Soluble receptor for advanced glycation end products protects from ischemia- and reperfusion-induced acute kidney injury

Acute kidney injury (AKI) is one of the important risk factors for the development of chronic kidney disease and end-stage renal disease (ESRD) (Basile et al., 2016). Renal ischemia and reperfusion (I/R), which leads to AKI, occurs in humans in settings such as trauma, circulatory arrest, major vascular surgery, and kidney transplantation (Eltzschig and Eckle, 2011). Renal I/R induces immune responses through inflammatory signaling transductions. Pattern recognition receptors, such as Toll-like receptors, recognize pathogen-related molecules and activate an inflammatory response. Recent studies have revealed that Toll-like receptors play a critical role even in I/R-induced sterile inflammation (Eltzschig and Eckle, 2011; Wu et al., 2010).

A full-length form of the receptor for advanced glycation end products (RAGE) is also a multiligand pattern recognition receptor (Bongarzone et al., 2017). RAGE binds advanced glycation end products, S100 proteins, high-mobility group box protein 1 (HMGB1), and β-sheet fibrillar material (Bongarzone et al., 2017; Schmidt et al., 2001; Sims et al., 2010). HMGB1, one of the damage-associated molecular patterns (DAMPs), is a common ligand for RAGE and Toll-like receptors. The HMGB1–RAGE association activates pro-inflammatory signal transduction via nuclear factor κ-light chain-enhancer of activated B cells, resulting in inflammatory responses (Batkulwar et al., 2015; Chen et al., 2016). It is known that soluble isoforms of RAGE (sRAGE) consist of cleaved isoforms of RAGE and endogenous secretory RAGE (esRAGE) (Prasad, 2019). Cleaved RAGE is derived by the proteolytic cleavage of full-length RAGE, whereas esRAGE is generated by alternative splicing of RAGE messenger RNA (mRNA) (Prasad, 2019; Yonekura et al., 2003; Kalea et al., 2009; Schmidt, 2015). RAGE has two ways of affecting DAMP signaling. The full-length signal-transducing RAGE binds DAMPs, eliciting inflammatory reactions; sRAGE captures DAMP-related ligands and inhibits the intracellular signal transductions as a decoy-type receptor (Bongarzone et al., 2017). sRAGE plays an important role in protecting organs in various pathologic conditions, such as acute lung injury, diabetic atherosclerosis, Alzheimer's disease, and septic shock (Park et al., 1998; Yamamoto et al., 2011; Sugihara et al., 2012; Blondonnet et al., 2017).

Although the relationship between RAGE and the pathophysiologic features of various kidney diseases, such as chronic unilateral ureteral obstruction and autosomal dominant polycystic kidney disease, has been reported, the effect of RAGE/sRAGE on AKI remains unclear (Gasparitsch et al., 2013; Lee et al., 2015). We therefore explored the pathophysiologic role of RAGE and sRAGE in a mouse model of AKI.

To assess whether RAGE or sRAGE was involved in renal tubular damage, we induced unilateral renal I/R injury both in Ager^−/− mice and WT control mice. Tubular damage, interstitial cell accumulation, and fibrosis were analyzed 2 and 7 days after I/R injury. Tubular damage, which consisted of cast formation, tubular necrosis, loss of the brush border, and tubular dilatation, was more severe in Ager^−/− mice than in control mice 2 and 7 days after I/R injury (Fig. 1A,B; P<0.05). Sirius Red staining showed that kidney fibrosis was more severe in Ager^−/− mice than in the control mice on day 7, although hydroxyproline levels were not significantly increased (Fig. 1C,D; P<0.0001). Furthermore, significant macrophage infiltration was also exaggerated in Ager^−/− mice in comparison with the control mice on day 2 (Fig. 1E,F; P<0.05).

To examine the expression level of each RAGE isoform in damaged tubular epithelial cells (TECs), we analyzed the gene expression of full-length RAGE and esRAGE in murine renal proximal tubular epithelial (mProx24) cells subjected to hypoxia (Fig. 2A). Hypoxia decreased the expression of genes coding full-length RAGE and esRAGE in mProx24 cells over time (Fig. 2B,C; P<0.05).

