Early Effects of Extracellular Vesicles Secreted by Adipose Tissue Mesenchymal Cells in Renal Ischemia Followed by Reperfusion: Mechanisms Rely on a Decrease in Mitochondrial Anion Superoxide Production

Abstract

Acute kidney injury (AKI) caused by ischemia followed by reperfusion (I/R) is characterized by intense anion superoxide (O₂^•−) production and oxidative damage. We investigated whether extracellular vesicles secreted by adipose tissue mesenchymal cells (EVs) administered during reperfusion can suppress the exacerbated mitochondrial O₂^•− formation after I/R. We used Wistar rats subjected to bilateral renal arterial clamping (30 min) followed by 24 h of reperfusion. The animals received EVs (I/R + EVs group) or saline (I/R group) in the kidney subcapsular space. The third group consisted of false-operated rats (SHAM). Mitochondria were isolated from proximal tubule cells and used immediately. Amplex Red™ was used to measure mitochondrial O₂^•− formation and MitoTracker™ Orange to evaluate inner mitochondrial membrane potential (Δψ). In vitro studies were carried out on human renal proximal tubular cells (HK-2) co-cultured or not with EVs under hypoxic conditions. Administration of EVs restored O₂^•− formation to SHAM levels in all mitochondrial functional conditions. The gene expression of catalase and superoxide dismutase-1 remained unmodified; transcription of heme oxygenase-1 (HO-1) was upregulated. The co-cultures of HK-2 cells with EVs revealed an intense decrease in apoptosis. We conclude that the mechanisms by which EVs favor long-term recovery of renal structures and functions after I/R rely on a decrease of mitochondrial O₂^•− formation with the aid of the upregulated antioxidant HO-1/Nuclear factor erythroid 2-related factor 2 system, thus opening new vistas for the treatment of AKI.

1. Introduction

Acute kidney injury (AKI) is one of the most severe systemic syndromes in internal medicine, with significant mortality rates, especially in intensive-care units (ICUs), where it accounts for more than 15% of the hospitalized patients [1] and 50% of those that are critically ill [2]. The economic impact is also very high; in the United States, it is estimated that it is greater than 20 billion dollars per year [3]. In the face of the controversy regarding several non-specific treatments that are offered nowadays and the severity of the outcomes [4,5], the search for therapeutic alternatives is growing worldwide. In the last few decades, replacements started focusing on cell therapies to prevent or slow the progression from AKI to chronic kidney disease (CKD). In the beginning, expectations were centered on renal resident adult stem cells [6,7], but this hope was almost totally dissipated when the existence of such cells was challenged [8,9]. Now, the use of progenitor cells is intensely focused on mesenchymal stromal cells (MSC) and the vesicles they secrete to the extracellular milieu (EVs). Recently, we demonstrated that EVs secreted by adipose MSC block the progression of cellular and molecular lesions 3 days after AKI provoked by ischemia followed by reperfusion (I/R) in rats [10].

The sudden temporary impairment of renal blood flow (ischemia) followed by restoration of circulation (reperfusion) is a central event in AKI of different etiologies [11], and intense production of anion superoxide (O₂^•−) occurs in the ischemic kidney followed by reperfusion (I/R) [12]. The exacerbated production of reactive O₂ species (ROS)—together with an important inflammatory component—is a central process in the physiopathogenesis of AKI-associated lesions as a consequence of I/R, which initiates a cascade of molecular events and aberrant signaling that culminates in tubular destruction [13,14], severe impairment of renal function, and high mortality or frequent progress to CKD and end-stage renal disease [15]. Increased Bax/Bcl2 ratios and increased lipid peroxidation are key pathological molecular events resulting from the exacerbated formation of O₂^•− during the I/R-induced AKI [14] because they can lead to mitochondrial damage and cellular necrosis.

The early moments of AKI onset and the response to treatments are crucial to define AKI evolution and prognosis [16]. Therefore, knowledge regarding the molecular and cellular events in these early stages can support the search for appropriate therapies and their targets, especially the mechanisms altered after the burst and during ongoing ROS formation. Considering that the effect of EVs secreted by adipose MSC, following subcapsular administration at the beginning of reperfusion, prevented or diminished tissular and functional damage 3 days after I/R [10], the present study aimed to investigate the early (24 h) EVs-induced molecular processes responsible for the late beneficial outcomes. We previously demonstrated that EVs stained with the fluorescent probe Vybrant DiD diffuse and are uniformly distributed in the kidneys 24 h after their injection [10]. We hypothesized that they could modulate processes linked to preventing oxidative damage before sustained transcriptional and translational renoprotective effects occur.

Specific objectives of this study were to investigate whether subcapsularly administered EVs, i.e., an acellular therapeutic approach, could modulate: (i) renal ROS production 24 h after I/R (bilateral arterial clamping for 30 min followed by 24 h of reperfusion), (ii) the O₂ consumption in different respiratory states to dissect the influence of I/R and EVs on electron fluxes and ATP synthesis, and (iii) the transcriptional status of injury molecular markers, antioxidant enzymes, and pro-inflammatory cytokines. Measurements of plasma levels of the lesional biomarkers’ urea and creatinine aimed to determine the renal damage status. Finally, we investigated the effect of EVs on apoptosis in vitro and on electrical potential difference across the inner mitochondrial membrane potential (Δψ), using cytometry and cultures of an established lineage of human renal proximal tubule cells (HK-2) subjected to 24 h of hypoxia.

