Relative Contribution of Intramyocellular Lipid to Whole-Body Fat Oxidation Is Reduced With Age but Subsarcolemmal Lipid Accumulation and Insulin Resistance Are Only Associated With Overweight Individuals Free

The global prevalence of type 2 diabetes is most apparent in older people (1), and it is estimated that the number of people over 65 years of age with diabetes will have increased 4.5-fold by 2050 (2). Gaining mechanistic insight into age-related insulin resistance and strategies to improve insulin sensitivity with age are clearly warranted. Although aging is associated with insulin resistance, age per se does not appear to cause insulin resistance (3–5). Several factors that likely contribute to age-related insulin resistance include increased abdominal adiposity and reduced physical activity (3,4), along with declines in muscle mass (6,7). Of note, intramyocellular lipid (IMCL) accumulates with age, particularly in subsarcolemmal (SSL) regions (8), and has been strongly associated with insulin resistance (9–12). Indeed, SSL lipid accumulation has been linked to the accumulation of metabolites, such as diacylglycerol (DAG) and ceramide, thought by some (13–15), but not others (16), to contribute to impaired insulin-stimulated muscle glucose uptake. Nevertheless, it remains contentious as to which factors associated with age influence IMCL accumulation.

The accumulation of IMCL and associated metabolites likely results from an imbalance between muscle lipid delivery and oxidation. Indeed, studies have demonstrated reduced free fatty acid (FFA) oxidation in older people compared with young, despite whole-body lipolysis and plasma FFA availability being greater at rest and during exercise at the same absolute and relative intensities (17,18). Linked to this, several studies have suggested that age-related blunting of FFA oxidation and increased IMCL accumulation are a result of reduced muscle mitochondrial content (8) and function (3,19,20) with age. However, increased adiposity and reduced habitual levels of physical activity also affect FFA flux and oxidation in older individuals (21), and studies to date have not controlled for these factors when investigating changes in muscle IMCL metabolism with age. Therefore, we investigated the effect of aging on whole-body and skeletal muscle lipid metabolism, with parallel characterization of muscle insulin sensitivity, in lean young and older individuals matched for estimated habitual physical activity levels and body composition. For determination of the effect of adiposity and reduced physical activity on the aging process, the older lean individuals were also compared with a group of older overweight individuals matched for lean mass. We hypothesized that an age-associated imbalance between FFA delivery and oxidation in skeletal muscle during exercise would only be observed in older overweight individuals, which would manifest as reduced IMCL oxidation and increased IMCL storage, particularly in the SSL region, and be associated with skeletal muscle insulin resistance.

Seven young lean (YL) (BMI <25 kg/m²), old lean (OL) (BMI <25 kg/m²), and old overweight (OO) (BMI >27 kg/m²) healthy, recreationally active male volunteers participated in the current study, which was approved by the University of Nottingham’s medical school ethics committee in accordance with the Declaration of Helsinki. Before taking part, all subjects underwent routine medical screening and completed a quality of life (SF-36) questionnaire indicating ability to perform physical activity. They also completed a general health questionnaire indicating their habitual frequency of performing moderate- to high-intensity physical activities including team sports, resistance exercise, running, cycling, and swimming (Table 1). Informed consent was obtained from all volunteers before participation in the study, and they were made aware that they were free to withdraw at any point. On a separate visit, all subjects performed an incremental exhaustive exercise test on an electronic-braked cycle ergometer (Excalibur, Lode Medical Technology, Groningen, the Netherlands) to determine VO_2max (Quark CPET, COSMED, Rome, Italy) and the workload that would elicit 50% VO_2max, which was confirmed in a familiarization visit at least 3 days later.

Subjects attended the laboratory on two occasions separated by at least 1 week. On the first occasion, they arrived at 0800 h after an overnight fast, having abstained from exercise and alcohol for the previous 48 h, in order to determine their body composition and insulin sensitivity. Trunk, leg, and arm composition using standardized regions were analyzed by a single operator using DEXA (Lunar Prodigy, GE Healthcare). Subjects then rested semisupine on a bed and underwent a 3-h euglycemic (4.5 mmol/L)–hyperinsulinemic (60 mU · m⁻² · min⁻¹) clamp (22) in combination with the intravenous infusion of 2-deoxy-d-glucose (2DG) (6 mg · kg⁻¹ · h⁻¹) to assess whole-body and skeletal muscle insulin sensitivity, respectively. 2DG is a glucose analog that closely resembles glucose in the characteristics of its transport but is metabolized by muscle to the 6-phosphate derivative (2DG6P). Thus, muscle 2DG6P is effectively trapped and its content can be determined as a direct measure of muscle glucose uptake (23). Indirect calorimetry (GEMNutrition Ltd., Daresbury, U.K.) was performed prior to and after 2 h of the clamp.

