Nrf2 Activation Does Not Protect from Aldosterone-Induced Kidney Damage in Mice

Nuclear factor erythroid 2-related factor 2 (Nrf2) is downregulated in chronic kidney disease (CKD). Activation of Nrf2 might be a therapeutic option in CKD. Here we investigate the effect of Nrf2 activation on aldosterone (Aldo)-induced renal injury. Wild-type (WT) mice, transgenic Keap1 hypomorphic (Nrf2ꜛ, genotype results in upregulation of Nrf2 expression) mice and WT mice treated with the Nrf2 activator sulforaphane (Sulf) received Aldo for 4 weeks. In Aldo-treated mice, kidneys were significantly heavier and pathologically altered, reflected by increased urinary albumin levels and tissue damage. In Nrf2ꜛ-Aldo mice the tubule damage marker NGAL was significantly decreased. Increased oxidative damage markers (8-OHdG, 15-isoprostane F2t) were measured in all Aldo-treated groups. Aldo-increased Nrf2 amounts were mainly found in the late tubule system. The amount of phosphorylated and thus putatively active Nrf2 was significantly increased by Aldo only in WT mice. However, expression of Nrf2 target genes NQO1 and HO1 was decreased in all Aldo-infused mice. GSK3β, which promotes Nrf2 degradation, was significantly increased in the kidneys of Aldo-treated WT mice. Neither genetic nor pharmacological Nrf2 activation was able to prevent oxidative injury induced by Aldo, probably due to induction of negative regulators of Nrf2.


Introduction
The main risk factors for the development of chronic kidney disease (CKD) are diabetes and hypertension. Hypertension is present in approximately 70% of individuals with moderate CKD in Germany [1]. In hypertension, the renin-angiotensin-aldosterone system (RAAS) is almost always activated, and is usually also causally involved in the development of hypertension. Therefore, treatment with RAAS inhibitors such as angiotensin converting enzyme inhibitors or mineralocorticoid receptor blockers is standard therapy for achieving the first therapeutic goal in hypertensive CKD patients, the lowering of systolic blood pressure to 120 mmHg [2]. Beyond that, the therapeutic options for treating CKD and especially stopping the progression of CKD are limited.
Several clinical studies focused on investigating the effects of activators of the transcription factor nuclear factor erythroid 2-related factor 2 (Nrf2) on progression of mainly diabetes-induced CKD [3,4]. This is based, on the one hand, on results of animal studies in which Nrf2 activation led not only to improved renal morphology, but also to an increase in renal function [5]. On the other hand, an increased glomerular filtration rate was observed in clinical antitumor studies after intake of the Nrf2 activator bardoxolone methyl (CDDO-Me) [6]. However, initial larger studies in CKD patients showed serious cardiac side effects of CDDO-Me, increased proteinuria and no arrest of CKD progression [3,7]. Recent studies include only patients without risk factors for heart failure [8]. It was suggested that the actual effects of Nrf2 activators in the kidney are not yet sufficiently understood for use of these compounds in the clinic [9,10].
In rats with aldosterone (Aldo)-induced hypertension and moderate kidney damage we found that the use of another Nrf2 activator, sulforaphane (Sulf), protected kidneys from

Double Staining
To localize pNrf2 in the kidney, double staining was performed to identify pNrf2 positives in different types of kidney cells. Staining of the first antigen (pNrf2) was performed as described above using diaminobenzidine (DAB) as a chromogen. After visualization of antibody binding, the protocol was repeated with an antibody against calbindin (1:200, #2173, Cell Signaling, Herts, UK) to identify distal tubular cells and cells of the early collecting duct. Proximal tubular cells were identified by the presence of the brush border, whereas glomeruli were identified by their capillary tuft (blue circles). Fifty glomeruli were analyzed per animal. The late collecting duct was identified by the absence of positive calbindin staining and brush border. The VECTOR ® VIP Peroxidase Substrate Kit was used to visualize the second antigen (SK-4600, Vector Lab, Burlingame, CA, USA), which produced a purple stain. Sections were counterstained with hematoxylin and dehydrated in descending alcohol concentrations. Pictures were taken at 400-fold magnification. The ratio of positive to negative nuclei for pNrf2 on 10 visual fields was assessed via ImageJ [30] by counting only nuclei positive for the specific kidney cell identifier.

