Opa1 Overexpression Protects from Early-Onset Mpv17-/--Related Mouse Kidney Disease.

Moderate overexpression of Opa1, the master regulator of mitochondrial cristae morphology, significantly improved mitochondrial damage induced by drugs, surgical denervation, or oxidative phosphorylation (OXPHOS) defects due to specific impairment of a single mitochondrial respiratory chain complex. Here, we investigated the effectiveness of this approach in the Mpv17-/- mouse, characterized by profound, multisystem mitochondrial DNA (mtDNA) depletion. After the crossing with Opa1tg mice, we found a surprising anticipation of the severe, progressive focal segmental glomerulosclerosis, previously described in Mpv17-/- animals as a late-onset clinical feature (after 12-18 months of life). In contrast, Mpv17-/- animals from this new "mixed" strain died at 8-9 weeks after birth because of severe kidney failure However, Mpv17-/-::Opa1tg mice lived much longer than Mpv17-/- littermates and developed the kidney dysfunction much later. mtDNA content and OXPHOS activities were significantly higher in Mpv17-/-::Opa1tg than in Mpv17-/- kidneys and similar to those for wild-type (WT) littermates. Mitochondrial network and cristae ultrastructure were largely preserved in Mpv17-/-::Opa1tg versus Mpv17-/- kidney and isolated podocytes. Mechanistically, the protective effect of Opa1 overexpression in this model was mediated by a block in apoptosis due to the stabilization of the mitochondrial cristae. These results demonstrate that strategies aiming at increasing Opa1 expression or activity can be effective against mtDNA depletion syndromes.

Introduction

There is an unmet need for new therapies for mitochondrial diseases. Given the extreme genetic, biochemical, and clinical heterogeneity of mitochondrial disorders, an ideal therapeutic approach should be applicable to more than one single clinical entity. Strategies with potential general applicability to several, if not all, mitochondrial diseases have been proposed over the past 10 years by us and others. These include approaches aimed at (1) increasing mitochondrial biogenesis; (2) clearing dysfunctional organelles by stimulating mitophagy; and, although more controversial, (3) by-passing the mitochondrial respiratory chain (MRC) complex defects by using alternative oxidases or (4) reversing the mitochondrial unfolded stress response. An additional promising approach is based on the moderate overexpression of Opa1, which encodes a master regulator of shape and function of mitochondrial cristae. We previously showed that this approach improved the clinical phenotype of the Ndufs4−/− mouse, which lacks a small subunit of the P module of the MRC complex I (cI), as well as of the muscle-restricted Cox15, which lacks an enzyme essential for the biosynthesis of COX-specific heme A. We proposed that the stabilization of the defective MRC and supercomplexes (SCs), due to tighter and tidier cristae morphology, mediated the therapeutic effect. In addition, the same approach protected mice from a number of insults leading to altered cristae morphology and decreased oxidative phosphorylation (OXPHOS) proficiency, including denervation, lipopolysaccharide (LPS)-induced liver damage, and ischemia reperfusion in heart and brain. However, the applicability of the Opa1 overexpression therapy to other genetically determined mitochondrial diseases, such as those due to instability of mitochondrial DNA (mtDNA), has not been investigated. Here, we sought to determine whether Opa1 overexpression could improve the molecular phenotype of the Mpv17 knockout (KO) mouse associated with mtDNA depletion, which mimics the molecular features of a severe mitochondrial syndrome in humans due to MPV17 mutations. MPV17 encodes a small and highly hydrophobic protein embedded in the inner mitochondrial membrane (IMM), whose function is still unknown. In humans, mutations in MPV17 cause hepatocerebral mtDNA depletion syndrome (OMIM: 266810256810), including Navajo neurohepatopathy. This syndrome is characterized by early-onset and profound mtDNA depletion in liver, later complicated by neurological failure and, in some cases, peripheral analgesia with corneal scarring. Milder reduction of mtDNA content in skeletal muscle and, occasionally, multiple mtDNA deletions, in addition to mtDNA depletion in liver, have also been reported., The patients surviving the acute hepatic failure, dominated by hypoglycemic crises and cirrhotic evolution of the liver parenchyma, develop a progressive ataxia and neurological impairment. The spectrum of disorders due to mutations in MPV17 has progressively expanded to include (1) juvenile- and adult-onset axonal sensorimotor polyneuropathy without hepatocerebral involvement;, (2) neuropathy and leukoencephalopathy with multiple mtDNA deletions in skeletal muscle; and (3) adult-onset neurohepatopathy-plus syndrome with multiple deletions of mtDNA in muscle. The common feature of all these clinical syndromes is mtDNA instability, pointing to a role of MPV17 in mtDNA maintenance. However, the function of the Mpv17 gene product remains unclear, although several hypotheses have been proposed. For instance, MPV17 has been shown to form a non-selective cation channel, affecting mitochondrial membrane potential and reactive oxygen species (ROS) production., Other data suggested that the lack of MPV17 leads to a substantial decrease in deoxyguanosine triphosphate (dGTP) and deoxythymidine triphosphate (dTTP) in mouse liver mitochondria, thus facilitating the incorporation of riboGTP (rGTP), which, in turn, distorts mtDNA and may lead to block of its replication. Accordingly, high levels of rGTP were detected in mtDNA from Mpv17−/− livers. Other data showed that reduced MPV17 expression was associated with impaired deoxythymidine monophosphate (dTMP) synthesis without affecting de novo or salvage synthesis of dTMP, suggesting that MPV17 may be involved in maintaining dTMP levels in mitochondria through a still-uncharacterized pathway that may be involved in transporting dTMP or one of its precursors from the cytosol to mitochondria. In zebrafish, Mpv17 has been related to pyrimidine nucleotide metabolism via impairment of dihydroorotate dehydrogenase. All these observations have been reported in single publications and do need further confirmation. Finally, data on Saccharomyces cerevisiae support a role for Sym1, the yeast ortholog of Mpv17, in the homeostatic control of tricarboxylic acid cycle (TCA) cycle intermediates.

