Whole-Body Proteolysis Rate Is Elevated in HIV-Associated Insulin Resistance

HIV-infected and HIV-seronegative men and women were recruited from the AIDS Clinical Trials Unit, the Infectious Disease Clinics, and the Volunteers for Health Program at Washington University School of Medicine (Table 1). None of the subjects had an AIDS diagnosis. Subjects were assigned to one of three groups: HIV seronegative with normal glucose tolerance (NGT; control; n = 10), HIV infected with NGT (HIV+NGT; n = 16), or HIV infected with IGT (HIV+IGT; n = 21). IGT was defined as fasting glucose 100–126 mg/dl or glucose 140–200 mg/dl 2 h after ingesting a 75-g glucose beverage (American Diabetes Association criteria). Glucose and fatty acid kinetics for 30 of these 47 subjects were reported previously in a study of men with HIV dyslipidemia (11).

Before enrollment, volunteers received a physical examination, including a medical history, fasting blood chemistry, lipid/lipoprotein and serum endocrine profile, a 2-h oral glucose tolerance test, and plasma HIV RNA quantitation (Roche Amplicor HIV-1 Monitor; Roche Diagnostics, Indianapolis, IN). Current anti-HIV medication regimens were compiled for the HIV-infected participants. Volunteers were excluded if they were taking medications or dietary supplements that affect amino acid metabolism (β-blocker, β-agonist, Ca²⁺ channel blocker, and corticosteroid), had a neuromuscular (severe peripheral neuropathy), or had another disorder that might affect amino acid metabolism. All subjects consumed <3 alcohol-containing beverages per week, were not infected with hepatitis C or B, reported not using recreational drugs for 6 months before enrollment, and were weight stable (<2% weight change in the 3 months before the study). None of the subjects took anabolic agents or appetite stimulants for at least 6 months before study. None of the subjects participated regularly in exercise activities that would constitute exercise training. The Human Studies Committee at Washington University School of Medicine approved the study, and all subjects provided informed consent before participating.

Subjects were admitted to the General Clinical Research Center (GCRC) for a 48-h period. For 3 days before admission and on day 1 of admission, participants consumed a standard weight-maintaining diet that contained defined, adequate amounts of energy and macronutrients: at least 250 g of carbohydrate per day and ∼12% protein, ∼55% carbohydrate, and ∼33% fat calories. Compliance with these guidelines was assessed using 3-day food recall records that were reviewed by a GCRC dietitian. This was done to reduce the effects of prior diet on substrate metabolism quantified during the glucose clamp (day 2 of the GCRC admission). During the GCRC admission, participants abstained from exercise, caffeine, and alcohol ingestion.

Whole-body fat and FFM were quantified using a Hologic QDR-2000 enhanced-array dual-energy X-ray absorptiometer (Waltham, MA). Images were processed by a certified technologist using Hologic software (version 5.64A) (20). Abdominal (subcutaneous and visceral) fat content was quantified using proton magnetic resonance imaging (Siemens, Iselin, NJ). Each was identified in serial axial images obtained at the level of the L₂-L₃ interspace, and their areas (in centimeters squared) were measured using Analyze Direct software (21). Areas from eight serial images of the abdomen were averaged. Intrahepatic lipid content was quantified using proton magnetic resonance spectroscopy (1.5T whole-body system; Magnetom Sonata; Siemens, Erlangen, Germany) (22). Intrahepatic lipid content (percent of water signal) was measured in 9 of 10 control, 13 of 16 HIV+NGT, and 16 of 21 HIV+IGT subjects.