On the second visit, volunteers again reported to the laboratory after an overnight fast and rested semisupine on a bed for infusion of NaH¹³CO₃ (Cambridge Isotope Laboratories) and [U-¹³C]palmitate (99% enriched; Cambridge Isotope Laboratories) bound to 4.5% human serum albumin (Zenalb 4.5, Bio Products Laboratory, Elstree, U.K.) at a ratio of approximately 3:1 (1.94:0.64 μmol/L).After a 63.75 μg/kg bolus of NaH¹³CO₃ to prime the bicarbonate pool (24), [U-¹³C]palmitate was infused at a rate of 0.19 mg · kg⁻¹ · h⁻¹ for 2 h, which then increased to 0.28 mg · kg⁻¹ · h⁻¹ at the onset of 1 h cycling exercise at 50% VO_2max.

During the first visit, arterialized venous blood (25) was obtained before and every 5 min throughout the clamp for measurement of blood glucose concentration (YSI 2300 STAT Plus Glucose & Lactate Analyzer, YSI Inc.) and every 30 min throughout the clamp for subsequent analysis of serum insulin using a solid-phase ¹²⁵I radioimmunoassay kit (human insulin assay, Merck Millipore) and plasma 2DG via gas chromatography–mass spectrometry (GC-MS) (MD800; Fisons, ThermoFisher Scientific, Loughborough, U.K.) (23). Needle biopsy samples were obtained from the vastus lateralis (26) before and immediately after the clamp and snapfrozen in liquid nitrogen. At a later date, 30 mg wet muscle was pulverized for analysis of 2DG6P content using a commercial spectrophotometric kit method (Cosmo Bio Co., Ltd., Tokyo, Japan) (27). In addition, ∼20 mg wet muscle from the baseline biopsy was used to determine muscle citrate synthase (CS) maximal activity spectrophotometrically (28) and carnitine palmitoyltransferase 1 (CPT1) maximal activity using a radioisotope assay (29). Total RNA was also extracted from ∼20 mg wet muscle tissue (Trizol reagent; ThermoFisher Scientific), and after generation of first-strand cDNA (SuperScript III kit; ThermoFisher Scientific), the relative abundance of mRNA of 12 genes from pathways involved in FFA oxidation and IMCL metabolism was determined using RT-PCR microfluidic cards (Applied Biosystems) (29).

