The Putative Drp1 Inhibitor mdivi-1 Is a Reversible Mitochondrial Complex I Inhibitor that Modulates Reactive Oxygen Species

Abstract

Mitochondrial fission mediated by the GTPase dynamin-related protein 1 (Drp1) is an attractive drug target in numerous maladies that range from heart disease to neurodegenerative disorders. The compound mdivi-1 is widely reported to inhibit Drp1-dependent fission, elongate mitochondria, and mitigate brain injury. Here, we show that mdivi-1 reversibly inhibits mitochondrial complex I-dependent O₂ consumption and reverse electron transfer-mediated reactive oxygen species (ROS) production at concentrations (e.g., 50 μM) used to target mitochondrial fission. Respiratory inhibition is rescued by bypassing complex I using yeast NADH dehydrogenase Ndi1. Unexpectedly, respiratory impairment by mdivi-1 occurs without mitochondrial elongation, is not mimicked by Drp1 deletion, and is observed in Drp1-deficient fibroblasts. In addition, mdivi-1 poorly inhibits recombinant Drp1 GTPase activity (K_i > 1.2 mM). Overall, these results suggest that mdivi-1 is not a specific Drp1 inhibitor. The ability of mdivi-1 to reversibly inhibit complex I and modify mitochondrial ROS production may contribute to effects observed in disease models.

Keywords: bioenergetics; brain; fission; fragmentation; mitochondria; neuron; respiration; reverse electron transfer; succinate; superoxide.

Figure 1. Mdivi-1 reversibly inhibits basal and maximal respiration at Complex I

(A) OCR traces for neurons or (B) COS-7 cells receiving mdivi-1 or DMSO vehicle (CTRL), FCCP (3 μM for neurons, 2 μM for COS-7) plus pyruvate (Pyr, 10 mM), and antimycin A (AA, 1 μM). Traces are mean ± SD from 3 wells and are representative of 4 independent experiments (2–3 wells each). (C) FCCP together with pyruvate or together with TMPD (0.4 mM) plus ascorbate (0.4 mM) were injected along with mdivi-1 (100 μM) or CTRL while measuring OCR. The Complex IV inhibitor azide (5 mM) was then injected. (D) Neurons were permeabilized by saponin (sap, 25 μg/ml), and OCR was stimulated by 1 mM ADP in the presence of pyruvate and malate (P/M, 5 mM each) or succinate (S, 5 mM) in the presence of rotenone (R, 0.5 μM). Mdivi-1 (50 μM) or vehicle control was then injected. (E) Neurons were treated with rotenone or mdivi-1 (50 μM) for 1 hr prior to OCR measurements. Rotenone or mdivi-1 was then either left on for the duration of the assay (“present”), or washed out and replaced with drug-free aCSF (“washout”). Following permeabilization by sap and addition of ADP with glutamate and malate (G/M, 5 mM each), 5 mM succinate was added. (F) Neurons or COS-7 cells were treated with DMSO or mdivi-1 (50 μM) for 1 or 5 hr. The drug was then washed out and maximal OCR was determined. See also Figure S1.

Figure 2. Mdivi-1 does not alter mitochondrial size in neuronal processes

(A) Representative immunofluorescence of Tom20 (red) and Drp1 (green) in neurons following treatment with DMSO vehicle (CTRL) or mdivi-1 (75 μM) for 1 or 5 hr (scale bars, 10 μm). Extrasomal Tom20 and Drp1 fluorescence is primarily localized to a meshwork of neuronal processes. (B) Frequency of binned mitochondrial areas for the treatments described in (A). (C) Pearson’s coefficient for Drp1 co-localized with mitochondrial Tom20 in neurons treated with CTRL or 75 μM mdivi-1. Results are mean ± SD, where ~800–1000 mitochondria were counted per treatment group per experiment, and 3 separate experiments were conducted. See also Figure S2.

Figure 3. Mdivi-1 is a poor inhibitor of human Drp1 GTPase activity

Substrate kinetics of recombinant human Drp1 was measured at 0–100 μM mdivi-1. Data were globally fit to an uncompetitive model yielding a K_i > 1.2mM. The data and SEM are for 3 independent preparations of Drp1. Residuals to the fit are shown above the graph. See also Figure S3.