After an overnight fast (from 2000), a catheter was inserted into an antecubital vein (0530) and used to administer stable isotope labeled tracers. A second catheter was inserted into a hand vein on the contralateral arm; the hand was heated (55°C) using a thermostatically controlled box to obtain arterialized venous blood samples. At 0700, primed- constant intravenous infusions of [6,6-²H₂]-glucose (22.5 μmol/kg; 0.25 μmol · kg⁻¹ · min⁻¹) and [1-¹³C]-leucine (7.6 μmol/kg; 0.13 μmol · kg⁻¹ · min⁻¹) were started. All tracers were from Cambridge Isotope Laboratories (Andover, MA). After a baseline period (0–210 min), a two-stage hyperinsulinemic-euglycemic clamp was started. During stage 1 of the clamp, a primed (80 mU/m² per min × 5 min; 40 mU/m² per min × 5 min), constant (20 mU/m² per min) infusion of regular human insulin was administered intravenously and continued for 2 h. During stage 2, a primed (160 mU/m² per min; 80 mU/m² per min), constant (40 mU/m² per min) infusion of human insulin was started and continued for 4 h. The ²H₂-glucose infusion rate was reduced by 50% of the basal infusion rate during stage 1 and by 75% during stage 2 to allow for the anticipated decline in glucose R_a during insulin infusion. Plasma glucose concentration was maintained at 5.5 mmol/l (100 mg/dl) by a variable-rate infusion of 20% dextrose containing 2.5% [6,6-²H₂]-glucose.

Blood samples were obtained before starting the tracer infusions to quantify background ²H and ¹³C enrichments and every 10 min during the last 30 min of the basal period and during each stage of the clamp to quantify hormone levels, substrate levels, and kinetics. Blood samples were obtained every 10 min during the clamp to quantify plasma glucose concentrations and used to adjust the variable 20% dextrose infusion rate.

An automated indirect calorimeter with a ventilated hood system (Sensormedics Deltatrac, Yorba Linda, CA) was used to quantify O₂ consumption and CO₂ production rates and to calculate respiratory exchange ratio during the last 15 min of the basal and stage 1 and 2 periods of the clamp. During the GCRC admission, urine nitrogen excretion was quantified and used to correct respiratory exchange ratio and substrate utilization rates for the contribution of protein utilization.

Plasma glucose concentration was quantified using an automated glucose analyzer (Yellow Spring Instruments, Yellow Springs, OH). Plasma insulin concentration was quantified using radioimmunoassay (23). Fasting serum nonesterified free fatty acid (FFA) levels were quantified using a commercial assay (Wako Chemicals, Richmond, VA). Fasting serum lipid/lipoprotein levels were quantified as described previously (21). Fasting serum interleukin (IL)-8 concentrations were quantified using a commercially available antibody bead–based 96-well microplate assay (Bio-Rad Laboratories, Hercules, CA) following the manufacturer’s protocol. Serum samples were diluted 1:4 in human serum diluent and analyzed in duplicate. Eight calibration standards (0.2–3,200 pg/ml) were analyzed in duplicate on each microplate. The coefficient of variation for replicate analyses was <7% for samples and standards in the concentration range 3–3,200 pg/ml and <16% for those with <3 pg/ml.

The tracer-to-tracee ratios for plasma ²H₂-glucose, ¹³C-leucine, and ¹³C-ketoisocaproic acid (KIC) were quantified using capillary gas chromatography–mass spectrometry (GC-MS; Agilent 6890N gas chromatograph and Agilent 5973N mass selective detector; Agilent, Palo Alto, CA) (20,24,25). For ²H₂-glucose enrichment quantitation, plasma proteins were precipitated with cold acetone, lipids were extracted into hexane, and the aqueous phase was dried (Speed-Vac; Savant Instruments, Farmingdale, NY). The heptafluorobutyric derivative of glucose was formed, and ²H₂-glucose enrichment was quantified using GC-electron ionization-MS and selective ion monitoring (mass/charge ratio [m/z] 519 and 521). For quantitation of plasma leucine concentration, an internal standard ([U-¹³C₆]-leucine) was added to the plasma before isolation. Plasma leucine was converted to the heptafluorobutyric propyl ester derivative; ¹³C-leucine enrichment was quantified using GC-MS in negative-chemical ionization mode (m/z 349 and 350), and leucine concentration was quantified by monitoring ions at m/z 349 and 355. In the same blood samples, plasma KIC was isolated, the trimethylsilyl derivative was formed, and ¹³C enrichment was measured using GC-electron ionization-MS and selected ion monitoring (m/z 232 and 233) (20). The GC-MS instrument response was calibrated using gravimetric standards of known isotope enrichment.

