Neurochemical Research

, Volume 36, Issue 2, pp 327–338

The Effects of Antidepressants on Mitochondrial Function in a Model Cell System and Isolated Mitochondria

Authors

    • Department of Medical Laboratory Sciences, Faculty of Allied Health SciencesThe Hashemite University
  • D. A. Kendall
    • School of Biomedical Sciences, University of Nottingham Medical SchoolQueen’s Medical Centre
  • T. E. Bates
    • School of Biomedical Sciences, University of Nottingham Medical SchoolQueen’s Medical Centre
    • New-Use Therapeutics Limited, BioCity Nottingham
ORIGINAL PAPER

DOI: 10.1007/s11064-010-0331-z

Cite this article as:
Abdel-Razaq, W., Kendall, D.A. & Bates, T.E. Neurochem Res (2011) 36: 327. doi:10.1007/s11064-010-0331-z

Abstract

The in vitro effects of antidepressant drugs on mitochondrial function were investigated in a CHOβ2SPAP cell line used previously to determine the effects of antidepressants on gene transcription (Abdel-Razaq et al., Biochem Pharmacol 73:1995–2003, 2007) and in rat heart isolated mitochondria. Apoptotic effects of clomipramine (CLOM), desipramine (DMI) and of norfluoxetine (NORF, the active metabolite of fluoxetine), on cellular viability were indicated by morphological changes and concentration-dependent increases in caspase-3 activity in CHO cells after 18 h exposure to CLOM, DMI and NORF. However, tianeptine (TIAN) was without effect. CLOM and NORF both reduced integrated mitochondrial function as shown by marked reductions in membrane potential (MMP) in mitochondria isolated from rat hearts. DMI also showed a similar but smaller effect, whereas, TIAN did not elicit any significant change in MMP. Moreover, micromolar concentrations of CLOM, DMI and NORF caused significant inhibitions of the activities of mitochondrial complexes (I, II/III and IV). The inhibitory effects on complex IV activity were most marked. TIAN inhibited only complex I activity at concentrations in excess of 20 μM. The observed inhibitory effects of antidepressants on the mitochondrial complexes were accompanied by a significant decrease in the mitochondrial state-3 respiration at concentrations above 10 μM. The results demonstrate that the apoptotic cell death observed in antidepressant-treated cells could be due to disruption of mitochondrial function resulting from multiple inhibition of mitochondrial enzyme complexes. The possibility that antimitochondrial actions of antidepressants could provide a potentially protective pre-conditioning effect is discussed.

Keywords

AntidepressantsMechanism of actionMitochondrial dysfunctionMitochondrial complexesApoptosisCaspase-3

Introduction

Depression has been suggested to be caused by alterations in monoaminergic neurotransmission and re-uptake systems leading to changes in post-receptor signalling [1]. However, there are problems with the simple monoamine theory; for example there is a significant delay between onset of treatment with antidepressant drugs which acutely modulate monoamine levels and their clinical effects, and it is likely that depression is a multifactorial disease with a complex aetiology. Amongst the possible targets for antidepressant drugs, the mitochondria have begun to attract attention.

Mitochondria play a fundamental role in energy metabolism, apoptosis and in intracellular signalling mechanisms. Over the last decade, increasing experimental data have been generated supporting the involvement of mitochondrial dysfunction in the pathogenesis of many neurodegenerative disorders [2], as well as cardiomyopathy, diabetes and various cancers [3, 4]. Many of these diseases have been linked to mutations in mitochondrial DNA and/or nuclear DNA encoding for mitochondrial proteins [5]. However, the pathophysiology of mitochondrial diseases is complex, and a number of mechanisms have been suggested to result in mitochondrial dysfunction, including increased generation of reactive oxygen species [6] and reactive nitrogen species [7], decreased anti-oxidant levels, decreased enzyme activities and changes in pro- and anti-apoptotic protein levels. In addition, the mechanisms that control the interactions between mitochondrial DNA and the nuclear genes encoding mitochondrial proteins are poorly understood [8], although it has been shown that stressors such as heat shock can up-regulate mitochondrial complex activities resulting in whole organ protection from ischaemia-reprefusion injury [9].

In the last few years, an alternative theory of depression has been proposed, called “the mitochondrial dysfunction hypothesis”, which suggests that impaired mitochondrial functions and/or defects in mitochondrial DNA are associated with psychiatric conditions such as schizophrenia [10], Alzheimer’s disease [11], bipolar disorder [12], major depression and affective spectrum disorders [13]. According to this theory, depression, mainly bipolar disorder, is regarded as a disease of mitochondrial energy metabolism [14]. This theory is based on the observation of abnormal brain energy metabolism in depressed patients, such as reductions in high-energy phosphates or phosphomonoesters (observed using 31P NMR spectroscopy) in the frontal and temporal lobes of untreated patients compared with control subjects [15]. In contrast, increased phosphomonoester levels were observed in bipolar patients treated with lithium carbonate [16]. It has also been shown that the tricyclic antidepressant clomipamine causes changes in a range of mitochondrial functions possibly by interacting with proteins in the electron transport chain [17] and this could be a feature common to other antidepressants.

