Effects of transaldolase exchange on estimates of gluconeogenesis in type 2 diabetes

Abstract

Transaldolase (TA) exchange overestimates gluconeogenesis measured with deuterated water (²H₂O). However, it is unknown whether TA differs in people with type 2 diabetes (T2DM). ²H₂O was ingested, and [1-¹³C]acetate and [3-³H]glucose were infused in T2DM (n = 10) and healthy nondiabetic (ND; n = 8) subjects. TA was assessed from the ratio of ¹³C3 to ¹³C4 glucose enrichment (¹³C3/¹³C4) measured by ¹³C NMR. Glucose turnover was measured before (∼16-h fast) and during hyperglycemic (∼10 mM) moderate-dose insulin (∼0.35 mU·kg⁻¹·min⁻¹) clamp. ¹³C3/¹³C4 in T2DM vs. ND was <1.0 and not different at baseline and clamp, indicating equivalent TA. To determine whether incomplete triose phosphate isomerase (TPI) exchange contributed to asymmetric ¹³C3/¹³C4, [U-¹³C]glycerol was infused in lieu of [1-¹³C]acetate during a separate visit in a subset of ND (n = 7) subjects. Ratio of ¹³C3/¹³C4 obtained following either tracer was <1.0 at baseline and during clamp, indicating that TPI exchange was essentially complete and did not contribute to asymmetric glucose enrichment. Uncorrected and corrected rates of gluconeogenesis were no different (P = not significant) in T2DM vs. ND both at baseline and during clamp. TA correction resulted in equivalent estimates of corrected gluconeogenesis in T2DM and ND that were ∼25–35% lower than uncorrected gluconeogenesis both at baseline and during the clamp. The asymmetric enrichment of glucose from ¹³C-gluconeogenic tracers is attributable to TA exchange and can be utilized to correct for TA exchange. In conclusion, TA exchange does not differ between T2DM and ND under fasting or hyperglycemic clamp conditions, and the ²H₂O method continues to provide an accurate estimation of gluconeogenesis.

after an overnight fast, plasma glucose is not symmetrically labeled from gluconeogenic carbon tracers such as [U-¹³C]glycerol (21) and [3-¹⁴C]lactate (25). Exchange of fructose 6-phosphate and triose phosphate mediated by transaldolase has been implicated by us (3, 7, 9) and others (12) in overestimation of rates of gluconeogenesis, utilizing the deuterated water method. In this method, glucose derived from glycogenolysis (GGL) via glucose 6-phosphate (G6P) source can become labeled in carbons 4, 5, and 6 from either a carbon-labeled gluconeogenic precursor or deuterated water independent of gluconeogenesis. This results in an overestimation of the gluconeogenic fraction and a corresponding underestimation of the glycogenolytic contribution to endogenous glucose production. This exchange is particularly important in the study of type 2 diabetes mellitus (T2DM) since it mimics the characteristic shift toward increased hepatic gluconeogenic activity during the progression of this disease. With a few exceptions (14), stable isotope tracer studies indicate an increased fractional gluconeogenesis in overnight-fasted T2DM subjects compared with healthy nondiabetic (ND) controls (6, 10, 15, 16, 28). Transaldolase (TA) exchange activity has been implicated in both ND and T2DM subjects from the depletion of position 5 relative to position 3 label following metabolism of [3,5-²H₂]galactose to glucose (7). However, it is not known whether transaldolase exchange activity is different between ND and T2DM under conditions of fasting and hyperglycemia (frequently seen with meals). This study was undertaken to compare TA exchange reaction in ND and T2DM subjects using [1-¹³C]acetate infusion under conditions of fasting and during a hyperglycemic moderate-dose insulin clamp to determine whether or not TA exchange activity explains differences in glucose enrichment from gluconeogenic substrates between ND and T2DM subjects.

We have demonstrated that TA activity occurs in healthy humans based on a higher enrichment of glucose carbon 4 over carbon 3 (i.e., ¹³C3/¹³C4 < 1.0) from [1-¹³C]acetate (3). In this study, the asymmetric labeling of glucose from this tracer could also be explained by incomplete exchange of the label between glyceraldehyde 3-phosphate and dihydroxyacetone phosphate mediated by incomplete triose phosphate isomerase (TPI; see Fig. 1). When the carbon label is derived from glycerol instead of acetate, incomplete TPI exchange results in the opposite glucose C3/C4 labeling asymmetry (i.e., ¹³C3/¹³C4 > 1.0), whereas TA activity promotes a higher enrichment of glucose carbon 4 relative to carbon 3 for both tracers (21). Thus, by comparing glucose ¹³C3/¹³C4 from both acetate and glycerol, the contributions of incomplete TPI exchange and TA activity to the glucose-labeling asymmetry can be resolved.

Fig. 1.

