Brain region-specific disruption of mitochondrial bioenergetics in cynomolgus macaques fed a Western versus a Mediterranean diet

Abstract

Mitochondrial dysfunction is evident in diseases affecting cognition and metabolism such as Alzheimer’s disease and type 2 diabetes. Human studies of brain mitochondrial function are limited to postmortem tissue, preventing the assessment of bioenergetics by respirometry. Here, we investigated the effect of two diets on mitochondrial bioenergetics in three brain regions: the prefrontal cortex (PFC), the entorhinal cortex (ERC), and the cerebellum (CB), using middle-aged nonhuman primates. Eighteen female cynomolgus macaques aged 12.3 ± 0.7 yr were fed either a Mediterranean diet that is associated with healthy outcomes or a Western diet that is associated with poor cognitive and metabolic outcomes. Average bioenergetic capacity within each brain region did not differ between diets. Distinct brain regions have different metabolic requirements related to their function and disease susceptibility. Therefore, we also examined differences in bioenergetic capacity between brain regions. Mitochondria isolated from animals fed a Mediterranean diet maintained distinct differences in mitochondrial bioenergetics between brain regions, whereas animals fed the Western diet had diminished distinction in bioenergetics between brain regions. Notably, fatty acid β-oxidation was not affected between regions in animals fed a Western diet. In addition, bioenergetics in animals fed a Western diet had positive associations with fasting blood glucose and insulin levels in PFC and ERC mitochondria but not in CB mitochondria. Altogether, these data indicate that a Western diet disrupts bioenergetic patterns across brain regions and that circulating blood glucose and insulin levels in Western-diet fed animals influence bioenergetics in brain regions susceptible to Alzheimer’s disease and type 2 diabetes.

NEW & NOTEWORTHY We show that compared with cynomolgus macaques fed a Mediterranean diet, a Western diet resulted in diminished bioenergetic pattern between brain regions related to blood glucose and insulin levels, specifically in brain regions susceptible to neurodegeneration and diabetes. In addition, fatty acid metabolism not directly linked to the TCA cycle and glucose metabolism did not show differences in bioenergetics due to diet.

INTRODUCTION

Dietary patterns have long been known to influence brain health and disease risk. Mediterranean diets rich in unsaturated fats, vegetables, fruits, nuts, and lean fish are considered protective against diseases affecting cognition, such as Alzheimer’s Disease (AD), and metabolic disruptions such as type 2 diabetes (T2D) (1–5). In contrast, Western diets abundant in saturated fats, sodium, fatty meat, and refined sugars are associated with an increased risk for cognitive decline, AD, and T2D (2, 6–10).

Perturbations in energy homeostasis are linked with neurodegenerative processes leading to diseases affecting cognition (11). Individual brain regions have different susceptibilities to disease, metabolic dysregulation, and mitochondrial dysfunction (12–15). In AD, the prefrontal cortex (PFC) and entorhinal cortex (ERC) are both affected. The PFC accumulates hallmark Aβ plaques in the second stage of the four-stage model of regional amyloid progression, indicating the PFC is vulnerable to amyloid pathology (16). Of note, brain regions susceptible to amyloid pathology express low levels of mitochondrial respiration genes (17). Other brain regions, such as the ERC, are susceptible to the accumulation of hallmark neurofibrillary tangles ahead of amyloid pathology (18). In fact, the ERC is one of the first brain regions to accumulate neurofibrillary tangles (18). In contrast, the cerebellum (CB), although it has some vulnerability to AD-related perturbations in homeostasis, also exhibits protective cellular processes against AD-induced damage (19, 20).

Similarly, in T2D, the PFC and ERC are vulnerable to dysregulation and damage. Executive functions controlled by the PFC are diminished, and cortical thickness in the PFC and ERC is reduced in T2D (21–23). The cerebellum also experiences alterations due to diabetes, but there is evidence that the cerebellum is more resilient to hypoglycemia and hyperglycemia than cortical regions (24). However, T2D is only diagnosed after glucose and insulin dysregulation have reached clinical thresholds (25). Therefore, the effects of glucose and insulin on mitochondrial bioenergetics in the brain at preclinical or risk stages are largely unknown.

In this study, we seek to understand how chronic consumption of two diets during middle age, one protective and one designed to promote metabolic dysregulation, affects mitochondrial bioenergetics in metabolically vulnerable and metabolically resilient brain regions. We also examine associations between glucose and insulin levels with brain region-specific bioenergetics.

To address these questions, we isolated mitochondria from the PFC, ERC, and CB of 18 middle-aged female cynomolgus macaques with 9 fed a Mediterranean diet and 9 fed a Western diet for 30 mo. We utilized high-resolution respirometry (HRR) to measure the oxygen consumption rates of isolated brain mitochondria exposed to substrates for oxidative phosphorylation (OXPHOS). HRR permits analysis of isolated mitochondria in response to the addition/titration of substrates, uncouplers, and inhibitors. Substrate-uncoupler-inhibitor titration (SUIT) protocols allow us to test specific electron entry points into the mitochondrial electron transfer system (ETS) to produce ATP to meet cellular energy demands. We used two complementary SUIT protocols to measure 11 parameters of mitochondrial bioenergetics. One protocol provided information on the activity of two dominant entry points into the ETS (Complex I and Complex II), whereas the other incorporated additional information about the utilization of fatty acids and glycerol-3-phosphate through two additional entry points (electron-transferring flavoprotein complex and glycerophosphate dehydrogenase).

