Alzheimer’s Disease Frontal Cortex Mitochondria Show a Loss of Individual Respiratory Proteins But Preservation of Respiratory Supercomplexes

Paula M. Kenney and James P. Bennett, Jr.
Neurodegeneration Therapeutics, Inc.
Charlottesville, VA

James P. Bennett, Jr. M.D.,Ph.D.
Neurodegeneration Therapeutics, Inc.
3050A Berkmar Drive
Charlottesville, VA 22901-3450
434-529-6457 (phone)
434-529-6458 (FAX)


Alzheimer’s disease (AD), the most common cause of sporadic dementia in adults, shows increased risk of occurrence with aging and is destined to become a major socio-medical tragedy over the next few decades. Although likely complex in origin, sporadic AD is characterized by a progressive and stereotyped neuropathology with aggregated protein deposition (esp beta amyloid (BA) and hyperphosphorylated tau (P-tau)) and neuronal degeneration. To date, prevention of BA synthesis or immune-mediated removal of BA have failed to alter AD progression. Development and testing of P-tau therapeutics is a work in progress.

AD brain tissues show multiple system deficits, including loss of respiratory capacity. In this study there were no apparent differences in mitochondrial mass between AD and CTL samples. We examined total homogenate and mitochondrial preparations of postmortem AD and age-matched CTL frontal cortex for relative levels of individual respiratory protein complexes by traditional Western immunoblotting and compared these to FPKM estimates of OXPHOS subunits gene expression. ANOVA revealed deficiencies of all respiratory complex subunits in AD. We also examined mitochondrial extracts with blue-native gel electrophoresis combined with immunoblotting for subunits of complexes I and III to search for “respiratory supercomplexes” (RSC’s). We found that levels of RSC’s did not differ between AD and CTL samples.

Mitochondrial preparations from end-stage AD brain tissue showed loss of ATP-producing respiration subunits but preservation of assembling respiratory subunits into RSC’s. Disease-altering therapies of early AD could include stimulation of mitochondrial biogenesis to overcome loss of respiratory subunits. (245 words).


Post-mitotic tissues such as brain require substantial production of ATP to meet energy requirements. Estimates are that 20-25% of cardiac output, metabolic fuels and oxygen are consumed by adult brains that typically constitute 2-3% of body weight. Both astrocytes and neurons participate in 2-deoxyglucose uptake by brain, a proxy of brain metabolism [1, 2].
Alzheimer’s disease (AD) is a mostly sporadically occurring, aging-related neurodegenerative condition of adults that is characterized in its early stages by brain regional loss of cerebral glucose utilization [3] and increased markers for oxidative stress [4-9]. These two findings suggest impairments of mitochondrial respiration, although other deficits, particularly increasing insulin resistance, may account for some of these changes.

Previously, we [10] and others [4] have shown deficits of mitochondrial respiration in postmortem AD brain. We carried out the present study to investigate the origins of these respiratory deficits in AD brain. There appears to be transcriptional repression in AD brain that accounts for some of the respiratory subunits’ deficits. In addition, using blue-native electrophoresis, we observed brain mitochondrial respiratory supercomplexes (RSC’s) for complexes I and III that were present at similar levels in both AD and CTL samples. Our findings suggest that AD brain mitochondria have an ATP-producing deficit that could be addressed by approaches to stimulate mitochondrial biogenesis.


Tissue Samples

Blocks of slow frozen cortical ribbon from human frontal cortex were obtained from the University of Virginia Brain Resource Facility. These samples were used in our earlier work [11].

Preparation of Mitochondria

Human frontal cortex mitochondrial fractions were prepared by a modification of the method of Lai and Clark [12] with the Isolation Buffer recommended by Jha, Wang and Auwerx [13]. 1 gm of slow frozen cortical ribbon from each case was minced, on ice, in 3 ml of Mitochondrial Isolation Buffer (MIB) [200 mM Sucrose, 1 mM Tris Base, 10 ml/liter 100 mM EGTA/Tris {100 mM EGTA adjusted to pH 7.4 using Tris powder} pH of final solution adjusted to 7.4 with 1 M HEPES, filtered and aliquots stored at -20° C. 1X Protease Inhibitor Cocktail, Mammalian (VWR, Radnor, PA) added at time of use.]. Minced tissue was homogenized, on ice, for 60 passes in a Dounce homogenizer (0.05 mm clearance) in 7 volumes MIB and centrifuged for 3 min at 1300 X g, 4° C. Supernatants were saved on ice while pellets were homogenized for an additional 30 passes in 6 volumes MIB and centrifuged 3 min at 1300 X g. Supernatants were combined and centrifuged for 80 min at 2,721 X g, 4° C. Pellets were washed with 10 volumes MIB and centrifuged for 80 min at 2,721 X g, 4° C. Final pellet were resuspended in 4 volumes MIB, protein concentrations determined using a Pierce™ Detergent Compatible Bradford Assay (VWR, Radnor, PA) and 100 ug aliquots centrifuged 20 min at 17,000 X g, 4° C. Supernatant were aspirated and mitochondrial pellets stored at -80° C.

