Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Permeabilization of mitochondrial membranes triggered by proapoptotic molecules and damage pathways is implicated as an important event in the control of cell death and survival and may be involved in both apoptosis and necrosis (for a recent review, see Kroemer and Reed, 2000). Inner membrane permeabilization causes dissipation of the mitochondrial membrane potential, uncoupling of oxidative phosphorylation, ATP depletion, and equilibration of small solutes and ions between the cytosol and the mitochondrial matrix. Solute influx into the matrix can cause swelling followed by rupture of the outer membrane and consequent efflux of proteins from the intermembrane space, including cytochrome c, procaspase 9, apoptosis-inducing factor, and endonuclease G. However, outer membrane permeabilization is not always related to inner membrane permeabilization but seems also to occur independently.

Several models of mitochondrial inner membrane permeabilization, specifically the mitochondrial permeability transition (MPT), are currently under discussion. Onset of the MPT is mediated by opening of a proteinaceous permeability transition (PT) pore that probably forms at contact sites between the inner and outer membranes. The PT pore is postulated to form through combination of the adenine nucleotide translocator of the inner membrane, the voltage dependent anion channel of the outer membrane, cyclophilin D (CypD) of the mitochondrial matrix and several other proteins, including the peripheral benzodiazepine receptor (outer membrane), creatine kinase (intermembrane space), hexokinase II (exterior side of the outer membrane), Bax and Bcl-2 (Zoratti and Szabò, 1995; Zamzami et al., 1998) An alternative model proposes that PT pores form by clustering of misfolded membrane proteins that are capped by chaperone-like regulatory and renaturing proteins such as CypD (He and Lemasters, 2002).

CypD binding to PT pores is required for the Ca²⁺-dependent onset of the MPT. CypD also confers sensitivity of the MPT to cyclosporin A (CsA), and CsA is an effective and specific inhibitor of the MPT at submicromolar concentrations. However, CsA also binds cyclophilin A, leading to inhibition of calcineurin. Calcineurin dephosphorylates the apoptogenic protein Bad and enables it to bind other antiapoptotic Bcl-2 family members and trigger release of cytochrome c and other proapoptotic proteins from the intermembrane space (Kroemer and Reed, 2000). These two mechanisms, inhibition of the MPT and inhibition of calcineurin, may account for the many reports of cytoprotective actions of CsA against necrotic and apoptotic cell death (for example, see Nieminen et al., 1995; Matsuura et al., 1996; Halestrap et al., 1997; Bradham et al., 1998; Seaton et al., 1998; Scheff and Sullivan, 1999; Li et al., 2000). Attempts to distinguish between these two possibilities have been made in a few cases (Zamzami et al., 1996; Trost and Lemasters, 1997; Seaton et al., 1998) using the analog N-methyl-4-valine-cyclosporin (also called PKF220-384 or SDZ220-384), which is not immunosuppressive (Zenke et al., 1993) but binds to cyclophilins and inhibits the MPT (Petronilli et al., 1994). However, PKF220-384 is no longer available to researchers. Here, we provide evidence that another nonimmunosuppressive cyclosporin derivative, NIM811 (Rosenwirth et al., 1994), a compound that is still available, blocks the MPT induced by calcium plus phosphate in the presence and absence of enhancing agents, such as 1-methyl-4-phenylpyridinium and rotenone, in a fashion comparable with CsA and PKF220-384. NIM811 also blocked apoptotic cell death and mitochondrial depolarization and inner membrane permeabilization in situ in hepatocytes exposed to tumor necrosis factor- (TNF) in a fashion nearly identical to CsA but without the cytotoxicity of CsA at high concentrations.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials

For photometric and fluorometric measurements, a Wallac Victor 1420 Multilabel Counter (PerkinElmer Wallac, Gaithersburg, MD) fitted with a 535DF35 filter or 485DF22 excitation/590DF35 emission filters, respectively (Omega Optical, Brattleboro, VT), was used.