The mitochondrial dysfunction observed in bipolar disorder has been suggested to be the cause of the abnormal expression of genes encoding for mitochondrial proteins [12, 18, 19] and depression could be viewed as a primary mitochondrial abnormality with alterations in monoaminergic activity occurring as a secondary effect on gene transcription. The antidepressant-induced changes in cyclic AMP and glucocorticoid-driven gene expression reported by Abdel-Razaq et al. [20] and others [21, 22, 23] could be mediated by the effects of antidepressants on mitochondria.

Previous researchers have reported conflicting results on the effect of antidepressants on mitochondrial functions especially on mitochondrial-induced apoptosis. Some studies have reported that antidepressants induce cell death by apoptosis [17, 2426] while others have reported the opposite, i.e. an anti-apoptotic effect [2729].

Therefore, the aims of the present study were to investigate the effects of antidepressants of different chemical structure on several mitochondrial parameters including mitochondrial membrane potential, mitochondrial electron transport chain complex activities, mitochondrial oxygen consumption and caspase-3 activity in rat-heart isolated mitochondria and/or Chinese hamster ovary (CHO) cells.

Materials and Methods

Materials

Cell culture reagents were from Sigma Chemicals (Poole, Dorset, UK) except for foetal calf serum (FCS) which was from PAA Laboratories (Teddington, Middlesex, UK). The caspase-3 enzyme substrate (Ac-Asp-Glu–Val–Asp-amino-methylcourmarin (Ac-DEVE-AMC), the mitochondrial uncoupler, carbonyl cyanide m-chlorophenyl hydrazone (CCCP) and Adenosine-5’-diphosphate potassium salt (ADP) were obtained from Calbiochem, USA. Dithiothreitol (DTT) and CHAPS were from Fisher, U.K. The caspase-3 enzyme inhibitor benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (z-VAD-fmk) was from R&D Systems, U.K. Rhodamine-123 was from Cambridge Bioscience, U.K. dodecyl (lauryl) maltoside was from Fluka Chemicals, U.K. Cytochrome-c was obtained from Roche, Germany. Norfluoxetine was a generous gift from Eli Lilly. All other reagents were supplied by Sigma Chemicals (U.K.) unless otherwise stated.

In Vitro Model Cell System

A Chinese hamster ovary (CHO) cell line (CHOβ2SPAP; McDonnell et al. [30]) supplied by Professor S.J. Hill (School of Biomedical Sciences, University of Nottingham) was used. This cell line stably expresses the human β2-adrenoceptor and a temperature-resistant secreted placental alkaline phosphatase reporter gene. These cells were used previously to determine the effects of antidepressants on CREB/CRE-mediated gene transcription [20, 23].

Cell Culture

CHO cells were grown to confluency, in T75 culture flasks using Dulbecco’s Modified Eagle Media F12 HAM (DMEM F12 HAM), containing 2 mM L-glutamine and 10% FCS (complete medium). Cells were incubated at 37°C in a humidified 5% CO2: 95% air incubator. For assay experiments CHO cells were sub-cultured in 6 well plates or 24-well plates and grown for 48 h prior to drug treatment. All drug treatments were made in serum-free medium (SFM) by growing CHO in DMEM F12 HAM medium, containing 2 mM L-glutamine only. Cells were generally passaged twice weekly at a ratio of 1:20. All cell culture was performed under sterile conditions in a class II, laminar airflow cabinet.

Janus Green Assay

CHO cells (1 × 106 ml) were grown in 6 well plates in DMEM F12 HAM complete medium for 48 h. For 18 h pre-incubations with antidepressants, all medium was aspirated and replaced with SFM containing antidepressants at the required concentrations. Controls were run in parallel. After 18 h incubation with antidepressant, SFM ± antidepressant was carefully removed and the cell monolayer of cultured cells was washed once with PBS. 1 ml of 50% ethanol was added per well to fix the cells and incubated at room temperature for 30 min. After 30 min, ethanol was pipetted off and cell monolayer washed with PBS. The PBS was removed and cells were stained with 0.2% (w/v) Janus green solution, covering the bottom of the well, and incubated at room temperature for 15 min. Janus green stain was then removed and the plates were washed until they run clear and dried. To elute Janus green from the cells, 1 ml of 0.5 M HCl was added per well and incubated at room temperature for a further 20 min, swirling the plates carefully to allow the stain to be evenly distributed. After 20 min, 150 μl was transferred from each sample in triplicate to a 96 well plate and the plate was read at 595 nm, on a Molecular Devices, SPECTRAmax 384 spectrophotometric plate reader.