On the second experimental visit, blood samples were obtained before and every 10 min during exercise and analyzed immediately for blood lactate concentration (YSI 2300 STAT Plus Glucose & Lactate Analyzer), and after centrifugation, plasma was stored at −80°C. Plasma treated with tetrahydrolipostatin (30 μg/mL plasma) was analyzed for total FFA (NEFA C kit, WAKO Chemicals, Neuss, Germany) on an automated analyzer (ABX Pentra 400, Horiba Medical, Ltd., Montpellier, France). Plasma separated from EGTA-treated blood was analyzed for [U-¹³C]palmitate and palmitate by TSQ triple quadrupole gas chromatography–mass spectrometry/mass spectrometry (ThermoFisher Scientific) and GC-MS (MD800; Fisons), respectively, after addition of a heptadecanoic internal standard and derivatization to their methyl esters (30). High-performance liquid chromatography with electrochemical detection was used to measure plasma epinephrine and norepinephrine concentrations (31). Breath samples were also collected every 10 min during exercise via one-way valve bags and introduced into vacuumed glass tubes (Exetainer, Labco Limited, Lampeter, U.K.) for subsequent ¹³CO₂ enrichment analysis by continuous-flow isotope-ratio mass spectrometry (AP2003; Breath Gas System, Analytical Precision, Cheshire, U.K.) (32). During the last 10 min of exercise when the ¹³CO₂ production was at a steady state and therefore no longer being retained by the muscle (negating the requirement for an acetate recovery factor), indirect calorimetry was performed (Quark CPET, COSMED). In addition, a vastus lateralis needle biopsy (26) was performed immediately before and after the exercise bout and processed within 10 s to minimize ex vivo changes in intracellular metabolism and contamination of the IMCL pool by extracellular adipocytes. A 5-mg portion buffered in ice-cold 3% gluteraldehyde and 0.1 mol/L sodium cacodylate (pH 7.4) was stored at 4°C for subsequent electron microscopy processing, and the remainder was immediately frozen in lipid nitrogen. Samples for transmission electron microscopy were fixed in 1% osmium tetroxide, dehydrated in graded ethanol series, and embedded in two resin blocks. Three ultrathin 70- to 90-nm sections were cut from each block, mounted on copper grids, and stained in uranyl acetate and lead acetate, with one section randomly selected to be visualized at ×4,200 magnification. Approximately 40 fields of view from up to 40 longitudinal fibers were systematically randomly selected by a blinded operator using the corners of copper grid squares as a guide. This method obtained at least six images per sample containing an SSL region, which was required for reproducible estimation of IMCL droplet (lipid droplet [LD]) characteristics. Images were analyzed using ImageJ to determine percentage of intermyofibrillar (IMF) and SSL area covered by LD, LD size, and total number of LDs per square micrometer of local tissue area, which have previously been shown (8) to produce values similar to three-dimensional stereology volume estimates (33). In addition, a portion of the pre-exercise biopsy was freeze dried, dissected free of visible blood and connective tissue, pulverized, and used for the quantification of DAG and ceramide. Briefly, 50 ng internal standard (1,3[d5]-15:0 DAG) was added to 5 mg muscle, from which total muscle lipids were extracted in CHCl₃:MeOH:H₂O and the most abundant DAG (diC16:0, C16:0/C18:1, and diC18:1) and ceramide (C16:0, C18:0, C18:1, C20:0, C24:1, and C24:0) species were quantified by liquid chromatography–mass spectrometry/mass spectrometry (Quattro Ultima, Micromass Ltd., Wythenshawe, U.K.) (34,35). Peak areas were normalized to the internal standard and converted to absolute concentrations using a standard curve specific for each species. A further portion of muscle powder was also used for the determination of muscle creatine, phosphocreatine, glycogen, lactate, and acetylcarnitine as previously described (36).

Insulin sensitivity index (SI_Clamp) was calculated using the equation of Matsuda and DeFronzo (22) of SI_Clamp = M/(G × ΔI), where steady-state (120–180 min) glucose disposal (M) is normalized for steady-state blood glucose concentration (G) (mmol/L) and the difference between fasting and steady-state plasma insulin concentrations (ΔI) (mU/L). Indirect calorimetry calculations both at rest and during exercise were performed according to nonprotein stoichiometric equations (37) and normalized to lean body mass (DEXA). The R_a, R_d, and oxidation of palmitate during the final 10 min of exercise were used to calculate total plasma FFA kinetics by dividing by the fractional contribution of plasma palmitate to total plasma FFA concentration, as previously described (38). The contribution of other fat sources was calculated by subtracting plasma FFA oxidation from total fat oxidation calculated via indirect calorimetry.

Differences between groups at baseline and within and between groups during exercise were analyzed using a one- and two-way ANOVA, respectively (GraphPad Prism 6.0, GraphPad Software, Inc.). When a significant main effect was observed, Tukey and Sidak post hoc tests were performed, respectively, to identify individual differences. Statistical significance was set at P < 0.05, and all values are presented as means ± SEM.

In line with the inclusion criteria, OL and YL had a similar body composition. However, OO had greater trunk, arm, and leg fat masses compared with OL and YL subjects but similar whole-body, arm, and leg fat-free mass (Table 1). Furthermore, self-reported levels of habitual physical activity were similar between OL and YL but less in OO compared with OL. Both OL and OO had similar absolute (mL/min) and relative (mL · kg lean body mass [lbm]⁻¹ · min⁻¹) VO_2max, but these were less than in YL, as were the corresponding absolute workload and heart rate at 50% VO_2max.