Rotenone and carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone (FCCP) were obtained from Sigma (Buchs, Switzerland), valinomycin and oligomycin from Sigma-Aldrich Chemie (Schnelldorf, Germany) and 1-methyl-4-phenylpyridinium iodide (MPP⁺) from RBI (Natick, MA). Safranin T was purchased from Fluka (Buchs, Switzerland). TNF was from R&D Systems, Inc. (Minneapolis, MN). Tetramethylrhodamine methylester and calcein AM were obtained from Molecular Probes (Eugene, OR). CsA, PKF220-384, NIM811, cyclosporin H (CsH; PKF037-839), and PSC833 are marketed by or are experimental compounds of Novartis.

Methods

Preparation of Mitochondria. Rat liver mitochondria were prepared by differential centrifugation essentially as described by Schnaitman and Greenawalt (1968). The liver of a rat weighing 300 to 400 g was washed three times with ice-cold isolation buffer (70 mM sucrose, 190 mM mannitol, 20 mM HEPES, 0.2 mM EDTA, brought to pH 7.5 with 1 M NaOH). The liver was cut into three to five pieces and then cross-chopped three times on a McIlwain tissue chopper (Mickle Laboratory Engineering Co., Gomshall, Surrey, UK). The chopped tissue was transferred into a 60-ml glass homogenizer fitted with a Teflon piston and further homogenized by two up-and-down strokes, diluted to 200 ml with ice-cold isolation buffer, and centrifuged at 650g for 10 min at 2°C. The supernatant, divided into eight ~25-ml portions, was recentrifuged at 7700g for 10 min at 2°C. The pellets were carefully suspended in 25 ml each of ice-cold washing buffer (same as isolation buffer, but without EDTA). The suspensions were recentrifuged at 7700g for 10 min at 2°C, the supernatants were discarded, and the pellets were suspended in 2 ml each of respiratory buffer (RB; 70 mM sucrose, 190 mM mannitol, 20 mM HEPES, 5 mM glutamate, and 0.5 mM malate, brought to pH 7.5 with 1 M NaOH). A total of about 20 ml of suspension was obtained, containing about 13 to 20 mg of mitochondrial protein/ml, as determined using a BCA assay (Pierce, Rockford IL). This suspension was stored on ice until further used.

Swelling Assay. Mitochondrial swelling caused by influx of solutes through open PT pores results in an increase in light transmission (i.e., a reduced turbidity; for discussion, see Zoratti and Szabò, 1995). This turbidity change offers a convenient and frequently used assay of the MPT by measurement of absorbance in mitochondrial suspensions. In the present study, the MPT induced by Ca²⁺ and inorganic phosphate (Ca²⁺/P_i) was monitored by absorbance changes at 535 nm by adapting the procedure of Cassarino et al. (1999) to a microtiterplate assay.

Measurement of Mitochondrial Membrane Potential. The same experimental conditions were used for the assessment of alterations of the mitochondrial membrane potential , except that safranin in 10 µl of RB was added after the cyclosporins at a final concentration of 15 µM. This concentration was determined beforehand as the optimal compromise between signal/baseline ratio and interference of safranin itself with swelling induced by Ca²⁺/P_i (safranin tended to enhance Ca²⁺/P_i-induced swelling at concentrations above 20 µM). Fluorescence readings were done using the filter combination 485DF22 excitation/590DF35 emission. Baseline values (F_baseline) reflecting the resting potential were measured before the addition of the tool compounds and/or Ca²⁺/P_i. After the repetitive measurements (F_X), FCCP at a final concentration of 1 µM was added, and fluorescence was measured again (yielding F_FCCP). Data were expressed as means of (F_FCCP  F_X)/(F_FCCP  F_baseline) for each well ± S.E.M.

Isolation and Culture of Hepatocytes. Hepatocytes were isolated from overnight-fasted male Sprague-Dawley rats (200-250 g) by collagenase perfusion of livers, as described previously (Gores et al., 1988). Cell viability routinely exceeded 90%, as determined by trypan blue exclusion. Hepatocytes were then cultured in Waymouth's MB-7521/1 medium containing 27 mM NaHCO₃, 2 mM L-glutamine, 10% fetal calf serum, 100 nM insulin, and 100 nM dexamethasone. For cell viability assay, hepatocytes were plated onto 24-well microtiter plates (Falcon, Lincoln Park, NJ) coated with 0.1% Type 1 rat-tail collagen at a density of 1.5 × 10⁵ cells/well in 1 ml of medium. For confocal microscopy, hepatocytes were cultured on collagen-coated 40-mm glass coverslips placed in 60-mm plastic Petri dishes at a density of 1 to 2 × 10⁶ hepatocytes/dish in 4 ml of medium. Hepatocytes were used after overnight (14-16 h) incubation in humidified 5% CO₂/95% air at 37°C.

