Insulin deficiency and intranasal insulin alter brain mitochondrial function: a potential factor for dementia in diabetes

Abstract

Despite the strong association between diabetes and dementia, it remains to be fully elucidated how insulin deficiency adversely affects brain functions. We show that insulin deficiency in streptozotocin-induced diabetic mice decreased mitochondrial ATP production and/or citrate synthase and cytochrome oxidase activities in the cerebrum, hypothalamus, and hippocampus. Concomitant decrease in mitochondrial fusion proteins and increased fission proteins in these brain regions likely contributed to altered mitochondrial function. Although insulin deficiency did not cause any detectable increase in reactive oxygen species (ROS) emission, inhibition of monocarboxylate transporters increased ROS emission and further reduced ATP production, indicating the causative roles of elevated ketones and lactate in counteracting oxidative stress and as a fuel source for ATP production during insulin deficiency. Moreover, in healthy mice, intranasal insulin administration increased mitochondrial ATP production, demonstrating a direct regulatory role of insulin on brain mitochondrial function. Proteomics analysis of the cerebrum showed that although insulin deficiency led to oxidative post-translational modification of several proteins that cause tau phosphorylation and neurofibrillary degeneration, insulin administration enhanced neuronal development and neurotransmission pathways. Together these results render support for the critical role of insulin to maintain brain mitochondrial homeostasis and provide mechanistic insight into the potential therapeutic benefits of intranasal insulin.—Ruegsegger, G. N., Manjunatha, S., Summer, P., Gopala, S., Zabeilski, P., Dasari, S., Vanderboom, P. M., Lanza, I. R., Klaus, K. A., Nair, K. S. Insulin deficiency and intranasal insulin alter brain mitochondrial function: a potential factor for dementia in diabetes.

Despite contributing ∼2% of body mass in a lean person, the brain consumes 20–25% of the body’s oxygen consumption, indicating high energy need. Although glucose is the predominant fuel, the limited glycolytic capacity renders neurons highly dependent on mitochondrial respiration for ATP generation for all their cellular functions (1). Moreover, impaired mitochondrial function has been implicated in the pathogenesis of neurodegenerative disease, such as Alzheimer’s disease (AD) (2). Although dementia has been more frequently associated with insulin resistance and type 2 diabetes (3), type 1 diabetes (T1D) also increases dementia risk (4); however, the mechanistic underpinnings of how reduced insulin action or deficiency contribute to cognitive impairment and AD are incompletely understood (5). It is possible that reduced insulin action in the presence of low/normal/high insulin in insulin-resistant people, including those with type 2 diabetes, may differ from absolute insulin deficiency that occurs in T1D. Here, we sought to determine whether insulin deficiency has any impact on mitochondrial function, specifically in areas of the brain rich in insulin receptors and involved in memory.

Although insulin is not necessary for glucose metabolism in brain, the brain has an extensive distribution of insulin receptors (6), and the role of insulin action in the brain influencing peripheral metabolism and behavioral functions has been reported (7). These observations have led to intranasal insulin being investigated as a possible therapy for cognitive impairment (8). In peripheral tissues, such as skeletal muscle (muscle), insulin deficiency has been shown to profoundly alter mitochondrial function, especially causing decreased ATP production and increased reactive oxygen species (ROS) emission, in mice (9, 10). Because insulin is a key regulator of mitochondrial biogenesis and function in muscle (11–13), it is possible that, as in muscle (9, 10), insulin deficiency may decrease mitochondrial ATP production and increase ROS emission in brain. Reduced availability of ATP adversely affects cellular functions, and excessive ROS damages the proteome. Brain tissue, like muscle, is mostly postmitotic and thus susceptible to accumulate protein damages, specifically to mitochondrial proteins. However, whether insulin deficiency influences mitochondrial function in brain remains to be elucidated.

Thus, we determined whether insulin deprivation in streptozotocin (STZ)-induced diabetic mice on insulin treatment causes declines in brain mitochondrial function compared with nondiabetic (ND) control mice. We hypothesized that insulin deficiency would decrease mitochondrial respiration and ATP production concomitant to increased ROS emission, especially in regions known to be rich in insulin receptors. Because brain also uses ketones and lactate as sources of fuel for energy needs, we also evaluated how inhibition of monocarboxylate transporters (MCTs), which act as transporters for ketones and lactate, influence mitochondrial biology during insulin deficiency. Furthermore, we determined the effect of intranasal insulin administration on ATP production to directly address insulin’s effect on brain mitochondria as reported in muscle (13). We also performed proteomic a survey to elucidate potential impacts of insulin deficiency and intranasal insulin on proteome homeostasis that may explain the regulatory role of insulin on the brain.

MATERIALS AND METHODS

Animals

The Mayo Clinic Institutional Animal Care and Use Committee approved all procedures. Male, 14- to 16-wk-old male C57BL/6J mice (The Jackson Laboratory, Bar Harbor, ME, USA) were acclimatized for 1 wk prior to all experiments. Water and chow (PicoLab 5053; Lab Supply, Fort Worth, TX, USA) were provided ad libitum.

STZ-induced diabetes was achieved as previously described (10). Briefly, mice were administered intraperitoneal injections of STZ (125 mg/kg) on 2 consecutive days. Five days following STZ injection, mice received subcutaneous insulin implants (LinShin; Toronto, ON, Canada) at 10 U/kg/d. After 3 wk of treatment, mice either continued treatment with insulin implants (STZ + insulin) or insulin implants were removed leading to the return of the diabetic phenotype (STZ + vehicle). This approach allows for the stabilization of body weight and blood glucose and ensures that the observed effects are the result of insulin deficiency rather than the toxic effects of STZ. After an additional 96 h, mice were euthanized by pentobarbital overdose. ND animals followed a similar protocol and received intraperitoneal sodium acetate injection rather than STZ and received blank implants. Depending on the analysis, the cerebrum, hippocampus, hypothalamus, or midbrain was excised per a mouse brain atlas (14). Plasma and cerebrum β-hydroxybutyrate (βHB) was measured using a colorimetric assay (Cayman Chemical; Ann Arbor, MI, USA).

A separate cohort of ND and STZ + vehicle mice were treated intranasally with vehicle (20% DMSO + 0.9% saline) or the MCT inhibitor α-cyano-4-hydroxy-cinnamic acid (4-CIN) (MilliporeSigma, Burlington, MA, USA) twice per day during the 96 h period between diabetes reestablishment and euthanasia. For awake intranasal delivery, the mouse’s head was restrained in a supine position with the neck in extension, as previously described (15). A total of 2 µg/20 μl 4-CIN or 20 μl vehicle was delivered per injection over both nares in 5 μl drops into alternating nostrils. An extra 5-μm drop was given if the subject forcibly ejected solution. Mice were held supine for 30 s after delivery to ensure all fluid was inhaled.

