Altered glucose kinetics occurs with aging: a new outlook on metabolic flexibility

Abstract

Our purpose was to determine how age affects metabolic flexibility and underlying glucose kinetics in healthy young and older adults. Therefore, glucose and lactate tracers along with pulmonary gas exchange data were used to determine glucose kinetics and respiratory exchange ratios [RER = carbon dioxide production (V̇co₂)/oxygen consumption (V̇o₂)] during a 2-h 75-g oral glucose tolerance test (OGTT). After an 12-h overnight fast, 28 participants, 15 young (21–35 yr; 7 men and 8 women) and 13 older (60–80 yr; 7 men and 6 women), received venous primed-continuous infusions of [6,6-²H]glucose and [3-¹³C]lactate with a ${\mathrm{H}}^{13}\mathrm{C}{\mathrm{O}}_3^{\mathrm{-}}$ bolus. After a 90-min metabolic stabilization and tracer equilibration period, volunteers underwent an OGTT. Arterialized glucose concentrations ([glucose]) started to rise 15 min post glucose consumption, peaked at 60 min, and remained elevated. As assessed by rates of appearance (Ra) and disposal (Rd) and metabolic clearance rate (MCR), glucose kinetics were suppressed in older compared to young individuals. As well, unlike in young individuals, fractional gluconeogenesis (fGNG) remained elevated in the older population after the oral glucose challenge. Finally, there were no differences in 12-h fasting baseline or peak RER values following an oral glucose challenge in older compared to young men and women, making RER an incomplete measure of metabolic flexibility in the volunteers we evaluated. Our study revealed that glucose kinetics are significantly altered in a healthy aged population after a glucose challenge. Furthermore, those physiological deficits are not detected from changes in RER during an OGTT.

NEW & NOTEWORTHY To determine metabolic flexibility in response to an OGTT, we studied healthy young and older men and women to determine glucose kinetics and changes in RER. Compared to young subjects, glucose kinetics were suppressed in older healthy individuals during an OGTT. Surprisingly, the age-related changes in glucose flux were not reflected in RER measurements; thus, RER measurements do not give a complete view of metabolic flexibility in healthy individuals.

INTRODUCTION

The World Health Organization (WHO) predicts that one in five people will be over 60 yr old by 2050 (1). Increased prevalence of chronic disease with aging places financial or other burdens on the older and general population (2). One of the most significant health threats to the aging population is the onset of type 2 diabetes (T2D), and as of 2020, 27% of the North American and Caribbean population over 65 yr have been diagnosed with diabetes, with the majority of this population being diagnosed with T2D (3).

A precursor to T2D is the loss of metabolic flexibility. In a seminal study using determinations of limb respiratory quotient (RQ), Kelley, Goodpaster, and colleagues (4) found that healthy nonobese subjects could easily switch between fat and carbohydrate (CHO) oxidation when transitioning from fasted to insulin-stimulated states. In contrast, obese nondiabetic subjects did not have the ability to transition between energy substrate utilization, and the investigators originated the term “metabolic inflexibility” to describe the unresponsive condition (5, 6). Since the original findings, numerous studies have linked prediabetic and T2D conditions to the loss of metabolic flexibility, i.e., “inflexibility” (4, 6–10).

Since its inception, the concept of metabolic flexibility has been broadened to include the ability to respond or adapt to conditional changes in metabolic demand such as glucose, insulin, or exercise challenges (11, 12). Furthermore, means to assess metabolic flexibility have evolved with more sophisticated measures of energy flux and energy substrate partitioning derived from isotope tracer technology (13). Additionally, transient pulmonary gas exchange ratios (RERs), but not limb RQ values, are typically measured. Importantly, data are lacking on changes in glucose flux rates that underlie assessments of metabolic flexibility that occur in older men and women.

The regulation of blood glucose concentration ([glucose]) within a tight range (euglycemia) is a key and highly controlled process that underpins metabolic flexibility. Euglycemia is maintained by a combination of physiological and behavioral processes: dietary carbohydrate consumption and endogenous glucose production via hepatic and renal gluconeogenesis (GNG) and hepatic glycogenolysis. Whereas some organs (erythrocytes, brain and nervous tissue) are obligatory glucose consumers, skeletal muscle and liver are the main sites for glucose regulation. Skeletal muscle is the main site for glucose disposal (14, 15), and the liver is the primary regulator of endogenous glucose production (16). As one ages there is a decreased ability to dispose of glucose (17), which may be caused by the inherent age-related loss of skeletal muscle mass (18, 19), and inability to regulate GNG via hepatic insulin insensitivity (20, 21). Consequently, both factors may contribute to the loss of metabolic flexibility with aging.

Although the effects of aging on metabolic systems have been addressed with findings such as increased oxidative stress (22–24) and mitochondrial dysfunction (25, 26), little is known about the effects of healthy aging on metabolic flexibility. Thus, our study aimed to assess metabolic flexibility and glucose regulation in aging by studying metabolic fluxes in healthy young and older men and women before and during an oral glucose tolerance test (OGTT). We hypothesized that in response to a glucose challenge 1) young participants would exhibit greater glucose clearance than older participants, 2) young participants would have decreased endogenous glucose production via gluconeogenesis, and 3) young people would have a higher dynamic range for substrate oxidation via RER measurements than older participants, i.e., be more metabolically flexible. In addition, we look to investigate whether our methods may help diagnose and define metabolic inflexibility in an aging population.

