Loss of Rictor in tubular cells exaggerates lipopolysaccharide induced renal inflammation and acute kidney injury via Yap/Taz-NF-κB axis

Our previous study demonstrated that the mammalian target of rapamycin complex 2 (mTORC2) signaling alleviates renal inflammation and protects against cisplatin-induced AKI. However, the underlying mechanisms for mTORC2 in regulating renal inflammation in AKI remain to be determined. In this study, we found that lipopolysaccharide (LPS) could activate mTORC2 signaling in NRK-52E cells, and blockage of mTORC2 signaling led to Yap/Taz degradation, which in turn activated NF-κB signaling and induced inflammatory cytokines secretion. Overexpression of constitutively active Taz (Taz-S89A) could attenuate the inflammation-amplified role of mTORC2 blockage. In mouse models, tubule-specific deletion of Rictor had higher blood urea nitrogen level, severe morphological injury as well as more inflammatory cells accumulation compared with those in their littermate controls. Overall, these results demonstrate that mTORC2 signaling protects against renal inflammation and dictates the outcome of AKI by modulating Yap/Taz degradation.


Introduction
Acute kidney injury (AKI) is an extremely lifethreatening clinical syndrome characterized by a rapid decrease in renal function. Tubulointerstitial inflammatory cells infiltration is the main characteristic of AKI. Accumulated evidence unveils that tubular cells play important roles in AKI-related inflammation. Tubular cells are always the initial site of injury caused by hypoxia or ischemia, which in turn secretes inflammatory mediators and recruits inflammatory cells infiltration, exacerbating renal injury [1][2][3][4] . However, the mechanisms by which tubular cells trigger kidney inflammation remain unclear.
The evolutionarily conserved Hippo signaling pathway is recognized as a regulator of cell proliferation, organ size, and tissue regeneration. MST1/2, SAV1, LATS1/2, MOB1A/B, Yes-associated protein (YAP) and its paralogue TAZ (also known as WWTR1) consist of the core components of the Hippo pathway. Among them, YAP and TAZ are the key downstream effectors of the Hippo pathway via MST1 and LATS kinases phosphorylation [5][6][7][8] . Recently, several studies have uncovered the roles of Yap/ Taz in regulating inflammatory and immune responses 9,10 . Deng et al. reported that phosphorylated Yap is sufficient to attenuate inflammatory responses in osteoarthritis by suppressing NF-κB signaling 10 . However, the mechanisms by which NF-κB signaling triggers Yap/ Taz inhibiting inflammatory response are not fully clarified.
In this study, we demonstrated that Rictor/ mTORC2 signaling is crucial for protecting against renal inflammation and lipopolysaccharide (LPS)-induced AKI through inhibiting Yap/Taz degradation and NF-κB nuclear translocation. Our studies unveiled a novel mechanism for Rictor/mTORC2 signaling in regulating renal inflammation and AKI.

LPS activates Rictor/mTORC2 signaling in tubular cells
Tubulointerstitial inflammation is the main characteristic of sepsis-induced AKI 19 . To determine the role of Rictor/mTORC2 signaling in the inflammatory response in AKI, first, we employed a model using LPS injection in CD1 mice. Blood urea nitrogen (BUN) assay and periodic acid-Schiff (PAS) staining revealed LPS-induced mice kidney dysfunction and morphological injury ( Fig. 1a-b). Western blot and immunostaining showed that p-Akt (Ser473) was dramatically increased in tubular cells at 1 day after LPS injection ( Fig. 1c-d). Then, we stimulated NRK-52E cells, a rat kidney epithelial cell line, with LPS (500 ng/ml) at different time durations as indicated. Western blot assay showed that the abundance of Rictor, p-Akt (Ser473), p-IKBα, and p-P65NFκB were upregulated after LPS treatment (Fig. 1e). Immunostaining for Rictor and P65NF-κB further confirmed these results (Fig.  1f). Taken together, these results suggest that Rictor/ mTORC2 signaling is activated in kidney epithelial cells after LPS treatment.

