Nrf2 deficiency promotes the progression from acute tubular damage to chronic renal fibrosis following unilateral ureteral obstruction

ABSTRACT

Background

Nuclear factor erythroid 2-related factor 2 (Nrf2) is a central mediator of cellular responses to oxidative stress. We hypothesized that Nrf2 modulates progression from acute tubular damage to renal fibrosis. We asked whether Nrf2 deletion increases renal injury in mice following unilateral ureteral obstruction (UUO).

Methods

We explored the time course of renal injury and Nrf2 expression in Nrf2^+/+ mice following UUO. We compared Nrf2^+/+ and Nrf2^−/− mice following UUO in tubular damage, transdifferentiation [vimentin, proliferating cell nuclear antigen (PCNA)], fibrosis [fibronectin, α-smooth muscle actin (SMA)], antioxidative and inflammatory responses. We studied Nrf2 in renal biopsies of patients with acute, subacute and chronic tubulointerstitial nephritis (TIN).

Results

In Nrf2^+/+ mice, renal Nrf2 expression and Nrf2-regulated glutamate-cysteine ligase catalytic (Gclc) and heme oxygenase-1 (Ho-1) were elevated, and renal injury occurred between 2 and 14 days after UUO. On Day 2 following UUO, in Nrf2^−/− mice compared with Nrf2^+/+ mice, tubular damage, apoptotic cell numbers, cleaved caspase3 and cleaved-poly ADP-ribose polymerase were increased. On Day 5, protein levels of vimentin and PCNA and the co-expressed cells of both proteins were increased. On Day 14, fibronectin and α-SMA protein levels were increased. Nrf2 deletion decreased expression of antioxidative genes (Gclc and Ho-1) and increased expression of inflammatory response genes (Tgfβ, Tnf, IL-6, IL-1β and F4/80). Finally, Nrf2 expression was upregulated in renal biopsies of patients with TIN.

Conclusions

Following UUO, Nrf2 deficiency increased tubular damage, transdifferentiation, fibrosis and inflammatory response while decreasing antioxidative responses. The renal protective role of Nrf2 in the development of tubulointerstitial fibrosis in UUO may be mediated by antioxidative and anti-inflammatory pathways.

INTRODUCTION

Renal tubulointerstitial fibrosis is a principal feature of progressive kidney injury and a major cause of end-stage renal disease (ESRD) [1]. Better understanding of pathogenesis of tubulointerstitial damage and deploying effective therapies may reduce the global burden of ESRD [2, 3].

Nuclear factor erythroid 2-related factor 2 (Nrf2) is a redox-sensitive transcription factor that regulates antioxidant proteins, cell cycle-related regulators and detoxification enzymes [4]. Nrf2 has shown a renal protective role in acute and chronic kidney disease (CKD) [5, 6] including acute kidney injury (AKI) models induced by rhabdomyolysis [7], ischemia/reperfusion and cisplatin [8], as well as, chronic injury models including pristine-induced lupus nephritis [9] and diabetic nephropathy [10].

The unilateral ureteral obstruction (UUO) mouse model is characterized by progressive tubulointerstitial fibrosis and reflects the biological processes present in clinical obstructive nephropathy [11]. Nrf2 activators attenuate UUO-induced kidney injury [12–16]. Nrf2 deletion aggravates the interstitial fibrosis at a single time point, Day 7 after UUO [17]. These results suggest that Nrf2 plays a protective role against tubulointerstitial fibrosis. However, the mechanisms by which Nrf2 deficiency prevents progression from acute tubular damage to chronic tubulointerstitial fibrosis remain to be clarified.

In this study, we aimed to verify whether Nrf2 deletion aggravates the progression of renal injury following UUO. We showed that renal Nrf2 expression is increased in wild-type (Nrf2^+/+) mice after UUO and in patients with tubulointerstitial nephritis (TIN). Nrf2 deficiency exacerbated tubular damage in the early phase, elicited maladaptive proliferation and differentiation in the transition phase and promoted fibrosis in the late phase of the UUO model.