Additionally, ND mice were administered vehicle (PBS + 0.5% bovine serum albumin) or insulin (Humulin R; Eli Lilly and Co., Indianapolis, IN, USA) intranasally twice per day for 14 d. A total of 2 U/20 μl insulin or 20 μl vehicle was delivered as previously described. Body composition was measured by MRI (EchoMRI, Houston, TX, USA), and plasma glucose and plasma insulin were determined by ELISA (Crystal Chem, Elk Grove Village, IL, USA).

Brain mitochondrial isolation

Mitochondria were isolated from the right and left cerebral hemispheres (with subcortical areas), as previously described (16). Briefly, brain was weighed and transferred into a tube containing 1 ml of ice-cold isolation medium (IM) and gently rinsed. The buffer was removed and IM containing protease was added (1 ml/100 mg tissue), and the tissue was minced and homogenized for 10 min. Homogenate was further diluted 1:4 with IM, centrifuged at 2000 g for 4 min, the resultant supernatant passed through cheesecloth, the filtrate collected and centrifuged at 9000 g for 10 min. The resulting pellet was suspended in 6 ml of IM containing 0.02% digitonin, homogenized for 10 min, and centrifuged at 9000 g for 5 min. The resultant pellet was washed with 1 ml of IM and centrifuged again at 9000 g for 5 min. The final pellet was resuspended in 125 μl of IM/100 mg of tissue.

Mitochondrial oxygen consumption and ROS production

Mitochondrial respiration and H₂O₂ production were measured simultaneously using Oxygraph-O2K-Fluorescence LED2-Module (Oroboros Instruments, Innsbruck, Austria), as previously described (16, 17). Oxygen consumption rate (OCR) and ROS production were measured in a 50 μl aliquot of isolated mitochondria suspension added to each 2 ml Oxygraph chamber and allowed to equilibrate. Mitochondrial respiration was measured devoid of substrates (state 1); in the presence of 10 mM glutamate, 2 mM malate, and 10 mM succinate (state 2); and 2.5 mM ADP (state 3). This was followed by addition of 2 μg/ml oligomycin to inhibit ATP synthase activity and induce state 4 respiration. Finally, 2.5 μM antimycin A was added to inhibit mitochondrial oxygen consumption and measure residual oxygen consumption. Mitochondrial H₂O₂ production was measured by continuous monitoring of Amplex Red oxidation (ThermoFisher Scientific, Waltham, MA, USA). Protein content from isolated mitochondria was determined using the DC Protein Assay (Bio-Rad Laboratories, Hercules, CA, USA).

Mitochondrial ATP production rate

Mitochondrial ATP production was measured using an enzymatic system containing hexokinase and glucose-6-phosphate dehydrogenase to convert ATP to NADPH through sequential formation of glucose-6-phaospahate and 6-phosphoglucolactone using glucose and NADP⁺ as previously described (18, 19). A Fluorolog 3 (Horiba Scientific, Piscataway, NJ, USA) spectrofluorometer was used to continuously measure the autofluorescence of NADPH. Ten microliters of isolated mitochondria suspension was added to a quartz cuvette with 2 ml of buffer Z containing (in millimolars) 110 K-MES, 35 KCl, 1 EGTA, 5 K₂HPO₄, 3 MgCl₂-6H₂O, and 5 mg/ml bovine serum albumin (pH 7.4, 295 mOsm) and 2.5 mM D-Glucose. The same stepwise titration protocol was used to induce states 1, 2, 3, and 4 as previously described. OCR, ROS emission, and ATP production were normalized per milligram of mitochondrial protein (reflective of mitochondrial protein quality) and per tissue wet weight (reflective of mitochondrial content).

Mitochondrial enzyme activities

Citrate synthase (CS) and cytochrome c oxidase (COX) activities were determined as previously described (13). Mitochondrial superoxide dismutase 2 (SOD2) activity was determined spectrophotometrically from the consumption of xanthine oxidase-generated superoxide radical by endogenous SOD2 (Cayman Chemical). Catalase (CAT) activity was determined spectrophotometrically by measuring peroxide removal (Cayman Chemical).

Insulin concentration determination

Frozen cerebral tissue was homogenized with 10 times w/v RIPA buffer with protease and phosphatase inhibitors (Roche, Basel, Switzerland). Homogenates were centrifuged at 14,000 rpm for 30 min at 4°C. The supernatant was collected and used as tissue lysate. Cerebral tissue lysate and plasma insulin concentrations were measured by ELISA according to the manufacturer’s recommendations (Crystal Chem).

Immunoblotting

Western blotting was performed as previously described (10). Primary antibodies were applied to membranes overnight at 4°C. Appropriate secondary antibody (Thermo Fisher Scientific) was applied for 1 h at room temperature, and proteins were detected by infrared fluorescence (Odyssey; Li-Cor Biosciences, Lincoln, NE, USA). Antibody dilutions were phosphorylated (p)AKT-Ser473 (1:1000; 9271; Cell Signaling Technology, Danvers, MA, USA), pGSK3B-Ser9 (1:1000; 9336; Cell Signaling Technology), AKT (1:2000; 9272; Cell Signaling Technology), GSK3B (1:2000; 9315; Cell Signaling Technology), dynamin-related protein 1 (DRP1; 1:1000; 101270; Santa Cruz Biotechnology, Dallas, TX, USA), mitofusin 1 (MFN1; 1:1000; PA5-67905; Thermo Fisher Scientific), MFN2 (1:1000; no. PA5-42171; Thermo Fisher Scientific), optic atrophy (OPA1; 1:1000; no. 59770; Novus Biologicals, ‎Littleton, CO, USA), oxidative phosphorylation cocktail (1:1000; no. MS604; Mitosciences, Eugene, OR, USA), peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α; 1:1000; no. 516557; MilliporeSigma), and β-actin (1:5000; no. 926-42212; Li-Cor Biosciences).

mRNA expression

RNA isolation was performed using an RNeasy Mini Kit (Qiagen, Venlo, The Netherlands) and cDNA synthesis was performed using the SuperScript IV First-Strand Synthesis System for RT-PCR (Thermo Fisher Scientific). Quantitative RT-PCR was performed in 384-well clear plates with 10 µl reaction volume using SYBR Green (Applied Biosystems, Foster City, CA, USA). Amplification conditions were 10 min at 60°C followed by 40 cycles of denaturing (95°C for 15 s) and annealing (60°C for 60 s) using a ViiA7 thermocycler (Applied Biosystems). Samples were amplified in duplicate with gene-specific primers (Table 1). Each plate included a repeated control on the plate (intra-assay control) and between plates (interassay control) along a no-template control, and 7-point relative standard curve spanning 4 log dilutions. mRNA expression values were quantified by the 2ΔΔCt method (20), whereby ΔCt = Gapdh Ct − target gene Ct.