MATERIALS AND METHODS

Overall Design

The overall design was detailed in Leija et al. (27) but is abstracted here for readers’ convenience. Human experimentation was approved by the University of California, Berkeley Committee for the Protection of Human Subjects (CPHS 2018-08-11312) and complies with the standards set by the Declaration of Helsinki. Verbal and written information was provided to prospective volunteers, and verbal and written informed consent was obtained. The study population consisted of 28 participants: 15 young individuals (7 men and 8 women) between the ages of 21 and 35 yr and 13 older individuals (7 men and 6 women) between the ages of 60 and 80 yr. Before being accepted into the study, volunteers were screened for metabolic and cardiovascular diseases. Screening involved a health history questionnaire, 3-day food record, basic metabolic panel, electrocardiogram, pulmonary function assessment, three-site skinfold measurements to assess body composition, and physical examination. For men skinfold sites were chest, abdomen, and thigh, and for women sites were triceps, suprailiac, and thigh. Subsequent to preliminary screening and physical examination, physical fitness was assessed by a continual, progressive leg cycle ergometer test to assess ventilatory threshold (VT) and peak oxygen consumption (V̇o_2peak) (Table 1). The cycle ergometer test entailed a 5-min warmup at 25-W external power output (PO). After the warmup, for the young volunteers we increased PO to 75 W, with subsequent increases of 30 W/min until pedaling rate fell below 50 rpm. For our older participants, after their warmup we increased PO to 50 W and increased it by 15 W/min until their pedaling rate decreased below 50 rpm.

Table 1.

Participant characteristics

Variable	Young	Older
Age, yr*	28.1 ± 1.4	70.6 ± 2.4
Body mass, kg	67.3 ± 5.2	72.6 ± 4.6
Body mass index, kg/m2*	23.5 ± 1.8	26.2 ± 1.3
Body fat, %*	14.9 ± 1.9	23.5 ± 1.4
FEV1/FVC, % P = 0.10	84.3 ± 2.2	79.8 ± 1.7
Absolute aerobic capacity (V̇o2peak), L·min−1*	2.8 ± 0.3	1.9 ± 0.2
Relative aerobic capacity (V̇o2peak), mL·kg−1·min−1*	41.1 ± 3.2	26.1 ± 1.9
Relative aerobic capacity to lean body mass (V̇o2-lbm), mL·kglbm−1·min−1*	48.2 ± 3.1	34.6 ± 2.3
Peak power output, W*	270 ± 20	170 ± 20
VT1, % V̇o2peak*	73 ± 2	61 ± 2
3-day diet records
Energy, kcal/day+	2,570 ± 122	2,277 ± 114
Carbohydrate, %	61 ± 2	60 ± 1
Fat, %	30 ± 1	30 ± 1
Protein, %	9 ± 1	10 ± 2

Values are means ± SE; n = 15 young, 13 older subjects; FEV₁, forced expiratory volume in 1 s; FVC, forced vital capacity; V̇o_2peak, peak oxygen consumption; peak power output; VT1, ventilatory threshold 1. *Significantly different from older individuals (P ≤ 0.05), +trended to be different from older individuals (P ≤ 0.10).

Participants could be classified as healthy and physically active but not athletes in training (12, 27–29). Self-reported 3-day food records were analyzed for caloric intake and macronutrient composition to ensure normal dietary habits (DietAnalysis Plus, version 6.1; ESHA Research, Salem, OR). Only individuals providing evidence of taking standard, balanced diets were entered into the study. For example, individuals following high-carbohydrate or ketogenic diets were excluded. Participants included in the study had a body mass index in the range of ≥18.5 and <30 kg/m², were nonsmokers, had a forced expiratory volume in 1 s (FEV₁)/forced vital capacity (FVC) of >70%, were diet and weight stable, and had fasting blood glucose levels <100 mg/dL and HBA1c levels <5.8%. Participants were excluded from the study if they were pregnant or had any preexisting metabolic conditions such as type 1 or type 2 diabetes or prediabetes or if they were deemed unfit to perform the graded exercise test. Volunteers were cleared for participation by a licensed physician. Screening preceded testing by at least 1 wk. In the case of young premenopausal women, testing occurred during the early follicular phase of their menstrual/uterine cycle (days 3–5 post menses).

Tracers

We administered primed-continuous infusions (CIs) of stable isotopically labeled glucose and lactate with a [¹³C]bicarbonate bolus (28). Specifically, [6,6-²H]glucose (D2-glucose, M + 2 signal) (labels lost in glycolysis) was infused to determine total glucose turnover. The blood glucose pool was primed with a 250-mg D2-glucose bolus and continuously infused at a rate of 2 mg/min. [3-¹³C]lactate (label lost in the tricarboxylic acid (TCA) cycle or recycled to glucose in the liver and kidneys) was infused to estimate gluconeogenesis (via Cori cycle, i.e.,[¹³C]glucose from [3-¹³C]lactate yields a M + 1 signal in the glucose isotopomer from mass spectrometry). For lactate, the priming bolus was 57.5 mg of [3-¹³C]lactate and the continuous infusion rate was 2.5 mg/min. As well, 136 mg of [¹³C]bicarbonate ($\mathrm{HC}{\mathrm{O}}_3^{\mathrm{-}}$) was infused to prime the bicarbonate pool. ¹³C-tracers were given as Na⁺ salts. Isotope tracer cocktails, including boluses and continuous infusates, were prepared by Mariner Advanced Pharmacy and Compounding Company (San Mateo, CA).

Procedures

The day before the OGTT, volunteers were asked to record and maintain their standard dietary pattern and refrain from strenuous physical exercise. After an overnight fast, participants reported to the laboratory at 5:00 am. A hand vein was catheterized for blood sampling; a heating pad was applied to obtain “arterialized” blood samples (30), and a saline drip was established to keep the blood sampling line open. The contralateral forearm or hand vein was then catheterized for tracer infusion.