Like mutant MPV17 patients, Mpv17−/− mice are characterized by profound depletion of mtDNA in liver and moderate to mild depletion in other organs, such as skeletal muscle, brain, and kidney. However, despite the profound liver mtDNA depletion and contrary to the severe liver failure that characterizes the patients, Mpv17−/− mice survive normally up to 1 year of age except for an invariant graying of their fur occurring around the sixth month of life. However, after the first year of life, they develop a focal segmental glomerulosclerosis, leading to death by 1.5–2.0 years of age. A yet-unexplained finding concerning the Mpv17−/− mouse model is the shift of the kidney failure, which led to early death a few weeks after birth in the original strain,, to a much later and slowly progressive condition that was observed and documented in mouse Mpv17−/− mice after several generations.

Results

Early-Onset Kidney Disease in Mpv17−/− Mice

In order to generate Mpv17−/−::Opa1 double recombinant mice, we mated Mpv17+/−, and Opa1 mouse strains (Figure S1). The latter expresses Opa1 to levels slightly higher than those in the controls (CTRs) by targeted insertion of an extra copy of Opa1 into the Hprt locus on the X chromosome.,, Surprisingly, we noticed that the first Mpv17−/− animals (without Opa1 extra copies) started to die around 8–9 weeks of age (Figure 1A). This was reminiscent of the original phenotype described for this strain, i.e., an early death due to focal segmental glomerulosclerosis. We termed this strain the “new” Mpv17−/− strain, compared to our later onset, slowly progressive “old” Mpv17−/− strain. To search for possible modifier genes responsible for the phenotype anticipation, we first analyzed two genes previously proposed to modify the severity of the clinical phenotype in some mouse models of mitochondrial dysfunction, i.e., Prkdc, encoding for the catalytic subunit of the DNA-dependent protein kinase (DNA-PK), and Nnt, encoding the nicotinamide nucleotide transhydrogenase. However, the Prkdc mutation described in BALB/c and other inbred strains was not present in our C57BL/6 colonies, and the Nnt mutation, present in the pure C57BL/6J strain, did not segregate with the disease (data not shown). In order to identify potential genetic modifiers, we also carried out a RNA-sequencing (RNA-seq) experiment by analyzing the total RNA extracted from the kidneys of four Mpv17−/− mice, obtained after the crossing with the Opa1 mice, and showing early-onset kidney failure compared with the RNA of four animals belonging to our original, long-surviving, Mpv17−/− strain, sacrificed at 1 month of age, i.e., in a pre-symptomatic stage in both strains. The results are summarized in the volcano plot in Figure S2A. We found that 23 genes were significantly varied in the new strain versus the original strain. Among these, the only mitochondrial ones were Slc25a48, encoding an uncharacterized carrier that was upregulated in the new strain, and Upp2, encoding the uridine phosphorylase 2 that was downregulated in the new strain. Other gene products differentially expressed in the two strains did not localize in mitochondria, and the most differently expressed target was a non-coding RNA, termed AC158975.2, which was downregulated in the new KO strain. These results are also reported in detail in Figure S2B.

Figure 1

Effects of Opa1 Overexpression on Survival Probability and Kidney Disease of Mpv17−/− Mice

(A) Kaplan-Meier analysis of WT, Opa1, Mpv17−/−, and Mpv17−/−::Opa1 mice (n = 30). Kaplan-Meier survival probability. Significance was assessed by the log-rank test. Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively: ∗∗∗∗p < 0.0001 (Mpv17−/− vs. WT); ∗∗p = 0.0067 (Mpv17−/−::Opa1 vs. WT); §§§§p < 0.0001 (Mpv17−/−::Opa1 vs. Mpv17−/−). (B) Histological characterization of WT, Opa1, Mpv17−/−, and Mpv17−/−::Opa1 kidneys. Scale bars: 50 μm for all the light microscopy staining except for PAS, for which the scale bars represent 20 μm. TEM indicates foot process in kidney of the indicated genotypes; dashed lines indicate height and width of foot process, and the asterisk indicates the glomerular basement membrane (GBM). Scale bars: 1 μm.

Opa1 Moderate Overexpression Prolongs the Lifespan of Mpv17−/− Mice by Preventing Early-Onset Glomerulosclerosis

To investigate the presence of kidney failure in our strains, we measured proteinuria by using semiquantitative strips. All tested “new” Mpv17−/− animals displayed high proteinuria compared to the CTR littermates (n = 8 per group) (Figure S3A). Necroscopic analysis of 8-week-old mice revealed that the kidneys were smaller and weighted less in Mpv17−/− versus CTR littermates (Figures S3B and S3C), which included both Mpv17+/+ and Mpv17+/− individuals. Hematoxylin and eosin (H&E) staining showed the presence of many degenerating glomeruli and some protein casts in the tubuli of Mpv17−/− (Figure 1B) but neither in the tubuli of CTR nor of Opa1 mice (Figure 1B). Periodic acid-Schiff (PAS) staining revealed the presence of collagen-related glycoproteins in the glomeruli, with partial obliteration of the capillary loops (Figure 1B). Picrosirius-red staining confirmed the massive presence of fibrotic tissue (Figure 1B). Finally, transmission electron microscopy (TEM) showed profoundly altered podocytes in the glomeruli with loss of the foot processes (Figures 1B and S3D). These data confirmed the presence of an early-onset and severe focal segmental glomerulosclerosis in the “new” Mpv17−/− mouse strain. However, Opa1 overexpression had a striking protective effect so that Mpv17−/−:: Opa1 mice lived significantly longer than naive Mpv17−/− littermates (median lifespan: 553 vs. 75 days) (Figure 1A) and showed only minor histological alterations (Figure 1B).