TABLE 1.

Quantitative RT-PCR primers for gene expression analysis

Gene	Sequence, 5′–3′
	Forward	Reverse
Gapdh	ACCCTTAAGAGGGATGCTGC	GTTCACACCGACCTTCACCA
Ins2	TCTTCTACACACCCATGTCCC	GGTGCAGCACTGATCCAC
Mct1	GCAGTGTTAGTCGGAGCCTT	TCACTGGTCGTTGCACTGAA
Mct2	TGCCCTTGGTTACTTCGTCC	CCAATGCACACCAAGAGCAC
Mct4	CTGGACCACCAATAGCAGGC	TGCGTTTCCATCTTTCAGCC

mtDNA quantification

DNA was extracted using a DNA mini kit (Qiagen) and analyzed for mtDNA copy number using ND1 and ND4 genes. The abundance of each target gene was normalized to 28S ribosomal DNA and assayed in duplicate by quantitative PCR (qPCR) using a ViiA7 thermocycler (Applied Biosystems), as previously described (21). Gene-specific primers are shown in Table 2.

TABLE 2.

Quantitative PCR primers used to assess mtDNA copy number

Gene	Sequence, 5′–3′
	Forward	Reverse
28S	TGGGAATGCAGCCCAAAG	CCTTACGGTACTTGTTGACTATCG
Nd1	AAGGAGAATCAGAATTAGTATCAGGGTT	TAGTACTCTGCTATAAAGAATAACGCGAAT
Nd4	TCCAACTACGAACGGATCCA	AAGTGGGAAGACCATTTGAAGTC

Lipid metabolites

Cerebral sphingolipids (SPH) were measured in ND and STZ + vehicle mice as previously described (22). Briefly, quantitative measurement of SPHs (C14:0-Cer, C16:0-Cer, C18:1-Cer, C18:0-Cer, C20:0-Cer, C22:0-Cer, C24:0- Cer, C24:1-Cer) were made using a Thermo TSQ Quantum Ultra mass spectrometer (Thermo Fisher Scientific) using positive ion electrospray ionization source with selective reaction monitoring. Chromatographic separation was performed using an Acquity Ultra Performance Liquid Chromatography (Waters, Milford, MA, USA).

NMRS analysis

Approximately 30 mg of cerebral tissue was pulverized and 100 µl of 6% HClO₄ was added. Tissue was ground for 30 s with a hand homogenizer, 200 µl 6% HClO₄ was added, and the mixture was shaken on vortex and frozen in liquid nitrogen. The sample was thawed and spun down at 10,000 g for 15 min. Supernatant was collected and the pellet was re-extracted with 100 µl of 6% HClO₄. Combined extracts were neutralized with 140 µl of 2 M KHCO₃. The mixture was spun down at 10,000 g for 15 min and supernatant (500 µl) was collected. One hundred milliliters of phosphate buffer (pH 7.4) and 50 ml of TSP-d4 solution in D₂O (1 mM) were added, and the sample was transferred to a 5 mm NMR tube.

NMR spectra were acquired on a Bruker 600 MHz Avance III HD spectrometer equipped with a BBI room temperature probehead and SampleJet auto sampler (Bruker, Billerica, MA, USA). [¹H] NMR spectra were recorded using 1D noesy pulse sequence (noesygppr1d), with calibrated 90° pulse (∼11 ms) and the following conditions: NS = 128; TD = 64 k; SW = 14 ppm; T = 298.2 K; AQ = 3.9 s; D1 = 5 s. Spectra were phase and baseline corrected using TopSpin v.3.5 software (Bruker). Metabolites were identified and quantified using Chenomx NMR suite v.8.3 software (Chenomx, Edmonton, AB, Canada) by fitting spectral lines of library compounds into the NMR spectra of tissue extract. Quantification was based on peak area of TSP-d4 signal. Metabolite concentrations were exported as micromolars in the NMR sample and recalculated as micromole per grams of tissue.

Proteomics analysis

Proteins were prepared for liquid chromatography–tandem mass spectrometry (MS/MS) (23). Label-free liquid chromatography–MS/MS data were acquired using a high-resolution nanoLC LTQ Orbitrap (Thermo Fisher Scientific). MaxQuant software (v.1.5.1.2; https://www.biochem.mpg.de/5111795/maxquant) was configured to match MS/MS against the RefSeq mouse sequence database (v.58; https://www.ncbi.nlm.nih.gov/refseq/) and identify peptides and proteins at 1% false discovery rate. Proteins with a corrected P value of ≤0.05 and absolute log₂ fold change of ≥0.5 were considered differentially expressed. Sample level protein oxidation was quantified from spectral counts (24). Proteomics data were compared with the MitoCarta 2.0 mouse mitochondrial protein database (25) and AlzGene database of AD-related proteins (26). Pathway analysis was performed using IPA software (Qiagen) (23). Fisher’s exact test was used to identify significantly affected pathways, and P values were expressed as –log₁₀ derivatives (e.g., P > 1.301 on –log₁₀ scales corresponds to P < 0.05).

Statistics

Results are presented as means ± sem. Significance was set at P < 0.05. Data comparing 2 groups were assessed by 2-tailed unpaired Student’s t test. When a single variable was compared across more than 2 groups, 1-way ANOVA was used. Repeated-measures 1-way ANOVA was used to assess within-group changes over time. To assess the effects of STZ, MCT inhibition, and their interactions, 2-way ANOVA was performed. Multiple comparisons Tukey test was used for post hoc analyses to determine significant differences between individual groups when appropriate. Analyses were conducted using JMP 10 (SAS Institute, Cary, NC, USA).

