The contralateral kidney presents with impaired mitochondrial functions and disrupted redox homeostasis after 14 days of unilateral ureteral obstruction in mice

In unilateral ureteral obstruction (UUO), both oxidative stress and mitochondrial dysfunction are related to cell death. The aim of this study has been to characterize profiles of enzyme antioxidant activities and mitochondrial functioning of the contralateral (CL) compared to UUO and Sham (false-operated) kidneys of Balb/c mice. Kidneys were resected 14 days after obstruction for immunohistochemical and cortical mitochondrial functioning assays. Antioxidant enzymes activities were investigated in mitochondria and cytosol. Oxygen consumption (QO2) and formation of O2 reactive species (ROS) were assessed with pyruvate plus malate or succinate as the respiratory substrates. QO2 decreased in CL and UUO in all states using substrates for complex II, whereas it was affected only in UUO when substrates for complex I were used. Progressive decrease in mitochondrial ROS formation–in the forward and reverse pathway at complex I–correlates well with the inhibition of QO2 and, therefore, with decreased electron transfer at the level of complexes upstream of cytochrome c oxidase. CL and UUO transmembrane potential responses to ADP were impaired with succinate. Intense Ca2+-induced swelling was elicited in CL and UUO mitochondria. Important and selective differences exist in CL antioxidant enzymes with respect to either Sham or UUO kidneys: CL kidneys had increased mitochondrial glutathione peroxidase and cytosolic catalase activities, indicative of compensatory responses in the face of an early altered ROS homeostasis (as detected by 4-hydroxynonenal), and of a significant tendency to apoptosis. In CL and UUO, upregulation of nuclear (erythroid-derived 2)-like 2 transcription factor (Nrf2), as well as of cytoplasmic and nuclear Kelch-like ECH-associated protein 1 (Keap1) in opposition to decreased heme oxygenase-1 (HO-1), suggest impairment of the Nrf2/Keap1/HO-1 system. It is concluded that chronic obstruction impairs mitochondrial function in CL and UUO, preferentially affecting complex II.


Introduction
Mitochondrial dysfunction participates in the initiation and progression of acute kidney injury (AKI) to chronic kidney disease (CKD) [1,2]. Experimental unilateral ureteral obstruction (UUO) provides an excellent model to study mitochondrial dysfunction since it reflects the progression of AKI to CKD [2]. In this model the contralateral non-obstructed kidney (CL), which is frequently used as an internal control [3], starts to grow within 20-30 days of ureteral ligation to compensate for the unilateral loss of renal function. The changes seen in the CL kidney may reflect structural adaptations occurring due to systemic angiotensin II perfusion [4], which could explain the long-term evolution of kidneys from medically complex living donors or that of a solitary kidney in children with congenital anomalies of the urinary tract [5,6].
Oxidative stress participates in the pathogenesis of UUO [7]. The balance between survival and cell death are controlled by antioxidant systems. Among them, the system nuclear factor (erythroid-derived 2)-like 2 transcription factor (Nrf2)/Kelch-like ECH-associated protein 1 (Keap1)/heme oxygenase-1 (HO-1) is a key defense system, acting as a master regulator of cellular redox homeostasis and mitochondrial physiology [8,9]. Interestingly, Nrf2 is downregulated in the CL kidney by the 14 th day after UUO [10], which was postulated to be a predictor of apoptosis [11] and late mitochondrial dysfunction [12].
Thus, the balance between life and death in injured renal tubular cells seems to rely on the balance between mitochondrial dysfunction and redox homeostasis. Despite the potential relevance of transient or permanent modifications in the mitochondrial physiology, however, there are no thorough studies regarding mitochondrial respiration and redox status of CL kidneys. The driving hypothesis of the present study was that the differential profiles of CL, UUO, and Sham groups relative to both mitochondrial respiration parameters and in situ antioxidant systems could represent death and survival profiles of tubular cell fate. We have therefore investigated respiration and the redox mitochondrial status after 14 days of obstruction, i.e. in a chronic model, at which time the kidney shows an important degree of peritubular fibrosis, the O 2 supply is compromised, and mechanostreching of epithelial cells triggers a series of signaling pathways, including those involved in intracellular ROS balance [13].
Eighty-seven mice were randomized into 2 groups; n = 39 in the Sham group and n = 48 in the group submitted to left ureteral obstruction (UUO). After euthanasia, left kidneys from UUO and Sham animals, and right kidneys from obstructed mice-the contralateral kidneys (CL)-were collected.
Surgical procedure UUO and Sham groups were operated as previously described [14].