Schematic highlighting the effects of incomplete triose phosphate isomerase (TPI) exchange and transaldolase (TA) exchange on the distribution of glucose ¹³C isotopomers (●, ¹³C; ○, ¹²C) and ¹³C enrichment of glucose carbon 3 and 4 positions (highlighted by the boxes) from [U-¹³C]glycerol and [1-¹³C]acetate. Glucose isotopomers favored by partial TPI exchange only (A) and TA activity in the presence of complete TPI exchange (B) are indicated by the bold arrows. When TPI exchange is complete, the 2 tracers generate equivalent glucose ¹³C3/¹³C4 ratios, and these are modified in the same manner by TA exchange (i.e., ¹³C3/¹³C4 < 1.0). DHAP, dihydroxyacetone phosphate; G6P, glucose 6-phosphate; FBP, fructose 1,6-bisphosphate; F6P, fructose 6-phosphate; GA3P, glyceraldehyde 3-phosphate.

Therefore, we also compared the ¹³C enrichment of glucose from [1-¹³C]acetate and [U-¹³C]glycerol in a subset of ND subjects to determine the extent to which asymmetric glucose C3 and C4 enrichments are determined by incomplete TPI exchange and by TA exchange.

RESEARCH DESIGN AND METHODS

After approval from the Mayo Institutional Review Board was given for our experiments, 11 T2DM and 10 ND subjects gave written informed consent to participate in the study. Data from one T2DM and two ND subjects were excluded from analyses due to insufficient ¹³C3 and ¹³C4 glucose enrichments for reliable quantification of ¹³C3/¹³C4. All subjects were in good health and at stable weights. None regularly engaged in exercise training programs. None of the first-degree relatives of the healthy subjects had a history of diabetes mellitus. At the time of screening, one of the T2DM subjects was being treated with lifestyle modification alone; eight T2DM subjects were being treated with metformin and two on a combination of a sulfonylurea and metformin. Thiazolidinedione and glucagon-like peptide-1-based therapies were exclusionary. Oral antihyperglycemic medications were discontinued 10 days prior to study. All subjects were instructed to maintain their body weight constant and follow a weight maintenance diet (55% carbohydrate, 30% fat, and 15% protein) ≥2 wk prior to study. Body composition was measured prior to the study visit in the Mayo Center for Translational Science Activities (CTSA) body composition core using Lunar iDXA software version 6.10 (GE Healthcare Technologies, Madison, WI), and waist/hip ratio was measured at screen visit. Subject characteristics are given in Table 1.

Table 1.

Subject characteristics

Characteristic	Healthy	T2DM
Age, yr	60 ± 12	64 ± 8.7
Sex (M/F)	4/4	6/4
Weight, kg	95.0 ± 21.5	86.1 ± 12.5
BMI, kg/m2	31.8 ± 4.7	30.6 ± 3.7
Lean body mass, kg	51.5 ± 11.7	49.8 ± 10.2
Fasting plasma glucose, mg/dl	90 ± 5	143.0 ± 49.4*
Hb A1c,% (mmol/mol)	5 ± 0.2 (31 ± 2)	7.1 ± 0.6 (54 ± 7)*
Waist/hip ratio	0.9 ± 0.1	1.0 ± 0.1

Values are means ± SD. T2DM, type 2 diabetes mellitus; M, male; F, female.

^*P < 0.05 vs. healthy.

Experimental Design

Subjects were admitted to the clinical research unit of the Mayo CTSA at ∼1700 on the evening before the study and provided a standard supper (10 calories/kg; carbohydrate/fat/protein, 55:30:15). Subjects then ingested 1.67 g/lean body wt ²H₂O in three equally divided doses at 2200, 2400, and 0200. Thereafter, the subjects were permitted sips of water containing ²H₂O if they so desired but otherwise remained fasting.

Subjects were awakened the following morning and catheters were placed in forearm veins for tracer infusion and sampling of arterialized venous blood as previously described (8). At 0600 (−180 min) infusions of [3-³H]glucose (12 μCi prime and 0.12 μCi/min continuous), and [1-¹³C]acetate (5.0 μmol·kg⁻¹·min⁻¹) or [U-¹³C]glycerol (0.5 μmol·kg⁻¹·min⁻¹) were started and continued until the end of study at 1300. At ∼0630, 1 g of acetaminophen was given and repeated at ∼0845 to enable measurement of urinary glucuronide. At time zero, somatostatin (60 ng·kg⁻¹·min⁻¹), insulin (0.35 mU·kg⁻¹·min⁻¹), and glucagon (0.65 ng·kg⁻¹·min⁻¹) were started to ensure constant and equal portal concentrations of insulin and glucagon (3, 8). Blood was sampled for glucose, [3-³H]glucose-specific activity, and hormones at −30 and 0, 60, 120, 180, 210, and 240 min. Pooled samples for [3-¹³C]glucose, [4-¹³C]glucose, and [5-²H]glucose and [2-²H]glucose enrichments were obtained (3).

An infusion of 50% dextrose containing [3-³H]glucose was started at time zero and given in amounts sufficient to clamp plasma glucose at ∼180 mg/dl, as described previously (8). In addition, the basal infusion of [3-³H]glucose was tapered beginning at time zero in a pattern that mimicked the anticipated changes in glucose production to minimize the changes in plasma glucose-specific activity, as described previously (2, 5).

