Regulation of mitochondrial dynamics in acute kidney injury in cell
culture and rodent models ([original](https://doi.org/10.1172%2FJCI37829)) ([raw](?raw))Mitochondrial fragmentation occurs in response to apoptotic stress in rat proximal tubular cells. Renal ischemia and nephrotoxicity are the major causes of acute kidney injury. To examine mitochondrial morphological changes under this condition, we transfected MitoRed into cultured rat proximal tubular cells (RPTCs) to label mitochondria with red fluorescence. The cells were subsequently subjected to azide-induced ATP depletion to model in vivo ischemia or cisplatin treatment for nephrotoxicity. As shown in Figure 1A, mitochondria in control cells were filamentous with a tubular or thread-like appearance and were often interconnected to form a network. During azide treatment, the mitochondrial network broke down and the mitochondria were fragmented into short rods or spheres. Mitochondrial fragmentation was azide-treatment time dependent, concurring with apoptosis (Figure 1B). Cisplatin also induced mitochondrial fragmentation in RPTCs in a time-dependent manner (Figure 1C). Notably, mitochondrial fragmentation during cisplatin treatment clearly preceded apoptosis and, as a result, fragmentation was observed at 4 hours, whereas apoptosis was not detected until 12 hours (Figure 1C). A time-lapse study further recorded mitochondrial fragmentation during azide-induced cell injury. A very rapid/sudden fragmentation of all mitochondria in individual cells was shown after about 2 hours of azide treatment (Supplemental Video 1; supplemental material available online with this article; doi:10.1172/JCI37829DS1). The results demonstrate an early and striking morphological change of mitochondria during tubular cell injury.
Mitochondrial fragmentation following ATP depletion and cisplatin treatment in RPTCs. RPTCs were transfected with MitoRed to fluorescently label mitochondria. The cells were then incubated with 10 mM azide in glucose-free medium to induce ATP depletion or treated with 20 μM cisplatin in cell culture medium. Mitochondrial morphology in MitoRed-labeled cells was evaluated by fluorescence microscopy to determine the percentage of cells that fragmented mitochondria. Apoptosis was assessed in these cells by cellular and nuclear morphology. (A) Representative images of mitochondrial morphology. Left panel, an untreated control RPTC showing long filamentous mitochondria with a thread-like appearance. Right panel, an azide-treated (3 hours) cell showing shortened punctate mitochondria. Scale bars: 5 μm. (B) Time courses of mitochondrial fragmentation and apoptosis during azide-induced ATP depletion. (C) Time courses of mitochondrial fragmentation and apoptosis during cisplatin incubation. Data in B and C are presented as mean ± SD; n = 3.
Mitochondrial fragmentation can be inhibited by Bcl2 but not by caspase inhibitors. While mitochondrial fragmentation could play a causative role in mitochondrial injury and apoptosis, it may also be a result of cell death. We therefore tested the effects of carbobenzoxy-valyl-alanyl-aspartyl-(O-methyl)-fluoromethylketone (VAD), a pan-caspase inhibitor that blocks apoptosis during ATP depletion and cisplatin treatment of RPTCs. Although caspase activation and apoptosis were inhibited (not shown), VAD did not suppress mitochondrial fragmentation during either azide or cisplatin treatments (Figure 2, A and B), suggesting that mitochondrial fragmentation is not secondary to caspase activation. We further tested the effects of Bcl2 on mitochondrial fragmentation during RPTC apoptosis. As shown in Figure 2, mitochondrial fragmentation induced by azide and cisplatin was suppressed in Bcl2-transfected cells. For example, azide treatment induced mitochondrial fragmentation in 70% of RPTCs, which was reduced to 20% in Bcl2-overexpressing RPTCs (Figure 2A). Interestingly, Bcl2 but not VAD attenuated azide-induced mitochondrial permeabilization, reflected by lower cytochrome_c_ release into cytosol (Figure 2C). The results suggest that Bcl2 may protect against mitochondria injury and apoptosis in part by suppressing mitochondrial fragmentation.
