Early Effects of Extracellular Vesicles Secreted by Adipose Tissue Mesenchymal Cells in Renal Ischemia Followed by Reperfusion: Mechanisms Rely on the Restoration of the Redox Tissular Environment

: Acute kidney injury (AKI) caused by ischemia followed by reperfusion (I/R) is characterized by intense anion superoxide (O 2•- ) production and oxidative damage. We investigated whether extracellular vesicles secreted by adipose tissue mesenchymal cells (EVs) administrated during reperfusion mitochondrial O 2•- formation after I/R. We used Wistar rats submitted 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 3rd group was of the false-operated rats (SHAM). Mitochondria were isolated from proximal tubule cells and immediately used. Amplex Red™ was used to measure mitochondrial O 2•- formation and MitoTracker® Orange to evaluate Δ  . In vitro studies were carried out by using human renal proximal tubular cells (HK-2) co-cultured or not with EVs under hypoxia conditions. Administration of EVs restored O 2•- formation to SHAM levels in all mitochondrial functional conditions. The expression of catalase and superoxide dismutase 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 recover the renal structure and function after I/R are related to the normalization of the mitochondrial redox environment. The intravesicular catalase is central in the preservation mechanisms that, with the aid of the upregulated antioxidant HO-1/Nuclear factor erythroid 2-related factor 2 system, depress early processes of cell death after I/R and open new vistas for the treatment of AKI.


Introduction
Acute kidney injury (AKI) is one of the more 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 higher than 20 billion dollars per year [3]. Facing the controversy regarding several non-specific available treatments nowadays and the severity of the outcomes [4,5], the search for therapeutic alternatives is growing worldwide. In the last few decades, the replacements started focusing on cell therapies to prevent or slow the progression from AKI to chronic kidney disease (CKD). In the beginning, the hope was centered on renal resident adult stem cells [6,7], which 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 reactive O2 species (ROS) occurs in the ischemic kidney and reoxygenation [12]. The exacerbated production of ROStogether with an important inflammatory componentis a central process in the physiopathogenesis of AKI-associate lesions as a consequence of I/R, which initiates a cascade of molecular events and aberrant signaling that culminates in tubular destruction [13], severe impairment of renal function and high mortality or frequent progress to CKD and endstage of renal disease [14].
The early moments of the AKI onset and the response to treatments are crucial to define the evolution and prognosis [15]. Therefore, knowledge regarding the molecular and cellular events in these early and short times could support the search for appropriate therapies and their targets, especially the mechanisms altered after the burst and continuity of ROS formation. Facing the effect of EVs secreted by adipose MSC, which 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 translation of renoprotective effects occur.
Bilateral clamping of renal arteries for 30 min was the in vivo model of I/R we used in rats. 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

Characterization of extracellular vesicles secreted by adipose tissue mesenchymal stromal cells
The mesenchymal stromal cells (MSC) secreted 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 electron transmission microscopy (TEM) representative images. The images at higher magnification allow detecting a well-delimited membrane that encircles non-homogeneous content. In Figure 1B, the detection of surface EVs markers [16] confirm the nature of the vesicles seen in panel A, with enrichment of exosomes markers such as tetraspanins and endosomal related proteins.

Mitochondria submitted 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 (Δ) are key for the potential 5 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 [17]. 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 co-cultured with EVs (on the right). The bar graph in Figure 2B quantifies the mitochondrial area showing its decrease in hypoxia condition and the recovery of the CTR mean value when the renal cells were cultured with EVs. When HK-2 cells previously exposed to 1% O2 during 24 h were co-cultured for an equal period of time with EVs, the CTR mitochondrial morphology was partially recovered (in the right). Figure 2D presents the distribution of the number of events normalized to mode, 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, highlighted the complete overlap in the intensity measured in CTR, HPX and HPX+EVs conditions, indicating that despite the mitochondrial morphological alterations the Δ is preserved under the assayed hypoxia 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 could be the target of potential beneficial effects of EVs, in early stages after hypoxic injuries.

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 have been measured in SHAM, I/R, and I/R+EVs rats after 24 h of reperfusion.  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). The 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 lower velocity, there was no difference among the 3 groups (average of 1,000 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 3 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, though at higher levels, was encountered when ATP synthesisand respirationwas 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 [18], was investigated. Proton leak in mitochondria (i.e. the return of H + from the interspace to the matrix through pathways different from the FoF1-ATPsintase) 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 [19][20][21]. Figure 5 shows that H + leak was similar in the 3 groups and, therefore, that the ischemic episode followed by reoxygenation did not affect the pathways through which H + flow back to the matrix through pathways different from the FoF1-ATPsintase. 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 2 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 [22,23]. KIM-1 was barely detectable in SHAM conditions and increased approximately 20 times 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 times after I/R without influence of EVs treatment ( Figure 6B). 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 expression of 2 pro-inflammatory cytokines: Interleukin-6 (IL-6) ( Figure 7A) and Tumor Necrosis Factor-alpha (TNF-) ( 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, once IL-6 dropped by half compared to I/R and TNF- 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.

The early anti-oxidative responses after EVs administration coexist 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 [24]. Thus we investigated the expression: (i) of a master regulator of cellular redox homeostasis and mitochondrial function [25,26], the heme oxygenase-1 (HO-1), and (ii) of 2 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 8 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 Figures 3 and 4, without any influence of the EVs content components. The profile was the opposite for CAT ( Figure 7D) and SOD ( Figure 7E) transcription, which was downregulated in I/R and, again without any EVs effects, indicating that I/R partially inhibited transcriptional processes unable to be restoredin the early phase of the damageby the factors that the vesicles carry.