Isolation of Rat Heart Mitochondria

Rat heart mitochondria were isolated from adult male Wistar rats using homogenisation and differential centrifugation as previously described [31].

Measurement of Mitochondrial Membrane Potential

Changes in mitochondrial membrane potential (MMP) were measured in isolated rat heart mitochondria at the excitation and emission wavelengths of 503 and 525 nm respectively using Rhodamine123 (Rh123), according to the method described in Athanasiou et al. [32]. A Hitachi F2500 Fluorescence Spectrophotometer with fitted stirrer and water-jacketed cuvette holder was used for these experiments, and data analysed and displayed using the FL-Solution software program. Initially, mitochondrial complex I-linked substrates [20 μl glutamate (10 mM/ml) and 20 μl malate (2.5 mM/ml)], 20 μl EGTA (100 μM/ml) and 40 μl bovine serum albumin (BSA; 2.5 mg/ml) were added to mitochondrial respiratory buffer (made up to 2,000 μl after accounting for the volume of Rd123 and mitochondria added), which had been warmed to 37°C, in a clean dry cuvette. The cuvette was then placed into the fluorimeter and stirred continuously (at 37°C) using a magnetic stirrer bar. After few seconds, Rh123 (f.c. 0.2 μM) was added and the florescence measured for a further 60 s (there are no mitochondria so the fluorescent dye, Rd123, remained unbound and therefore a sharp increase in fluorescence was observed).

After approximately 100 s, isolated rat heart mitochondria (f.c. 0.5 mg protein/ml) and incubated for a further 5 min. A decrease in fluorescence was observed due to sequestering of the fluorescent dye, Rd123, by the intact functioning mitochondria. When a stability MMP was observed (as a stable baseline fluorescence), 20 μl of antidepressant (f.c. 50 μM) was added. As the MMP decreases, Rh123 dissociates from the mitochondria and fluorescence increases. Finally, 2 μl of the mitochondrial uncoupler carbonyl cyanide m-chlorophenyl hydrazone (CCCP; f.c. 10 nM) was added to the cuvette at the end of the experiment which resulted in a marked decrease in MMP. As it is a protonphore, CCCP completely collapses the MMP. Thus, any decrease in fluorescence caused by CCCP after the addition of antidepressants indicates that the mitochondria had not completely lost their membrane potential.

Preparation of Cell Lysates for Mitochondrial Enzymes Assays

CHO cells (1 × 106 ml−1) were grown in DMEM F12 HAM complete medium for 48 h. For 18 h preincubations with antidepressant all medium was aspirated and replaced with SFM containing antidepressants at the required concentrations. After 18 h incubation with antidepressant, the cells were washed once with PBS and harvested using non-enzymatic dissociation solution (from Sigma Chemicals, U.K.). Cells were centrifuged at 200 g for five minutes and the resultant cell pellets ware resuspended in 200 μl ice cold PBS before transferring the resuspended cells into microfuge tubes and centrifuging at 200 g for five minutes. The PBS was removed and the cell pellets were snap-frozen in liquid nitrogen and stored at −80°C until required.

On the day of experimentation, the cell pellets were thawed rapidly using a water bath at 37°C, then immediately snap-frozen in liquid nitrogen three times to lyse the cells. The lysed cell pellets were resuspended in 80 ml PBS. The resultant homogenates were kept on ice and used for measuring mitochondrial complexes activities according to Bates et al. [33] in a 250 μl reaction mixture in 96-well plates.

Measurement of Mitochondrial Complex I, Complex II–III and Complex IV Activity

Mitochondrial complex I activity, complex II-III activity and complex IV activities were all measured spectrophotometrically in a 96 well plate at 37°C on a Molecular Devices, SPECTRAmax 384 plate reader using methods described in Daley et al. [17]. Specifically, complex I (NADH-ubiquinone oxidoreductase) activity was measured as the rotenone-sensitive decrease in NADH concentration at 340 nm. Mitochondrial complex II/III (succinate cytochrome-c reductase) activity was measured as the succinate-dependent antimycin-A sensitive reduction of cytochrome-c at 550 nm. Mitochondrial complex IV (cytochrome oxidase) activity was measured as the decrease in the concentration of reduced cytochrome-c at 550 nm.

Caspase-3 Assay

The activity of caspase-3 was measured as the cleavage of a fluorescent AMC tag from a synthetic peptide substrate, Ac-DEVD-AMC (fluorogenic caspase-3 substrate). Cleavage of the AMC label by caspase-3 resulted in an increase in fluorescence at 460 nm.