Inhibition of mitochondrial fragmentation and membrane permeabilization by Bcl2 and not by VAD. Wild-type and Bcl2-overexpressing RPTCs were transfected with MitoRed to label mitochondria. The cells were then treated with azide (10 mM, 3 hours) or cisplatin (20 μM, 16 hours) in the absence or presence of 100 μM VAD. Mitochondrial morphology in individual cells was evaluated by fluorescence microscopy to determine the percentage of cells with mitochondria fragmentation. (A) Mitochondrial fragmentation during azide-induced ATP depletion. (B) Mitochondrial fragmentation during cisplatin incubation. (C) Cytochrome c (Cyt.c) release during azide treatment. Cells were fractionated to collect cytosolic fraction for immunoblot analysis of cytochrome_c_. Data in A and B are presented as mean ± SD; n ≥ 3. *P < 0.05, significantly different from untreated control;#P < 0.05, significantly different from azide- or cisplatin-treated RPTCs. Results show that Bcl2 (but not caspase inhibitors) can suppress mitochondrial fragmentation and outer membrane permeabilization. Ctrl, control.
Drp1 is activated and translocates to mitochondria early during azide-induced ATP depletion. The morphology of mitochondria is determined by 2 opposing processes, fission and fusion (20–22). Accordingly, mitochondrial fragmentation observed in our study could be a result of activated fission, suppressed fusion, or both. We demonstrated the activation of Drp1, a fission protein, during azide treatment of RPTCs. As shown in Figure 3A, compared with untreated cells, 1–3 hours of azide treatment induced an accumulation of Drp1 in the mitochondrial fraction. Azide-induced Drp1 translocation was also confirmed by dual immunofluorescence staining of Drp1 and Fis1, an integral mitochondrial protein involved in fission (20–22). As shown in Figure 3B, colocalization in Drp1 and Fis1 was shown in some azide-treated cells. Of note, consistent with previous reports (27, 28), Drp1 appeared to be restricted to specific sites along the mitochondrial membrane, where fission may have been occurring.
Drp1 translocation to mitochondria during azide-induced ATP depletion. (A) Immunoblot analysis of Drp1 in mitochondrial and cytosolic fractions. RPTCs were subjected to 0 to 3 hours of 10 mM azide treatment for ATP depletion and then fractionated into mitochondrial and cytosolic fractions for immunoblot analysis of Drp1. (B) Dual immunofluorescence staining of Drp1 and Fis1 in azide-treated cells. After 3 hours of ATP depletion by azide treatment, RPTCs were fixed for immunofluorescence of Drp1 and Fis1. Drp1 and Fis1 signals were examined by confocal microscopy, and both separate and merged images are shown. The results show the translocation of a portion of Drp1 to mitochondria, overlapping with Fis1 during ATP depletion. Scale bars: 5 μm.
Dominant-negative Drp1 inhibits mitochondrial fragmentation during RPTC injury. To determine whether Drp1 is indeed involved in mitochondrial fragmentation during tubular cell apoptosis, we tested the effects of dominant-negative Drp1 (DN-Drp1) that had a K38A point mutation (28). RPTCs were transfected with DN-Drp1 or empty vector and then treated with azide or cisplatin. As shown in Figure 4A, DN-Drp1 reduced mitochondrial fragmentation from 64% to 28% during azide treatment. Similarly, DN-Drp1 significantly inhibited mitochondrial fragmentation during cisplatin incubation (Figure 4B). Representative images are shown in Figure 4C. Untreated cells had typical filamentous mitochondria; after azide treatment, the mitochondria became fragmented and punctate, but the cells transfected with DN-Drp1 retained their filamentous mitochondria. These data suggest a role for Drp1 and associated fission in mitochondrial fragmentation during tubular cell apoptosis.