Figure 4. Mdivi-1-induced respiratory inhibition is not mimicked by Drp1 KO or dependent upon Drp1 expression

WT (A) or Drp1 KO (B) MEFs were treated with mdivi-1 (25–100 μM) or vehicle followed by FCCP (3 μM) plus pyruvate (10 mM) and antimycin A (AA, 1 μM) while OCR was measured. Traces here and below are mean ± SD of 3 wells and representative of at least 3 experiments. (C) Maximal OCR calculated immediately following FCCP plus pyruvate addition, expressed as a percentage of the maximal rate in vehicle-treated cells (mean ± SD, n=3). (D) Drp1 KO MEFs were permeabilized with 5 μg/mL saponin, along with ADP (1 mM) and either pyruvate and malate (5 mM each) or succinate (5 mM) in the presence of rotenone (0.5 μM). MEFs were then treated with 50 μM mdivi-1 or vehicle (CTRL). OCR is expressed as a percentage of OCR prior to drug injection. Results are mean ± SD, n=3. * p<0.05 compared to control. (E) Electron micrographs of WT or Drp1 KO MEFs treated with CTRL or mdivi-1 (50 μM) for 16 min. Scale bars = 1 μm. (F) Cumulative frequency distribution plot for mitochondrial length in WT or Drp1 KO MEFs ± mdivi-1. Bar graph inset is mean mitochondrial length ± SD. Results are from 50 mitochondria from 2 different experiments. See also Figures S4–S6.

Figure 5. Yeast Ndi1 NADH dehydrogenase prevents respiratory inhibition by mdivi-1

(A) Ndi1 and β-actin protein levels in COS-7 cells (COS) transfected with the NDI1 gene and selected for NDI1 expression compared to parental COS. (B) COS or NDI1-transfected COS (Ndi1) were treated with vehicle (CTRL), rotenone (1 μM), or piericidin A (100 nM), followed by FCCP (2 μM) plus pyruvate (10 mM), and then antimycin A (1 μM) while OCR was measured. (C) OCR traces for Ndi1-COS-7 or parental COS-7 cells treated with DMSO or mdivi-1, followed by the drugs in (B). Traces in (B) and (C) are mean ± SD of 3 wells and representative of 3 experiments. (D) Bar graph showing quantification of traces in (C). See also Figure S7.

Figure 6. Mdivi-1 increases ROS in WT and Drp1 KO fibroblasts but not neurons

(A) DHE fluorescence (arbitrary units, A.U.) was measured over time in neurons. Rotenone (1 μM) or mdivi-1 (50 μM) was added as indicated (drug). Numbers are rates (A.U./min, mean ± SEM) before and after drug addition. (B) The rate of DHE fluorescence change following drug addition was divided by the basal rate to determine fold change. Piericidin A was added at 0.5 μM. DMSO was the vehicle for mdivi-1 and EtOH was the vehicle for rotenone and piericidin A. (C) DHE fluorescence measured over time in WT or Drp1 KO MEFs. Mdivi-1 was 50 μM. Numbers are rates before and after mdivi-1 addition. (D) and (E) The rate of DHE fluorescence change following drug addition was calculated in either WT (D) or Drp1 KO (E) MEFs as in (B). All data in bar graphs are mean ± SEM, n=5. * p<0.05.

Figure 7. Mdivi-1 modulates ROS production by brain mitochondria as predicted by Complex I inhibition

(A) Representative traces depicting H₂O₂ production by isolated brain mitochondria oxidizing glutamate and malate (G/M, 5 mM each) before and after addition of DMSO (CTRL) or mdivi-1 (50 μM), followed by addition of ADP (2 mM), and then rotenone (1 μM). (B) Quantification of the data in (A) (mean ± SD, n=5). * p<0.05 compared to control, @ p<0.05 compared to the respective condition with no ADP, # p<0.05 for mdivi-1 rate after ADP compared to control rate after ADP. Each rotenone (Rot) bar is the mean of 2 experiments. (C) Representative traces depicting RET-mediated H₂O₂ production by brain mitochondria incubated with succinate, followed by addition of CTRL or mdivi-1, and then ADP (2 mM). (D) Quantification of the data in (C) (mean ± SD, n=3). * p<0.05 compared to control.