Plasma glucose and leucine R_a were calculated by dividing each tracer infusion rate by the average tracer-to-tracee ratio obtained during the last 30 min of each stage of the clamp. Glucose R_d was calculated as the sum of endogenous glucose R_a plus infused dextrose. Nonoxidative glucose R_d was determined by subtracting carbohydrate oxidation rate from whole-body glucose R_d. The average plasma ¹³C-KIC enrichment value obtained during the last 30 min of each stage of the clamp was used to calculate the rate of whole-body proteolysis (20,26). Kinetic rates were expressed per kilogram FFM. Homeostasis model assessment (HOMA) was calculated as described (27).

Descriptive characteristics, fasting lipid/lipoprotein levels, liver enzymes, and endocrine profiles were compared using Kruskal-Wallis nonparametric tests. Several of these variables had nonhomogeneous variance; thus, a nonparametric test was used and reported values represent as means ± SD. One-way ANOVA was used to compare hormone and substrate levels and substrate kinetics during the clamp among the three groups. In general, clamp-related metabolic parameters varied less, had more homogeneous variance than the descriptive parameters, and reported values represent means ± SE. When significant differences were noted (P < 0.05) among the three groups, a Tukey honestly significant difference post hoc test was used to identify which groups differed. Pearson correlation coefficients were calculated and used to assess bivariate associations between continuous variables.

The three groups were similar (P = NS) with respect to sex distribution, age, BMI, and FFM. HIV+NGT and HIV+IGT subjects had similar plasma HIV RNA levels and CD4⁺ T-cell numbers, but HIV+IGT subjects had a longer duration of known HIV infection and treatment with antiretroviral therapy than HIV+NGT (Table 1). The median plasma HIV RNA level for HIV+NGT and HIV+IGT subjects was 0; 63% HIV+NGT and 81% HIV+IGT subjects had plasma HIV RNA <400 copies/ml. Median CD4⁺ T-cell numbers were not different (496 vs. 501 cells/μl for HIV+NGT vs. HIV+IGT, respectively, P = NS).

Seven of 16 HIV+NGT subjects were naïve to HAART. Of the remaining nine HIV+NGT subjects, all were receiving nucleoside reverse transcriptase inhibitors (six on zidovudine, two on stavudine, and one on lamivudine plus didanosine); six of these were also receiving a non-nucleoside reverse transcriptase inhibitor (five on efavirenz and one on nevirapine), and three were receiving protease inhibitors (one on nelfinavir and two on ritonavir-boosted lopinavir). In HIV+IGT subjects, 2 of 21 were naïve to HAART. Of the remaining 19 subjects, all were receiving nucleoside reverse transcriptase inhibitors (7 on zidovudine, 7 on stavudine, 2 on lamivudine plus didanosine, 1 on lamivudine plus abacavir, and 2 on abacavir plus didanosine); 11 of these also received a non-nucleoside reverse transcriptase inhibitor (7 on efavirenz and 4 on nevirapine), and 12 received protease inhibitors (5 on nelfinavir, 2 on ritonavir-boosted lopinavir, 1 on ritonavir-boosted amprenavir, 1 on ritonavir-boosted saquinavir, and 3 on indinavir).

Visceral adipose tissue area and liver lipid content were higher in HIV+IGT subjects than the other two groups (Table 1). Liver lipid content and visceral adipose area were correlated (P < 0.001, r = 0.66). Liver lipid content was also directly correlated with fasting insulin levels (P = 0.006, r = 0.44) and with HOMA (P = 0.004, r = 0.46) and inversely correlated with glucose R_d (P = 0.03, r = −0.36) during clamp stage 2.