Suppression of mitochondrial fragmentation during RPTC injury by DN-Drp1. RPTCs were cotransfected with MitoRed and DN-Drp1 or empty vector. Cells were then incubated with 10 mM azide for 3 hours or 20 μM cisplatin for 16 hours. (A) Effects of DN-Drp1 on mitochondrial fragmentation during azide-induced ATP depletion. (B) Effects of DN-Drp1 on mitochondrial fragmentation during cisplatin treatment. (C) Representative images of mitochondria. Left upper panel, untreated control cells showing long thread-like filamentous mitochondria; middle upper panel, fragmented mitochondria in vector-transfected cells after azide treatment; right upper panel, cells transfected with DN-Drp1 retaining filamentous mitochondria after azide treatment. Higher-magnification images of the framed areas are shown in the bottom panels. Scale bars: 5 μm (upper panels); 1 μm (lower panels). Data in A and B are presented as mean ± SD;n ≥ 3. *P < 0.05, significantly different from untreated control; #P < 0.05, significantly different from treated group transfected with empty vector.
Inhibition of cytochrome c release during RPTC injury by DN-Drp1. Next, we determined whether blocking mitochondrial fragmentation with DN-Drp1 would inhibit cytochrome c release. RPTCs were cotransfected with MitoRed and DN-Drp1 or empty vector and then subjected to azide-induced ATP depletion. Cytochrome c release was examined by immunofluorescence. Of note, the transfection efficacy in RPTCs was 20%–30%. Thus, our examination was focused on the transfected (MitoRed labeled) cells to determine the effects of DN-Drp1. Representative images are shown in Figure 5A. In control cells, cytochrome c was sequestered in mitochondria and overlapped with the MitoRed signal. After azide treatment, mitochondria became fragmented and cytochrome c was released from mitochondria and distributed throughout the cytosol. However, DN-Drp1–transfected cells retained filamentous mitochondria and maintained cytochrome c in the organelles. Cell counting showed that DN-Drp1 significantly inhibited cytochrome c release during azide treatment, reducing cytochrome _c_–released cells from 61% to 26% (Figure 5B). In contrast, DN-Drp1 did not suppress azide-induced Bax accumulation in mitochondria (Figure 5C). The results suggest that Drp1-mediated mitochondrial fragmentation contributes to mitochondrial damage and cytochrome c release, although it does not affect initial Bax activation or translocation.
Inhibition of cytochrome c release during ATP depletion by DN-Drp1. RPTCs were cotransfected with MitoRed and DN-Drp1 or empty vector and then treated with 10 mM azide for 3 hours. The cells were fixed for immunofluorescence of cytochrome c or Bax. The examination was focused on the transfected (MitoRed labeled) cells to determine the effects of DN-Drp1. (A) Representative images of MitoRed and cytochrome_c_ staining. The staining was examined by confocal microscopy in the same cells. In untreated cells (control), cytochrome c staining colocalized with MitoRed in filamentous mitochondria. Following azide treatment, mitochondria in vector-transfected cells became fragmented and cytochrome c was released into the cytosol. Cells transfected with DN-Drp1 retained their filamentous mitochondria, and cytochrome_c_ was retained in mitochondria. Scale bars: 5 μm. (B) Quantification of the effects of DN-Drp1 on cytochrome_c_ release. The localization of cytochrome c in MitoRed-transfected cells was evaluated to determine the percentage of cells that released cytochrome c into cytosol. (C) Quantification of the effects of DN-Drp1 on Bax translocation to mitochondria. The localization of Bax in MitoRed-transfected cells was evaluated to determine the percentage of cells that showed Bax accumulation in mitochondria. Data inB and C are presented as mean ± SD;n ≥ 3. *P < 0.05, significantly different from untreated control; #P < 0.05, significantly different from azide-treated cells that were transfected with empty vector.
Inhibition of RPTC apoptosis by DN-Drp1. Once released, cytochrome c activates caspases to result in apoptosis (29). By blocking mitochondrial fragmentation and cytochrome c release (Figures 4 and 5), DN-Drp1 was expected to inhibit caspase activation and apoptosis. To test this, we cotransfected RPTCs with GFP and wild-type Drp1, DN-Drp1, or empty vector. The cells were then subjected to azide treatment followed by recovery in normal medium to analyze apoptosis and caspase activation. By TUNEL assay, azide induced 51% apoptosis in vector-transfected cells, which was decreased to 27% in DN-Drp1–transfected cells (Figure 6A). Consistently, DN-Drp1 suppressed the development of apoptotic morphology in transfected (GFP-labeled) cells, including cellular shrinkage and nuclear fragmentation (Figure 6B). DN-Drp1 also suppressed caspase activation (not shown). In addition, DN-Drp1 reduced cisplatin-induced apoptosis from 55% to 32% (Figure 6C).