2. Results

2.1. Characterization of Extracellular Vesicles Secreted by Adipose Tissue Mesenchymal Stromal Cells

The mesenchymal stromal cells (MSC) of adipose origin (ADMSC) secreted the EVs presented in Figure 1A. They constitute a population of microvesicles (diameter above 120 nm) and exosomes (diameter ranging 50–120 nm), as demonstrated by the transmission electron microscopy (TEM) representative images. The images at higher magnification reveal a well-delimited membrane that encircles non-homogeneous content. The raw images are presented in Figure S1A,B. In Figure 1B, the detection of surface EV markers [17] confirms the nature of the vesicles seen in panel A, with enrichment of exosome markers such as tetraspanins and endosomal-related proteins. In a previous study [10], we demonstrated the presence of the MSC markers CD73, CD90, and CD105 (but not CD146) in the EVs isolated from cultures of ADMSC (see Figure S2).

2.2. Mitochondria Subjected to Hypoxic Damage Are Key Targets for the Extracellular Vesicles Secreted by Mesenchymal Cells

Processes that contribute to the maintenance of the electrical potential difference across the inner mitochondrial membrane (Δψ) are key for the potential recovery of damaged cells with high metabolic demand and elevated rates of O₂ consumption. This is the case for renal proximal tubule cells, in which mitochondrial integrity is required for the synthesis of ATP in a nephron segment where active processes of transport, mainly of Na⁺, take place at high rates [18]. Figure 2A demonstrates the normal mitochondrial morphology in HK-2 cells cultured in CTR conditions (i.e., the cells are maintained in normoxia, on the left), the appearance of pathological myelin figures in mitochondria from cells cultured in hypoxia (in the middle of the panel), and the recovery of the normal mitochondrial morphology when HK-2 cells were subsequently co-cultured with EVs (on the right). The raw images are presented in Figure S3A–C.

The bar graph in Figure 2B quantifies the mitochondrial area, showing its decrease in hypoxia and the recovery of the CTR mean value when the renal cells were subsequently cultured with EVs.

Figure 2C shows representative images of MitoTracker™ Orange-labeled (at three different magnifications). The mitochondrial network around the nucleus is dense and intertwined within HK-2 cells in CTR (a qualitative indication of preserved Δψ; on the left) and rarified in the cells subjected to hypoxia for 24 h (in the middle). When HK-2 cells previously exposed to 1% O₂ for 24 h were subsequently co-cultured with 1% O₂ plus EVs for an equal period of time, the CTR mitochondrial morphology was partially recovered (on the right). Figure 2D presents the distribution of the number of events normalized to the modal value, which were recorded in flow-cytometry analysis of HK-2 cells also stained with MitoTracker™ Orange. Mean fluorescence intensity (MFI), which was also analyzed by flow cytometry, highlights the complete overlap in the intensity measured in CTR, HPX, and HPX + EVs conditions, indicating that despite the mitochondrial morphological alterations, Δψ is preserved under the hypoxia assay conditions. Figure 2E (mean ± SEM) quantifies the values of mitochondrial fluorescence intensities in arbitrary units (AU). These data suggest that organization and structure of renal mitochondria may be the target of potential beneficial effects of EVs, in early stages after hypoxic injuries.

2.3. Subcapsular Administration of Extracellular Vesicles Secreted by Adipose Tissue Mesenchymal Cells Restores Normal Values of Anion Superoxide Formation by Renal Mitochondria after Ischemia/Reperfusion

The rates of O₂^•− formation by isolated renal mitochondria in different respiratory states were measured in SHAM, I/R, and I/R + EVs rats after 24 h of reperfusion. Figure 3 presents representative recordings of resorufin fluorescence variations after successive additions of substrates, uncoupler, and inhibitors to mitochondrial suspensions. These recordings correspond to H₂O₂ formation after dismutation of the O₂^•− that is produced as a consequence of premature transfer to O₂ of the electrons removed during oxidation of endogenous substrates and succinate. Comparison of Figure 3A,B shows that the velocity of O₂^•− formation increases by 50% after addition of succinate (traces between peaks 1 and 2) in mitochondria isolated from I/R rats, which returned to the SHAM value when EVs were injected at the beginning of reperfusion (Figure 3C). Quantification of the velocities when mitochondria oxidize succinate in non-phosphorylating conditions is given in Figure 3D. It can be seen that EVs treatment totally recovers the control formation of H₂O₂.

After adding 0.1 mM ADP, a condition in which phosphorylation to ATP occurs at low velocity, there was no difference among the three groups (average of 1000 pmol H₂O₂ × mg⁻¹ × min⁻¹ over the entire period) (data not shown). When electron transfer and ATP formation were accelerated by the addition of a higher ADP concentration (1 mM) (Figure 4A), H₂O₂ formation decreased in the three groups (compare with Figure 3D) and the 50% stimulation in the I/R group with respect to SHAM appeared again, as well as the recovery of the SHAM velocity in the I/R + EVs group. A similar profile of stimulation/recovery, but at higher levels, was encountered when ATP synthesis—and respiration—was blocked by oligomycin (Figure 4B). When the respiration became accelerated and the H⁺ electrochemical gradient dissipated by the addition of the uncoupler FCCP (Figure 4C), the H₂O₂ formation decreased in all groups. Still, the profile of increase by I/R and total normalization by EVs persisted. Finally, when electron fluxes were greatly diminished by inhibition of Complex III by Antimycin A, the rate of H₂O₂ formation increased by 100% with respect to SHAM, and the decrease promoted by EVs was partial (Figure 4D).

