Age-associated changes in human CD4⁺ T cells point to mitochondrial dysfunction consequent to impaired autophagy

Protein expression in human CD4⁺ T Cells

We examined the protein profiles from cytoplasmic extracts of human CD4⁺ T cells obtained from young (21-34 years-old) and older (68–83 years-old) male participants of the BLSA (Supplementary Table 1). Total CD4⁺ T cells were isolated from apheresis-derived peripheral blood mononuclear cells (PBMCs). Total CD4⁺ T lymphocytes, without distinction between naïve and memory subclass, were used for protein expression analysis because of the high number of cells needed to obtain enough cytoplastic extract to perform the proteomic study. The proportion of naïve CD4⁺ T cells (based on CD45RA⁺ expression) in the younger or older groups were comparable (Supplementary Figure 1), reflecting earlier published observations [28, 29]. Cytoplasmic extracts from CD4⁺ T cells were labeled with isobaric tags for relative and absolute quantification (iTRAQ) of proteins by liquid chromatography-mass spectrometry (LC-MS/MS). We compared 18 cytoplasmic extracts from older participants against four extracts from younger participants using four different iTRAQ 8plex kits. Equal amounts of cytoplasmic extract were used. One control sample from young participants was included in all four iTRAQ sets and used for protein quantitative normalization. To enhance comparability, technical controls (duplicate samples) were included within the same iTRAQ and also between iTRAQs (Supplementary Table 1). About 3,510 proteins were identified from each iTRAQ experiment. Data were analyzed within each iTRAQ 8plex and also in combination. Twenty-nine proteins (Table 1), ranked by significance (only proteins with p < 0.05 shown), were found to be up-regulated (ratio >1.3) in CD4⁺ T cells from older donors compared to younger donors after meeting additional analytical criteria (≥ 2 unique peptides, high confidence of the peptide sequences, and ion score of more than 30; see Materials and Methods). Interestingly, 23 of these were mitochondrial proteins (16 belonged to the electron transport chain; Figure 1A and Supplementary Table 2).

Figure 1. (A) Mitochondrial electron transport chain proteins that we found to be more highly expressed in the old are indicated in red by their location (Complex II, IV, and V) in the cartoon schema. (B) Gene pathways identified for proteins upregulated in the old using DAVID to convert protein identifiers to their genes (DAVID Bioinformatics Resources 6.7 (https://david.ncifcrf.gov). These genes fall into pathways related to oxidative phosphorylation, neurodegenerative disease, and integration of energy metabolism (Oxidative phosphorylation [16 from KEGG pathway database], Neurodegenerative disease [16 from KEGG], Cardiac muscle contraction [6 from KEGG], Integration of energy metabolism [16 from REACTOME pathway database], Diabetes [16 from REACTOME], Metabolism of nucleotides [8 from REACTOME]).

Using the Database for Annotation, Visualization and Integrated Discovery (DAVID), we found that proteins increased in CD4⁺ T cells from older participants mapped to several molecular pathways, namely oxidative phosphorylation (16 from KEGG pathway database), cardiac muscle contraction (6 from KEGG), integration of energy metabolism (16 from REACTOME pathway database), and metabolism of nucleotides (8 from REACTOME) (Figure 1B). These data demonstrate that mitochondrial proteins and particularly proteins that pertain to electron transport chain complexes accumulate with age in CD4⁺ T lymphocytes.

Gene expression in human CD4⁺ T Cells

To understand whether the higher accumulation of mitochondrial proteins in the cytoplasm of CD4⁺ T lymphocytes from older compared to younger participants was due to increased transcription/translation or rather to defective elimination, we studied gene expression in human CD4⁺ T cells by using the Parametric Analysis of Gene Enrichment (PAGE) analysis [29, 30] of normalized hybridization data from cRNA. Cells from 33 donors were analyzed divided into a young group (5 men and 3 women, ages 25–38 years-old) and an older group (19 men and 6 women, ages 70–82 years-old), (Supplementary Table 3). Expression pathways were transformed into Z-scores, and significant differences (p-value ≤ 0.05) between young (Y) and old (O) donors were identified (Figure 2, Supplementary Figure 2A, 2B).

Figure 2. Gene expression analysis showing top 51 up- and down-regulated biological pathways in CD4⁺ T cells between older (70 to 80 years-old) and younger (20 to 30 years-old) men. Mitochondria-related and oxidative phosphorylation-related pathways were the most down-regulated pathways in older compared to younger donors. Differences of Z-scores between younger and older participants are shown on the X-axis. Each row denotes a different pathway (p ≤ 0.05 and FDR ≤ 0.3). N = 5 young, 19 old donors.