After setup and participant catheterization, background breath sampling for oxygen consumption (V̇o₂) and carbon dioxide production (V̇co₂) occurred to determine RER (= V̇co₂/V̇o₂). Furthermore, arterialized blood was collected for determinations of background endogenous isotopic enrichments (IEs) as well as insulin and counterregulatory hormone levels. Subsequently, priming boluses of D2-glucose, [3-¹³C]lactate, and ${\mathrm{H}}^{13}\mathrm{C}{\mathrm{O}}_3^{\mathrm{-}}$ were given, and CIs of D2-glucose and [3-¹³C]lactate commenced. A 90-min isotope equilibration period occurred with simultaneous arterialized blood and breath samples collected at 75 and 90 min of CI as done previously (31–33).

OGTT

After 90 min of resting CI, the participants consumed a 296-mL solution containing 75 g of D-glucose (Azer Scientific, catalog no. 10-0-75) within 1–2 min. After the drink, a 2-h timer was initiated, and arterialized blood and expired air were taken for determinations of isotopic enrichments and glucoregulatory hormones at 5, 15, 30, 60, 90, and 120 min after consumption of the glucose drink.

Determinations of Isotopic Enrichments

Isotopic enrichments of blood metabolites were analyzed as previously described (28, 31–35). For glucose analyses, 200 μL of whole blood was placed in 400 μL of 200-proof ethanol and spun at 10,000 g for 1 min. A liquid-to-liquid ethanol extraction was performed, and the organic layer containing glucose was extracted, transferred to a 2-mL glass vial, and dried under nitrogen (N₂) gas. The dried glucose was derivatized with 100 μL of a 2:1 mixture of acetic anhydride and pyridine. The glass vial containing the derivatized glucose was sealed and heated at 60°C for 20 min. After 20 min, the sample was dried under N₂ and then reconstituted in 100–500 μL of ethyl acetate.

Isotopic enrichments (IEs) of glucose penta-acetate derivative were measured on an Agilent 6890/5973 gas chromatograph-mass spectrometer (GC-MS) utilizing positive chemical ionization (PCI) and selected ion monitoring (SIM) using a DB-17 GC column. The initial oven temperature was set to 110°C and increased by 35°C every minute until 280°C was reached and held for 5 min. The mass-to-charge ratios (m/z) 331 (nonlabeled glucose), 332 (M + 1 isotopomer, [1-¹³C]glucose), and 333 (M + 2 isotopomer, D₂-glucose) were monitored for the glucose penta-acetate derivative. Selected ion abundances were compared against external standard curves to calculate mole percent excess isotopic enrichments. Data on arterialized lactate and glucose fluxes in young subjects have been reported (27). Blood lactate concentration and kinetics data on young and older individuals will be reported separately (J. A. Arevalo, R. G. Leija, A. D. Osmond, C. C. Curl, J. J. Duong, M. J. Huie, U. Masharani, and G. A. Brooks, unpublished observations).

Determination of Whole Blood Glucose Concentration

Blood samples for the determination of whole blood glucose concentrations were immediately deproteinized in 7% perchloric acid. Each sample was neutralized with 2 N NaOH, and concentrations were subsequently determined enzymatically via hexokinase and glucose-6-phosphate dehydrogenase (36).

Calculations

Glucose flux rates, i.e., rates of appearance (Ra, mg·kg⁻¹·min⁻¹) and disposal (Rd, mg·kg⁻¹·min⁻¹) and metabolic clearance rate (MCR, mL·kg⁻¹·min⁻¹), were calculated from the equations of Steele modified for use with stable isotopes (37):

$$\text{Ra}=\frac{F-\mathrm{V}\left(\frac{{\mathrm{C}}_1+{\mathrm{C}}_2}{2}\right)\left(\frac{\mathrm{I}{\mathrm{E}}_2-\mathrm{I}{\mathrm{E}}_1}{t_2-t_1}\right)}{\left(\frac{\mathrm{I}{\mathrm{E}}_1+\mathrm{I}{\mathrm{E}}_2}{2}\right)}$$ (1)

$$\text{Rd}=\mathrm{Ra}-\mathrm{V}\left(\frac{{\mathrm{C}}_2-{\mathrm{C}}_1}{t_2-t_1}\right)$$ (2)

$$\text{MCR}=\frac{\mathrm{Rd}}{\frac{({\mathrm{C}}_1+{\mathrm{C}}_2)}{2}}$$ (3)

where F represents isotope infusion rate (D2-glucose; mg·kg⁻¹·min⁻¹); V is the volume of distribution for glucose (180 mL·kg⁻¹); C₁ and C₂ are concentrations (mg·L⁻¹) at sampling times t₁ and t₂, respectively; and IE₁ and IE₂ are the excess isotopic enrichments of glucose at these sampling times.

The rate of lactate conversion to glucose, gluconeogenesis (GNG) in milligrams per kilogram per minute, was calculated as previously described (38), as derived from Zilversmit et al. (39):

$$\text{GNG}=\frac{\left[\mathrm{Lactate}\hspace{0.17em}\mathrm{Ra}\hspace{0.17em}\times \hspace{0.17em}\mathrm{I}{\mathrm{E}}_{\mathrm{Glucose}}\right]}{\mathrm{I}{\mathrm{E}}_{\mathrm{Lactate}}}\hspace{0.17em}\times \hspace{0.17em}\mathrm{H}$$ (4)

where lactate Ra is the rate of appearance of lactate (J.A. Arevalo, R.G. Leija, A.D. Osmond, C.C. Curl, J.J. Duong, M.J. Huie, U. Masharani, G.A. Brooks, unpublished observations); IE_Glucose is the isotopic enrichment of the M + 1 glucose isotopomer; IE_Lactate is the isotopic enrichment of lactate; and H is the Hetenyi factor to correct for loss of label in the tricarboxylic acid cycle during GNG [1.45 before and 1.0 during the OGTT, respectively (31, 40)]. The fraction of glucose Ra from GNG, or fractional gluconeogenesis (fGNG), is calculated below and was adapted from previous research (38):

$$\text{fGNG}=\frac{\mathrm{GNG}}{\mathrm{Glucose}\hspace{0.17em}\mathrm{Ra}}\hspace{0.17em}\times \hspace{0.17em}100$$ (5)

