A Link between Mitochondrial Dysregulation and Idiopathic Autism Spectrum Disorder (ASD): Alterations in Mitochondrial Respiratory Capacity and Membrane Potential

Autism spectrum disorder (ASD) is a neurological disorder triggered by various factors through complex mechanisms. Research has been done to elucidate the potential etiologic mechanisms in ASD, but no single cause has been confirmed. The involvement of oxidative stress is correlated with ASD and possibly affects mitochondrial function. This study aimed to elucidate the link between mitochondrial dysregulation and idiopathic ASD by focusing on mitochondrial respiratory capacity and membrane potential. Our findings showed that mitochondrial function in the energy metabolism pathway was significantly dysregulated in a lymphoblastoid cell line (LCL) derived from an autistic child (ALCL). Respiratory capacities of oxidative phosphorylation (OXPHOS), electron transfer of the Complex I and Complex II linked pathways, membrane potential, and Complex IV activity of the ALCL were analyzed and compared with control cell lines derived from a developmentally normal non-autistic sibling (NALCL). All experiments were performed using high-resolution respirometry. Respiratory capacities of OXPHOS, electron transfer of the Complex I- and Complex II-linked pathways, and Complex IV activity of the ALCL were significantly higher compared to healthy controls. Mitochondrial membrane potential was also significantly higher, measured in the Complex II-linked pathway during LEAK respiration and OXPHOS. These results indicate the abnormalities in mitochondrial respiratory control linking mitochondrial function with autism. Correlating mitochondrial dysfunction and autism is important for a better understanding of ASD pathogenesis in order to produce effective interventions.


Introduction
Autism spectrum disorder (ASD) is a heterogeneous neurodevelopmental disorder characterized by a combination of impairments in social communication and interaction, sensory anomalies, repetitive behaviors, and varying levels of intellectual disability. The global prevalence of ASD has increased dramatically 20-to 30-fold over the past decades [1] and the World Health Organization estimated that 1 in 160 children is affected by ASD [2]. The recent increase in ASD underscores the importance of expanding research into risk factors and effective interventions. Whilst most studies focus on effective interventions on promoting communication, social skills, and education [3], the underlying cause of idiopathic autism remains unclear. Up to the present, no significant underlying cause and no pharmacological medication is currently established as a treatment. Tracking the increase in ASD prevalence poses unique challenges because of its complex nature, lack of diagnostic biomarkers, and challenging diagnostic criteria.
Over the last decade, increasing attention has been paid to the mitochondrial physiology that may underlie some of the symptoms of ASD. Often mitochondrial dysfunction healthy and the ethnicity of the family was classified as non-Hispanic or Latino. The DNA of the ASD patient was tested for mutation of the Fragile-X syndrome gene, the FMR1 gene, and no known mutation was found.

High-Resolution Respirometry
All experiments were performed using a modular instrument for high-resolution respirometry (HRR) and fluorometry, the Oxygraph-2k (O2k; Oroboros Instruments, Innsbruck, Austria). The temperature of the experimental chamber was kept at 37 • C with constant stirring at 750 rpm to ensure a homogeneous oxygen distribution in the chamber without cell disruption. Air calibration of the oxygen sensors was performed daily. DatLab 7 software (Oroboros Instruments) was used for real-time data acquisition and analysis. The oxygen concentration was kept at and below air saturation (normoxic conditions). The O2k chambers were filled with 2.0 mL MiR05 equilibrated at 37 • C. Titrations were performed manually by injection into O2k chambers using precalibrated Hamilton microsyringes. Substrate concentrations and respiratory states are listed in Tables 1 and 2, respectively. Oxygen consumption rates were normalized to cell number and expressed as O 2 flow per cell [amol·s −1 ·cell −1 ]. Table 1. Substrates, uncoupler, and inhibitors used in respirometry protocol to induce different respiratory states [20,21].

Substrates Function Concentration
Digitonin ( ET, E Oxygen consumption in the non-coupled state at optimum uncoupler concentration, ET capacity.