Steady-state serum insulin and glucose disposal during the euglycemic-hyperinsulinemic clamp for OL, YL, and OO were 119.6 ± 7.0, 117.7 ± 7.8, and 137.5 ± 4.3 mU/L and 57.8 ± 5.6, 65.1 ± 5.6, and 41.6 ± 5.2 µmol · kg lbm⁻¹ · min⁻¹ (P < 0.01, OL and YL vs. OO), respectively. As such, OL and YL had similar SI_Clamp values that were 57% (P < 0.05) and 86% (P < 0.01) greater than OO, respectively (Fig. 1A). Furthermore, muscle 2DG6P accumulation during the clamp was not different between OL and YL, but was less than half that of YL in OO (P < 0.01) (Fig. 1B), and OL and YL had similar steady-state plasma 2DG concentrations during the clamp, which were less than OO (50.0 ± 1.9 and 48.8 ± 2.2 vs. 72.6 ± 2.2 µmol/L, respectively; P < 0.05) (Fig. 1C). Insulin-stimulated resting energy expenditure increased by >10% in both OL (P < 0.01) and YL (P < 0.05) but did not change in OO (P = 1.0) (Fig. 1D). There were no differences in muscle DAG species between groups with the exception of diC18:1, which was lower in OL and YL compared with OO (both P < 0.05) (Fig. 2A). Similarly, most muscle ceramide species did not differ except C20:0, which was lower in YL compared with OO (P < 0.01) (Fig. 2B).

Whole-body energy expenditure during the last 10 min of 1 h of exercise at 50% VO_2max was lower in OL and OO compared with YL (both P < 0.05) (Fig. 3A), but the relative contribution from total fat oxidation to energy expenditure was similar (42.4 ± 3.1, 40.1 ± 4.6, and 43.9 ± 6.5%, respectively). Nevertheless, the oxidation of fat from sources other than plasma FFA (i.e., predominantly from IMCL) was almost threefold lower in OL and OO compared with YL (both P < 0.05) (Fig. 3A), such that the relative contribution of these sources to total fat oxidation was around half that of YL (38.7 ± 7.7 and 45.0 ± 7.9 vs. 71.9 ± 3.1%, respectively; P < 0.01) (Fig. 3A). Plasma FFA R_a was similar between OL and OO (24.2 ± 2.9 vs. 24.3 ± 5.3 µmol · kg lbm⁻¹ · min⁻¹), and greater compared with YL (13.8 ± 2.3 µmol · kg lbm⁻¹ · min⁻¹; P < 0.05), but there were no differences in plasma FFA concentration (0.62 ± 0.06, 0.58 ± 0.09, and 0.45 ± 0.08 for OL, OO, and YL, respectively). Plasma FFA R_d was also similar between OL and OO but greater in OL compared with YL (P < 0.05) (Fig. 3B). In contrast, whereas the percentage of plasma FFA R_d oxidized was similar between OL and YL (54.4 ± 5.9 and 52.7 ± 3.4%, respectively) (Fig. 3C), it was lower in OO (42.2 ± 1.2%) compared with OL (P < 0.05) and YL (P = 0.07) (Fig. 3C).

From similar baseline concentrations, plasma norepinephrine increased to a similar steady state in OL and OO throughout 1 h of exercise and was ∼1.5-fold greater than the steady-state concentration achieved in YL (both P < 0.05, respectively) (Fig. 3D). However, there were no differences between groups in baseline or steady-state plasma epinephrine (0.25 ± 0.02 to 0.53 ± 0.11, 0.25 ± 0.06 to 0.34 ± 0.06, and 0.24 ± 0.04 to 0.43 ± 0.06 nmol/L) or blood lactate (0.71 ± 0.06 to 1.50 ± 0.29, 0.97 ± 0.13 to 2.05 ± 0.31, and 0.87 ± 0.09 to 1.73 ± 0.28 mmol/L) concentrations in OL, OO, and YL, respectively.

The area of SSL region covered by LD was similar between OL and YL at rest and did not change during exercise (Fig. 4A). However, SSL area covered by LD in OO was almost threefold greater at rest compared with YL (P < 0.05) and increased during exercise (P < 0.05), such that postexercise it was greater than both OL (P < 0.05) and YL (P < 0.01) (Fig. 4A). This was predominantly due to a 25% increase in average SSL LD size in OO (P = 0.05) (Fig. 4B). In contrast, exercise caused a decrease (P < 0.01) in both the number of IMF LD (0.024 ± 0.001 to 0.017 ± 0.003, 0.022 ± 0.003 to 0.015 ± 0.002, and 0.023 ± 0.03 to 0.018 ± 0.003 LD/µm² for OL, YL, and OO, respectively) and area covered by LD (Fig. 4C). The latter was isolated to a 40% reduction in IMF area covered by LD in YL (P < 0.05) (Fig. 4C). Average IMF LD size was 45% greater in OO compared with OL and YL postexercise (both P < 0.01) (Fig. 4D).