Quantification of 8-OhdG and 15-Isoprostane F 2t in Urine
8-hydroxy-2 -deoxyguanosine  in urine was quantified using the DNA Damage ELISA Kit (StessMarq Biosciences Inc., Victoria, BC, Canada) according to the manufacturer's protocol. Urinary 15-isoprostane F 2t levels were determined with the Urinary Isoprostane ELISA Kit (EA85, Oxford Biomedical Research, Rochester Hills, MI, USA) following the manufacturer's protocol. 15-isoprostane F 2t was related to urinary creatinine.

Preparation of Cytosolic and Nuclear Protein Fractions
To detect translocation of specific proteins into the nucleus, a cytosolic and a nuclear fraction were prepared. To the cytosolic (10 mM Tris-HCl, 50 mM NaCL, 500 mM sucrose, 0.1 mM EDTA, and 0.5% Triton-X) and nuclear lysis buffers (10 mM Tris-HCl, 500 mM NaCL, 0. 2 mM EDTA, 1% Nonident P40, and 1% Tergitol) were added 1 mM DTT, 1 mM PMSF, 1 mM sodium orthovanadate and 1× protease inhibitor cocktail (Sigma, Taufkirchen, Germany) just before the start of the experiment. All work was performed at 0-4 • C. For extraction, frozen kidney tissue was first minced and homogenized with a pestle in cytosolic lysis buffer for 2 min and then incubated for an additional 5 min. The homogenate was centrifuged at 1000× g for 10 min to separate the nuclear proteins (pellet) from the cytosolic proteins (supernatant). After a wash step, the supernatant representing the cytosolic fraction was removed and stored on ice. After another centrifugation (1000× g, 4 min), the remaining pellet was resuspended in 300 µL of cytosolic lysis buffer. After washing, the pellet was resuspended in 150 µL of nuclear lysis buffer and sonicated. After centrifugation at 14,000× g for 10 min, the supernatant representing the nuclear fraction was removed. Fifty micrograms of the protein lysate of the cytosolic fraction and 10 µg of the protein lysate of the nuclear fraction were used for a Western blot. The cytosolic and nuclear fractions were stored at −80 • C until further use.

Quantitative RT-PCR
mRNA was extracted from 20-40 mg frozen kidney tissue using the QIAcube (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Isolated mRNA was transcribed into cDNA using the High Capacity cDNA Reverse Transcription Kit (Thermo Fisher, Waltham, MA, USA). Quantitative Real-Time PCR (qRT-PCR) was performed with 20 µg cDNA and the SensiMix SYBR Hi-ROX Mastermix (Bioline GmbH, Luckenwalde, Germany) using the CFX96 Real-Time System (BIO-RAD, Hercules, CA, USA). Primer sequences utilized for gene expression analysis are listed in Table S2. Relative expression levels of target genes were normalized to the housekeepers GAPDH and β-actin and calculated using the comparative CT method with the analysis software Bio-Rad CFX Manager 3.1 (BIO-RAD, Hercules, CA, USA).

Statistics
Data from seven to eight animals per group are expressed as mean ± standard error of the mean (SEM). GraphPad Prism 6 (GraphPad Software, La Jolla, CA, USA) was used for statistical analyses. Data were tested for normal distribution using the Kolmogorov-Smirnov test with Dallal-Wilkinson-Liliefor p values. Groups of each mouse strain were tested for significance (normal distribution) among themselves or against control using analysis of variance (one-way ANOVA) followed by Tukey's correction. Values not normally distributed were tested for significance using the Kruskal-Wallis test followed by Dunn's multiple comparison test. Significant differences between mouse strains (controls or Aldo-infused animals) were tested for significance using 2-way ANOVA followed by Tukey's correction. A p value ≤ 0.05 was considered significant.

