Peripheral Mononuclear Cell Mitochondrial Function Associates with T‐Cell Cytokines in Parkinson's Disease

Abstract

Background

Parkinson's disease (PD) is the most common neurodegenerative movement disorder and one of the world's fastest‐growing neurological diseases. Although the exact causes of PD are unknown, mitochondrial dysfunction and inflammation may have significant roles in disease progression. As well as being prevalent in the brain, there is also evidence that peripheral mitochondrial dysfunction and inflammation occur in PD. However, if/how peripheral mitochondrial dysfunction and inflammation are linked is still unclear.

Objectives

This study aimed to determine the extent that mitochondrial dysfunction in peripheral immune cells is associated with inflammation in PD.

Methods

The study comprised of 35 controls and 35 PD patients that were age and sex matched. Flow cytometry was used to assess mitochondrial content and superoxide production in mononuclear cells, in the presence and absence of the mitochondrial stressor antimycin A. Serum inflammatory cytokines were measured by ELISA.

Results

Superoxide levels were significantly increased in PD patient mononuclear cells at baseline, and PD mononuclear cells had an impaired response to antimycin A. Immune cell superoxide levels correlated with serum cytokines associated with T‐cell responses, namely interleukin (IL) IL‐12, interferon‐γ, and IL‐17A.

Conclusions

Results show that mitochondrial dysfunction is prevalent in PD immune cells and may contribute to an inflammatory phenotype. © 2025 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Parkinson's disease (PD) is the most common neurodegenerative movement disorder worldwide, with a current prevalence of over 6 million individuals that is projected to double over the next 30 years. ¹ PD is a synucleinopathy, pathologically defined by the formation and accumulation of α‐synuclein enriched Lewy bodies and Lewy neurites in various brain regions, along with the progressive degeneration of dopamine producing neurons in the substantia nigra. PD symptoms are heterogeneous, having diverse motor and non‐motor symptoms that become more prevalent during different disease stages. ² PD diagnosis relies on the presence of cardinal motor symptoms, namely bradykinesia with rigidity and/or tremor, but non‐motor symptoms including autonomic dysfunction, neuropsychiatric features, sleep disturbances, and hyposmia can present many years prior to classical motor symptoms. ³ , ⁴ Tracking the clinical trajectory of PD has remained difficult because of an incomplete understanding of the underlying pathological mechanisms and a lack of objective biomarkers to inform diagnosis and prognosis.

Mitochondrial dysfunction and increased oxidative stress have emerged as key pathological mechanisms involved in the process of dopaminergic neuron death and are pathophysiological features that are observed in both genetic and sporadic PD. ⁵ , ⁶ , ⁷ Normal mitochondrial function is essential to the central nervous system (CNS) because of its significant role in energy production, calcium regulation, mitophagy, and homeostasis. Subsequently, impaired mitochondrial function impacts several cellular pathways, leading to increased reactive oxygen species (ROS), adenosine triphosphate (ATP) depletion, decreased mitochondrial complex 1 enzyme activity and caspase 3 activation. ⁸ In addition to the CNS, mitochondrial dysfunction may also be prevalent in the periphery of PD patients. A study of mitochondrial function and glycolysis in peripheral blood monocular cells (PBMCs) in sex matched healthy controls, found increased ROS levels and reduced mitochondrial content in PD monocytes. ⁹ Increased mitochondrial respiration has been demonstrated in PD patient PBMCs, ¹⁰ as well as higher levels of intracellular ATP, ¹¹ and significantly altered expression of genes involved in mitochondrial oxidative phosphorylation and mitochondrial quality control. ¹¹ , ¹² Mitochondrial DNA damage is also elevated in PBMCs from patients with PD. ¹³ These results suggest that mitochondrial function in peripheral mononuclear cells occurs in PD and may contribute to disease pathophysiology and/or have biomarker utility.

There is also evidence for low grade peripheral inflammation in PD. ¹⁴ , ¹⁵ Although not observed in every study, inflammatory cytokines can be increased in PD patient blood samples, ¹⁶ and PD patient immune cells may be hypersensitive to and/or have altered responses to innate immune stimuli. ¹⁷ , ¹⁸ , ¹⁹ Systemic activation of the innate immune system with lipopolysaccharide is commonly used to induce neurodegeneration in rodent models, ²⁰ , ²¹ and higher levels of peripheral inflammatory cytokines may exacerbate the progression of PD. ²² , ²³ Inflammatory bowel disease (IBD) can increase the risk of developing PD, but this risk is ameliorated in IBD patients taking strong immunosuppressing therapies. ²⁴ , ²⁵ Therefore, there is a strong link between inflammation and PD, although the exact drivers of inflammation and the mechanisms by which it contributes to PD remain unclear.