Estimation of Hepatic Glycogenolysis

Estimations were based on the report of Stender et al. (41), who, using magnetic resonance spectroscopy (MRS) technology, showed that most of an oral glucose load was sequestered and held by the liver as glycogen, with glucose release commencing 30 min after the glucose challenge. Consequently, our assumption was that hepatic glycogenolysis was the main contributor to glucose Ra commencing 30 min post glucose consumption.

$$\text{Hepatic\hspace{0.17em}glycogenolysis}=\mathrm{Glucose}\hspace{0.17em}\mathrm{Ra}\hspace{0.17em}\left(\mathrm{grams}\right)-\mathrm{GNG}\hspace{0.17em}(\mathrm{grams})$$ (6)

where Glucose Ra is the total amount of glucose that was produced in grams from 30 min until the end of the study and GNG is the total amount of glucose that came from gluconeogenesis in grams from 30 min onward.

Determinations of Insulin and Glucoregulatory Hormones

Blood samples for the determination of plasma hormone concentrations were collected in tubes containing EDTA and protease inhibitors (aprotinin and a dipeptidyl peptidase-4 (DPP-4) inhibitor]. Concentrations of plasma insulin and glucagon were determined with commercially available enzyme-linked immunosorbent assay (ELISA) kits (ALPCO, Salem, NH).

Determination of Metabolic Flexibility

Parameters of respiratory gas exchange were determined with a Rudolph breathing valve and a Parvo Medics TrueOne 2400 metabolic cart (Salt Lake City, UT). Metabolic flexibility was defined as the difference between peak and baseline RER (ΔRER), utilizing breath samples taken during the last 5 min of each 10-min sampling period. RER was determined using 5-s averages of V̇o₂ and V̇co₂ during the aforementioned 5-min period.

$$\text{Metabolic\hspace{0.17em}flexibility}=\Delta \text{RER}=\mathrm{RE}{\mathrm{R}}_{\mathrm{Peak}}-\hspace{0.17em}\mathrm{RE}{\mathrm{R}}_{\mathrm{Baseline}}$$ (7)

Because of technical difficulties with the collection of expired air and metabolic cart functioning, we report metabolic flexibility data on 13 young and 13 older individuals.

Statistical Analyses

We used GraphPad Prism 3.0 software (GraphPad Software Inc., San Diego, CA) for assessments of statistical significance. A mixed-model repeated-measures ANOVA (time × group) was used to detect mean differences for glucose concentrations, Ra, Rd, MCR, rate of GNG, fGNG, and RER. t Tests were used to detect mean differences for the descriptive statistics (age, height, weight, body fat %, lean body mass %, aerobic capacity, etc.). Statistical significance was set at α = 0.05, and Tukey’s post hoc tests were used to make pairwise comparisons within the repeated measures; values are represented as means ± SE.

RESULTS

Subject Characteristics

Compared to the young population, older individuals had a reduced aerobic capacity [V̇o_2peak and V̇o_2peak relative to lean body mass (V̇o_2peak-lbm); P ≤ 0.001] and diminished muscular power output (P ≤ 0.001) (Table 1). Furthermore, the older subjects had a higher body fat percentage (P ≤ 0.001). As determined from data on days of screening and experimentation, participants were weight stable. Additionally, the total energy and macronutrient composition of the young and older individuals’ diets were not different and did not change between screening and the day before OGTT (Table 1).

Blood Glucose Concentrations Before and During an OGTT

Baseline, 12-h fasting arterial glucose concentrations were similar in young and older cohorts. After a 15-min lag, blood [glucose] rose from baseline values following the oral glucose challenge (Fig. 1). Blood [glucose] peaked after 60 min for both older and young participants (P ≤ 0.001) and remained elevated for the remainder of the trial (P ≤ 0.001), while [glucose] remained elevated for the remainder of the trial until 90 min, when the younger individuals decreased blood glucose concentrations compared to older individuals (P ≤ 0.05). The [glucose] area under the curve was not significantly different between the young (16,016 ± 1,096 mg/dL × min, 95% confidence interval 13,868 to 18,165 mg/dL × 120 min) and older (18,378 ± 1,275 mg/dL × min, 95% confidence interval 15,879 to 20,878 mg/dL × 120 min; P > 0.05) participants.

Figure 1.

Changes in arterialized blood glucose concentration ([glucose]) over the course of an oral glucose tolerance test (OGTT) for young (solid line, n = 15) and older (dashed line, n = 13) subjects. Values are means ± SE. #Significantly different from baseline values (P ≤ 0.05), $significantly different from peak glucose concentration (P ≤ 0.05), *significantly different from older individuals (P ≤ 0.05), +trended to be different from older individuals (P ≤ 0.10). Sampling time 0 represents an average of values determined 75 and 90 min after the initiation of isotope tracer infusion. The 5 min and subsequent sampling times represent the time after consumption of 75 g of D-glucose. Arterialized [glucose] increases more in older than young adults during an OGTT.

Additionally, for the older participants, the 2-h blood glucose (144.9 ± 9.94 mg/dL) was slightly above what the ADA classifies as “prediabetic” (>140 mg/dL) after an OGTT. In this regard it is important to note that we report values of arterialized, not venous, blood. Importantly, in a previous study on young men arterialized blood values were 15 mg/dL higher than in simultaneously sampled venous blood 120 min after glucose consumption (42). Hence, the older persons in our study should not be considered to be prediabetic.