ROX, Rox
Residual oxygen consumption, measured after inhibition of the ETS.
(E − P)/E Relative E − P excess capacity, defines the limitation of OXPHOS capacity exerted by the phosphorylation system.
(P − L)/P OXPHOS coupling efficiency, combining the effects of coupling and limitation by the phosphorylation system.

Cytochrome c Oxidase (CIV) Activity Assay
Cells in MiR05 were titrated into each chamber, followed by titration of digitonin (7.5 µg·mL −1 for NALCL, 27.5 µg·mL −1 for ALCL). CIII was inhibited by the addition of antimycin A (2.5 µM). Stepwise titration of the protonophore CCCP (0.5 µM steps) leads to proton leakage across the mitochondrial inner membrane. Ascorbate (2 mM) was added before the addition of TMPD (0.5 mM) to avoid uncontrolled autoxidation and to maintain TMPD in a reduced state. The artificial substrate TMPD reduced cytochrome c. The activity of CIV was later inhibited by the addition of sodium azide (100 mM). Chemical background due to the autoxidation of ascorbate, TMPD, and cytochrome c was assessed after the inhibition of CIV by sodium azide. Chemical background correction was performed using DatLab 7. The CIV activity was expressed as a function of oxygen concentration. The protocols for measuring CIV activities in NALCL and ALCL were shown in Figure 1a and Figure 1b respectively.

Respiration
The addition of the NADH-linked substrates pyruvate (P; 5 mM) and malate (M; 2 mM) induced non-phosphorylating LEAK respiration, N(GM) L . Subsequently, N-OXPHOS capacity, N(GM) P , was measured after the addition of a saturating concentration of ADP (2.5 mM). Cytochrome c (10 µM) was added to test the integrity of the mitochondrial outer membrane damage. Glutamate (10 mM) was added to stimulate multiple NADH hydrogenases, N(PGM) P . The addition of succinate (S; 10 mM) stimulated the OXPHOS capacity of the combined N-and S-pathways (NS P ). Stepwise titration of the CCCP (0.5 µM steps) was used to measure the capacity of the ETS (NS E ). Then, CI was inhibited by rotenone (0.5 µM) to measure S-ET capacity (S E ). The inhibition of CIII by antimycin A (2.5 µM) provided a measure of Rox. For further methodological details see [21]. The protocols for measuring respiration in NALCL and ALCL were shown in Figure 2a and  are the protocols for measuring CIV activities in NALCL and ALCL, respectively. Titrations: Cell, dig (digitonin, cell permeabilization), Ama (antimycin A, CIII inhibition), CCCP (oxidative phosphorylation uncoupling), As (ascorbate, maintaining TMPD in a reduced state), TMPD (reducing cytochrome c), c (cytochrome c integrity of outer mt-membrane), and Azd (azide, CIV inhibition). Blue plots indicate O2 concentration and red plots are the O2 consumption expressed per cell. (c) The total O2 consumption rate was baseline-corrected for autoxidation after inhibition of CIV. Results are expressed as means ± S.D. * denotes p < 0.05 compared to NALCL as determined by t-test.

Respiration
The addition of the NADH-linked substrates pyruvate (P; 5 mM) and malate (M; 2 mM) induced non-phosphorylating LEAK respiration, N(GM)L. Subsequently, N-OXPHOS capacity, N(GM)P, was measured after the addition of a saturating concentration of ADP (2.5 mM). Cytochrome c (10 µM) was added to test the integrity of the   After plasma membrane permeabilization, five respiratory states were sequentially established to measure NADH-linked LEAK respiration with pyruvate and malate, N L , OXPHOS capacity, N P , NS-pathway OXPHOS capacity, NS P , and NSand S-pathway ET capacity, NS E and S E , where S indicates the succinate pathway. Results are expressed as means ± S.D. * denotes p < 0.05 while ** denotes p < 0.01 compared to NALCL as determined by t-test.

