Immortalized Parkinson's disease lymphocytes have enhanced mitochondrial respiratory activity

Abstract

In combination with studies of post-mortem Parkinson's disease (PD) brains, pharmacological and genetic models of PD have suggested that two fundamental interacting cellular processes are impaired - proteostasis and mitochondrial respiration. We have re-examined the role of mitochondrial dysfunction in lymphoblasts isolated from individuals with idiopathic PD and an age-matched control group. As previously reported for various PD cell types, the production of reactive oxygen species (ROS) by PD lymphoblasts was significantly elevated. However, this was not due to an impairment of mitochondrial respiration, as is often assumed. Instead, basal mitochondrial respiration and ATP synthesis are dramatically elevated in PD lymphoblasts. The mitochondrial mass, genome copy number and membrane potential were unaltered, but the expression of indicative respiratory complex proteins was also elevated. This explains the increased oxygen consumption rates by each of the respiratory complexes in experimentally uncoupled mitochondria of iPD cells. However, it was not attributable to increased activity of the stress- and energy-sensing protein kinase AMPK, a regulator of mitochondrial biogenesis and activity. The respiratory differences between iPD and control cells were sufficiently dramatic as to provide a potentially sensitive and reliable biomarker of the disease state, unaffected by disease duration (time since diagnosis) or clinical severity. Lymphoblasts from control and PD individuals thus occupy two distinct, quasi-stable steady states; a 'normal' and a 'hyperactive' state characterized by two different metabolic rates. The apparent stability of the 'hyperactive' state in patient-derived lymphoblasts in the face of patient ageing, ongoing disease and mounting disease severity suggests an early, permanent switch to an alternative metabolic steady state. With its associated, elevated ROS production, the 'hyperactive' state might not cause pathology to cells that are rapidly turned over, but brain cells might accumulate long-term damage leading ultimately to neurodegeneration and the loss of mitochondrial function observed post-mortem. Whether the 'hyperactive' state in lymphoblasts is a biomarker specifically of PD or more generally of neurodegenerative disease remains to be determined.

Keywords: AMPK; ATP; Complex I; Lymphoblast; Lymphocyte; Mitochondria; OXPHOS; Oxidative stress; Parkinson's disease; ROS; Respiration.

Fig. 1.

Alterations to parameters of mitochondrial function in PD lymphoblasts. (A) Reactive O₂ species levels are elevated in lymphoblasts from individuals with PD. Intracellular ROS levels were measured in lymphoblasts from PD and control individuals using MAK142 (Deep Red) fluorescence. Except for one control cell line (which was assayed only once), the mean normalized fluorescence from 10⁵ cells was measured in duplicate in at least three independent experiments. The ROS fluorescence in the PD lines (n=30) was significantly elevated compared with controls (n=9) (single-sided Welch test). (B) Mitochondrial membrane potential is unaltered in iPD lymphoblasts. The relative mitochondrial membrane potential (Δψ_m) was measured in lymphoblasts from PD and control individuals using the ratio of MitoTracker Red CMXRos (Δψ_m-dependent) to MitoTracker Green (mitochondrial mass-dependent) fluorescence. Each PD (n=30) and control (n=9) cell line was assayed in duplicate in at least three independent experiments and means were calculated. The mitochondrial membrane potential was not significantly different in the PD and control samples (two-sided Welch t-test). (C) Steady-state ATP levels are elevated in lymphoblasts from individuals with PD. Steady-state ATP levels were assayed using a luciferase-based luminescence assay in lymphoblasts from PD and control individuals. Each PD (n=30) and control (n=9) cell line was assayed in duplicate in at least three independent experiments and means were calculated. The steady-state ATP levels in the PD lines were elevated significantly (single-sided Welch test). Error bars are standard errors of the mean (s.e.m.).

Fig. 2.