Inhibition of ATP depletion–induced apoptosis by DN-Drp1. (A) Effects of DN-Drp1 on azide-induced apoptosis. RPTCs were cotransfected with GFP and DN-Drp1 or empty vector. Cells were then treated with 10 mM azide for 3 hours followed by 2 hours recovery. After treatment, cells were subjected to TUNEL assay. Cells were examined by fluorescence microscopy to determine the percentage of apoptosis (TUNEL positive) in transfected (GFP labeled) cells. (B) Representative cell morphology. RPTCs were transfected with wild-type or DN-Drp1 and then subjected to 3 hours of azide treatment followed by 2 hours recovery. Cells were stained with Hoechst 33342 and examined by fluorescence microscopy. Scale bars: 5 μm. (C) Effects of DN-Drp1 on cisplatin-induced apoptosis. RPTCs were cotransfected with GFP and DN-Drp1 or empty vector and then incubated with 20 μM cisplatin for 16 hours. Cells were stained with Hoechst 33342 for examination by fluorescence microscopy to determine the percentage of apoptosis in transfected (GFP labeled) cells. Data in A and C are presented as mean ± SD; n ≥ 3. *P < 0.05, significantly different from untreated control;#P < 0.05, significantly different from vector-transfected cells treated with azide or cisplatin.
siRNA knockdown of Drp1 blocks mitochondrial fragmentation, cytochrome c release, and apoptosis. We further confirmed the role of Drp1 in mitochondrial regulation and apoptosis by using an RNA interference approach. Stable Drp1-siRNA cell lines, including R3 and R24, were generated by transfection of RPTCs with a specific short hairpin siRNA of Drp1 (30). Drp1 knockdown in R3 and R24 cells was verified by immunoblot analysis (Figure 7A). In response to azide treatment, both R3 and R24 cells showed significantly lower mitochondrial fragmentation than parental RPTCs (Figure 7B). Moreover, these cells demonstrated less apoptosis and cytochrome_c_ release (Figure 7, C and D). Cisplatin-induced mitochondrial fragmentation, cytochrome c release, and apoptosis were also suppressed in these cells (data not shown).
siRNA knockdown of Drp1 inhibits mitochondrial fragmentation, cytochrome_c_ release, and apoptosis following ATP depletion in RPTCs. (A) RPTCs were transfected with Drp1 short hairpin siRNA to select 2 stable cell lines: R3 and R24. Knockdown of Drp1 in R3 and R24 cells was verified by immunoblot analysis. (B) RPTCs, R3 cells, and R24 cells were transfected with MitoRed and then incubated with 10 mM azide for 3 hours. Cells were examined by fluorescence microscopy to enable counting of cells with mitochondrial fragmentation. (C) RPTCs, R3 cells, and R24 cells were incubated with 10 mM azide for 3 hours followed by 2 hours recovery to evaluate apoptosis by morphological criteria. (D) RPTCs, R3 cells, and R24 cells were incubated with 10 mM azide for 3 hours. Cells were then fractionated to collect the cytosolic fraction for immunoblot analysis of cytochrome_c_. Data in B and C are presented as mean ± SD; n ≥ 3. *P < 0.05, significantly different from untreated control RPTCs;#P < 0.05, significantly different from azide-treated RPTCs.