Tissue harvesting
After 14 days, mice received the same anesthesia protocol and euthanasia by exsanguination. Left kidneys from UUO and Sham, and the right kidneys from obstructed animals (CL), were harvested resected and maintained on ice for mitochondria isolation and assays (Sham, n = 23; CL, n = 32; UUO, n = 32), and for measurement of anti-oxidant enzymes activities (n = 10, each group). For histology, kidneys (n = 6, each group) were resected and prepared as previously described [15].

Isolation of the cortex corticis mitochondria
Cortex corticis was used because: (i) of its predominant oxidative metabolism [16] and (ii) >95% of its cell population corresponds to proximal convoluted tubule cells [17]. Mitochondria were isolated by differential centrifugation from kidney cortex corticis of UUO, Sham and CL groups immediately after resection, as previously described [18]. Enrichment with mitochondria was assessed by succinate dehydrogenase assays [19,20]. For respiration and swelling assays, the isolation method and the medium used were those described elsewhere [18], and mitochondria were immediately used. For measurement of ROS and Δψ m , the isolation medium contained 320 mM sucrose, 10 mM TRIS-HCl, 1 mM EDTA (disodium salt), 1 mM EGTA (free acid) and 0.1% (w/v) BSA, the final pH being adjusted to 7.4 with TRIS buffer. Respiration gave identical values when mitochondria were isolated in this medium. Mitochondrial protein was quantified by the Lowry method [21]. Table 1 presents the mitochondrial protein yield, which can be considered indicative of the amount of mitochondria in the tubular cells from the cortex corticis tissular segment. The recovery of mitochondrial protein was similar in Sham and CL groups, whereas it was 35% less in UUO. This lower value is a consequence of the smaller mass of the atrophied UUO kidney, as indicated by the decreased total homogenate protein in this group, and the similar ratio mitochondrial protein/total cortex corticis homogenate protein in the 3 groups.

Cytosolic and mitochondrial fractions for antioxidant enzymes assays in vitro
The cytosolic fractions of proximal tubules were obtained from the supernatant recovered after mitochondrial sedimentation [18], followed by a further centrifugation at 104,000 g to spin down plasma membranes [20]. Controls for residual subcellular contaminants were carried out as previously described [22]. The mitochondrial fractions for these assays were prepared after washing samples obtained as in [18] to remove the sucrose used in their final suspension. Mitochondria suspended in a solution containing 5 mM TRIS-HCl (pH 7.4), 150 mM NaCl and 1 mM EDTA were gently homogenized in a Potter Elvehjem homogenizer with a Teflon pestle before being centrifuged at 750 g for 10 min at 4˚C. The mitochondria-containing supernatants and the cytosolic fractions were stored at -70˚C until use.

Oxygen consumption (QO 2 ) measurements
The QO 2 of isolated mitochondria was measured using a high-resolution respirometer (Oroboros, Innsbruck, Austria) [23]. Mitochondria were suspended in the basic assay medium (37˚C) containing 320 mM mannitol, 10 mM TRIS-HCl (pH 7.4), 0.1% (w/v) BSA, 4 mM MgCl 2 , 80 μM EDTA and 8 mM phosphate-TRIS. The suspensions were supplemented with: (i) 5 mM pyruvate plus 2.5 mM malate, substrates for complex I; or (ii) 0.5 μM rotenone to inhibit mitochondrial complex I, followed by 10 mM succinate, substrate for complex II. The QO 2 under phosphorylating conditions was measured following the addition of 150 μM ADP and then 2 μg/ml oligomycin. The uncoupled QO 2 was assessed in the presence of FCCP (3 additions of 0.5 μM or a single addition of 1.5 μM).