Opa1 Overexpression Prevents OXPHOS Defects by Increasing mtDNA Content in Kidney

Profound mtDNA depletion in liver and more moderate in kidney is a hallmark of Mpv17−/− mice. We thus investigated whether Opa1 overexpression was able to increase mtDNA content in these tissues. We found that liver mtDNA content was doubled in Mpv17−/−::Opa1 mice compared to “new” Mpv17−/− animals (10.92 ± 2.33 vs. 4.31 ± 2.28; p < 0.01; n = 6 per genotype), but it was still markedly lower than in wild-type (WT) or Opa1 littermates (100 ± 33 and 106 ± 20, respectively; p < 0.0001; n = 6) (Figure S4A). Accordingly, the spectrophotometric activities of MRC complexes cI, cIII2, and cIV in liver homogenates were no different between Mpv17−/−::Opa1 and Mpv17−/−, and both were significantly lower than in the CTR and Opa1 groups (Figure S4B), despite Opa1 overexpression being confirmed in the liver of Mpv17−/−:: Opa1 versus Mpv17−/− animals (Figure S4C).

We then analyzed the kidneys, the most affected organ in this “new” Mpv17−/− model characterized by early-onset disease. First, we confirmed that Opa1 was overexpressed in Opa1 and Mpv17−/−::Opa1 compared to WT and Mpv17−/− by real-time PCR (Figure S5A) and western blot immunovisualization (Figures S5B and S5C). Notably, the amount of OPA1 was reduced in Mpv17−/− kidneys but was similar to that of Opa1 in Mpv17−/−::Opa1. Second, we measured mtDNA content. Similarly to that in liver, mtDNA content was doubled in Mpv17−/−:: Opa1 versus Mpv17−/− kidneys (69 ± 13.5 vs. 33 ± 21.1). The mtDNA amount of double recombinant Mpv17−/−:: Opa1 animals was, indeed, not significantly different from that of WT and Opa1 mice (100.0 ± 13.0 and 90.0 ± 15.3, respectively) (Figure 2A).

Figure 2

Effects of Opa1 Overexpression on mtDNA Content and Respiratory Chain of Mpv17−/−

(A) mtDNA copy number in WT, Opa1, Mpv17−/−, and Mpv17−/−::Opa1 kidneys (5 mice per genotype). Data are shown as mean ± SD. Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by one-way ANOVA with Tukey’s post hoc multiple comparison test: ∗∗∗∗p < 0.0001 (Mpv17−/− vs. WT); ∗p = 0.01 (Mpv17−/−::Opa1 vs. WT); §p = 0.0137 (Mpv17−/−::Opa1 vs. Mpv17−/−). (B) MRC activities normalized to citrate synthase (CS) in kidney homogenates of WT, Opa1, Mpv17−/−, and Mpv17−/−::Opa1 mice (n = 4 mice per genotype). Data are shown as mean ± SD. Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by one-way ANOVA with Tukey’s post hoc multiple comparison test. For cI: ∗∗∗∗p < 0.0001 (Mpv17−/− vs. WT); ∗∗∗p = 0.009 (Mpv17−/−::Opa1 vs. WT); §p = 0.04515 (Mpv17−/−::Opa1 vs. Mpv17−/−). For cII: ∗∗p = 0.0095 (Mpv17−/− vs. WT); §p = 0.0479 (Mpv17−/−::Opa1 vs. Mpv17−/−). For cIII: ∗∗∗p = 0.0005 (Mpv17−/− vs. WT); ∗∗∗p = 0.0002 (Mpv17−/−::Opa1 vs. WT). For cIV: ∗p = 0.0121 (Mpv17−/− vs. WT). (C) High-resolution respirometry on isolated kidney mitochondria from WT, Opa1, Mpv17−/−, and Mpv17−/−::Opa1 mice (n = 4 mice per genotype). Data are shown as mean ± SD. Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by one-way ANOVA with Tukey’s post hoc multiple comparison test. For cI: ∗∗p = 0.0032 (Mpv17−/− vs. WT); §§p = 0.0071 (Mpv17−/−::Opa1 vs. Mpv17−/−). For uncoupled: ∗p = 0.0402 (Mpv17−/− vs. WT); §p = 0.0236 (Mpv17−/−::Opa1 vs. Mpv17−/−). (D) Western-blot immunovisualization of MRC complex subunits in kidney homogenates from the indicated genotypes. (E) Densitometric analysis of the western blot immunovisualization shown in (D) (n = 3 WT; n = 3 Opa1; and n = 4 Mpv17−/−). Data are shown as mean ± SD. Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by one-way ANOVA with Tukey’s post hoc multiple comparison test. For cI: ∗∗∗p = 0.0007 (Mpv17−/− vs. WT); §§p = 0.057 (Mpv17−/−::Opa1 vs. Mpv17−/−). For cIV: §p = 0.04 (Mpv17−/−::Opa1 vs. Mpv17−/−). For COXVB: ∗∗∗p = 0.0005 (Mpv17−/− vs. WT); §§p = 0.003 (Mpv17−/−::Opa1 vs. Mpv17−/−). For cV: §p = 0.0146 (Mpv17−/−::Opa1 vs. Mpv17−/−). (F) 1D-BNGE analysis of DDM-solubilized kidney mitochondria from the different genotypes. (G) Densitometric analysis of the western blot immunovisualization shown in (F); n = 4 WT, 4 Opa1, 7 Mpv17−/−, and 7 Mpv17−/−::Opa1. Data are shown as mean ± SD. Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by two-way ANOVA with Tukey’s post hoc multiple comparison test. For cI: ∗p = 0.0218 (Mpv17−/− vs. WT); §§p = 0.0045 (Mpv17−/−::Opa1 vs. Mpv17−/−). For cIV: ∗∗∗p = 0.0002 (Mpv17−/− vs. WT); §p = 0.0432 (Mpv17−/−::Opa1 vs. Mpv17−/−). (H) Histochemical staining for COX in kidneys of WT, Opa1, Mpv17−/−, and Mpv17−/−::Opa1 mice. Scale bars: 10 μm.