As anticipated, fasting insulin levels and HOMA were higher in HIV+IGT subjects than the other two groups (Table 2). Fasting blood glucose levels were in the normal range but greater in HIV+IGT than control subjects. Serum IL-8 levels were higher in HIV+IGT subjects than the other two groups. Serum IL-8 levels were directly correlated with visceral adiposity (P = 0.01, r = 0.39) and with insulin resistance (HOMA; P = 0.01, r = 0.38) and inversely correlated with glucose R_d during stage 2 of the clamp (P = 0.03, r = −0.34). Plasma triglyceride concentrations were greater in HIV+IGT subjects than the other groups. Fasting serum FFA levels were higher in HIV+IGT than HIV+NGT subjects. Serum FFA concentrations were inversely correlated with whole-body glucose R_d measured during low (P = 0.005, r = −0.41) and high (P = 0.05, r = −0.29) insulin doses and directly correlated with basal insulin levels (P = 0.02, r = 0.31), HOMA index (P = 0.003, r = 0.42), and liver lipid content (P = 0.05, r = 0.31). Fasting glucagon, leptin, cortisol, testosterone, epinephrine, and norepinephrine levels were not different among the groups.

Mean basal leucine concentration in HIV+IGT subjects was higher than in control subjects and higher than control and HIV+NGT subjects during clamp stages 1–2. Mean basal plasma insulin concentration in HIV+IGT subjects was higher than control and HIV+NGT subjects and higher than HIV+NGT subjects during clamp stage 1. Insulin levels during clamp stage 2 were not different among the groups. The primary difference in the HIV+IGT group was higher absolute leucine levels, despite elevated basal insulin levels, which persisted across all insulin levels. The insulin-induced reduction in leucine concentration per unit increase in insulin concentration (slope from basal to stage 1 and from stage 1 to stage 2, Fig. 1) was not different among the groups. This indicates basal leucine levels were resistant to the leucine-lowering actions of basal insulin.

Glucose kinetics, plasma leucine R_a, and whole-body proteolysis rate (calculated using ¹³C-KIC; the immediate intracellular deamination product of leucine) were measured at basal, low, and high plasma insulin levels, while plasma glucose levels were constant (5.5 mmol/l). Basal glucose R_d (micromoles per kilogram FFM per minute) was similar among the groups but was lower in HIV+IGT subjects than HIV+NGT and control subjects during clamp stages 1 and 2 (Table 3). When normalized to the circulating insulin concentration glucose R_d (micromoles per kilogram FFM per minute per microunit insulin per milliliter) was lower in HIV+IGT subjects than HIV+NGT and control subjects during basal and stage 1 and 2. During clamp stage 1, endogenous glucose R_a was higher in HIV+IGT than the other groups.

The absolute values for leucine R_a and whole-body proteolysis rate were greater (P ≤ 0.005) in HIV+IGT subjects than in HIV+NGT and control subjects during all stages of the clamp (Fig. 1). Leucine R_a and whole-body proteolysis rates were not different between control and HIV+NGT subjects during basal or clamp conditions. Despite higher basal insulin concentrations in the HIV+IGT group, the absolute leucine R_a and whole-body proteolysis rate were ∼15% higher than in HIV+NGT and control subjects. Similarly, despite higher insulin concentrations during clamp stage 1, the absolute leucine R_a and whole-body proteolysis rate in HIV+IGT subjects remained ∼15% higher than that in HIV+NGT subjects. During the clamp, the insulin-induced suppression of leucine R_a and whole-body proteolysis rate was not different among the three groups. This implies that basal leucine and protein turnover rates were dysregulated in HIV+IGT subjects, and this elevated rate of leucine release persisted throughout the clamp. However, the absolute magnitude of the suppression in leucine R_a and whole-body proteolysis rate per unit increase in insulin during the clamp was similar among the groups. Plasma ¹³C-KIC/¹³C-leucine ratio tended to decrease with increasing insulin levels (P = NS), but the ratios were not different among the groups at any insulin level (data not shown). Whole-body proteolysis rates were directly correlated with glucose R_a at all insulin concentrations (P < 0.01, r = 0.36–0.47) and fasting FFA levels (P = 0.03, r = 0.32) and inversely correlated with glucose R_d during clamp stage 1 (P = 0.04, r = −0.31).