Specific deletion of Rictor in tubular cells aggravates LPS-induced AKI in mice
To elucidate the role of tubular cell mTORC2 activation in LPS-induced AKI, we generated a mouse model with Rictor gene deletion in tubular cells via the Cre-Loxp system as previously reported 18 . Mice were then injected with LPS to induce AKI. As shown in Fig. 2a, the knockouts developed higher BUN levels compared with their littermate controls at 1 day after LPS injection. In addition, loss of brush border, enlargement of tubular lumen, and tubular cell loss were more severe in the KO mice than those in controls (Fig. 2b, c). Thus, these results demonstrated that tubular-specific ablation of Rictor aggravates tubular cell injury and AKI in mice after LPS injection.

Ablation of Rictor in tubular cells promotes LPS-induced kidney inflammation
Considering that inflammation is a crucial component of LPS-induced AKI, we next assessed the role of Rictor ablation in tubular cells on renal infiltration of inflammatory cells after LPS injection. First, we examined the infiltration of renal L6b + neutrophils and CD3 + T lymphocytes. At 1 day after LPS injection, the number of L6b + neutrophils and CD3 + T lymphocytes significantly increased in the kidney from Tubular-Rictor −/− mice compared with the controls (Fig. 3a-c). In addition, NF-κB nuclear translocation in tubular cells dramatically increased in Tubular-Rictor −/− kidneys compared with the controls (Fig. 3a, d). Second, we examined the expression levels of renal proinflammatory cytokines. Quantitative, real-time RT-PCR (qRT-PCR) showed that the mRNA expression of TNF-α, Rantes, and IL-6 was upregulated in the KO kidneys compared with those from their littermate controls after LPS injection (Fig. 3e). Together, it is concluded that specific deletion of Rictor in tubular cells exacerbates renal inflammation after LPSinduced AKI.

Blockade of Rictor/Akt axis exacerbates LPS-induced NF-κB signaling activation
To elucidate the role of mTORC2 signaling in inflammatory response in vitro, we cultured NRK-52E cells, a rat kidney proximal tubular epithelial cell line, transfected with Rictor small interfering RNAs (siRNAs). Rictor protein expression was downregulated about 70% compared with scramble siRNA (Supplemental Fig. 1A). We then treated the cells with LPS (500 ng/ml) for different durations as indicated. Western blot assay showed that knocking down Rictor expression markedly upregulated LPS-induced NF-κB signaling activation (Fig. 4a). Knocking down Rictor could also upregulate the mRNA abundance of inflammatory cytokines, including TNF-α, Rantes, and IL-6 ( Fig. 4b). To decipher the role of Akt, the major downstream molecule for mTORC2 signaling, in regulating NF-κB signaling activation, we also treated NRK-52E cells with Akt1/2 inhibitor to block the Rictor/ Akt axis, which was followed by LPS treatment, p-Akt (ser473) expression was markedly suppressed (Supplemental Fig. 1B). Western blot assay and real-time RT-PCR analysis showed that blocking Akt can significantly promote NF-κB signaling activation (Fig. 4c, d). Immunostaining for P65NF-κB further confirmed these results (Fig.  4e, f). Overall, these findings suggest that blockading Rictor/mTORC2 promotes LPS-induced NF-κB signaling activation.