Sample Preparation for SDS-Page

0.3 g slow frozen human cortical ribbon from each case was homogenized for 30 passes in modified Radioimmunoprecipitation Assay (RIPA) buffer (50 mM Tris HCl, pH 7.4; NP-40, 0.25% sodium deoxycholate, 150 mM sodium chloride, 1mM PMSF and 1X Protease Inhibitor Cocktail, Mammalian [VWR, Radnor, PA] added at time of use.). Alternatively, 600 ug of mitochondrial fraction pellets from each case were resuspended in 300 ul RIPA buffer. All RIPA suspensions were sonicated for 4 min in a water bath sonicator, held on ice for 30 min with vortexing every 5 min and centrifuged at 15,000 X g for 10 min. Supernatants were saved, protein concentrations determined using a Pierce™ Detergent Compatible Bradford Assay and 20 ug (mitochondrial fraction or 40 ug (whole brain) aliquots stored at -80° C.

SDS-PAGE Immunoblots

BioRad (Hercules, CA) complete XT Sample Buffer added to 40 ug cortex RIPA lysate or 20 ug mitochondrial RIPA lysate, heated for 5 min at 95°C, loaded on BioRad Criterion XT 4-12% Bis-Tris gels and electrophoresed for 30 to 40 min 200 v in XT-MES buffer. Gels were equilibrated 10 min in transfer Buffer and proteins transferred to PVDF membranes using an Invitrogen iBlot (Life Technologies, Carlsbad, CA) program P3 for 7 min. Transferred membranes were fixed in 8% Acetic Acid, washed with water and air-dried. Following rehydration (3 x 5 min methanol, 3 x 5 min water), membranes were blocked for 1 hr in Li-Cor Blocking Buffer (Lincoln, NE), incubated 2 hr in primary antibodies, washed, incubated 30 min in secondary antibodies, washed and imaged on a Li-Cor Odyssey Imager. All primary antibodies were from abcam (Cambridge, MA) and all secondary antibodies were from Li-Cor.

Blue Native-PAGE Immunoblots

Blue Native-PAGE of mitochondrial fractions was preformed according to the method of Jha, Wang and Auwerx [13] using a XCell SureLock Mini-Cell, Native Sample Prep Kit, NativePAGE Anode buffer, NuPage Transfer Buffer and NativeMark Unstained Protein Standard all from Fisher Scientific (Atlanta, GA). Anode, cathode and sample buffers were prepared, fresh, on the day of use. Briefly, 100 ug mitochondrial pellets were solubilized in 40 ul each 8 g/g digitonin sample buffer for 20 min on ice, centrifuged 10 min at 20,000 X g, 4°C and 15 ul of supernatant transferred to two fresh tubes. 2 ul Coomassie G250 Sample Additive was added to each tube and prepared samples loaded in wells of Invitrogen NativePage 3-12% Bis-Tris Gels. Gels were run for 25 min at 150 v in Dark Blue Cathode Buffer (0.044 g Coomassie Brilliant Blue G-250 in 220 ml NativePage anode buffer). Dark Blue cathode buffer was removed, Light Blue Cathode Buffer (20 ml Dark Blue Cathode Buffer in 180 ml NativePage Anode Buffer) added and gels run for an additional 3.5 hr at 250 v. Final gels were equilibrated 15 min in NuPage Transfer Buffer and proteins transferred to PVDF membranes using an Invitrogen iBlot program P3 for 14 min. Membranes were washed with water, destained/fixed for 3 x 5 min in 25% Acetic Acid/50% methanol (Liu et al, 2017), washed with water and air-dried. Following rehydration, membranes were immunolabelled and imaged as above. All primary antibodies were from Abcam (Cambridge, MA) and all secondary antibodies were from Li-Cor.