CHO cells (1 × 106 ml) were grown in DMEM F12 HAM complete medium for 48 h. For 18 h pre-incubations with antidepressant all medium was aspirated and replaced with SFM containing antidepressants at the required concentrations. After 18 h incubation with antidepressant, the cells were washed once with PBS and suspended in lysis buffer (100 mM HEPES, 100 mM Sodium chloride, 10 mM DTT, 1 mM EDTA, 0.1% (w/v) CHAPS and 10% (w/v) Sucrose, pH 7.4 at 37°C). 10 μl of cell lysates was added, in duplicate, to 190 μl lysis buffer in a 96 well plate. In order to differentiate between specific caspase 3 activity and non-specific caspase 3 specific activity, one μl z-VAD-fmk (pan-caspase inhibitor, 10 mM) was added and incubated to one set of samples for 30 min at 37°C. Then two μl Ac-DEVE-AMC (caspase 3-specific fluorogenic substrate, 10 mM) was added to all wells and incubated for a further hour at 37°C. The production of AMC was measured at room temperature at excitation and emission wavelengths of 360 and 460 nm respectively using a fluorescence spectrophotometer (FLUOstar Galaxy from BMG Lab Technologies). Background readings and untreated controls were run in parallel.

Mitochondrial Oxygen Consumption Rate Measurements

Isolated rat heart mitochondria were used as an in vitro system to study the effects of antidepressants on mitochondrial oxygen consumption rate. The measurement of oxygen consumption rates was performed according to Bates et al. [6] using a custom-made incubation chamber with a water jacket and a micro-Clark electrode (Yellow Springs Instrument Co., OH, U.S.A.) fitted into the top of the chamber. The incubation medium was stirred constantly using an electromagnetic stirrer and magnetic stirring bar. The oxygen consumption studies were conducted at 30°C in mitochondrial respiration buffer (125 mM KCl, 2 mM KB2BHPOB4, 1 mM MgCl2, 1 μM EGTA and 20 mM Tris, pH was adjusted to 7.2 at 37°C). Approximately 0.2 mg of mitochondrial protein and 0.5 of fat-free BSA (to bind any free fatty acids in the preparation that can uncouple the mitochondria) were added into an incubation chamber in a total volume of 250 ml of respiration medium. State-3 respiration was induced by the addition of ADP, and the respiratory control ratio was calculated from the ratio of the state-3/state-4 oxygen consumption rates, with and without ADP, respectively.

232.5 μl of mitochondrial respiration buffer was added to the chamber and mitochondria were allowed to stir for approximately 2 min. 2.5 μl malate (5 mM), 2.5 μl glutamate (5 mM), 2.5 μl BSA (50 mg/ml) and 10 μl of mitochondria (approximately 0.25 mg protein) were added to the respiration buffer and allow stirring and mixing for two minutes. Antidepressants were added to the oxygen electrode using a Hamilton syringe, and state-3 respiration was induced by the addition of 3 μl of ADP (25 mM). The oxygen electrode and the incubation chamber were washed thoroughly with ethanol and double distilled water before the next incubation was carried out.

Western Immuno-Blotting for HSP70

Following exposure to antidepressants or vehicle, CHO cells were suspended and gently mixed in 300 μl of lysis buffer containing complete a protease inhibitor cocktail (Roche), then centrifuged at 15,000g at 4°C for 10 min. Supernatants (cytosolic fractions) were pipetted into clean microfuge tubes (with 10 μl samples being used for protein determinations). Samples were then boiled in 2× Laemmli solubilisation buffer containing 2.5% bromophenol blue at 95°C for 5 min, and then centrifuged at 15,000g for one minute.

Western blotting was carried out methods detailed in Abdel-Razaq et al. [20] using a rabbit polyclonal anti-HSP70 primary antibody supplied by ABCAM (Cambridge, U.K.) specific for the inducible form of Hsp70. A polyclonal goat anti-rabbit secondary antibody conjugated with horse-radish-peroxidase (HRP) supplied by DakoCytomation (Denmark) was used to bind with the primary antibody. The protein bands were detected with enhanced chemiluminescence using a GS-710 Imaging Densitometer (Bio-Rad) and processed using the Quantity One image analysis software.

Protein Measurements

Cell lysate and mitochondrial protein concentrations were assayed for total protein content using the Bio-Rad Lowry DC protein assay with bovine serum albumin as a concentration standard.

Statistical Analysis

Data are presented as mean ± standard error of the mean (SEM) of three or more independent experiments. Statistical analysis was performed using a one-way ANOVA followed by Dunnett’s multiple comparison tests. Concentration/response curve fitting was performed by nonlinear regression using the GraphPad Prism computer program (GraphPad, San Diego, CA, USA).