Glucose Kinetics Before and During a Fasting OGTT

Although fasting arterial glucose Ra values were similar between age groups, we found several time- and age-related changes in blood glucose kinetics during the OGTT. There was a significant interaction between aging and glucose Ra (F = 4.082, P ≤ 0.05), with main effects for both time (F = 83.091, P ≤ 0.001), and age (F = 4.307, P ≤ 0.05). Figure 2 shows a linear increase in glucose Ra that commenced soon after ingestion of the drink for the young population, reaching significance after 15 min (P ≤ 0.10, 5 min; P ≤ 0.05 from 15 to 120 min). In the older population there was a 30-min lag before glucose Ra rose (P ≤ 0.05 from 30 to 120 min post glucose challenge). Additionally, after 60 min glucose Ra tended to be higher in the young compared to older participants (P = 0.07), reaching significance 90 min and onward post glucose consumption (P ≤ 0.05). Noteworthy was that glucose Ra did not follow the same pattern as [glucose] (Fig. 1 and Fig. 2).

Figure 2.

Glucose rate of appearance (Ra) before and after an oral glucose tolerance test (OGTT) for young (solid line, n = 15) and older (dashed line, n = 13) subjects. Values are means ± SE. #Significantly different from baseline (P ≤ 0.05), *significantly different from older individuals (P ≤ 0.05), ^trended to be different from baseline (P ≤ 0.10), +trended to be different from older individuals (P ≤ 0.10). Glucose Ra is suppressed in aging during an OGTT.

We found a significant interaction between aging and glucose Rd (F = 4.097, P ≤ 0.05), with main effects for both time (F = 120.79, P ≤ 0.001) and age (F = 5.415, P ≤ 0.05). Figure 3 shows a linear increase of Rd that commenced soon after ingestion of the glucose drink for the young population, reaching significance after 15 min (P ≤ 0.10, 5 min; P ≤ 0.05 from 15 to 120 min), whereas there was a 30-min lag before glucose Rd rose in the older group (P ≤ 0.05 from 30 to 120 min post glucose challenge). Furthermore, within 15 min of glucose ingestion, Rd was significantly higher in the young (P ≤ 0.05) compared to older individuals. From 15 to 30 min, there was no difference in Rd, yet from 60 min onward young individuals had a higher glucose Rd, reaching statistical significance after 90 min (P ≤ 0.05).

Figure 3.

Glucose rate of disappearance (Rd) before and after an oral glucose tolerance test (OGTT) for young (solid line, n = 15) and older (dashed line, n = 13) subjects. Values are means ± SE. #Significantly different from baseline (P ≤ 0.05), *significantly different from older individuals (P ≤ 0.05), ^trended to be different from baseline (P ≤ 0.10), +trended to be different from older individuals (P ≤ 0.10). Glucose Rd is suppressed in aging during an OGTT.

Whereas measures of glucose Ra and Rd demonstrated linear increases over time in all study participants after the glucose challenge, glucose clearance (MCR) demonstrated a multiphasic response (Fig. 4). Specifically, there was a significant interaction between aging and MCR (F = 3.815, P ≤ 0.05), with main effects for both time (F = 22.645, P ≤ 0.001) and age (F = 6.352, P ≤ 0.05). For the first 5 min after the glucose challenge, young individuals had an immediate increase in glucose clearance over baseline, followed by a significant decrease after 15 min, which was followed by a continuous rise over the last 60 min of the trial. Unlike the young individuals, glucose clearance in the older group decreased for the first 60 min after the glucose challenge. After this diminished clearance rate in older individuals, MCR returned to baseline by the end of the experimental trial. Interestingly, unlike the glucose Ra and Rd, except for the 30 min sampling time, the young group cleared glucose at a higher rate at almost every sampling time than did older individuals (P ≤ 0.05).

Figure 4.

Glucose metabolic clearance rate (MCR) before and after an oral glucose tolerance test (OGTT) for young (solid line, n = 15) and older (dashed line, n = 13) subjects. Values are means ± SE. #Significantly different from baseline (P ≤ 0.05), *significantly different from older individuals (P ≤ 0.05), ^trended to be different from baseline (P < 0.10). Glucose MCR is suppressed in aging during an OGTT.

Gluconeogenesis and Hepatic Glycogenolysis before and during an OGTT

We examined the effects that a glucose challenge had on gluconeogenesis as determined by the conversion from lactate into glucose, the main gluconeogenic precursor (43, 44). Twelve-hour fasting GNG rates were higher in young compared to older adults (P ≤ 0.05), yet GNG rates were relatively unchanged in the young group post challenge (Fig. 5; P ≤ 0.05 at 60 min). Inversely, GNG rates rose in older individuals post glucose challenge (P < 0.05, at 60 and 90 min) to meet those in the young.

Figure 5.

The rate of lactate conversion to glucose (GNG) before and after an oral glucose tolerance test (OGTT) for young (solid line, n = 13) and older (dashed line, n = 13) subjects. Values are means ± SE. #Significantly different from baseline (P ≤ 0.05), ^trended to be different from baseline (P ≤ 0.10). Fasting lactate conversion to glucose before a glucose challenge is suppressed in aging. *Significantly different from older individuals.

To further explore the effect of a glucose challenge, we determined the fraction of GNG (fGNG) that contributed to total glucose Ra. Although there was not a significant interaction between age and fGNG, there was a significant decrease in fGNG over time (F = 2.926, P ≤ 0.05), with no main effect of age. The young population demonstrated a steady decrease in fGNG from baseline to the end of the trial, reaching significance from 90 min post challenge onward (P ≤ 0.05; Fig. 6). Inversely, after the oral glucose challenge fGNG persisted in the older population such that young persons had lower fGNG 90 min after the glucose challenge (P ≤ 0.05).

Figure 6.

Fractional gluconeogenesis (fGNG; %) before and after an oral glucose tolerance test (OGTT) for young (solid bar, n = 13) and older (dashed bar, n = 13) subjects. Values are means ± SE. #Significantly different from baseline (P ≤ 0.05), *significantly different from older individuals (P ≤ 0.05), +trended to be different from older individuals (P ≤ 0.10). Elevated fGNG persists during an OGTT in older persons.