Lipid oxidation rate was similar in all groups during basal conditions but was greater in HIV+IGT subjects than the other groups during clamp stage 2 and greater than in HIV+NGT subjects during clamp stage 1 (Table 3). Whole-body carbohydrate oxidation rate tended to be lower in HIV+IGT subjects than the other two groups during clamp stage 2, but this did not achieve statistical significance (P = 0.08) (data not shown).

In HIV-infected individuals with IGT, we found greater absolute plasma leucine concentrations, greater rates of leucine appearance in the plasma, and greater rates of whole-body proteolysis under fasting insulin and hyperinsulinemic conditions than in HIV+NGT and healthy control subjects. These markers of fasting branched-chain amino acid (BCAA) metabolism were elevated in HIV+IGT subjects despite fasting insulin levels that were two to two and a half times greater than in HIV+NGT and control subjects. This indicates that HIV+IGT subjects were resistant to the antiproteolytic actions of basal (postabsorptive) insulin levels. However, it appears that HIV+IGT subjects were sensitive to the leucine-lowering, antiproteolytic actions of hyperinsulinemia achieved during the clamp. During hyperinsulinemia, absolute levels of leucine, leucine R_a, and whole-body proteolysis rate remained significantly higher in HIV+IGT subjects, but the magnitude of the insulin-induced suppression for each of these parameters was identical among the groups. This is contrasted with glucose regulation in HIV+IGT subjects, where absolute basal glucose R_a and R_d (micromoles per kilogram FFM per hour) were similar among the groups, but insulin action on glucose R_a and R_d was blunted in HIV+IGT subjects during insulin infusion. Taken together, these findings indicate that dysregulated fasting leucine and protein metabolism, along with impairments in glucose and fatty acid metabolism (9,11,13–16,18), may be integrated and represent metabolic syndromes associated with HIV infection and HAART.

Fasting leucine R_a and whole-body proteolysis rate were not adequately suppressed by fasting insulin levels in HIV+IGT subjects. This is contrary to what is typically observed in type 2 diabetes, where fasting insulin levels adequately maintain (suppress) leucine R_a and whole-body proteolysis rate in the normal range, while glucose R_a is increased. Since the slope of the insulin “dose-response curve” was similar among the groups (Fig. 1), a factor other than insulin resistance may have been responsible for the elevated basal leucine levels and kinetics. Despite similar absolute suppression of leucine R_a during insulin infusion, leucine R_a and whole-body proteolysis were not suppressed to normal levels in HIV+IGT subjects, suggesting that maximal responsiveness of leucine metabolism to insulin may have been impaired. We did not achieve a clear plateau in suppression of leucine R_a at the highest insulin level; thus, we cannot definitely claim that maximal responsiveness was blunted.

Advanced HIV disease is characterized by elevated plasma glutamine R_a and elevated whole-body proteolysis (20,25,28–30). All subjects were weight stable for at least 6 months and had well-controlled, asymptomatic HIV infection, making it less likely that HIV wasting or progressive viral replication was responsible for the elevated proteolysis rate. It is unlikely that increased sympathetic nervous system activity played a role because plasma epinephrine and norepinephrine concentrations were not different among groups. HIV+IGT subjects had several features of hypercortisolism including higher visceral adipose area, hepatic steatosis, and glucose intolerance; however, plasma cortisol levels were not different among groups. Viral protein R acts as a coactivator of the human glucocorticoid receptor, and it is possible that this might link the features of hypercortisolism with HIV+IGT subjects (31). However, this would not explain the differences in leucine metabolism between the two HIV-infected groups. It is possible, but unlikely, that HAART contributed to the elevated proteolytic rate in HIV+IGT subjects. Use of HAART is associated with maintenance of skeletal muscle mass, and we have previously found that HAART improves, but does not completely normalize, muscle protein synthetic or proteolytic rates in HIV-infected individuals (20). Individual components of HAART cannot be excluded as contributing factors to dysregulated leucine/protein metabolism (32); however, this cross-sectional study was not designed, or powered, to identify specific components of HAART that might disrupt amino acid metabolism.