MATERIALS AND METHODS

Mouse model

Animal studies were conducted under a protocol approved by the Institution Animal Committee. Male Nrf2^+/+ (n = 54) and Nrf2^−/− mice (n = 36) were divided at age 10 weeks into four groups: (i) sham Nrf2^+/+, (ii) UUO Nrf2^+/+, (iii) sham Nrf2^−/− and (iv) UUO Nrf2^−/−. Nrf2^−/− mice developed as described previously [18]. UUO surgery was performed by two-point ligation of the left ureters and the ligated ureter was cut. Sham-operated mice underwent the same procedure without ureter ligation [19].

Experiment 1 was to observe time course of kidney injury and Nrf2 expression in Nrf2^+/+ mice after UUO. Experiment 2 was to compare Nrf2^−/− with Nrf2^+/+ mice on Days 2, 5 and 14 after UUO (Supplementary data, Figure S1). Animals were euthanized after kidney perfusion. Half kidney was fixed in 4% paraformaldehyde for paraffin embedding and the other half was stored at −80°C for future use.

Human kidney tissues

Nine patients with biopsy-proven TIN including acute, sub-acute and chronic cases (n = 3 in each group) (Table 1) were recruited from the Nephrology Department between 2015 and 2017 March. Diagnoses were made by a nephron-pathologist following published criteria [20]. Normal kidney tissue was obtained at the time of renal cancer surgery in the Urology Department. These tissues were located at least 5 cm away from the tumors. All subjects provided written consent approved by the Institutional Review Boards.

Histology and immunostaining

Kidney sections cut from human (2 µm) and mouse (5 µm) paraffin-embedded kidney blocks were deparaffinized and rehydrated. Tissues were stained with periodic acid–Schiff (PAS), Masson-trichrome, TdT-mediated dUTP nick-end labeling (TUNEL), immunohistochemical (IHC) and immunofluorescent (IF) staining.

Quantification of tubular injury on PAS sections and TUNEL-positive cells were performed on 10 randomly selected fields from each kidney. Necrotic tubules were defined by the loss of brush border, cellular vacuolization and nuclear detachment into the tubular lumen. The percentage of necrotic tubules in one field was determined. The average percentage of all fields was indicated as tubular injury extent [21]. The average of TUNEL-positive cells per high power field was expressed as apoptosis degree [22].

For IHC, kidney sections underwent pretreatment described above, antigen retrieval and blocking nonspecific reaction, followed by incubation with antibodies against Nrf2, fibronectin or F4/80 overnight at 4°C. After washing, sections were incubated with biotin-conjugated goat IgG for 30 min at room temperature (RT), and reacted with streptavidin-conjugated peroxidase for 30 min at RT followed by visualization with a DAB Kit (Zsbio, China) [23]. To quantify Nrf2 expression in human biopsies, 10 random fields per kidney were digitally imaged. The positive staining was measured by Image Pro Plus software and expressed as a percentage of positive Nrf2 area within the total captured area [24].

For IF, pretreated kidney sections were incubated with antibodies against vimentin, proliferating cell nuclear antigen (PCNA) or α-smooth muscle actin (α-SMA) at 4°C overnight, then incubated with Alexa-568/Alexa-488 donkey IgG (Thermo Scientific, Rockford, IL, USA) at RT for 1 h. After washing, the slides were mounted with VECTASHIELD^® Mounting Medium containing 4′,6-diamidino-2-phenylindole (Vector Labs, Burlingame, CA, USA). Images were captured by immunofluorescent microscopy (Nikon Corporation, Tokyo, Japan).

Immunoblotting

Kidney total proteins were extracted using Radio-Immunoprecipitation Assay (RIPA) buffer with protease inhibitors, and the concentrations were determined by Bicinchoninic acid (BCA) assay. Equal amount of protein was separated by SDS-PAGE and transferred onto PVDF membranes (Millipore Immobilon-P, Darmstadt, Germany). After blocking with 5% milk, membranes were incubated with primary antibodies against Nrf2, glutamate-cysteine ligase catalytic (Gclc) subunit, heme oxygenase-1 (Ho-1), cleaved caspase3 (C-caspase3), poly ADP-ribose polymerase (PARP), vimentin, PCNA, fibronectin, α-SMA or F4/80 overnight at 4°C. The blots were incubated with peroxidase-conjugated goat IgG for 60 min at RT. The antibody–antigen reactions were detected by High-sig ECL Western Blotting Substrate and visualized by the Tanon 5500 imaging system. The protein loading variation was normalized by α-tubulin, β-actin or GAPDH. The blot density was analyzed by NIH ImageJ software. The protein level is expressed as the ratio of blot density from individual protein to its housekeeping antibody (the information for all antibodies seen in supplemental data, Table S1).