Because peripheral immune cells are major mediators of the inflammatory response, and these cells may have dysfunctional mitochondria in PD, then it is possible that mitochondrial dysfunction itself may be an inflammatory driver. Indeed, a large body of evidence now indicates that mitochondria are powerful regulators of innate immunity affecting cytokine responses downstream of many immune pathways, including toll‐like receptors, Rig‐1 like receptors, and the inflammasome. ²⁶ , ²⁷ Therefore, this study aimed to determine the extent to which mitochondrial dysfunction in PD patient peripheral immune cells may associate with serum levels of inflammatory cytokines.

Patients and Methods

Participant Details

For optimization of the mitochondrial flow cytometry assay, buffy coat samples were provided by the Australian Red Cross Lifeblood association with ethical approval granted by the University of Sydney (2017/847). PD and neurologically normal control subjects were recruited via the PD research clinic at the Brain and Mind Centre, University of Sydney with ethical approval granted by University of Sydney (2017/826). All recruited participants provided written informed consent. PD patients were diagnosed according to clinically established criteria. ²⁸ The Movement Disorder Society Unified Parkinson's Disease Rating Scale part three, Montreal Cognitive Assessment (MoCA), Mini Mental State Examination (MMSE), Hoehn and Yahr staging scale (H&Y) and disease duration were recorded at the time of blood collection by a trained neurologist. Demographic and clinical data are provided in Table 1. Information on anti‐inflammatory medication use was recorded and is provided in Supplementary Table S1, with seven control and 13 PD participants reporting anti‐inflammatory medication use, predominantly the daily use of aspirin.

TABLE 1.

Demographic and clinical data of PD patients and controls

	Control	PD	P value
Age (y)	67 ± 1.48	72 ± 1.81	0.05
Sex (%M)	57%	57%	>0.99
Disease duration	N/A	9.21 ± 1.34	N/A
MDS‐UPDRS‐III	4.97 ± 0.94	34.65 ± 2.89	<0.0001****
MMSE	28.41 ± 0.29	27.53 ± 0.60	0.63
MoCA	26.82 ± 0.45	28.06 ± 0.85	0.16
LEDD	N/A	794.45 ± 143.75	N/A
H&Y	N/A	2.15 ± 0.18	N/A
Post thaw PBMC viability (%)	85 ± 1.5	81 ± 1	0.02*
Sample storage time (mo)	19 ± 3	11 ± 2	0.03*

Note: Data shows mean ± standard error. Significance was determined at P value <0.05. *P < 0.05.

****P < 0.0001.

Abbreviations: PD, Parkinson's disease; MDS‐UPDRS‐III, Movement Disorder Society Unified Parkinson's Disease Rating Scale part three; MMSE, Mini Mental State Examination; MOCA, Montreal Cognitive Assessment; LEDD, levodopa equivalent daily dose; H&Y, Hoehn and Yahr; PBMC, peripheral blood monocular cells.

PBMC Isolation from Buffy Coat

Optimization was performed using buffy coat samples from healthy blood donors. Up to 60 mL of buffy coat was carefully layered into 2 × 50 mL tubes containing 15 mL of Ficoll (Sigma Aldrich, MO) and centrifuged at 400× g for 30 minutes at 21°C (acceleration at 4 and deceleration at 0). The PBMC layers from both 50 mL tubes were transferred into a new 50 mL tube. A total of 40 mL of warmed complete RPMI media (RPMI supplemented with 10% heat inactivated fetal bovine serum (FBS) and 1× penicillin and streptomycin solution) was added to the new tube, which was then centrifuged at 300 × g for 10 minutes at 21°C (acceleration 9 and deceleration at 9). The supernatant was discarded and PBMCs resuspended in complete RPMI media for counting using 0.4% trypan blue stain (ThermoFisher Scientific, MA) and a Countess automated cell counter (ThermoFisher). Following counting, cells were again centrifuged at 300 × g for 10 minutes at 21°C (acceleration at 9 and deceleration at 9), the supernatant discarded, and PBMCs resuspended in cryopreservation media comprised of RPMI, 20% FBS and 10% dimethyl sulfoxid, in cryovials at a density of 1 × 10⁷ cells/mL and placed into a MrFrosty Freezing Container (ThermoFisher) containing isopropanol (Sigma‐Aldrich) overnight in a −80°C freezer. PBMC samples were then transferred to a vapor phase liquid nitrogen tank until required for experiments.