RESULTS

Phenotype features and brain metabolites

STZ-injected mice became hyperglycemic by the fifth d postinjection. Insulin treatment normalized blood glucose and partially restored body weight (Fig. 1A, B). Following 96 h of insulin withdrawal, hyperglycemia returned in STZ + vehicle, whereas STZ + insulin maintained euglycemia. Plasma and cerebral βHB were increased in STZ + vehicle compared with STZ + insulin and ND (Fig. 1C). NMR spectroscopy analysis of cerebral homogenates revealed ∼7-fold greater glucose in STZ + vehicle compared with ND (Fig. 1D). NMR spectroscopy detected increased lactate in STZ + vehicle suggesting increased glycolysis. The tricarboxylic acid cycle metabolites glutamate, glutamine, and fumarate were also elevated in STZ + vehicle. Consistent with increased branched-chain amino acids in plasma during insulin deprivation, valine was increased in STZ + vehicle. Additionally, choline, the precursor for the neurotransmitter acetylcholine, and phosphocholine, which is important for cell membrane integrity, were decreased in STZ + vehicle. Analysis of lipid metabolites showed that C18:0, C24:0, and C24:1 ceramides and sphingosine were increased and C16:0 ceramide tended (P = 0.06) to be increased in cerebral homogenates from STZ + vehicle compared with ND (Fig. 1E).

Figure 1.

A, B) Insulin treatment (STZ + insulin) restored body weight (A) and blood glucose (B) in STZ-induced diabetic mice to levels of ND mice, whereas the removal of insulin treatment (STZ + vehicle) promoted the reestablishment of the diabetic phenotype. C) βHB in plasma and cerebral homogenates was elevated in STZ + vehicle and corrected by STZ + insulin to levels of ND. D) Direct comparison between STZ + vehicle and ND mice showed alterations in metabolites in cerebral homogenates as measured by NMR spectroscopy. E) STZ + vehicle increased ceramide species C18:0, C24:0, and C24:1, and sphingosine compared with ND. F) Ins2 mRNA expression did not differ between groups in the hypothalamus (Hypo), hippocampus (Hippo), or midbrain. G) Insulin concentration in cerebral tissue was decreased in STZ + vehicle: note that 6 of 8 STZ + vehicle samples failed to reach a detection limit of ∼8 ng/g tissue (shown here as 0). H) Plasma insulin concentration was decreased in STZ + vehicle. Data are means ± sem (n = 9/group for body weight and blood glucose; n = 6–7/group for βHB measurements; n = 6/group for NMR spectra analysis and lipid metabolites; n = 5–7/group for mRNA expression analysis; n = 8/group for insulin concentration measurements). *P < 0.05, **P < 0.01, ***P < 0.001.

Prior reports demonstrate that insulin can be produced in the brain through the expression of Insulin 2 (27, 28). Assessment of Insulin 2 mRNA in the hypothalamus, hippocampus, and midbrain showed no differences, suggesting STZ did not influence the local production of insulin in brain (Fig. 1F). In contrast, assessment of insulin in cerebral tissue homogenates showed that insulin concentration was ∼24 ng/g tissue in ND and STZ + insulin mice, whereas insulin levels were below the limit of detection in 6 of 8 samples from STZ + vehicle mice (Fig. 1G). Plasma insulin concentration was also very low in STZ + vehicle mice (Fig. 1H). These data suggest reductions in brain insulin levels in STZ + vehicle are the result of impaired insulin secretion from the pancreas rather than impairments in local insulin production.

Insulin deprivation alters brain mitochondrial function

We hypothesized that insulin deficiency would alter the function of mitochondria isolated from the cerebrum. No differences in OCR were observed between groups, although state 4 (leak) respiration tended to be greater in STZ + vehicle compared with ND (P = 0.056) when normalized to tissue weight (Fig. 2A) thus explaining a decline in the respiratory control ratio [(RCR) state 3/state 4], suggesting decreased mitochondrial coupling efficiency in STZ + vehicle (Fig. 2B). Consistent with the notion of increased proton leak, state 3 mitochondrial ATP production rate (MAPR) was decreased in STZ + vehicle compared with ND and STZ + insulin when normalized to tissue weight (Fig. 2C). Mitochondrial complex V (ATP synthase) protein in cerebral homogenates was lower in diabetic mice during insulin deprivation and was not restored by insulin treatment (Fig. 2D). Although uncoupled respiration is often accompanied by increased mitochondrial ROS emission, no differences in H₂O₂ emission (Fig. 2E) or the percentage of oxygen consumption measured as H₂O₂ emission (Fig. 2F) were detected. Similar results for mitochondrial OCR, MAPR, and ROS emission were observed when normalized to mitochondrial protein content (data not shown). Of interest, endogenous antioxidants SOD2 and CAT activities in cerebral homogenates were increased in STZ + vehicle compared with ND and STZ + insulin (Fig. 2G, H), indicating that insulin deficiency increases antioxidant defense to prevent against excess oxidative stress.

Figure 2.

Insulin deprivations impairs brain mitochondrial physiology. A) Cerebral mitochondrial oxygen consumption (JO₂) did not differ between ND, insulin deficient (STZ + vehicle), and insulin treated (STZ + insulin) mice. B) However, mitochondrial coupling efficiency, as measured by RCR was decreased in STZ + vehicle compared with ND. C) State 3 MAPR was reduced in STZ + vehicle but corrected by insulin treatment. D) Protein expression of mitochondrial respiratory complex V (CV) was decreased in STZ + vehicle and STZ + insulin compared with ND. E–J) Neither mitochondrial H₂O₂ emissions (E) nor oxygen consumption as a percentage of ROS (F) differed between groups. Mitochondrial SOD2 (G) and CAT (H) activities in cerebral homogenates were greater in STZ + vehicle compared with ND and STZ + insulin. CS activity was reduced by insulin deprivation but corrected by insulin treatment in the hypothalamus (Hypo) (I), whereas COX activity was lower in STZ + vehicle compared with ND mice in the hypothalamus and hippocampus (Hippo) (J). Data are means ± sem (n = 9/group for mitochondrial function measures; n = 8 for Western blot measurement; n = 5–7 for SOD, CAT, CS, and COX activities). *P < 0.05.

Insulin receptor expression varies considerably throughout the brain, with the hypothalamus and hippocampus having high expression and the midbrain having relatively low expression (6). We observed decreased CS activity in the hypothalamus of STZ + vehicle compared with ND and STZ + insulin mice (Fig. 2I) and decreased COX activity in the hypothalamus and hippocampus of STZ + vehicle compared with ND mice (Fig. 2J). However, CS and COX activities were unaltered in the midbrain, supporting a notion that insulin deficiency has a greater effect on mitochondrial function in brain regions rich in insulin receptors These results demonstrate that insulin deprivation alters brain mitochondrial function, although these alterations are region-specific and at least partially restored by insulin replacement.