mTORC2 signaling attenuates the NF-κB pathway by inhibiting Yap/Taz degradation
Previous studies have demonstrated that Yap/Taz play key roles in regulating inflammatory and immune responses 9 . In addition, Yap can also inhibit TNF-αinduced NF-κB nuclear localization 10 . To verify this finding, we created a model with LPS injection in CD1 mice. Western blot assay showed that the abundance of Yap/Taz was reduced in LPS-administered mice compared with the vehicle control (Fig. 5a). In cultured NRK-52E cells, we also found that Yap/Taz expression was reduced in response to LPS treatment (Fig. 5b). To further decipher the role of Yap/Taz induction in attenuating NF-κB signaling activation, we transfected NRK-52E cells with Taz-S89A expression plasmid. Western blot and real-time RT-PCR results showed that LPS-induced NF-κB signaling activation was considerably compromised in cells expressing Taz-S89A (Fig. 5c), which we further confirmed by immunostaining for P65NF-κB (Fig. 5d). Thus, Yap/Taz acts as an important inflammatory regulator.
Our published study demonstrated that Rictor/ mTORC2 signaling stimulates Yap/Taz transcription in fibroblasts. Therefore, we predicted that in tubular cells, mTORC2 signaling would attenuate the NF-κB pathway through modulating Yap/Taz expression. To verify this supposition, we transfected NRK-52E cells with Rictor siRNA or Akt1/2 inhibitor to block Rictor/mTORC2/Akt axis. We found that blocking the Rictor/mTORC2/Akt axis significantly reduced the total protein levels of YAP and TAZ after LPS treatment (Fig. 6a, b). We also treated NRK-52E cells with lactacystin to inhibit proteasomal degradation. Western blot assay showed that the reduced Taz protein levels were restored when cells were treated with lactacystin (Fig. 6c). In addition, the blockage of the Rictor/mTORC2/Akt axis could not decrease mRNA abundance of Yap/Taz (Fig. 6d). Next, we transfected NRK-52E cells with Taz-S89A to decipher the role of Taz induction in mediating the Rictor/mTORC2-modulated NF-κB pathway. The results showed that expression of TAZ-S89A could largely decrease the LPS-induced overactivation of the NF-κB pathway caused by Rictor siRNA transfection (Fig. 6e, f). Immunostaining for P65NF-κB further confirmed these results (Fig. 6g). Overall, these data demonstrate that mTORC2 signaling attenuates the NF-κB pathway by inhibiting Yap/Taz degradation.