Characteristics of Keap1 Hypomorphic (Nrf2Ţ) Mice
Hypomorphic Keap1 (Nrf2Ţ) mice showed significantly decreased expression of Keap1 in the kidney at both the mRNA and protein levels (Figure 1a,b). This did not affect Nrf2 gene expression (Figure 1c), which was not expected because Keap1 regulates Nrf2 primarily posttranslationally. However, an increased amount of Nrf2 protein could be measured in the mice (Figure 1d).

Characteristics of Keap1 Hypomorphic (Nrf2ꜛ) Mice
Hypomorphic Keap1 (Nrf2ꜛ) mice showed significantly decreased expression of Keap1 in the kidney at both the mRNA and protein levels (Figure 1a,b). This did not affect Nrf2 gene expression (Figure 1c), which was not expected because Keap1 regulates Nrf2 primarily posttranslationally. However, an increased amount of Nrf2 protein could be measured in the mice (Figure 1d).

Blood Pressure Changes and Clinical Characteristics
Only the wild-type (WT) mice infused with 125 µg/kg Aldo per day showed a temporary significant increase in blood pressure up to 155 mmHg until the 15th day after implantation (Figure 2a). After this time point, however, the blood pressure decreased to the level of the WT control (C) group by the end of the experiment. In the Aldo-infused WT mice treated additionally with Sulf in the drinking water, blood pressure did not increase above the level of the WT control. Quantification of the aldosterone levels in the serum of the six mouse groups showed a significant increase in all Aldo-treated groups ( Figure 2b).

Blood Pressure Changes and Clinical Characteristics
Only the WT mice infused with 125 µg/kg Aldo per day showed a temporary significant increase in blood pressure up to 155 mmHg until the 15th day after implantation (Figure 2a). After this time point, however, the blood pressure decreased to the level of the WT control (C) group by the end of the experiment. In the Aldo-infused WT mice treated additionally with Sulf in the drinking water, blood pressure did not increase above the level of the WT control. Quantification of the aldosterone levels in the serum of the six mouse groups showed a significant increase in all Aldo-treated groups (Figure 2b).  A difference in body weight could not be detected between the six different groups (Table 1). However, compared to the WT mice, the Nrf2ꜛ mice had a higher kidney weight already basally. Aldo infusion in the WT mice and in the Nrf2ꜛ mice resulted in a significant increase in relative kidney weight. Aldo treatment in the Nrf2ꜛ mice resulted in an even greater increase in kidney weight, which was even significantly higher than in the Aldo-infused WT mice. Additional treatment of the Aldo-exposed WT mice with Sulf also resulted in an increase in kidney weight compared with the WT control and Aldo-only treatment. The relative heart weights of WT control and Nrf2ꜛ control mice did not differ basally. Sulf had no impact on relative heart weight, but Aldo infusion resulted in a significant increase in relative heart weight compared with WT control in both WT and Nrf2ꜛ A difference in body weight could not be detected between the six different groups (Table 1). However, compared to the WT mice, the Nrf2Ţmice had a higher kidney weight already basally. Aldo infusion in the WT mice and in the Nrf2Ţmice resulted in a significant increase in relative kidney weight. Aldo treatment in the Nrf2Ţmice resulted in an even greater increase in kidney weight, which was even significantly higher than in the Aldo-infused WT mice. Additional treatment of the Aldo-exposed WT mice with Sulf also resulted in an increase in kidney weight compared with the WT control and Aldoonly treatment. The relative heart weights of WT control and Nrf2Ţcontrol mice did not differ basally. Sulf had no impact on relative heart weight, but Aldo infusion resulted in a significant increase in relative heart weight compared with WT control in both WT and Nrf2Ţmice. Collagen deposition was significantly increased in Aldo-treated Nrf2 animals, while Aldo-treated WT animals showed no change at all. Table 1. Overview of weight, renal function and histological parameters of WT and Nrf2Ţmice after 28 days of treatment. Shown in each case is the mean ± SEM. The parameters TSI, MSI and GSI were semi-quantitatively evaluated in the renal cortex of Aldo-treated WT and Nrf2Ţmice and their controls. For every parameter, 20 visual fields or 50 glomeruli per animal were analyzed. n = 7 for the Nrf2Ţ-C group for the determination of albumin, KIM-1 and NGAL, n = 8 for all other parameters and groups. Aldo: aldosterone, GSI: glomerular sclerosis index, KIM: kidney injury molecule-1, C: control, MSI: mesangiolysis index, NGAL: neutrophil gelatinase-associated lipocalin, Nrf2: nuclear factor erythroid 2-related factor 2, Sulf: sulforaphane, TSI: tubulointerstitial sclerosis index, WT: wild-type.