To be consistent with proteomic data, we first compared gene expression only between young and old men. Contrary to our proteomic analysis, PAGE identified “mitochondria-related” and “oxidative phosphorylation-related” pathways were amongst the most down-regulated biological pathways in older compared to younger individuals (Figure 2). That is, for many proteins found to be overrepresented in older persons CD4⁺ T lymphocytes, gene transcripts were instead underrepresented in CD4⁺ T lymphocytes from older persons. The same trends were noted when the analysis was restricted to women or included both sexes (Supplementary Figure 2A, 2B). Thus, proteomic and gene expression results tended to be inversely correlated.

Differences of mitochondrial respiration between young and old CD4⁺ T Cells

To understand whether overrepresentation of mitochondrial proteins in CD4⁺ T lymphocytes from older compared to younger individuals comes from functional or dysfunctional mitochondria, we measured mitochondrial respiration in CD4⁺ T cells of 7 young (ages 22–35 years-old) and 7 old (ages 80–93 years-old) male participants (Supplementary Table 4A) using high-resolution respirometry OROBOROS Oxygraph-2k (O2k). We challenged CD4⁺ T lymphocytes from younger and older participants with several mitochondrial inhibitors in sequence to estimate mitochondrial reserve capacity, nonmitochondrial oxygen capacity, ATP-linked respiration and proton leak and we used these values to derive the Bioenergetic Health Index (BHI), a summary index of mitochondrial function in cells [31]. The BHI is derived from calculating a ratio of positive aspects of mitochondrial bioenergetic function (i.e. reserve capacity and ATP-linked respiration) to potentially deleterious aspects of mitochondrial bioenergetic function (i.e. non-mitochondrial oxygen consumption and proton leak) [31].

BHI was higher in cells from younger compared to older participants (Supplementary Figure 3, representative mitochondrial respiration measurements), except for one outlier whose flow cytometric profile was different from the other 13 samples and the percentage of memory cells (central and effector) much higher than that percentage in the other 6 older donors (59% compared to a range 26–48%) (Figure 3A and Supplementary Table 4B). After accounting for differences due to different percent of memory cells between groups, the BHI differences between cells from younger and older subjects was statistically significant (p = 0.036) (Figure 3B). Of note, the majority (85%) of CD4⁺ T cells from younger participants had higher BHI compared to cells from older donors. Conversely, nonmitochondrial respiration, typically attributed to the action of cyclooxygenases, lipoxygenases and NADPH oxidases that are considered negative indicators of bioenergetic health, was significantly higher (p = 0.049) in CD4⁺ T lymphocytes from older compared to younger participants [31] (Figure 3C). We also found that the reserve mitochondrial capacity was significantly higher (p = 0.045) in cells from younger compared to older participants (Figure 3D). Cumulatively, these observations demonstrate that despite increased abundance of mitochondrial proteins, mitochondrial function of CD4⁺ lymphocytes is lower in older compared to younger individuals.

Figure 3. (A) Calculated bioenergetic health index (BHI) from young and old CD4⁺ T cells. The BHI is derived from calculating a ratio of positive aspects of mitochondrial bioenergetic function (i.e. reserve capacity and ATP-linked respiration) to potentially deleterious aspects of mitochondrial bioenergetic function (i.e. non-mitochondrial oxygen consumption and proton leak). Cellular mitochondrial function was determined using high-resolution respirometry with oligomycin, FCCP, rotenone, and antimycin A. For BHI, one outlier of 14 participants was noted and a comparison of the old and young BHI with this outlier was not significantly different (p=0.19). (B) In further analysis, FACS showed the outlier subject had a higher percentage of total memory CD4⁺ T cells (59% compared to a range 26-48%) than other participants. Adjusting for the average percentage of memory CD4⁺ T cells, the calculated the BHI was significantly higher for younger compared to older participants (*p = 0.036). (C) Nonmitochondrial respiration was found to be significantly higher in CD4⁺ T cells from older compared to younger individuals (*p = 0.049). (D) Reserve capacity was significantly higher in cells from young compared to older participants (*p = 0.045). (A–D) P-values were calculated by Welch’s t-test, a variation of the Student’s t-test that does not make the assumption of equal variance in the two compared samples [55]. Error bars reflect the standard error of the mean (±SEM). N = 7 young, 7 old donors.

Morphology of mitochondria and autophagic vacuoles in CD4⁺ T cells from young and old participants

We hypothesized that the apparent discrepancy between mitochondrial protein abundance and function reflected the persistence of dysfunctional mitochondria. To gain insight into cause of mitochondrial dysfunction, we visualized mitochondria number, mass (area) and shape in naïve and memory CD4⁺ T lymphocytes from younger (2 men and 3 women; ages 22–34 years-old) and older (4 men; ages 70–80 years-old) participants (Supplementary Table 5) using TEM. In both younger and older participants, memory CD4⁺ T lymphocytes had significantly higher mitochondria counts (p = 0.0267, p = 0.0008, respectively) than naïve CD4⁺ T lymphocytes (Figure 4C). These observations were consistent with early studies in mouse CD8⁺ memory T cells (32). The counts of structures identified as mitochondria were not significantly different between CD4⁺ naïve or memory T lymphocytes from older and younger participants (Figure 4A–4C). We calculated mitochondrial surface area as a surrogate for mitochondrial mass [32]. In cells from young individuals, average mitochondrial area of naïve and memory CD4⁺ T cells from young individuals were not significantly different (Figure 4D). Conversely, in cells from older individuals, average mitochondrial areas of naïve CD4⁺ T cells were significantly lower than those from CD4⁺ T memory cells (p = 0.0030). Though we noted a trend towards lower mitochondrial area in older compared to younger naïve CD4⁺ T cells, such a difference was not statistically significant in this limited cohort.