Insulin deprivation alters mitochondrial protein expression and increases oxidative post-translational modifications

Because reductions in the mitochondrial proteome contribute to decreased MAPR in insulin-deficient muscle (10), we hypothesized that impaired MAPR in the insulin-deficient brain may be supported by similar reductions in mitochondrial proteins. We chose to compare ND and STZ + vehicle mice to address our research question as we are aware that systemic insulin treatment increases peripheral insulin concentration in comparison with ND and does not fully normalize SPHs in STZ mice with its impact on insulin sensitivity (29), thus making our interpretation of the data complex. Further, to maintain consistency with our functional measurements we assessed cerebral, rather than region-specific, protein expression. A proteomic survey identified 96 differentially expressed proteins between STZ + vehicle and ND, of which 76 and 20 were up-regulated and down-regulated, respectively, in STZ + vehicle (Fig. 3A). A complete list of differentially expressed proteins is provided in Supplemental Table S1. Paradoxically, we observed a net increase in mitochondrial proteins (14 up-regulated, and 2 down-regulated) in STZ + vehicle. Of note, decreased MFN1 and increased mitochondrial fission process 1 in STZ + vehicle suggest increased mitochondrial fragmentation during insulin deficiency, which may contribute to reduced mitochondrial function despite a possible compensatory increase in mitochondrial proteins. Follow-up experiments revealed decreased MFN1, MFN2, and dominant OPA1, proteins promoting mitochondrial fusion, and increased DRP1, which promotes mitochondrial fission, in the cerebrum, hypothalamus, and hippocampus, but not the midbrain, in STZ + vehicle (Fig. 3B). These results indicate that insulin deficiency alters mitochondrial dynamics in a region-specific manner, with regions with higher levels of insulin receptor displaying greater alterations in fission and fusion proteins.

Figure 3.

The effect of insulin deprivation (STZ + vehicle) on the cerebral proteome. A) Compared with ND mice, STZ + vehicle influenced the expression of 16 mitochondrial proteins (14 up- and 2 down-regulated), increased 6 proteins linked to cholesterol and lipid synthesis, and altered proteins associated with neurotransmission and neural development (9 up- and 2 down-regulated) and registered in the AlzGene database of AD-associated proteins (5 up- and 3 down-regulated) (n = 3/group). B) Insulin deficiency altered protein expressions of mitochondrial fusion and fission proteins in the cerebrum, hypothalamus, and hippocampus, but not the midbrain (n = 6/group). C, D) Insulin deficiency also increased select canonical pathways (C) and functional annotations (D) compared with ND. Only significantly affected proteins are shown on the volcano plot [false discovery rate (FDR)–corrected P < 0.05 and absolute log₂ fold change ≥0.5]. Histogram bars show select significantly affected pathways [−log₁₀ P > 1.301 (P < 0.05)]. *P < 0.05, **P < 0.01.

Additionally, we observed 5 up-regulated (APOC3, HLA, HMGCS2, CELF2, GSTZ1) and 3 down-regulated (BCAM, HMGCS1, GC) proteins with established roles in AD development (26). Similarly, proteins involved in neurotransmission and synaptic function (e.g., down-regulation of GFAP) were also altered in STZ + vehicle, whereas proteins important for cholesterol biogenesis and lipid metabolism (30) were up-regulated in STZ + vehicle.

Irreversible post-translational modifications (PTMs) of proteins mostly represent critical damage that may impair protein function. We observed 12 proteins with oxidative PTMs in STZ + vehicle that were unmodified in ND (Table 3). Of these proteins, amyloid β protein precursor intracellular domain associated protein-1 (AIDA-1) modulates amyloid β processing (31), protein phosphatase 2A (PP2A) directly regulates τ phosphorylation and is critical in neurofibrillary degeneration (32), and MEK2 is implicated in neurofibrillary degeneration in AD (33). These results indicate that increased protein damage may contribute to the increased risk for neurodegenerative diseases observed in diabetes. Furthermore, pathways analysis showed increased ketogenesis, cholesterol biosynthesis, and lipid metabolism networks (Fig. 3C) and displayed evidence for increased psychiatric and mood disorders (Fig. 3D) in STZ + vehicle compared with ND. A complete list of proteins in these pathways is provided in Supplemental Table S2.

TABLE 3.

Insulin deficiency causes oxidative PTMs

Protein	Function	Oxidative spectral count
		ND	STZ + vehicle	P
AIDA-1	Modulation of amyloid β processing	0	26.28	2.00E−06
Myosin-VI (MYO6)	Clathrin-mediated endocytosis and vesicle trafficking; regulation of synaptic structure	0	26.08	2.83E−08
Ubiquitin-fold modifier-conjugating enzyme 1 (UFC1)	Ubiquitin conjugating enzyme	0	25.91	2.24E−07
Heat shock protein 105 kDa (HS105)	Protein folding	0	25.83	4.93E−08
Peptidyl-prolyl cis-trans isomerase H (PPIH)	Protein folding	0	25.33	4.50E−09
PP2A	Regulation of tau phosphorylation	0	24.65	3.48E−06
ERI1 exoribonuclease family member 3 (ERI3)	Regulations of nucleic acid binding and exonuclease activity	0	24.61	2.46E−11
Potassium channel protein (KAT3)	Voltage-gated potassium channel subunits	0	24.49	1.43E−08
MEK2	Implicated in neurofibrillary degeneration in AD	0	24.44	7.55E−07
Phosphatidylserine decarboxylase 1 (PSD1)	Phospholipid metabolism; promotes mitochondrial fusion	0	23.34	2.09E−09
Regulation of nuclear pre-mRNA domain-containing protein 1A (RPR1A)	Cell-cycle and transcription regulation	0	23.28	1.54E−09
Putative ATP-dependent RNA helicase DHX57 (DHX57)	Nucleic acid binding; helicase activity	0	22.99	2.54E−09

Blind detection of PTMs was performed using mass spectrometry in cerebral tissue of ND and insulin-deficient (STZ + vehicle) mice (n = 3/group).