Mitochondrial Membrane Potential
The Fluorescence-Sensor Blue of the O2k-Fluorescence LED2-Module was used with filter sets for safranin. Two sensors were inserted through the front windows of the O2kchambers. Following air calibration, the chamber illumination was switched off. HRR provides simultaneous measurement of respiration and the safranin signal in each chamber [20]. The polarization voltage regulating light intensity was set at 500 mV at gain 1000. Safranin is a fluorophore with an excitation wavelength of 495 nm and an emission wavelength of 587 nm, and it is a lipophilic cation that accumulates in mitochondria depending on the inside negative potential in energized mitochondria [19,20]. Upon accumulation in the matrix, safranin undergoes a change in absorption and self-quenching of fluorescence [25,26]. Fluorescence intensity is linearly related to mtMP [27]. For calibration, a 200 µM stock solution of safranin was titrated in five steps to obtain final concentrations of 0.25, 0.5, 1.0, 1.5, and 2.0 µM, obtaining a linear increase of the fluorescence signal as a function of safranin concentration in the chamber. Cells were then added to the chambers. In the absence of respiratory inhibitors, the mtMP builds up on the basis of endogenous substrates, and a corresponding amount of the dye accumulates in the mitochondrial matrix (i.e., the initial decline of the fluorescence signal). In addition, a fraction of safranin binds non-specifically to cell membranes. Therefore, the free safranin concentration is lower than the total safranin concentration added to the chamber [20]. As safranin inhibits N-pathway OXPHOS capacity [19,20], only S-linked respiration was measured. S L(n) was induced by adding succinate (10 mM) in the presence of rotenone (0.5 µM) that inhibited CI. The titration of rotenone prevents the accumulation of oxaloacetate, which is a potent inhibitor of CII. Subsequently, S P was measured after the addition of a saturating concentration of ADP (2.5 mM). S L(Omy) was induced by adding oligomycin (2.5 µM) that inhibited F 1 F 0 -ATPase (ATP synthase). Stepwise titration of the CCCP (0.5 µM steps) was used to measure S E . The inhibition of CIII by antimycin A (2.5 µM) provided a measure of Rox. Figure 3c,d were the protocols for measuring mitochondrial membrane potential (mtMP) using safranin concentration in NALCL and ALCL, respectively. The Rox-corrected respiration was shown in Figure 3e.

Data Analysis
Oroboros DatLab 7 was used to calculate respiration corrected for instrumental background oxygen flux and graphical presentation of experimental data. Three variants of cell lines for ALCL and NALCL were used and three independent experiments were performed for each protocol type. Data are expressed as means ± standard deviation. Statistical analyses using SPSS statistical software version 16 and t-test were applied to determine the significant differences among the groups, where p < 0.05 was considered significant.

Data Analysis
Oroboros DatLab 7 was used to calculate respiration corrected for instrumental background oxygen flux and graphical presentation of experimental data. Three variants of cell lines for ALCL and NALCL were used and three independent experiments were performed for each protocol type. Data are expressed as means ± standard deviation. Statistical analyses using SPSS statistical software version 16 and t-test were applied to determine the significant differences among the groups, where p < 0.05 was considered significant. . Succinate-linked respiration of NALCL and ALCL measured simultaneously with mitochondrial membrane potential in the presence of safranin. (a,b) are the protocols for measuring respiration in NALCL and ALCL, respectively, while (c,d) are the protocols for measuring mitochondrial membrane potential, mtMP using safranin concentration in NALCL and ALCL, respectively. Before the cells were added, safranin was calibrated up to 2 µM. The initial drop to 1.6 µM safranin is mainly due to unspecific binding upon the addition of cells. Low safranin concentrations indicate high mtMP. Titrations: Cell, dig (digitonin, cell permeabilization), Rot (rotenone, CI inhibition), S (succinate, S-LEAK respiration, S L(n) ), ADP (S-OXPHOS capacity, S P ), Omy (oligomycin, S-LEAK, S L(Omy) ), CCCP (S-ET capacity, S E ), Mna (malonate, CII inhibition), and Ama (antimycin A, CIII inhibition, Rox). Oxygen consumption was corrected for Rox. Blue plots indicate O 2 concentration, red plots are the O 2 consumption expressed per cell, and neon plots indicate safranin concentrations. (e) The Rox-corrected respiration and (f) the relative safranin signal: ROUTINE state, R, and in four respiratory states in permeabilized cells: S-linked LEAK respiration in the absence of adenylates, S L(n) , S-linked OXPHOS capacity, S P , S-linked LEAK respiration after inhibition of ATP synthase by oligomycin, S L(Omy) , and S-linked ET capacity, S E . Results are expressed as mean ± S.D. * denotes p < 0.05 while ** denotes p < 0.01 compared to NALCL as determined by Student's t-test.