METHODS

Experimental Model and Study Design

This study utilized brain tissue samples from 18 randomly selected middle-aged female cynomolgus macaques (Macaca fascicularis) from a larger parent study (26). Only female animals were used in this study because the parent study focuses on cardiovascular disease, which disproportionally affects women’s health. Alzheimer’s disease also disproportionately affects women’s health and specifically mitochondria in women with Alzheimer’s disease (26, 27). Demographic information is summarized in Table 1. Animals were housed in small social groups of four, in pens measuring 3.3 m × 3.3 m × 3.3 m, on a 12-h/12-h light/dark schedule, with water available ad libitum for 38 mo. The animals were fed either a Western or Mediterranean diet designed and produced by the Primate Nutrition and Diet Laboratory at Wake Forest School of Medicine for 30 mo. Macronutrient compositions and cholesterol content were similar between the two diets. Protein:fat:carbohydrate caloric ratios were 15.6:30.6:53.7 in Western diet and 16.4:31.7:51.9 in Mediterranean diet (Supplemental Table S1, all Supplemental material is available at https://doi.org/10.6084/m9.figshare.14529459) (26). In the Western diet, protein and fat were derived primarily from animal sources, whereas protein and fat were derived mostly from plant sources in the Mediterranean diet. This resulted in diet differences in fatty acid content. The Mediterranean diet was enriched in monounsaturated fatty acids and had a lower n-6:n-3 fatty acid ratio (3:1), less salt, and more fiber than the Western diet. The Western diet was enriched in saturated fats and had a n-6:n-3 of 15:1. Brain tissue was collected at the time of necropsy. Animals were sedated with intramuscular ketamine hydrochloride (15 mg/kg). Then, intravenous sodium pentobarbital (∼13 mg/kg) was administered to achieve surgical anesthesia followed by exsanguination in accordance with guidelines established by the Panel on Euthanasia of the American Veterinary Medical Association. All animal manipulations were performed according to the guidelines of state and federal laws, the US Department of Health and Human Services, and the Animal Care and Use Committee of Wake Forest School of Medicine (ACUC #A12-195; #A15-180).

Table 1.

Characteristics of middle-aged female cynomolgus macaque cohort

	Total (n = 18)	Mediterranean Diet (n = 9)	Western Diet (n = 9)	P Value‡
Age, mean (SD) [range], yr	12.3 (0.7) [1.2–13.7]	12.2 (0.8) [11.2–13.7]	12.4 (0.6) [11.5–13.2]	0.45
Weight,* mean (SD) [range], kg	3.6 (1.3) [2.0–7.1]	3.3 (1.2) [2.0–5.2]	3.8 (1.4) [2.7–7.1]	0.45
BMI,* mean (SD) [range], kg/m2	45.8 (11.1) [32.5–76]	42.8 (8.8) [32.5–58.8]	48.8 (12.8) [37.4–76]	0.26
Fasting blood glucose,† mean (SD) [range], mg/dL	84.6 (16.8) [61–120]	89.4 (20.1) [64–120]	79.7 (12) [61–96]	0.23
Fasting blood insulin,† mean (SD) [range], mlU/L	59.7 (45.3) [6.4–175.3]	52.4 (33.5) [16.6–116.7]	68.0 (57.1) [6.4–175.3]	0.50

BMI, body mass index.

^*Age, weight, and BMI were recorded at necropsy following 30 mo of diet intervention.

^†Fasting glucose and insulin levels were measured using plasma samples collected at intervention month 26.

^‡An unpaired two-tailed Student’s t test was used to compare differences between diet groups.

Body Mass Measurements

Body length, measured from the suprasternal notch to the pubic symphysis, was recorded during month 27 of diet treatment, and body weight was measured at time of necropsy. Body mass index (BMI; in kg/m²) was measured using the calculation described previously (28).

Insulin, Glucose, and Insulin Sensitivity Measurements

Intravenous fasting glucose and insulin levels were measured during treatment phase month 26 as previously described (29). Briefly, animals were fasted for 18 h before sedation with intramuscular ketamine hydrochloride (15 mg/kg body weight). Glucose concentration was determined by colorimetric assay using reagents and instrumentation (ACE Alera Autoanalyzer) from Alfa Wasserman Diagnostic Technologies (30). Insulin concentration was determined using an ELISA assay (Mercodia).

Mitochondrial Isolation

Brain tissues were kept in ice-cold phosphate buffered saline for 1 h before processing. Approximately 80–100 mg of wet-weight tissue was used per brain region. The tissue was homogenized using 70 strokes of a glass-on-glass type B Dounce homogenizer. The mitochondria were isolated using differential centrifugation as previously described (31). Briefly, the steps were executed as follows: centrifuge 1,400 g for 7 min at 4°C; decant supernate through cheesecloth and retain; centrifuge 1,400 g for 7 min at 4°C; retain supernate; centrifuge 10,000 g for 5 min at 4°C; retain pellet; resuspend in 600 µL of mitochondrial isolation buffer (17.5 mM sucrose, 55 mM mannitol, 1.25 mM KH₂PO₄, 1.25 mM MgCl₂, 0.5 mM HEPES, 0.25 mM EGTA, 0.05% fatty acid free BSA, 7.4 pH); centrifuge 800 g for 3 min at 4°C; retain supernate; centrifuge 10,000 g for 5 min at 4°C; and resuspend pellet in Mir05 (Oroboros, Innsbruck, Austria), a respiration medium containing 0.5 mM EGTA, 3 mM MgCl₂, 60 mM lactobionic acid, 20 mM taurine, 10 mM KH₂PO₄, 20 mM HEPES, 110 mM d-sucrose, and 1 g/L fatty acid free BSA. Protein concentration of the isolated mitochondrial preparations was measured using a Nanodrop (ND-ONE-W, Thermo Scientific).

High-Resolution Respirometry

High-resolution respirometry (HRR) was performed using an Oroboros O2k oxygraph (Oroboros, Innsbruck, Austria). The chambers were filled with 2.1 mL of MiR05. The chamber equilibrated with room oxygen concentration at 37°C for at least 30 min. Then 0.9–3.8 µg of isolated mitochondria was added to each chamber. Respirometric values were normalized to protein. Three O2k oxygraphs were used concurrently, one for each brain region. Two substrate-uncoupler-inhibitor titration (SUIT) protocols were used for each region, one protocol per chamber.

In chamber A, we used Protocol 1, a SUIT protocol modified from a previously described protocol (32). Final chamber concentrations were follows: glutamate (10 mM) + malate (2 mM), ADP (2.5 mM) + Mg²⁺ (1.5 mM), cytochrome c (20 µM), succinate (10 mM), oligomycin (1 µM), FCCP (1.5–3.5 µM), rotenone (0.5 µM), and antimycin A (2.5 µM). FCCP was titrated to increase the concentration in the chamber in 0.5 µM increments until achieving maximal respiration. Thus, the number of titrations varied per sample.