Results

In order to determine the effect of antidepressants on cellular viability, CHO cells were incubated with 50 μM CLOM and NORF (Fig. 1, panel A and B respectively). The CHO cells underwent distinct morphological changes (rounding and shrinkage) in about 20 min and most of cells were rounded after 30 min. Such morphological changes are consistent with the CHO cells undergoing apoptosis.
https://static-content.springer.com/image/art%3A10.1007%2Fs11064-010-0331-z/MediaObjects/11064_2010_331_Fig1_HTML.gif
Fig. 1

Effect of CLOM (50 μM, panel a) or NORF (50 μM, panel b) on cellular morphology of CHO cells continuously filmed under a light microscope using a Leica DMIRE2 time lapse microscope

Moreover, a colourimetric Janus green assay was performed on CHO cells treated with various concentrations of antidepressants in order to quantify the effect of antidepressants on cell viability and proliferation. Janus green is a basic azo dye which binds to protein and DNA, and the binding of Janus green is linearly correlated with cell number [34]. Figure 2 indicates that there was a concentration-dependent decrease in the number of viable cells after 18 h exposure to CLOM, DMI and NORF. However, numbers of viable TIAN-treated cells were no different to control cells.
https://static-content.springer.com/image/art%3A10.1007%2Fs11064-010-0331-z/MediaObjects/11064_2010_331_Fig2_HTML.gif
Fig. 2

Effect of antidepressants on CHO cells viability, after 18 h treatment. Data are presented as means ± SEM of triplicate samples (P = 3). *P < 0.05; **P < 0.01 compared with control (100%)

In order to verify the apparent apoptotic effect of antidepressants, CHO cells were exposed to increasing concentrations of CLOM, DMI, NORF or TIAN followed by measurement of caspase-3 activity. During apoptosis signalling, the latent forms of caspases undergo rapid proteolytic processing and activation, and caspase-3 activation is a well known marker of apoptosis in mammalian cells [17, 35]. Figure 3 shows a significant concentration-dependent increase in caspase-3 activity in CHO cells treated with DMI, CLOM and NORF. TIAN did not, however, change caspase-3 activity.
https://static-content.springer.com/image/art%3A10.1007%2Fs11064-010-0331-z/MediaObjects/11064_2010_331_Fig3_HTML.gif
Fig. 3

Relative caspase-3 activity in CHO cells exposed to a range of antidepressant concentrations for 18 h period after 60 min incubation with caspase-3 substrate (Ac-DEVE-AMC). Data points are means ± SEM of triplicate samples (n = 3). *P < 0.05; **P < 0.01 compared with the control (basal) caspase-3 activity

The acute effects of antidepressants on integrated mitochondrial function were assessed by monitoring MMP using Rh123 in mitochondria isolated from rat hearts. Hearts were used since it would have been very difficult to isolate sufficient mitochondrial protein from the CHO cells. Figure 4 shows the effect of various antidepressants on MMP. Both CLOM and NORF caused a marked increase in fluorescence (indicating a decrease in MMP). DMI also showed a similar but smaller effect, whereas, TIAN did not provoke any significant change in MMP.
https://static-content.springer.com/image/art%3A10.1007%2Fs11064-010-0331-z/MediaObjects/11064_2010_331_Fig4_HTML.gif
Fig. 4

Effect of antidepressants (50 μM) on MMP in rat heart isolated mitochondria. Rd123 Rodamine123, CCCP carbonyl cyanide-m-chlorophenyl hydrazone (mitochondrial uncoupler). Panels represent single representative experiments repeated on at least two further occasions with essentially identical results

It is possible that the apoptotic cell death observed in antidepressant-treated cells could result from the effects of antidepressants on mitochondria. Since any disruption of the mitochondrial membrane could result in the release of mitochondrial cytochrome-c which represents a critical event in activation of a cascade of apoptotic caspases during apoptosis [36]. The effects on MMP indicated that the antidepressants could act via modulation of enzymes in the mitochondrial electron transport chain (ETC.). This was then investigated by measuring the enzymatic activities of the ETC. complexes in cell lysates after treating CHO cells with antidepressants for an 18 h-period using established spectrophotometric assays [17].

Mitochondrial complex I activity was decreased with all antidepressants used at concentrations greater than 20 μM (Fig. 5). NORF showed the greatest inhibition of complex I activity.
https://static-content.springer.com/image/art%3A10.1007%2Fs11064-010-0331-z/MediaObjects/11064_2010_331_Fig5_HTML.gif
Fig. 5

Effect of antidepressants on mitochondrial complex I activity in CHO cells after 18 h treatment. Data are presented as means ± SEM of triplicate samples (n = 3). *P < 0.05; **P < 0.01 compared with control

Mitochondrial complex II/III activity was decreased with all of the antidepressants used, except for TIAN (Fig. 6). The inhibition of complex II/III activity occurred at concentration as low as 1 μM in case of CLOM and DMI. NORF showed significant inhibition at concentrations in excess of 10 μM.
https://static-content.springer.com/image/art%3A10.1007%2Fs11064-010-0331-z/MediaObjects/11064_2010_331_Fig6_HTML.gif
Fig. 6

Effect of antidepressants on mitochondrial complex II/III activity in CHO cells, after 18 h treatment. Data are presented as means ± SEM of triplicate samples (n = 3). *P < 0.05; **P < 0.01 compared with control