Mitochondrial Respiration
In this study, the first respiration rate measured in both cell lines is the ROUTINE respiration which represents energy demand under steady-state conditions. ROUTINE respiration, in ALCL, was significantly higher (12.8 ± 1.3 amol·s −1 ·cell −1 ) compared to NALCL (8.1 ± 1.8 amol·s −1 ·cell −1 ). A digitonin concentration, Dig, of 7.5 µg·mL −1 was optimal for NALCL and 27.5 µg·mL −1 for ALCL for complete permeabilization of the plasma membrane without affecting the mitochondrial membranes.
Further stimulation of respiration with succinate activates convergent entry through the S-linked pathway via CII in addition to the N-pathway through CI, which results in NSlinked OXPHOS capacity (NS P ).
The NS P of NALCL increased 1.7-fold to 25.3 ± 2.1 amol·s −1 ·cell −1 , indicating a strongly additive effect. The NS P of ALCL was significantly higher than in NALCL (44.3 ± 1.2 amol·s −1 ·cell −1 ), which is a 1.9-fold increase with respect to N P .
After inhibition of the N-pathway by rotenone, S-ET capacity (S E ) was 20.1 ± 5.5 and 32.2 ± 3.2 amol·s −1 ·cell −1 in NALCL and ALCL, respectively. The NS-linked P/E ratio (NS P/E ) in NALCL (0.21) was higher than in ALCL (0.01). The P/E coupling control ratio describes the function of the phosphorylation system as a crucial controller limiting OXPHOS capacity. A ratio of 1.0 indicates no limitation by the phosphorylation system.

Mitochondrial Membrane Potential
MtMP was measured in the S-pathway simultaneously with respiration in the same chamber. Rotenone and succinate were added simultaneously to establish the LEAK state in the absence of exogenous ADP and ATP (no adenylates), S L(n) . S L was similar in NALCL and ALCL (4.7 ± 1.7 and 6.3 ± 1.4 amol·s −1 ·cell −1 , respectively). OXPHOS capacity, S P was significantly lower in NALCL than ALCL (6.6 ± 2.5 and 12.8 ± 2.0 amol·s −1 ·cell −1 , respectively). In both cell lines, the inhibition of ATPase by oligomycin inhibited LEAK respiration, S L(Omy) , was below the level of S L(n) , indicating an overestimation of LEAK in the absence of inhibition of the phosphorylation system due to some recycling of endogenous ATP to ADP by ATPases (Figure 3a,b). OXPHOS coupling efficiencies were 0.29 and 0.51 in the absence of oligomycin, but 0.53 and 0.65 in the presence of oligomycin for NALCL and ALCL, respectively. The S-pathway ET capacity in the mtMP protocol was obtained by stepwise titration of CCCP up to reaching maximum respiration.
In the S-pathway with electron entry into the Q-junction through CII and only two coupling sites downstream, the P/E ratios increased to 0.67 and 0.60 in NALCL and ALCL, respectively, indicating the lower limit in the phosphorylation system. On the other hand, the LEAK/OXPHOS (S L/P ) ratio depicts respiration triggered by membrane leakiness in comparison to OXPHOS capacity. The mitochondrial membrane leakiness appeared higher in NALCL with an L/P ratio of 0.72, while in ALCL, the L/P ratio was 0.49.
During ROUTINE respiration of intact cells, the safranin signals showed no significant difference between NALCL and ALCL (Figure 3f). After plasma membrane permeabilization, the mtMP increased in both cell lines as S-linked respiration was stimulated by the addition of succinate in the presence of Rot, as reflected by a decline of safranin fluorescence. The mtMP in ALCL showed a higher increase than in NALCL during LEAK (p < 0.01) and OXPHOS (p < 0.05; Figure 3f). There was no significant difference between mtMP in NALCL and ALCL during ET capacity measurement.