Cell Viability Assay. Viability of hepatocytes cultured on multiwell plates was monitored by propidium iodide fluorometry using a fluorescence scanner (FLUOstar 403; BMG LabTechnologies, Durham, NC). Briefly, hepatocytes in 24-well plates were incubated with 30 µM propidium iodide. Fluorescence from each well was measured using excitation and emission wavelengths of 544 nm (25-nm band pass) and 590 nm (35-mm band pass), respectively. For each experiment, an initial fluorescence measurement (A) was made 20 min after addition of propidium iodide and then at intervals thereafter. Individual experiments were terminated with 375 µM digitonin to permeabilize all cells, and a final fluorescence measurement (B) was obtained 20 min later. The percentage of viable cells (V) was calculated as V = 100(B  X)(B  A), where X is fluorescence at any given time. Cell killing in this assay corresponds to that assessed by trypan blue nuclear staining (Nieminen et al., 1992).

Induction of Apoptosis. Rat hepatocytes were infected with an adenovirus (Ad5IB) expressing an IB super-repressor that blocks the antiapoptotic NFB signaling pathway to sensitize hepatocytes to TNF-induced apoptosis, as described previously (Bradham et al., 1998), and 2 µM t-butylhydroperoxide was added to accelerate the response to TNF. Briefly, 2 h after plating of hepatocytes, the culture medium was changed to hormonally defined medium (HDM) containing 30 plaque-forming units/cell of Ad5IB for 2 h at 37°C. After culturing overnight, 2 µM t-butylhydroperoxide was added, followed by 30 ng/ml TNF 1 h later. After 16 h of exposure to TNF, cell viability was determined by propidium iodide fluorometry.

Confocal Microscopy. For confocal microscopy, TNF-treated hepatocytes were incubated in HDM supplemented with 25 mM HEPES, pH 7.4, to stabilize pH during confocal measurements. Approximately 30 min before imaging, the hepatocytes were loaded with 250 nM tetramethylrhodamine methyl ester (TMRM) and 1 µM calcein-AM in TNF-free HDM for 15 min. After loading, the hepatocytes were washed, and the original TNF-containing HDM was added back. The coverslips were then mounted on the stage of a Zeiss LSM-410 inverted laser scanning confocal microscope in a Focht Chamber System (Bioptechs, Butler, PA) to maintain temperature at 37°C. The green fluorescence of calcein and the red fluorescence of TMRM were excited simultaneously with the 488- and 568-nm lines of an argon-krypton laser. Fluorescence was divided by a 568-nm emission dichroic reflector and recorded by separate photomultipliers through 515 to 565 nm band pass and 590-nm long pass barrier filters. The argon laser was operated at approximately 50% of full power, and the laser output was attenuated with 0.3 to 1% neutral density filters to minimize photodamage and photobleaching. Pinholes were set to 0.9 airy units in the red and green channels to maximize z-axis resolution, which was less than 1 µm.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Comparison of Effects of Ca²⁺/P_i on Swelling and . The effects of graded concentrations of Ca²⁺ in combination with 1 mM P_i on swelling in the absence or presence of 15 µM safranin and on as evidenced by changes in fluorescence at 485/590 nm are shown in Fig. 1A-C. The results show a good temporal correspondence between loss of absorbance and increase in safranin fluorescence as reflected by the decrease of the quotient (F_FCCP  F_X)/(F_FCCP  F_baseline) (compare Fig. 1, A and C). Note that the decrease of this quotient occurred more abruptly than the loss of absorbance. This was a consistent finding (see Figs. 3 and 5).