Analytical Methods

Samples at −30 and 0 min were combined for baseline and 210 and 240 min for clamp measurements of ²H and ¹³C enrichments by ²H and ¹³C NMR analysis of monoacetone glucose (MAG). All blood samples were immediately placed on ice, centrifuged at 4°C, separated, and stored at −80°C until analyses. Plasma glucose was analyzed using a GM9 Analox glucose analyzer (Analox Instruments, London, UK). Plasma insulin, C-peptide, and glucagon concentrations and [3-³H]glucose-specific activity were measured as described previously (2, 3).

¹³C-excess enrichment of glucose carbons 3 and 4 from [1-¹³C]acetate was measured by quantitative ¹³C and ¹H NMR analysis (see Fig. 2) of the MAG derivative, as described previously (19, 20). ¹³C-excess enrichment of glucose carbons 3 and 4 from [U-¹³C]glycerol was measured from partially saturated and no nuclear Overhauser enhancement (nOe-enhanced) proton-decoupled ¹³C NMR spectra by analyzing the carbon 3 and carbon 4 isotopomer signals (see Fig. 3). Briefly, the singlet component of each signal was assumed to represent the 1.11% natural abundance ¹³C, and the doublet components were assumed to represent ¹³C enrichment of plasma glucose from [U-¹³C]glycerol. Excess ¹³C enrichment was calculated as the ratio of doublet to singlet multiplied by 1.11% (21). The ¹³C NMR analyses were performed with a BrukerAvance III 600 system equipped with a 5-mm TCP-QNP cold probe. ²H NMR spectra were acquired at 50°C with a 14.1T Varian VNMR system, as described previously (19, 20). ²H enrichment was calculated by comparing the hexose positional ²H signal intensities with those of the MAG methyl signals enriched to 2% ²H (19). NMR signals were quantified using the NUTS NMR spectral analysis program (Acorn NMR, Fremont, CA).

Fig. 2.

²H NMR and ¹³C NMR spectra of monoacetone glucose derivative from plasma glucose of a healthy subject administered with ²H₂O and infused with [U-¹³C]glycerol and [1-¹³C]acetate in separate visits. Spectra were obtained at baseline and during a clamp. Nos. above the signals indicate their position in the glucose molecule. The chemical shift axes of the spectra are omitted for clarity.

Fig. 3.

High-resolution view of the baseline [U-¹³C]glycerol ¹³C NMR spectrum shown in Fig. 2. The carbon 3 and carbon 4 isotopomer signals and their 1- and 2-bond coupling constants are shown in expanded form.

Calculations

All rates are expressed per lean body mass. Glucose turnover was calculated using the steady-state equations of Steele, as described previously (27). The fraction of plasma glucose derived from TA exchange was calculated from the ratio of excess glucose carbon 3 to carbon 4 ¹³C enrichments (¹³C3/¹³C4) (3). Endogenous glucose production (EGP) was calculated by subtracting the glucose infusion rate required to maintain hyperglycemia from total glucose appearance. Gluconeogenesis was calculated by multiplying the ratio of deuterium enrichment of positions 5 and 2 of glucose (²H₅/²H₂) times endogenous glucose production. GGL was calculated by subtracting gluconeogenesis from EGP, as described previously (8). TA-corrected rates of gluconeogenesis and GGL were calculated as published previously (3, 4).

Statistical Analysis

Data in the text and figures are expressed as means ± SE. Values from −30 and 0 min were averaged as basal and at 210 and 240 min as clamp for statistical analysis and representation. Student's paired one-tailed t-test was used to test the hypothesis that plasma glucose ¹³C3/¹³C4 is less than one, and Student's unpaired two-tailed t-test was used to test differences between groups. A P value <0.05 was considered statistically significant.

RESULTS

Effects of TA Exchange

Plasma glucose, insulin, C-peptide, and glucagon concentrations.

Fasting plasma glucose concentrations (Fig. 4) were higher (P < 0.0001) in subjects with T2DM than in ND subjects at baseline (10.74 ± 0.72 vs. 5.34 ± 0.15 mM) but were matched [P = not significant (NS)] during the clamp (10.00 ± 0.52 vs. 9.43 ± 0.13 mM) (Fig. 4). Fasting insulin concentrations (Fig. 4) were significantly higher (P < 0.05) in T2DM than in ND subjects at baseline (59.34 ± 7.30 vs. 34.65 ± 5.59 pM) but did not differ (P = NS) during the clamp (134.94 ± 4.64 vs. 126.30 ± 7.99 pM). Similarly, C-peptide concentrations (Fig. 4) were significantly higher (P < 0 .05) in T2DM than in ND subjects before the clamp (1.32 ± 0.13 vs. 0.97 ± 0.13 nM) but were no different (P = NS) during the clamp (0.16 ± 0.02 vs. 0.16 ± 0.02 nM), indicating that there was suppression of endogenous insulin secretion in both groups. Glucagon concentrations (Fig. 4) were significantly higher (P < 0 .05) in T2DM than in ND subjects before the clamp (78.67 ± 5.37 vs. 57.56. ± 5.09 nM) and, as expected, were not different (P = NS) during the clamp (65.17 ± 2.71 vs. 56.25 ± 3.20 nM).