In chamber B, we used Protocol 2, a modified SUIT protocol developed by Oroboros (33). The sequence of additions is as follows: ADP (2.5 mM) + Mg²⁺ (1.5 mM), octanoylcarnitine (5 mM), malate (2 mM), cytochrome c (20 µM), pyruvate (5 mM), glutamate (10 mM), succinate (50 mM), glycerol-3-phosphate (10 mM), FCCP (1.5–4 µM), rotenone (0.5 µM), and antimycin A (2.5 µM). FCCP was titrated to increase the concentration in the chamber in 0.5 µM increments until achieving maximal respiration. Thus, the number of titrations varied per sample.

Whole Tissue Western Blot

From each region, 10–11 mg of frozen brain tissue specimens were thawed and lysed by sonication in 150-µL Pierce RIPA Buffer (89900, Thermo Scientific) containing 1:100 protease inhibitor cocktail (P8340, Sigma, St. Louis, MO) and kept on ice for 30 min. The lysate was then spun at 21,000 g for 20 min at 4°C. Following lysis, the protein concentration of the supernatant was measured using a BCA assay kit (23223, Thermo Scientific). Then 15 µg of isolated mitochondrial protein per sample was loaded into a 15-well NuPAGE 12% Bis-Tris Gel (NP0343BOX, Invitrogen, Carlsbad, CA) and electrophoresed at 200 V for ∼70 min. PageRuler Plus Prestained Protein Ladder, 10–250 kDa (Cat. No. 26619, Thermo) was used to mark molecular weights. Following gel electrophoresis, the proteins were transferred to a PVDF membrane using transfer system at conditions recommended by the manufacturer for 1 h. The membrane was then blocked with 5% BSA for 1 h. The OXPHOS antibody cocktail (1:2,000; Cat. No. ab110411, Abcam, RRID:AB_2756818) containing anti-complex I subunit NDUFB8-18kDa, anti-complex II subunit SDHB-30kDa, anti-complex III subunit UQCRC2-48kDa, anti-complex subunit IV COX II-22kDa, and anti-Complex V subunit ATP5A-54kDa primary mouse monoclonal antibodies, the anti-VDAC1/Porin-31kDa primary rabbit polyclonal primary antibody (1:2,000; Cat. No. ab1589, Abcam, RRID:AB_2214787), and the anti-GAPDH-37kDa primary rabbit polyclonal primary antibody (1:1,000; Cat. No. ab9485, Abcam, RRID:AB_307275) were incubated with the membrane overnight at 4°C. The following day, the membrane was washed four times, 10 min each, in 0.1% TBS-T followed by incubation with (1:5,000) anti-mouse secondary antibody (NEF822E001EA, Perkin Elmer, Boston, MA) and (1:5,000) anti-rabbit secondary antibody (NEF812E001EA, Perkin Elmer, Boston, MA) for 1 h at room temperature. Subsequently, the membrane was washed four times, 10 min each, in 0.1% TBS-T and imaged. Images were taken using Image Lab software (v. 5.2.1). Images were analyzed using ImageJ (Schneider, Rasband, & Eliceiri, 2012) software. See Supplemental Fig. S1 for representative blot.

Citrate Synthase Assay

Citrate synthase activity was determined according to the manufacturer’s instructions (Citrate Synthase Assay Kit; Cat. No. CS0720, Sigma). Briefly, brain tissue was homogenized at pH 7.4 in prechilled CelLytic MT (Cat. No. C3228, Sigma) with protease inhibitor cocktail (Cat. No. P8340, Sigma). The homogenate was centrifuged at 12,000 g for 10 min at 4°C. The supernatant containing the protein was collected. Protein concentration was measured using BCA Kit (Cat. No. 23223, Thermo Scientific). Citrate synthase activity was measured by continuous spectrophotometric rate determination at 412 nm. Each sample was run in triplicate.

Statistical Analysis

To determine differences in bioenergetic capacities between brain regions and diets with O² flux bioenergetic measurements, citrate synthase activity, and western blots, we used a mixed-effects models of analysis with brain region defined as a random effect. Marginal means and exact P values are reported for all comparisons of mean respiration of each brain region with diet and of each diet within brain region. Mixed-effects models were conducted using SAS software, version 9.4 (Cary, NC). Pearson’s correlation coefficients were determined using GraphPad Prism Software, version 8.3.0, for Windows (San Diego, CA), and statistical significance is defined as P ≤ 0.05 and trending relationships are defined as P ≤ 0.1. Percent of total uncoupled respiration by fatty acid oxidation (FAO) was determined with a one-way ANOVA using Daniel’s XL Toolbox, version 7.3.4 (http://www.xltoolbox.net).

RESULTS

Characteristics of the Nonhuman Primate Cohort

Female cynomolgus macaques were used in this study. Age, body weight, body mass index (BMI), fasting blood insulin levels, and fasting blood glucose levels are summarized in Table 1. There were no statistically significant cross-sectional differences between these characteristics. However, this ancillary study examining nine animals administered the Mediterranean diet and nine administered the Western diet was not adequately powered to detect differences in these measurements. The parent study reported increased body weight after 6 mo and hyperinsulinemia after 2.5 yr in animals fed the Western diet but not in animals fed the Mediterranean diet (26).

Bioenergetic Profiling of Isolated Brain Mitochondria

We compared the bioenergetic profiles of isolated prefrontal cortex mitochondria (iPFCm), isolated entorhinal cortex mitochondria (iERCm), and isolated cerebellum mitochondria (iCBm) using two high-resolution respirometry substrate-uncoupler-inhibitor titration (SUIT) protocols. The sequence of substrate-uncoupler-inhibitor additions used in this study, their interpretations, and abbreviations are summarized in Table 2. For additional information on the terms and abbreviations adopted for this manuscript, refer to Gnaiger (34). The hashed boxes in Fig. 1 mark bioenergetic parameters measured during the assay. Additionally, before analysis, residual oxygen consumption (ROX) was subtracted from each bioenergetic parameter.

Table 2.