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Effects of graded Ca2+ concentrations on swelling in the absence (A) or presence (B) of safranin, and on safranin fluorescence (C) Swelling: 20 µl of mitochondrial preparation and 160 µl RB were added to each well and the baseline values (0 min in the graphs) were measured. Then, 10 µl of Ca2+ (graded concentrations) were added with a 12-tip pipette at 10-s intervals per row, followed by 10 µl of Pi 2 min later, also at 10-s intervals per row. Repeated absorbance measurements were started exactly 2 min after addition of Ca2+ to the first row. Data are means ± S.E.M. of quadruplicates. Safranin fluorescence: 20 µl of mitochondrial preparation, 150 µl of RB, and 10 µl of safranin (final concentration, 15 µM) were added to each well and the baseline fluorescence (Fbaseline) was measured. The values depend on the concentrations of mitochondrial proteins and range between about 50,000 and 70,000 fluorescence units among different experiments. Ca2+/Pi were added and repetitive fluorescence measurement started as above, yielding Fx values. Fluorescence values obtained after full depolarization ranged between 170,000 and 230,000 among different experiments. After termination of the repetitive measurements, 10 µl of FCCP was added and the plate remeasured once, yielding FFCCP. Data represent (FFCCP  FX)/(FFCCP  Fbaseline) for each well ± S.E.M. of quadruplicates.

Calibration of Dependence of Safranin Fluorescence. Experiments were run in parallel with 10, 15, and 20 µM safranin in RB, isolation buffer, or the buffer used by Reers et al. (1991) (200 mM sucrose, 20 mM mannitol, 20 mM 4-morpholinopropanesulfonic acid, 1 mM EDTA). The results were very similar; therefore, only those obtained with 15 µM safranin in RB are shown in Fig. 2. A sigmoid relationship between the change in safranin fluorescence and was observed with a practically linear section between 60 and 160 mV. These results showed that safranin fluorescence measurements faithfully reflected changes in in a relevant potential range.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Calibration of dependence of safranin fluorescence on . To each well of a 96-well plate, 20 µl of mitochondrial preparation containing 200 to 400 µg of mitochondrial protein, 140 µl of RB, 20 µl of RB containing 20 µM rotenone, 200 nM valinomycin, and 50 µM oligomycin (yielding final concentrations of 1 µM rotenone, 20 nM valinomycin and 2.5 µM oligomycin) and 10 µl of RB containing 300 µM safranin (final concentration, 15 µM) were added and the plate read at 485/590 nm, yielding Fbaseline. Then, 10 µl of RB containing graded concentrations of KCl (to reach final concentrations of 0.1, 0.3, 0.8, 1.5, 3, 10, and 20 mM K+, four wells for each K+ concentration) were added at intervals of 10 s to each row from top to bottom using a 12-tip pipette. The plate was put into the reader and measurements started such that the first well was read 2 min after pipetting KCl to the first row, yielding F2min. Thereafter, 10 µl of RB containing 20 µM FCCP (final concentration 1 µM) were added and the plate remeasured, yielding FFCCP. From these data, (FFCCP  F2min)/(FFCCP  Fbaseline) was calculated for each well and plotted against , calculated using the Nernst equation.

View larger version (47K): [in this window] [in a new window]   Fig. 3.   Prevention of Ca2+/Pi-induced swelling and changes by cyclosporin derivatives. The procedure was exactly as described in the legend of Fig. 1. To 20 µl of mitochondrial preparation, 140 µl of RB and 20 µl of RB containing appropriate concentrations of CsA (A and B), CsH (A and B), PKF220-384 (C and D), and NIM811 (E and F) (figures in captions, in micromolar) were added, baseline absorbance was measured, Ca2+ (final concentration, 100 µM) and Pi (final concentration, 1 mM) were added, and repetitive absorbance readings were started. For measurements of safranin fluorescence, 130 instead of 140 µl of RB and 10 µl of safranin (final concentration, 15 µM) were added, Fbaseline was read, Ca2+ and Pi were added, FX read was repetitively, FCCP in 10 µl of RB (final concentration 1 µM) was added and reread. Data are absorbance readings (A, C, E, and G) and (FFCCP  FX)/(FFCCP  Fbaseline) (B, D, F, and H) ± S.E.M. of quadruplicates. PSC833 (G and H) was measured in a separate experiment, in which a Ca2+ concentration of 200 µM was used; everything else was identical.