Because there were differences in glucose Ra and fasting gluconeogenesis between young and older participants, we decided to investigate how aging affected hepatic glycogenolysis post glucose consumption. We found that there was a significant increase in glycogenolysis in the younger (23 ± 7 g) compared to older (17 ± 7 g) individuals during the course on an OGTT (P ≤ 0.05).

Glucoregulatory Hormonal Response during an OGTT

The alterations in glucoregulatory hormonal response will be reported in detail elsewhere (A.D. Osmond, R.G. Leija, J.A. Arevalo, C.C. Curl, J.J. Duong, M.J. Huie, U. Masharanai, G.A. Brooks, unpublished observations) but for the reader’s convenience salient points are abstracted here. Baseline plasma insulin levels in the young population (23.8 ± 0.7 pM) were significantly lower than those in the older population (34.5 ± 1.1 pM; P ≤ 0.05). In young individuals insulin peaked at 60 min (406.5 ± 15.0 pM), whereas the older group did not peak until the end of the trial (299.1 ± 8.6 pM). After peaking at 60 min, plasma insulin levels in the young individuals started to decline, but values did not return to baseline levels by the end of the test (347.9 ± 11.5 pM).

Baseline glucagon levels were similar in the young (9.8 ± 0.2 pM) and older (6.5 ± 0.2 pM) individuals. After the ingestion of the glucose load, both young and older participants’ glucagon levels continued to fall until the end of the trial at 120 min (young 3.7 ± 0.2 pM, older 3.1 ± 0.1 pM).

Metabolic Flexibility as Assessed by RER During an OGTT

Fasting RER values were similar in young and older individuals. Overall, there was a significant increase in RER 60 min post glucose ingestion in both groups, which remained elevated for the remainder of the study (Fig. 7). RER increased over time (F = 66.76, P ≤ 0.001) and was higher in the younger compared to older individuals 60 min post glucose consumption (P ≤ 0.05). Although we observed no differences in RER_baseline and RER_peak between age groups (Table 2), in healthy young and older individuals there were no differences in metabolic flexibility as classically defined.

Figure 7.

Respiratory exchange ratio (RER) before and after an oral glucose tolerance test (OGTT) for young (solid line, n = 13) and older (dashed line, n = 13) subjects. Values are means ± SE. *Significantly different from older individuals (P ≤ 0.05), #significantly different from baseline (P ≤ 0.05). No differences in RER-measured energy substrate partitioning between young and older individuals during fasting and after an oral glucose challenge.

Table 2.

Pulmonary gas exchange measures of metabolic flexibility

Variable	Young	Older
RERbaseline	0.82 ± 0.02	0.82 ± 0.02
RERpeak	0.95 ± 0.03	0.95 ± 0.05
Metabolic flexibility	0.13 ± 0.05	0.13 ± 0.04

Values are means ± SE; n = 13 young, 13 older persons. RER_baseline, baseline respiratory exchange ratio (RER); RER_peak, peak RER. Metabolic flexibility = ΔRER = RER_peak − RER_baseline. No age- or sex related significant differences identified.

DISCUSSION

We assessed the effects of aging on metabolic flexibility following an oral glucose challenge. For that purpose, we employed a standard OGTT procedure with isotope tracer technology and pulmonary respiratory assessments. In brief, we report the following findings: 1) pulmonary RER provides an incomplete view of metabolic flexibility, 2) assessing glucose clearance and fGNG may be better alternatives to RER for evaluating metabolic flexibility, 3) immediate glucose clearance is the precursor for the enteric phase of the postprandial lactate shuttle (PLS) (27), and 4) the continuous rise in glucose 30 min after an oral glucose challenge is a component of the systemic phase of the PLS. These findings are discussed sequentially.

The Classic, RER-Based Definition of Metabolic Flexibility Might Not Be Optimal

Metabolic flexibility can be a useful tool for characterizing an individual’s capacity for energy substrate partitioning. Kelley and colleagues (4) showed that metabolically inflexible individuals may have passed an OGTT with normal glucose levels, yet leg RQ determinations were unresponsive to insulin stimulation. Building on those seminal findings, many studies of metabolic flexibility have utilized controlled hyperinsulinemic-euglycemic clamp technique to examine glucose kinetics in conjunction with assessments of pulmonary RER as indicators of metabolic flexibility (4, 7, 9, 10). However, studies of metabolic flexibility have not often compared results obtained from healthy young and older persons (45). Extant studies offer age group comparisons with some type of metabolic deficit, such as obesity, diabetes, or combination of the two (4, 7, 9, 10, 46). Thus, the effect of aging on metabolic flexibility in healthy individuals is difficult to parse from results of past research.

We assessed metabolic flexibility utilizing an OGTT rather than a hyperinsulinemic-euglycemic clamp (clamp). Clamp and OGTT procedures each have advantages and proponents. Whereas the clamp procedure allows for tight regulation of blood [glucose] via titration of insulin, the OGTT is less controlled but better mimics a real-life situation that relies on an individual’s ability to produce insulin for clearing an exogenous, not vascularly infused glucose load. Healthy study participants undergoing clamp procedures typically reach an RER around 1.00, whereas metabolically inflexible individuals are incapable of reaching an RER of 1.00 (4, 7, 9, 10). In our study healthy participants reached a peak RER of 0.94, not ≥1.0, thus making the measurement more reflective of an endogenous, mixed substrate metabolic response. Most relevant to the present study, during a hyperinsulinemic-euglycemic clamp procedure glucose is infused intravenously, thus bypassing the gut; this is a problem for studying phenomena related to postprandial lactate shuttling, particularly the first-pass enteric PLS phase.