Fig. 4.

Plasma glucose, insulin, C-peptide, and glucagon concentrations observed before and during the clamp on [1-¹³C]acetate infusion in healthy (□) and type 2 diabetes mellitus (T2DM; ▲) subjects and on[U-¹³C]glycerol infusion in healthy subjects (gray squares). An insulin infusion was started at time zero.

Glucose infusion rate and [3-³H]glucose-specific activity.

The glucose infusion rate (Fig. 5) required to maintain hyperglycemia (∼10 mM) during the clamp was significantly lower (P < 0.05) in T2DM than in ND subjects (10.63 ± 3.08 vs. 33.39 ± 8.93 μmol·kg⁻¹·min⁻¹) (Fig. 5).

Fig. 5.

Top: glucose infusion rate required to maintain target glucose concentrations (∼10 mM) during [1-¹³C]acetate infusion in healthy (□) and T2DM subjects (▲) and [U-¹³C]glycerol infusion in healthy subjects (gray squares). Bottom: plasma [3-³H]glucose-specific activity during [1-¹³C]acetate infusion in healthy (□) and T2DM subjects (▲) and [U-¹³C]glycerol infusion in healthy subjects (gray squares).

Plasma [3-³H]glucose-specific activity (Fig. 5) remained constant throughout the study in both groups, enabling measurement of glucose fluxes using steady-state equations.

Plasma glucose ¹³C enrichments and ratio of ¹³C3/¹³C4 obtained from [1-¹³C]acetate.

¹³C NMR spectra of plasma MAG obtained following infusion of [1-¹³C]acetate demonstrate a consistently higher ¹³C NMR signal intensity of carbon 4 over carbon 3 (see Fig. 2 and Table 2). Since these spectra were acquired under quantitative conditions (i.e., no nuclear Overhauser enhancement during proton decoupling and full relaxation of the ¹³C signals between pulses) as in previous studies (3), these data (Table 2) indicate an excess ¹³C enrichment of carbon 4 over carbon 3 or ratio of ¹³C3/¹³C4 < 1.0. This is consistent with TA activity, where unlabeled G6P derived from GGL (under baseline conditions) or a combination of GGL and inflow of plasma glucose (under clamp conditions) is enriched in positions 4, 5, and 6 from labeled trioses. The ²H enrichment patterns that are revealed by the ²H NMR spectra also suggest that glucose ¹³C enrichment from [1-¹³C]acetate occurs because of TA exchange. For both ND and T2DM groups, there are substantial GGL and GGL/glucose contributions to EGP under baseline and clamp conditions, respectively, as seen by the ratio of positions 5 and 2 enrichments (Fig. 2 and Table 3), thereby providing a significant source of unlabeled G6P for TA exchange.

Table 2.

Plasma [3-¹³C]glucose/[4-¹³C]glucose ratios observed during transaldolase exchange and triose phosphate isomerase exchange

[1-13C]Acetate Infusion	13C3 enrichment	13C4 enrichment	13C3/13C4
Baseline
Healthy	0.29 ± 0.03	0.39 ± 0.03	0.72 ± 0.04
T2DM	0.26 ± 0.05	0.38 ± 0.06	0.66 ± 0.02
Clamp
Healthy	0.30 ± 0.03	0.43 ± 0.06	0.72 ± 0.05
T2DM	0.36 ± 0.06	0.52 ± 0.07	0.66 ± 0.03
[U-13C]glycerol infusion
Healthy (baseline)	0.46 ± 0.03	0.56 ± 0.02	0.81 ± 0.05
Healthy (clamp)	0.23 ± 0.02	0.32 ± 0.04	0.72 ± 0.04

Data are presented as means ± SE. ¹³C enrichment of glucose in carbons 3 and 4 in healthy subjects infused with [1-¹³C]acetate and [U-¹³C]glycerol and type 2 diabetic subjects with [1-¹³C]acetate alone under baseline and clamp conditions.

Table 3.

Deuterium enrichment in positions 5 and 2 of plasma glucose

[1-13C]Acetate Infusion	2H2 Enrichment	2H5 Enrichment (Uncorrected)	2H5/2H2 (Uncorrected)	2H5 Enrichment (Corrected)	2H5/2H2 (Corrected)
Baseline
Healthy	0.51 ± 0.03	0.29 ± 0.02	0.57 ± 0.03	0.20 ± 0.01	0.40 ± 0.01
T2DM	0.52 ± 0.03	0.29 ± 0.02	0.55 ± 0.02	0.19 ± 0.02	0.36 ± 0.02
Clamp
Healthy	0.24 ± 0.04	0.14 ± 0.03	0.53 ± 0.03	0.10 ± 0.03	0.39 ± 0.04
T2DM	0.46 ± 0.02	0.25 ± 0.02	0.56 ± 0.03	0.17 ± 0.02	0.37 ± 0.03
[U-13C]glycerol infusion
Healthy (baseline)	0.47 ± 0.02	0.27 ± 0.02	0.56 ± 0.02	0.22 ± 0.03	0.45 ± 0.05
Healthy (clamp)	0.20 ± 0.02	0.11 ± 0.01	0.54 ± 0.02	0.08 ± 0.01	0.39 ± 0.04