MCT inhibition increases ROS emission during insulin deficiency

The paradoxical observation of decreased MAPR and mitochondrial coupling in STZ + vehicle despite no differences in ROS emission led us to ask if compensatory mechanisms during insulin deficiency prevent increased oxidative stress. In STZ + vehicle compared with ND and STZ + insulin, we observed increased Mct1, 2, and 4 mRNA in the hypothalamus and increased Mct1 and 2 mRNA in the hippocampus but not midbrain (Fig. 4A), which is indicative of increased ketone and lactate transport in STZ + vehicle in brain regions with high insulin receptor expression. Because βHB and lactate were increased in STZ + vehicle, we hypothesized that elevations in ketone and lactate during insulin deficiency protect against increased ROS emission and mitigate oxidative stress. Thus, we intranasally injected the MCT inhibitor 4-CIN in ND and STZ + vehicle mice to determine how blocking ketones and lactate transport influenced the function of mitochondria isolated from the cerebrum. STZ + vehicle increased state 4 (leak) respiration when normalized to tissue weight (Fig. 4B) and decreased RCR (Fig. 4C) independent of 4-CIN. Corroborating our initial experiment, state 3 MARP was decreased in STZ + vehicle compared with ND when normalized to tissue weight (Fig. 4D). The impairment in MAPR in STZ + vehicle was further accentuated by 4-CIN demonstrating that ketones and lactate are significant energy sources in the insulin-deficient brain, whereas 4-CIN trended (P = 0.09) to decrease MAPR in ND. In STZ + vehicle, 4-CIN increased ROS emission when normalized to tissue weight, whereas no effect of 4-CIN on ROS emission was observed in ND (Fig. 4E). ROS emission was also increased in STZ + vehicle by 4-CIN when analyzing ROS production as a percentage of oxygen consumption (Fig. 4F). Similar results for mitochondrial OCR, MAPR, and ROS emission were observed when normalized to mitochondrial protein content (data not shown). Intriguingly, in cerebral homogenates, 4-CIN prevented increases in SOD2 and CAT activities during STZ + vehicle, whereas no effect was observed in ND (Fig. 4G, H). Because βHB and lactate were increased in the STZ + vehicle cerebrum, these results demonstrate that metabolite transport via MCTs, presumably of ketones and lactate, prevents impairments in brain ROS production during insulin deficiency, which may be attributed to increased antioxidant capacity.

Figure 4.

A) MCT1, 2, and 4 mRNA expressions were up-regulated by insulin deficiency (STZ + vehicle) compared with ND and insulin-treated (STZ + insulin) mice in the hypothalamus and hippocampus but not midbrain. B) Independent of intranasal treatment with control (Con) or the MCT inhibitor 4-CIN, STZ + vehicle increased state 4 respiration compared with ND mice. C) Similarly, mitochondrial coupling efficiency, as measured by RCR was decreased in STZ + vehicle independent of intranasal treatment. D) State 3 MAPR was decreased in STZ + vehicle compared with ND when treated with Con. Additionally, 4-CIN further decreased MAPR in STZ + vehicle, whereas it had no significant effect on MAPR in ND. E–H) In STZ + vehicle, 4-CIN increased H₂O₂ emission (E) and oxygen consumption as a percentage of ROS (F), whereas 4-CIN did not influence ROS production in ND. In STZ + vehicle, 4-CIN treatment prevented increases in mitochondrial SOD2 (G) and CAT (H) activities in cerebral homogenates. Data are means ± sem (n = 5–7/group for mRNA expression analysis; n = 7–8/group for mitochondrial function measures; n = 6/group for SOD2 and CAT measurements). *P < 0.05, **P < 0.01.

Intranasal insulin increases mitochondrial ATP production

To investigate the direct effect of insulin on brain mitochondrial function, insulin was intranasally administered to ND mice. Intranasal insulin administration did not change systemic insulin (control: 0.72 ± 0.08 vs. insulin: 0.71 ± 0.09 ng/ml; P = 0.92) and glucose (115.8 ± 5.5 vs. 111.3 ± 3.6 mg/dl; P = 0.50) concentration nor did it alter body weight (29.8 ± 0.2 vs. 29.4 ± 0.3 g; P = 0.21), body fat percentage (12.9 ± 0.4 vs. 12.4 ± 0.5%; P = 0.51), or food intake (11.8 ± 0.4 vs. 11.4 ± 0.3 kcal/d; P = 0.40), indicating that the selected dose, delivery route, or both, did not influence peripheral metabolism. Insulin increased AKT phosphorylation at Ser-473 and GSK-3β phosphorylation at Ser-9 in the hippocampus, a region selected for its key involvement in memory (Fig. 5A), indicating that insulin impacted activated signaling molecules. Moreover, we noted increased mitochondrial respiration from isolated cerebral mitochondria in states 2 and 3 when normalized to tissue weight (Fig. 5B), although no differences were observed when normalized to mitochondrial protein concentration (Fig. 5C). No difference in RCR was present (Fig. 5D), but enhanced state 3 MAPR was noted when normalized to tissue weight (Fig. 5E), although this difference was abated when normalized to mitochondrial protein concentration (Fig. 5F), consistent with a notion that mitochondrial protein content increased. Insulin did not influence ROS emission (Fig. 5G) or the percentage of oxygen consumption accounted for as ROS emission (Fig. 5H). To assess whether and how intranasal insulin influenced mitochondrial content, we measured mtDNA copy number and PGC-1α protein content, a master regulator of mitochondrial biogenesis, in the hippocampus. Insulin also increased Nd1 and Nd4 expression (Fig. 5I) as well as PGC-1α protein (Fig. 5J).

Figure 5.

Intranasal insulin increases mitochondrial respiration and ATP production. A) Intranasal insulin increased AKT and GSK-3β phosphorylation in the hippocampus. B, C) Intranasal insulin increased cerebral mitochondrial oxygen consumption (JO₂) when normalized to tissue weight (B) but not when normalized to mitochondrial protein content (C). D) Mitochondrial coupling efficiency, as measured by RCR, was not influenced by intranasal insulin. E, F) Intranasal insulin increased MAPR during state 3 respiration when normalized to tissue weight (E) but not when normalized to mitochondrial protein content (F). G, H) Mitochondrial H₂O₂ emissions (G) and oxygen consumption as a percentage of ROS (H) were not influenced by intranasal insulin. I, J) Intranasal insulin increased mtDNA copy (I) and PGC-1α protein expression (J) in the hippocampus. K) Intranasal insulin influenced the expression of 14 mitochondrial proteins (9 up- and 5 down-regulated) and proteins influencing neurotransmission (16 up- and 9 down-regulated) in the hippocampus. L) Intranasal insulin also increased canonical pathways and functional annotations associated with increased neurotransmission. Data are means ± sem (n = 8–10/group for mitochondrial function measures; n = 6/group for mtDNA measurements; n = 6/group for Western blot measurements). For proteomics data, only significantly affected proteins are shown on the volcano plot [false discovery rate (FDR)–corrected P < 0.05 and absolute log₂ fold change ≥0.5] (n = 6/group). Histogram bars show select significantly affected pathways [−log₁₀ P > 1.301 (P < 0.05)]. *P < 0.05.