The rate of O₂ consumption by mitochondrial respiration is elevated in lymphoblasts from individuals with PD. Oxygen consumption rates (OCR in pmol/min) were measured using a Seahorse Extracellular Flux Analyzer for 8×10⁵ lymphoblasts from PD and control individuals. In each experiment the OCR was measured before (basal respiration) and after successive addition of oligomycin (ATP synthase inhibitor), CCCP (uncoupling protonophore), rotenone (complex I inhibitor) and antimycin A (complex III inhibitor), allowing determination of each of the components of respiration as shown in panel (A). In each experiment, data was collected and averaged from four separate wells for each individual cell line. A typical example of an experiment with standard errors reflecting the between-well variance within the experiment is shown in panel (B). Each PD (n=30) and control (n=9) cell line was assayed in at least three independent experiments and means were calculated. From this data, we determined the OCR attributable to (C) basal respiration, (D) ATP synthesis, (E) uncoupled (maximum) respiration, (F) complex I activity, (G) complex II activity, (H) mitochondrial activities other than ATP synthesis (‘proton leak’), (I) processes not driven by electron transport (‘non-mitochondrial’) and (J) the spare capacity (excess of uncoupled over basal respiration). Significance probabilities are shown for each of the respiratory parameters. All were elevated significantly in PD lymphoblasts compared with controls (P<0.05, single-sided Welch t-test). Error bars are s.e.m. (K) Multiple regression analysis of the relationship between ROS production and the parameters of respiration shown in panels C to J showed that ROS levels were correlated with the rate of basal respiration. The analysis incorporated dummy variables to allow distinction between control and PD lymphoblasts and successive removal of least significant coefficients until only significant coefficients remained. The only remaining significant coefficient (apart from the intercept) was that shown, relating ROS levels to basal metabolic rate, so that both control and PD values lay on the same regression line. The significance of the final regression is shown (P<0.05, two-sided t-test).

Fig. 3.

Regression relationships amongst the components of basal and maximum (uncoupled) respiration. Multiple regression analysis was conducted for each of the illustrated respiratory parameters against either the basal respiration rate (top row) or the maximum (uncoupled) respiration rate and dummy variables distinguishing PD and control groups (bottom row). Least significant variables were removed successively until only significant regression coefficients (P<0.05) remained in the model. All regressions were highly significant, but there were no significant differences in the regressions (slope or intercept) between PD and control groups. Plotted regression lines were fitted by the least squares method. Each point represents the mean result from at least three independent experiments on lymphoblasts from a single participant. OCR=O₂ consumption rate. Electron transport OCR is the sum of the contributions of complex I and complex II to the maximum OCR.

Fig. 4.