RNA sequencing of AD and CTL tissues

RNA sequencing (RNAseq) and bioinformatics analyses were carried out as described in an earlier publication.[11]


Figure 1 shows plots of immunoblot results of VDAC (porin) levels normalized to tissue beta actin levels. There were no differences between CTL and AD frontal cortex samples. This result indicates no differences in estimates of mitochondrial mass between AD and CTL samples.

Figure 1. No differences in apparent mitochondrial mass between CTL and AD frontal cortex samples. Shown are mean +/- SEM values for beta actin-normalized VDAC/porin levels in frontal cortex samples from CTL and AD whole tissue samples. There were no significant differences between CTL and AD samples.


Figure 2 shows OXPHOS blots of semi-purified mitochondrial preparations, extracted with RIPA buffer, electrophoresed/blotted/immunostained as described in Methods and normalized to mitochondrial mass (VDAC/porin).

Figure 2. AD mitochondrial preparations from frontal cortex show reductions in all OXPHOS subunits representative of complexes I-V. Shown are immunoblot results for mitochondrial preparations of frontal cortex from CTL and AD samples, extracted with RIPA buffer, electrophoresed and immunoblotted as described in Methods and normalized to mitochondrial mass (VDAC). All OXPHOS subunits showed significant reductions in AD mitochondria (2-way ANOVA), and several subunits showed individual declines in AD (t-test). (top) OXPHOS subunit levels are expressed as % mean CTL levels. (bottom) OXPHOS subunits raw signals (VDAC-normalized) are shown.


There were significant reductions in the AD samples, both with regard to all the individual respiratory subunits examined (2-way ANOVA), and multiple individual complex subunits (t-tests). These results indicate a loss of individual respiratory subunits in the mitochondria. An image of OXPHOS complexes immunoblot can be found in Supplemental Figure 1.

Supplemental Figure 1. Shown is a representative immunoblot for OXPHOS subunits imaged on the Li-Cor near-infrared laser scanner. MW markers (in kDa) are on the right; individual OXPHOS subunits are indicated on the left.


Figure 3 shows results for immunostaining of Complex IV, subunit 6b, which is a nuclear genome-encoded subunit.

Figure 3. CTL and AD frontal cortex crude mitochondrial fractions’ have reductions of levels of CIV, subunit 6b. Shown are the WB results (mean +/- SEM) for signals of CIV, subunit 6b, normalized to VDAC levels, obtained from crude mitochondrial fractions of frontal cortex as described in Methods. There was a significant reduction (by t-test) in signal for the AD samples.


We carried out this analysis because all of the subunits of our OXPHOS antibody cocktail were made against nuclear genome-encoded subunits, except for Complex IV, which was made against subunit 2 that is mtDNA-encoded. We found that VDAC-normalized CIV, subunit 6B levels were significantly reduced in AD samples (t-test).
Figure 4 shows the relationships among VDAC-normalized OXPHOS subunits in frontal cortex whole homogenates compared to their gene expression values from Cufflinks analysis of RNAseq data (shown as FPKM (fragments per kilobase per million bases sequenced)).

Figure 4. CTL and AD frontal cortex samples’ OXPHOS gene expression (as FPKM values) compared to Western blot immunostaining reveal few correlations. Note there is little if any positive relationship except for the CIII subunit (UQCRC2).


In general there were no obvious relationships between gene expression FPKM values and Western blot signals for the OXPHOS subunits, in both CTL and AD samples. The one exception was for complex III (UQCRC2 subunit), where in both CTL (top) and AD (bottom) there appeared to be linear relationships
Figure 5 shows analyses of CI respiratory supercomplexes (RSC) (top) and CIII RSC (bottom) as % of total CI or CIII.

Figure 5. CTL and AD frontal cortex mitochondrial samples show no differences in levels of total respiratory supercomplexes (RSC) of CI and CIII. Shown are mean +/- SEM of immunoblot data from frontal cortex mitochondrial preparations of CTL and AD samples, immunostained for CI subunit (TOP) or CIII subunit (BOTTOM). In both cases, the majority of immunostaining was in the RSC bands.


Note that most of the signals for both CI and CIII was found in RSC. Images of RSC can be found in Supplemental Figure 2.

Supplemental Figure 2. Shown are representative immunoblots for RSC’s of CI (TOP) and CIII (BOTTOM) imaged on the Li-Cor near-infrared laser scanner.



AD is already a major socio-medical problem that will only worsen as more of our population ages. To date, no single approach has altered the course of decline in AD, although a multi-component, lifestyle change program has shown promise [14]. These difficulties in achieving success in altering the course of decline in AD likely have multiple origins, including the possibility that AD at the molecular level may be a heterogenous disorder in spite of common macroscopic neuropathological changes.