An interesting feature was encountered when H⁺ leak, one of the most important mechanisms of antioxidant defense in mitochondria [19], was investigated. Proton leak in mitochondria (i.e., the return of H⁺ from the interspace to the matrix through pathways different from the F_oF₁-ATPsynthase) can be estimated from the difference between the QO₂ in the presence of oligomycin and the residual QO₂ after inhibition of Complex III by Antimycin A [20,21,22]. Figure 5 shows that H⁺ leak was similar in the three groups and, therefore, that the ischemic episode followed by reoxygenation did not affect the pathways through which H⁺ flows back to the matrix through pathways different from the F_oF₁-ATPsynthase.

2.4. Twenty-Four Hours after Ischemia Followed by Reperfusion, the Acute Renal Lesions Still Persist

To see whether the early normalization of the redox status of renal cortical mitochondria was associated with amelioration of the acute lesions provoked by the ischemia followed by reperfusion, we investigated the expression of two biomarkers of kidney damage 24 h after the injury. They were Kidney Injury Molecule-1 (KIM-1) and Neutrophil Gelatinase-Associated Lipocalin (NGAL), which are considered sensitive and specific markers of proximal tubule lesions [23,24]. KIM-1 was barely detectable in SHAM conditions and increased approximately 20-fold in the renal cortex corticis of I/R rats, a level that remained unmodified in the group I/R + EVs (Figure 6A). As in the case of KIM-1, the expression of NGAL was very low in SHAM rats, increasing more than 15-fold after I/R without influence of EVs treatment (Figure 6B). The persistent pathological dysfunction of AKI is confirmed by the elevated levels of urea (87.8 ± 2.0 mg/dL vs. 57.8 ± 1.0 mg/dL in SHAM; p = 0.0017) and creatinine (0.47 ± 0.01 mg/dL vs. 0.28 mg/dL in SHAM; p = 0.0001) in blood, which were not modified by EVs administration in the 24 h window of lesion (100.3 ± 10.4 mg/dL for urea, NS with respect to I/R, p = 0.0001 vs. SHAM; 0.52 ± 0.05 mg/dL for creatinine, NS with respect to I/R, p < 0.0001 vs. SHAM) (one-way ANOVA, followed by Tukey’s test; n = 5–7).

Since hypoxia and exacerbated ROS formation can trigger and positively feed inflammatory processes in kidney tubule cells and their interstitial surroundings [13], the next step was to study the gene expression of two pro-inflammatory cytokines: Interleukin-6 (IL-6) (Figure 7A) and Tumor Necrosis Factor-alpha (TNF-α) (Figure 7B). Both cytokines were remarkably upregulated after ischemia followed by 24 h of reperfusion to 700% and 100% higher levels, respectively, with respect to SHAM values. The effects of the EVs administration were different for each cytokine: IL-6 dropped by half compared to I/R and TNF-α remained unmodified in I/R + EVs rats, probably due to different influences of redox environment alterations in the release of these cytokines. This point will be discussed below.

2.5. The Early Anti-Oxidative Responses after EVs Administration Co-Exist with Unmodified Transcriptional Activation of Key Enzymes Involved in Redox Regulation

It is well known that enzyme-catalyzed antioxidant mechanisms control the balance between cell death and survival in acute renal lesions [25]. Thus, we investigated the gene expression: (i) of a master regulator of cellular redox homeostasis and mitochondrial function [26,27], the heme oxygenase-1 (HO-1), and (ii) of two enzymes that catalyze the sequential split of H₂O₂ to H₂O and O₂ after its formation by dismutation of O₂^•−: catalase (CAT) and superoxide dismutase (SOD), respectively. The expression of HO-1 in proximal tubules increased by 300% after I/R compared to SHAM. The upregulated levels remained unmodified by EVs injection (Figure 7C), evidence that I/R triggered an antioxidant mechanism of defense that could participate in the early restoration of the mitochondrial redox status presented in Figure 3 and Figure 4, without any influence of the EVs components. The profile was the opposite for CAT (Figure 7D) and SOD (Figure 7E) transcription, which was downregulated in I/R again without any EVs effects, indicating that I/R partially inhibited transcriptional processes unable to be restored—in the early phase of the damage—by the factors that the vesicles carry.

2.6. Apoptotic Processes Are Partially Reversed When Renal Proximal Tubule Cells Are Co-Cultured with Extracellular Vesicles after Hypoxia

We investigated whether EVs can prevent cell death in a lineage of kidney proximal tubule cells besides their beneficial influence on the increased redox stress. HK-2 cells subjected to hypoxia for 24 h were analyzed by Annexin/PI staining (Figure 8). Hypoxia, which mimics ischemia in vitro, increased the number of apoptotic cells, especially of those in late-stage, i.e., of ANX⁺PI⁺ cells (by approximately threefold; Figure 8A,B,E), and increased by 100% the number of cells in an early stage of apoptosis (ANX⁺PI⁻) (Figure 8A,B,D). The co-culture with EVs in the reoxygenation phase significantly decreased the process of early cell death (Figure 8C,D), but not that of late apoptosis (Figure 8C,E). When the number of total apoptotic cells (early and late stage) (Figure 8F) was considered, the mean value of the group (HPX + EVs) (6.2%) was statistically similar to those of the groups CTR (4.0%) and HYX (8.3%).

3. Discussion

The main finding in this study was that subcapsular administration of MSC-derived EVs at the moment of reperfusion after ischemia totally recovers the normal ROS formation by mitochondria from renal proximal tubule cells in all functional respiratory states. Recovery of basal and controlled production of ROS is required for proper mitochondrial function, including ATP synthesis and energy delivery for a varied ensemble of cell processes, especially ion transport [28], and this was achieved within 24 h after I/R by the EVs that diffused from the renal subcapsular region to the entire parenchyma, as we recently demonstrated by in vivo biofluorescence approaches [10].