We then analyzed the activities of the MRC complexes in kidney homogenates from animals of the different genotypes. cI- and cIV-specific activities normalized to citrate synthase (CS) activity were significantly higher in Mpv17−/−::Opa1 versus Mpv17−/− samples, although they were still significantly lower than in WT and Opa1 animals (Figure 2B). No differences were detected between Mpv17−/−::Opa1 versus Mpv17−/− in cIII2 activity, which was in both strains lower than in CTR animals. Finally, cII, which was significantly higher in Mpv17−/− kidneys, was comparable to CTR and Opa1 animals in Mpv17−/−::Opa1 mice. In addition, cI-dependent (rotenone-sensitive) and uncoupled oxygen consumptions, which were both reduced in Mpv17−/− mice, were normal in Mpv17−/−::Opa1 kidney mitochondria (Figure 2C).

Given the aforementioned results, we analyzed the protein levels of several MRC subunits. In agreement with the spectrophotometric activities, cI subunit NDUFB8, cIV subunits mt-CO2 and COX5B, and cV subunit ATP5A were increased in Mpv17−/−::Opa1 versus Mpv17−/− mice (Figures 2D and 2E). No significant changes were detected in cII subunit SDHA and cIII2 subunit UQCR2. Finally, we analyzed respiratory holocomplexes in dodecyl-maltoside (DDM)-solubilized kidney mitochondria by first-dimension blue native gel electrophoresis (1D-BNGE). The amounts of both cI and cIV, which were decreased in Mpv17−/− mice, became comparable to those of CTRs in Mpv17−/−::Opa1 (Figures 2F and 2G). However, western blot analyses of 1D-BNGE in digitonin-treated isolated kidney mitochondria did not show any significant changes in the amount of cI- and cIII2-containing SCs, suggesting that the amelioration of the phenotype in the Mpv17−/−::Opa1 mice was not related to an increase in the stability of MRC SCs (Figures S5D and S5E). These data were further confirmed by in-gel activity for cI and cI-containing SCs (Figure S5F). Histochemical analysis of kidney sections revealed a decrease in the intensity of COX-specific staining in virtually all glomeruli and uneven distribution of COX staining in the tubuli (Figure 2H) of Mpv17−/−, as previously reported.

Decreased mtDNA content and expression were previously reported to correlate to mitochondrial coenzyme Q (CoQ) levels. However, we did not find any difference in CoQ levels in homogenates from Mpv17−/− and CTR kidneys (Figure S5G).

We then used laser-captured microdissected glomeruli and tubuli from animals of the different genotypes. We analyzed both COX-positive and COX-negative tubuli from Mpv17−/− and Mpv17−/−::Opa1 animals. By quantitative PCR, we found that mtDNA content was particularly low in Mpv17−/− glomeruli and COX-negative tubuli (18% and 29% of the CTRs, respectively) (Figures S6A and S6B, respectively). In contrast, no differences were detected in COX-positive tubuli compared to CTRs (Figure S5C). Notably, the mtDNA content in both glomeruli and tubuli was markedly increased in Mpv17−/−::Opa1 versus Mpv17−/− mice, although, in both cases, it remained significantly lower than in WT and Opa1 animals (60% and 68% respectively).

These results indicate that Opa1 overexpression is preventing mtDNA depletion and mitochondria dysfunction in specific regions of Mpv17−/− mouse kidney.

Opa1 Overexpression Prevents Mitochondrial Fragmentation and Preserves Cristae Ultrastructure in Mpv17−/− Mice

TEM analysis of the kidneys at low magnification showed a significant reduction in the number of cells in the glomeruli of Mpv17−/− kidney sections compared to WT, Opa1, and Mpv17−/−::Opa1 (Figures 3A and 3B), with accumulation of abnormal mitochondria with altered cristae in glomerular cells (Figures 3C and 3D). In addition, a mixed population of normal and altered mitochondria was present in different areas of the tubuli of Mpv17−/− mice (Figure S7).

Figure 3

TEM Analysis of Mpv17−/− Kidneys

(A) Representative low-magnification TEM micrographs from mouse kidneys of the indicated genotypes showing the glomeruli. Scale bars: 200 nm. (B) Quantification of the number of nuclei per glomerulus in WT, Mpv17−/−, and Mpv17−/−::Opa1 kidneys. Data for number of nuclei per glomerulus are plotted as mean ± SD (n = 3). Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by one-way ANOVA followed by Tukey’s multiple comparison test. ∗∗p = 0.0021 (Mpv17−/− vs. WT and Mpv17−/−::Opa1 vs. WT); §§§§p < 0.0001 (Mpv17−/−::Opa1 vs. Mpv17−/−). (C) Representative TEM images at high magnification showing the accumulation of damaged mitochondria in Mpv17−/− glomeruli. (D) Quantification of the number of damaged mitochondria in glomeruli of the indicated genotypes. Data plotted for number of damaged mitochondria per cell are expressed as mean ± SD (n = 3). Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by one-way ANOVA followed by Tukey’s multiple comparison test. ∗∗∗p = 0.0021; ∗p = 0.0403 (Mpv17−/− vs. WT and Mpv17−/−::Opa1 vs. WT); §§§p = 0.0001; §p = 0.0403 (Mpv17−/−::Opa1 vs. Mpv17−/−).

Given the reduced number of cells in Mpv17−/− glomeruli, we carried out additional experiments in isolated primary podocytes. Since no differences were observed between WT and Opa1 mice, we did not produce Opa1 podocytes.