Steady-state AMPK activity, mitochondrial mass and genome copy number are unchanged in PD lymphoblasts but oxphos mRNA and protein expression are elevated. (A) AMPK activity is not altered in PD lymphoblasts. Each PD (n=30) and control (n=8) cell line was assayed in duplicate in each of at least two independent experiments and means were calculated. AMPK activities in the PD cells were not elevated significantly (single-sided Welch t-test). Quantitative western blots also revealed no significant change in the ratio of activated AMPK (phosphorylated α1 subunit) to total AMPK in PD cells (Fig. S1A). (B) Mitochondrial mass is not elevated in PD lymphoblasts. Mitochondrial mass was measured in lymphoblasts from PD and control individuals using MitoTracker Green fluorescence. Each PD (n=30) and control (n=9) cell line was assayed in at least three independent experiments and means were calculated. The fluorescence in the PD lines was not significantly elevated (single-sided Welch test). (C) Mitochondrial genome copy number is unchanged in PD lymphoblasts. Relative mitochondrial genome copy number was measured in duplicate in semi-quantitative RT-PCR as the difference in qPCR threshold cycle number between the mitochondrial genes encoding ND1 and ND4 (mitochondrially encoded subunits of complex I) and the nuclear gene encoding β2-microglobulin. Each PD (n=30) and control (n=8) cell line was assayed in two or three independent experiments and means were calculated. The mitochondrial genome number was not significantly different in the PD and control cell lines (single-sided Welch tests). Very similar results were obtained using the nuclear 18S rDNA genes as the genome loading control (not shown). (D) Mitochondrial gene expression is elevated in PD lymphoblasts. Quantitative reverse transcription PCR was used to assay ND1 and ND4 mRNAs (mitochondrially encoded subunits of complex I) relative to the nuclear-encoded cytoplasmic 18S rRNA (which provided the internal loading control). Each PD (n=26) and control (n=7) cell line was assayed in duplicate in one to five independent experiments and mean threshold cycle numbers (C_t) were calculated. For all cell lines the differences between mRNA expression levels for both subunits and the corresponding averages for the control cell lines were determined. These were found to be elevated significantly in the PD cells (single-sided Welch test). Similar results were found in a subset of samples using the β2-microglobulin mRNA as the internal control (not shown). (E,F) Steady-state levels of OXPHOS proteins are elevated in PD lymphoblasts. (E) Crude lymphoblast protein extracts were separated on SDS-PAGE and western blotted using commercial antibodies. The blot shown is an example with fluorescence signals as shown for the ATP5A and SDHB bands, normalized to the corresponding α-tubulin signal in the same track and expressed relative to the average control values. (F) Background-subtracted fluorescence signals for each band were normalized against the background-subtracted α-tubulin signal in the same track and expressed relative to the average control value. The ATP5A and SDHB levels were significantly higher in PD cells (n=11 PD and n=6 controls, single-sided Welch test). Similar results were found in experiments with 19 PD samples and four control samples using an anti-β-actin antibody to detect actin as the internal loading control (Fig. S1B,C,D). Error bars are s.e.m.

Fig. 5.

Regression relationships between mitochondrial respiratory activity, patient age and disease duration. The time since first diagnosis was recorded as a surrogate measure of disease duration. For plotting and analysis purposes, control participants were deemed to have disease durations of 0. Planar surfaces were fitted by least squares to show the relationships between patient age and disease duration and (A) basal OCR, (B) OCR attributable to mitochondrial ATP synthesis, (C) maximum uncoupled OCR and (D) maximum uncoupled OCR attributable to complex I. In multiple linear regressions incorporating successive removal of least significant coefficients and the use of dummy variables to distinguish between PD and control groups, the only significant regressions were those relating the maximum OCR (P=8.09×10⁻⁴, two-sided t-test) and complex I activity (P=2.09×10⁻³, two-sided t-test) to PD patient age (Fig. S2). Basal OCR, ATP synthesis, complex II activity, the proton leak and non-respiratory OCR were not significantly affected by patient age or disease duration (P>0.1, two-sided t-test).

Fig. 6.

ROC curve analysis of the use of respiration rates of lymphoblasts as diagnostic tests for PD. The optimum test threshold was a maximum OCR of 335.84 pmol/min (A,B), a basal OCR of 228.92 pmol/min (C,D) or a complex I activity of 263.55 pmol/min (E,F). Lymphoblasts exceeding this threshold indicate PD with the indicated specificities (proportion of control individuals correctly excluded) and sensitivities (proportion of PD individuals correctly diagnosed). Areas under the curve (a measure of test reliability) were 0.9926 (95% confidence interval: 0.975-1.0, DeLong's test) for the maximum OCR, 0.985 (95% confidence interval: 0.953-1.0, DeLong's test) for the basal OCR and 0.989 (95% confidence interval: 0.964-1.0, DeLong's test) for complex I activity.

Fig. 7.

PD and control lymphoblasts occupy distinct regions in multidimensional metabolic space. The dramatic differences between PD and control lymphoblasts in respiratory and associated biochemical parameters can be visualized in the form of distinct regions of a multidimensional metabolic space. The figure illustrates this by depicting the median ellipsoidal surface fit to the results for three of the respiratory measures (OCR in pmol/min) that separated the two groups most clearly – complex I activity in the x axis, the maximum respiratory capacity in the y axis and the basal metabolic rate in the z axis. The existence of these two different metabolic states was confirmed in principal components analysis (Fig. S4), which accounts for and removes the effect of cross-correlations between variables.