Real-time PCR

Kidney total RNAs was extracted with Trizol reagent (Life Technologies, Carlsbad, CA, USA) and the concentration was measured with Nanodrop 2000. RNA (50 ng) was subjected to reverse transcription using PrimeScript RT Reagent Kit and followed by PCR with SYBR Premix Ex Taq (Takara, Dalian, China) [25] for E-cadherin, Gclc, Ho-1, vimentin, fibronectin, α-SMA, transforming growth factor β-1 (Tgfβ1), tumor necrosis factor (Tnf), interleukin 6 (IL-6) and IL-1β. Primers were designed using Primer Express (Applied Biosystems, Carlsbad, CA, USA) and synthesized by Life Technologies (Shanghai, China) (Supplementary data, Table S2). Real-time fluorescence was detected with QuantStudio 6 Flex Quantitative PCR System (Applied Biosystems). The difference of gene expression between two groups was expressed as a ratio of each group to sham Nrf2^+/+ mice group. The mRNA levels were expressed as 2^−ΔΔCt (ΔCt: Gapdh Ct − individual gene Ct) [26].

Statistical analyses

All data are expressed as mean ± standard deviation. Differences between groups were analyzed by one- or two-way ANOVA with Bonferroni multiple comparison tests GraphPad Prism software. A P-value <0.05 was accepted as statistically significant.

RESULTS

Time course of renal injury and Nrf2 expression after UUO in Nrf2^+/+ mice

Loss of brush border and cellular vacuolization were seen on Days 2–5; nuclear detachment into the tubular lumen was found on Days 5–21; thinned inner medulla, scarred cortical surface and increased interstitial cell numbers were observed on Day 14; and widened interstitial distance appeared on Day 21 after UUO (Figures 1A and B and 3). The weight ratio of obstructed to contralateral kidney decreased on Days 14–21 (Figure 1C). E-cadherin mRNA, a marker of differentiated tubular cell phenotype, significantly decreased from Days 2–21 (Figure 1D). Vimentin mRNA, a marker of tubular cell transdifferentiation, increased at Day 5 and peaked at Day 21 (Figure 1E). Fibronectin and α-SMA mRNA expression in two fibrotic factors were significantly increased from Days 14–21 (Figure 1F).

Total Nrf2 protein increased at Day 5, peaked at Day 14 and returned to the baseline at Day 21 in Nrf2^+/+ mice after UUO compared with Day 0 or sham-operated mice on western blot (Figure 2A and B). The protein levels of Gclc and Ho-1, Nrf2 downstream genes, started increasing earlier than Nrf2 at Day 2, with Ho-1 maintaining higher expression until Day 21 after UUO (Figure 2A and B). In IHC staining, Nrf2-positive staining can be detected as early as Day 2, which mainly located in nuclei and less in cytoplasm of healthy-looking or slightly injured tubules. By Day 5, cytoplasmic Nrf2 staining was further increased while nuclear Nrf2 decreased in the tubules with brush border loss and cellular vacuolization. By Day 14, positive Nrf2 was limited to the tubules with nuclear detachment into the tubular lumen. By Day 21, positive Nrf2 disappeared from disordered and dilated tubules (Figure 3).

Hence, we selected Days 2, 5 and 14 to investigate the role of Nrf2 in Nrf2^−/− mice following UUO.