PBMC Isolation from Whole Blood

Up to 18 mL of whole blood was collected by venepuncture into sodium heparin tubes (BD Biosciences, NJ). PBMCs were isolated from whole blood following the SepMate manufacturer instructions (STEMCEll Technologies, Canada). Specifically, whole blood was gently mixed with an equal volume of phosphate‐buffered saline (PBS) containing 2% heat inactivated low endotoxin qualified FBS (ThermoFisher) in a 50 mL tube. Diluted blood was then carefully transferred to a 50 mL SepMate tube (STEMCELL Technologies) containing 17 mL of density gradient medium. Samples were centrifuged at 1200× g for 10 minutes at 21°C (acceleration at 9 and deceleration at 9) and the enriched PBMC layer transferred to a new 50 mL tube. PBMCs were washed, counted, and cryopreserved as above.

Serum Collection

Serum from each participant was collected in 8.5 mL Serum Separator Tubes (SST) II Advanced tubes (BD Biosciences). Serum samples were incubated at room temperature for 30 minutes to allow blood to clot prior to centrifugation at 1500× g for 15 minutes at 21°C (acceleration at 9 and deceleration at 5). Serum was then transferred into a 15 mL tube, mixed by inversion, partitioned into 500 μL aliquots and immediately placed at −80°C until required for experiments.

Flow Cytometry Measurement of Mitochondrial Function

Cryopreserved PBMC samples were thawed in a 37°C bead bath until small ice crystals remained. Cells were then slowly pipetted into a 15 mL centrifuge tube containing 7 mL of pre‐warmed complete RPMI media. Samples were centrifuged at room temperature at 300 × g for 10 minutes, supernatant removed, and cells resuspended in 8 mL of complete RPMI media. Cell count and viability was determined using 0.4% trypan blue stain (ThermoFisher, T10282) and a Countess automated cell counter (ThermoFisher, C10228). Cells were plated at a density of 1 × 10⁶/mL and left to recover from thawing for 2 hours in a 37°C tissue culture incubator. Samples receiving treatment were stimulated with 2 μm of antimycin A for 2 hours. Following treatment, 150 μm of fluorescence‐activated cell sorting (FACS) buffer (1× PBS, 1 mM ethylenediaminetetraacetic acid, 25 mM HEPES, 1% heat inactivated FBS) was added to the cells and samples were centrifuged at 300 × g for 5 minutes. Cells were then stained, shielded from light, with 5 μM MitoSOX Red (ThermoFisher, M36008) and 100 nm MitoTracker Deep Red (ThermoFisher, M22426) for 30 minutes at 37°C. Cells were then washed twice with FACS buffer and stained with Live/Dead viability dye (BD Biosciences, 5656388) at 1:1000 dilution and FcR Blocking Reagent (Miltenyi Biotec, Germany, 130–059‐901) at a 1:20 dilution for 30 minutes at 4°C, shielded from light. Following this incubation, cells were again washed with FACS buffer and then incubated with PE‐Cy7 conjugated CD14 primary antibody (BD Biosciences, 557742) at 1:40 dilution for 30 minutes at 4°C, shielded from light. Cells were then finally washed twice again with FACS buffer, resuspended in 300 μL of FACS buffer and transferred to 5 mL FACS tubes (Falcon, 352235). Samples were kept on ice and analyzed immediately on a Cytek Aurora Spectral Analyser (Cytek Biosciences, CA) with 100,000 events recorded per sample. FCS files were analyzed using FlowJo Version 10.8 (BD Biosciences).