Measurement of reactive O 2 species generation (ROS)
Mitochondria were suspended in the basic medium described below supplemented with 5 μM Amplex Red and 2 U/ml horseradish peroxidase to catalyze the oxidation of the non-fluorescent Amplex Red to the fluorescent resorufin, with 1:1 stoichiometry reduction of the H 2 O 2 produced [18]. Thus, the rate of H 2 O 2 formation corresponded to that of resorufin detected in a Cary Eclipse fluorometer (Varian, Palo Alto, CA), excitation at 563 nm/emission at 587 nm. The basic reaction medium contained 320 mM mannitol, 10 mM HEPES-TRIS, 4 mM KH 2 PO 4 , 4 mM MgCl 2 , 80 μM EDTA (disodium salt), 1 mM EGTA (free acid) and 0.1% (w/v) BSA, with the final pH being adjusted to 7.2 with TRIS. Immediately before the assay, the medium was supplied with a concentrated SOD solution (in 0.1 M phosphate-Tris buffer pH 7.0) to obtain a final concentration of 60 U/ml. SOD was added to circumvent the possibility of endogenous SOD being rate-limiting for the dismutation of the O 2 �released to the intermembrane space, the transport process that can occur directly from complex III, or via the voltage-dependent anion channels [24,25]. S1 Table shows that H 2 O 2 production was the same in the absence or presence of SOD. ROS formation was measured in media containing: (i) 5 mM pyruvate plus 2.5 mM malate, (ii) 0.5 μM rotenone plus 10 mM succinate, or (iii) 10 mM succinate without rotenone. In these 3 combinations of respiratory substrates, the reaction media were then successively supplied with 150 μM ADP, 2 μg/ml oligomycin and 1.5 μM FCCP.

Evaluation of mitochondrial membrane potential (Δψ m )
The fluorescent probe safranine O was used to estimate Δψ m [18,26,27] established between the mitochondrial matrix and the intermembrane space. Its fluorescence (5 μM) was monitored using a Hitachi model 4700 fluorometer (Tokyo, Japan) with excitation at 495 nm/emission 586 nm. The other components of the medium were as for the ROS assay in the presence of rotenone and succinate, but without oligomycin. Its aim was to explore how UUO affects (in CL and UUO kidneys): (i) the velocity of H + transfer from the matrix and formation of the Δψ m after energization and electron flux along the complexes, and (ii) the velocity of the nonspecific dissipation of the gradient after addition of FCCP.

Measurement of mitochondrial swelling in vitro
Mitochondrial swelling in vitro was measured before (in the presence of contaminant Ca 2+ ) and after 5 successive additions of CaCl 2 (5, 10, 20, 50 and 100 μM) to mitochondria suspended (0.4 mg/ml) in a medium containing 125 mM sucrose, 65 mM KCl, 10 mM HEPES-KOH (pH 7.2) and 5 mM succinate. Changes in light-scattering due to swelling were measured by following the decrease in optical density at 520 nm. Mitochondria were incubated for 1 min in the presence of contaminant Ca 2+ , with successive additions being made at intervals of 2 min.

Activity of the antioxidant enzymes in vitro
Glutathione-peroxidase (GPx). GPx activity was measured according to Flohé and Günzler [28]. The choice of tert-butyl hydroxiperoxide (T-BHP) for the assay in vitro was intended to eliminate the interference of catalase, which does not react with this peroxide. In a medium (37˚C) containing 100 mM HK 2 PO 4 /H 2 KPO4 buffer (pH 7.0), 1 mM EDTA (disodium salt), 1.2 mM T-BHP, 0.15 mM NADPH, 0.5 mM GSH, 0.24 U/ml glutathione reductase and 15 μg/ ml protein (mitochondrial or cytosolic), GPx activity was quantified following over 5 min the decrease of absorbance at 340 nm due to NADPH oxidation. One unit of GPx corresponds to 1 μmol oxidized NADPH/min. Superoxide dismutase (SOD). SOD activity was measured in the mitochondrial and cytosolic fractions of the proximal cells by the method of Crapo et al. [29]. The reaction medium (37˚C) contained 50 mM HNa 2 PO 4 /H 2 NaPO 4 buffer (pH 8.0), 0.1 mM EDTA (disodium salt), 10 μM KCN, 50 μM xanthine plus 2 μg/ml xanthine oxidase (the O 2 �generating system), and 20 μM cytochrome c, in the absence or presence of the mitochondrial or cytosolic fractions (15 μg/ml). Oxidation of cytochrome c was followed over 5 min at 550 nm. SOD activity was calculated by subtracting the slope of the cytochrome c oxidation linear recording in the presence of cell fractions from that in their absence. One unit of SOD corresponds to 50% reduction in the cytochrome c oxidation/min.