Summary of SUIT protocols and bioenergetic parameters

SUIT	Bioenergetic Parameter	Electron Path into ETS	Explanation
SUIT Protocol 1
Cat, catalase			Converts H2O2 into O2*
G, glutamate M, malate			Complex I substrates
D, ADP Mg, Mg2+	CI	Complex I	Complex I-linked OXPHOS† respiration
Cyt, cytochrome c		Complex I	Tests mitochondrial outer membrane integrity
S, succinate	CI + CII	Complex I + Complex II	Complex I + Complex II-linked OXPHOS† respiration
Oligo, oligomycin	LEAK	Complex I + Complex II	Oligomycin induced LEAK‡ respiration
U, FCCP	Max1	Complex I + Complex II	Complex I + Complex II-linked ETS§ respiration
Rot, rotenone	CII(E)	Complex II	Complex II-linked ETS§ respiration
Ama, antimycin A	ROX		Nonmitochondrial residual oxygen consumption
SUIT Protocol 2
Cat, catalase			Converts H2O2 into O2*
D, ADP Mg, Mg2+			Ensures mitochondria are not ADP limited
Oct, octanoylcarnitine		FAO	Fatty acid β-oxidation (FAO) substrate
M, malate	FAO	FAO	Fatty acid β-oxidation (FAO)-linked OXPHOS† respiration; low malate minimizes inhibitory effect of acetyl-CoA accumulation on FAO
Cyt, cytochrome c		FAO	Tests outer membrane integrity
P, pyruvate		FAO + Complex I	Complex I substrate
G, glutamate	FAO + CI	FAO + Complex I	FAO + Complex I-linked OXPHOS† respiration
S, succinate	FAO + CI + CII	FAO + Complex I + Complex II	FAO + Complex I + Complex II-linked OXPHOS† respiration
Gp, glycerol-3-phosphate	FAO + CI + CII + GpDH	FAO + Complex I + Complex II + Glycerol-3-phosphate dehydrogenase	FAO + Complex I + Complex II + Glycerol-3-phosphate dehydrogenase-linked OXPHOS† respiration
U, FCCP	Max2	FAO + Complex I + Complex II + Glycerol-3-phosphate dehydrogenase	FAO + Complex I + Complex II + Glycerol-3-phosphate dehydrogenase-linked ETS§ respiration
Rot, rotenone	CII + GpDH(E)	Complex II + Glycerol-3-phosphate dehydrogenase	Complex II + Glycerol-3-phosphate dehydrogenase-linked ETS§ respiration
Ama, antimycin A	ROX		Nonmitochondrial residual oxygen use

ROX, residual oxygen consumption; SUIT, substrate-uncoupler-inhibitor titration.

^*H₂O₂ added with catalase increases oxygen concentration in the chamber. H₂O₂ was not required for any experiments in this study.

^†Oxidative phosphorylation (OXPHOS) respiration: oxygen use when the proton motive force generated by the electron transfer system is used to make ATP via ATP synthase.

^‡LEAK respiration: oxygen use when protons move from the intermembrane space to the matrix not facilitated by ATP synthase.

^§Electron transfer system (ETS) respiration: oxygen use when the proton motive force is eliminated by FCCP freely transporting protons across the inner membrane. Therefore, ETS activity is “uncoupled” from phosphorylation of ADP by ATP synthase.

Figure 1.

Representative high-resolution respirometry bioenergetic profiles. The thick black line indicates respiration reported as O₂ flux per mass (pmol·s⁻¹·mg⁻¹). The thin gray line indicates oxygen concentration in µM. Dark-gray hashed boxes indicate the portion of O₂ flux per mass selected to analyze for each bioenergetic parameter. A: substrate-uncoupler-inhibitor titration (SUIT) Protocol 1 marked with 5 bioenergetic parameters of interest [CI, CI + CII, LEAK, Max1, and CII(E)] and residual oxygen consumption (ROX). B: SUIT Protocol 2 marked with 6 bioenergetic parameters of interest [FAO, FAO + CI, FAO + CI + CII, FAO + CI + CII + GpDH, Max2, and CII + GpDH(E)] and ROX. SUIT abbreviations: Cat, catalase; GM, glutamate and malate; DMg, ADP and Mg²⁺; Cyt, cytochrome c; Succ, succinate; O, oligomycin; U, FCCP; Rot, rotenone; AA, antimycin A; Oct, octanoylcarnitine; P, pyruvate; G, glutamate. Bioenergetic parameter abbreviations: CI, Complex I respiration; CII, Complex II respiration; LEAK, Leak respiration; Max, Maximal ETS respiration; CII(E), Complex II ETS respiration; ROX, Residual oxygen consumption; FAO, respiration due to fatty acid oxidation; GpDH, respiration due to Glycerol-3-phosphate dehydrogenase. Refer to Table 2 for full summary of SUIT protocols and bioenergetic parameters.

We used mixed-effects models of analysis in SAS software, version 9.4, to evaluate the effect of the Mediterranean diet compared with the Western diet on bioenergetic capacity in each brain region. Oxygen consumption rates of brain-derived isolated mitochondria from animals fed the Western diet were significantly elevated in three bioenergetic parameters [Complex I (CI) in iERCm, CI in iCBm, and FAO in iCBm]. The remaining 30 comparisons in iPFCm were not significantly different between diet groups (Fig. 2). Notably, the standard deviation of bioenergetic values was greater for animals fed the Western diet compared with the Mediterranean diet (Fig. 2). This indicates that although the average bioenergetic capacity within each brain region did not differ significantly due to diet, animals administered the Western diet had a broader range of bioenergetic capacities compared with animals administered the Mediterranean diet. The broader range of bioenergetic capacity in the Western diet-fed animals was especially prominent in iPFCm but was also notable in iERCm and iCBm.

Figure 2.

Comparing isolated mitochondria bioenergetics from the PFC, ERC, and CB brain regions of female cynomolgus macaques fed either a Mediterranean or Western diet. A–K: high-resolution respirometry of isolated mitochondria measuring individual bioenergetic parameters. A mixed-effects model of analysis in SAS software, version 9.4, was used to evaluate the effect of the Mediterranean diet compared with the Western diet on bioenergetic capacity in each brain region due to diet. *P ≤ 0.05. Data are presented as means ± SD. Exact P values are recorded in Supplemental Table S2. n = 9 (PFC Med FAO and FAO-CI n = 8). CB, cerebellum; CI, Complex I respiration; CII, Complex II respiration; CII(E), Complex II ETS respiration; ERC, entorhinal cortex; FAO, respiration due to fatty acid oxidation; GpDH, glycerol-3-phosphate dehydrogenase; iCBm, isolated cerebellum mitochondria; iERCm, isolated entorhinal cortex mitochondria; iPFCm, isolated prefrontal cortex mitochondria; Med, Mediterranean diet; PFC, prefrontal cortex; West, Western diet.