Effects of Cyclosporin Derivatives on Swelling and Changes Caused by Ca²⁺/P_i. The effects of graded concentrations of CsA, PKF220-384 and NIM 811 on Ca²⁺/P_i-induced swelling and loss of were compared. Figure 3 shows that all these compounds caused a dose-dependent inhibition of swelling and loss of with partial inhibition beginning at about 0.2 µM. Full inhibition by CsA and NIM811 occurred at 1 µM and greater. PKF220-384 was a little more potent than CsA and NIM811, causing full protection against the MPT at 0.5 µM.

Interaction of Cyclosporins with the Effects of Rotenone and MPP⁺. Rotenone at 0.1 µM did not induce swelling or affect when added to the mitochondria in the absence of Ca²⁺/P_i. However, in the presence of Ca²⁺/P_i, rotenone delayed swelling by about 6 min without affecting the decrease in caused by the Ca²⁺/P_i (not shown). At 1 µM, rotenone added alone caused a slow, gradual loss of (Fig. 4B) but did not cause swelling (Fig. 4A). In combination with Ca²⁺/P_i, an immediate and complete collapse of occurred, but the swelling induced by Ca²⁺/P_i was almost completely prevented. When CsA, PKF220-384, or NIM811 was combined with rotenone plus Ca²⁺/P_i, no swelling was observed; collapse of was initially as complete as after rotenone plus Ca²⁺/P_i alone. However, partially recovered thereafter to a maximum extent after about 12 min and then slowly subsided again (Fig. 4, A and B). With 10 µM rotenone, the results were similar except that the rebound of did not occur (not shown).

View larger version (30K): [in this window] [in a new window]   Fig. 4.   Combination with rotenone (A and B) and MPP+ (C and D). To 20 µl of mitochondrial preparation, 130 µl of RB and 20 µl of RB containing 10 µM CsA, PKF220-384, or NIM811 (to reach final concentrations of 1 µM) were added, baseline absorbance was measured, MPP+ (final concentration, 600 µM) or rotenone (rot; final concentration 1 µM) in 10 µl of RB was added at 10-s intervals per row, followed likewise 2 min later by Ca2+ (final concentrations, 50 µM in the MPP+ experiment and 200 µM in the rotenone experiment) and again 2 min later by Pi (final concentration, 1 mM), and repetitive absorbance readings started. For measurements of safranin fluorescence, 120 instead of 130 µl of RB and 10 µl of safranin (final concentration, 15 µM) were added, Fbaseline was read, Ca2+ and Pi was added, FX was read repetitively, and FCCP in 10 µl of RB (final concentration, 1 µM) was added and reread. Data are absorbance readings (A and C)) and (FFCCP  F2min)/(FFCCP  Fbaseline) values (B and D) ± S.E.M. of quadruplicates.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Inhibition by CsA and Nim811 of TNF-induced apoptosis in primary cultured rat hepatocytes. Hepatocytes infected with Ad5IkB were treated with t-butylhydroperoxide and TNF, as described under Experimental Procedures, in the presence of 0 to 10 µM CsA or Nim811. Cell killing was assessed after 16 h by propidium iodide fluorometry.

Protection against TNF-Induced Apoptosis by NIM811. Many cells, including cultured rat hepatocytes, resist TNF-mediated apoptosis via activation of an NFB survival pathway (Wang et al., 1996; Bradham et al., 1998). To make hepatocytes sensitive to TNF, we infected with an IB (S34A/S36A)-superrepressor expressing adenovirus (Ad5IB) that prevents activation of antiapoptotic NFB signaling after TNF addition (Iimuro et al., 1998). We also exposed the Ad5IB-infected hepatocytes to a small nontoxic concentration of t-butylhydroperoxide (t-BuOOH, 2 µM), because preliminary experiments determined that pretreatment in this way accelerated onset of apoptosis after subsequent TNF addition (T. Qian and J. J. Lemasters, unpublished observations). Ad5IB and t-butylhydroperoxide alone or in combination at the concentrations used did not cause cell killing in the absence of TNF.