First, we assessed the purity of the culture by immunostaining with an anti-NPHS2 antibody, which recognizes a specific marker for podocytes, and used an anti-cytokeratin antibody to exclude the presence of epithelial cells (Figure S8). We then quantified mitochondria length and found a significant decrease in the average size of Mpv17−/− mitochondria compared to WT, while it was normal in Mpv17−/−::Opa1 podocytes (Figures 4A and 4B). In keeping with these data, we found that the amount of MFN2, a protein of the outer mitochondrial membrane playing a key role in mitochondrial fusion, was markedly decreased in Mpv17−/− kidney homogenates but was comparable to WT in Mpv17−/−::Opa1 kidneys (Figures 4C and 4D). Next, we quantified mitochondrial ultrastructure in isolated podocytes. A significantly reduced number of cristae and increased cristae junction (CJ) width were found in Mpv17−/− versus WT cells (Figures 5A–5C). In contrast, both number and CJ width were similar to WT in Mpv17−/−::Opa1 cells. Similar to what was observed for MFN2, MIC19, a component of the mitochondrial contact site and cristae organizing system (MICOS) complex involved in cristae shaping, was markedly decreased in Mpv17−/− kidney mitochondria but was comparable to WT in Mpv17−/−::Opa1 kidneys (Figures 5D and 5E). In addition, overexposure of the 1D-BNGE membranes revealed accumulation of the free ATP-synthase F1 particle in Mpv17−/− cells, which was not observed in Mpv17−/−::Opa1 (Figure 5F). These results suggest that Opa1 overexpression preserves mitochondrial ultrastructure in the absence of Mpv17, possibly by increasing mitochondria fusion and leading to the stabilization of the MICOS complex and ATP synthase structure.

Figure 4

Analysis of Isolated Podocytes

(A) Confocal micrographs indicating mitochondrial morphology (TOM20 in red) and nuclei (DAPI in blue) of podocytes from described genotypes. Maximum intensity projection of z stacks is indicated. Scale bars: 10 μm. (B) Mitochondrial length analysis from (A). Data are plotted as means ± SD (n = 3). Total mitochondria (particles) per condition: 5,663 for WT, 5,107 for Mpv17−/−, and 4,360 for Mpv17−/−::Opa1+/+. Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by unpaired t test. ∗∗∗p = 0.0003 (Mpv17−/− vs. WT); §p = 0.0001 (Mpv17−/−::Opa1 vs. Mpv17−/−). (C) Western blot immunovisualization with an anti-MFN2 antibody in kidney homogenates. An anti-VDAC antibody was used as a loading control. (D) Densitometric analysis of the western blot in (C). Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by two-way ANOVA with Tukey’s post hoc multiple comparison test. ∗p = 0.0218 (Mpv17−/− vs. WT); §§p = 0.0045 (Mpv17−/−::Opa1 vs. Mpv17−/−).

Figure 5

Analysis of Mitochondrial Cristae Morphology

(A) Representative TEM images of isolated podocytes for the indicated genotypes. Scale bars: 200 nm. (B) Quantification of the number of cristae per mitochondrial area from TEM pictures (n = 3 per genotype). Data plotted are mean ± SD (n = 3) of CJ maximum width. Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by one-way ANOVA with Tukey’s post hoc multiple comparison test. ∗∗∗∗p < 0.0001 (Mpv17−/− vs. WT); §§p = 0.0068 (Mpv17−/−::Opa1 versus Mpv17−/−). (C) Quantification of CJ maximum width from TEM pictures (n = 3). Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by one-way ANOVA with Tukey’s post hoc multiple comparison test. ∗p = 0.0182 (Mpv17−/− vs. WT); §p = 0.0109 (Mpv17−/−::Opa1 vs. Mpv17−/−). (D) Western blot immunovisualization using an anti MFN2 antibody. VDAC was used as a loading control. (E) Densitometric analysis of the western blot in (D). Data are shown as mean ± SD. Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by unpaired two-tailed Student’s t test. ∗∗p = 0.0058 (Mpv17−/− vs. WT); §p = 0.0037 (Mpv17−/−::Opa1 vs. Mpv17−/−). (F) 1D-BNGE using anti-cI (NDUFB8) and anti-cV (ATP5A) antibodies. cII subunit SDHA was used as a loading control. Note the accumulation of ATP synthase F1 subunits in Mpv17−/− mitochondria.

Opa1 Overexpression Prevents Mpv17-Dependent Glomerulosclerosis by Blocking Apoptosis

We then investigated whether the correction of mitochondrial ultrastructure can lead to the rescue of the glomerulosclerosis and lethality of Mpv17−/− mice. As mentioned earlier, a significant reduction in the number of cells in the glomeruli was observed at low magnification in TEM analysis of Mpv17−/− kidney sections compared to the WT, Opa1, and Mpv17−/−::Opa1 (Figures 3A and 3B). This result, which suggests increased cell death, prompted us to investigate whether Opa1 overexpression blocked apoptosis in Mpv17−/− kidneys. Cleaved caspase-3 immunohistochemistry showed many positive cells, especially in the glomeruli of “new” Mpv17−/− kidneys. In contrast, only a few positive cells were detected in Mpv17−/−::Opa1 samples (Figure 6A). These results were further confirmed by quantification of cleaved caspase-3 in isolated podocytes (Figure 6B).

Figure 6

Caspase-3 Staining in Kidney Sections and Isolated Podocytes

(A) Immunohistochemical staining with anti-caspase-3 antibody. Scale bars: 50 μm (upper row) and 20 μm (lower row). (B) Confocal micrographs from primary podocytes derived from described genotypes showing cytoplasm staining (GAPDH in green), apoptotic cells (cleaved caspase-3 in red) and nuclei (DAPI in blue). Maximum intensity projection of z stacks is indicated. Scale bars: 50 μm. On the right panel, quantification of caspase-3-positive cells. Data are presented as mean ± SD. Symbols ∗ and § represent the significance levels versus WT and Mpv17−/−, respectively, calculated by one-way ANOVA with Tukey’s post hoc multiple comparison test. ∗∗∗∗p < 0.0001 (Mpv17−/− vs. WT, and Mpv17−/−::Opa1 vs. WT); §§§§p < 0.0001 (Mpv17−/−::Opa1 versus Mpv17−/−).