Serum Inflammatory Cytokine Measurement

Meso Scale Discovery (MSD) S‐Plex Proinflammatory (MSD, K15396S), and U‐Plex Interferon (MSD, K1509) kits were used to measure serum cytokines. Serum was thawed, centrifuged at 2000× g for 3 minutes at 4°C and kept on ice while assay reagents were bought to room temperature. For the pro‐inflammatory 9‐plex assay, 25 μL of undiluted serum sample was added per well. For interferon measurements, serum was diluted 2‐fold as recommended by the manufacturer and 50 μL added per well. Assays were performed as per manufacturer's instructions. Plates were read on the MESO QuickPlex SQ 120MM plate reader (Meso Scale Discovery, MD). The calibration curves used to calculate the serum cytokine concentrations were established by fitting the signals from the calibrators using a four‐parameter logistic model with a 1/Y 2 weighting. Data was obtained and analyzed using the MSD Discovery Workbench software Version 4.0. All calibrators and samples were performed in duplicates with average values used for the final statistical analysis.

Statistical Analysis

Mann–Whitney U tests were used to assess demographic and clinical differences between PD patients and controls. For mitochondrial flow cytometry and cytokine measures, comparisons between PD and control were performed using Students t test. A repeated measures t test was used to determine the increase in superoxide production in response to antimycin A. For correlation analyses, Spearman, Pearson, or Partial Correlations covarying for anti‐inflammatory use were performed as indicated. The effects of viability, sex and storage time on measured outcomes was performed by including these variables as covariates in analysis of variance (ANCOVA) tests. For all statistical tests, significance was accepted at P < 0.05. Statistical analysis was performed using SPSS (IBM, NY), with graphs generated using Prism Software (GraphPad, CA).

Data Sharing

Data supporting the outcomes of this manuscript are available from the corresponding author on reasonable request.

Results

In House Establishment of Live Cell Flow Cytometry Mitochondrial Assay

We first sought to establish in house, a flow cytometry assay for mitochondrial function that has previously been described. ⁹ PBMCs from healthy blood donors were used to set up the mitochondrial assay gating on CD14 positive monocytes (example gating strategy shown in Supplementary Fig. S1A,B). A time course of incubating PBMCs with the fluorescent mitochondrial probes indicated that 30 minutes was the time point for maximal MitoTracker signal (Supplementary Fig. S1C), whereas MitoSOX continued to rise over the time course (Supplementary Fig. S1D). Based on these results a 30‐minute incubation was chosen for both probes so MitoSOX and MitoTracker could be plexed together in the same assay. As a positive control, the mitochondrial respiratory chain inhibitor antimycin A was able to increase MitoSOX signal under the chosen conditions (Supplementary Fig. S1E).

Increased Mitochondrial Content and ROS in PD Mononuclear Cells

Mitochondrial function was then assessed in PBMCs from control and PD patients. Under baseline conditions, there was a significant increase in the fluorescence intensity from the MitoSOX probe, indicating increased superoxide in PD samples compared to controls, regardless of whether cells were gated on all viable PBMCs (Fig. 1A), CD14 positive monocytes (Fig. 1B), or CD14 negative cells (predominantly lymphocytes) (Fig. 1C). The same result of significantly increased fluorescent intensity in PD samples was also measured for the MitoTracker probe, indicating increased mitochondrial content, for all PBMCs (Fig. 1D), CD14 positive monocytes (Fig. 1E), and CD14 negative cells (Fig. 1F). A significant positive correlation was seen between levels of MitoSox and Mitotracker for the cell populations across the three different gates used (Fig. 1G–I).

FIG. 1.

Significantly elevated mitochondrial superoxide levels and mitochondrial content in Parkinson's disease (PD) mononuclear cells. Flow cytometry reveals increased levels of mitochondrial superoxide in PD patient (red dots) peripheral blood monocular cells (PBMCs) (A), CD14 positive monocytes (B), and CD14 negative cells (C) compared to controls (blue dots). Elevated levels of mitochondrial content were also observed in PD PBMCs (D), CD14 positive monocytes (E), and CD14 negative cells (F). Significant Pearson correlations between mitochondrial superoxide and mitochondrial content were observed for PBMCs (G), CD14 positive monocytes (H), and CD14 negative cells (I). Significance was determined at P value <0.05. *P < 0.05, **P < 0.01. Graphs show individual data points as well as the mean and standard error for each group. [Color figure can be viewed at wileyonlinelibrary.com]