We found that electron leak, i.e., the premature transfer of electrons to O₂ forming O₂^•− [29], is the major early mitochondrial molecular event after non-septic AKI caused by I/R. The rate of O₂^•− production rose by more than 50% after mitochondrial energization by addition of succinate, the respiratory substrate of Complex II (Figure 3 and Figure 4), even though this complex per se is not a source of O₂^•- formation [30,31]. This substrate was chosen for the following reasons. First, during ischemia, the destruction of tubule cells releases fatty acids that are oxidized, forming FADH₂ from FAD at the level of Complex II; when reoxygenation occurs, the electrons it carries are transferred to Complexes III and IV. At the level of Complex III, there is an intense production of O₂^•− [30] that can be carried to the cytosol by voltage-dependent anion channels [32]. The second reason for choosing succinate was that O₂^•− formation at the level of Complex I resulted from reverse transfer of electrons from FADH₂ [33], thus activating a new site for the mitochondrial generation of ROS. The importance of the EVs-induced total blockade of the uncontrolled O₂^•− production and tissue damage emerges from the observation that damage of cytosolic structures is propagated to intact mitochondria, which amplifies the vicious circle of mitochondrial destruction [34].

Of particular interest is the effect of I/R—and of EVs—when O₂^•− formation is assayed in phosphorylating conditions after the addition of 1 mM ADP (Figure 4A). The production of O₂^•− after I/R decreased to less than 15% when compared with the values obtained in the presence of succinate alone (Figure 3D), as a consequence of the dissipation of the Δ_eleqH⁺ during the simultaneous ATP synthesis [33,35]. The complete recovery of control rates of O₂^•− production points to the beneficial influence of EVs in maintaining ATP synthesis and recovering the lower velocity of O₂^•− formation, thus avoiding the amplified damage of I/R through the impairment of ATP synthesis. The similarity between the profiles in phosphorylating conditions and with FCCP (uncoupled) (Figure 4A,C) supports the view that dissipation of Δ_eleqH⁺ is the mechanism on which relies the decrease in the overall rate of O₂^•− formation. The 50% increase of O₂^•− formation after I/R could be attributed to the intensification of the pro-oxidant microenvironment and electron leak in the vicinity of the Fe.S-containing 14 central subunits of Complex I [36], which alternate between successive cycles of oxidation/reduction during the catalysis of electron transport [37]. Possibly, the premature transfer of electrons to O₂ occurs to a lesser extent in Complex III, conceivably because it contains fewer Fe.S centers in its dimeric structure [38].

The view of selective oxidative damage of the electron transport system as a critical early mechanism for mitochondrial dysfunction is reinforced by the observation that O₂^•− production increases by more than 80% in the I/R group in comparison with SHAM when phosphorylation is blocked by oligomycin (Figure 4B), whereas H⁺ leak—which is estimated by the QO₂ in this condition and is considered a central antioxidant defense [19,22]—remains unmodified (Figure 5). Since increased O₂^•− enhances H⁺ leak, thus favoring uncoupling, and uncoupling decreases O₂^•− [39], it may be that the lack of effects of I/R observed in Figure 5 results from protective feedback mutually involving the two processes.

The modest partial recovery of the rate of O₂^•− formation in the group I/R + EVs when compared to SHAM, as well as the high acceleration caused by I/R when the ETS is blocked by Antimycin A at the level of Complex III (Figure 4D), are probably due to the complexity of the pathways for the electron fluxes when this inhibition occurs. When Complex III is blocked, the electrons flow from succinate toward Complex IV through the alternative pathway that requires high spatial organization. It involves reverse transfer of electrons to Complex I [33] that is part of the supercomplex I₁ + III₂ + IV₁ [40,41,42], followed by electron tunneling through the mobile pool of coenzyme Q towards Complex IV. It is likely that under intense pro-oxidant activity and lipid peroxidation, this pool is easily destructured, favoring electron leak, increasing further O₂^•− formation and disrupting repair mechanisms.

The recovery of the normal mitochondrial redox status by EVs seems to be a very early process that is not accompanied by an overall improvement of the injuries caused by I/R. Neither gene expression of KIM-1 nor that of NGAL, two key lesional biomarkers of acute renal injury, was modified by EVs (Figure 6), thus confirming that intense tubular lesions persist. Since 72 h after I/R, the expression of the tubular lesional biomarker KIM-1 was strongly downregulated in rats given EVs [10], and this was accompanied by almost normalization of the high Bax/Bcl2 ratio, an indicator of mitochondrial recovery [43], it may be proposed that early preservation of a physiological redox mitochondrial microenvironment during the first 24 h of reperfusion after ischemia helped to avoid elevated lipid peroxidation and, therefore, activation of Bax-related proteins with later apoptosis. In addition, this ensemble of events occurred in the middle of extensive tubular damage, as demonstrated by the maintenance of I/R high levels of KIM-1 and NGAL after administration of EVs. In the case of NGAL, an opposite view deserves consideration. Its upregulated expression could be viewed as indicative of an antiapoptotic response that can help later recovery, as proposed for an in vivo endotoxin-induced model of AKI in rats [44].

Tubular damage is accompanied by a persistent inflammatory response, as evidenced by the elevated levels of the cytokines IL-6 and TNF-α (Figure 7), whose expression increased in several models of AKI [45,46]. The elevated level of pro-inflammatory TNF-α, which remained unmodified in rats receiving EVs, indicates immune infiltration [47] and is probably responsible for the only partial decrease of IL-6 levels [48]. The amount remaining after 24 h could contribute to reducing lipid peroxidation and oxidative damage [49]. Possibly, the partial selective early decrease of the master regulator IL-6 [50] by EVs results from the release of anti-inflammatory factors contained within the EVs that were demonstrated in our previous proteomic studies [10].