Mitochondrial complex IV activity showed a significant inhibition following all treatments, except for TIAN (Fig. 7). However, the magnitude of inhibition of complex IV activity was greater than for the other complexes and no detectable enzyme activity remained at higher concentrations of antidepressants.
https://static-content.springer.com/image/art%3A10.1007%2Fs11064-010-0331-z/MediaObjects/11064_2010_331_Fig7_HTML.gif
Fig. 7

Effect of antidepressants on mitochondrial complex IV activity in CHO cells after 18 h treatment. Data are presented as means ± SEM of triplicate samples (n = 3). *P < 0.05; **P < 0.01 compared with control

The effect of antidepressants on mitochondrial oxygen consumption (respiration) was measured in isolated rat heart mitochondria using glutamate and malate as substrates according to Bates et al. [6]. None of the antidepressants used showed a significant effect on state-4 respiration (without ADP). However, in the presence of ADP (state-3 respiration), all antidepressants showed a significant decrease in the oxygen consumption rate at concentrations above 10 μM compared to the control values (untreated mitochondria; Figs. 8, 9, 10, 11).
https://static-content.springer.com/image/art%3A10.1007%2Fs11064-010-0331-z/MediaObjects/11064_2010_331_Fig8_HTML.gif
Fig. 8

Effect of CLOM on mitochondrial oxygen consumption. Data points are means ± SEM of triplicate samples (n = 5). **P < 0.01 compared with control

https://static-content.springer.com/image/art%3A10.1007%2Fs11064-010-0331-z/MediaObjects/11064_2010_331_Fig9_HTML.gif
Fig. 9

Effect of DMI on mitochondrial oxygen consumption. Data points are means ± SEM of triplicate samples (n = 5). **P < 0.01 compared with control

https://static-content.springer.com/image/art%3A10.1007%2Fs11064-010-0331-z/MediaObjects/11064_2010_331_Fig10_HTML.gif
Fig. 10

Effect of NORF on mitochondrial oxygen consumption. Data points are means ± SEM of triplicate samples (n = 5). *P < 0.05; **P < 0.01 compared with control

https://static-content.springer.com/image/art%3A10.1007%2Fs11064-010-0331-z/MediaObjects/11064_2010_331_Fig11_HTML.gif
Fig. 11

Effect of TIAN on mitochondrial oxygen consumption. Data points are means ± SEM of triplicate samples (n = 5). *P < 0.05; **P < 0.01 compared with control

HSP70 is a well known member of the family of heat shock proteins which are critical for normal cellular functions and are also involved in the normal functioning of mitochondria [37]. The expression levels of HSP70 in CHO cells preincubated for 18 h with various concentrations of CLOM, DMI, NORF or TIAN were measured by immuno-blotting. Densitometric analysis of HSP70 blots normalized to the amount of β-actin protein showed that CLOM, DMI and NORF induced a slight, but not statistically significant, increase in HSP70 at 50 μM (data not shown).

Discussion

Mitochondria play a central role in normal cellular functions since cells depend on mitochondrial oxidative phosphorylation to provide them with the majority of their energy requirements, in the form of ATP. Therefore, any disturbance or abnormality in the function of mitochondria may lead to cellular dysfunction and/or death.

A number of studies have reported that antidepressants might cause impairment in both mitochondrial function and ion compartmentation [26, 38]. It has also been demonstrated that treatment of animals with imipramine in vitro inhibited the respiratory activity of rat liver and brain mitochondria. However, in vivo treatment for up to 2 weeks has been reported to show significant stimulation of the respiratory activity of rat liver and brain mitochondria [39]. These conflicting data may be due to the drugs causing an acute inhibition of mitochondrial function [17], but, in the longer term results in up-regulation of mitochondrial function and whole organ function, as has been seen in rat heart following a whole body thermal shock [9].

In the present study, the antimitochondrial effects of some antidepressants were investigated in CHO cells and in isolated rat heart mitochondria. CHO cells treated with CLOM and NORF, showed distinct morphological changes (Fig. 1) characterised by rounding and shrivelling as observed in apoptotic cells. These morphological changes, characteristic of apoptosis caused by the antidepressants, were confirmed biochemically by a concentration-dependent increase in caspase-3 activity in the CHO cells exposed to CLOM, DMI and NORF (Fig. 3). These observations are consistent with those of Qi et al. [40] who reported that acute exposure of rat glioma cells to DMI caused inhibition of cellular proliferation in a concentration- and time-dependent manner, which resulted in apoptotic cell death via a caspase-3-dependent pathway. In contrast however, TIAN did not cause any significant change in caspase 3 activity.

Other studies have reported the opposite effect for antidepressants. For example, Li and Luo [28] reported that DMI, at concentrations of 1 and 5 μM, resulted in an anti-apoptotic effect in cultured rat phaeochromocytoma (PC12) cells by antagonizing corticosterone-induced apoptosis. Similarly, it has also been reported that antidepressant treatment with TIAN reduces stress-induced apoptosis in the hippocampus [29].