Figure 4. (A) TEM image of CD4⁺ T cell showing mitochondria, and (B) mitochondrion area (size) outlined in blue. (C) Memory (M) CD4⁺ T cells had a significantly higher mitochondria number than naïve (N) cells in cells from both young (Y) and old (O) (*p = 0.0267, ***p = 0.0008, respectively). (D) In memory CD4⁺ T cells of both young and old, the mitochondrial area was significantly greater compared to naïve CD4⁺ T cells (**p = 0.0030). Arbitrary Unit (A.U.). (C, D) P-values were calculated by Student’s t-test (two-tailed) using GraphPad PRISM 7 software. Error bars reflect the standard error of the mean (±SEM). N = 5 young, 4 old donors.

To determine if the higher abundance of mitochondrial proteins could be a consequence, at least in part, of altered mitochondrial turnover by autophagy, we carried out morphometric analysis of the two main subcellular compartments related to autophagy, autophagosomes (AP) and autolysosomes (AL), from young and old naïve and memory CD4⁺ T cells. In autophagy, cytosolic components are sequestered inside AP that then fuse with lysosomes (forming AL), where degradation takes place. This process of AP maturation to AL has been shown to be compromised in different cells from old rodents [33]. We set out to analyze possible differences in the total content of autophagic compartments (autophagic vacuoles; AV) and in their state of maturation using TEM and morphometry. For this analysis we used 54 TEM images (15 of naïve cells from 5 young donors, 12 of naïve cells from 4 older donors, 15 of memory cells from 5 young donors, and 12 of memory cells from 4 older donors; Figure 5A–5D and Supplementary Figure 4A, example images). The total count of AV (sum of AP and AL) was significantly higher (p = 0.0006, p = 0.0001, respectively) in naïve CD4⁺ T lymphocytes subsets from older compared to younger participants. This increase was mostly as result of higher AP abundance (p = 0.0001) in naïve CD4⁺ T lymphocytes subsets from older compared to younger participants (Figure 5E, 5F), whereas we found no age-related difference in the number of autolysosomes in both cell types (Figure 5G). These data support differences in AP maturation in CD4⁺ T cells from old individuals. In fact, when we quantified the fraction of AV in the AP or AL state, we found that both memory and naïve CD4⁺ T cells from older individuals displayed higher percentage of AV in an immature (AP) state compared to younger individuals (Figure 5H). These differences reached statistical significance in memory CD4⁺ T cells from older individuals (p = 0.022) where clusters of AP were a common occurrence (examples in Supplementary Figure 4A panels a, c and d). AP-occupied areas expressed in absolute terms or as percentage of total cytosol area were significantly higher (p = 0.0005, p = 0.0002, respectively) in older than younger participants in naïve CD4⁺ T cells but not in memory CD4⁺ T cells (Supplementary Figure 4E and 4H). In further support that the increase in AV was mostly due to reduced maturation/clearance rather than increased AP biogenesis, transcriptome analysis indicated that expression of most autophagy genes was not different in CD4⁺ T cells from both age groups, and Beclin-1 (BECN1), that participates in autophagy induction, was substantially underrepresented in CD4⁺ T lymphocytes from older compared to younger participants (Figure 5I). Only ATG7 was overrepresented in the older group, but whether this is a compensatory reaction or related with some of the non-autophagy related functions of ATG7 recently described [34], will require future investigation.