Differences in Bioenergetic Capacity between Brain Regions Are Disrupted in Animals Administered a Western Diet

We used mixed-effects models of analysis to examine differences in bioenergetic capacity between brain regions. This statistical method calculates the difference in bioenergetic capacity between iPFCm and iERCm, iPFCm and iCBm, and iERCm and iCBm within each animal. These differences are then grouped by diet and brain region to determine if there is a significant difference in bioenergetic capacity between brain regions and if this difference between brain regions remains consistent between diets. This was repeated for each bioenergetic parameter.

SUIT Protocol 1 examines bioenergetic capacity driven by mitochondrial Complex I and Complex II.

Bioenergetic capacity in isolated brain mitochondria from animals fed the Mediterranean diet was significantly different between brain regions. Across all five SUIT Protocol 1 bioenergetic parameters, iPFCm respiration was greater than iERCm respiration, iPFCm was greater than iERCm respiration iCBm, and iERCm respiration was not different from iCBm respiration (Fig. 3, A–E, and Supplemental Table S3).

Figure 3.

Comparing iPFCm, iERCm, and iCBm bioenergetics across brain regions of female cynomolgus macaques fed either a Mediterranean or Western diet. A–K: high-resolution respirometry of isolated mitochondria measuring individual bioenergetic parameters. A mixed-effects model of analysis in SAS software, version 9.4, was used to evaluate the effect of the Mediterranean diet compared with the Western diet on bioenergetic capacity between brain regions due to diet. Each point represents the difference between the regions in one animal. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001. Exact P values are recorded in Supplemental Table S3. Data are presented as means ± 95% confidence interval, n = 9 (PFC Med FAO and FAO-CI n = 8). CI, Complex I respiration; CII, Complex II respiration; CII(E), Complex II ETS respiration; FAO, respiration due to fatty acid oxidation; GpDH, glycerol-3-phosphate dehydrogenase; iCBm, isolated cerebellum mitochondria; iERCm, isolated entorhinal cortex mitochondria; iPFCm, isolated prefrontal cortex mitochondria; Med, Mediterranean diet; PFC, prefrontal cortex; West, Western diet.

Bioenergetic capacity in isolated brain mitochondria from animals fed the Western diet was different between brain regions, but to a lesser extent, and differences between iPFCm and iCBm respiration were either not present or not as pronounced as they were in animals fed the Mediterranean diet. Overall, iPFCm respiration was greater than iERCm respiration, iPFCm respiration was not different from iCBm respiration except in the LEAK parameter (CI), and iERCm respiration was not different from iCBm respiration (Fig. 3, A–E, and Supplemental Table S3).

SUIT Protocol 2 examines bioenergetic capacity driven by fatty acid metabolism (FAO) through the electron-transferring flavoprotein complex (cETF) in conjunction with Complex I, Complex II, and glycerol-3-phosphate dehydrogenase (GpDH).

Bioenergetic capacity in isolated brain mitochondria from animals fed the Mediterranean diet was significantly different between brain regions and aligned with results from SUIT Protocol 1. Overall, iPFCm respiration was greater than iERCm respiration except in the FAO parameter, iPFCm respiration was greater than iCBm respiration except in the Complex II (CII) + GpDH(E) parameter and iERCm respiration was greater than iCBm respiration in FAO and FAO + CI, but not different for the remaining bioenergetic parameters (Fig. 3, F–K, and Supplemental Table S3).

Bioenergetic capacity in isolated brain mitochondria from animals fed the Western diet was different between some brain regions for some bioenergetic parameters. The overall strength of differences in respiration was diminished in mitochondria from animals fed the Western diet compared with respiration from animals fed the Mediterranean diet and some strong differences observed in mitochondria from animals fed the Mediterranean diet were not present in animals fed the Western diet. Overall, iPFCm respiration was greater than iERCm respiration for FAO + CI + CII and Max2 but not for the other bioenergetic parameters, iPFCm respiration was greater than iCBm respiration for FAO and FAO + CI + CII + GpDH but not for the other bioenergetic parameters, and iERCm respiration was greater than iCBm respiration for FAO and FAO + CI + CII + GpDH but not for the other bioenergetic parameters (Fig. 3, F–K, and Supplemental Table S3).

Fasting blood glucose and insulin levels are related to bioenergetic capacity of isolated brain mitochondria in animals administered a Western diet but not a Mediterranean diet.

Next, we examined if measures of metabolic heath such as fasting blood glucose and insulin levels relate to brain bioenergetic capacity. We used Pearson’s correlations to compare fasting blood glucose and insulin levels to each bioenergetic parameter from each brain region and separated the analysis by diet.

Fasting blood glucose and insulin levels in animals fed a Mediterranean diet were not correlated to the bioenergetic capacity in iPFCm, iERCm, or iCBm (Tables 3 and 4).

Table 3.

In animals fed the Western diet, fasting blood glucose levels are related to bioenergetics in PFC and ERC

	Pearson’s r Values of Fasting Glucose vs. Bioenergetic O2 Flux (P Value)*†
Bioenergetic Parameters	PFC		ERC		CB
	Med	West	Med	West	Med	West
CI	0.010 (0.98)	0.77 (0.016)	−0.21 (0.59)	0.58 (0.10)	−0.098 (0.80)	0.034 (0.93)
CI + CII	0.087 (0.82)	0.65 (0.061)	0.076 (0.85)	0.32 (0.41)	0.086 (0.83)	−0.021 (0.96)
LEAK	−0.032 (0.94)	0.26 (0.50)	0.037 (0.92)	−0.093 (0.81)	0.033 (0.93)	−0.18 (0.65)
Max1	0.15 (0.71)	0.62 (0.078)	0.027 (0.95)	0.33 (0.38)	0.16 (0.68)	−0.043 (0.91)
CII(E)	0.10 (0.79)	0.53 (0.14)	0.081 (0.84)	0.20 (0.61)	−0.18 (0.65)	−0.030 (0.94)
FAO	−0.095 (0.82)	0.72 (0.030)	−0.012 (0.76)	0.26 (0.50)	−0.21 (0.58)	0.13 (0.74)
FAO + CI	−0.057 (0.89)	0.72 (0.030)	−0.026 (0.95)	0.39 (0.31)	−0.017 (0.97)	0.045 (0.91)
FAO + CI + CII	0.075 (0.85)	0.61 (0.079)	0.0020 (0.96)	0.33 (0.39)	0.28 (0.46)	0.084 (0.83)
FAO + CI + CII + GpDH	0.92 (0.81)	0.58 (0.10)	0.049 (0.90)	0.30 (0.43)	0.32 (0.40)	0.094 (0.81)
Max2	0.018 (0.96)	0.59 (0.098)	−0.033 (0.93)	0.28 (0.47)	0.19 (0.62)	0.019 (0.96)
CII + GpDH(E)	0.11 (0.77)	0.48 (0.20)	0.027 (0.94)	0.19 (0.63)	0.26 (0.50)	−0.061 (0.88)