Because of the brain’s disproportionately large energy substrate use, oxygen consumption and ATP requirements, and OXPHOS deficits found in AD brain and other tissues, we and others have conducted multiple studies to determine potential origins of these respiratory deficiencies. AD brain and peripheral tissues demonstrate deficits in electron transport/ATP production and increases in oxidative stress.[3-9, 15-17]

There appear to be multiple origins of these respiratory problems. AD mitochondrial DNA is functionally altered, as determined in cybrid studies [9], and shows excessive mutations [9]. AD mitochondria show reduced respiratory capacity and increased oxidative stress damage. One of the earliest detectable changes in AD subjects (who have a clinical precursor state known as “mild cognitive impairment”) is the loss of cerebral glucose metabolism [3], consistent with but not diagnostic of reduced oxidative phosphorylation. These observations suggest that loss of respiratory capacity occurs early in AD pathogenesis and may comprise a focal point of therapeutic intervention.

Using protein immunoblotting we found no reduction in apparent mitochondrial mass in AD frontal cortex brain samples, but we did note reductions in levels of individual respiratory protein subunits in semi-purified mitochondrial subfractions.
There was no apparent decline in levels of respiratory supercomplexes (RSC’s), which we detected with blue-native gel electrophoresis. As far as we are aware, ours is the first report of RSC’s in postmortem human brain. Prior studies have involved fresh (not frozen) mitochondria isolated from rodent brain [13] or human fibroblasts [18, 19]. Comparing our images with those published, we observed more “smear” of higher MW RSC’s in both AD and CTL samples, likely arising from use of frozen tissues and long-term storage of samples.

For some but not all respiratory subunits, there was a correlation between levels of mRNA for that subunit and levels of subunit protein in total homogenate subfractions. These findings suggest either a loss of transcriptional drive (gene expression) or accelerated degradation of mRNA, such as might occur with microRNA’s. Conceivably both processes could be taking place. Further studies are needed to sort out the mechanisms of respiratory subunit loss.

There are many limitations in our study. Chief among these are the use of postmortem frozen tissues that have been stored and have significant (but generally unavoidable) postmortem delays. Another limitation is our use of advanced/end-stage AD samples. There is significant neurodegeneration in multiple forebrain areas of AD, and it is likely that much of the neuronal mitochondrial pathology is missing from our samples. We may be examining changes of “survivor” neurons, in addition to astroglia, and hope that it will be possible in the future to repeat our study in earlier cases (particularly MCI).

Are there therapeutic implications of our findings? One of the most straightforward therapeutic approaches would be to stimulate mitochondrial biogenesis (mitobiogenesis), even though “defective” mitochondria might result. An increase of mitochondrial mass, even if composed of reduced respiratory capacity/mitochondria, could increase ATP production to normal (or even above normal) levels.

Would increasing “defective” mitos increase oxidative stress? Stimulating mitobiogenesis could be combined with anti-oxidant therapy. It is not clear what therapeutic effect, if any, would accrue in such circumstances, but since the therapeutic index for stimulating mitobiogenesis is favorable, this seems like a worthwhile experiment. Peripheral tissues such as platelets or fibroblasts could serve as biomarkers for effective of stimulating mitobiogenesis in individuals.