We then evaluated the expression of genes coding for pro-inflammatory mediators after hypoxia was induced in mProx24 cells (Fig. 3A). Hypoxia caused mRNA upregulation of Hmgb1 and Tnfa in mProx24 cells (Fig. 3B). However, the addition of sRAGE decreased Hmgb1 and Tnfa mRNA levels in hypoxic mProx24 cells (Fig. 3B). We also assessed cellular damage in hypoxic mProx24 cells with or without the addition of sRAGE (Fig. 3C). The sRAGE induced proliferation of hypoxic mProx24 cells (Fig. 3D; P<0.05). The addition of sRAGE induced the proliferation of TECs after their subjection to hypoxia and was most effective at 40 µg/ml (Fig. 3E; P<0.05). Moreover, we assessed the expression of genes coding for pro-inflammatory mediators after hypoxia in primary TECs (Fig. 3F). Hmgb1 and Tnfa mRNA were expressed more in primary TEC-derived Ager^−/− mice after hypoxia than in the control mice (Fig. 3G; P<0.05).

We then explored the renoprotective effects of sRAGE in vivo. Unilateral renal I/R was induced in WT mice with or without sRAGE administration. Tubular damage was assessed 2 days later (Fig. 4A). Among the B6 mice subjected to I/R, sRAGE administration reduced tubular damage more than that caused by phosphate-buffered saline (PBS; Fig. 4B,C; P<0.05).

Next, to determine whether sRAGE has therapeutic potential in another model, we assessed tubular damage from anti-glomerular basement membrane (anti-GBM) glomerulonephritis in mice on day 7, in accordance with the protocol that comprised of preparing anti-GBM glomerulonephritis model mice, administering sRAGE, and then sacrificing them on day 7. (Fig. 5A). Kidneys from WT mice that received sRAGE and from Ager^−/− mice were evaluated with the injection of anti-GBM antibody. Tubular damage was more severe in Ager^−/− mice than in WT mice. In mice that received sRAGE, tubular damage was less severe than in the other two groups (Fig. 5B,C).

This study aims at exploring the pathophysiologic role of RAGE and sRAGE in tubular injury in vivo and in vitro. In an AKI model, we demonstrated that renal tubular damage was more severe in Ager^−/− mice than in control mice. The expression of full-length RAGE and esRAGE mRNA was downregulated in TECs under conditions of hypoxia. Furthermore, the administration of sRAGE caused downregulation of pro-inflammatory mediators and induced proliferation of hypoxic TECs. Finally, sRAGE protected against renal tubular damage in models of both AKI and anti-GBM glomerulonephritis.

Recent studies have revealed that RAGE is involved in the pathogenesis of I/R-induced injury to the lungs (Sharma et al., 2013), heart (Wang et al., 2018), and brain (Liu et al., 2020). However, few studies have focused on the pathophysiologic roles of RAGE in AKI. We demonstrated that renal tubular damage was exacerbated in Ager^−/− mice subjected to I/R in comparison with control mice and was less severe in WT mice subjected to I/R and sRAGE administration. In support of our hypothesis, sRAGE reportedly acted as a decoy in inhibiting RAGE-related inflammatory signal transduction (Bongarzone et al., 2017). Moreover, sRAGE expression was abundant in mouse kidneys, although the expression levels of isoforms of RAGE differed according to the species and organs (Harashima et al., 2006; Jules et al., 2013).

RAGE was reported to be upregulated in various types of chronic kidney disease, such as diabetic nephropathy, hypertensive nephropathy, and obesity-related glomerulopathy (D'Agati and Schmidt, 2010). The serum levels of sRAGE are also known to increase in patients with ESRD (Kalousova et al., 2006). We demonstrated that the expression of full-length RAGE and esRAGE mRNA was downregulated by hypoxia induced in mProx24 cells over time. This result suggests that acute hypoxia induces downregulation of RAGE and sRAGE in the renal tubules, which results in failure to inhibit inflammatory signal transductions and may exacerbate renal tubular damage. However, the molecular mechanisms of RAGE and sRAGE expression remain unclear in TECs subjected to hypoxia.