CB, cerebellum; CI, Complex I respiration; CII, Complex II respiration; CII(E), Complex II ETS respiration; ERC, entorhinal cortex; ETS, electron transfer system; FAO, respiration due to fatty acid oxidation; GpDH, glycerol-3-phosphate dehydrogenase; LEAK, LEAK respiration; Max, maximal ETS respiration; Med, Mediterranean diet; PFC, prefrontal cortex; ROX, residual oxygen consumption; West, Western diet.

^*P ≤ 0.05 is considered significant and P ≤ 0.10 is considered trending toward significance.

^†n = 9 (PFC Med FAO and FAO-CI n = 8).

Table 4.

In animals fed a Western diet, fasting blood insulin levels are related to bioenergetics in PFC and ERC

	Pearson’s r Values of Fasting Insulin vs. Bioenergetic O2 Flux (P Value)*†
Bioenergetic Parameters	PFC		ERC		CB
	Med	West	Med	West	Med	West
CI	−0.13 (0.73)	0.41 (0.32)	−0.0022 (0.995)	0.48 (0.23)	−0.033 (0.93)	−0.012 (0.98)
CI + CII	−0.13 (0.74)	0.58 (0.13)	0.10 (0.80)	0.55 (0.16)	−0.12 (0.75)	0.27 (0.53)
LEAK	−0.022 (0.96)	0.85 (0.0076)	0.020 (0.96)	0.62 (0.099)	−0.33 (0.39)	0.46 (0.25)
Max1	−0.035 (0.93)	0.58 (0.13)	0.12 (0.77)	0.52 (0.18)	−0.098 (0.80)	0.24 (0.58)
CII(E)	−0.11 (0.78)	0.70 (0.054)	0.13 (0.75)	0.58 (0.13)	−0.18 (0.65)	0.35 (0.39)
FAO	−0.39 (0.33)	0.42 (0.30)	−0.015 (0.97)	0.84 (0.0094)	−0.21 (0.59)	0.19 (0.66)
FAO + CI	−0.34 (0.40)	0.35 (0.39)	0.088 (0.82)	0.59 (0.12)	−0.14 (0.71)	0.14 (0.73)
FAO + CI + CII	−0.20 (0.60)	0.60 (0.11)	0.14 (0.72)	0.62 (0.10)	−0.23 (0.55)	0.30 (0.46)
FAO + CI + CII + GpDH	−0.21 (0.58)	0.64 (0.089)	0.17 (0.66)	0.69 (0.058)	−0.22 (0.56)	0.36 (0.38)
Max2	−0.21 (0.59)	0.67 (0.071)	0.13 (0.73)	0.70 (0.054)	−0.24 (0.54)	0.37 (0.37)
CII + GpDH(E)	−0.27 (0.48)	0.72 (0.046)	0.19 (0.62)	0.73 (0.042)	−0.18 (0.64)	0.41 (0.31)

CB, cerebellum; CI, Complex I respiration; CII, Complex II respiration; CII(E), Complex II ETS respiration; ERC, entorhinal cortex; ETS, electron transfer system; FAO, respiration due to fatty acid oxidation; GpDH, respiration due to glycerol-3-phosphate dehydrogenase; LEAK, LEAK respiration; Max, maximal ETS respiration; Med, Mediterranean diet; PFC, prefrontal cortex; ROX, residual oxygen consumption; West, Western diet.

*P ≤ 0.05 is considered significant and P ≤ 0.10 is considered trending toward significance.

†n = 9 (PFC Med FAO and FAO-CI n = 8).

Fasting blood glucose and insulin levels had a significant (P ≤ 0.05) or trending (P ≤ 0.10) positive correlation with iPFCm and iERCm bioenergetics (Tables 3 and 4). Measures of CI, CI + CII, Max1, FAO, FAO + CI, and Max2 parameters correlated with fasting blood glucose levels in iPFCm and measures of the CI parameter correlated with fasting blood glucose levels in iERCm. Fasting blood glucose levels in animals fed the Western diet did not correlate with any bioenergetic parameters in iCBm. Measures of LEAK, Complex II ETS respiration [CII(E)], FAO + CI + CII + GpDH, Max2, and CII + GpDH(E) parameters correlated with fasting blood insulin levels in iPFCm and measures of LEAK, FAO, FAO + CI + CII, FAO + CI + CII + GpDH, Max2, and CII + GpDH(E) correlated with fasting blood insulin levels in iERCm. Fasting blood insulin in animals fed the Western diet did not correlate with any bioenergetic parameters in iCBm.

Differences in Mitochondrial Complex Proteins and Citrate Synthase Activity between Diets

Next, we examined the expression of proteins representing the five major mitochondrial complexes (Complex I, Complex II, Complex III, Complex IV, and Complex V) normalized to VDAC/Porin. We also looked at VDAC/Porin expression normalized to cellular control (gapdh) to determine if there was difference in mitochondrial content due to diet and between brain regions. We found that in the PFC, Complex III and Complex V were increased in animals fed the western diet as compared with animals fed the Mediterranean diet. In the ERC, Complex I, Complex III, Complex IV, and Complex V all had decreased expression in animals fed the Western diet as compared with animals fed the Mediterranean diet indicating different effects of diet in different brain regions. The CB showed no differences in expression of any of the complexes. Also of note, Complex II and VDAC/Porin showed no differences between diets for any of the regions (Fig. 4, Supplemental Table S4, and Supplemental Fig. S1). The lack of differences in VDAC/Porin suggests there is no difference in mitochondrial content between diets within each brain region. There was also no difference in citrate synthase activity between diets.

