. (44), using a euglycemic hyperinsulinemic clamp, reported a marked improvement in whole-body glucose uptake in three subjects with non-HIV lipodystrophy. Using an insulin suppression test, Oral et al. (10) also found a significant improvement in insulin sensitivity in patients with non-HIV lipodystrophy. In contrast Lee et al. (11) reported only a trend to improvement in patients with HIV lipodystrophy. Conceivably, the failure to see a significant increase in whole-body glucose uptake in our study, in contrast to those reported in non-HIV lipodystrophy (10, 44), may have reflected the relatively mild degree of peripheral insulin resistance at baseline in some of our subjects. However, whereas we did not observe a consistent increase in peripheral insulin sensitivity, we did demonstrate reductions in fasting insulin and glucose levels and suppression of hepatic glucose production, glycogenolysis, and gluconeogenesis both during fasting and in response to hyperinsulinemia during the clamp, suggesting that leptin treatment improved hepatic insulin sensitivity. Therefore, our results emphasize the importance of considering site-specific effects in metabolic studies of leptin.
Leptin administration also led to a decrease in the rate of lipolysis and FFA levels during both fasting and hyperinsulinemia, suggesting an improvement in insulin sensitivity in adipose tissue. Overexpression of leptin in a transgenic, lipoatrophic mouse model has been shown to decrease lipolysis (45), and a trend to a decrease in lipolysis was observed in three patients with non-HIV lipodystrophy who were treated with leptin (44).. It has been suggested that excess lipolysis is an important metabolic feature of HIV lipodystrophy (46).
The design of the current study does not allow us to determine whether the improvements noted after 6 months of leptin treatment are dose or time dependent. In general, the metabolic results after 3 months of treatment fell between those measured at baseline and 6 months. In contrast, the major reduction in VAT appeared to occur between 3 and 6 months, suggesting that a higher dose may be required to achieve a reduction in VAT. These results also raise the possibility that the metabolic effects of leptin are at least in part directly induced by leptin and not a secondary result of the decrease in VAT. However, we did not measure adiponectin, an adipocytokine that is inversely associated with VAT and positively with insulin sensitivity (47). Adiponectin increased during treatment with rosiglitazone (32) or TH9507 (29) in patients with HIV-associated fat accumulation but did not change during leptin treatment in the study of Lee et al. (11). Although subjects were hypoleptinemic at baseline, the dose of leptin used during months 3-6 produced serum leptin levels that were above the upper limit of the normal range for lean adult men (about 14 ng/ml) and thus might be considered supraphysiological. Although leptin was well tolerated, even at the higher dose, the loss of weight and lean tissue seen during treatment cannot be considered to be positive effects.
Two small studies of the metabolic effects of leptin in HIV-infected patients with lipoatrophy and hypoleptinemia have now been completed, with some similarities and differences. Lee et al. (11) used a dose of 0.02 mg/kg twice daily in a placebo-controlled crossover design. The dose of leptin used in the study of Lee et al. was in between the two doses used in our open-label study, and the treatment duration was shorter (2 months in the study of Lee et al. vs. 6 months in our study). Seven subjects were treated, and five subjects completed both phases of the Lee study, whereas all eight subjects who entered our study completed 6 months of treatment. Based on the use of surrogate measures of insulin sensitivity (fasting insulin, homeostasis model assessment of insulin resistance, insulin suppression test), Lee et al. concluded that leptin improved insulin sensitivity. Although we, too, observed decreases in fasting insulin, our study used the criterion method for measuring whole-body glucose uptake, the euglycemic hyperinsulinemic clamp, along with stable isotope tracer technology, and found that although leptin improved hepatic insulin sensitivity, it had no effect on peripheral insulin sensitivity. Thus, our results provide novel information on the specific sites of leptin's effects on insulin sensitivity. Subjects in both studies had decreases in weight and trunk fat, increases in HDL-C, and trends to decreases in LBM. However, in contrast to reductions in triglyceride levels and LDL-C in our study, no such changes were seen in the study of Lee et al. (11). The design of the study by Lee et al. required subjects to have triglyceride levels between 300 and 1000 mg/dl in the absence of lipid lowering therapy at baseline, and two subjects were withdrawn during the study when triglycerides exceeded 1000 mg/dl. We cannot speculate on whether the absence of decreases in triglycerides and LDL-C in their study was a result of this or other differences in study design such as the use of a lower dose of leptin or a shorter duration of treatment. Overall, we conclude that the optimal therapeutic dose of leptin and effective treatment period are not yet known.