During apoptosis, the proapoptotic factor BAX translocates from cytoplasm to the mitochondrial outer membrane, promoting the release of cytochrome c, which, in turn, interacts with APAF1, activating the caspase cascade (Figure 7A). Caspase-3 activation also causes the cleavage of the pro-autophagic factor Beclin-1 in three fragments of 50, 37 and 35 kDa. Cleaved BECLIN-1 cannot activate autophagy, and the 50-kDa N-terminal fragment re-localizes to both nucleus and mitochondria, enhancing the apoptotic cascade.28, 29, 30 We found that BAX expression was increased in Mpv17−/− mice, whereas its levels was comparable to WT in Mpv17−/−::Opa1 littermates (Figures 7B and 7C). In addition, although several cleaved forms of BECLIN-1 were present in WT kidney homogenates, the three previously described, caspase-3 dependent, BECLIN-1 fragments were clearly predominant in Mpv17−/− kidney homogenates. In contrast, the same pattern as in the WT was detected in the Mpv17−/−::Opa1 samples (Figure 7D).

Figure 7

Interplay between Apoptosis and Autophagy through Regulation of BECLIN-1

(A) Schematic representation of the pathway connecting apoptosis and autophagy via modulation of BECLIN-1 processing. Details are given in the text. (B) Western blot immunovisualization of kidney homogenates using an anti-BAX antibody. (C) Densitometric analysis of the western blot in (B). Data are presented as mean ± SD (n = 4 WT; n = 8 KOS). Symbols ∗ and § represent the significance levels of Mpv17−/−::Opa1 versus WT and Mpv17−/−, respectively, calculated by one-way ANOVA with Tukey’s post hoc multiple comparison test. ∗p = 0.0176 (Mpv17−/− versus WT); §§§p = 0.0008 (Mpv17−/−::Opa1versus Mpv17−/−). (D) Western blot analysis of kidney homogenates using an anti-BECLIN-1 antibody. Note the different pattern of BECLIN-1 processing in Mpv17−/− compared to WT and Mpv17−/−::Opa1 kidneys.

Discussion

Alterations in mitochondrial ultrastructure have an impact on mtDNA stability, as demonstrated, for instance, by the presence of multiple deletions in patients carrying mutations in OPA1, MFN2, and ATAD3. We showed that moderate overexpression of Opa1 was able to rescue the mtDNA depletion in the “new” genetic background Mpv17−/− kidneys, resulting in a markedly prolonged median lifespan from 75 to 553 days. However, the maximum lifespan was similar in Mpv17−/− and Mpv17−/−::Opa1 mice, suggesting that Opa1 overexpression slowed down but did not completely stop the disease progression, as confirmed by the histological data. One hypothesis that could explain the reason why about 25% of the “new” Mpv17−/− model survived as much as the Mpv17−/−::Opa1 littermates is that not all of them would have inherited the genetic modifier(s) or combination of polymorphisms leading to the more severe phenotype.

Although the MPV17 function remains unclear, mutations in its encoding gene constitute a prominent cause of mtDNA instability syndromes in humans. We previously found that mitochondrial cristae ultrastructure was disrupted in the liver of Mpv17−/− mice and that mtDNA was decreased to less than 10% of the CTR littermates. However, this was not associated with an overt liver dysfunction unless challenged with a ketogenic high-fat diet, which led to cirrhosis and mouse liver failure. We also reported that Mpv17−/− mice developed a late-onset focal segmental glomerulosclerosis, leading to death at around 1.5–2.0 years of age. The early-onset glomerulosclerosis observed after crossing with Opa1 mice was unexpected and remarkably similar to the phenotype originally described, characterized by focal segmental glomerulosclerosis, accumulation of protein casts in the tubuli, and early death. Also the strain we originally reported (a pure C57/6N) developed the same phenotype, but much later, after 12–18 months of life. One question remains unanswered; that is, why 90% mtDNA depletion in liver has little consequence on liver function, whereas partial mtDNA depletion in kidney leads to a very severe, fatal phenotype. One possible explanation is that the kidney is particularly susceptible to oxidative stress. Although we did not measure ROS production in our Mpv17−/− mice, this was previously reported in the same mouse model. However, the sudden shift in the phenotype prompted us to hypothesize the presence of causative variants in one of the supposed genetic modifiers of the Mpv17−/− mitochondrial phenotypes. The first candidate gene was Prkdc, encoding the catalytic subunit of the DNA-PK, which forms a heterotrimer with the Ku p70/YRCC6-p86/XRCC5 proteins in DNA double-strand break repair and recombination. A spontaneous mutation R2140C in Prkdc was found in some mouse strains, leading to profound decrease in the amount of the protein, which was proposed to modify the severity of Mpv17 kidney disease. However, our mice did not carry the Prdkc hypomorphic mutant allele.

A second potential modifier of the Mpv17−/− phenotype was the nicotinamide nucleotide transhydrogenase (Nnt), which regulates NAD(P)H levels by coupling hydride transfer between NAD(H) and NADP(+) to proton translocation across the IMM, thus regulating redox poise. C57BL/6J mice carry two spontaneous mutations in the Nnt gene: a missense (methionine-to-threonine) mutation in the mitochondrial leader sequence and an in-frame deletion that removes exons 7–11, resulting in the expression of a truncated, non-functional protein. The mutant Nnt allele was previously proposed to impact on the severity of mtDNA mutations. However, although we detected both the WT and truncated forms of the Nnt gene in the individuals of our colony, these variants did not segregate with the disease. These results conclusively suggest that other modifiers, or combinations of polymorphisms, may be responsible for the shift in the phenotype. RNA-seq analysis showed a few quantitative differences between the “old” KO and the “new” KO strains, including only two mitochondrial targeted gene products, an unknown transporter (Slc25a48) and uridine phosphorylase 2 (Upp2). These observations, as well as the few hundred SNP variants identified in the same RNA-seq experiment, warrant validation and further investigation.