Resting skeletal muscle glycogen (Fig. 5A), phosphocreatine (Fig. 5B), and lactate (Fig. 5C) content was similar between OL and YL and did not change measurably during exercise, whereas acetylcarnitine content increased during exercise by approximately sevenfold (P < 0.001) and threefold (P < 0.05), respectively. However, resting muscle glycogen and phosphocreatine content were lower (P < 0.05) and muscle lactate content more than doubled during exercise (P < 0.05) in OO. Nevertheless, there were no significant differences in maximal CS (116.8 ± 12.6, 94.7 ± 8.8, and 84.1 ± 8.3 nmol · mg protein⁻¹ · min⁻¹, respectively) or CPT1 (2.3 ± 0.3, 1.8 ± 0.3, and 2.0 ± 0.1 nmol · mg protein⁻¹ · min⁻¹, respectively) activities between OL, YL, and OO, respectively, although the former tended to be greater in OL versus OO (P = 0.08).

The relative expression of 12 skeletal muscle transcripts involved in fatty oxidation and IMCL turnover is presented in Table 2. HADHB and PLIN2 expression was greater in OL compared with YL (P < 0.05), whereas ACACB, SPTLC1, and DGKD expression was lower in YL compared with OO (all P < 0.05). Furthermore, PLIN2 gene expression was greater in OL versus OO (P < 0.05), respectively.

Insulin resistance is closely related to IMCL accumulation, and both are associated with increasing age. However, it remains to be determined to what extent perturbations in IMCL metabolism are related to the aging process per se or secondary to age-related changes in lifestyle. Thus, by matching young and older volunteers for body composition and self-reported habitual physical activity levels, the current study demonstrates that lean older individuals display IMCL content and insulin sensitivity comparable with those in their younger counterparts. On the other hand, aging per se appeared to cause an exacerbated lipolytic response to exercise due, at least in part, to an increased sympathetic response. Coupled with increased adiposity and reduced habitual physical activity levels in an age-matched group, this resulted in SSL IMCL accumulation and may mechanistically help explain the association between increased IMCL and skeletal muscle insulin resistance in older individuals.

In line with several studies that suggest aging per se does not cause insulin resistance (3–5), there was no difference in whole-body glucose disposal, skeletal muscle 2DG6P accumulation, or the energy expenditure response during a euglycemic-hyperinsulinemic clamp between old and young individuals matched for body composition and self-reported physical activity in the current study. Furthermore, the finding that whole-body and skeletal muscle insulin action was reduced in old overweight individuals with a similar lean body mass but lower self-reported physical activity supports the notion that lifestyle factors are more influential in the development of age-related insulin resistance (3–5,39–41). A possible link between these factors and reduced skeletal muscle insulin sensitivity is the accumulation of SSL IMCL (9,10) and associated lipid metabolites such as DAG and ceramide (13–15). Indeed, whereas there was no difference in SSL IMCL between OL and YL, SSL IMCL was more than twofold higher in the OO individuals, which is in agreement with a two- and threefold greater content observed in lean sedentary older individuals (8) and patients with type 2 diabetes (9), respectively. However, although the skeletal muscle content of the predominant DAG and ceramide species were not different between OL and YL, they were also not greater in OO, with the exception of diC18:1 DAG and C20:0 ceramide. Indeed, total muscle DAG and ceramide do not correlate well with insulin sensitivity, but specific lipid species, particularly sarcolemmal saturated DAG, may influence insulin action (42).