Figure 5. (A–D) Representative images showing autophagosomes and autolysosomes from young (Y) and old (O) naïve (N) and memory (M) CD4⁺ T cells. Circles in red indicate autolysosomes, in yellow indicate autophagosomes. (E) The number of autophagic vacuoles, the combination of autophagosomes and autolysosomes, was increased in naïve CD4⁺ T cells from older as compared to younger individuals (***p = 0.0006) and also increased in memory CD4⁺ T cells from young compared to naïve CD4⁺ T cells from young individuals (*p = 0.0424). (F) Autophagosomes were significantly higher in both naïve and memory CD4⁺ T cells from older individuals compared to naïve and memory CD4⁺ T cells of younger individuals (***p = 0.0001, *p = 0.0217, respectively). (G) No significant differences were found in the number of autolysosomes from young and old naïve and memory CD4⁺ T cells. (H) Significant differences were found in autophagic vacuole (AV) maturation of autophagosomes (AP) transitioning to autolysosomes (AL) between memory CD4⁺ T cells from older individuals compared to younger individuals (*p = 0.022). No differences were found for autophagic vacuole maturation for naïve CD4⁺ T cells. (E–H) P-values were calculated by Student’s t-test (two-tailed) using GraphPad PRISM 7 software. Error bars reflect the standard error of the mean (±SEM). N=5 old, 4 young donors. (I) Gene expression data for autophagy pathway genes in CD4⁺ T cells from older compared to young individuals showing up-regulation (above 0-1.5 Z- ratio) of some autophagy-related genes (ATG7, WIPI1, MAP1LC3A, ATG12, TMEM1, ATG3, ATG4D). Interestingly, BECN1 was down-regulated (-2.8 Z-ratio) in older compared to young individuals. N= 8 young, 25 old donors.

Overall, the TEM data showed in CD4⁺ naïve and memory T lymphocytes from older participants the permanence of dysmorphic mitochondria and higher abundance of undegraded immature AP. In many instances, mitochondria were observed inside the AP that accumulated in CD4⁺ T cells from older patients (Supplementary Figure 4A panels b–e), suggesting that part of the altered mitochondria were sequestered from the cytosol but failed to undergo degradation. These findings could explain why the higher content of mitochondrial proteins in older patient cells did not associate with higher number or size of mitochondria.

Mitophagy detection in CD4⁺ T cells

To better understand the basis for morphological differences between mitochondria present in CD4⁺ T cells from older compared to younger individuals, we assessed mitophagy flux using flow cytometry and a mitochondria-specific fluorescent dye whose emission intensity increases as organelle pH acidifies [35]. Using this method, mitochondria in cytoplasm can be distinguished from those within immature autolysosomes and autolysosomes because of their lower pHs [36]. We incubated CD4⁺ T cells from 24 donors (Supplementary Table 6) ranging in age from 23–90 years-old with Mtphagy Dye for 30 minutes, and then further cultured the cells for 18 hours with no added pharmacologic agents. We found that a greater proportion of mitochondria from older individuals had intermediate fluorescence intensity compared to cells from younger individuals (Figure 6A, 6B, top row marked ‘int’ for a representative younger and older individual [see Supplementary Figure 5A, 5B for dot plots for a representative younger and older individual]; quantified for all individuals in Figure 6C), supporting that mitochondria were healthier in the younger individuals and that, at a given time, the amount of mitochondria requiring to undergo mitophagy and already trapped in autophagic compartments was higher in cells from older individuals. To compare mitophagy flux in both groups of individuals, we forced mitochondrial degradation by culturing the cells for 18 hours in the presence of Carbonyl cyanide m-chlorophenyl hydrazone (CCCP, a mitochondrial transport uncoupler [37]), and found that this agent was still able to efficiently drive mitochondria into a high fluorescent intensity (low pH) compartment in CD4⁺ T cells from older donors (Figure 6A, 6B, middle row marked ‘hi’ for a representative younger and older individual; quantified for all individuals in Figure 6D). The results after adding CCCP support that cargo-recognition and mitophagy induction were still preserved in CD4⁺ T cells from older donors, and that as in the case of the autophagic flux analysis, the higher presence of mitochondria in intermediate pH compartments in the older group was a consequence of impaired maturation/degradation. We confirmed that bafilomycin A₁ treatment, which prevents vacuolar acidification [38], increased the fraction of dye in the intermediate fluorescence (immature autophagic compartments) in both groups of age (quantified for all individuals in Figure 6E), but differences between CD4⁺ T cells from young and older individuals were still noticeable, in support that defective clearance of autophagic compartments in the older group may not be due to differences in acidification but rather in cargo degradation. We infer that a greater proportion of mitochondria in CD4⁺ T cells from the elderly are already ‘stuck’ in autophagic compartments due to poor clearance in these compartments.

Figure 6. (A and B) Representative histograms from a young and old donor for all conditions using Mtphagy Dye (20,000 cells recorded). Summary data showing percentage of CD4⁺ T cells of young and old donors divided by fluorescent intensity (low [‘lo’], intermediate (‘int’), and high [‘hi’]) of Mtphagy Dye using, © Mtphagy Dye alone, (D) CCCP administration, and (E) Bafilomycin A₁ administration. ****p = 0.0001, ***p = 0.0002 and 0.0003, **p = 0.0081. (C–E) P-values were calculated by Student’s t-test (two-tailed) using GraphPad PRISM 7 software. Error bars reflect the standard error of the mean (±SEM). N = 12 young, 12 old donors.