Mitochondrial fragmentation and apoptosis in primary proximal tubular cells are inhibited by DN-Drp1. To further substantiate the role of mitochondrial fragmentation in renal injury, we examined mitochondrial fragmentation during cisplatin treatment of primary proximal tubular cells. Primary cultures of proximal tubular cells had highly filamentous mitochondria (Figure 8A). Upon cisplatin treatment, however, the mitochondria became fragmented and punctate (Figure 8A). The fragmentation was inhibited by DN-Drp1 (Figure 8A). Cell counting indicated that DN-Drp1 reduced mitochondrial fragmentation from 55% to 31% (Figure 8B). DN-Drp1 also suppressed cisplatin-induced apoptosis in the primary cells (Figure 8C). To determine whether Drp1 suppression affects upstream signaling during cisplatin treatment, we examined cisplatin-induced p53 phosphorylation and p53-upregulated modulator of apoptosis α (PUMA-α) induction (31) in RPTCs stably transfected with Drp1 siRNA. Downregulation of Drp1 in the siRNA cells was verified above, as shown in Figure7A. As shown in Supplemental Figure 1, Drp1 siRNA cells showed p53 and PUMA-α induction similar to that of the parental wild-type RPTCs, suggesting that Drp1 suppression does not affect upstream signaling during cisplatin treatment. Together, these data strongly support a role for Drp1-mediated mitochondrial fragmentation in tubular cell apoptosis.
Mitochondrial fragmentation and its inhibition by DN-Drp1 in primary cultures of proximal tubular cells. Proximal tubular cells were isolated from renal cortex of male C57BL/6 mice for primary culture. Primary cells were cotransfected with MitoRed and DN-Drp1 or empty vector. Cells were then incubated with 50 μM cisplatin for 24 hours. Mitochondrial morphology was examined by fluorescence microscopy. Apoptosis in transfected cells was evaluated by morphological criteria. (A) Representative mitochondrial morphology. Scale bars: 5 μm. Insets show higher magnification of the framed areas. (B) Effects of DN-Drp1 on mitochondrial fragmentation. (C) Effects of DN-Drp1 on apoptosis. Data in B and C are presented as mean ± SD;n ≥ 3. *P < 0.05, significantly different from untreated control; #P < 0.05, significantly different from cisplatin-treated vector-transfected cells.
Mitochondrial fragmentation occurs in vivo in proximal tubular cells following renal ischemia. To examine mitochondrial fragmentation in vivo, C57BL/6 mice were subjected to 30 minutes of bilateral clamping to induce renal ischemia followed by brief (15 minutes) reperfusion. Kidneys were collected for EM. EM micrographs of proximal tubular cells from both cortex and outer stripe of outer medulla were selected for evaluation. Due to the orientation of mitochondria, proximal tubular cells from control animals usually displayed 10%–20% long (≥ 2 μm) filamentous mitochondria at the basal lateral side, whereas the perinuclear mitochondria were cross-sectioned and thus appeared “fragmented” (Figure 9A). After ischemia/reperfusion, however, many proximal tubular cells completely fragmented their mitochondria into small, punctate suborganelles (Figure9A). For quantification, we determined the percentage of cells that had completely lost their long mitochondria. As shown in Figure 9B, sham control had approximately 7% proximal tubular cells with fragmented mitochondria, which was increased to 42% during renal ischemia/reperfusion.
2D EM analysis of mitochondrial fragmentation in kidney tissues. C57BL/6 mice (male, ~8 weeks) were subjected to 30 minutes of bilateral renal ischemia followed by 15 minutes of reperfusion (Ischemia) or control sham operation (Ctrl). Kidneys were fixed in situ via vascular perfusion and processed for EM. (A) EM micrographs of a control and an ischemically injured proximal tubular cell. Scale bars: 1 μm. Asterisks indicate elongated (>2 μm) mitochondria. (B) Quantification of mitochondrial fragmentation. Mitochondrial length was measured in individual tubular cells to determine the percentage of cells that showed filamentous mitochondria less than 1% long (>2 μm). A total of 90 cells in control and 160 cells in ischemic kidneys from 4 animals were evaluated.
To confirm mitochondrial fragmentation in ischemically injured tubular cells, we performed 3D reconstruction of mitochondria from 100 serial section EM micrographs. The control tubular cell showed many filamentous mitochondria at the basolateral side in 2D EM (not shown); in the reconstruction, we purposely selected a perinuclear area where mitochondria appeared “fragmented” (Figure 10A). The 3D image showed clearly that these mitochondria were actually long and filamentous (Figure 10B). In sharp contrast, mitochondria in the ischemic cells were completely fragmented (Figure 10, C and D). These data indicate that mitochondrial fragmentation indeed occurs in vivo in renal tubular cells during ischemic injury.