Mitochondrial cristae are controlled by a molecular network formed by the MICOS complex, the F1Fo-ATP synthase, the IMM pleiotropic OPA1 protein, and the non-bilayer-forming phospholipids cardiolipin and phosphatidylethanolamine. OPA1 has been shown to be epistatic to MICOS, and its protective effect against mitochondrial dysfunction is mediated by the stabilization of cristae morphology as well as ATP synthase. We found that Opa1 overexpression was able to correct mitochondrial ultrastructure and increase mtDNA content in the kidney of “new” Mpv17−/− mice. mtDNA is bound to the IMM, and altered cristae morphology has an impact on mtDNA maintenance.,39, 40, 41 Stabilization of the cristae morphology by increased OPA1 expression can, thus, lead to increased mtDNA levels. Interestingly, OPA1 protein levels were reduced in the “new” Mpv17−/− mice, possibly as a consequence of cristae alterations. In keeping with this, Opa1 mRNA levels were normal in Mpv17−/− mice. However, while the effect on liver mtDNA was modest, the mtDNA increase in kidney was sufficient to preserve the activities of MRC complexes to WT levels. In addition, our data show that the correction of mitochondrial ultrastructure also stabilized respiratory complexes and SCs. The only exception to this is cIII, whose activity was similar in Mpv17−/−::Opa1 and Mpv17−/− kidneys. One possible explanation is that cIII is more efficiently incorporated in respiratory SCs and/or engaged in respiration, which is, indeed, normal in oxygen consumption studies and in-gel activities (Figure S4), thus contributing to the therapeutic effect. However, the C57BL/6 background makes it particularly difficult to draw conclusions, as SCs have been reported to be almost absent in this background, particularly in kidneys. However, we acknowledge that this is a very controversial issue. Anyway, it should also be noted that Opa1 overexpression led to ATP synthase stabilization. The 1D-BNGE western blot data, indeed, show the accumulation of free F1 ATP-synthase particles in Mpv17−/−, which were absent in Mpv17−/−::Opa1 kidneys, confirming that Opa1 overexpression contributes to the stabilization of OXPHOS. These results are in agreement with recent published data. The normalization of mitochondrial ultrastructure explains the reduced apoptosis in Mpv17−/−::Opa1 versus Mpv17−/− kidneys and podocytes. This process may be boosted by the block of BECLIN-1 processing.

Overall, our results further support the deleterious role of cristae disruption in the pathogenic mechanisms leading to organ failure in mitochondrial disorders and the correcting effects of moderate overexpression of Opa1, which can be exploited for future therapeutic approaches.

Materials and Methods

RNA-Seq Analysis

RNA was extracted by using a standard TRIzol-based protocol from kidney tissue. RNA-seq was carried out by the Next Generation Sequencing (NGS) facility, Department of Pathology, University of Cambridge, Cambridge, UK. A detailed description of the methodology followed for the collection and analysis of data is reported in the Supplemental Materials and Methods.

Antibodies

Antibodies used were: mouse monoclonal anti-GAPDH (1:1,000; ab8245; Abcam), rabbit monoclonal anti-TOM20 Alexa Fluor 647 (1:500; ab209606; Abcam), Alexa Fluor 488 Goat anti-mouse (1:300; A11004; Invitrogen), Alexa Fluor 647 Goat anti-rabbit (1:300; A27040; Invitrogen), rabbit polyclonal anti-NPHS2 (1:300; ab50339; Abcam), mouse monoclonal anti-CYTOKERATIN (1:300; #ab86734; Abcam), and anti-cleaved caspase-3 (1:200; Cell Signaling, #9661).

Chemicals

All chemicals were from Merck-Sigma Aldrich.

Animals

All animal experiments were carried out in accordance with the UK Animals (Scientific Procedures) Act 1986 (PPL: P6C20975A) and EU Directive 2010/63/EU. Mpv17+/− and Opa1 original strains were on C57BL/6N and C57BL/6J backgrounds, respectively. Mpv17+/−::Opa1 mice were mated to generate Mpv17−/−::Opa1 animals. Mpv17+/−::Opa1+/+ and Mpv17+/+::Opa1+/+ were used as controls, as no differences were observed between Mpv17+/+ and Mpv17+/− mice. Since the Opa1 transgene is located on the X chromosome, only homozygous females and hemizygous males were considered in this study. The animals were maintained in a temperature- and humidity-controlled animal-care facility with a 12-h:12-h light:dark cycle and free access to water and food and were sacrificed by cervical dislocation.

Isolation of Primary Podocytes

Kidneys were removed from 7- to 10-day-old mice, washed in Ca2- and Mg2-free Hank’s medium (Thermo Fisher Scientific), and treated with 1.5 mg/mL collagenase type I (Sigma-Aldrich) for 1 min at 37°C. The reaction was stopped by growth medium consisting of Dulbecco’s modified Eagle’s medium/F-12 (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS; GIBCO), 1× insulin-transferrin-selenium-ethanolamine (Thermo Fisher Scientific), 1 μM hydrocortisone (Sigma Aldrich), 1× penicillin-streptomycin, and 2 mM L-glutamine (Sigma-Aldrich).