Study populations

Samples were derived either from BLSA participants or normal donors. Participants enrolled in the BLSA (NIA protocol #03-AG-0325) are volunteers who at the time of study enrollment were “healthy” based on strict eligibility criteria. BLSA participants tend to be healthier than the general population. Young normal donors are volunteers who have donated blood through the Tissue Procurement for Biomedical Research protocol (NIA protocol #03-AG-N322) or apheresis material through the Cytapheresis of Volunteer Donors protocol (NIA protocol #03-AG-N316). All individuals were consented for their donations and protocols have been reviewed by the appropriate institutional review board at NIH.

Proteomics

Subjects

Twenty-two men (21–83 years-old) were enrolled. Samples were taken from apheresis material from older donors from the BLSA and younger normal donors through the NIA Cytapheresis of Volunteer Donors protocol. This study required a large quantity of WBCs that could only be obtained by apheresis.

Cell preparation

PBMCs were isolated from apheresis packs using Ficoll-Paque Plus (GE Healthcare, Piscataway, NJ, USA) density gradient. CD4⁺ T cells were obtained by positive selection using anti-human CD4 microbeads (Miltenyi Biotec, Auburn, CA, USA). Cells were kept overnight at 4 °C.

Protein extraction

6×10⁷ CD4⁺ T cells in 18 mL culture medium were plated. Cells were rested overnight at 4 °C in RPMI 1640 + 1% penicillin/streptomycin, 1% L-glutamine (P/S/G) + 10% fetal bovine serum (FBS) medium. Cells were then incubated for 4 hours in 5% CO2 at 37 °C. Cells were collected by centrifugation and used to prepare cytoplasmic extracts (CE) [29, 30]. CE proteins were quantified using the Bradford protein assay.

Protein digestion and labeling with iTRAQ reagent

Briefly, 50 μg of each extract were mixed with six volumes of cold acetone (−20 °C) for overnight precipitation. The samples were centrifuged at 4°C, acetone was removed, and pellets were air dried. The pellets were then resuspended in dissolution buffer (0.5 M triethylammonium bicarbonate [TEAB], 0.2% SDS), then reduced and alkylated before digesting overnight at 37 °C using sequencing-grade modified trypsin (Promega, Maddison, WI) at a protein-to-enzyme ratio of 20:1 according to iTRAQ 8plex labeling protocol (ABSciex, Foster city, CA). Using the iTRAQ kit (ABSciex, Foster City, CA), protocol peptides were then labeled with iTRAQ tags as seen in Supplementary Table 1. One common reference sample was used in all four different iTRAQ 8plex experiments, which would help in normalization among iTRAQ runs and also to compare samples from one iTRAQ to another. We also used a technical control sample within an iTRAQ run and used three technical control samples between iTRAQ runs. After stopping the labeling reaction, the resulting labeled peptides were all combined and dried. The samples were then resuspended in 0.1% formic acid to a final volume of 1 mL and then desalted using Waters Oasis HLB 1cm cartridges (Milford, MA) per the manufacturer’s instruction. The eluate was then dried and resuspended in reverse phase solvent buffer A (10 mM TEAB in water, pH 8.4-8.6). The combined peptide digest was fractionated offline by basic reverse phase fractionation followed by online reverse phase liquid chromatography-tandem mass spectrometry.

Peptide fractionation using basic reverse phase chromatography and LC-MS/MS analysis

The pooled mixture of iTRAQ-labeled peptides was fractionated using basic reverse phase chromatography on an Agilent 1200 HPLC using a XBridge C18 Column, 5µm, 2.1x100 mm (Waters, Milford, MA). A 60 min linear gradient from 0% to 40% Buffer B (10 mM TEAB in 90% Acetonitrile, pH 8.4-8.6)) was used to separate the peptides with a flow rate of 200 µL/min with a total run time of 90 minutes. Fractions were collected at 1-minute intervals on a 96-well microplate with a total of 90 fractions. The fractions were then pooled to a total of 12 fractions. Each fraction was completely dried using a SpeedVac (Thermo Fisher Scientific, San Jose, CA) and then resuspended in 0.1% formic acid before injection for LC-MS/MS analysis. Fractions were collected at 1-minute intervals on a 96-well microplate with a total of 90 fractions. The fractions were then pooled to a total of 12 fractions. Each fraction was completely dried using a SpeedVac and then resuspended in 0.1% formic acid before injection for LC-MS/MS analysis. Liquid chromatography-tandem mass spectrometry was performed using the Eksigent nanoLC-Ultra 1D plus system (Dublin, CA) coupled to an LTQ Orbitrap Elite mass spectrometer (Thermo Fisher Scientific, San Jose, CA). Ten µL of peptide digest was first loaded onto a Zorbax 300SB-C18 trap column (Agilent, Palo Alto, CA) and then separated on a reverse-phase PicoFrit analytical column (New Objective, Woburn, MA) using a 40-min linear gradient of 5–40% acetonitrile in 0.1% formic acid with a flow rate of 250 nL/min. LTQ-Orbitrap Elite settings were as follows: spray voltage 1.5kV; full MS range: m/z 300-2000, data-dependent acquisition. A full MS-scan was followed by six data-dependent MS2 scans for precursor ions.