Data are presented as means ± SE. Deuterium (²H) enrichments of glucose on hydrogens 5 (²H₅) and 2 (²H₂) from ingested ²H₂O and the ²H₅/²H₂ ratio uncorrected and corrected for transaldolase exchange in healthy and T2DM subjects infused with [1-¹³C]acetate and in healthy subjects infused with [U-¹³C]glycerol alone under baseline and clamp conditions.

Rates of EGP and glucose disappearance.

Rates of EGP (Fig. 6) were significantly higher (P < 0.05) in T2DM than in ND subjects at baseline (19.78 ± 0.75 vs. 17.53 ± 0. 44 μmol·kg⁻¹·min⁻¹) but similar (P = 0.06) during clamp (7.50 ± 0.69 vs. 4.27 ± 1.23 μmol·kg⁻¹·min⁻¹) (Fig. 6). Rates of glucose disappearance are equal to glucose production at baseline. Therefore, rates of glucose disappearance (Fig. 6) were significantly higher (P < 0.05) in T2DM than in ND subjects at baseline (19.78 ± 0.75 vs. 17.53 ± 0.44 μmol·kg⁻¹·min⁻¹) but lower during clamp (18.61 ± 3.10 vs. 35.85 ± 6.22 μmol·kg⁻¹·min⁻¹). The higher rates of EGP and glucose disappearance in T2DM than in ND subjects at baseline need to be interpreted in the light of higher glucose and insulin concentrations in T2DM during this period. In contrast, when glucose and insulin concentrations were matched between groups during the clamp, there were no statistically significant differences in rates of EGP, but rates of glucose disappearance were lower in T2DM than ND subjects.

Fig. 6.

Rates of endogenous glucose production and glucose disappearance observed at baseline and clamp during [1-¹³C]acetate infusion in healthy (open bars) and T2DM subjects (black bars) and [U-¹³C]glycerol infusion in healthy subjects (gray bars). *P < 0.05 vs. healthy acetate.

Rates of gluconeogenesis and GGL uncorrected and corrected for TA exchange.

Corrected rates of gluconeogenesis (Fig. 7) in T2DM were lower (P < 0.05) than uncorrected rates of gluconeogenesis both prior to (7.32 ± 0.63 vs. 10.50 ± 0.78 μmol·kg⁻¹·min⁻¹) and during the clamp (2.71 ± 0.21 vs.4.11 ± 0.40 μmol·kg⁻¹·min⁻¹) (Fig. 7). Corrected rates of gluconeogenesis (Fig. 7) were lower (P < 0.05) than uncorrected rates of gluconeogenesis prior to (6.65 ± 0.43 vs. 9.54 ± 0.46 μmol·kg⁻¹·min⁻¹) but not during the clamp (1.68 ± 0.62 vs. 2.48 ± 0.78 μmol·kg⁻¹·min⁻¹) in ND subjects. The fractional difference between uncorrected and corrected rates of gluconeogenesis were not different (P = NS) between T2DM and ND subjects at baseline or during clamp, suggesting that the contribution of TA exchange to the estimate of fractional gluconeogenesis is no different between groups under these experimental conditions.

Fig. 7.

Contribution of gluconeogenesis (black bars) and glycogenolysis (open bars) to endogenous glucose production (EGP) observed during [1-¹³C]acetate infusion before and during the clamp in healthy and T2DM subjects, calculated using the uncorrected and corrected ratios of the deuterium enrichment on position 5 and position 2 of plasma glucose. *P < 0.05 vs. uncorrected gluconeogenesis.

Corrected rates of GGL (Fig. 7) in T2DM were higher (P < 0.05) than uncorrected rates of GGL both prior to (12.45 ± 0.58 vs.9.18 ± 0.65 μmol·kg⁻¹·min⁻¹) and during the clamp (4.80 ± 0.48 vs.3.40 ± 0.47 μmol·kg⁻¹·min⁻¹). In contrast, corrected rates of GGL (Fig. 7) were higher (P < 0.05) than uncorrected rates of GGL prior to (10.88 ± 0.36 vs. 7.99 ± 0.55 μmol·kg⁻¹·min⁻¹) but not during the clamp (2.59 ± 0.71 vs. 1.79 ± 0.46 μmol·kg⁻¹·min⁻¹) in ND subjects.

Effects of TPI Exchange

Plasma glucose, insulin, C-peptide, and glucagon concentrations.