PBMC Viability and Storage Time Can Impact Mitochondrial Measurements

As the post‐thaw viability of the PBMCs was significantly different between the control and PD groups (Table 1), we determined the impact of this on mitochondrial measures by including viability as a covariate in analysis of variance tests. Even though a live‐dead stain was included for gating on viable cells, post thaw viability had a significant effect on MitoSox signal in all PBMCs (P = 0.008) and CD14 negative cells (P = 0.013), but not in CD14 positive monocytes (P = 0.347). In all cases, however, the MitoSox signal was still significantly increased in the PD group (Supplementary Fig. S2A–C). Viability did not have a significant effect on MitoTracker signal in all PBMCs (P = 0.121), CD14 negative cells (P = 0.168), or CD14 positive monocytes (P = 0.247). However, when viability was included as a covariate, increased MitoTracker signal in the PD group only trended to increase and was no longer statistically significant (Supplementary Fig. S2A–C). To determine if storage time of the PBMCs was impacting on viability we calculated the storage time in months for each of the PBMC samples obtained. We found there was a significant difference in storage time. However, it was control samples that had been stored on average for longer than PD samples (Table 1), indicating that reduced viability of the PD PBMCs was not simply because of longer time in storage. To determine how storage time impacted on mitochondrial measures we again used analysis of variance using storage time as a covariate. Storage time also had a significant effect on MitoSox signal in all PBMCs (P = 0.009), CD14 negative cells (P < 0.001), and CD14 positive monocytes (P = 0.035). In all cases, the MitoSox signal was still significantly increased in the PD group (Supplementary Fig. S3A–C). Storage time also had a significant effect on MitoTracker signal in all PBMCs (P = 0.008), CD14 negative cells (P = 0.003), and CD14 positive monocytes (P = 0.006). When storage time is included as a covariate there was no longer a significant increase in MitoTracker signal in the PD group (Supplementary Fig. S3A–C). These results confirm that mitochondrial superoxide levels are increased in PD immune cell populations, but that PBMC viability and storage time are potential confounders that should be considered and matched in subsequent studies.

Mitochondrial Superoxide Levels Correlate to Age and MoCA

To determine if increased mitochondrial superoxide in the PD PBMCs associated with the obtained clinical variables, MitoSOX fluorescence and clinical measures were correlated. In those with PD, there were significant positive correlations between age at collection for MitoSOX fluorescence in all PBMCs, CD14 positive monocytes, and CD14 negative cells (Supplementary Table S2). There was also a significant negative correlation between MitoSOX fluorescence and MoCA scores, but this only occurred for the CD14 negative cells (Supplementary Table S2). There was no association between mitochondrial superoxide and dopamine medication, disease duration, motor symptom severity scales, or the MMSE (Supplementary Table S2). We also performed analysis of variance using sex and age as covariates, with age having a significant effect on MitoSOX signal in all PBMCs (P = 0.001), CD14 positive monocytes (P = 0.003), and CD14 negative cells (P < 0.001). For all cell types, mitochondrial superoxide levels were still increased with PD following adjustment (Supplementary Fig. S4A–C). Sex had no significant effect on MitoSOX signal and neither age nor sex had a significant effect on MitoTracker signal for all cell types. When age and sex were included as a covariates there was no longer a significant increase in MitoTracker signal in the PD group (Supplementary Fig. S4D–F).

PD Monocytes Have a Blunted Response to the Mitochondrial Stressor Antimycin A

We next determined whether superoxide production in response to antimycin A differed between PD and control mononuclear cells. Antimycin A had a significant effect to increase superoxide production in the control cells regardless of whether cells were gated on all viable PBMCs (Fig. 2A), CD14 positive monocytes (Fig. 2B), or CD14 negative cells (Fig. 2C). However, in PD samples, antimycin A had a reduced effect to increase superoxide production in all PBMCs and CD14 negative cells, and superoxide production was not increased at all in the PD CD14 positive monocytes (Fig. 2D–2F). We also examined the effect of antimycin A on MitoTracker signal. There was a small, but statistically significant effect of antimycin A to increase MitoTracker signal in control PBMCs (Fig. 2G), CD14 positive monocytes (Fig. 2H), or CD14 negative cells (Fig. 2I). In contrast, in PD samples there was no significant effect of antimycin A on MitoTracker Signal (Fig. 2J–L).

FIG. 2.