Catalase activity
Catalase activity was measured in the mitochondrial and cytosolic fractions of proximal cells by the method of Aebi [30]. . Twenty high quality photomicrographs (2048×1536 pixel buffer) were captured from non-overlapping renal cortical areas using a 40× objective. Data acquisition and analysis were blinded in all cases. Surface density of cytoplasmic cytochrome c, coenzyme Q-10B, Keap1, HO-1, caspase 3, and active caspase 3. The tubular reactive zones for these proteins were expressed as a percentage of tissue surface area in renal cortical tubules.
Nuclear cytochrome c, Nrf2, Keap1, HO-1, caspase 3, and active caspase 3. The percentage of reactive tubular nuclei to the total area in the renal cortical tubules represents the nuclear labeling index.

Statistical analyses
Histomorphometrical measurements were done by one investigator in a blinded manner. The data-mitochondrial parameters and enzyme activities-were analyzed using GraphPad Prism 6.01 software (GraphPad Software, Inc., La Jolla, CA) and were expressed as means ± standard error of the mean (SEM). The differences between the groups in the figures were analyzed using one-way ANOVA. Differences were considered significant at p<0.05, with asterisks indicating the level of significance: � p<0.05, �� p<0.01, ��� p<0.001, ���� p<0.0001; ns, indicates no significant differences. For the sake of better visualization, different lower-case letters after SEM values in the tables indicate statistical differences between the mean values (p<0.05). In the tables analyzed by using unpaired t-test, the asterisks indicate the level of significance.  Mitochondrial ROS generation decreases in CL and UUO kidneys. Cortical mitochondrial production of ROS in vitro was measured by the rate of H 2 O 2 formation [18]. Assays were carried out in the presence of: (i) substrates for complex I (Fig 1), (ii) succinate in the presence of rotenone (Fig 2), or (iii) succinate in the absence of rotenone (Fig 3). Table 2 compares the results obtained using succinate in the presence and absence of rotenone. Determinations were first made in the basal conditions (no exogenous substrates) and then in the presence of ADP with either substrate (when ATP is formed), when ATP synthesis is blocked (oligomycin), and when respiration was uncoupled from ATP synthesis by FCCP (Figs 1A-1E and 2A-2F). This combination of substrates/inhibitors was assayed because O 2 �formation depends on the magnitude of Δψ m and substrate concentrations [31,32]; also because O 2 �formation also occurs at the level of complex I by a reversal flux of electrons from complex II when succinate is oxidized [31,33]. In all respiratory states, mitochondrial ROS formation was highest in the Sham group, significantly depressed in the CL group, and very low in the UUO group (Figs 1 and 2). The results with the use of succinate alone show increased production of ROS in the absence of rotenone in all groups (Table 2). These differences disappeared in the Sham mice after addition of ADP, oligomycin and FCCP, but remained significantly higher in the UUO group. The difference persisted only after addition of oligomycin in the case of the CL group.

4-Hydroxy-2-nonenal (4-HNE) immunoreactivity demonstrates lipid peroxidation in CL and UUO kidneys
Lipid peroxidation is the major consequence of oxidative damage, which can be detected using an antibody against the major peroxidation product, 4-HNE [34]. Sham kidneys show a faint staining of tubular cells and CL tubular cells show a clearly visible cytoplasmic staining, whereas UUO kidneys show diffuse and intense cytoplasmic reactivity (Fig 4A-4C).