In conclusion, results from this open-label pilot study demonstrate that treatment with leptin was well tolerated and led to significant improvements in dyslipidemia, lipolysis, and hepatic insulin sensitivity that were accompanied by a marked decrease in VAT. These improvements occurred without any apparent exacerbation of peripheral lipoatrophy. These findings suggest that leptin may have therapeutic potential in HIV-infected patients with metabolic and morphologic alterations in association with lipoatrophy and warrant a larger, placebo-controlled trial in this population.
Baseline characteristics and clinical course
Total body fat was less than 20% in six subjects and less than 10% in two (Table 1). Values for leg fat in all eight subjects were below the threshold for the lowest decile for seronegative controls in a large survey (22), providing objective evidence of peripheral lipoatrophy. By design, all subjects had fasting serum leptin levels less than 3 ng/ml, and the majority were less than 2. All subjects had stable CD4 cell counts at baseline. Seven subjects were on thymidine analog nucleoside reverse transcriptase inhibitors (stavudine or zidovudine). Seven subjects were also on lipid-lowering agents, and five were on testosterone replacement. Subject 8 had developed type 2 diabetes while on antiretroviral therapy before enrollment and was on rosiglitazone (8 mg daily). The respective doses and regimens of antiretroviral therapy, lipid-lowering agents, testosterone, and rosiglitazone remained unchanged throughout the study. All subjects tolerated the study procedures, were fully adherent to twice-daily injections, and completed the 6-month trial.
Serum leptin levels
Serum leptin levels, measured 8-10 h after the preceding dose of leptin, increased in a dose-dependent manner, with mean values of 2.7 ± 0.6, 7..7 ± 1.9, and 21.3 ± 5.3 ng/ml at baseline and months 3 and 6, respectively (P = 0.002).
Effects of leptin on body composition
VAT decreased in all subjects (Fig. 1). The average decrease in VAT at month 6 was 32% (P = 0.001). A more variable effect on abdominal SAT was noted, with an overall change that was not significant (Table 2). An average weight loss of nearly 2 kg was largely attributable to decreases in LBM and trunk fat as measured by DEXA (Table 2). Thigh muscle cross-sectional area tended to decrease, consistent with the decreases in LBM by DEXA. Notably, neither limb (arm + leg) fat measured by DEXA nor thigh SAT measured by MRI decreased during treatment with leptin. Liver fat tended to decrease, but the change was not statistically significant. Two subjects had very low levels of liver fat at baseline (<2%), which left little margin for further reductions.
Effects of leptin on lipid metabolism
Total cholesterol, LDL-C, and non-HDL-C levels decreased significantly, with mean decreases of 15, 20, and 19% respectively, by month 6 (Table 3). HDL-C increased in six of eight subjects, with an average increase of 23%. Triglyceride levels decreased significantly under fasting, hyperinsulinemic (clamp), and postprandial conditions (Fig. 2). Whole-body lipolysis decreased under both fasting and hyperinsulinemic conditions, and the ability of insulin to suppress lipolysis increased with leptin (Table 3). Similarly, fasting FFA levels tended to decrease with leptin, and the ability of insulin to suppress FFA increased significantly.
Effects of leptin on glucose metabolism
Although fasting glucose levels did not change during leptin treatment, fasting insulin concentrations decreased significantly (Table 3). Glucose levels measured 2 h after an oral glucose load did not change significantly with leptin, and 2-h insulin levels tended to decrease. Using the current classification system (23), two of the seven nondiabetic subjects had impaired fasting glucose both before and after 6 months of leptin treatment. Three of the seven nondiabetic subjects had impaired glucose tolerance during the pretreatment oral glucose tolerance test, which normalized after 6 months of leptin. Fasting and 2-h glucose levels in the subject with diabetes who was on stable rosiglitazone treatment met the criteria for diabetes both before and after leptin treatment. Although there was no significant change in glucose AUC in response to leptin treatment, insulin AUC decreased by approximately 30% (P = 0.046). Insulin-mediated glucose uptake tended to increase, but results were not consistent among subjects and failed to achieve statistical significance.
When measured under fasting conditions, endogenous glucose production decreased significantly during leptin treatment, an effect that can be attributed primarily to a reduction in glycogenolysis (Table 3). During the clamp, leptin treatment improved the ability of insulin to suppress endogenous glucose production, with significant decreases in both glycogenolysis and gluconeogenesis.
Energy expenditure and intake
There were no significant changes in resting energy expenditure, adjusted for LBM, or lipid or carbohydrate oxidation rates under fasting and hyperinsulinemic conditions during the study (data not shown). Likewise, self-reported energy intake did not change significantly.