Additionally, a proteomic survey identified 9 up-regulated compared with 5 down-regulated mitochondrial proteins following insulin treatment in the hippocampus (Fig. 5K). A complete list of differentially expressed proteins is provided in Supplemental Table S3. These data suggest that insulin increases mitochondrial content, in turn increasing OCR and ATP production. Proteomic pathway analysis also identified increased CREB and nNOS signaling pathways with insulin treatment (Fig. 5L). A complete list of proteins in these pathways is provided in Supplemental Table S4. Similarly, insulin up-regulated proteins with functions related neuronal development and neurotransmission. Collectively, these findings provide potential mechanistic insight into the therapeutic benefits of intranasal insulin.

DISCUSSION

The novelty of the current study is the demonstration of the regulatory role of insulin and insulin deficiency on mitochondrial ATP production and how ketone and lactate action ameliorate oxidative stress and contribute to ATP production in the insulin-deficient brain. Insulin deprivation, unlike in peripheral tissues, decreased cerebral mitochondrial ATP production without any concurrent increases in ROS emission but increased endogenous antioxidant defense. Insulin deprivation, as expected, increased ketones and lactate, and we demonstrated that intranasal administration of 4-CIN, which blocks ketone and lactate transport, decreased ATP production and antioxidants while increasing ROS emission, supporting the hypothesis that ketone and lactate not only contribute to ATP production but also counteract oxidative stress during insulin deficiency. Insulin deprivation decreased mitochondrial oxidative enzymes not only in cerebrum but also in the hypothalamus and hippocampus, which are rich in insulin receptors. Insulin deprivation also decreased mitochondrial fusion proteins (MFN1, MFN2, and OPA1) and increased fission protein (DRP1), suggesting that insulin deficiency adversely affects mitochondrial dynamics. Moreover, intranasal insulin administration enhanced PGC-1α protein, mitochondria proteins, and mitochondrial ATP production, offering evidence supporting insulin’s direct role in promoting brain mitochondrial biogenesis and ATP production. The current study also showed important changes in the brain proteome following insulin deprivation that offer insight into how insulin deficiency may contribute to AD and other dementias.

In the current study, we maintained STZ-diabetic mice on insulin treatment for 3 wk to ensure the changes that we observed were not related to direct effect of STZ or weight loss. Similar to muscle (9, 10), we observed impaired ATP production on insulin withdrawal in brain. However, previous reports in STZ-diabetic rats, without being maintained on insulin, reported no differences in brain ATP levels (34, 35) unlike the reduction in ATP production in the current study. These discrepancies with the current study may stem from differences in species, the lack of insulin treatment in the previous studies, and the methodology to assess ATP production vs. content. Surprisingly, no significant increase in ROS emission was detected in brain mitochondria in the current study, whereas SOD2 and CAT activities were increased, which contrasts to the ∼3-fold increase in ROS emission seen in insulin-deficient muscle mitochondria (10). However, our observations agree with a previous report in STZ-diabetic mice showing increased SOD activity in whole brain homogenates (36). We cannot exclude that the maximal respiration conditions in our ex vivo mitochondrial ROS emission measurements may not be sufficiently sensitive to detect small changes in ROS occurring during insulin deficiency and may not reflect in vivo levels of oxidative stress. Additionally, by only measuring mitochondrial H₂O₂ emission, it is possible that we failed to detect other sources of ROS that may have contributed to increased antioxidant defense systems during insulin deficiency.

Previous reports show that ketones protect against chemical-induced oxidative stress in isolated neurons (37, 38), supporting the results in the current animal study. Brain ketone and lactate transport is increased in T1D during hypoglycemia (39); however, their effects on oxidative stress in the diabetic state have not been reported. Of interest, during ketosis glucose utilization in the brain has been shown to decrease by ∼10% for each moles-per-liter increase in plasma ketones (40). Our finding that MCT inhibition increased ROS emission in STZ + vehicle mice supports a notion that increased ketone and lactate utilization has a neuroprotective effect during insulin deficiency by protecting from hyperglycemia-related oxidative stress. Enhanced antioxidant effects are a proposed mechanism by which ketones provide neuroprotection against degenerative diseases such as AD (41). Our finding that MCT inhibition further impairs ATP production in STZ + vehicle indicate that ketones and lactate constitute important energy sources in the insulin-deficient brain. Overall, these findings support that brain, unlike muscle, is much more efficiently shielded from irreversible damages by antioxidant defense and in sustaining ATP production during insulin deficiency.

Reductions in mitochondrial ATP production during insulin deficiency were accompanied by the paradoxical increase of 13 mitochondrial proteins. This response differs from that of insulin-deficient muscle, which displays substantially decreased abundance of mitochondrial proteins (10). In the current study, the down-regulation of proteins promoting mitochondrial fusion and up-regulation of DRP1 supports increased mitochondrial fragmentation during insulin deficiency. Thus, despite increased mitochondrial protein expression, the functions of these proteins may be impaired because of mitochondrial fragmentation, leading to damaged mitochondria and reduced ATP production during insulin deficiency. Because the inhibition of DRP1 prevented reductions in ATP production in hippocampal neurons from db/db insulin-resistant mice (42), insulin may enhance brain mitochondrial ATP production by promoting proper mitochondrial dynamics. However, because we did not control for hyperglycemia in the CNS of STZ + vehicle mice, we cannot distinguish between the relative effects of insulin deficiency and hyperglycemia, which has been shown to increase oxidative stress and promote mitochondrial fission in neurons (43, 44). Future controlled experiments are needed to directly assess the independent contribution of hyperglycemia on brain mitochondrial function.

In addition to low cerebral insulin levels and hyperglycemia, additional factors, such as impaired blood–brain barrier (BBB) integrity and neurovascular damage, also could potentially contribute to alterations in mitochondrial function and dynamics. Studies in STZ-diabetic rats show compromised BBB integrity and increased microvascular damage, which is only partially corrected with insulin treatment (45). Compromised BBB integrity in STZ-diabetes has been shown to increase permeability to sucrose (46) and serum proteins such as albumin (47) but decrease permeability to choline (48). Interestingly, compromised BBB integrity in STZ-diabetes is also associated with increased insulin transport kinetics independent of plasma insulin levels (49). However, given the very low plasma insulin concentrations in STZ + vehicle, the current study levels of cerebral insulin concentrations in STZ + vehicle were often below the level of detection.