However, there is increasing evidence that several antidepressants, mainly selective serotonin-reuptake inhibitors (SSRIs) and tricyclic antidepressants (TCAs) induce cell death by caspase-3-dependent apoptosis, in different cell lines [24, 25, 41, 42]. Moreover, some widely used antidepressants such as CLOM, imipramine and citalopram have been demonstrated to possess anti-neoplastic effects due to their causing apoptosis selectively in tumour cells [17, 42].

The MMP is the transmembrane electrochemical (proton) gradient which is generated by the mitochondrial ETC. complexes during the process of oxidative phosphorylation. This proton gradient is in turn responsible for the synthesis of ATP by the ATP synthase (complex V, which is not part of ETC.). For this reason, MMP is critical for the maintenance of mitochondrial functionality and is a measure of integrated mitochondrial function, and the general energy status of the cell. In isolated rat-heart mitochondria some antidepressants particularly NORF and CLOM caused the MMP to decrease (Fig. 4). Since MMP is a product of the proton gradient across the inner mitochondria membrane due to the transfer of protons outside the mitochondrial matrix by complexes I, III and IV, changes in MMP might be due to inhibitory effects of antidepressants on these complexes.

At concentrations greater than 20 μM all antidepressants used (CLOM, DMI, NORF and TIAN) decreased complex I activity, by about 50% of the control activity. Drug effects on complex II/III activity showed more selectivity, since CLOM, DMI and NORF, but not TIAN, decreased activity at concentrations as low as 10 μM with more than 70% inhibition at antidepressant concentrations greater than 20 μM. The reduction in the activity of complex II/III could be due to a direct effect of the antidepressants as suggested by Daley et al. [17] who reported that CLOM selectively inhibited complex III. Similarly complex IV activity showed a pronounced inhibition at high antidepressant concentrations (more than 90% in the presence of DMI and more than 99% in the presence of CLOM and NORF). Similar results were also reported by Hroudova and Fisar [43] who measured the inhibitory effects of some antidepressants and mood stabilizers on the activities of respiratory electron transport chain complexes in crude mitochondrial fraction isolated from pig brain.

However, the data are not entirely consistent with a recent study that reported that CLOM did not induce any significant effect on complex IV activity [17], although this could be a time-dependant effect, as a sequential decrease in different complex activity has been observed in some models where a particular complex has been inhibited [44, 45]. Complex IV is responsible for transferring electrons to molecular oxygen atoms in addition to pumping more protons across the inner mitochondrial membrane to the intermembrance space. Genetic alterations of this complex are a common cause of mitochondrial diseases. For example, it has been demonstrated that inhibition of complex IV activity is involved in the pathogenesis of many neurodegenerative disease such as Alzheimer’s disease [46].

The present data suggest that complex I may be more sensitive to antidepressant inhibition than other ETC. complexes since all antidepressants reduced its activity. Recently, a significant role of complex I dysfunction/inhibition has been recognised since it has been found that many human mitochondrial diseases (e.g. Parkinson’s disease) involve structural and functional defects at the level of this enzyme complex [8, 33, 47].

Interestingly, TIAN which did not produce any change in cell morphology, inhibited only complex I activity. It has been reported that TIAN, which is metabolized mainly by beta-oxidation of its heptanoic side chain, inhibited both mitochondrial beta-oxidation and the TCA cycle [48]. However, the metabolic pathways controlling oxidative phosphorylation and the TCA cycle are linked. Therefore, inhibition of complex I activity will cause an increase in the intra-mitochondrial NADH/NAD+ ratio and slow down the TCA cycle by feedback inhibition, as NADH inhibits the activities of pyruvate dehydrogenase, isocitrate dehydrogenase, α-ketoglutarate dehydrogenase, and citrate synthase [49]. However, TIAN-treated cells can still survive even at high TIAN concentrations since complex I can be bypassed during the oxidative phosphorylation process to generate enough ATP for cell survival via direct input of reducing equivalents from FADH2 into complex II, or in addition it may be that glycolysis is either up-regulated or able to support cellular ATP demands in the absence of normal levels of oxidative phosphorylation. However, CLOM, DMI and NORF induced a significant inhibition of all mitochondrial enzyme complexes, which may make it difficult for the cells to survive, as significant amounts of cytochrome c would be released due to decreases in MMP in the CHO cells, which would initiate apoptosis.

The observed inhibitory effect of antidepressants on the mitochondrial ETC. complexes could be due to direct binding of the antidepressants to the complexes which in turn would lead to an increase in reactive oxygen species (ROS) production at the inhibited complexes as seen in the case of drugs of the cannabinoid family [50]. These ROS may damage the mitochondria, causing a decrease in MMP and a loss of mitochondrial integrity, possible due to opening of the permeability transition pore and or a movement of cytochrome-c from the mitochondria to the cytosol. However, there are insufficient data in the literature describing the effects of antidepressants on the mitochondrial complexes activities to enable definite conclusions to be drawn.