Apoptotic processes are partially reverted when renal proximal tubule cells are cocultured 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 in the increased redox stress.
HK-2 cells submitted 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 3 times; Figure   8A, B, E), and by 100% the number of cells in an early stage of apoptosis (ANX + ) ( 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).

Discussion
The main finding in this study was that subcapsular administration of MSCderived 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 a proper mitochondrial function, including ATP synthesis and energy delivery for a varied ensemble of cell processes, especially ion transport [27], 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 biofluorescense approaches [10].
We found that electron leak, i.e. the premature transfer of electrons to O2 forming O2 •- [28] is the major early mitochondrial molecular event after non-septic AKI caused by I/R.  [29] that can be carried to the cytosol by voltage-dependent anion channels [31]. The second reason was that O2 •formation at the level of Complex I resulted from reverse transfer of electrons from FADH2 [32], 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 [33].
Of particular interest is the effect of I/Rand of EVswhen O2 •formation was 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 eleqH + during the simultaneous ATP synthesis [32,34].  [35], which alternate between successive cycles of oxidation/reduction during the catalysis of electron transport [36]. Possibly, the premature transfer of electrons to O2 occurs to a lesser extent in Complex III, possibly because it contains fewer Fe.S centers in its dimeric structure [37].
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 was blocked by oligomycin ( Figure 4B), whereas H + leak which is estimated by the QO2 in this condition and is considered a central antioxidant defense [18,21] remained unmodified ( Figure 5). Since increased O2 •enhances H + leak, thus favoring uncoupling, and uncoupling decreases O2 •- [38], 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 only 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), is 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 [32] that is part of the supercomplex I1+III2+IV1 [39][40][41], 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 impairing 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 expression of KIM-1 nor the NGAL, two key lesional biomarkers of acute renal injury, were 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 Bcl2/Bax ratio, an indicator of mitochondrial recovery [42], 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. And, also, 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 could help later recovery, as proposed for an in vivo endotoxin-induced model of AKI in rats [43].
Tubular damage is accompanied by the persistent inflammatory response as evidenced by the elevated levels of the cytokines IL-6 and TNF- (Figure 7), which expression increased in several models of AKI [44,45]. The elevated pro-inflammatory TNF-, which remained unmodified in rats receiving EVs, indicates immune infiltration [46] and is probably responsible for the only partial decrease of IL-6 levels [47], which remaining amount after 24 h could contribute to reducing lipid peroxidation and oxidative damage [48]. Possibly, the partial selective early decrease of the master regulator IL-6 [49] 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 •in all respiratory states in the first 24 h promoted by EVs was not due to transcriptional up-regulation of the enzymes that, sequentially, catalyzes the formation of H2O2 from O2 •-(SOD) and its conversion to H2O and O2 (CAT), because their levels remain the same encountered in I/R ( Figure 7D, E). These enzymes became downregulated as the ROS production increased in I/R [50][51][52], 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 [53], as demonstrated for the CAT from normal and tumoral cells from different origins, together with the destabilization of preexisting mRNA [24].
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 [54]. Due to the role of HO-1 in the regulation of processes such as inflammation and apoptosis [26], its early upregulation without further modification by EVs appears to be central in the crossroad between apoptosis and survival collaborating with the rapid EVs-mediated protective antioxidant mechanisms in mitochondria discussed above.
The images and the bar graph in Figure 8 revealed that EVs transferred to renal cellsat least in partthe molecular machinery able to stimulate mechanisms of repair previously characterized by proteomic studies [10]. The antioxidant effects evidenced by the decrease in O2 •formation and upregulation of HO-1, early protective mechanisms of EVs, resulted in decreased apoptosis in renal cells cultured with EVs after they were submitted to extreme anoxia. It is worth mentioning that the upregulation of these mechanisms is related to cell survival and proliferation [55,56].
One of the key factors shuttled by EVs secreted by MSC is the catalase, whose inactivation suppresses the beneficial effects of EVs [57]. Its release after diffusion of Kingdom) and anti-mouse (1:5,000; GE Healthcare). The procedures for immunodetection were as recently described [59], except that the protein immunosignals were detected using the ImageQuant LAS 4000 system (GE Healthcare, Chicago, IL, USA).

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 and maintained at 23 ± 1 o C on a 12:12 h light-dark cycle and allowed to acclimatize during 1 week, with free access to a commercial chow for rats (Purina Agribands, 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) false operated (SHAM), (ii) ischemia/reperfusion (I/R), and (iii) I/R that received subcapsularly 2  10 9 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 [60]. 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 posterior 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 reactive O2 species (ROS) formation and O2 consumption, the mitochondria of the external portion of the cortex (cortex corticis) were isolated as previously described [60] with slight modifications. In this segment of the renal tissue, more than 90% of the cell population corresponds to proximal tubules [61]. 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 of 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 o 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]. HK-2 cells were also used to qualitatively investigate the energetic mitochondrial state -represented by the transmembrane potential Δafter hypoxia and the influence of co-culture with EVs, as recently described [59]. HK-2 cells from the

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 o C in the solution above descript 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 Negative controls without reverse transcriptase were carried out in parallel with each run.
The reversal 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)    Means were compared using one-way ANOVA followed by Tukey´s test. ****P < 0.0001; NS: not significant.     and HPX+EVs (n = 9). *P < 0.05; **P < 0.01; ***P < 0.001; NS: not significant (one-way ANOVA followed by Tukey´s test. EVs that, after subcapsular administration and diffusion into the renal parenchyma [10], reach the tubular segments injured by I/R. By releasing several factors, including the catalase they carry [57], they contribute to maintaining the normal local redox state existing in the absence of injury. With mitochondrial and cytoplasmic redox homeostasis restored, the mitochondrial processes required for ATP synthesis are preserved and, therefore, the appropriate ATP supply is preserved for the transport demands and maintenance of tubular structures.