Fasting plasma glucose concentrations (Fig. 4) in ND subjects were no different between [1-¹³C]acetate and [U-¹³C]glycerol studies (5.34 ± 0.15 vs. 5.17 ± 0.12 mM, P = NS; Fig. 4). Similarly, plasma glucose concentrations were also matched during the clamp (9.43 ± 0.13 vs.9.36 ± 0.40 mM, P = NS) on either occasion (Fig. 4). Plasma insulin concentrations (Fig. 4) were similar (P = NS) on both acetate as well as glycerol study days both prior to (34.65 ± 5.59 vs. 34.80 ± 4.61 pM) and during the clamp (126.30 ± 6.36 vs. 127.97 ± 4.37 pM). Similarly, C-peptide concentrations (Fig. 4) were no different (P = NS) on acetate vs. glycerol study days either before (0.97 ± 0.13 vs. 0.88 ± 0.11 nM) or during the clamp (0.16 ± 0.02 vs. 0.16 ± 0.03 nM), indicating comparable suppression of endogenous insulin secretion on both study days. Glucagon concentrations (Fig. 4) were no different (P = NS) on either study day before (57.56 ± 5.09 vs. 56.35 ± 3.79 nM) or during the clamp (56.25 ± 3.20 vs. 60.71 ± 3.35 nM).

Glucose infusion rate and [3-³H]glucose-specific activity.

The glucose infusion rate (Fig. 5) required to maintain hyperglycemia (∼10 mM) during the clamp was no different in ND subjects on acetate vs. glycerol (33.39 ± 8.93 vs. 34.07 ± 5.96 μmol·kg⁻¹·min⁻¹; Fig. 5) study days.

Plasma [3-³H]glucose-specific activity (Fig. 5) remained constant throughout the study on both days, enabling measurement of glucose fluxes using the steady-state equation.

Plasma glucose ¹³C enrichments and ratio of ¹³C3/¹³C4 obtained from [1-¹³C]acetate and [U-¹³C]glycerol.

Following the infusion of [U-¹³C]glycerol, ¹³C NMR spectra of glucose observed showed ¹³C enrichment of glucose carbons 3 and 4 as ¹³C/¹³C spin-coupled doublets flanking central singlet carbon 3 and 4 signals, as shown in Fig. 3. The carbon 4 doublet was further split into two equal components by long-range coupling to carbon 6 (21), consistent with the [4,5,6-¹³C₃]glucose isotopomer being the principal contributor to carbon 4 ¹³C enrichment. This was also confirmed by inspection of the carbon 5 and 6 isotopomer signals (data not shown). The monoacetone carbon 3 resonance does not have significant coupling with carbon 1, and hence, the carbon 3 enrichment appeared as a simple doublet, representing nearest-neighbor coupling with carbon 2. Analysis of the carbon 2 multiplet showed that [1,2,3-¹³C₃]glucose isotopomer constituted the majority (>80%) of ¹³C enrichment of the “upper” triose moiety of glucose and was the principal contributor to carbon 3 excess ¹³C enrichment. ¹³C3/¹³C4 ratios were <1.0 in all ND subjects on glycerol study visit at baseline and during clamp (Table 2 and Fig. 8). Similarly, ¹³C3/¹³C4 ratio derived from [1-¹³C]acetate observed was also <1.0 in all ND subjects studied (Fig. 8) and was not significantly different (P = NS) from those obtained with [U-¹³C]glycerol (Table 2 and Fig. 8). The concordance of [1-¹³C]acetate- and [U-¹³C]glycerol-derived ¹³C3/¹³C4 values is consistent with TA activity in the setting of nearly complete TPI exchange, as explained in Fig. 1.

Fig. 8.

Plasma [3-¹³C]glucose to [4-¹³C]glucose ratio observed at baseline (left) and during the clamp (right). *P < 0.001 vs. 1.

Rates of EGP and glucose disappearance.

Rates of endogenous glucose production in ND subjects were no different (P = NS) at baseline (17.53 ± 0.44 vs. 17.38 ± 0.40 μmol·kg⁻¹·min⁻¹) on either acetate or glycerol study days (as shown in Fig. 6). Similarly, suppression of EGP during clamp was no different (P = NS, 4.27 ± 1.23 vs. 3.33 ± 0.74 μmol·kg⁻¹·min⁻¹) on either acetate or glycerol study days (Fig. 6).

At baseline, EGP equals glucose disappearance. Rates of glucose disappearance were no different (P = NS) prior to clamp (Fig. 6). During the clamp, rates of glucose disappearance were similar on both acetate vs. glycerol study days (35.85 ± 6.22 vs. 36.58 ± 5.26 μmol·kg⁻¹·min⁻¹), suggesting comparable insulin-induced glucose uptake on both occasions (Fig. 6).

Rates of gluconeogenesis and GGL uncorrected and corrected for TA exchange reaction.