Hypoxia induction was reported to cause upregulation of pro-inflammatory mediators, such as HMGB1, TNF-α, and monocyte chemotactic protein-1 in the kidneys and human leukocytes (Bai et al., 2018; Zhang et al., 2019). We obtained similar results in hypoxic mProx24 cells, and the addition of sRAGE downregulated the mRNA expression of pro-inflammatory mediators. Furthermore, the induction of hypoxia has been shown to inhibit the growth of rat TECs (Zhu et al., 2018). We found that cell proliferation was also inhibited in hypoxic mProx24 cells; moreover, cell proliferation was restored by the addition of sRAGE. This result suggests that sRAGE may reduce the inflammation and enhance the proliferation of TECs in conditions of hypoxia. In addition, the expression of pro-inflammatory mediators was enhanced in Ager^−/−-derived primary TECs in comparison with WT-derived primary TECs. These findings would support our hypothesis that sRAGE may protectively act against I/R-induced AKI.

We showed that sRAGE treatment also reduced renal damage in a mouse model of AKI. In support of our hypothesis, sRAGE administration has been reported to reduce acute organ damage, such as acute lung injury, and sepsis (Yamamoto et al., 2011; Blondonnet et al., 2017). Moreover, it is known that anti-GBM glomerulonephritis causes not only glomerular damage, but also tubular damage (Andres et al., 1978). We confirmed the therapeutic effect of sRAGE in models other than tubular damage caused by I/R injury. These results indicated that sRAGE has protective roles in tubular injury, regardless of the underlying cause of tubular damage. Thus, sRAGE could be a promising therapeutic option for kidney injury.

This study had several limitations. First, the molecular mechanisms involved in RAGE and AKI caused by I/R have not been fully evaluated. Because HMGB1 is involved even in aseptic inflammation, such as that caused by I/R, we assume that the HMGB1–RAGE signal was responsible for kidney injury in our model. However, we did not analyze the interaction between RAGE and ligands such as HMGB1. Moreover, the roles of other pattern recognition receptors, such as Toll-like receptors, need to be elucidated.

In conclusion, we have shown the protective role of sRAGE in renal I/R-induced tubular damage (Fig. 6). Hypoxic stimulation downregulated the expression of full-length RAGE/esRAGE in TECs, which in turn might reduce the capture ability of DAMPs, such as HMGB1. These uncaptured DAMPs would have exacerbated tubular injury in our model. Hence, sRAGE administration showed a renoprotective effect in tubular injury. These findings help clarify molecular mechanisms and indicate novel therapeutic options for AKI.

We purchased male C57BL/6J (WT) mice from Charles River Japan (Yokohama, Japan). RAGE-deficient [Ager knockout (Ager^−/−); C57BL/6J background] mice were produced by crossbreeding heterozygous mutant mice (Myint et al., 2006). The C57BL/6J and Ager^−/− mice were housed and bred at Kanazawa University, Kanazawa, Japan. All animal experiments were conducted in accordance with the guidelines for animal care of Kanazawa University and were approved by the Institute for Experimental Animals, Kanazawa University Advanced Science Research Center (registration number, AP-153391). This study is reported following the recommendations of the ARRIVE guidelines (https://arriveguidelines.org).

We induced kidney I/R injury as previously described (Iwata et al., 2012). To induce ischemia, we clamped the left renal pedicle with a nontraumatic clip (Natsume Seisakusho, Tokyo, Japan) after the induction of anesthesia. The clip was removed after 50 min. To control body temperature, we placed the mice on a 37.0°C heating pad throughout the procedure.

We prepared mouse glomerulus according to the method of Krakower and Greenspon (Krakower and Greenspon, 1951) and anti-mouse GBM antibodies as previously described (Wada et al., 1996). To confirm specificity, we performed in vitro indirect immunofluorescence using fluorescein isothiocyanate-conjugated anti-rabbit immunoglobulin G (Organon Teknika Corporation, Durham, NC, USA) on frozen sections of normal mouse kidneys. Sharp linear immunofluorescence was obtained along the GBM.

We induced anti-GBM glomerulonephritis in mice as previously described (Wada et al., 1996) with some modifications. In both WT and Ager^−/− mice, 0.5 ml of nephrotoxic serum was intraperitoneally injected on days –1 and 0. Each mouse was sacrificed on day 7. The mice with anti-GBM glomerulonephritis were divided into three groups: the WT group, in which WT mice received vehicle PBS by intraperitoneal injection on days 1, 3, and 5; the Ager^−/− group, in which Ager^−/− mice received vehicle PBS by intraperitoneal injection on days 1, 3, and 5; and the WT+sRAGE group, in which WT mice received recombinant sRAGE (50 μg each time) by intraperitoneal injection on days 1, 3, and 5.