The present study utilized an OGTT, a standard clinical diagnostic tool in clinical practice. For inclusion in our study, we recruited healthy, physically active participants into both young and older age groups. Surprisingly, as classically assessed by determinations of RER we found no differences in metabolic flexibility between the older and young participants. However, our results did reveal that healthy older participants demonstrated reduced glucose clearance and disposal rates compared to the young individuals. Using euglycemic-hyperinsulinemic clamp technology, Meex and colleagues (10) observed that metabolically inflexible individuals with T2D had suppressed glucose Rd compared to healthy individuals. Although based on RER the older participants in our study did not have altered metabolic flexibility compared to the younger cohort, we did show that older participants had suppressed glucose Rd and MCR during the OGTT. The latter results lead us to suspect that the individuals may be metabolically inflexible. Furthermore, related to our results is that Basu et al. (17) and Jackson et al. (47) have also provided evidence that healthy older individuals have reduced glucose Rd during an OGTT compared to a younger cohort. Again, because RER was not measured, investigators were unable to assess metabolic flexibility using measures of pulmonary gas exchange. Therefore, measures of glucose Rd and MCR could be signs of compromised glucose clearance in older subjects that precede changes in ΔRER.

Although aging did not alter metabolic flexibility as determined from delta RER, an explanation for the decreased glucose Rd and MCR in older individuals may stem from age-related reduction in lean body mass (18). Although our older volunteers were determined to be healthy and physically active, there was an age-predicted change in body mass distribution, with our older individuals having a higher percent body fat and decreased percent lean body mass (Table 1). Because skeletal muscle is the primary binding site for insulin (15) and glucose disposal (14), decreases in lean body mass could hamper one’s ability to clear glucose. Along with others, we showed that aging decreased glucose Ra and Rd accompanied by lower lean body mass percentage (17, 47). Hence, even after correcting for lean body mass, we showed that glucose kinetics were impacted by aging.

With aging there is also a notable loss of endurance capacity (48, 49). In our study aerobic capacity and peak power output were 63% lower in our older individuals, and these deficits were still evident when correcting for lean body mass. Previously, we demonstrated that endurance training increased glucose Rd and MCR during the same relative metabolic challenge (i.e., exercise bouts) in longitudinal training studies that included both young (31) and older (50) participants. Exercise training has been shown to increase mitochondrial density (51) and glucose (52–54) and lactate (55) transporter abundances; thus, the decreased glucose Rd and MCR observed in our older adults may, at least partly, be attributed to the loss of mitochondrial density and solute transporters yielding lower endurance capacities compared to those of young subjects.

Fractional Gluconeogenesis as a Marker for Metabolic Health in Older Adults

Gluconeogenesis, in both fasted and fed states, has significant implications for overall health (35, 56–59) and may provide insight into hepatic insulin insensitivity (20). Previously we have shown that exercise training increases fasting fGNG in men (31), indicating increased hepatic insulin sensitivity. In contrast, T2D patients show lower GNG during fasting and increased GNG under hyperglycemic conditions (60, 61). There was an increased fasting GNG rate and a notable elevation of fasting fGNG in younger compared to older individuals, which may be attributed to elevated endurance capacity in our younger subjects. We also observed a significant decrease in fGNG from baseline by the end of the trial in our young volunteers, which was not evident in the older cohort. Although insulin and glucagon levels were mostly similar between the two groups during the trial, our older individuals exhibited elevated gluconeogenic rates and increased relative contribution of GNG to glucose Ra, fGNG, compared to younger participants during the OGTT. This could be a sign of compromised hepatic insulin sensitivity in our older adults. Importantly, in our aged cohort, the larger contribution of GNG to glucose Ra, fGNG, may be an indicator of metabolic inflexibility, which was also not detected with RER measurements.

Immediate Glucose Clearance during an OGTT, a Pathway for the Enteral Phase of the Postprandial Lactate Shuttle?

In the first 5 min post glucose consumption, we observed an increased glucose clearance rate with a simultaneous increased lactate appearance in young adults (27). However, those changes were not observed in older individuals (J. A. Arevalo, R. G. Leija, A. D. Osmond, C. C. Curl, J. J. Duong, M. J. Huie, U. Masharani, and G. A. Brooks, unpublished observations). Interestingly, in young adults, during the OGTT the rise in glucose MCR and lactate Ra occurs before the spike in arterial [glucose] that occurred 15 min after glucose consumption. A possible explanation for this is that during the first pass of circulation through the gut and liver there is an immediate conversion of glucose to lactate in the gut as hypothesized by the enteral lactate shuttle (27). Previously, in experiments where glucose was directly infused intestinally into rats, lactate concentrations rose immediately throughout the lumen and small intestine (62), thus providing evidence that glucose is converted into lactate by the lumen and intestinal wall. Additionally, after a gastric glucose load to rats implanted with a portal vein catheter, a porto-peripheral lactate gradient occurred (63). Although in the present study we could not measure glucose or lactate Ra at the portal vein, we did observe an immediate increase in glucose MCR with a simultaneous rise in lactate Ra, without a rise in glucose concentrations or changes in substrate utilization via RER in the younger participants. The initial increase in glucose MCR without an increase in glucose Ra is consistent with results of Stender et al. (41), who observed that most glucose in an OGTT is immediately sequestered by the liver.