Protein identification/data analysis

Protein identification and quantitation using the MS/MS spectra was performed using the Proteome Discoverer v1.3 software (Thermo Fisher Scientific, San Jose, CA) using Mascot search engine (Matrix Science, Boston, MA). The search criteria were set to the following: database, Swiss-Prot (Swiss Institute of Bioinformatics); taxonomy, human; enzyme, trypsin; miscleavages, 3; variable modifications, oxidation (M), deamidation (NQ), iTRAQ 8plex tyrosine; fixed modifications, (MMTS) methyl methanethiosulfonate ©, N-terminal iTRAQ 8plex, iTRAQ 8plex lysine, MS peptide tolerance, 10 ppm; MS/MS tolerance, 0.05 Da. Proteins where fold changes in protein expression were greater than 1.3 or less than 0.7 were examined.

Statistical analysis

Data were analyzed within each iTRAQ 8plex and also with all combined. Differences between young and old were determined by Student’s t-test (two-tailed) and proteins that met the following criteria were ranked by significance (only p < 0.05 shown). Proteins were selected that were up- or down-regulated (ratio >1.3 or <0.7 and had a p < 0.05). Averages of ratios were calculated between old and young (averages [old] / averages [young]) after normalization using the common control sample used across iTRAQ 8plex experiments. Two hundred and twenty-four proteins had ratios of 1.3 or greater and 56 proteins had ratios of <0.7 or smaller. Proteins were then selected according their p-value (p < 0.05). One hundred and four proteins fitted the criteria indicated above out of 3,510 proteins after this three-step analysis. Twenty-nine of these 104 proteins were determined to be significantly up-regulated in the old after accounting for unique peptides (≥ 2, similar peptides sequences not counted), high confidence of the peptide sequences (medium and low confidence peptide sequences eliminated), and ion score of more than 30. DAVID v.6.7 (bioinformatics tool) was then used to convert between protein identifiers/gene. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD016039. .

Gene expression

Thirty-three donors (24 men and 9 women) were enrolled, age range of 25–82 years-old (Supplementary Table 3). All blood samples were taken from older donors from the BLSA and one younger normal donor through the NIA Cytapheresis of Volunteer Donors protocol.

Cell preparation

PBMCs were isolated from apheresis packs using Ficoll-Paque Plus (GE Healthcare, Piscataway, NJ, USA) density gradient. CD4⁺ T cells were obtained by positive selection using anti-human CD4 microbeads (Miltenyi Biotec Auburn, CA, USA) [29]. CD4⁺ T cells in RPMI 1640+1% P/S/G+10% FBS medium were incubated for 4 hours at 37ºC (5% CO₂ incubator). CD4⁺ T cells were then washed and collected by centrifugation. Cell pellets were then lysed in 350 μl RLT buffer (QIAGEN Inc, Valencia, CA, USA) for total RNA extraction [29, 30].

Total RNA purification and microarray hybridization

Total RNA was extracted from frozen cell pellets containing 5×10⁶ cells using the Qiagen RNeasy Mini Kit (QIAGEN Inc, Valencia, CA, USA) as previously described [29].

Microarray hybridizations

Total RNA was used to generate biotin-labeled single-strand RNA (cRNA) using the Illumina Total Prep RNA Amplification Kit (Ambion, Austin, TX, USA) according to the manufacturer's recommendations as previously described [30]. 0.75 μg of biotin-labeled cRNA was hybridized for 16 hours at 58^oC to Illumina's Sentrix HumanRef-8 Expression BeadChips (Illumina, San Diego, CA, USA). The arrays were washed and blocked, and then the labeled cRNA was detected by streptavidin-Cy3 staining. Arrays were scanned using an Illumina BeadStation 500X Genetic Analysis Systems scanner and the image data extracted using Illumina BeadStudio software, version 3.0 [29].

Microarray data analysis

Microarray data were analyzed using DIANE 6.0 as previously described [29, 30]. Raw microarray data were subjected to filtering by the detection p-values and then normalized by Z-transformation with log signal values. The data were further tested for significant changes as previously described [29]. Individual genes with Z-ratio ≥ 1.5, p-value ≤ 0.05 and false discovery rate (FDR) ≤ 0.3 were considered significantly changed. The PAGE algorithm was employed for gene set enrichment analysis by using all the genes in each sample as input against and the data set supplied by Gene Ontology Institute and pathway gene set of the MIT Broad Institute molecular signature database. For each relevant comparison, the lists of differentially expressed genes and Z-ratios were entered into the PAGE Pathway Analysis software to organize them according to known biological pathways. The enrichment Z-scores for each functional grouping were calculated based on the chance of mRNA abundance changes predicting these interactions and networks by Z-test. Pathways p- ≤ 0.05 and pathways FDR ≤ 0.3 were the cutoff criteria for significant pathway selection. Up and down top 50 pathway Z-scores were averaged amongst young (Y) (20-30) and old (O) (70-80) donors, and Z-score differences between young and old are shown on the X-axis. Each row denoted a different pathway (p ≤ 0.05 and FDR ≤ 0.3).