Discussion
mtD is often linked with oxidative stress as damaged mitochondria not only produce more reactive oxygen species (ROS, including H 2 O 2 ) and reactive nitrogen species through OXPHOS, but mitochondria are also becoming more vulnerable to oxidative stress [28]. In the present study, we investigated the mtD concerning CIV activity, mitochondrial respiration through N-pathway (CI-linked), S-pathway (CII-linked), and both pathways as combined NS-pathway (CI-and CII-linked), and mtMP as the key markers of mitochondrial function.
The first respiratory state measured in this study is ROUTINE respiration. ROUTINE respiration is the respiration controlled by intrinsic energy demand. It represents energy demand under steady-state conditions. During ROUTINE respiration, cell respiration relies on endogenous substrates in the cells only.
The higher ROUTINE respiration rate in ALCL as compared to in NALCL showed that the intrinsic energy demand under steady-state conditions in ALCL is higher as compared to in NALCL. This situation is considered to be a compensatory response to the increased intrinsic energy demand in ALCL. Ideally, bioenergetic efficiency is achieved when there is a balance between energy supply and energy demand [29]. In response to changes in energy demand, mitochondria respond by adjusting both their capacity and efficiency of ATP production [29]. This might be reflected by the higher OXPHOS capacity in ALCL as compared to in NALCL. In an intact cell with high energy demand, nutrient utilization and its availability have very significant control over respiration as compared to ATP turnover [29].
It has been proposed that the increased respiration rate leads to higher oxidative stress and alterations in mitochondrial energy metabolism [30]. The higher OXPHOS capacity in ALCL might be linked to the over replication and a higher mtDNA copy number in ASD than in controls [31]. These conditions were observed in cells in response to oxidative stress [31]. mtDNA encodes some of the proteins of the complexes of the electron transfer system (ETS). Thus, an increase in mtDNA copy number possibly leads to an increased expression of ETS complexes, resulting in a higher respiratory ET capacity. This, however, needs further clarification via real-time PCR to investigate the level of expression of ETS genes encoded in the mtDNA. mtD from mtDNA mutations could lead to ETS instability, electron leakage, and increased ROS production [32]. Higher OXPHOS-and ET capacities would result in increased production of H 2 O 2 and other ROS-inducing oxidative stress. A high level of H 2 O 2 production is frequently observed in ASD patients compared to controls [33]. Our results are in line with another study that reported a 40%-50% increase in maximal respiratory rates in lymphoblasts of ASD patients as compared to the non-autism relative [34]. However, clarification of this hypothesis needs to be done via sequencing of the mtDNA of both cell lines.
CIV is the terminal respiratory complex of the ETS, which converts O 2 to H 2 O. Diminished activity of ETS complexes has been reported as a prevalent mitochondrial abnormality associated with ASD [4]. For example, lowered activities of CI, CIII, and CIV were found in ASD studies using muscle cells [35] and in the brain from post-mortem ASD patients [36,37]. However, CIV mitochondrial-dependent oxygen consumption was not different in peripheral blood lymphocytes from children with autism versus controls [33]. On the other hand, increased activity of CIV was detected in brain tissue [38,39]. Similarly, in the present study of human lymphoblastoid cell lines, the CIV activity was higher in ALCL compared to NALCL. This occurs to cope with the increase in the ET capacity of the NADH-and succinate pathways (CI and CII), which results in higher production and the transfer of electrons toward CIV.
MtMP reflects the electrical potential difference across the mitochondrial inner membrane, where transmembrane pH difference is the driving force for ATP synthesis. Alterations in mtMP are strongly linked to the control of electron transfer and ATP synthesis [21,33,34] and are involved in apoptosis and necrosis [40]. In our study, the mtMP in ALCL was higher in S-OXPHOS respiration as compared to NALCL. This is in agreement with previous studies, which found that the mtMP in ASD is higher compared to controls [28,41]. Theoretically, the higher the mtMP would reflect the higher the energy capacity of the mitochondrial inner membrane and the higher the synthesis of ATP [42]. However, a high electric field is energetically expensive to maintain; thus, ion leakage would occur to compromise the mtMP [42]. This is reflected by a higher S-LEAK value in ALCL compared to NALCL. Mitochondrial ROS production by the mitochondrial transport system is increased at high membrane potential [42][43][44]. Fluctuations of the mtMP between different respiratory states are larger in ALCL than in NALCL. Such fluctuations may have deleterious effects on cell physiology [42]. Thus, high mtMP is potentially harmful to mitochondria and consequently to the cell [45].
On another note, classic mtD biomarkers are not always accurate. The elevation of pyruvate levels does not always reflect lower ET capacity. It could be due to impaired pyruvate metabolism and TCA cycle inhibition. Thus, an enzyme marker (e.g., CIV) can be used as a diagnostic tool for ASD and other mitochondrial diseases. It is also worth noting that ASD is a spectrum disorder in which different effects and causes may account for the link between mtD and various observed symptoms. Our findings suggest an involvement of mitochondria, particularly the complexes in the electron transport chain, such as CI, CII, and CIV, in the pathogenesis of ASD, specifically in idiopathic autism. This is tightly linked to the production of ROS by mitochondria as the possible underlying cause of ASD as CI is the biggest contributor of ROS production under pathological conditions, and both respiration rate and mtMP are proportional to H 2 O 2 production [39,41,42].
These initial findings lead us to further directions that might include the exploration of mitochondrial elements in a bigger picture, such as measurement of mitochondrial density and determination of mitochondrial copy number. Other mitochondrial markers (e.g., measurement of the activity of citrate synthase) should be further explored for a wider understanding of the involvement of mitochondrial function in idiopathic autism. Altogether, these results indicate the correlation between mitochondrial function abnormalities and ASD. Novel and extended approaches are required to improve the diagnosis of mitochondrial function in ASD and find better intervention strategies.