The decrease in O₂^•− promoted by EVs in all respiratory states in the first 24 h was not due to transcriptional upregulation of the enzymes that, sequentially, catalyze the formation of H₂O₂ from O₂^•− (SOD) and its conversion to H₂O and O₂ (CAT), because their levels remain the same as those encountered in I/R (Figure 7D,E). These enzymes became downregulated as the ROS production increased in I/R [51,52,53], thus worsening the prognosis of tubular damage. It may be that the sudden decrease in the pO₂ during the ischemia together with the formation of O₂^•− during reperfusion resulted in the shutdown of transcription promoters [54], as demonstrated for the CAT from normal and tumoral cells from different origins, together with the destabilization of preexisting mRNA [25].

Of particular relevance is the upregulation of HO-1 by I/R, which was not modified in I/R that received EVs (Figure 7C). The mRNA levels 24 h after I/R are similar to those encountered after 72 h [10], when immunohistochemistry analysis revealed small areas positively stained for HO-1. This correlation may indicate that the upregulation of this antioxidant gene is an essential part of the early intrinsic response of kidney cells facing the oxidative stress caused by I/R. This protective response is associated to a previous step of activation of the Nuclear factor (erythroid-derived 2)-like 2 (Nrf2) and translocation to the nucleus, which is ensured by rapid, numerous, and interacting pathways [55]. Due to the role of HO-1 in the regulation of processes such as inflammation and apoptosis [27], its early upregulation without further modification by EVs appears to be located at a crossroad between apoptosis and survival, collaborating with the rapid EVs-mediated protective antioxidant mechanisms in mitochondria discussed above. This view is supported by recent observations demonstrating that administration of anti-ischemic drugs such as meldonium increases the expression of Nrf2 with simultaneous decrease of lipid peroxidation and of the Bax/Bcl2 ratios in a model of I/R in rats [14].

The images and the bar graph in Figure 8 revealed that EVs transferred to renal cells—at least in part—the molecular machinery able to stimulate mechanisms of repair previously characterized by proteomic studies [10]. It is clear, however, that the overall benefit of EVs is partial, because the sum of cells in early and late apoptosis that were cultured with EVs was not different from either the CTR group or the HPX group. The antioxidant effects evidenced by the decrease in O₂^•− formation and upregulation of the HO-1 transcription, early protective mechanisms of EVs, resulted in in a partial decrease in apoptosis in renal cells cultured with EVs after they were subjected to extreme anoxia. It is worth mentioning that the upregulation of these mechanisms is related to cell survival and proliferation [56,57]. One of the key factors shuttled by EVs secreted by MSC is catalase, whose inactivation suppresses the beneficial effects of EVs [58]. Its release after diffusion of EVs into renal parenchyma [10] may be responsible for suppressing the excess of O₂^•− formed during I/R.

4. Materials and Methods

4.1. Ethical Statement

All experimental procedures were approved by the Committee of Ethics in Animal Experimentation of the Federal University of Rio de Janeiro (CEUA) (protocol N° A02/16-61-15) and performed under the Committee’s guidelines, following the Uniform Requirements for Manuscripts Submitted to Biomedical Journals. The animal study is reported in accordance with ARRIVE guidelines [59].

4.2. Isolation and Characterization of Extracellular Vesicles Secreted by Adipose Mesenchymal Stromal Cells

Human adipocyte mesenchymal stromal cells (MSC) were purchased from Lonza (Basel, Switzerland). They were cultured in dishes with Dulbecco´s modified Eagle medium (DMEM) (Life Technologies, Carlsbad, CA, USA) at 37 °C under an atmosphere of 5% CO₂ in air. When cells attained 80% confluence, they were washed twice with phosphate-buffered saline (PBS) at pH 7.2 and starved for 14–16 h in Roswell Park Memorial Institute (RPMI) medium without phenol red (Thermo Fisher Scientific, Waltham, MA, USA). The culture medium was then centrifuged at 1713× g for 20 min at room temperature to remove cell debris and apoptotic bodies, and the supernatant was centrifuged at 100,000× g for 2 h at 4 °C. The recovered sediment contained exosomes and microvesicles (40–100 nm and 100–500 nm, respectively; Figure 1A), i.e., a heterogeneous population of EVs (resuspended in 200 µL RPMI containing 1% dimethyl sulfoxide and immediately stored at −80 °C).

For the morphological characterization of the vesicles by transmission electron microscopy, the pellets obtained as above were resuspended in PBS (pH 7.2) and adhered onto glow-discharged formvar-coated copper grids (300 mesh; EMS, Hatfield, PA, USA) for 10 min. Excess solution was removed using filter paper (Whatman No 1; Merck, Darmstadt, Germany). Then, the grids were negatively stained with 1% (w/v) aurothioglucose in water for 30 s and dried with the same filter paper. The samples were observed on a Tecnai-Spirit electron microscope (Thermo Fisher Scientific, Waltham, MA, USA) operating at 120 kV and equipped with a 2K camera (Veleta, Olympus, Münster, Germany).

The EVs surface antigens CD9, CD63, TSG101, and CD81, were assayed by Western blotting (Figure 1B) using the corresponding antibodies. CD9 (#ab92726, anti-rabbit, at 1:100 dilution) was from Abcam (Cambridge, United Kingdom). CD63 (#sc-5275), TSG101 (#sc-136111), and CD81 (#sc-70803) were from Santa Cruz Biotechnology (Dallas, TX, USA) (anti-mouse in the 3 cases at 1:100 dilution). The secondary antibodies were anti-rabbit (1:2500; GE Healthcare, Cheshire, United Kingdom) and anti-mouse (1:5000; GE Healthcare). The procedures for immunodetection were as recently described [60], except that the protein immunosignals were detected using the ImageQuant LAS 4000 system (GE Healthcare, Chicago, IL, USA).