Histopathological Changes in the Kidney and Renal Function
During organ removal, hydronephrosis of the right kidney was observed in two of eight mice in the Nrf2Ţcontrol group and in four of eight mice in the Nrf2Ţ-Aldo-infused group (these kidneys were not used for further analyses). Significant pathologic damage to the tubular system was recorded in Aldo-treated WT animals, and included infiltration of inflammatory cells, atrophy of basement membranes and fibrotic changes ( Table 1, TSI). Nrf2Ţmice were not affected, neither basally nor by Aldo. Aldo-treated WT mice also showed a significant increase in mesangiolysis (Table 1, MSI), whereas all mice, including those treated only with Sulf and the Nrf2Ţcontrol, presented increased glomerular sclerosis ( Table 1, GSI).
Aldo treatment resulted in a significant increase in drinking and urine volume (Table 1) in all groups compared with the matched controls and the control of the respective other mouse strain (WT-C vs. Nrf2Ţ-Aldo and Nrf2Ţ-C vs. WT-Aldo). Creatinine in serum was not different between the groups. Aldo alone did not affect creatinine clearance. However, treatment with Sulf resulted in significantly lower creatinine clearance compared with WT control and combination treatment (WT-Aldo + Sulf) and the Nrf2Ţmice showed a slightly lower creatinine clearance basally. Albumin was used as a marker for glomerular damage in this experiment, while urinary NGAL and KIM-1 served as markers of tubular damage. All Aldo-treated animals had significantly elevated albumin levels, with the Nrf2Ţ-Aldo group presenting the highest levels. Aldo infusion resulted in a tenfold increase in KIM-1 to creatinine ratio in WT mice only, whereas Nrf2Ţmice treated with Aldo had significantly lower levels. All Aldo-treated animals had significantly elevated NGAL levels, with the Nrf2Ţ-Aldo group here presenting the lowest levels. In the co-treated mice, Sulf treatment had no effect at all on the three markers.

Oxidative Stress Markers
Systemic oxidative stress was determined by the excretion of 15-isoprostane F 2t , a marker of lipid peroxidation, and 8-OHdG, a marker of oxidative damage to nucleic bases, in the collection urine of mice. 15-isoprostane F 2t was significantly increased in all Aldo-treated animals. Higher 8-OHdG excretion was detected in WT and Nrf2Ţmice compared with the WT-Sulf group and the Nrf2Ţ-C group, respectively (Figure 3a,b). As another reasonably stable marker for oxidative stress, the DNA damage marker γH2AX was quantified on paraffin sections of the kidneys. Here, a localization of damage in renal cortex and medulla is feasible. Compared with the WT-C group, there were approximately twice as many γH2AX-positive nuclei in the renal cortex of Sulf-treated WT mice and in the Nrf2Ţ-C group (Figure 3c,d). Aldo resulted in a significant increase of 4-and 6-fold in γH2AX-positive nuclei in all three Aldo-treated groups compared with the respective control groups. In the medulla, there was no elevated damage after Aldo treatment. As a possible source of oxidative stress, we examined the change in expression of NADPH oxidase isoform 2, which was significantly upregulated in the two Aldo-treated WT groups (Figure 3e).