Figure 4.

Comparing PFC, ERC, and CB mitochondrial complexes, VDAC/Porin, and citrate synthase activity from brain tissue of female cynomolgus macaques fed either a Mediterranean or Western diet. A–E: abundance of protein normalized to VDAC/Porin. F: abundance of protein normalized to GAPDH. G: citrate synthase activity. H: images of Western blot bands for each complex. A mixed-effects model of analysis in SAS software, version 9.4, was used to evaluate the effect of the Mediterranean diet compared with the Western diet. Each point represents the difference between the regions in one animal. *P ≤ 0.05, **P ≤ 0.01, and ****P ≤ 0.0001. Exact P values are recorded in Supplemental Table S4. Full representative blot is presented in Supplemental Fig. S1. Data are presented as means ± SE, n = 7 for Mediterranean diet for all brain regions, n = 6 for Western diet for PFC and ERC, and n = 7 for Western diet for CB (due to extra sample availability). Lanes are cut from original image to present in a logical order. None of the images were manipulated in any way. CB, cerebellum; ERC, entorhinal cortex; Med, Mediterranean diet; PFC, prefrontal cortex; West, Western diet.

DISCUSSION

We previously reported skeletal muscle fiber bioenergetic capacity in this animal cohort is significantly elevated in animals administered a Western diet compared with a Mediterranean diet (35). Although the data presented here from brain mitochondria only show statistically significant elevated respiration with Western diet in 3 of the 33 bioenergetic parameters, the remaining 29 parameters were still elevated compared with animals on the Mediterranean diet (Fig. 2). These data suggest that Western diets may systemically elevate bioenergetic capacities, but the duration of 30 mo may not have been long enough to elicit robust elevations in brain bioenergetic capacity, however, to a different extent in different tissues. Skeletal muscle bioenergetics may be more readily affected by diet and a duration of 30 mo. Importantly, skeletal muscle bioenergetics was measured in permeabilized muscle fiber bundles, whereas this study used isolated mitochondria. Isolated mitochondria assays are normalized to protein concentration and negate potential differences in mitochondrial content that may be present at the tissue level. Moreover, isolated mitochondria lose the context of the cellular environment and therefore primarily report bioenergetic differences driven by the electron transfer system (36). Altogether, these can contribute to differences in the findings we report between skeletal and brain bioenergetics.

It is well accepted that there are intrinsic regional differences in mitochondrial function, enzyme activity, protein expression levels, number, and distribution in both healthy and diseased brains (37). Our results, similar to a previous study (38), highlight that in nondiseased animals administered a healthy Mediterranean diet, the brain maintains distinct differences in bioenergetic capacity between brain regions. Unique to this study, we examined the bioenergetics of brain mitochondria subject to a Western diet condition. The parent study of these animals documented peripheral metabolic dysregulation in the form of elevated insulin resistance and triglyceride levels in animals fed the western diet (26). Beyond the experiments with this animal cohort, the Western diet is well documented to promote metabolic dysfunction in the periphery and central nervous system (39–42). The animals administered a Western diet in this study had diminished differences in bioenergetic capacity between brain regions compared with the animals administered the Mediterranean diet with the exception of respiration driven by FAO (Fig. 3F and Supplemental Table S3).

These results suggest that factors influenced by diet may be regulating or influencing patterns in bioenergetic capacity. We hypothesized that circulating blood glucose and insulin levels could be potential mediators of bioenergetic function in animals administered the Western diet. Insulin resistance due to diet effects has been mechanistically linked to mitochondrial health in animals and insulin and glucose homeostasis are disrupted by diet, particularly high-fat and Western-style diets (40, 43, 44). In support of this hypothesis, our results indicate that fasting blood glucose and insulin levels relate to brain mitochondrial bioenergetics in animals fed a Western diet but not in animals fed a Mediterranean diet. In particular, we found that bioenergetic relationships with glucose and insulin levels were restricted to mitochondria isolated from the cortical brain regions (iPFCm and iERCm), whereas iCBm bioenergetics did not relate to glucose or insulin levels. As previously mentioned, the PFC and ERC are vulnerable to metabolic perturbance and also show mitochondrial dysfunction in early disease (45, 46). Interestingly, the relationships are not mediated by elevations in fasting glucose or insulin, as the animals in both diet groups had comparable levels of fasting glucose and insulin. Rather, we hypothesize that the observed relationships with glucose and insulin to bioenergetics in animals administered a Western diet may be due to increased sensitivity to glucose and other substrates in these metabolically vulnerable brain regions in animals administered a Western diet. Although glucose and insulin sensitivity is diminished in diabetes progression and is a risk factor for Alzheimer’s disease (47), these animals are not on the diabetic spectrum. Thus, the observed sensitivity is apparent before any disease pathology. Furthermore, it should be noted that this study is focused on intrinsic ETS activity on activation by various fuels, measured using isolated mitochondria. Alternatively, the Mediterranean diet may have had a protective effect that blunted the influences of glucose and insulin on mitochondrial bioenergetic capacity. Altogether, these results suggest that circulating glucose and insulin influence mitochondrial bioenergetics and physiology in animals consuming a Western diet and may prime the mitochondria to use alternative sources of fuel not regulated by glucose and insulin levels.

Contextual environment of the mitochondria dictates their energetic output and ensures bioenergetics remain in tune with regional demands (37). Loss of region-specific differences in bioenergetics in animals fed the Western diet is related to fasting glucose and insulin levels that influence the bioenergetic output of mitochondria in animals fed a Western diet in the PFC and ERC, two regions susceptible to decline in Alzheimer’s disease. Therefore, homeostatic bioenergetic needs of the cells may not be controlling the bioenergetic output in these brain regions. If the intrinsic energy needs of the cell are not precisely met, this could lead to disruptions in normal cell function. With time, this repeated diversion from homeostatic levels of energy production could damage cells, either by not providing them with enough energy or over providing and causing excess ROS production to further damage the cell via mitochondrial Complex I and Complex III LEAK (48). Although this study takes place in midlife, dysregulation of homeostatic energy production is evident and may cause metabolic disruptions. Later in life, the processes for repairing damage become less robust and cells may not be able to recover, leading to cognitive deficits (49).