4.3. Animals and Treatments

We utilized adult male Wistar rats weighing 200–300 g, available at the Central Animal Facilities Health Sciences Center, Federal University of Rio de Janeiro (Rio de Janeiro, Brazil). The animals were housed in a room maintained at 23 ± 1 °C on a 12:12 h light–dark cycle and allowed to acclimatize for 1 week, with free access to a commercial chow for rats (Purina Agribrands, Paulínia, Brazil) and filtered tap water. As stated above, the experimental protocols followed the guidelines and were approved by the local ethical committee.

The rats were randomly divided into 3 groups: (i) sham operated (SHAM), (ii) ischemia/reperfusion (I/R), and (iii) I/R that received subcapsular injections of 2 × 10⁹ EVs (in 150 µL PBS) at the moment of reperfusion (I/R + EVs). The group I/R received the same volume of saline. The I/R protocol was as previously described with slight modifications [61]. Briefly, the rats were anesthetized with 0.2 mL ketamine 10% plus 0.1 mL xylazine 2% (Syntec, Santana de Parnaíba, Brazil) intraperitoneally, and the anesthesia was confirmed by pressing the hind legs. The abdominal cavity was opened, the renal pedicles were dissected, and the renal arteries were clamped for 30 min using a stainless-steel clamp. The pedicle was smoothly manipulated in the SHAM group after visualization without dissection.

The rats received a xylocaine topical ointment and were placed in individual cages under the same previous conditions for 24 h. They were sacrificed by decapitation, and both kidneys were immediately removed. For the experiments intended for studies of ROS formation and O₂ consumption, the mitochondria of the external portion of the cortex (cortex corticis) were isolated as previously described [61] with slight modifications. In this segment of the renal tissue, more than 90% of the cell population corresponds to proximal tubules [62]. Briefly, the cortical fragments were washed twice using a solution containing 250 mM sucrose, 10 mM HEPES-Tris (pH 7.4), 2 mM Na₂EDTA, and 0.15 mg/mL trypsin inhibitor (Sigma-Aldrich, St. Louis, MO, USA), which was used in the following steps. After gentle manual homogenization using a glass Potter-Elvehjem homogenizer (Merck), the total homogenates were centrifuged at 4 °C. First, at 600× g for 5 min to sediment intact cells, cell debris, and nuclei, recovering the supernatants that were immediately centrifuged for 10 min at 12,000× g to obtain mitochondria-enriched sediments. These sediments were resuspended in 5 mL of the above-described solution and centrifuged again at 12,000× g to obtain a washed pellet containing mitochondria, finally resuspended in 300 µL of the above solution and immediately used. For the experiments intended for mRNA analysis and relative quantitative expression, the whole cortex region was processed as recently described [10].

4.4. Renal Cells Cultures and Hypoxia/Reoxygenation Protocol

Immortalized cells from human kidney proximal tubules (HK-2 lineage; ATCC, Manassas, VA, USA) were cultured in 6-well plates (5 × 10⁵ cells per well) in Keratinocyte Serum-Free Medium (K-SFM) (Thermo Fisher Scientific, Waltham, MA, USA) supplied with 5% (v/v) bovine fetal serum (Gibco™, Thermo Fisher Scientific, Waltham, MA, USA), under an atmosphere of 5% CO₂ in air at 37 °C, until they reached 80% confluence. The cells were then divided into 2 lots and cultured in serum-free DMEM, under normoxia (19% O₂, 5% CO₂ in air) or hypoxia (1% O₂, 5% CO₂ in N₂), at 37 °C for 24 h. Next, the plates were cultured for 24 h at 37 °C under normoxia. Some of the cells previously subjected to 24 h hypoxia conditions were co-cultured with 2 × 10⁹ EVs. Thus, we obtained 3 groups: (i) CTR, 24 h normoxia → 24 h normoxia; (ii) HPX, 24 h hypoxia → 24 h normoxia; (iii) HPX + EVs, 24 h hypoxia → 24 h normoxia in co-culture with EVs.

4.5. Flow Cytometry

The 3 populations of HK-2 cells were detached by adding 1 mL of PBS containing 0.05% (w/v) porcine pancreatic trypsin (Trypsin-EDTA, Thermo Fisher Scientific, Waltham, MA, USA). The cell suspensions were centrifuged at 321× g for 5 min at room temperature, and the pelleted cells were resuspended in 100 µL of a solution containing 10 mM HEPES-Tris (pH 7.4), 140 mM NaCl, and 2.5 mM CaCl₂ (solution C from Annexin V kit, Thermo Fisher Scientific, Waltham, MA, USA). The samples (50 µL of the suspension) were supplied with 3 µL of Annexin V (solution A from Annexin V kit) and 3 µL of a solution containing 0.15 mM propidium iodide (10-fold diluted solution B from Annexin V kit). The cytometry analyses were carried out immediately after 15 min at room temperature using an Accuri™ C6 flow cytometer (Accuri Cytometers Inc./Becton Dickinson, Franklin Lakes, NJ, USA). The remaining 50 µL of the cell suspensions were supplied with 50 nM MitoTracker™ Orange (Molecular Probes, Eugene, OR, USA) diluted in DMEM, incubated for 15 min, mixed with 700 µL RPMI, centrifuged at 321× g for 5 min, and washed 3 times with this solution. The final pellet was resuspended in 300 µL RPMI, and immediately observed in the same Accuri™ C6 flow cytometer.