Proteomic analysis also demonstrated that insulin deprivation is accompanied by irreversible PTMs of several proteins that cause τ phosphorylation or neurofibrillary degeneration that are pathognomonic of AD. Several reports show τ hyperphosphorylation (50–52) and decreased PP2A activity (50) during insulin deprivation. The current study found that reduced PP2A activity during insulin deficiency can be attributed to irreversible damage to this protein. Moreover, our study showed that other observed PTMs could cause neurofibrillary degeneration and amyloid processing, such as modifications to AIDA-1 and MAPK2K2. Oxidative damage to these proteins provides potential mechanistic insight to explain increased τ phosphorylation and amyloid plaques in T1D.

Our proteomic survey also revealed networks indicative of increased cholesterol and lipid metabolism during insulin deficiency. Brain cholesterol turnover is increased in AD and other neurodegenerative diseases and may contribute to their pathogenesis (53, 54) and potentially relate to reduced insulin action. The proteomic evidence for increased lipid metabolism during insulin deficiency is supported by our observation of elevated sphingosine and ceramide concentrations, which are central to numerous neurodegenerative diseases (55). Because sphingosine acts as a substrate for de novo ceramide synthesis (56), these changes may reflect increased de novo ceramide synthesis within the brain. Altered SPH metabolism is hypothesized to contribute to depression pathology (57), whose proteomic pathways were also increased by insulin deficiency, which is consistent with findings reporting mood disorders in T1D humans (58). Thus, alterations in lipid metabolism may link T1D with a variety of neurologic disorders, including cognitive impairment and depression. However, region-specific differences as STZ-diabetic rats are reported to show increased (59), and decreased (60) cholesterol content in the hippocampus and hypothalamus, respectively.

Because insulin deficiency is accompanied by numerous metabolic changes, which may contribute to altered mitochondrial function independent of insulin action, and previous studies (12, 13, 61) showed that insulin has a stimulatory effect on mitochondrial biogenesis and function, we investigated the effects of intranasal insulin to isolate the sole effect of insulin on brain mitochondrial function. We noted that insulin increased mitochondrial respiration and ATP production, indicating a direct action of insulin on brain mitochondrial function and expanding upon in vitro findings that insulin increased mitochondrial respiration in cortical neurons (62). The observation that intranasal insulin increased PGC-1α protein, mtDNA copy number, and proteins supports that insulin promotes mitochondrial biogenesis in brain. Our observation of the increase in neural development and neurotransmission proteomic pathways provides potential mechanistic insight to reports showing promising effects of intranasal insulin for increasing cognitive performance in humans (8, 63). Additionally, CREB and nNOS signaling pathways, which were both increased with insulin treatment, have been shown to mediate learning and memory (64, 65). Because impaired mitochondrial ATP production and oxidative damage is central to many neurodegenerative diseases, one mechanism by which insulin may enhance cognitive performance is through increasing mitochondrial content and ATP production.

Our study is focused on different regions of brain involving insulin receptors and affected by AD. We appreciate that insulin receptors are expressed in several types of cells in the CNS, such as neuron, astrocyte, microglia, and oligodendrocyte. The current study does not address which types of cell are affected by insulin. Identification of the cell types responsible for this insulin-induced amelioration of mitochondrial functions should be considered in the future. Our primary aim was to assess the effect of insulin deficiency at the organ level using ex vivo approaches. Future in vivo studies applying NMR spectroscopy and in vitro cell-based studies will offer more insight on how insulin deficiency influences mitochondria function. Our functional measurements were performed on isolated mitochondria from the cerebrum because of insufficient tissue size available in regions like the hypothalamus and hippocampus. Future studies need to expand on our findings on isolated mitochondrial function from pooled samples from the specific brain regions with rich or poor insulin receptor expression.

In summary, we demonstrate that insulin deprivation in STZ-diabetic mice leads to significant impairment of mitochondrial function, whereas ROS emission is mitigated by ketone and lactate transport to the brain. Reduced ATP production and oxidative enzymes likely affect many neuronal functions. These changes occur in association with increased expression of proteins affecting cholesterol and SPH metabolism, increased ceramide content, and with increased oxidative damage to proteins involved in AD pathogenesis. Moreover, as direct evidence of insulin’s effect on mitochondria, we find that intranasal insulin promotes mitochondrial ATP synthesis, which may at least partially explain its therapeutic potential. Recent viewpoints highlight the importance of mitochondrial metabolism for maintaining brain function and ketosis during metabolic stress may enhance neural networks and improve stress resistance (66, 67). Collectively, our data indicate insulin as a pivotal regulator of brain mitochondrial biogenesis and energy production. Moreover, these studies clearly demonstrate that brain, unlike peripheral tissues, such as muscle, has many compensatory mechanisms including increased reliance on ketones and lactate to prevent rapid deteriorations in mitochondrial function.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.

Glossary

βHB

β-hydroxybutyrate

4-CIN

α-cyano-4-hydroxy-cinnamic acid

AD

Alzheimer’s disease

AIDA-1

amyloid β protein precursor intracellular domain associated protein-1

BBB

blood–brain barrier

CAT

catalase

COX

cytochrome c oxidase

CS

citrate synthase

DRP1

dynamin-related protein 1

HClO₄

perchloric acid

IM

isolation medium

MAPR

mitochondrial ATP production rate

MCT

monocarboxylate transporter

MFN1

mitofusin 1

MS/MS

tandem mass spectrometry

ND

nondiabetic

OCR

oxygen consumption rate

OPA1

optic atrophy

PGC-1α

peroxisome proliferator-activated receptor gamma coactivator 1-alpha

PP2A

protein phosphatase 2A

PTM

post-translational modification

RCR

respiratory control ratio

ROS

reactive oxygen species

SOD2

superoxide dismutase 2

SPH

sphingolipid

STZ

streptozotocin

T1D

type 1 diabetes

AUTHOR CONTRIBUTIONS

G. N. Ruegsegger, S. Manjunatha, and K. S. Nair designed experiments and wrote the manuscript; G. N. Ruegsegger, S. Manjunatha, P. Summer, S. Gopala, P. Zabeilski, S. Dasari, P. M. Vanderboom, and K. A. Klaus conducted experiments and reviewed the manuscript; G. N. Ruegsegger, S. Manjunatha, S. Dasari, and K. S. Nair analyzed data; and I. R. Lanza contributed to the establishment of the ATP production assay and reviewed the manuscript.