CL, like UUO kidneys, have impaired mitochondrial functioning
Oxygen consumption (QO 2 ). The quantified QO 2 data are given in Table 3 (pyruvate plus malate data) and 4 (succinate data). With the substrates for complex I, only the UUO  Table 3. Oxygen consumption (pmol O 2 × mg -1 × s -1 ) of kidney mitochondria from Sham, CL and UUO mice. Assays were carried out with addition of pyruvate and malate − substrates for complex I − under phosphorylating and non-phosphorylating conditions (states 3 and 4: with and without ADP, respectively), with ATP synthesis blocked (oligomycin), and in the uncoupled state of respiration (FCCP). Uncoupled respiration was measured after adding 1.5 μM FCCP. group had accentuated respiratory damage under non-phosphorylating, phosphorylating and uncoupled conditions (Table 3). In contrast, when the substrate for complex II was used (Table 4), QO 2 was also severely impaired in the CL mitochondria, which had a lower phosphorylating capacity (indicated by a depressed response to ADP), and uncoupled respiration was also inhibited. Except for assays after oligomycin, the depression of respiration in all states was much more accentuated in the UUO group in the presence of succinate. The respiration control ratio, RCR, which is considered to be an estimate of mitochondrial coupling, was: (i) similar in the 3 groups with pyruvate/malate, despite the decreased response to ADP, because inhibition of respiratory states 3 and 4 was similar; (ii) depressed in the CL and UUO groups when succinate was oxidized. Fig 5 gives the findings of representative experiment that allows simultaneous visualization of O 2 concentration decay, mitochondrial QO 2 and, in particular, the response to ADP. Profiles of O 2 concentration decay and QO 2 in the 3 different groups after successive additions of rotenone, succinate, ADP, oligomycin, and FCCP can be seen in Fig 5A-5C. The response to ADP (phosphorylative state) in the 3 groups can be compared by the differences in the slopes of the dashed (state 3) and dotted lines (state 4) [35], which were inserted into the traces of [QO 2 ] decay (Fig 5D-5F). Transmembrane electric potential gradient (Δψ m ). The mitochondrial transmembrane electrochemical potential gradient for H + (matrix negative and alkaline) is the proton motive force (Δρ) generated across the inner mitochondrial membrane during electron flux along the mitochondrial complexes [36]. The transmembrane electric potential Δψ m was similar in the 3 groups (Fig 6A-6D) though its formation was slower in mitochondria from the CL and UUO kidneys, demonstrated by the different rates in the decrease in fluorescence after addition of succinate (Fig 6A-6C). In close agreement with that found in response of QO 2 to ADP, the partial depolarization was inhibited by 50% in CL and 70% in UUO after addition of a micromolar ADP pulse (Fig 6E).

Additions
Mitochondrial swelling. In isolated mitochondria, opening of the permeability transition pore (mPTP) leads to mitochondrial swelling [37]. Ca 2+ -induced swelling was investigated by measuring the decrease in optical density (Table 5). In the presence of contaminant Ca 2+ (~10 μM, as quantified by atomic mass spectrometry) [20], absorbance was~50% lower in the CL and UUO groups compared to the Sham group. Further addition of CaCl 2 led to a progressive decrease of absorbance in the Sham group, which was less accentuated in the CL and UUO groups. There were no differences between CL and UUO in all range of Ca 2+ concentrations.

Altered redox homeostasis in CL and UUO kidneys: different profiles depending on the enzymes classes and the intracellular compartment
We measured the activities of glutathione peroxidase (GPx), superoxide dismutase (SOD) and catalase in mitochondria and cytosol from the cortical region (Fig 7). GPx activity increased only in the mitochondria of the CL group and decreased in the cytosol of the UUO group in relation to the other 2 groups (Fig 7A and 7B). Mitochondrial SOD activity increased in the UUO group compared with Sham, had an intermediate value in the CL group, with a mirror image of activity in the cytosolic enzyme (Fig 7C and 7D). Cytosolic catalase activity was  crippled in UUO kidneys, with a significant increase in the CL group compared to the Sham group (Fig 7E).

Cytochrome c-nuclear reactivity increased in CL and UUO kidneys
After mPTP opening, cytochrome c is released, activating caspase 3 and promoting cell death [38]. Immunohistochemical assays showed a cytoplasmic punctate staining of cytochrome c with no difference among the 3 groups of kidneys (Fig 8A-8C and 8E). However, a very high  level of nuclear cytochrome c was seen in UUO kidney (Fig 8C and 8F) and lower, though significantly different, in the CL group (Fig 8B and 8F) compared with the Sham group, in which there was no staining (Fig 8A). Horseradish peroxidase-based immunohistochemistry is not sensitive enough to determine mitochondrial content of cytochrome c; it can only quantify diffuse cytosolic and nuclear cytochrome c-containing areas.

The coenzyme Q-binding protein, CoQ10B, is diminished in CL and UUO kidneys
Coenzyme Q is essential in shuttling electrons among different mitochondrial complexes [39], also being one of the most potent lipophilic antioxidants [40]. Whereas Sham kidney have a diffuse reactivity for CoQ10B antibody in all tubular cells, CL kidney showed heterogeneous and diminished cytoplasmic reactivity, which was further decreased in the dilated tubules from UUO kidney (Fig 9A-9C).