The effects of antidepressants on mitochondrial oxygen consumption rates were measured since changes in the ETC. are manifested by decreasing the ability of mitochondria to utilize oxygen. The data showed that all antidepressants used decreased the state-3 respiration rates (in the presence of ADP) but not state-4 (without ADP). The decrease in oxygen consumption was more than 50% of the control rate in the case of CLOM, DMI and NORF at drug concentrations more than 20 μM (Figs. 8, 9, 10, respectively), whereas a concentration of 40 μM TIAN was needed to reach 50% inhibition (Fig. 11). The inhibition of the state-3 oxygen consumption rates almost reached the values of the state-4 oxygen consumption rates, except in the case of TIAN exposure.

It is possible that the decrease observed in mitochondrial oxygen consumption was due to the decreased activity of complex IV. Although TIAN did not show any inhibition in the complex IV activity, it still decreased state 3 respiration rate. However, the inhibition of state 3 respiration in the case of TIAN did not exceed more than 60% of the control. These observations are consistent with mitochondrial control theory [51] which proposes that the level of activity of any particular ETC. complex must be decreased to a critical “threshold” before any of the three accepted measures of integrated mitochondrial function (i.e. mitochondrial oxygen consumption, MMP or ATP production rate), are significantly decreased.

In contrast to our data, a study by Katyare and Rajan [39] reported a significant stimulation of state 3 respiration rate in rat brain mitochondria following imipramine administration in vivo, which was accompanied with increased mitochondrial content of cytochrome b and c. However, these in vivo effects might reflect adaptive changes in response to acute physiological and/or pharmacological challenges as described in Sammut et al. [9].

The heat shock protein HSP70 (sometimes referred to as HSP72) is thought to play an important role in protection and tolerance to a number of stressors, including exposure to cytotoxic drugs, glucose deprivation, virus infection, alterations in intracellular redox state and oxidative stress [52], and it is involved in the transport and assembly of mitochondrial membrane proteins, including respiratory complexes. However the levels of HSP70 in CHO cells treated with CLOM, DMI, NORF or TIAN did not show any significant change even at high concentrations (50 and 100 μM).

Although it is well accepted that nuclear genes have control over some mitochondrial activities, it has also been recognized, conversely, that mitochondria can influence nuclear gene expression [53]. For example, it has been reported that mitochondrial dysfunction induces activation of CREB/CRE-mediated gene transcription via an intracellular signalling pathway that affects several kinase-dependent pathways, particularly the cyclic AMP signalling pathway [54]. The effect of mitochondria on the expression of specific nuclear genes depends on the mitochondrial functional state and the condition of mitochondrial DNA [55]. This phenomenon has been called “molecular retrograde communication” and implies that drugs having direct effects on the mitochondria can initiate adaptive changes in cell function via altered nuclear gene transcription. It is, therefore, possible that some of the changes in gene expression accompanying antidepressant treatments might result from initial interaction with the mitochondria, in addition to any effects on synaptic neurotransmitter availability resulting from their accepted acute mechanisms of action [20, 23]. Antidepressants are proposed to protect stress-vulnerable neurones [56] and the mitochondrial challenge by the drugs could be viewed as being equivalent to a pre-conditioning effect in which stimulation below the threshold of injury results in subsequent protection.

This innovative idea is consistent with the hormetic dose responses reported by Calabrese et al. [57] who describes the concept of hormetic dose effects mediated by endogenous cellular defense mechanisms which could provide neuroprotection against stresses, particularly those related to the mitochondrial redox signaling. Thus, this could represent as a novel target for therapeutic intervention in neurodegenerative diseases such as depression at low exposure doses of antidepressants.

In this context the mitochondria have been proposed to be the “master integrators” of neuroprotection [58] and antimitochondrial agents such as the respiratory chain inhibitor nitroproprionic acid [59] are known to protect neurones from subsequent ischemic damage. Conversely, a more extensive inhibition of mitochondrial function could be responsible for some of the unwanted side-effects of antidepressant drugs. The in vitro data reported herein are supported by reports that antidepressants acutely reduce brain oxidative metabolism in vivo [60] perhaps leading to subsequent protection.

Conclusions

In summary, the data presented clearly show that antidepressant drugs of different structures have common anti-mitochondrial effects which, we propose, can promote changes in gene transcription which could be part of the mechanism for initiating some of the adaptive responses that underlie the clinical effects of the drugs but could also underlie unwanted side effects.

Acknowledgments

The authors are grateful for the support of the Hashemite University (Jordan) and the University of Nottingham (UK).

Copyright information

© Springer Science+Business Media, LLC 2010