Tubular damage on Day 2 after UUO in Nrf2^−/− mice

In Nrf2^−/− mice compared with Nrf2^+/+ mice, protein expression of C-caspase3 and cleaved-PARP (C-PARP), two apoptotic markers, increased on Day 2 after UUO; in contrast, full-length PARP showed no change on western blots (Figure 4A and B). PAS staining showed more severe tubular damage in Nrf2^−/− mice compared with wild-type mice (Figure 4C and D). TUNEL-positive cells increased 10-fold in Nrf2^+/+ mice after UUO compared with sham group, and increased 3-fold in Nrf2^−/− compared with Nrf2^+/+ mice (Figure 4C and E).

Transdifferentiation and proliferation on Day 5 after UUO in Nrf2^−/− mice

On Day 5, UUO induction increased vimentin and PCNA proteins in Nrf2^+/+ mice compared with sham group, and further increased both proteins in Nrf2^−/− compared with Nrf2^+/+ mice (Figure 5). Vimentin mRNA also was increased after UUO (Supplementary data, Figure S2). In addition, Nrf2 deficiency appeared to increase tubular and interstitial co-expressed cells with vimentin and PCNA (Figure 5D).

Renal fibrosis on Day 14 after UUO in Nrf2^−/− mice

On Day 14 after UUO, tubulointerstitial fibrosis appeared to increase after UUO in Nrf2^+/+ and Nrf2^−/− mice on Masson stain. Renal fibronectin and α-SMA, two fibrosis indicators, increased in Nrf2^−/− compared with Nrf2^+/+ mice on western blot and immunostaining (Figure 6).

Time course of antioxidative and inflammatory response after UUO

To explore the mechanisms by which Nrf2 plays a renoprotective role in UUO, we examined antioxidative and inflammatory responses at Days 2, 5 and 14 after UUO. Antioxidative response genes Gclc and Ho-1 were up-regulated in both Nrf2^+/+ and Nrf2^−/− mice starting Day 2. Nrf2 deficiency reduced Gclc and Ho-1 mRNA levels (Figure 7A and B). The inflammatory cytokines Tgfβ1 and Tnf mRNA gradually increased from Days 2 to 14 (Figure 7C and D). IL-6 and IL-1β mRNA were remarkably elevated on Day 14 (Figure 7E and F). Further Nrf2 deletion aggravated the production of the four inflammatory cytokines.

We further examined F4/80, a macrophage marker, on Day 14 after UUO. The protein levels of F4/80 increased in both Nrf2^−/− and Nrf2^+/+ mice. Nrf2 deficiency further increased F4/80 expression compared with Nrf2^+/+ mice (Figure 8).

Nrf2 expression in renal biopsies from patients with TIN

Lastly, we studied Nrf2 in TIN patients. Tubular damage and interstitial fibrosis appeared more severe in chronic compared with acute cases on PAS and Masson stain. Nrf2-positive staining was increased in both nuclei and cytoplasm in acute TIN (Figure 9B3), and gradually decreased and limit in cytoplasm of sub-acute and chronic TIN with quantification (Figure 9C3–D3 and Supplementary data, Figure S3). This pattern of Nrf2 staining in human TIN kidneys was similar to that of UUO mice.

DISCUSSION

We found that Nrf2 up-regulated in UUO-treated mice and in renal biopsies from TIN patients. Nrf2 deficiency aggravated tubular injury in early phase, promoted maladaptive proliferation and dedifferentiation in transition phase and increased fibrogenesis in late phase after UUO. Nrf2 deletion decreased antioxidative genes and increased profibrotic inflammatory cytokines.

UUO is a classic model of renal interstitial fibrosis [11]. It starts with hemodynamic and metabolic changes, followed by tubular injury, transdifferentiation and proliferation and oxidative and inflammatory responses. These cellular responses ultimately lead to renal fibrosis [27]. Nrf2 modulates antioxidant genes by interacting with antioxidant response element. Under normal conditions, Nrf2 is confined to the cytoplasm associated with the suppressor protein Keap1 and then degraded by the ubiquitin–proteasome pathway. In disease condition, oxidative and electrophilic stress factors stimulate dissociation of the Nrf2–Keap1 complex, promote Nrf2 release and translocation into the nucleus to up-regulate Nrf2-mediated antioxidant genes to protect cell from injury [28]. The up-regulation of Nrf2 can protect against UUO-induced oxidative stress response [29].