Glomeruli were isolated by sieving through the 100-μm mesh (Fisherbrand) and seeded in 100-mm dishes precoated overnight at 4°C with collagen type IV (Sigma-Aldrich). Cells were kept at 37°C in a 5% CO2 atmosphere for 7 days. Additional sieving was performed through the 40-μm mesh resuspending the cells with trypsin-EDTA. Cells were plated on dishes or coverslips treated with poly-L-lysine (Sigma-Aldrich) and, the next day, fixed for the respective technique. Cell characterization was performed by immunofluorescence using the podocyte (NPHS2) and epithelial cell (cytokeratin) markers.

Immunofluorescence and Imaging

Cells were fixed with 3.7% formaldehyde in F-12 complete media and permeabilized with 0.1% in PBS with 0.1% Triton X-100 and 0.05% sodium deoxycholate before staining with primary and secondary antibodies in blocking solution (5% goat serum). Coverslips were mounted with ProLong Glass with or without DAPI (Invitrogen).

Images were acquired randomly for each slide using a Dragonfly Spinning Disk imaging system (Andor Technologies), composed with a Nikon Ti-E microscope, a Nikon 100× TIRF ApoPlan or 20× ApoPlan objective, and an Andor iXon EMCCD or Zyla sCMOS camera. The z stacks were acquired using Fusion software (Andor Technologies), and the 3D images were analyzed and exported using Imaris software (Bitplane). For mitochondria length measurements of 3, preparation of podocytes (n = 3) per group was used, and several regions of interest (ROIs) were drawn in the peripheric regions of the mitochondrial network using ImageJ. Automatic threshold was used, and “particles” (mitochondria) bigger than 0.05 μm2 were analyzed. Cleaved caspase-3 analysis was performed by manually counting the number of DAPI- and GAPDH-positive cells and caspase-3-positive cells (n = 3).

TEM

For TEM analysis, mice were subjected to perfusion with the fixative solution (2.5% glutaraldehyde [Sigma-Aldrich], 2% paraformaldehyde [MP Biomedicals] in 0.1 M PB buffer [Sigma-Aldrich]), 1-mm3 pieces of kidney were kept in fixative solution overnight, and samples were washed several times in 0.1 M phosphate buffer for the following days. For fibroblasts, 1 × 106 fibroblasts were fixed in the fixative solution for 1 h at room temperature (RT) in agitation. Cells were scraped, centrifuged (2,500 rpm for 10 min at 4°C), and washed 3 times with 0.1 M PB buffer. Pelleted cells were sent to the Electron Microscopy Platform (Scientific and Technological Centers, University of Barcelona, Barcelona, Spain) for sample preparation. Regions containing the glomeruli were selected, and ultrathin (55-nm) sections were cut, for kidney and cells, and mounted on 200-mesh copper grids with carbon-supported film. Image acquisition (Digital Micrograph, v1.85.1535, Gatan, Pleasanton, CA, USA) was performed with a transmission electron microscope (FEI Tecnai G2 Spirit Twin) coupled with an charge-coupled device (CCD) camera (Orius-SC200B, Gatan, Pleasanton, CA, USA). Quantification of number and width of cristae was performed manually using ImageJ; 10 cells were analyzed for each genotype (n = 3), number of cristae was counted per visible mitochondria area, and maximum CJ width was determined by randomly measuring 3–4 CJs per mitochondria. Foot process was measured by the ratio between maximum height and width touching the glomerular basement membrane. The mean of this ratio (foot process aspect ratio) was calculated from 21 randomly chosen segments per group (n = 3).

Immunoblotting

Mouse tissues were homogenized in 10 vol of 10 mM potassium phosphate buffer (pH 7.4). Mitochondrial-enriched fractions were collected after centrifugation at 800 × g for 10 min in the presence of protease inhibitors and were frozen and thawed three times in liquid nitrogen. Protein concentration was determined by the Lowry method. Aliquots, 30 μg each, were run through a 12% SDS-PAGE and electroblotted onto a polyvinylidene fluoride (PVDF) membrane, which was then immunodecorated with different antibodies.

For BNGE analysis, brain mitochondria, isolated as previously described, were resuspended in 1.5 M aminocaproic acid, 50 mM Bis-Tris-HCl (pH 7), and 4 mg DDM per milligram of proteins and incubated for 5 min on ice before centrifuging at 20,000 g at 4°C. 5% Coomassie G250 was added to the supernatant. 100 μg was separated by 4%–12% gradient BNGE and electroblotted on nitrocellulose membranes for immunodetection.

Spectrophotometric Analysis of Respiratory Chain Activities

Tissues were snap-frozen in liquid nitrogen and homogenized in 10 mM phosphate buffer (pH 7.4). The spectrophotometric activity of cI and CS was measured as previously described.

Morphological Analysis

For histological and immunohistochemical analyses, animals were fixed in 10% neutral buffered formalin (NBF) for a few days at RT and then included in paraffin wax. 4-μm-thick sections were used for analysis. H&E staining was performed by standard methods. Immunohistochemistry was performed using the Novolink Polymer Detection System and specific antibodies against the indicated proteins.

Statistical Analysis

All numerical data are expressed as mean ± SD. One- or two-way ANOVAs with Tukey correction were used for multiple comparisons (see figure legends for details); Kaplan-Meier distribution and log-rank test were used for survival probability analysis. Differences were considered statistically significant when p < 0.05.

Author Contributions

C.V., L.S., M.L.-S., and M.Z. designed the study. M.L.-S. performed the in vivo experiments, the biochemical assays, the oxygraphy experiments, and the molecular analysis. R.C. performed the histological and histochemical analyses. C.B. and M.L.-S. performed the experiments on primary podocytes. C.B. and A.Y. performed the TEM experiments. G.B.-C. performed the CoQ pool analysis. M.L.-S., C.B., and R.C. analyzed the data. C.V., M.L.-S., C.B., and M.Z. wrote the manuscript.

Conflicts of Interest

The authors declare no competing interests.