4.6. Analysis of Mitochondrial Morphology and Qualitative Evaluation of the Electrical Potential Difference across the Inner Mitochondrial Membrane from Renal Cells in Co-Culture with Extracellular Vesicles

HK-2 cells were fixed in 2.5% (v/v) glutaraldehyde with 4% (v/v) formaldehyde in 0.1 M cacodylate buffer (pH 7.2) for 24 h. Following a wash in cacodylate buffer 0.1 M, the samples were post-fixed in 1% OsO₄ (w/v) plus 0.8% (w/v) K₄(CN)₆Fe in 0.1 M cacodylate buffer for 40 min, dehydrated in acetone series, and embedded in epoxide resin (Embed 812 Resin, EMS, Hatfield, PA, USA). Ultrathin sections (70 nm) were obtained using a Leica EM UC7 ultramicrotome (Leica, Wetzlar, Germany), collected onto 200 mesh copper grids (EMS), and stained for 20 min in 5% (w/v) uranyl acetate and 5 min in Reynolds’ lead citrate. After, the samples were observed on a Tecnai-Spirit electron microscope (FEI Company, Eindhoven, Netherlands) operating at 120 kV, equipped with a 2k CCD camera (Veleta, Olympus). Profiles of mitochondria in randomly selected cells were evaluated with ImageJ software (U.S. National Institutes of Health, Bethesda, MD, USA), and we compared the average area among the different experimental groups. Due to the polymorphic shape of the cells, we did not measure the volumetric density.

HK-2 cells were also used to qualitatively investigate the mitochondrial energetic state—represented by the transmembrane potential Δψ—after hypoxia and the influence of co-culture with EVs, as recently described [59]. HK-2 cells from the three experimental groups described above were placed in plates of 24 wells, incubated with the fluorophore MitoTracker™ Orange (Molecular Probes), and assayed as follows for immunofluorescence visualization. Since the HK-2 cells do not adhere to glass, the coverslips at the bottom of the plates were washed with 400 µL Attachment Factor Cascade Biologics™ AF (Thermo Fisher Scientific, Waltham, MA, USA) and immediately dried at 37 °C for 30 min. The cells (8 × 10⁴ per well, suspended in 500 µL DMEM without serum) were placed onto 13 mm diameter coverslips deposited at the bottom of the wells, washed twice with PBS, and supplied with 500 µL of the fluorophore diluted in a modified Krebs solution containing 120 mM NaCl, 4 mM KCl, 1.4 mM MgCl₂, 2.5 mM CaCl₂, 6 mM glucose, and 10 mM HEPES (pH adjusted to 7.4 with Tris), and incubated for 20 min at 37 °C. Then, the cells were washed 3 times for 5 min with the same solution and fixed with paraformaldehyde 4% (w/v) for 15 min at room temperature, sheltered from the light. After removal of the fixative, the cells were rewashed 3 times for 5 min using the same solution. The coverslips were carefully removed from the plate with the aid of a small forceps, supplied with 20 µL DAPI (5 µg/mL in a PBS-glycerol 40% (v/v) solution at pH 7.4 adjusted with Tris). The coverslips were mounted onto glass slides, sealed with colorless nail polish, and stored at −20 °C for 24 h. The images were collected using AxioVision 4.8.2 software in an ApoTome microscope (ApoTome Axion Imager.M2, Carl Zeiss Inc., Jena, Germany) at 554 nm (excitation) and 576 nm (emission).

4.7. Time Course of Reactive O₂ Species Formation

The generation of mitochondrial O₂^•− was quantified by measuring the changes in fluorescence when the fluorescent resorufin is formed by the peroxidase (2 U/mL)-catalyzed oxidation of the Amplex Red™ probe (5 mM, Thermo Fisher Scientific, Waltham, MA, USA), after dismutation of O₂^•− into H₂O₂ in the presence of excess superoxide dismutase (60 U/mL, Sigma-Aldrich), as recently described [63]. Generation of O₂^•− was evaluated at different respiratory states after substrate–uncoupler–inhibitor–titration (SUIT): (i) 10 mM succinate (substrate for Complex II); (ii) 0.1 mM ADP; (iii) 1 mM ADP; (iv) 0.2 µg/mL oligomycin (to inhibit ATP formation and to stimulate H⁺ leak); (v) 1.5 µM carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone (FCCP) to completely dissipate the H⁺ gradient; (vi) 2.5 µM antimycin to inhibit Complex III. The basic reaction medium before SUIT contained: 10 mM HEPES-Tris at pH 7.4, 320 mM mannitol, 4 mM KH₂PO₄, 4 mM MgCl₂, 0.08 mM Na₂EDTA, 1 mM EGTA (free acid), and 1 mg/mL fatty acids free BSA. The final pH was adjusted to 7.4 by adding Tris; the mitochondrial protein concentration was 0.1 mg/mL. Recordings were acquired at 563 nm (excitation) and 587 nm (emission). Calibration curves were obtained using successive pulses of 10 nM H₂O₂, thus allowing the conversion of arbitrary fluorescence units into pmol H₂O₂/mg per min. The stoichiometry 1 O₂^•−: 1 H₂O₂ allowed quantification of O₂^•− formation rate.