Alterations in the Nrf2/Keap1)/HO-1 system in CL and UUO kidneys
CL and UUO tubules have increased amounts of nuclear phosphorylated Nrf2. UUO kidneys (Fig 10C) had the highest level of nuclear Nrf2, which was absent in Sham kidneys (Fig 10A), with intermediate levels in CL kidneys (Fig 10B).
UUO kidneys have the highest levels of cytoplasmic and nuclear Keap1 staining. UUO kidneys had a diffuse intense cytoplasmic and nuclear reactivity (Fig 11C, 11E and 11F), Sham kidneys (Fig 11A, 11E and 11F) possessed faint reactivity in the apical tubular regions with rare nuclear staining, and CL kidneys (Fig 11B, 11E and 11F) had intermediate surface reactivity for Keap1 in the cytosol, with barely detectable staining of the nuclei.

Caspase 3 and active caspase 3 alterations in CL and UUO groups
Representative immunostainings of caspase 3 and active caspase 3 in the cytosol and nuclei from cortical tubular cells are presented in Fig 13. Whereas scarce or absent reactivity was found in Sham, an important nuclear staining was encountered in CL. UUO mice had elevated levels of both enzyme forms (Fig 13A-13F). The quantifications seen in Fig 13G-13J allow to perceive the important upregulation of nuclear caspase 3 and cytosolic and nuclear active caspase 3 in CL, which in the 3 cases attained approximately 50% of the highest UUO levels.