Accession number

The microarray GEO accession numbers for the data reported in this paper is GSE131407.

Mitochondrial respiration

Subjects

Fourteen Caucasian men (23–82 years-old) were enrolled for this study. All blood samples were taken from old donors selected from the BLSA and younger donors selected from the NIA Tissue Procurement for Biomedical Research protocol. All participants were fasting at time of blood draw.

Cell preparation

PBMCs were isolated from 5 of 10 mL heparinized whole blood tubes. CD4⁺ T cells were isolated from PBMCs using the immunomagnetic negative selection kit (EasySep Human CD4⁺ T cells Enrichment Kit; StemCell Technologies, Vancouver, BC, Canada) using RoboSep, a fully automated cell separator (StemCell Technologies) according to the manufacturer's protocol.

Flow cytometry

CD4⁺ T cells were stained using the following fluorescent antibodies: CD4 (FITC), CD45RA (APC), CD62L (PECy7) from BD Bioscience. Cells were then analyzed by flow cytometry (BD FACSCANTO II). Data were analyzed to define % of naïve, central memory (CM), effector memory (EM) CD4⁺ T cell percentages using FLOWJO software (version 10).

Mitochondrial respiration

Mitochondrial respiration was measured using high-resolution respirometry (Oxygraph-2 k; Oroboros Instruments, Innsbruck, Austria) [31]. The O2k is a 2-chamber high resolution respirometer for monitoring oxygen consumption from small amounts of biological samples. A suspension of CD4⁺ T cells in culture medium (RPMI 1640 medium, 10% FBC, 1% P/S/G) was added to the Oxygraph-2k chambers at a concentration of 5×10⁶ cells in 2 mL. After 10 minutes of stabilization at 37 °C, the coupling control protocol was applied to evaluate the cellular routine respiratory state, the mitochondrial coupling state, non-coupled respiratory capacity, and rotenone-insensitive or residual oxygen consumption [52]. Briefly, Oligomycin (2 μg/mL) was added to inhibit ATP synthase and induce non-phosphorylating (LEAK) respiration (electron flow coupled to proton pumping to compensate for proton leaks). Maximal respiratory capacity of the electron transfer system was subsequently obtained by titrating the uncoupler FCCP (injected stepwise up to 1–1.5 μM). Finally, rotenone (0.5 μM) and antimycin-A (2.5 mM) were added to measure the non-coupled respiratory capacity, and rotenone-insensitive or residual oxygen consumption.

Bioenergetic health index (BHI)

The BHI, a summary index of mitochondrial function in cells [31] was calculated for each sample. The BHI is derived from calculating a ratio of positive aspects of mitochondrial bioenergetic function (i.e. reserve capacity and ATP-linked respiration) to potentially deleterious aspects of mitochondrial bioenergetic function (i.e. non-mitochondrial oxygen consumption and proton leak) as follows:

$\text{BHI}=\log =\frac{\left(\text{reserve}\text{capacity}\right)\times \left(\text{ATP-linked}\right)}{\left(\text{non-mitochondrial}\right)\times \left(\text{proton\hspace{0.17em}leak}\right)}$

Transmission Electron Microscopy (TEM)

Nine donors (6 men and 3 women) were analyzed for this study, age range 22–80 years-old. Older subjects for this study were selected from the BLSA, younger donors were from the NIA Tissue Procurement for Biomedical Research protocol.

Cell preparation

PBMCs were isolated from 50 mL apheresis heparinized blood using Ficoll-Paque Plus (GE Healthcare, Piscataway, NJ, USA) density gradient.

CD4⁺ T cells

CD4⁺ T cells were isolated from PBMCs using the immunomagnetic negative selection kit (EasySep Human CD4⁺ T cells Enrichment Kit; StemCell Technologies, Vancouver, BC, Canada) using a fully automated cell separator (RoboSep, StemCell Technologies) according to the manufacturer's protocol.

Flow cytometry

CD4⁺ T cell were stained using the following fluorescent antibodies: CD4 (FITC), CD45RA (APC), CD62L (PECy7) from BD Bioscience, then cells were sorted to naïve and memory cells using a BD FACSAria flow cytometer.

TEM

Five to seven million naïve and also 5–7 × 10⁶ memory CD4⁺ T cells were fixed in 5% glutaraldehyde (GA), 6 mM MgCl_2, in 0.2 M sodium cacodylate buffer for 10 minutes at 4 °C then kept in 2.5% GA, 3 mM MgCl_2, in 0.1 M sodium cacodylate buffer (pH 7.2-7.4) under continuous motion (rotary mixer) overnight at 4°C.