In this study, the nuclear Nrf2 staining increased in some healthy appearing tubules and slightly injured tubules on Day 2 (Figure 3B), earlier than the elevated Nrf2 total protein on western blot at Day 5 (Figure 2). This suggests that Nrf2 translocated to the nuclei in the early phase after UUO. The nuclear translocation might protect tubular damage by up-regulating Nrf2 downstream antioxidative genes (Gclc and Ho-1) in mRNA and protein at Day 2 after UUO (Figures 2 and 7A and B). On Day 5, the elevated Gclc and Ho-1 up-regulated Nrf2 protein, and more cytoplasmic and less nuclear positive Nrf2 staining in the injured tubules, might contribute to balance adaptive and maladaptive repairing process of injury. Even though the increased Nrf2 expression was mainly located in cytoplasmic on Day 14 (Figure 3D4), increased protein of Gclc and Ho-1 (Figure 2) and mRNA of Ho-1 on Day 14 (Figure 7B) suggested that Nrf2 may play renal protective role through antioxidant activity in the late phase of cellular repair. Our findings were consistent with other reports that Nrf2 protected kidney from injury through antioxidative response [13, 29, 30]. On Day 21, baseline level of Nrf2 protein (Figures 2 and 3E) suggests that severely damaged cells lost Nrf2 renal protection due to sustained UUO.

So we explored the mechanisms of Nrf2 renal protection at Days 2, 5 and 14 after UUO. Apoptosis is the earliest event of UUO-induced fibrosis [30]. Nrf2 is a key regulator of antioxidant responses that protect against cell death [31, 32]. Liu et al. [8] reported that Nrf2 deficiency worsened ischemia and cisplatin-induced AKI in Nrf2^−/− mice and that Nrf2 activator (CDDO-imidazolide) attenuated ischemic AKI [33]. The Yamamoto group recently reported that tubular-specific Nrf2 activation protects tubular damage progress against ischemia and reperfusion [34]. Most current studies have focused on late fibrosis [13, 17], and few have reported early changes before fibrosis in UUO animals [12, 35]. We found that apoptotic cells, C-caspase3 and C-PARP increased on Day 2 after UUO in Nrf2^−/− compared with Nrf2^+/+ mice (Figure 4). This suggests that the increased apoptosis might contribute to the late renal fibrosis in Nrf2^−/− mice with UUO.

After a single episode of tubular damage, adaptive repair processes promote cell proliferation and restore normal tissue architecture. On the other hand, maladaptive repair processes such as activation of pericytes and myofibroblast contribute to progressive fibrosis [36]. After sustained UUO, damaged tubular cells produce reactive oxygen species (ROS) that lead to the failed differentiation of regenerating tubular cells. Furthermore, interstitial fibroblasts are activated, myofibroblasts are generated and progressive fibrosis ensues [37]. This transition phase is characterized by downregulating E-cadherin and upregulating vimentin [38]. Tubular and interstitial co-expression of vimentin and PCNA was seen in the repair phase after ischemia and reperfusion [39]. It is the first report that vimentin and PCNA increased in Nrf2^+/+ mice on Day 5 after UUO with an even greater increase in Nrf2^−/− (Figure 5A–C). In addition, the co-expression of vimentin and PCNA increased in tubules, particularly in interstitium of Nrf2^−/− mice (Figure 5D), prior to fibrosis formation on Day 7 as reported previously [16, 35]. The increased tubular co-expression of vimentin and PCNA indicated that the regenerative tubular cells tend to de-differentiate, and the elevated interstitial co-expression might be a precursor population of myofibroblasts that contribute to fibrosis formation as reported in ischemic AKI [40]. Our findings suggest that maladaptive proliferation and transdifferentiation on Day 5 might trigger late fibrosis formation in Nrf2^−/− mice with UUO.