Why IMCL accumulates, particularly in the SSL region, is not clear, but several studies have demonstrated reduced FFA oxidation in older individuals despite increased whole-body lipolysis and FFA availability compared with young at rest and during exercise (17,18). Indeed, although the relative contribution of fat oxidation to total energy expenditure during exercise was not different between the young and old groups of the current study, there was an elevated plasma norepinephrine, FFA R_a, and FFA R_d response to exercise at the same relative intensity observed in both the lean and overweight older individuals, suggesting an effect of age per se on whole-body responses. This would fit with previous reports that age-associated increments in norepinephrine are independent of habitual physical activity and likely due to increased sympathetic activity rather than reduced norepinephrine clearance (43). As a consequence, the relative contribution of IMCL to fat oxidation was reduced in both old lean and overweight compared with young individuals. This is remarkable given there was presumably a greater lipolytic stimulus to IMCL by norepinephrine in the older individuals (44,45) and would suggest a potent inhibitory effect of plasma-derived FFA or a blunted contraction-induced IMCL hydrolysis. Furthermore, a novel finding of the current study was that, assuming similar rates of adipose tissue FFA reesterification (where FFA released from adipose tissue is reincorporated in a futile cycle), OL individuals were able to oxidize a larger proportion of the excess FFA delivered during exercise compared with the OO individuals, in whom it deposited in SSL LDs. This not only suggests that a more general, chronic imbalance between skeletal muscle FFA delivery and oxidation may contribute to IMCL accumulation but also provides evidence for distinct roles of the localized IMCL pools. For example, the reduction in the number of IMF LDs during exercise in the younger individuals suggests that this pool is used for muscle contraction, possibly in an “all or nothing” fashion, whereas the deposition in the SSL pool suggests a role in buffering/trafficking of FFA influx (46) and perhaps insulin resistance. Interestingly, an improvement in insulin sensitivity has been previously observed with reduced SSL but not IMF IMCL after 10–12 weeks of exercise training where the capacity to oxidize FFA was increased (9,10).

In addition to reduced energy expenditure, such as was observed during insulin-stimulated conditions of the current study, several mechanisms may explain the apparent inability of skeletal muscle of older overweight men to oxidize excess FFA delivery. For example, it has been suggested that aging is associated with impaired in vivo (19) and in vitro (3,20) skeletal muscle mitochondrial ATP production, as well as a reduction in mitochondrial content (8), independently of adiposity. However, there was no difference in skeletal muscle maximal CS activity, maximal and relative CPT1 activity, or phosphocreatine, glycogen, lactate, or acetylcarnitine metabolism during exercise between the lean old and young participants in the current study, all of which are markers of in vivo muscle oxidative capacity. On the other hand, the disparity between OL and OO participants in the ability to oxidize excess fatty acids may be due to differences in partitioning of skeletal muscle lipid and a diversion of fatty acids from oxidation toward synthesis of IMCL and other lipid species. For example, OL individuals had a greater mRNA expression of perilipin 2 (PLIN2), an LD bound protein involved in IMCL hydrolysis, and β-hydroxyacyl-CoA dehydrogenase (HADHB), an intramitochondrial enzyme that catalyzes a rate-limiting step in β-oxidation, whereas OO had greater mRNA expression of acetyl-CoA carboxylase 2 (ACACB), which produces malonyl-CoA and inhibits CPT1, the rate-limiting step for fatty acid entry into mitochondria. The PLIN2 expression in particular would fit previous reports in overweight individuals of impaired IMCL turnover and FFA release toward mitochondrial oxidation (44,47). Furthermore, OO had a greater expression of DAG kinase Δ (DGKD), which phosphorylates DAG to produce phosphatidic acid, and serine palmitoyltransferase (SPTLC1), a rate-limiting step in ceramide synthesis, compared with young lean individuals. Both of these observations fit with the greater muscle content of some of the DAG and ceramide species in the current study. A similar gene expression pattern has also been previously observed in insulin-resistant individuals (48), but how this translates into protein content/activity and whether it is cause or effect require further investigation.

In conclusion, it is our assertion that increased IMCL (4,8,19) and reduced insulin sensitivity, mitochondrial capacity, and fat oxidation (3,17–21,49) often observed in older individuals are likely due to lifestyle factors rather than aging per se as commonly reported. However, age per se appears to increase the systemic sympathetic response to exercise and cause exacerbated adipose tissue lipolysis. Compounded by greater adiposity, the increased FA delivery appears to cause SSL IMCL accumulation in physically inactive older individuals. Thus, targeted strategies to reduce muscle lipid delivery and improve adipose tissue function may be warranted, particularly as physical inactivity appears to worsen the inability to suppress adipose tissue lipolysis in older individuals (50).

Acknowledgments. The authors thank Sara Brown (David Greenfield Human Physiology Unit, The University of Nottingham) and Denise Christie (School of Life Sciences Imaging and Microscopy Facility, The University of Nottingham) for technical assistance.

Funding. This study was funded by the Dunhill Medical Trust (R211/0711).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. C.C. and F.B.S. drafted the manuscript. All of the authors researched data and reviewed and edited the manuscript. F.B.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.