After buffer rinse, samples were postfixed in 1% osmium tetroxide, 0.8% potassium ferrocyanide in 0.1 M sodium cacodylate buffer (1 hour) on ice in the dark. Following a DH₂O rinse, plates were stained with 2% aqueous uranyl acetate (0.22 μm filtered, 1 hour in the dark), dehydrated in a graded series of ethanol, and embedded in Eponate 12 (Ted Pella) resin. Samples were polymerized at 60°C overnight.

Thin sections, 60 to 90 nm, were cut with a diamond knife on the Reichert-Jung Ultracut E ultramicrotome and picked up with 2×1 mm formvar coated copper slot grids. Grids were stained with 1% tannic acid, 2% uranyl acetate in 50% methanol and 0.4% lead citrate before imaging on a Philips CM120 at 80 kV. Images were captured with an AMT XR80 CCD (8 megapixel) camera.

Images

Twenty to 40 cell images were taken per sample. For each cell, we obtained whole cell images (direct magnification 17,500×) and also partial cell images of mitochondrial areas (one to three depending on mitochondrial areas; Direct Mag 33,000×). We processed 368 cell images and counted 3,059 mitochondria. We traced the perimeter of each mitochondrion to find out the area (size). We analyzed 54 images for morphometric analysis.

Morphometric analysis

AV were identified using previously established criteria [53, 54]. AV (vesicles, 0.5 mm) were classified as AP when they met two or more of the following criteria: double membranes (complete or at partially visible), absence of ribosomes attached to the membrane’s cytosolic side, luminal density similar to cytosol, and/or identifiable organelles or regions of organelles in their lumen. Vesicles of similar size but with a single membrane (or less than 40% of the membrane having a double portion), luminal density lower than the surrounding cytosol, or multiple single membrane-limited vesicles containing light or dense amorphous material were classified as AL. Total AV were composed of the sum of AP and AL. Maturation of AV was calculated as the percent of AL making up AV.

Mitophagy detection in CD4⁺ T cells

Subjects

Twenty-four donors (12 men [6 Caucasian and 6 African American] and 12 women [6 Caucasian and 6 African American]) were enrolled for this study. Blood samples were donated from “healthy’ volunteers selected from the BLSA and Tissue Procurement for Biomedical Research protocols. All blood draws were from fasting participants.

Cell preparation

PBMCs were isolated from four 10 mL heparinized whole blood tubes. CD4⁺ T cells were isolated from PBMCs using the CD4⁺ negative selection kit (StemCell Technologies, Vancouver, Canada) and RoboSep (StemCell Technologies) according to the manufacturer's protocol. Resting CD4⁺ T cells in 10 mL RPMI 1640 complete medium (10% FBS and 1% P/S/G) were incubated for 2 hours at 37°C. Cells were divided into 0.5×10⁶ cells in 0.5 mL 1640 media for four different conditions for mitophagy detection.

Mitophagy detection

Mitophagy detection experiments were performed using the Mitophagy Detection Kit (Dojindo Molecular Technologies, Inc, Rockville, MD, USA) for live cells using manufacturer's instruction with small modifications. Briefly, 2×10⁶ CD4⁺ T cells were divided into four tubes containing serum-free RPMI. Tube #1 was for unstained cells that followed all washing and media changing processes. 100 nmol/l Mtphagy Dye working solution was added to tubes #2, 3, 4 and then all tubes were incubated at 37°C for 30 minutes. The cells were then washed with serum-free medium. After discarding the supernatant, complete medium (RPMI 1640, 10% FBS, 1% P/S/G) was added to all the tubes. Ten μmol/l CCCP (Sigma-Aldrich) mitophagy-inducer was then added to tube #3, 100 nmol/l Bafilomycin A₁ (Sigma-Aldrich) autophagy inhibitor was added to tube #4. All tubes were then incubated at 37 °C for 18 hours. After 18 hours, cells were washed with FACS buffer. Mtphagy Dye fluorescence detection by flow cytometry (BD FACSCANTO II) was performed using 488 nm for excitation and 695 nm for emission (this corresponds to the PerCP cy5.5 channel). Data were analyzed using FLOWJO software (version 10).

The Mitophagy Detection Kit was designed for easy use for adhesive cells such as HeLa cells to detect in live cell imaging of mitochondrial autophagy using confocal microscopy. Since CD4⁺ T cells are primary suspension cells, tubes were used instead of the plates for fluorescence detection by flow cytometry. Some images were taken using confocal microscopy to examine the intensity (weak and strong) of the Mtphagy Dye for mitophagy detection but the quantification experiments were performed using flow cytometry (BD FACSCANTO II) and analyzed using FLOWJO V10 software.