Nrf2 Expression, Activation and Target Gene Regulation
No significant difference in Nrf2 abundance in nuclear and cytosolic kidney samples was observed between groups, except for the Aldo-treated Nrf2Ţgroup (Figure 4a,b). Here, the amount of cytosolic Nrf2 was significantly increased compared with the WT control and the WT-Aldo group. To obtain better local resolution of Nrf2 levels, renal tissue was immunohistochemically stained with an antibody against Nrf2 (Figure 4c-e). A 4-to 8-fold higher expression of Nrf2 was observed in the medulla compared with the cortex, but significantly altered abundance of Nrf2 was observed only in the cortex, increased in all Aldo-treated groups.
When the phosphorylated form of Nrf2, pNrf2, which is ostensibly regarded as the active form of the transcription factor, is considered, we found its significant increase only in the Aldo-treated Nrf2Ţgroup in the cytosolic fraction (Figure 5a,b). In contrast, in the nuclear and the cytosolic extracts of the Aldo-treated WT mice, interindividual variations prevented the finding of significant changes of pNrf2. Here, immunohistochemical staining revealed a significantly increased amount of pNrf2 only in the cortex of WT-Aldo mice (Figure 5c,d). An even more detailed localization of pNrf2 according to kidney structure or cell type (Figure 5e-i) showed a significant increase in the distal tubule, early and later collecting duct of all Aldo-treated groups and also in the later collecting duct in the Nrf2Ţ-C group.

Nrf2 Expression, Activation and Target Gene Regulation
No significant difference in Nrf2 abundance in nuclear and cytosolic kidney samples was observed between groups, except for the Aldo-treated Nrf2ꜛ group (Figure 4a,b). Here, the amount of cytosolic Nrf2 was significantly increased compared with the WT control and the WT-Aldo group. To obtain better local resolution of Nrf2 levels, renal tissue was immunohistochemically stained with an antibody against Nrf2 (Figure 4c-e). A 4-to 8-fold higher expression of Nrf2 was observed in the medulla compared with the cortex, but significantly altered abundance of Nrf2 was observed only in the cortex, increased in all Aldo-treated groups. Considering the gene and protein expressions of the target genes of Nrf2, it is striking that only the target exclusively regulated by Nrf2, NADPH quinone dehydrogenase 1 (NQO1), was significantly downregulated at both gene and protein levels in the Aldotreated WT groups and upregulated in both Nrf2Ţgroups (Figure 6a). Furthermore, thioredoxin reductase 1 (Trxr1) and superoxide dismutase 1 (Sod1) showed significant downregulation at the gene level in the Aldo-treated WT group and heme oxygenase 1 (HO1) at the protein level in all Aldo-treated groups (Figure 6b-d). The catalytic subunit of glutathione cysteine ligase (γGCLC) showed no difference between groups (Figure 6e).

Keap1-Independent Regulation of Nrf2
Because downregulation of Nrf2 targets was found despite increased Nrf2 and pNrf2 levels in Aldo-treated WT animals, known negative regulators of Nrf2 activity were examined. The heterodimerization partner of Nrf2, MafK, was not changed in nuclear extracts of Aldo-treated WT animals (Figure 7a). The transcriptional regulator Bach1 is a competitor to Nrf2 in dimerization with MafK, but was also not changed in any group (Figure 7b). The kinase GSK3β, which targets Nrf2 to proteasomal degradation via phosphorylation at Ser 335 and 338, was changed neither in the cytosol (Figure 7c) nor in the nucleus (Figure 7d). In contrast, its inactive form phosphorylated at serine 9 was significantly increased in the cells of both Aldo-treated WT groups (Figure 7e). The Src family tyrosine kinase FYN, which can also mark Nrf2 for proteasomal degradation via phosphorylation, showed a significant reduction in nuclear extracts in both groups of Nrf2 mice compared with WT animals (Figure 7f).
Here, the amount of cytosolic Nrf2 was significantly increased compared with the WT control and the WT-Aldo group. To obtain better local resolution of Nrf2 levels, renal tissue was immunohistochemically stained with an antibody against Nrf2 (Figure 4c-e). A 4-to 8-fold higher expression of Nrf2 was observed in the medulla compared with the cortex, but significantly altered abundance of Nrf2 was observed only in the cortex, increased in all Aldo-treated groups.