And finally, it has not escaped our notice that the shapes of the plots in Fig. 1 ([glucose]) and Fig. 2 (glucose Ra) are dissimilar. We take the results to mean that for maintaining euglycemia the liver plays major roles in both enteric and systemic PLS phases. Activity of the PLS buffers the effect of an oral sugar load on blood [glucose] by converting glucose to lactate. This can be seen by the rapid rise in arterial [lactate] and Ra (27) compared to the delayed rise in [glucose] and Ra (Fig. 1 and Fig. 2). As explained above, during the first, enteral phase of the PLS glycolysis in the gut buffers the entry of glucose entering the systemic circulation. Also, sequestration of the glucose load by the liver limits rapid entry of glucose into the systemic circulation. During the second and more persistent PLS phase, hepatic glucose release (41) and glucose conversion to lactate (27) modulate glucose Ra in a manner to maintain euglycemia (Fig. 1). The diminished glycogenolysis in the older subjects could possibly be attributed to compromised glucoregulation by the liver and may indicate metabolic inflexibility.

Limitations

In this study, we focused on the effects of an oral glucose load on the parameters of glucose flux to assess effects of age on metabolic flexibility. For this purpose, tracers were introduced into the systemic circulation via venous catheters, and “arterialized” blood samples were collected from a contralateral, warmed hand vein. Previous studies using an OGTT have shown glucose Ra and Rd to peak ∼60 min after glucose ingestion and then level off and remain stable for the remainder of the trial (64, 65). In contrast, our results showed that after a lag there occurs a continuous rise in glucose Ra for the remainder of the 120-min period observed. A possible explanation for differences between present and extant differences is the use of venous versus arterialized sampling sites, with venous blood giving lower overall concentrations resulting from the influences of various tissue compartments (42, 65).

In the present study, we reported total glucose kinetics after ingestion of 75 g of a standardized, non-isotopically labeled OGTT glucose drink widely used experimentally and clinically. In retrospect, we could have used a high-carbohydrate mixed meal with a spiked [¹³C]glucose tracer in conjunction with an intravenous primed-continuous infusion of D2-glucose to measure both exogenous and endogenous glucose fluxes as previously done (17, 66). Had we used a [¹³C]glucose tracer, we would not have been able to simultaneously utilize a [¹³C]lactate tracer to examine the effects of an OGTT glucose load on fGNG. This is because the ¹³C signals from both glucose and lactate would mix into the same carbon pool, consequentially leading to uninterpretable blood and breath results. Importantly, by measuring gluconeogenesis, we captured changes in endogenous glucose production during an OGTT as seen in previous research (64, 65). Moreover, because of our study design we were able to show that older subjects were unable to suppress fGNG during an oral glucose challenge; that observation would not have been made without use of the [3-¹³C]lactate tracer. As mentioned previously (vide supra), GNG is a crucial measurement for metabolic health.

Furthermore, we could not determine the absolute peak in glucose flux or oxidation rates after ingestion of a glucose load because we did not detect a diminution following peak glucose Ra, Rd, and MCR by the end of the trial. We could have extended our study like others to include 4 to even 7 h of observations post glucose ingestion (17, 47), but as we were attempting to detect clinically relevant data, we used the standard 2-h sampling time implemented during a 75-g oral glucose tolerance test. Thus, we deemed 2 h as an appropriate end to trials. And finally, because we did not infuse a [¹³C]glucose tracer, our observation of increased fGNG in older individuals might have resulted from decreased capacities for oxidative glucose disposal.

Importantly, our data and interpretations are consistent with those of Stender et al. (41), who showed that rapid hepatic sequestration and conversion of glucose to glycogen accounted for “most” of the glucose load during the first 25–30 min of an OGTT. In Leija et al. (27) we estimated short-term (30 min) hepatic glucose sequestration to account for 82% of the glucose load. This was followed by hepatic glucose release and disposal by various means including conversion of released glucose to lactate during the subsequent 90 min of OGTTs.

And finally, because pulmonary gas exchange has been widely used to assess metabolic flexibility we admit to being disappointed by an inability to show clear differences in RER between healthy young and older individuals following an oral glucose challenge. Several factors need to be considered. First, that we did not find differences in RER in postchallenge study participants might be attributable to the absence of metabolic diseases in the older cohorts. Hence the results may be appropriate, thus preserving some viability for clinical use of pulmonary gas exchange to detect metabolic inflexibility. Second, others might be cleverer in using RER in combination with other technologies such as fatty acid or glycerol isotopic tracers or other tools to assess metabolic flexibility in research and clinical settings. And third, it may be that RQ and RER methodologies may variably, and independently, suffer from technical difficulties in measuring CO₂ contents of arterial and venous blood as well as expired air during nonsteady metabolic states such as hyperinsulinemic-euglycemic clamp, OGTT, and exercise studies.

Conclusions

Compared to young adults, healthy older individuals exhibit suppressed glucose disposal and clearance after consumption of a 75-g oral glucose challenge. Additionally, the inability of older individuals to limit gluconeogenesis from lactate following an oral glucose challenge is a component of decreased metabolic flexibility in aging. Importantly, the age-related changes in glucose flux that we observed in older individuals were not reflected in changes in pulmonary RER or arterial glucose concentrations.

DATA AVAILABILITY

Data will be made available upon reasonable request.

GRANTS

This work was supported by NIH Grant R01 AG059715 to G.A.B. and a grant from the UC, Berkeley Center for Research and Education on Aging (CREA).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

G.A.B. conceived and designed research; C.C.C., R.G.L., J.A.A., A.D.O., J.J.D., M.J.H., and G.A.B. performed experiments; C.C.C., R.G.L., J.A.A., A.D.O., M.A.H., and G.A.B. analyzed data; C.C.C., R.G.L., J.A.A., A.D.O., J.J.D., M.J.H., U.M., and G.A.B. interpreted results of experiments; C.C.C. and R.G.L. prepared figures; C.C.C., R.G.L., and G.A.B. drafted manuscript; C.C.C., R.G.L., J.A.A., A.D.O., J.J.D., M.J.H., U.M., M.A.H., and G.A.B. edited and revised manuscript; C.C.C., R.G.L., J.A.A., A.D.O., M.J.H., U.M., M.A.H., and G.A.B. approved final version of manuscript.