Safety and tolerability
Leptin treatment was well tolerated by all subjects. No clinical or laboratory adverse events were reported, including no changes in serum liver enzymes or hematology (Table 4). There were no injection site reactions or discomfort, or reports of headache or other somatic complaints that had been reported in other studies of leptin (10, 24). CD4 T cell counts and viral loads remained stable in all subjects. HIV RNA levels were undetectable in six of the eight subjects at both baseline and month 6. Two subjects had detectable viremia at both baseline and month 6; of these, one experienced a 0.3 log increase and the other a 0.05 log decrease during treatment.
Subjects and Methods
Subjects and study protocol
Medically stable HIV-infected men with clinical evidence of lipoatrophy, hypoleptinemia (serum leptin <3.0 ng/ml), and hypertriglyceridemia (fasting triglyceride levels >/=300 but <1000 mg/dl; or, if on hypolipidemic therapy, documentation of prior triglyceride >/=300 mg/dl) were recruited for study. These eligibility criteria were based on those used in a trial of leptin therapy in non HIV-infected patients with congenital or acquired lipodystrophy (10). Subjects who were on hypolipidemic therapy at baseline were required to have been on stable treatment for at least 3 months. Other eligibility criteria included demonstration of normal adrenal and thyroid function by measurement of fasting serum cortisol and TSH levels and dexamethasone suppression testing. Of 70 individuals with lipoatrophy who underwent screening, 13 met all of these eligibility criteria and eight enrolled. The primary reasons for screen failures were leptin greater than 3 ng/ml (71%) and triglycerides less than 300 mg/dl (52%). The study was approved by the Committee on Human Research of the University of California, San Francisco, and written informed consent was obtained from each subject.
Before treatment, subjects were hospitalized in the General Clinical Research Center (GCRC) at San Francisco General Hospital Medical Center for 5 d to undergo comprehensive studies of carbohydrate and lipid metabolism. Body fat content and distribution were also assessed. After this initial evaluation, treatment with recombinant methionyl human leptin (Amgen Inc., Thousand Oaks, CA) was begun. For the initial 3 months, leptin was administered at a dose of 0.01 mg/kg twice daily, a dose that was expected to increase plasma leptin levels to the low-normal range.
After 3 months, the subjects were readmitted to the GCRC for a repeat 5-d metabolic evaluation, and the dose was then increased to 0.03 mg/kg twice daily. This latter dose was expected to achieve plasma leptin levels in the high-normal range. After 6 months of treatment, subjects were readmitted to the GCRC for a repeat of all of the studies that were performed at baseline and 3 months. Subjects also reported to the GCRC at months 1, 2, 4, and 5 for outpatient assessments that included collection of fasting blood samples, measurement of weight, an updated medical history, and safety assessment.
During each admission, subjects were fed a fixed diet. Energy intake was set at a level to maintain weight; the diets provided a protein intake of 1.5 g/kg • d or greater and carbohydrate intake of 40-50% of total energy content. Subjects were asked to prepare written food intake diaries for 3-7 d before each inpatient admission, and energy and macronutrient intake was analyzed using Food Processor software (version 7.0, ESHA Research, Salem, OR).
Height was measured using a stadiometer. Subjects were weighed under fasting conditions on a calibrated scale. Visceral (VAT) and sc (SAT) abdominal adipose tissue areas at L4-L5 and midthigh cross-sectional muscle and sc fat area were measured by wide-slice magnetic resonance imaging (MRI; Magnetom 1.5T clinical MRI system; Siemens, Malvern, PA) and analyzed using a customized software program written in the interactive data language platform (IDL Research Systems, Inc., Boulder, CO). Fat and lean body mass (LBM) were measured by whole-body dual-energy x-ray absorptiometry (DEXA; Lunar model DPX, Madison, WI) with regional analysis as described previously (13).
Intrahepatic lipid content was measured in the right lobe of the liver by proton magnetic resonance spectroscopy in the aforementioned 1.5T magnet with a loop surface coil and using an image-guided localized proton spectra with a voxel size of approximately 30 cm3. Volume-selective shimming was performed and spectra were recorded by a STEAM technique (repetition time = 3000 msec, echo time = 15 msec, acquisitions 256, averages 1) with and without frequency-selective water suppression. The recorded signals were analyzed using J-MRUI fitting software (14). Chemical shifts were measured relative to water signal intensity at 4.8 ppm. Methylene signal intensity, which represents intracellular triglycerides in the liver (15), was assigned at 1.3-1.5 ppm.