Antimycin A stimulation of control and Parkinson's disease (PD) patient samples. Stimulation of control participant samples without (blue dots) or with (red dots) 2 μM of mitochondrial respiratory chain inhibitor antimycin A for 2 hours significantly increased mitochondrial superoxide levels in control subject peripheral blood monocular cells (PBMCs) (A), CD14 positive monocytes (B), and CD14 negative cells (C). In PD patient cells, antimycin A stimulation significantly increased mitochondrial superoxide in PBMCs (D), but not CD14 positive monocytes (E). An effect of antimycin A stimulation to increase superoxide levels was also observed in PD CD14 negative cells (F). The same antimycin A treatment also significantly increased Mitotracker signal in control PBMCs (G), CD14 positive monocytes (H) or CD14 negative cells (I), but not PD patient PBMCs (J), CD14 positive monocytes (K), or CD14 negative cells (L). Significance was determined at P < 0.05. *P < 0.05, **P < 0.01, ****P < 0.001. Graphs show individual data points as well as the mean and standard error for each group. [Color figure can be viewed at wileyonlinelibrary.com]

Mononuclear Cell Superoxide Levels Correlate to Serum Inflammatory Cytokines

We next determined if mitochondrial dysfunction associated with levels of inflammatory cytokines. Measured cytokines tended to be higher in PD serum, although none were statistically significantly different between the two groups, including if participants taking anti‐inflammatory medications were excluded from the analysis (Supplementary Table S3). Across all participants and covarying for anti‐inflammatory medication use, there was a significant positive correlation with serum levels of interleukin (IL)‐12 and interferon‐γ (IFN‐γ) and superoxide production in all mononuclear cell populations (Table 2). Serum levels of IL‐17A and IL‐2 also correlated with superoxide production in the total PBMC and non‐monocyte PBMCs groups (Table 2). Inflammatory cytokines also correlated with each other (Supplementary Table S4). These results suggest that superoxide production in mononuclear cells may contribute to peripheral inflammation.

TABLE 2.

Correlations between superoxide cell populations and pro‐inflammatory cytokine measures of PD patients and controls.

	PBMC	CD14+ monocytes	CD14− cells
IFN‐γ	r = 0.301 P = 0.015 *	r = 0.271 P = 0.029*	r = 0.279 P = 0.024*
IL‐1β	r = 0.189 P = 0.389	r = 0.123 P = 0.327	r = 0.151 P = 0.327
IL‐2	r = 0.320 P = 0.009**	r = 0.181 P = 0.148	r = 0.349 P = 0.004**
IL‐4	r = −0.157 P = 0.212	r = −0.086 P = 0.497	r = −0.102 P = 0.418
IL‐6	r = −0.144 P = 0.253	r = −0.028 P = 0.823	r = −0.082 P = 0.514
IL‐10	r = 0.044 P = 0.727	r = 0.117 P = 0.354	r = 0.101 P = 0.422
IL‐12	r = 0.290	r = 0.310	r = 0.395
	P = 0.019*	P = 0.012*	P = 0.001**
IL‐17A	r = 0.281 P = 0.024**	r = 0.165 P = 0.188	r = 0.291 P = 0.018*
TNF‐α	r = 0.002 P = 0.988	r = 0.012 P = 0.926	r = 0.1027 P = 0.832
IFN‐α	r = 0.083 P = 0.499	r = 0.052 P = 0.674	r = 0.039 P = 0.750
IFN‐β	r = −0.102 P = 0.407	r = −0.114 P = 0.357	r = −0.079 P = 0.521

Note: Partial correlation coefficients and P‐values between PBMC, CD14+, and CD14− superoxide measures and pro‐inflammatory cytokines IFN‐γ, IL‐1β, IL‐2, IL‐4, IL‐6, IL‐10, IL‐12, IL‐17A, and TNF‐α of PD patients and controls. Significance was determined at P < 0.05. *P < 0.05.

**P < 0.01.

Abbreviations: PD, Parkinson's disease; PBMC, peripheral blood monocular cells; IFN, interferon; IL, interleukin; TNF‐α, tumor necrosis factor α.