Conclusions
Many mitochondrial diseases and mtD are characterized by lower activities of respiratory complexes and mitochondrial respiratory capacities, whereas we observed an opposite pattern in agreement with a few other studies on ASD. Thus, linking mtD to ASD should be revisited, and instead of mitochondrial 'dysfunction', mitochondrial 'dysregulation' represents a more accurate term in the context of ASD pathogenesis, especially in the case of idiopathic autism. Changes in mitochondrial respiratory capacities can be caused by a modification of mitochondrial quality or mitochondrial density. Both mechanisms may be implicated in the increased OXPHOS-and ET capacities, CIV activity, and mtMP in the autism lymphoblastoid cell lines, and may be linked to ROS production and oxidative stress, which needs further clarification and an interesting point to explore. These bioenergetic characteristics imply a disruption of nutrient homeostasis as a potential cause of the pathological symptoms prevalent in idiopathic autism.

Limitations & Future Direction
This is a preliminary study that uses only one pair of LCL; ALCL versus NALCL, comparing the mitochondrial functions which are the CIV activity, respiration of intact and permeabilized cells, specifically in the ROUTINE, LEAK, OXPHOS and ETS respirations, and mtMP. This study provides an insight into the differences in mitochondrial function between cells derived from an autistic child compared to his apparently healthy sibling. The findings might provide an overview of a subset of ASD. The generalization on ASD as a whole, therefore, must be made cautiously.
Future studies will, therefore, need to consider the gene mutation on both cell lines specifically in the mutations in the mtDNA.

Data Availability Statement:
The data presented in this study are available upon request from the corresponding author as the data is not available elsewhere.