4.8. Oxygen Consumption by Isolated Mitochondria from Kidney Proximal Tubule Cells

We measured oxygen consumption (QO₂) by isolated mitochondria using a high-resolution O₂ electrode (Oxygraph-2K, OROBOROS Instruments, Innsbruck, Austria) at 37 °C in the solution described above for mitochondrial O₂^•− determination, except that the components required to detect H₂O₂ formation were omitted, and 0.06 mg/mL mitochondrial protein was used. The QO₂ assays were run in parallel with H₂O₂ assays. All mitochondrial preparations were assayed to determine the respiratory control ratio (RCR) and, therefore, to evaluate the coupling between electron fluxes and ATP synthesis. The RCR was calculated from the ratio between the QO₂ in the presence of 1 mM ADP and the QO₂ after the addition of oligomycin. Despite differences in the absolute values of QO₂ (lower in I/R mitochondria), the RCR averaged 3.0 in all groups, with an interassay variation coefficient of ~10%. The QO₂ after successive additions of oligomycin and Antimycin A was used to estimate the H⁺ leak from the mitochondrial space back to the matrix [20,21,22].

4.9. RNA Isolation, Reverse Transcription and Real Time Quantitative Polymerase Chain Reaction (qRT-PCR)

Small fragments of kidney cortex were suspended in 500 µL of Lysis/binding buffer (miRNA isolation kit mirVana™, Thermo Fisher Scientific, Waltham, MA, USA) in RNAse-free microtubes and homogenized using the dissociator TissueLyser LT provided with a 5 mm diameter bead (Qiagen, Hilden, Germany). After 7 min of intense agitation, 400 µL of the homogenate was mixed with 40 µL of miRNA homogenate additive (miRNA isolation kit mirVana™), vigorously vortexed, and then placed on ice for 10 min. The suspensions were then supplied with the same volume of chloroform, vortexed again at 12,000× g for 5 min at room temperature, to recover supernatants of 1 mL that were mixed with 100% ethanol (1.25 mL ethanol:1 mL sample), gently homogenized with a micropipette and centrifuged at 12,000× g for 15 s at room temperature using microtubes supplied with the filter provided by the kit (miRNA isolation kit mirVana™). After removing the liquid, the filters were first washed with 700 µL of the miRNA wash solution 1, centrifuged for 15 s at 12,000× g at room temperature, and then washed twice using 500 µL of the miRNA wash solution 2/3 (both solutions from the kit mentioned above). Finally, the filters were immersed in 100 µL of RNAse-free H₂O at 95 °C, immediately centrifuged for 30 s at 13,000× g, and the liquids, after removal of the filters, were stored at −80 °C.

The mRNAs were obtained using 10 µL of the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems™, Foster City, CA, USA)—with components freshly mixed—and 10 µL samples containing 20 ng/µL mRNA. The RNA of these samples was quantified using the NanoDrop™ ND-1000 (Thermo Fisher Scientific, Waltham, MA, USA), as recently described [10]. The cDNAs were synthesized in the thermocycler PCR Thermal Cycler (Applied Biosystems™), with cycles of 10 min at 25 °C, 120 min at 37 °C, 5 min at 85 °C, and 1 min at 4 °C, before storage at −20 °C. Negative controls without reverse transcriptase were carried out in parallel with each run.

The reverse transcription followed by the real-time quantitative polymerase chain reaction was performed in a single step using 10 µL of Power SYBR Green® PCR Master Mix (Applied Biosystems™) and 10 µL of the cDNA-containing solution and the primers (0.25 ng/µL and 100 nM, respectively). The sequence-specific oligonucleotides [10] were from Eurofins Genomics (Ebersberg, Germany) (

Table 1. List of the primers used in qRT-PCR analysis.

mRNA	Sequence
rSOD1 F1	AAGAGAGGCATGTTGGAGACC
rSOD1 R1	ACGGCCAATGATGGAATGCT
rCAT F1	CAGCTCCGCAATCCTACACC
rCAT R1	GGACATCGGGTTTCTGAGGG
rIL6 F1	AAGCCAGAGTCATTCAGAGC
rIL6 R1	GTCCTTAGCCACTCCTTCTG
rTNFA F1	CTTCTCATTCCTGCTCGTGG
rTNFA R1	TGATCTGAGTGTGAGGGTCTG
rKIM1 F1	ACCTGATCAGACAGAGTGTGC
rKIM1 R1	ATCTACAGAGCCTGGAAGAAGCA
rHO-1 F1	AGGTGCACATCCGTGCAGAG
rHO-1 R1	CTTCCAGGGCCGTATAGATATGGTA
rNGAL F1	GGGCTGTCCGATGAACTGAA
rNGAL R1	CATTGGTCGGTGGGAACAGA
rGAPDH F1	GCCAAAAGGGTCATCATCTC
rGAPDH R1	GGCCATCCACAGTCTTCT

). Their amplifications were followed using the ViiA™ 7 Real-Time System (Applied Biosystems™) after a stage of 10 min to reach 95 °C, followed by cycles of 15 min at 95 °C, 60 min at 60 °C, and 15 min at 95 °C.

4.10. Statistical Analysis

The mean values of the different parameters investigated in the 3 experimental groups of rats and in the 3 HK-2 cells assays were analyzed using one-way ANOVA followed by Tukey’s test. In the case of RQ data, the mean of each SHAM group was taken as 1.0, and the individual values from the 3 groups were expressed as a fraction or as a multiple of this value [63,64,65]. This allowed the SEM of the unitary value of SHAM reference to be calculated.

5. Conclusions

We demonstrated that the early and central mechanism by which EVs protect renal structure and function after I/R is the decrease of O₂^•− production to control levels and, therefore, the normalization of the mitochondrial redox environment. The antioxidant components of EVs are central in the preservation mechanisms that, with the aid of the upregulated antioxidant HO-1, depress early processes of mitochondrial damage and cell death after I/R. The proposed molecular events elicited by EVs are depicted in Figure 9.