Keap1-Independent Regulation of Nrf2
Because downregulation of Nrf2 targets was found despite increased Nrf2 and pNrf2 levels in Aldo-treated WT animals, known negative regulators of Nrf2 activity were examined. The heterodimerization partner of Nrf2, MafK, was not changed in nuclear extracts of Aldo-treated WT animals (Figure 7a). The transcriptional regulator Bach1 is a competitor to Nrf2 in dimerization with MafK, but was also not changed in any group (Figure 7b). The kinase GSK3β, which targets Nrf2 to proteasomal degradation via phosphorylation at Ser 335 and 338, was changed neither in the cytosol (Figure 7c) nor in the nucleus (Figure 7d). In contrast, its inactive form phosphorylated at serine 9 was significantly increased in the cells of both Aldo-treated WT groups (Figure 7e). The Src family tyrosine kinase FYN, which can also mark Nrf2 for proteasomal degradation via phosphorylation, showed a significant reduction in nuclear extracts in both groups of Nrf2 mice compared with WT animals (Figure 7f).

Discussion
Nrf2 activation was only minimally able to protect kidneys of mice from damage induced by Aldo. Neither pharmacological activation of Nrf2, which proved highly ineffective in this model, nor genetic activation of Nrf2 prevented systemic or local oxidative stress and also had little effect on renal damage markers. In WT animals treated with Aldo, there was broad downregulation of the Nrf2 system.
To examine the effects of Nrf2 activation on moderate renal injury, we chose an Aldo concentration that did not persistently increase blood pressure significantly above normal but produced measurable histological and physiological changes in the kidney [15]. These changes to the kidney were different in the treatment groups and showed higher damage to the glomeruli and moderate protection of the tubules in the Aldo-treated Nrf2Ţanimals, whereas the Nrf2 activator Sulf appeared to provide mild protection of the glomeruli in the WT mice, but none to their tubuli. Overall, Nrf2Ţanimals had decreased creatinine clearance, increased albumin excretion and a tendency to hydronephrosis from the outset. Risk factors for the development of hydronephrosis include ureteral stenosis and heart failure and also increased water and sodium intake [31], which was the case for our animals. Hydronephrosis has not been previously reported in the Nrf2Ţmouse, but only in a mouse in which Keap1 was specifically deleted in the renal epithelium [32]. Here, all animals studied already suffered from the renal alteration without treatment. In our study, hydronephrosis was more pronounced in the Aldo-treated group, which also had an increased drinking volume. The Nrf2Ţmouse seems to be predisposed here, as this phenomenon was not found in the Aldo-treated WT mice despite similar drinking volume, and because a quarter of the Nrf2Ţcontrol mice already showed hydronephrosis, possibly caused by the NaCl administered to all mice in the drinking water. All of this suggests a negative effect of Keap1 knockdown on renal development and/or function, which has not been reported previously [27,[33][34][35]. Furthermore, we saw increased collagen deposition in the hearts of animals with Nrf2 activation, which may be a contributing factor to the cardiac side effects of the Nrf2 activator CDDO-Me. It is known that complete deletion of Keap1 in Nrf2Ţmice leads to death from malnutrition due to hyperkeratosis of the esophagus shortly after birth [36]. Until our report, only studies showing positive effects of Keap1 knockdown in renal damage models, such as protection from tubular damage in reperfusion or ureteral obstruction, had been published [35,37]. Our animals also showed some protection of tubuli, however, only by the genetic modification and not by the activator. This contradicts observations from studies performed with another Nrf2 activator, CDDO-Me, which was able to protect tubules at least during shorter treatment periods [38,39].
The increased Aldo-induced albumin excretion reveals the sensitivity of Nrf2Ţanimals to glomerular damage, which has already been reported from three other renal injury models and in which increased fibrosis in the glomerulus was observed, as in the animals in this work [34]. There, loss of podocyte processes was additionally seen. Sulf somewhat reduced albuminuria in the WT animals, as previously shown for Sulf and also curcumin in other models [5,11].