Recombinant mouse soluble RAGE was provided by Yasuhiko Yamamoto (Kanazawa University, Kanazawa, Japan) (Yamamoto et al., 2011).

The murine TEC line mProx24 was provided by Takeshi Sugaya (St. Marianna University School of Medicine, Tokyo, Japan). In each in vitro experiment, the cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Thermo Fisher Scientific, Waltham, MA, USA) with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin for cell culture and in DMEM with 1% FBS after a 24-h period of serum starvation.

Primary murine renal TECs from WT and Ager^−/− mice were generated in accordance with the method described in Wuthrich et al. (1990), with some modifications. The kidneys were stripped of blood cells by being washed in saline. The kidney cortices from WT and Ager^−/− mice were cut into small pieces and then digested in defined K1 medium containing 4 mg/ml of collagenase at 37°C for 1 h. The digested kidney pieces were washed with a cell strainer (mesh diameters of 100 and 40 μm). The cortical tubular cells were spun down at 1500 rpm for 5 min and washed again. The cell pellet was resuspended in defined K1 medium. The cell suspension was placed on cell culture Petri dishes and incubated at 37°C. The experiments were performed after the cells had reached 80%–90% confluence. Primary TECs were stimulated in serum-free K1 medium for 10 h starvation and then placed in a chamber with 5% O₂ for 8 h according to the protocol (Fig. 3F).

A sample of kidneys from each mouse was fixed in 10% buffered formalin (pH 7.2) and embedded in paraffin. We stained 5-μm sections with periodic acid–Schiff and Azan reagents. The percentage of proximal tubules at the corticomedullary junction that displayed proximal tubule dilation, a loss of the brush border, the presence of casts, and cellular necrosis were counted. The specimens were evaluated in a blinded manner from at least 10 different kidney sections (magnification, ×200) for each sample. Staining with F4/80 (catalog no. MF48000; Invitrogen, Carlsbad, CA, USA) and Sirius Red stain was performed as previously described (Iwata et al., 2012).

To isolate total cellular RNA from the cultured cells, we used the High Pure RNA Isolation Kit (Roche Diagnostics K. K., Tokyo, Japan) and the ISOSPIN cell and tissue RNA (NIPPON Gene, Tokyo, Japan). We performed quantitative real-time polymerase chain reaction (PCR) analysis with iQ™ SYBR^® Green Supermix (catalog no. 170-8885; BioRad, Hercules, CA, USA) using the ViiA™7 Real-Time PCR System (Thermo Fisher Scientific). The following primers were used: Ager (catalog no. Mm_Ager_1_SG, QuantiTect Primer Assay; Qiagen, Hilden, Germany), HMGB1 (catalog no. Mm00849805_gH, TaqMan Gen Expression Assay; Applied Biosystems, Foster City, CA, USA), Tnfa (catalog no. Mm_TNF_1_SG, QuantiTect Primer Assay; Qiagen), and Actb (catalog no. 4352341E, Mouse ACTB Endogenous Control; Applied Biosystems). Data were analyzed according to the delta-delta Ct method (Iwata et al., 2017). Semiquantitative PCR for full-length RAGE, esRAGE, and β-actin were performed. The PCR products were analyzed by electrophoresis on 2% agarose in 1× TAE buffer (100 V, 20 min). Data were analyzed using the ImageJ software (https://imagej.nih.gov/ij/).

TEC proliferation was determined using the Cell Counting Kit-8 (catalog no. CK04; Dojindo, Kumamoto, Japan) in accordance with the manufacturer's instructions, as previously described (Iwata et al. 2010).

I/R-injured kidneys were taken from each mouse to assess the amount of kidney collagen. The hydroxyproline assay ware performed according to the standard protocol of our laboratory, as previously described (Sakai et al., 2013). Assay results were expressed as micrograms of hydroxyproline per kidney.

The data are expressed as means±s.e.m. To perform the statistical analysis, we used the two-tailed unpaired student's t-test to compare the two groups; we also used one-way analysis of variance with Tukey's multiple comparison test to compare more than two groups (GraphPad Prism 8 software). P-values <0.05 were considered statistically significant.

We thank Takeshi Sugaya for providing the mProx24 cell line.