Fibrosis is the outcome of progressive CKD. Renal fibrogenesis appeared in kidney cells and infiltrated macrophages [41]. The macrophages can produce profibrotic cytokines [42] and transform to myofibroblast producing fibronectin [43]. F4/80 increase from Days 3 to 14 after UUO [43] and distribute interstitially on Day 14 after ischemia and reperfusion [34]. Nrf2 can alleviate inflammatory phenotypes by inhibiting transcription of IL-6 and IL-1β in macrophages rather than up-regulating antioxidant proteins to eliminate ROS [44]. In our study, fibronectin, α-SMA and Masson-stained fibrosis increased on Day 14 after UUO in Nrf2^−/− compared with Nrf2^+/+ mice (Figure 6). Meanwhile, Nrf2 deletion increased expression of Tgfβ1, Tnf, IL-6, IL-1β and F4/80 (Figures 7 and 8). This is the first report about the increase of macrophages infiltration and profibrotic cytokine release on Day 14 after UUO in Nrf2^−/− mice. Our data suggest that the UUO-induced fibrosis in Nrf2^−/− mice may be aggravated by over-production of inflammatory cytokines.

Nrf2 shows protection in several kidney diseases with Nrf2^−/− mice [8–10, 17] or Nrf2 activators [13, 33, 45–47]. In UUO models, various Nrf2 activators like sulforaphane, sinomenine [17], dimethylfumarate [13], epigallocatechin-3-gallate [14], fimasartan [15] and oleanolic acid [12] attenuate UUO-induced renal fibrosis [12, 13, 17]. However, the majority of these studies used Day 7 after UUO as the single endpoint. Only oleanolic acid showed a decreased apoptosis at Day 3 [12]. One recent study showed more renal fibrogenesis at Day 7 in Nrf2^−/− mice [17]. We studied multiple time points to evaluate the role of Nrf2 with Nrf2^−/− mice. Our data provided more evidence that Nrf2 protected renal injury via anti-apoptosis, balancing proliferation and transdifferentiation, antioxidative responses and anti-profibrotic inflammation in different stages of UUO.

Nrf2 is increased in renal biopsies of patients with lupus nephritis or diabetic nephropathy [9, 10]. Since it is difficult to obtain human kidneys from different stages of obstructive nephropathy, we used acute, subacute and chronic TIN renal biopsies to explore Nrf2 expression in human tubulointerstital injury. This is the first report of Nrf2 expression in TIN patients. Nrf2 increased in acute TIN and gradually decreased in subacute and chronic TIN (Figure 9). The TIN patients’ data suggested that Nrf2 might also play an important role in the development of human tubulointerstitial diseases.

In this study, Nrf2 showed a protection role in UUO-induced progressive renal fibrosis by antioxidative response and anti-inflammation (Figures 7 and 8). However, we believe that the renal-protective mechanism of Nrf2 is more complicated because of a time-accumulative effect in UUO-induced fibrosis (Figure 10). Applying tubular-specific Nrf2 knockout mice will help clarify the mechanism of Nrf2 in UUO-induced renal injury.

In conclusion, our data suggest that Nrf2 plays an important role in the progression from acute damage to renal fibrosis in UUO animals. The renoprotective role of Nrf2 might be mediated by decreasing apoptosis and antioxidative response in the early phase, attenuating maladaptive proliferation and transdifferentiation in the repair process and anti-inflammation and anti-fibrosis in the late phase. The potential therapeutic target of Nrf2 in the progression of ESRD needs further investigation.

AUTHORS' CONTRIBUTIONS

W.K. conducted all experiments and drafted the manuscript. H.Z. designed the experiments and edited the manuscript. N.L. provided patients’ pathological diagnoses. L.Y, H.W., L.W., J.F. and J.P. revised the experimental design. C.J., G.G. and J.L. collected kidney biopsy tissue from patients. M.Y. provided Nrf2^−/− mice and reviewed this manuscript.

SUPPLEMENTARY DATA

Supplementary data are available online at http://ndt.oxfordjournals.org.

ACKNOWLEDGEMENT

The authors thank Dr Jeffrey Kopp from Intramural Research Program, NIDDK, NIH, Bethesda, MD for English editing.

FUNDING

This research was supported by the Chinese Nature Science Foundation (81370835 to H.Z., 81573106 to J.P. and 81402635 to J.F.), Pandeng Scholar of Liaoning Province (2013222 to H.Z. and J.P.) and Startup Funding of China Medical University (to J.P.).

REFERENCES