As expected for the Nrf2Ţmouse strain used, the expression of Keap1 was already decreased by more than 50% by insertion of the loxP sites in the absence of activating Cre expression [33,37]. That the decreased amount of Keap1 basally has no direct effect on Nrf2 expression has also been observed previously by another group, but does not indicate anything about the posttranslational regulation of Nrf2 [37]. Looking closer at the localization of Nrf2, in these groups of animals, as in the preparatory experiment [15], a significantly increased Nrf2 amount was shown in all Aldo-treated groups only in the cortex, not in the medulla of the kidney. Too large interindividual differences among the animals prevented an evaluation of the presence of Nrf2 in the nucleus. Therefore, the presence of phosphorylated and thus usually labeled active pNrf2 was additionally quantified in nuclear and cytosolic extracts but also in the different cells of the kidney. Indeed, phosphorylation is not essential for translocation to the nucleus [40]. Increased phosphorylation of Nrf2 was confirmed in the cortex of Aldo-treated WT animals and, at greater resolution, increased phosphorylation in the nuclei of the distal tubules and collecting duct in all Aldo-treated groups and, at the latter site, also in the Nrf2Ţcontrol group. Sulf alone failed to alter phosphorylation status. Differential distribution of Nrf2 but also of other components of the antioxidant defense system within the different nephron segments was reported by several groups, with frequently better equipment found in the proximal tubule [41,42].
Again, as previously reported, Aldo treatment in mice did not induce Nrf2 target genes as we previously observed in rats [11,15], consistent with reports of downregulated Nrf2 targets from the human situation in CKD and other animal CKD models [5,16]. Unexpectedly, based on our previously conducted rat study with Sulf as the Nrf2 activator [11], Sulf had no effect at all on Nrf2 target genes, either alone or in combination with Aldo. It is unlikely that the reason for this was the chosen oral route of administration, which represented less suffering for the animals. Oral application was supposed to compensate for the short half-life of Sulf of about 2 h, which was also the time point where the highest concentration of Sulf was detected in kidneys of mice after a single oral administration [43,44]. This kind of application had worked very well in our rats, in which it was shown that 80% of orally administered Sulf was bioavailable [11,45]. The choice of the Sulf dose used was in the range of published doses which varied from 0.5 to 12.5 mg/kg × day, with one report in which a higher dose of 10 mg/kg × day produced no effects compared with the lower dose of 2 mg/kg × day [46]. As some effects were noted in the animals, it cannot be assumed that the Sulf was not incorporated at all, but possibly the dose was not appropriate for this model. The assumption that Sulf may be less effective in the kidney than CDDO-Me was not expected based on its use and effects in animal models of renal injury [11,47]. That we do not find a clear protective effect against Aldo-induced damage, as seen in other models of renal injury, is partly because of the adverse effects of Aldo on Nrf2 signaling discussed below. On the other hand, the type of renal injury also plays a crucial role. Protective effects of Nrf2 activators have been seen mainly in acute injury and in models that primarily induce tubular injury [35,37,39]. When glomerular damage is involved, reports that find no protection by Nrf2 activation accumulate, as we observe in our animals [34,46]. Thus, it may be that only patients with specific renal diseases that do not involve marked proteinuria benefit from Nrf2 activation.
In contrast to the situation in the WT mice, in the Nrf2Ţmice, there was a clear induction of the most specific Nrf2 target, NQO1. Other targets such as HO1 regulated by other transcription factors besides Nrf2 [48] were decreased or only tended to be increased, such as TrxR1 or γGCLC. In most cases, Aldo enhanced the response already seen basally, with the exception of SOD1. Other studies using Nrf2Ţmice all showed induction of NQO1 at the protein or mRNA level, suggesting sustained activation of Nrf2 signaling in the kidney [33][34][35]37]. Additionally, microarray analysis showed that gene expression of most Nrf2 target genes was already basally increased in Nrf2Ţmice, which included the targets NQO1, TrxR1, γGCLC, and HO1, also studied in this work [35].