Anti‐Inflammatory Medication Use Increases Mitochondrial Content

Finally, we determined the effect of anti‐inflammatory medication use on flow cytometry measures of mitochondrial function by splitting the cohort into those taking, or not taking, medication. When analyzed this way, there was no significant effect of anti‐inflammatory use on mitochondrial superoxide production for either the whole PBMCs (Fig. 3A), CD14 positive monocytes (Fig. 3B), or non‐CD14 positive cells (Fig. 3C). However, a significant increase in MitoTracker signal intensity was seen for people taking anti‐inflammatory medications for whole PBMCs (Fig. 3D), CD14 positive monocytes (Fig. 3E), and non‐CD14 positive cells (Fig. 3F).

FIG. 3.

Effect of anti‐inflammatory medication on mitochondrial function. The cohort was stratified into people not taking (blue dots) or currently taking (red dots) anti‐inflammatory medications and flow cytometry measures of mitochondrial function were re‐analyzed. There was no significant effect of anti‐inflammatory medications on mitochondrial superoxide levels measured using the MitoSox probe in peripheral blood monocular cells (PBMCs) (A), CD14 positive monocytes (B), or CD14 negative cells (C). Mitochondrial content measured using the MitoTracker probe was significantly increased PBMCs (D), CD14 positive monocytes (E), or CD14 negative cells (F). Significance was determined at P < 0.05. **P < 0.01. Graphs show individual data points as well as the mean and standard error for each group. [Color figure can be viewed at wileyonlinelibrary.com]

Discussion

In this study, we aimed to replicate previous findings of impaired mitochondrial function in peripheral immune cells from patients with PD and determine the extent to which immune cell mitochondrial function may contribute to altered levels of blood cytokines. Previous studies have measured peripheral mitochondrial dysfunction in PD patients using skin fibroblasts, ²⁹ , ³⁰ , ³¹ , ³² immortalized lymphocytes, ³³ and PBMCs, ⁹ , ¹⁰ with results collectively suggesting an increased respiratory rate, increased intracellular ATP, increased mitochondrial membrane potential, and increased ROS occurring in PD cells. The results from the current study confirm peripheral mitochondrial dysfunction in PD, with an increase in superoxide ROS measured in PD patient immune cells. Increased superoxide levels were seen in the peripheral monocyte and non‐monocytic gated groups, in contrast to the study of Smith and colleagues, ⁹ who found increased superoxide specifically in the monocyte population. Reasons for this difference remain unresolved, but it should be noted that although flow cytometry assays were performed similarly, blood collection and PBMC processing protocols were different between the studies. Both studies also focused on specifically gating on monocyte populations, and the inclusion of additional surface marker antibodies to better define mitochondrial dysfunction in other PBMC populations would be useful. Finally, a strong positive correlation was observed in the current study between age and PBMC superoxide levels in PD patients. It should be noted that the current study had on average an older PD cohort than the study by Smith and colleagues, ⁹ which focused on PD with a very short disease duration. Therefore, mitochondrial dysfunction may be selectively detectable in monocytes in the early disease course, and then in all peripheral immune cells with disease progression. Future studies could specifically investigate this proposal.

In the current study, we also found an increase in mitochondrial content, as measured using the MitoTracker probe, although the extent of the increase was complicated by longer time in storage of control PBMC samples and a reduced viability of PD PBMC samples. A recent study using a high content imaging approach showed that the number of mitochondria per cell was increased in PD fibroblasts. ²⁹ Impaired mitophagy resulting in the reduced clearance of defective mitochondria has long been associated with PD. ³⁴ Mitophagy is also impaired in fibroblasts from PD patients, ³¹ and therefore, impaired mitophagy comprises one potential explanation for increased mitochondria in PD cells. Alternately, a compensatory increase in mitochondrial biogenesis to maintain cellular homeostasis may be occurring. Intriguingly, higher levels of mitochondrial content were observed in participants taking anti‐inflammatory medications, predominantly aspirin, and aspirin is suggested to promote mitochondrial biogenesis. ³⁵ Both explanations require further investigation, but the observation that PD immune cells show an impaired response to antimycin A indicates that either way, not all mitochondria in PD immune cells seem fully functional.