[1] Nehlig A, Coles JA (2007) Cellular pathways of energy metabolism in the brain: is glucose used by neurons or astrocytes? Glia 55, 1238-1250.
[2] Figley CR, Stroman PW (2011) The role(s) of astrocytes and astrocyte activity in neurometabolism, neurovascular coupling, and the production of functional neuroimaging signals. Eur J Neurosci 33, 577-588.
[3] Mosconi L, Pupi A, De Leon MJ (2008) Brain glucose hypometabolism and oxidative stress in preclinical Alzheimer’s disease. Ann N Y Acad Sci 1147, 180-195.
[4] Ferrer I (2009) Altered mitochondria, energy metabolism, voltage-dependent anion channel, and lipid rafts converge to exhaust neurons in Alzheimer’s disease. J Bioenerg Biomembr 41, 425-431.
[5] Dias C, Barbosa RM, Laranjinha J, Ledo A (2014) Evaluation of Mitochondrial Function in the CNS of Rodent Models of Alzheimer’s Disease – High Resolution Respirometry Applied to Acute Hippocampal Slices. Free Radic Biol Med 75 Suppl 1, S37.
[6] Kumar A, Singh A (2015) A review on mitochondrial restorative mechanism of antioxidants in Alzheimer’s disease and other neurological conditions. Front Pharmacol 6, 206.
[7] Grimm A, Schmitt K, Eckert A (2016) Advanced Mitochondrial Respiration Assay for Evaluation of Mitochondrial Dysfunction in Alzheimer’s Disease. Methods Mol Biol 1303, 171-183.
[8] Islam MT (2017) Oxidative stress and mitochondrial dysfunction-linked neurodegenerative disorders. Neurol Res 39, 73-82.
[9] Onyango IG, Khan SM, Bennett JP, Jr. (2017) Mitochondria in the pathophysiology of Alzheimer’s and Parkinson’s diseases. Front Biosci (Landmark Ed) 22, 854-872.
[10] Young-Collier KJ, McArdle M, Bennett JP (2012) The dying of the light: mitochondrial failure in Alzheimer’s disease. J Alzheimers Dis 28, 771-781.
[11] Bennett J, Keeney P (2017) Micro RNA’s (mirna’s) may help explain expression of multiple genes in Alzheimer’s Frontal Cortex. Journal of Systems and Integrative Neuroscience 3, 1-9.
[12] Lai JC, Clark JB (1978) Isocitrate dehydrogenase and malate dehydrogenase in synaptic and non-synaptic rat brain mitochondria: a comparison of their kinetic constants. Biochem Soc Trans 6, 993-995.
[13] Jha P, Wang X, Auwerx J (2016) Analysis of Mitochondrial Respiratory Chain Supercomplexes Using Blue Native Polyacrylamide Gel Electrophoresis (BN-PAGE). Curr Protoc Mouse Biol 6, 1-14.
[14] Bredesen DE, Amos EC, Canick J, Ackerley M, Raji C, Fiala M, Ahdidan J (2016) Reversal of cognitive decline in Alzheimer’s disease. Aging (Albany NY) 8, 1250-1258.
[15] Mancuso M, Orsucci D, Siciliano G, Murri L (2008) Mitochondria, mitochondrial DNA and Alzheimer’s disease. What comes first? Curr Alzheimer Res 5, 457-468.
[16] Zhang L, Zhang S, Maezawa I, Trushin S, Minhas P, Pinto M, Jin LW, Prasain K, Nguyen TD, Yamazaki Y, Kanekiyo T, Bu G, Gateno B, Chang KO, Nath KA, Nemutlu E, Dzeja P, Pang YP, Hua DH, Trushina E (2015) Modulation of mitochondrial complex I activity averts cognitive decline in multiple animal models of familial Alzheimer’s Disease. EBioMedicine 2, 294-305.
[17] Coskun P, Helguera P, Nemati Z, Bohannan RC, Thomas J, Samuel SE, Argueta J, Doran E, Wallace DC, Lott IT, Busciglio J (2017) Metabolic and Growth Rate Alterations in Lymphoblastic Cell Lines Discriminate Between Down Syndrome and Alzheimer’s Disease. J Alzheimers Dis 55, 737-748.
[18] Salvador-Severo K, Gomez-Caudillo L, Quezada H, Garcia-Trejo JJ, Cardenas-Conejo A, Vazquez-Memije ME, Minauro-Sanmiguel F (2017) Mitochondrial proteomic profile of complex IV deficiency fibroblasts: rearrangement of oxidative phosphorylation complex/supercomplex and other metabolic pathways. Bol Med Hosp Infant Mex 74, 175-180.
[19] Baertling F, Sanchez-Caballero L, van den Brand MAM, Wintjes LT, Brink M, van den Brandt FA, Wilson C, Rodenburg RJT, Nijtmans LGJ (2017) NDUFAF4 variants are associated with Leigh syndrome and cause a specific mitochondrial complex I assembly defect. Eur J Hum Genet 25, 1273-1277.

Author contributions and conflicts of interest

PMK and JPB designed the study. PMK performed all protein extractions and mitochondrial preparations, all Western blots and generated all primary immunoblot data. JPB helped analyze the primary data, performed all bioinformatics studies (RNAseq) and wrote the manuscript drafts. Both authors have reviewed and agree with the manuscript final version. All data were analyzed “blind”, solely by case number and not diagnosis.
Neither author has any conflicts of interest with this manuscript.

Data availability

All primary data are the property of Neurodegeneration Therapeutics, Inc. and are available upon request to the corresponding author (JPB) after completion of a suitable material transfer agreement (MTA).