View larger version (72K): [in this window] [in a new window]   Fig. 6.   TNF-induced onset of the MPT in cultured hepatocytes. Sensitized hepatocytes were loaded with 200 nM TMRM (a potential-indicating fluorophore) and 1 µM calcein AM (a probe of membrane permeability) and exposed to TNF, as described under Experimental Procedures and Fig. 5. The red fluorescence of TMRM and green fluorescence of calcein were imaged by confocal microscopy. Note that after 6 h of exposure to TNF, mitochondria depolarized and the inner membrane became permeable, as indicated by release of TMRM fluorescence from individual mitochondria and redistribution of calcein from the cytosol into mitochondria.

View larger version (106K): [in this window] [in a new window]   Fig. 7.   Inhibition by NIM811 of TNF-induced onset of the MPT in cultured hepatocytes. Sensitized hepatocytes were treated with TNF in the presence of 2 µM Nim811 and loaded with TMRM and calcein, as described in Fig. 6. In the presence of Nim811, the MPT was blocked, as indicated by no loss of red mitochondrial TMRM fluorescence and the absence of redistribution of calcein fluorescence into mitochondria.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Comparative studies of drug effects on the MPT by means of absorbance measurements using the conventional cuvette technique are cumbersome, because the mitochondria have to be freshly prepared by differential centrifugation. By the time they are ready, almost half a day is spent. Moreover, they can be used only for a few hours, because their susceptibility toward induction of the MPT increases with the time elapsed since their preparation (see below). This limits the number of samples that can be evaluated per experiment. Therefore, it was useful to set up an assay that allowed simultaneous measurement of several samples. The most straightforward way to achieve this goal was the microtiterplate format. We developed microtiterplate assays of mitochondrial swelling and and used them in the present study to compare the effects of CsA with those of its nonimmunosuppressive derivatives PKF220-384 and NIM811. Swelling was assessed by absorbance decreases, and mitochondrial by the fluorescence of safranin, a cationic fluorophore that accumulates into polarized mitochondria in proportion to (Akerman and Wikstrom, 1976). Uptake of safranin leads to fluorescence quenching that can be measured with the fluorescence reader (Fiskum et al., 2000).

To obtain a baseline absorbance providing a suitable absorbance difference in the microtiterplate reader upon MPT induction, a concentration of 1 to 2 mg/ml mitochondrial protein in the wells was necessary. Although a lower concentration of mitochondrial protein would have sufficed for the measurements with safranin, it seemed essential for comparability between absorbance and to maintain this mitochondrial concentration and all other parameters constant. Accordingly, a relatively high safranin concentration was needed to provide good fluorescence signals. The calibration of the safranin response using K⁺ diffusion potentials (Fig. 2) indicated that safranin fluorescence was sensitive to changes over the potential range of interest for the conditions used.

The susceptibility of rat liver mitochondria to MPT induction by Ca²⁺ measured by absorbance can vary from one preparation to another. The minimal Ca²⁺ concentration causing maximal absorbance loss within 12 min after exposure ranged between 75 and 300 µM in over 50 experiments but was 100 or 150 µM in about 70% of the cases. Moreover, this susceptibility increased with the time elapsed after mitochondrial isolation. On average, the MPT occurred twice as fast after 3.5 h compared with 30 min after finishing the preparation of mitochondria. Although variation from one preparation to another occurred, the increase of susceptibility with time after isolation always occurred. This should be borne in mind when the presented absorbance and data are compared, because they were not obtained simultaneously, but sequentially. Although matching `absorbance' and `safranin' plates were always read immediately one after the other in this sequence, a time difference of 30 to 40 min could not be avoided. Therefore, a small lead of the changes over the corresponding absorbance decreases should not be interpreted as significant. For these reasons, we conclude from the obtained data that changes in absorbance and after exposure of mitochondria to Ca²⁺/P_i occurred virtually simultaneously within the temporal resolution of our experimental setup. This correspondence was maintained if the effects of Ca²⁺/P_i were enhanced by the prooxidant, t-BuOOH.