Discussion
The results demonstrate profound functional mitochondrial alterations in the CL kidney of mice when the UUO kidney has tissue alterations similar to those found in CKD [13]. The considerable recent advance comes from taking organ biopsies for mitochondrial oximetric studies in a preserved cell environment [41]. However, we have used here isolated mitochondria: (i) to investigate mitochondrial ROS formation without interference from the cytosolic environment, and (ii) for the assessment of mitochondrial swelling in vitro with raising Ca 2+ concentrations.
We have shown that in chronic UUO, kidneys from CL and UUO mice have differential profiles regarding mitochondrial respiration and ROS generation (as also in enzymes involved    in the redox status) compared to Sham mice. QO 2 in vitro allowed determination of the degree of mitochondrial dysfunction in the different respiratory states. Isolated cortical CL and UUO mitochondria had striking functional alterations that cannot be correlated, respectively, to humoral stimuli for compensatory growth in the case of the remaining kidney after uninefrectomy [42] or with chronic ischemia in the case of UUO. The decrease in QO 2 in all states with succinate in CL and UUO, and only in UUO when pyruvate and malate are oxidized (Fig 5;  Tables 3 and 4), as well as the intense and maximal mitochondrial swelling in the presence of a very low Ca 2+ (~10 μM; Table 5), point to intense overall impairment in mitochondrial structure and function in CL and UUO cases, especially at the level of the electron pathway starting at complex II.
The Δψ m generated after mitochondrial energization in CL and UUO kidneys is notably similar to that encountered in Sham kidneys, although the delay in its formation and the loss in the response of Δψ m to ADP (Fig 6) correlate well with the impaired response of respiration to ADP (Fig 5; Tables 3 and 4). These alterations are possibly provoked by the decrease in CoQ10 (Fig 9), in the velocity of electron flux, and in the functioning of the F 0 F 1 -ATPsynthase, as indicated by a parallel diminution of the FCCP-stimulated respiration and in the RCR. These functional changes could be associated with a decrease in the abundance of respiratory complexes [43] and in mitochondrial DNA [44], which have been reported in obstructive nephropathy, but not as far as we know in the CL kidney from UUO animals.
With respect specifically to respiration, the profiles seen in Sham, CL and UUO mitochondria after QO 2 was measured following addition of pyruvate plus malate or succinate had differences indicative of a more accentuated effect of obstruction on the electron fluxes through the branch starting at complex II in both kidneys of obstructed mice. This conclusion is supported by the observation that QO 2 of CL mitochondria was unaffected in all respiratory states when pyruvate and malate were present (Table 3), whereas it was significantly decreased in succinate (Table 4). Moreover, an inhibition of 90% on the QO 2 assessed in mitochondria respiring with succinate in the presence of oligomycin (only 50% with pyruvate plus malate) in the UUO group suggests a major impairment of proton leak [45,46] that seems to be sensitive to the redox status of the intramitochondrial NADH pool. The same preferential damage of the succinate-activated branch became evident when uncoupled respiration with the different substrates was assayed.
From the present results, it clearly emerges that there is an intriguing functional difference at the mitochondrial level between the CL kidney from UUO mice and the remaining kidney of uninephrectomized rats, even though both undertake compensatory renal growth [4]. Among these differences, the increased rate of respiration found in the uninefrectomy model deserves special mention, and also in the profile of ROS formation and abundance of antioxidant enzymes [47]. The progressive decrease in ROS formation from Sham ! CL ! UUO kidneys in the different respiratory states with the use of substrates for complexes I and II (Figs 1-3) contrasts with that reported in the remaining kidney after uninephrectomy [48], and in the CL organ ("intact opposite") from UUO kidneys [49]. The diminution found here matches well with the inhibition of QO 2 with succinate as the respiratory substrate (Table 4), and therefore with a decreased possibility of premature electron transfer to O 2 upstream of the cytochrome c oxidase, especially in the branch that starts at complex II.
The comparison of ROS production with succinate, either in the absence or presence of rotenone (Table 2), showed up 4 important peculiarities that shed light on the possible different impact on CL and UUO mitochondria, in terms of possible mechanisms that had been differentially modified. First, the huge increase in succinate alone in the absence of rotenone, which augmented 60, 100 and 400% from Sham ! CL ! UUO kidneys shows that O 2 �formation by reverse electron transport at complex I [31][32][33] is progressively stimulated as the lesions progress. Since this pathway is considered the major route of mitochondrial redox signaling in physiological and pathological conditions [32], the differences could mean that the dissimilar lesions found by immunohistochemistry (Figs 4, 10 and 11) rely-at least in part-in upregulated O 2 �formation by reverse electron flux at complex I, potentiated by the downregulated CoQ pool (Fig 9). The second differences that deserve mention are those related to ROS production assessed in the presence of ADP. Again, only in UUO mitochondria was O 2 �formation lower when rotenone was added, which is suggestive of obstruction stimulating the reversal pathway due to modifications in Δψ and in the NADH pool redox state, which are the thermodynamic driving forces of this pathway [32]. Third, comparison of ROS formation in Sham, CL and UUO isolated mitochondria treated with oligomycin in the absence and presence of rotenone also points to a similarity between CL and UUO kidneys. The highly significant increase in ROS when reverse electron flux is allowed to occur and ATP synthesis is blocked is also indicative of the special relevance of complex I in the redox status of both kidneys from unilaterally obstructed mice. Finally, comparison of ROS formation in the absence of rotenone, but with FCCP present, also corroborates the view regarding the importance of complex I in an altered redox signaling when chronic lesions have become established in the UUO kidney.
How then can we combine decreased mitochondrial ROS formation (Figs 1-3) with the prominent mirror-image increase in 4-HNE from Sham ! CL ! UUO seen in renal tissue (Fig 4)? It is possible that the immunofluorescence images are relics of a previous ROS burst from different origins soon after obstruction, which persists after 14 days, being sustained by extra-mitochondrial production originating in arterial vessels and inflammatory cells [11,47,49] present in the interstitium (seen in Fig 4C).