3D image of mitochondria in control and ischemically injured tubular cells. C57BL/6 mice (male, ~8 weeks) were subjected to 30 minutes of bilateral renal ischemia followed by 15 minutes of reperfusion or control sham operation. Kidneys were fixed in situ via vascular perfusion and processed to collect 100 serial sections of a representative region at 45 nm/section for EM. EM micrographs of serial section no. 50 were shown for 2D image. For 3D image, EM images of the 100 serial sections were aligned for 3D reconstruction using the Reconstruct software. (A) 2D EM image of a control tubular cell. (B) 3D EM image of the same control cell as shown in A. (C) 2D EM image of an ischemically injured tubular cell. (D) 3D EM image of the same ischemic cell as shown in C. Note: The numbered mitochondria shown in A and B correspond respectively with those inC and D. In addition, some numbered mitochondria in 2D images are masked in the 3D images.
Amelioration of ischemic and cisplatin nephrotoxic renal injury and tubular apoptosis by mdivi-1, a pharmacological inhibitor of Drp1. Cassidy-Stone and colleagues have recently screened several chemical libraries and identified mdivi-1 as an efficacious inhibitor of mitochondrial division that operates by selectively inhibiting Drp1 (32). To determine the role of Drp1 and mitochondrial fragmentation in vivo, we examined the effects of mdivi-1 in the mouse model of ischemia/reperfusion (Figure 11). Ischemia/reperfusion induced a rapid loss of renal function, as indicated by increases in serum creatinine and blood urea nitrogen (BUN) (Figure 11, A and B), which was partially but significantly reduced in animals pretreated with mdivi-1 (Figure 11, A and B). Consistently, mdivi-1 ameliorated tubular damage in renal cortical and outer medulla tissues, as determined by histological examination (Figure 11, C and D). We further analyzed renal apoptosis by TUNEL assay. Cell counting showed that ischemia/reperfusion induced 24 apoptotic cells/mm2 renal tissue in mice that were pretreated with vehicle solution but only 11 in mdivi-1–pretreated mice (Figure 11E). Using EM, we confirmed that mdivi-1 partially suppressed ischemia-induced mitochondrial fragmentation in proximal tubular cells, from 43% to 32%.
Amelioration of ischemic renal injury and tubular apoptosis by mdivi-1, a pharmacological inhibitor of Drp1. C57BL/6 mice were injected with 50 mg/kg mdivi-1 or vehicle solution for 1 hour and then subjected to 30 minutes of bilateral renal ischemia followed by 48 hours of reperfusion. Control animals were subjected to sham operation without renal ischemia. Blood samples and renal tissues were collected for analysis. (A) Serum creatinine. (B) BUN. (C) Representative renal histology. Insets show higher magnification. Scale bars: 80 μm; 20 μm (insets). (D) Quantification of tubular damage. The percentage of damaged renal tubules was determined for each animal for histology scoring as described in Methods. (E) Tubular apoptosis. Renal tissues were subjected to TUNEL assay to enable counting positive cells to indicate apoptosis. Data are presented as mean ± SD;n ≥ 5. *P < 0.05, significantly different from sham control. #P < 0.05, significantly different from ischemic group injected with vehicle solution.
We further analyzed mitochondrial fragmentation in a mouse of cisplatin nephrotoxicity (15, 33). It was shown that cisplatin (30 mg/kg) treatment for 3 and 4 days led to mitochondrial fragmentation in 33% and 53% of proximal tubular cells, respectively. We then determined the effects of mdivi-1 on cisplatin-induced renal injury and nephrotoxicity. As shown in Supplemental Figure 2, both BUN and serum creatinine increases during cisplatin treatment were reduced by daily mdivi-1 injections. Consistently, mdivi-1 ameliorated renal tissue damage, especially in renal tubules. Cisplatin-induced apoptosis was also decreased by mdivi-1 (Supplemental Figure 2). Collectively, the results support an important role for Drp1-mediated mitochondrial fragmentation in the pathogenesis of both ischemic and nephrotoxic acute kidney injury.