Thus, although increased levels of Nrf2 and also of pNrf2 were found in the kidney, no induction of Nrf2 targets was detected in the WT mice. This suggests that either dimerization partners were lacking or not present in the appropriate proportion for the unhindered activity of Nrf2 or pNrf2 in the nucleus, or that the function of Nrf2 was negatively regulated. For example, the formation of a heterodimer with sMaf is essential for the binding of Nrf2 to the ARE [24]. However, sMaf proteins can also form homo-or heterodimers with each other or with other proteins, such as Bach1, which inhibits transcription of ARE-regulated genes [49]. Nguyen et al. demonstrated that overexpression of MafK represses the expression of ARE-dependent catalase in a dose-dependent manner [50]. The MafK level was indeed slightly higher in the nuclear fraction of Aldo-treated WT groups, which may have resulted in an inhibition of expression, whereas no clear picture emerged for Bach1.
The potential negative regulator GSK3β was slightly increased in the nucleus, where its phosphorylation of Nrf2 would lead to its nuclear export with subsequent degradation, whereas no change in GSK3β was detected in the cytosol. In the Aldo-targeted WT groups, even the inactive phosphorylated form of GSK3β, pGSK3β, was significantly increased and therefore cannot be responsible for the decreased expression of Nrf2 targets. Higher expression of GSK3β was observed in a folic acid-induced CKD model in mice after 4 weeks and resulted in increased oxidative stress but lower Nrf2 accumulation in the nucleus and decreased HO1 expression [51]. This study further demonstrated that after initial damage to the kidney, the Nrf2 pathway was activated, but this activation progressively decreased over the next 4 weeks, coupled with an increase in GSK3β protein levels. Furthermore, these effects were also found in renal biopsies from CKD patients [51]. The important role of GSK3β is underlined by a study in which inhibition of GSK3β was able to protect against Aldo-induced inflammation and fibrosis in the kidney and heart of mice [52]. Sulf-only-treated animals showed no significant change in GSK3β or pGSK3β, although Sulf theoretically not only has an effect on the Keap1-Nrf2 interaction but can also affect Nrf2 regulation via deactivation of the GSK3β kinase. Thus, inactivation of GSK3β via its phosphorylation was observed in the heart of Sulf-treated mice [53] and a negative effect of Sulf on GSK3β activity was also shown in diabetic rats [54]. A clearer picture could be obtained if regulators were also examined at the cell type level rather than in total kidney extracts.
The last negative regulator we examined in this work was the kinase FYN, which can eject Nrf2 from the nucleus by phosphorylation and is itself regulated by GSK3β [55]. Here, despite the analysis of whole kidney extracts, there was a significant difference between the WT and Nrf2Ţanimals in the expression of regulators of Nrf2, because the FYN level was slightly higher in the WT animals, whereas it was already present at a significantly lower level basally in the Nrf2Ţ-C group and was also significantly decreased compared to the Aldo-WT animals after Aldo treatment. This could explain the difference in Nrf2 target expression between the two mouse strains.
A limitation of our study is that specific Nrf2 activators were not yet available at the time of the experiment. It would be very interesting in the future to investigate the responses of the kidney to the specific protein interaction inhibitors. Furthermore, the very high interindividual variations of the groups were a hindrance to obtaining significant results from more of the parameters studied. The unexpected lack of protection of the kidneys of Aldo-treated animals against damage by Nrf2 activation could be explained with the help of recent publications, the findings of which are supported by our study.

Conclusions
In conclusion, neither the Nrf2 activator Sulf as used in the model presented here nor genetic upregulation of Nrf2 activity provided protection against Aldo-induced moderate renal injury. In contrast, negative effects of constitutive genetic Nrf2 upregulation were observed with respect to the glomeruli and the filtration barrier, which was also the case, only to a lesser extent, in the Sulf-treated mice. In addition, the hearts of the Sulf-treated animals showed an increase in collagen deposition, as did the animals with genetically activated Nrf2. Treatment with Aldo resulted in extensive inhibition of Nrf2-regulated antioxidant defenses in the kidney. This was not the case with genetic Nrf2 activation for most Nrf2 targets, possibly due to a lower presence of the negative Nrf2 regulator FYN in the nucleus and an inactive GSK3β in the cytosol. Our results highlight the unclear effects of Nrf2 activation on the body in kidney injury. However, we are one step closer to understanding the reduced Nrf2 activity in kidney injury, which may be due to an activated GSK3β/FYN axis.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.