Glucose tolerance and insulin sensitivity
On d 3 of the inpatient study, an oral glucose tolerance test was performed after an overnight fast and consumption of a minimum of 100 g carbohydrate per day during the previous 3 d. After collection of fasting blood samples, subjects drank a solution containing 75 g glucose in 300 ml. Blood samples were collected at 30, 60, 90, 120, and 180 min after ingestion of the glucose solution for determination of plasma glucose and serum insulin levels. Integrated areas under the curve (AUCs) for glucose and insulin were calculated by the trapezoidal method.
Whole-body insulin sensitivity was measured on d 5 by the euglycemic hyperinsulinemic clamp technique (16). After the baseline measurements were completed, insulin (Humulin; Eli Lilly & Co., Indianapolis, IN), bound to human albumin, was infused at a rate of 40 mU/m2 • min for 180 min.. Blood samples were collected at 5-min intervals from a retrograde iv line placed in a hand that was warmed in a heated box at 50-55 C. Whole-blood glucose concentrations were determined by the glucose oxidase method (Stat glucose analyzer; Yellow Springs Instruments, Yellow Springs, OH). A variable infusion of 20% dextrose (labeled with [U-13C] glucose) was adjusted to maintain plasma glucose concentrations at the baseline level. Blood samples were collected at 30-min intervals and the serum was frozen and batched for measurement of insulin. Insulin sensitivity was calculated as a measure of whole-body glucose uptake during the final hour of the clamp and adjusted for endogenous glucose production, steady-state serum insulin levels, and LBM (17).
Stable isotope tracer studies
Endogenous glucose production, gluconeogenesis, glycogenolysis, and whole-body lipolysis were measured under conditions of fasting and hyperinsulinemia (during the clamp) by stable isotope infusion. On the evening of d 4, catheters were placed in the antecubital or forearm veins of each arm for infusion of the isotope solutions and periodic blood sampling. At 0430 h the following day, primed, continuous infusions of [U-13C] glucose (1.2 mg/kg • h), and [2-13C] glycerol (15 mg/kg LBM per hour) were begun. Steady-state blood sampling was performed between 0800 and 0830 h (fasting measurements) and during the final 30 min of the clamp (hyperinsulinemia). Endogenous glucose production (Ra glucose) and lipolysis (Ra glycerol) were calculated using standard dilution techniques (18). Gluconeogenesis was calculated using mass isotopomer distribution analysis (19, 20). Glycogenolysis was calculated by subtracting the rate of gluconeogenesis from that of glucose production.
Oxygen consumption and carbon dioxide production were measured under fasting conditions and during the clamp by indirect calorimetry (DeltaTrac metabolic cart; SensorMedics, Yorba Linda, CA). Resting energy expenditure and substrate oxidation rates were calculated using stoichiometrically derived equations (21).
Fasting blood samples were collected at 0800 h on d 4 for measurement of total, direct low-density lipoprotein (LDL-C) and high-density lipoprotein cholesterol (HDL-C) concentrations, serum leptin levels, CD4 lymphocyte counts, and HIV RNA levels. Triglyceride and free fatty acid (FFA) concentrations were measured on d 5 under fasting and hyperinsulinemic (clamp) conditions. Postprandial triglyceride levels were also measured at 1900 h on d 5, after subjects had consumed a standardized lunch at 1200 h (upon conclusion of the clamp) and snack at 1500 h.
Triglycerides and total cholesterol were measured using Thermo Infinity reagents and HDL-C using the Thermo Trace/DMA HDL Precipitating kit (Thermo Electron, Melbourne, Australia). Direct LDL-C was measured using the Wako L-Type LDL-C reagents and FFA using the Wako NEFA kit (Wako Chemical USA, Richmond, VA). Serum insulin and leptin levels were measured by RIA (Linco Research Inc., St. Charles, MO) in the Core Hormone Laboratory of the GCRC at San Francisco General Hospital. CD-4 T cell counts were measured by flow cytometry, in the clinical laboratory at San Francisco General Hospital. HIV-1 RNA levels were quantified using the Amplicor HIV-1 monitor test (detection limit <75 copies/ml; Roche Diagnostics, Branchburg, NJ).
Data are presented as means ± SEM unless otherwise noted. For normally distributed data, results from pretreatment, month 3, and month 6 assessments were analyzed by one-way repeated-measures ANOVA with Holm-Sidak pairwise multiple comparisons for significance of differences between time points. Nonnormally distributed data were analyzed by repeated-measures ANOVA on ranks with multiple comparisons by Tukey test. Statistical analyses were performed using SigmaStat version 3.5 (Systat Software, Inc., Point Richmond, CA). A two-sided P < 0.05 was considered statistically significant.