Uncorrected and corrected rates of gluconeogenesis in ND subjects were similar (P = NS) on acetate and glycerol study days both prior to (9.54 ± 0.46 vs. 9.82 ± 0.56 μmol·kg⁻¹·min⁻¹, 6.65 ± 0.43 vs. 7.85 ± 0.89 μmol·kg⁻¹·min⁻¹) and during the clamp (2.48 ± 0.78 vs. 1.83 ± 0.62 μmol·kg⁻¹·min⁻¹, 1.68 ± 0.62 vs. 1.32 ± 0.32 μmol·kg⁻¹·min⁻¹) (data not shown).

Uncorrected and corrected rates of glycogenolysis in ND subjects were no different (P = NS) on acetate and glycerol study days either prior to (7.99 ± 0.55 vs. 7.56 ± 0.36 μmol·kg⁻¹·min⁻¹, 10.88 ± 0.36 vs. 9.53 ± 0.76) or during the clamp (1.79 ± 0.46 vs. 1.50 ± 0.31 μmol·kg⁻¹·min⁻¹, 2.59 ± 0.71 vs. 2.01 ± 0.42 μmol·kg⁻¹·min⁻¹) in ND subjects (data not shown). These data indicate that there is nearly complete TPI exchange and that overestimation of gluconeogenesis is due to TA exchange alone.

DISCUSSION

The present study demonstrates that TA exchange is equal in people with and without type 2 diabetes both in the overnight-fasted state and during matched hyperglycemia and portal insulin concentrations. The deuterated water method is utilized extensively for estimation of rates of gluconeogenesis (1, 6, 11, 13, 15, 17, 26). Together with glucose tracer methodology, this is important for understanding the progression of type 2 diabetes and for evaluating the effects of various therapeutic strategies (viz., medications) intended to mitigate excessive hepatic glucose production by inhibiting gluconeogenesis. Therefore, validation of tracer-derived gluconeogenesis measurements is possible in principle, for example, by comparison with gluconeogenic fluxes obtained by subtraction of real-time glycogenolysis and EGP (18, 23, 24), a measurement that is insensitive to TA activity. However, outside of a few specialized clinical research centers, this approach is not available, and in many cases it does not provide reliable estimates of gluconeogenesis under clamp conditions. Therefore, recent studies have focused on the analysis of glucose enrichment from ¹³C tracers that are coadministered with ²H₂O. Apart from providing a wealth of additional metabolic information for estimating gluconeogenic fluxes in various settings (12, 22, 29), these tracers also can be used to identify asymmetric enrichment of glucose from TA activity.

A relevant limitation of using [1-¹³C]acetate as a tracer for determining TA activity is that the observed glucose ¹³C3/¹³C4 ratio can be explained by incomplete equilibration of the ¹³C label into dihydroxyacetone acetone phosphate and glyceraldehyde 3-phosphate pools by TPI, as illustrated in Fig. 1. To address this, we infused [U-¹³C]glycerol, a tracer that enters the triose phosphate pool distally to [1-¹³C]acetate and would generate glucose with a higher enrichment of carbon 3 over carbon 4 if TPI exchange were incomplete, as shown in Fig. 1.

Ratio of ¹³C3/¹³C4 obtained during fasting (∼16 h) via [1-¹³C]acetate was ∼0.7, whereas the values obtained with [U-¹³C]glycerol were ∼0.8. In comparison, the enrichment differences for overnight-fasted and 24-h-fasted subjects administered with [U-¹³C]propionate were 0.92 and 0.79, respectively (12). Whereas current data from our [U-¹³C]glycerol study are in reasonable agreement with these values, the [1-¹³C]acetate data tend to show more variability. We are not sure whether this reflects real differences in hepatic acetate metabolism over that of glycerol and propionate systematic differences in NMR quantification of excess enrichments from ¹³C singlet vs. spin-coupled isotopomer signal difference due to experimental conditions or a combination of all of the above. For assessment of ¹³C3/¹³C4 from [1-¹³C]acetate, when the signal contribution from excess ¹³C enrichment is small (<10% of the natural abundance signal or <0.1% excess ¹³C enrichment), relatively small systematic errors in measurement of the carbon 3 and carbon 4 resonances propagate much larger errors in estimates of ¹³C3/¹³C4. For reasons not currently ascertainable, analysis of these very low enrichment samples yielded implausibly low values of ¹³C3/¹³C4. Because of this, we excluded analysis of subjects infused with [1-¹³C]acetate that had ¹³C3 or ¹³C4 glucose excess enrichments of <0.1%. ¹³C NMR isotopomer analyses are more robust in that the ¹³C excess enrichment information is directly obtained from the ratio of different ¹³C NMR signals rather than from subtraction of two signals of similar intensities. The precision of isotopomer analysis is limited only by the signal-to-noise ratio of the weaker ¹³C signal component of the resonance (usually the multiplet). We chose to measure the C3 and C4 multiplet signals over the others (i.e., C2 vs. C5 or C1 vs. C6) because they provide the highest signal-to-noise ratio for a given ¹³C enrichment level and are contributed largely by the principal [1,2,3-¹³C₃]- and [4,5,6-¹³C₃]glucose isotopomers formed from [U-¹³C]glycerol. In principle, an isotopomer analysis of ¹³C acetate incorporation into the two triose halves of glucose could also be determined by replacing [1-¹³C]acetate with [1,2-¹³C₂]acetate.