MPP⁺ at 600 µM enhanced the absorbance loss caused by Ca²⁺/P_i and similarly enhanced the Ca²⁺/-induced dissipation of , in agreement with the report of Cassarino et al. (1999) who ascribed this to an increase in the production of oxygen radicals. Rotenone, in contrast, delayed and at higher concentrations prevented Ca²⁺/-induced swelling, also as described by Cassarino et al. (1999). This inhibition by rotenone is probably caused by its inhibition of NADH-linked respiration (Chernyak and Bernardi, 1996). More efficient than 600 µM MPP⁺, rotenone rapidly and completely dissipated in the presence of Ca²⁺/P_i. Alone, rotenone, like MPP⁺, caused a slow, gradual loss of . It seems plausible to attribute these effects of rotenone and MPP⁺ on to complex I inhibition. The difference between MPP⁺ and rotenone on changes to supports the idea that the neurotoxicity and effect on the MPT of MPP⁺ are not simply caused by complex I inhibition (Cassarino et al., 1999; Nakamura et al., 2000).

Ca²⁺/-induced swelling and dissipation were prevented by CsA, NIM811 and PKF220-384. PKF220-384 was about twice as potent as CsA and NIM811. The concentration-response relationships were very steep. The formyl peptide receptor antagonist CsH (de Paulis et al., 1996) was ineffective on both swelling and loss of at 10 µM, but the P-glycoprotein inhibitor PSC833 did block the MPT at 30 µM, in agreement with its weak interaction with cyclophilin D (Perkins et al., 1998). In additional experiments not shown here, CsA, NIM811, and PKF220-384 at 1 µM also completely prevented swelling and dissipation induced by the combination of Ca²⁺/P_i with the enhancing agents atractyloside, PK11195, and the prooxidant t-butylhydroperoxide. Ca²⁺/-induced swelling enhanced by 0.1 µM FCCP (which by itself is ineffective despite lowering ) was also completely abolished by CsA, NIM811, and PKF220-384 without alteration of the effect of the uncoupler on . Although the cyclosporins abolished the induction of swelling by the combination of MPP⁺ and Ca²⁺/P_i, the loss of was attenuated but not entirely prevented. The pattern obtained with rotenone was different in that the cyclosporins did not affect the initial collapse of by rotenone + Ca²⁺/P_i but did promote a subsequent partial repolarization.

NIM811 also inhibited apoptosis in Ad5IB-sensitized hepatocytes exposed to TNF. Cytoprotection by NIM811 was associated with prevention of the MPT, as indicated by the blockade of mitochondrial depolarization and inner membrane permeabilization that otherwise occurred after TNF treatment. Prevention by NIM of apoptosis and the MPT in cultured hepatocytes was the same as that observed previously by CsA (Bradham et al., 1998). These findings support the conclusion that a CsA- and NIM811-sensitive MPT occurs during apoptotic signaling in TNF-treated hepatocytes. This MPT causes mitochondrial depolarization and inner membrane depolarization and leads to mitochondrial swelling and release of proapoptotic intermembrane proteins, such as cytochrome c.

The cytoprotective effects of CsA and NIM811 were identical only at lower concentrations (0.5-2 µM). At higher concentrations (5-10 µM), CsA lost efficacy. Such a biphasic dose-response curve for CsA has been described previously in experiments with isolated myocytes, perfused hearts, and hepatocytes (Nazareth et al., 1991; Griffiths and Halestrap, 1993; Qian et al., 1997). At high concentrations, cyclosporin A is cytotoxic (Ruiz-Cabello et al., 1994; Jiang and Acosta, 1995). This toxicity may overcome the beneficial effect of CsA on the MPT. By contrast, the efficacy of NIM811 was not lost at higher concentrations. The differential efficacy of NIM811 and CsA may be related to the different effects of these cyclosporin derivatives on calcineurin. Thus, loss of protection by CsA at high concentration may be cytotoxicity caused by inhibition of calcineurin. For this reason, NIM811 seems to be more efficacious than CsA in preventing injury in intact cells.

In conclusion, the present results suggest that the nonimmunosuppressive cyclosporin derivative, NIM811, affects the MPT in much the same way as CsA and PKF220-384, another nonimmunosuppressive derivative. The latter was used in the past to discriminate between MPT- and calcineurin-related effects of CsA in studies of cell protection but is no longer available. NIM811 may fill this gap.