The only bioenergetic parameter that did not lose distinct regional differences in bioenergetic capacity with animals administered the Western diet was respiration driven by FAO. In addition to producing acetyl-CoA for processing in the TCA cycle and producing NADH to donate electrons to the ETS via Complex I, FAO also donates electrons to the ETS via the electron-transferring flavoprotein complex (cETF) (34). Unlike the other pathways evaluated in this study, the cETF does not rely on TCA cycle metabolism or TCA cycle products to deliver electrons to the ETS (34). FAO contribution through electrons donated by the cETF makes up only 5%–15% of coupled respiration in this study depending on the brain region in our study (Supplemental Fig. S2) indicating, and consistent with literature, that glucose pathways are the preferred source of energy production (50, 51). Nevertheless, studies indicate that a shift in brain metabolism from glucose OXPHOS metabolism to fatty acid OXPHOS metabolism paired with glycolysis is apparent in aged and Alzheimer’s disease-affected brains (52–54). In animals administered a Western diet, the fatty acid OXPHOS metabolism maintained the same distinct differences in bioenergetic capacity between brain regions as the animals administered the Mediterranean diet whereas all other bioenergetic parameters had diminished differences in bioenergetic capacity. This suggests that regional patterns of FAO metabolism are preserved compared with glucose metabolism in animals administered the Western diet, potentially allowing fatty acids to become an alternative source of fuel as glucose metabolism becomes dysregulated.

There were no differences in citrate synthase activity, a marker of mitochondrial content/volume, or VDAC/Porin abundance between diets (Fig. 4, Supplemental Table S4, and Supplemental Fig. S1), in line with the muscle fiber bundle respirometry from these animals (35). However, there were differences in abundance of ETS proteins between diets. In the PFC, Complex III and Complex V had elevated expression in animals fed the Western diet as compared with animals fed the Mediterranean diet. Conversely, in the ERC, several complexes had decreased expression of ETS proteins in animals fed the Western diet as compared with animals fed the Mediterranean diet. There were no differences in the abundance of proteins from mitochondrial complexes between the diets in the CB. Interestingly, there was no difference in Complex II expression between diets for any of the brain regions. Citrate synthase and Complex II are both part of the TCA cycle, indicating differences in respirometry or respirometric patterns and regulation between regions is due to the ETS system. Additionally, no differences in vdac between diets indicates there is similar mitochondrial content, but the abundance of ETS proteins can vary despite no differences in overall mitochondrial content. This could indicate that any trend toward increased respiration may be related to the structure or quality of mitochondria. Overly nutrient-rich environments, such as in Western diets, are characterized by fragmented mitochondria, inefficient ATP production, and excess ROS production (55). The proteins from the complexes that have differences in expression are proteins particularly active in ROS production (56). Differences in protein expression from the ROS-producing complexes between a health-promoting and metabolic disease-promoting diets indicate that there may be dysregulation in ROS production as well. Complex V (ATP synthase) abundance is also altered in the PFC and ERC indicating ATP production is likely altered in the Western diet as compared with the Mediterranean diet. Long-term availability of ATP could hinder cellular processes or change the preferred energy-producing method (57), thereby further altering metabolism.

A design element to consider is that the diets in this study were macronutrient and cholesterol content matched. The source and composition of the fat were different between diets but fat percentage between diets is comparable and the diets would not be classified as “high fat” (58). It is important to remember that the terms “high-fat diet” and “Western diet” are often used interchangeably. The results presented in this study are due to differences in diet composition and perhaps likely the differences in fatty acid composition, not differences in fat percentage.

There are limitations to this study that should be considered when interpreting the results. First, fasting blood glucose and insulin levels were measured after 26 mo of diet administration. All respirometric assessments of brain mitochondria were performed on study completion and necropsy at 30 mo. This difference in timing may influence relationships between mitochondrial bioenergetics and fasting blood glucose or insulin levels. Second, SUIT Protocol 1 induced LEAK respiration with the addition of oligomycin. Oligomycin is known to decrease uncoupled respiration (59). The resulting measures of MAX1 and CII(E) bioenergetic capacities are therefore likely lower than if oligomycin was not included in the assay. We treated all samples the same to allow comparisons between animals while acknowledging this limitation related to the measurement of LEAK respiration.

The findings presented in this manuscript impact our current understanding of brain metabolism, particularly as it relates to risk factors for Alzheimer’s disease. Indeed, patients with type 2 diabetes are at greater risk for Alzheimer’s disease and cognitive impairment (60). The work presented here highlights two key circulating factors that are dysregulated in diabetes, glucose and insulin, as they relate to dysregulation of cortical mitochondrial bioenergetics in brain regions vulnerable to neurodegeneration. These results were driven by the administration of a Western diet known to promote development of diabetes and poor cognitive outcomes, but not with the administration of a Mediterranean diet associated with positive glycemic and cognitive outcomes. Interestingly, fatty acid-driven bioenergetics is not different between the diets, highlighting perhaps that as glucose-driven bioenergetics dysregulates, fatty acid metabolism remains intact and becomes an alternative fuel source to meet cellular demand for energy within the brain. Future work is needed to confirm these findings in other animal models and to examine the driving factors behind increased sensitivity to insulin and glucose in cortical mitochondria before any disease onset.

SUPPLEMENTAL DATA

Supplemental Tables S1–S4 and Fig. S1: https://doi.org/10.6084/m9.figshare.14529459.v8.

GRANTS

This work was supported by National Institutes of Health Grants R01 AG054523 (to A.J.A.M.), R01 AG061805 (to A.J.A.M.), R56 AG057864 (to A.J.A.M.), R01 HL087103 (to C.A.S.), R01 HL122393 (to T.C.R.), T32 GM127261, the Wake Forest Claude D. Pepper Older Americans for Independence Center (P30 AG021332), and the Wake Forest Alzheimer’s Disease Research Center (P30 AG049638).

DISCLOSURES

A.J.A. Molina is a board member of Aeva Bioscience. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

S.C., T.C.R., C.A.S., and A.J.A.M. conceived and designed research; K.A.A., G.M., and Z.G. performed experiments; K.A.A. and J.B. analyzed data; K.A.A. interpreted results of experiments; K.A.A. prepared figures; K.A.A. drafted manuscript; K.A.A., G.M., J.B., C.A.S., and A.J.A.M. edited and revised manuscript; K.A.A., G.M., J.B., and A.J.A.M. approved final version of manuscript.