We have demonstrated previously that the pattern of glucose 123 vs. 456 enrichment from [U-¹³C]glycerol is far more sensitive to incomplete TPI exchange compared with tracers that enter the triose phosphate pool as glyceraldehyde 3-phosphate, such as [1-¹³C]acetate and [U-¹³C]propionate (21). For [U-¹³C]glycerol, incomplete TPI exchange generates glucose with ¹³C3/¹³C4 > 1.0, i.e., opposing the effect of TA. The overall tendency for a higher ¹³C3/¹³C4 ratio from [U-¹³C]glycerol compared with [1-¹³C]acetate is consistent with TPI exchange that approaches 100% completion.

As described recently, the effects of TA on measurements of gluconeogenesis by the ²H₂O method can be corrected by analyzing ¹³C3/¹³C4 from coadministration of a ¹³C tracer (3, 12, 21). Since equilibration of the ¹³C label between glyceraldehyde 3-phosphate and dihydroxyacetone phosphate (DHAP) via TPI is essentially complete, the ¹³C tracer can be a DHAP precursor (i.e., [¹³C]glycerol or dihydroxyacetone) or a glyceraldehyde 3-phosphate precursor (¹³C-propionate, acetate, or bicarbonate), although glucose enrichment from the latter group is less sensitive to incomplete TPI exchange (21).

Consistent with our previous observation in healthy humans and in prediabetes (3, 4), we have now observed that a comparable fraction of the hepatic G6P pool also undergoes TA exchange in people with type 2 diabetes. TA exchange explains the labeling asymmetry of plasma glucose from gluconeogenic carbon tracers and contributes to the enrichment of glucose position 5 from ²H₂O. The fraction of hepatic G6P that undergoes TA exchange is relatively constant and does not differ with varying degrees of insulin sensitivity, ranging from healthy subjects to highly insulin-resistant patients with established type 2 diabetes, nor with varying doses of insulin (3, 4, 7, 9). Since the effect of TA exchange on the source of EGP depends on the contribution of glycogenolysis to EGP, its effects are proportionally highest under conditions of abnormally high hepatic G6P production from glycogenolysis. For example, if glycogenolysis contributed 75% of EGP and gluconeogenesis 25%, the presence of a modest level of TA exchange activity (20% exchange of hepatic G6P) would result in apparent glycogenolytic and gluconeogenic contributions of 60 and 40%, respectively. In other words, the uncorrected gluconeogenic flux would be 1.6 times higher than the corrected value.

In summary, the present data indicate that infusion of [1-¹³C]acetate results in comparable estimates of TA exchange in healthy subjects and in people with type 2 diabetes both in the overnight-fasted state and during acute hyperglycemia with moderate-dose insulin infusion. We also have demonstrated that asymmetric enrichment of glucose from [U-¹³C]glycerol is also consistent with TA activity and nearly complete TPI exchange. Together with our current and previous observations and those of others, these data provide strong evidence that TA exchange results in a significant fraction of glycogenolytic G6P (15–35%) becoming enriched in position 5 from ²H₂O, resulting in a modest overestimation of fractional gluconeogenesis and underestimation of glycogenolytic contribution to EGP. However, this exchange activity is equivalent for healthy, prediabetic, and T2DM subjects and therefore does not explain differences in gluconeogenic and glycogenolytic fluxes between these subject groups. However, it may contribute to the variability in these flux estimates within each subject group. Thus, the ²H₂O method continues to provide reliable estimates for quantifying gluconeogenesis and glycogenolysis in various physiological and pathophysiological states.

GRANTS

This study was supported by the US Public Health Service (DK-29953, DK-50456, and UL1-TR000135) and the Mayo Clinic and Fundaçãopara a Ciência e a Tecnologia (FCT; PTDC/SAU-MET/11198/2009, RECI/QEQ-QFI/0168/2012). The NMR spectrometers are part of the National NMR Network and were purchased in the framework of the Portuguese National Programme for Scientific Re-equipment, contract REDE/1517/RMN/2005, with funds from POCI 2010 and FCT.

DISCLOSURES

Dr. Rita Basu is the guarantor of this work, had full access to all the data, and takes full responsibility for the integrity of data and the accuracy of data analysis. There are no conflicts of interest, financial or otherwise, to declare for any of the authors.

AUTHOR CONTRIBUTIONS

A.R., S.D., A.B., and R.B. performed the experiments; A.R., S.D., F.C., A.R.S., A.F., J.G.J., and R.B. analyzed the data; A.R., S.D., A.B., J.G.J., and R.B. edited and revised the manuscript; A.R., S.D., A.B., J.G.J., and R.B. approved the final version of the manuscript; S.D., A.B., J.G.J., and R.B. interpreted the results of the experiments; S.D., J.G.J., and R.B. prepared the figures; S.D., J.G.J., and R.B. drafted the manuscript; A.B., J.G.J., and R.B. contributed to the conception and design of the research.