Discussion
In this study, by employing a mouse model with tubular cell-specific deletion of Rictor, we demonstrated that Yap/Taz is a critical mediator for Rictor/mTORC2 in modulating renal inflammation and LPS-induced AKI. This study deciphers the mechanism for Rictor/ mTORC2 signaling in regulating renal inflammation and AKI.
Sepsis-induced AKI manifests as a dramatic decline in the glomerular-filtration rate (GFR), as well as tubular dysfunction. The adaptive responses of tubular cells to an injurious simulation elicit a range of inflammatory mediators from these cells, which results in serious kidney injury [1][2][3][4][20][21][22][23] . LPS, as an exogenous toxin, is widely used to induce systemic inflammation that mimics many of the initial clinical features of sepsis, which could induce acute kidney dysfunction and inflammatory cells infiltration [24][25][26] . The NF-κB pathway is a key regulator in the production and secretion of inflammatory cytokines [27][28][29] . However, the mechanisms by which tubular cells trigger kidney inflammation remain unclear.
It is known that mTORC2 might play a crucial but discrepant role in regulating inflammation and NF-κB signaling in diverse cell types [30][31][32][33] . Brown et al. reported that mTORC2 can negatively regulate LPS-induced inflammatory response 34 . Rictor is a defining member of mTORC2 that interacts extensively with mTOR and contributes to mTORC2 assembly 13,35,36 . Our previous study reported that ablation of Rictor in the tubule aggravates inflammatory cells infiltration and kidney injury in cisplatin-induced AKI 18 . In this study, Rictor fl/fl mice were crossed with Ksp1.3/Cre− transgenic mice to generate mouse model with tubular cell-specific deletion of Rictor, in which Cre recombinase is expressed mainly in the collecting ducts, loops of Henle, and distal tubules. In this genetic model, we found~5% deletion of Rictor in proximal tubules and~80% deletion in distal tubules 18,37 . LPS-induced nephrotoxicity might involve all segments of the tubule 38 . Therefore, our results suggest a critical role for Rictor/mTORC2 signaling in distal tubules during LPS-induced AKI. As we reported previously, Rictor/ mTORC2 is dispensable for tubular cell injury under physiological conditions 18 . Basing from several lines of evidence, we found that Rictor deletion promoted LPSinduced inflammation. The mouse model with tubular cell-specific ablation of Rictor was characterized by more neutrophils and T-lymphocytes accumulation, NF-κB activation in tubular cells, and inflammatory cytokines secretion in the kidneys after LPS injection. In NRK-52E cells, blockage of Rictor/mTORC2 signaling exacerbated LPS-induced inflammatory cytokines secretion. Together, the results suggest that activation of Rictor/ mTORC2 signaling is sufficient for preventing LPSinduced renal inflammation and AKI. It should be pointed out that in addition to tubular cells, our published studies reported that Rictor/mTORC2 signaling activation in fibroblasts may protect against tubular cell death and ischemia/reperfusion-induced AKI 39 . However, activation of Rictor/mTORC2 in kidney fibroblasts may lead to interstitial fibrosis in mice 17,40 . Therefore, mTORC2 signaling plays a dual role during kidney injury, and targeting mTORC2 for patients with chronic kidney disease should be cautious.
Hippo signaling is identified to control cell proliferation, organ size, and tissue regeneration 8,41 . Yap and its paralogue Taz are the transcriptional coactivators that are the key effectors of the pathway. Activation of Hippo signaling leads to the phosphorylation of Yap/Taz. By contrast, inactivation of Hippo signaling increases Yap/Taz nuclear localization, where they interact with transcriptional enhancer factor TEF (also known as TEAD) to activate target gene expression, including CTGF, MYC, and Cyr61 5,7 . Substantial evidence demonstrated that NF-κB is recognized as one of the important targets of Yap/Taz, Hippo signaling plays an anti-inflammatory role in regulating innate immunity and autoimmunity 10,42,43 . Moreover, our published study demonstrated that Rictor/ Fig. 3 Loss of Rictor in tubular cells aggravates renal inflammation after LPS administration. a Representative immunofluorescent staining images for Ly6b, CD3, and P65NF-κB among groups as indicated. White arrows indicate the NF-κB p65-positive tubular cells. Scale bar = 20 μm. b-d Quantitative determination of Ly6b + , CD3 + , and P65NF-κB + cells among groups as indicated. Each vertical bar represents the mean ± SEM (n = 3-6) analyzed by one way analysis of variance (ANOVA) followed by post hoc comparisons by Student-Newman-Keuls test. *p < 0.05 compared between different groups. e Real-time qRT-PCR analysis showing the mRNA abundance for TNF-α, Rantes, and IL-6 in Rictor +/+ and the knockout mice after LPS administration. Each vertical bar represents the mean ± SEM (n = 6) analyzed by students t test. *P < 0.05 vs. CDH16 + -Rictor +/+ mice treated with LPS. mTORC2 signaling stimulates Yap/Taz transcription in fibroblasts 17 . In this study, we found that LPS can decrease the protein expression of Yap/Taz through mTORC2 signaling-mediated Yap/Taz degradation. The conclusion is supported by the following evidence. First, in NRK-52E cells, the expression of Yap and Taz was significantly suppressed by Rictor siRNA or Akt1/2 inhibitor after LPS stimulation. Second, inhibiting proteasomal degradation with lactacystin could abolish Akt1/ 2 inhibitor-induced Taz degradation after LPS administration. Moreover, blockage of the Rictor/mTORC2/Akt axis could not decrease the mRNA abundance of Yap/Taz. Third, overexpression of constitutively active Taz could largely decrease the LPS-induced overactivation of the NF-κB pathway and inflammatory cytokines secretion caused by Rictor siRNA transfection. However, the underlying mechanisms of Rictor/mTORC2 signaling in regulating the Yap/Taz degradation need further investigation. It should be noted that Rictor/mTORC2 respectively plays transcriptional-or degradational-regulatory roles in fibroblasts and tubular cells, suggesting the distinct regulatory mechanisms in different cell types of the kidney.
In summary, our study demonstrated that Rictor/ mTORC2 signaling protects against renal inflammation and LPS-induced AKI through inhibiting Yap/Taz degradation and NF-κB nuclear translocation. Activation of mTORC2 signaling may potentially provide a new strategy for AKI.