Another aim of the current study was to determine if immune cell mitochondrial dysfunction may be associated with increased serum levels of inflammatory cytokines in PD. A number of studies have reported increased cytokines in PD patients, ¹⁶ although such results have not always been consistently observed because of the challenges of measuring cytokines across different platforms and the widespread use of anti‐inflammatory medications in elderly populations. In the current study a significant group difference was not observed between PD and controls. This is likely due, at least in part, to the much larger variation in inflammatory cytokine levels in the PD group. This large variation is consistent with emerging data suggesting that inflammation in PD may be driven by a subgroup of patients. For example, PD patients can be stratified into high and low inflammatory subgroups based on blood cytokine levels, with the higher inflammatory subgroup showing faster clinical disease progression. ²² Nonetheless, a strong positive correlation was seen between circulating levels of IL‐12 and IFN‐γ and immune cell mitochondrial dysfunction. IL‐12 has previously been implicated in PD ³⁶ and indeed, a main function of IL‐12 is inducing the production of IFN‐γ from T‐cells. ³⁷ Interestingly, the production of IFN‐γ from T‐cells following IL‐12 stimulation may also be dependent on mitochondrial ROS, ³⁸ therefore, it cannot be concluded from the current study if IL‐12 may drive superoxide production in immune cells or vice versa. We also found correlations between mitochondrial superoxide levels and serum levels of IL‐17A and IL‐2, but only in the non‐CD14 positive cells. IL‐17A is produced by T‐cells, particularly the Th17 subset of helper T‐cells. IL‐17A is predominantly a pro‐inflammatory cytokine ³⁹ and in rodent PD models has been linked to exacerbated neurodegeneration. ⁴⁰ Treatment with anti‐IL‐17A therapy has also been shown to reduce the risk of developing PD in people with inflammatory bowel disease. ²⁵ Like IL‐12, IL‐17A can also induce mitochondrial ROS, ⁴¹ and therefore, further experiments to determine if neutralizing cytokines can improve mitochondrial function in PD could be of interest. Although it should be noted that anti‐inflammatory medications used in the current study, predominantly aspirin, did not have an effect to suppress mitochondrial superoxide production. IL‐2 is also produced by T‐cells and acts to regulate T cell differentiation. IL‐2 expression can be regulated by mitochondrial ROS, ⁴² and interestingly, studies in PD patients have shown impaired secretion of IL‐2 in response to activation of peripheral immune cells. ⁴³ , ⁴⁴ Altered levels of IL‐2 can affect circulating T‐cell populations, and studies have shown that PD patients do indeed exhibit changes in T‐cell profiles. ⁴⁵ In particular, anti‐inflammatory regulatory T‐cell (T_reg) populations are decreased in PD patients. ⁴⁵ IL‐2 is an important mediator of T_reg differentiation, ⁴⁶ and treatment with IL‐2 has shown neuroprotection in rodent models of PD by increasing T_reg populations to attenuate neuroinflammation. ⁴⁷ , ⁴⁸ A recent study has also demonstrated that mitochondrial ROS are important for T‐cell migration, and that impaired migration of PD T‐cells in a transwell assay is associated with an inability to upregulate superoxide production in response to activation. ⁴⁹ Therefore, there is a clear link between mitochondrial superoxide production, IL‐2, and T‐cells, and the inclusion of surface T‐cell markers would be important in future studies.

In summary, we have identified a preliminary link between peripheral immune cell mitochondrial dysfunction and circulating cytokines related to T‐cell responses. This may help to understand the role of peripheral immune cells in PD pathogenesis and/or help with the biological stratification or staging of PD, as has been recently proposed. ⁵⁰ The study has limitations to acknowledge. Although larger than most published studies investigating peripheral immune cell mitochondrial function in PD, the sample size is still small, and the study was only performed cross‐sectionally at a single site. Mitochondrial function was assessed in a limited capacity using the commercially available MitoSox and MitoTracker dyes, and the known caveats of using these dyes should be considered. ⁵¹ Future studies addressing these limitations and expanding mechanistic insight to further understand the relationship between immune cell mitochondria dysfunction and peripheral inflammation are warranted.

Author Roles

1. Research Project: A. Conception, B. Organization, C. Execution; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript Preparation: A. Writing of the First Draft, B. Review and Critique.

F.A: 1B, 1C, 2A, 3A

P.Y : 1B, 1C, 3A

J.G : 1B, 2C, 3A

M.C: 1B, 1C, 3A

G.M.H: 1B, 2C, 3B

S.J.G.L: 1A, 1C, 2C, 3B

N.D: 1A, 1B, 2A, 2B, 3A, 3B

Supporting information

Data S1. Supporting Information.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.