A key finding is that proteolytic rates were elevated during fasting conditions and were related to impaired insulin action on glucose metabolism. Blunted insulin-mediated suppression of glucose R_a is likely due to increased proteolysis and amino acid availability rather than HIV infection or its treatment. This notion is supported by a recent report that elevated whole-body proteolytic rates are related to elevated rates of gluconeogenesis in obese HIV-negative subjects (33). The relationship between whole-body proteolysis and glucose production may be explained by increased hepatic and renal conversion of alanine and glutamine to glucose (34) in HIV+IGT subjects. This may increase reliance on gluconeogenic pathways, rather than glycogenolysis, for glucose production in HIV+IGT subjects.

In agreement with others, we found that fasting FFAs and liver lipid content were elevated and positively correlated in HIV+IGT subjects (11,13,35). The correlations between FFA levels, glucose, and leucine kinetics in HIV+IGT subjects raise the possibility that elevated FFA flux and their byproducts can accumulate in insulin-sensitive tissues and disrupt normal insulin signaling pathways. The pathways that regulate insulin-mediated glucose disposal in skeletal muscle are disrupted in HIV+IGT subjects (17,18,36). Impairments along the insulin signaling pathway were not evaluated in the current study; thus, we cannot identify a primary cellular defect. Our findings suggest that there may be a defect in the insulin signaling pathway, which is common to the regulation of both glucose and BCAA metabolism. Possibilities might include insulin-induced intramyocellular kinase activation cascades that regulate protein synthesis (e.g., Akt, glycogen synthase kinase 3β, mTOR, p70^s6kinase, elongation-initiation factors-2B and -4E, and 4E-binding protein-1) and protein breakdown (Forkhead transcription factors [31], nuclear factor-κB, ubiquitin-proteasome pathway, atrogin, and myostatin). A role for these factors in the pathogenesis of HIV+IGT subjects needs to be examined further.

We observed higher serum IL-8 levels in HIV+IGT subjects, and this is consistent with findings in type 2 diabetes (37), obesity (38), and nonalcoholic steatohepatitis (39). It is possible that the chronic inflammatory state associated with HIV infection, and reflected as elevated fasting serum IL-8 levels, may have contributed to dysregulated glucose, leucine, and protein metabolism in HIV+IGT subjects. IL-8 enhances vascular smooth muscle cell proliferation, monocyte adhesion to endothelial cells, and is believed to have proatherogenic properties (40). IL-8 is produced and released from human subcutaneous and visceral adipose tissue (41) and subcutaneous adipose tissue obtained from HIV-infected individuals with fat redistribution and insulin resistance. Higher serum IL-8 may reflect increased production and release from the larger visceral adipose depot in HIV+IGT subjects, represent a biomarker for inflammation in visceral adipose tissue, or be an indicator/signal for increased cardiovascular disease risk in HIV+IGT subjects.

In summary, we found that patients with HIV infection and IGT have markedly higher rates of fasting proteolysis. Insulin signaling pathways that are common to the regulation of glucose and BCAA metabolism are potential sites for the pathogenesis of HIV-related metabolic syndromes. Lean tissue proteolysis may further worsen HIV hyperglycemia by providing amino acids for gluconeogenesis or, in the case of leucine, directly impair insulin-mediated glucose disposal in muscle. These pathways and the role of specific antiretroviral agents in the pathogenesis of this phenomenon are unclear but deserve further study.

This work was supported by National Institutes of Health Grants RR00036 (GCRC); RR00954 (Biomedical Mass Spectrometry Resource); DK020579 (Diabetes Research and Training Center); DK56341 (Clinical Nutrition Research Unit); AI25903 (AIDS Clinical Trials Unit); DK49393, DK54163, and DK59531 (to K.E.Y.); DK59532 (to W.G.P.); DK59534 (to S.K.); DK066977 (to W.T.C.); and RR19508, AG00078, and DK63683 (to D.N.R.).

We thank the participants for their altruism and patience. Sherry Lassa-Claxton MS, RD, Mary Hoffman, RN, Erin Laciny, and Amanda DeMoss were research coordinators. Jennifer Chen, Sam Smith, Amanda Becker, Freida Custodio, and Junyoung Kwon provided analytical expertise.