Mice and animal models
Mouse model with tubular cell-specific deletion of Rictor gene was generated as previously reported 18 . Briefly, homozygous Rictor floxed mice (C57BL/6J background) from Dr Magnuson (University of Vanderbilt) were bred with the Ksp1.3/Cre transgenic mice from Jackson lab (012237, C57BL/6J background). After several times of cross breeding, mice with genotyping Rictor fl/fl , Cre +/− were considered as knockouts. The Age and gender matched Rictor fl/fl , Cre −/− mice of the same litter were considered to be the control littermates. Tubule-Rictor−/− or Tubule-Rictor+/+ mice at the age of 6W-8W were selected to establish the AKI model by single intraperitoneally injection of LPS (7.5 mg/Kg). For some experiments, CD1 male mice weighing~20-25 g were also injected intraperitoneally with 7.5 mg/kg LPS to induce AKI. Mice were euthanized at 24 h after LPS administration. Blood and kidney samples were harvested for further analysis. All of the mice were housed under specific pathogen-free (SPF) facilities in Animal Center of Nanjing Medical University according to institutional animal care guidelines with approval from the Nanjing Medical University' Institutional Animal Care and Use Committee.

Serum BUN assay
Blood urea nitrogen (BUN) was measured using a QuantiChrom Urea Assay kit (cat: DIUR-500, Hayward, CA) as per the manufacturer's instructions.

Histology and immunohistochemistry
Mouse kidney specimens were pre-fixed with 10% neutral formalin overnight and were paraffinembedded. Renal sections (3-µm thick) were prepared for periodic acid-Schiff (PAS) staining. The severity of renal injury was evaluated as follows, tubular necrosis, loss of brush borders, cast formation, and tubular dilatation. A semi-quantitative score was implemented to evaluate each tissue. Accordingly, the injury score was scored with: 0, ≤10% of the injury area stained; 1, 11-25% stained; 2, 26-50% stained; 3, 51-75% stained; and 4, >75% stained. More than ten microscope fields (magnification, ×400) were randomly selected from each section under a light microscope and we calculated the mean score. Slides were captured on a Nikon Eclipse 80i microscope connected to a digital camera (DS-Ri1, Nikon).

RNA extraction and detection of mRNA
Total RNA of cultured cells and mouse kidneys were extracted using Trizol reagent (Invitrogen) according to the manufacturer's instruction, and concentration was determined by measuring optical absorbance at 260 nm. Subsequently, 1 μg of total RNA was used to synthesize cDNA with a ReverTra Ace qPCR RT Kit (Vazyme, Nanjing, China). Quantitative real-time PCR (qRT-PCR) was carried out to measure gene expression by using an Applied Biosystems 7300 Real-time PCR System and real-time PCR assay (Vazyme). Realtive fold changes were calculated using the 2δCt method, where △CT = CT gene − CT control .

Statistical analysis
The data from the present study are presented as mean ±standard error (S.E.M.). SigmaStat software (Jandel Scientific Software) was carried out to perform statistical analysis of the data. The students t test was used for comparisons between two groups. One-way analysis of variance (ANOVA) followed by the Student-Newman-Keuls test was assessed for comparisons among multigroup. We regarded a p < 0.05 (two-side) as statistically significant.