Differences from previous results regarding oxidative stress obtained by other laboratories deserve special discussion. Increased ROS production in UUO kidneys without determination of the precise source was reported by studying end-products of lipid peroxidation [11], stress response proteins [10,50], or with the use of fluorescence probes [51]. These differing results compared with the direct mitochondrial measurements presented in Figs 1-3 probably reflect events of undefined origin that no longer occur in mitochondria after 14 days, as mentioned above.
In relation to antioxidant enzymes, our results also indicate important and selective differences in CL kidneys either with respect to Sham or UUO kidneys, differences that were also found when compartmentalization of cytosolic and mitochondrial enzymes are compared. The increase in mitochondrial GPx activity ( Fig 7A) and cytosolic catalase activity (Fig 7E) in CL kidneys are indicative of molecular and specific compensatory responses-which are synchronous in cytosol and mitochondria, as well as being consistent with the concept of renal counterbalance-in face of an altered ROS homeostasis ( Fig 4B) and a significant tendency to apoptosis (Fig 8). These synchronized responses might ensure late compensatory hypertrophy, vasodilatation and adaptive changes in the renin/angiotensin system, as detected in CL kidneys [49,52].
The results in Fig 10 showing increased levels of the master regulator of redox homeostasis phosphorylated nuclear Nrf2 [53] are in line with the view of an intense initial oxidative stress that results in induction of its dose-dependent upregulation by 4-HNE [54,55]. This upregulation seems to be the initial event of a cascade that culminates in the protective upregulation of antioxidant enzymes in CL kidneys (Fig 7). As proposed for UUO kidneys [12], the deregulated Nrf2/Keap1 signaling pathway in CL mitochondria could affect NF-κB signaling-mediated inflammation, autophagy, apoptosis and-in terms of apoptosis-related proteins-the Bax/ Bcl-2 ratio. The increase in apoptotic cells in renal cortex in CL and UUO kidneys (S2A-2C and S2G Fig)-which were barely detectable in Sham kidneys-support the idea that, besides the modifications provoked in redox signaling, the abnormal Nrf2/Keap1 abundance and localization also impacts cell survival in both kidneys of the obstructed mice.
The increased cytosolic Keap1 in both CL and UUO kidneys (Fig 11) seems at first sight be contradictory with a similar increment in nuclear Nrf2: in a model of CKD, Western blotting analysis indicated an inverse relationship between Nrf2 and Keap1 in chronic tubulointerstitial nephropathy [56]. This unexpected and huge increase in nuclear Keap1 in UUO kidneys, which was barely detectable in CL kidneys, constitutes a differential in the lesions induced in the 2 kidneys of obstructed mice, which could mean that the regulation of Keap1 synthesis and shuttling function continues to be completely altered when chronic lesions become fully established. It is hypothesized that this late increase in nuclear Keap1 and, possibly of its constitutive hyperactivity, results from deregulated synthesis and turnover of the protein during the intense renal oxidative stress caused by the obstruction during the first days, as described by Chung et al. [12].
In the case of persistent ureteral obstruction with impaired renal blood circulation, hypoxia could also induce nuclear enrichment of a non-activated Nrf2, a hypothesis that receives support from the observation that the abundance of HO-1 is the mirror-image (Fig 12) of Nrf2 and Keap1 in Sham, CL and UUO kidneys. Other regulatory proteins could promote dissociation of Nrf2 from Keap1, thereby decreasing Nrf2 trafficking toward the proteasomal pathway and increasing its half-life in the nuclear compartment. p62 can directly interact with Keap1 in oxidative stress conditions, resulting in increased levels of Nrf2 in the nucleus [57]. The use of an antibody against phosphorylated Nrf2 indicates to us that its preferential accumulation in the nucleus is a consequence of a kinase-mediated phosphorylation of Nrf2 [58,59], which could also contribute to its dissociation form Keap1.
The unexpected finding of cytochrome c nuclear translocation (Fig 8) seems to indicate a possible evolution to apoptotic cell death, which is presumed to happen at least in UUO mice, even though in our model a few-though increased when compared to Sham-ApopTag reactive cells were found (S2 Fig). These findings are apparently more contradictory when few Apoptag reactive cells coexists with the huge upregulation of nuclear caspase and caspase 3 in CL and, even more, in the UUO group. However, in some situations, cells could survive despite high levels of active caspase 3, as demonstrated in sensory neurons from long term experimental diabetic rats [60]. Moreover, a recent study demonstrated that cytochrome c nuclear translocation in response to DNA damage is implicated in the attenuation of nucleosome assembly, thus increasing the time available for DNA repair [61]. On the basis of these observations, we can propose that the increased caspase 3 and active caspase 3 (Fig 13) could indicate an adaptive response to the altered microenvironment, thus explaining why CL kidneys can recover their structure and function.
Finally, we conclude that CL kidneys stand at the crossroad between death and survival, and that, despite the impaired mitochondrial physiology after 14 days, recovery of overall renal function may rely in upregulation of key antioxidant enzymes.

Limitations of the study
While we have investigated QO 2 and ROS formation after mitochondrial energization ate the level of complexes I and II, this was not at the level of complexes III and IV. We also need to assess the expression, abundance and organization of mitochondrial supercomplexes involving different combinations of complexes I, II, III and IV coexisting with their monomeric forms [62]. These combinations could be differentially altered in CL and UUO kidneys, which deserves further investigation. Data represent mean ± SEM (n = 6), submitted to one-way ANOVA test followed by Tukey's test. �� p<0.01, ���� p<0.0001. Apoptosis and proliferation were barely detected in Sham kidney. CL kidney presented with increased tubular cells apoptosis and tubular cell proliferation, while in the UUO kidney there was a more accentuated increase in both the number of apoptotic and PCNA + tubular cells. (TIF) S1 File. Immunohistochemistry and immunofluorescence. (DOCX) S1 Table. Comparative analysis of ROS production (pmol H 2 O 2 × mg -1 × min -1 ) by kidney mitochondria from Sham, CL and UUO mice in the presence of succinate and rotenone, and without (−) or with SOD (+). (DOCX)