Six to eight-week-old C57BL/6 male mice were purchased from Southern Medical University Animal Center (Guangzhou, China) and maintained under the standard environment with normal light/dark cycle and sterile conditions. Mice were performed by UUO or UIRI surgery, as described in previous studies (Zhou et al., 2018). For UUO model, the left ureter of C57BL/6 male mice was ligated by a 4–0 silk suture in a midline abdominal incision, under general anesthesia. Mice were randomly assigned to three groups: 1) sham control, 2) UUO, 3) UUO + SB225002 (Tocris Bioscience, No. 2725). SB225002 was intraperitoneally injected at 2.5 mg/kg/d for 7 days. Mice were euthanized 7 days after UUO operation. Kidney tissues were harvested for the various analyses.

For UIRI model, C57BL/6 male mice were carried out by an established protocol (Zhou et al., 2018). Firstly, left renal pedicle was clipped by a microaneurysm clamp (item no. 18051–35; Fine Science Tools, Cambridge, United Kingdom). In the whole process of ischemia, mouse body temperature was controlled between approximately 37°C and 38°C through a metal bath system. After removing the microaneurysm clamps, reperfusion of the left ischemic kidney was visually confirmed. 10 days later, the whole right kidney was excised under sterile conditions. Mice were randomly divided into four groups: 1) sham + control-shRNA, 2) UIRI + control-shRNA, 3) UIRI + control-shRNA + IL-8 (R&D systems, 208-IL), 4) UIRI + p16^INK4A-shRNA + IL-8. IL-8 was injected at 50 ng each mouse per day through the tail vein for 11 days. Mouse p16^INK4A-siRNA sequence (5′-CAC​CAG​AGG​CAG​UAA​CCA​UTT-3′) was constructed on an shRNA plasmid vector (control-shRNA). The aboved mice were injected with shRNA expression plasmid (control-shRNA or p16^INK4A-shRNA) by tail vein injection with a rapid and large volume plasmid solution as reported (Luo et al., 2018). Mice were euthanized 11 days after UIRI surgery. Serum and renal tissues were harvested for experimental analyses. All matters related to animal experiments were authorized by the Animal Ethics Committee at the Nanfang Hospital, Southern Medical University.

Human proximal tubular epithelial cells (HKC, clone-8) were provided by Dr. Lorraine C. Racusen (Johns Hopkins University, Baltimore, MD, United States), and were cultured as described previously (Zhou et al., 2015). HKC-8 cells were stimulated with recombinant human TGF-β1 protein [(R&D Systems, Minneapolis, MN, United States) (5 ng/ml)], SB225002 [(R&D) Systems, (0.2 μM)], or ICG-001 [(847591-62-2, Chemleader, Shanghai, China) (10 μM)], respectively. Human CXCR2 siRNA sequence was 5′-GAA​CCA​GAA​UCC​CUG​GAA​ATT-3′. The transfection of siRNA into HKC-8 cells was carried out with lipofectamine 2000 (Life Technologies, Carlsbad, CA, United States). Control plasmid (pcDNA3) or CXCR2 expression plasmid (pFlag-CXCR2) was also transfected into HKC-8 cells using lipofectamine 2000 (Life Technologies, Carlsbad, CA, United States), according to instructions.

Protein expression was detected by western blot analysis (Zhou et al., 2018). All primary antibodies used in this study were as follows: anti-fibronectin (F3648; Sigma-Aldrich), anti-active-β-catenin (19807T; Cell Signaling Technologies), anti-PGC-1α (ab54481; Abcam), anti-p16^INK4A (AFS779; BD Transduction Laboratories), anti-TFAM (PB0413; BOSTER), anti-TOMM20 (ab186735; Abcam), anti-COX1 (SAB1301619; Sigma-Aldrich), anti-p19^ARF (SC-1665; Santa Cruz Biotechnology), anti-CPT1A (ab128568; Abcam), anti-γ-H2AX (ab26350; Abcam), anti-CXCR2 (ab14935; Abcam), anti-CXCR1 (BA0755-1; BOSTER), anti-α-tubulin (RM 2007; Ray Antibody Biotech, Beijing, China), anti-β-actin (RM 2001; Beijing Ray Antibody Biotech), anti-GAPDH (RM 2001; Ray Antibody Biotech), anti-flag-tag (KM3002; Sungene Biotech Co.).

Total RNA from renal tissues and HKC-8 cells were extracted using TRIzol RNA reagent (Life Technologies, Grand Island, NY, United States). qRT-PCR experiment was carried out with an ABI PRISM 7000 instrument (Applied Bio-systems, Foster City, CA, United States). The primers used in qRT-PCR were shown in Supplementary Table S1.

Paraffin section (6 μm) were stained using Sirius Red Staining kit (DC0040; Leagene), according to the manufacturer’s instruction. Paraffin kidney sections (3 μm) were stained with periodic acid-schiff (PAS) reagent according to the instructions. Immunohistochemical staining was performed with standard protocol. All antibodies used were as follows: anti-CXCR2 (Ab61100; Abcam), anti-p16^INK4A (ab189034; Abcam), rabbit anti β-catenin (ab15180; Abcam), anti-fibronectin (F3648; Sigma-Aldrich), anti-γ-H2AX (ab26350; Abcam).

Immunofluorescence staining was performed with routine protocol (Zhou et al., 2018). Frozen kidney sections or the cover slides of cell culture were fixed with 4% paraformaldehyde at room temperature for 15 min. Primary antibodies were incubated after blocking with 10% donkey serum at room temperature for 1 h. All primary antibodies used were as follows: anti-fibronectin (F3648; Sigma-Aldrich), anti-TOMM20 (ab186735; Abcam), anti-γ-H2AX (ab26350; Abcam), anti-CXCR2 (Ab61100; Abcam), anti-β-catenin (610154; BD Transduction Laboratories), anti-Lotus Tetragonolobus Lectin (LTL) (FL-1321; VECTOR Laboratories), anti-Peanut Agglutinin (PNA) (FL-1071; VECTOR Laboratories), anti-Dolichos Biflorus Agglutinin (DBA) (FL1031; VECTOR Laboratories). Then the slides were treated with Cy3-or Cy2-conjugated secondary antibodies (Jackson Immuno-Research Laboratories, West Grove, PA, United States) for 1 h at room temperature. Nuclei were stained using DAPI (Sigma-Aldrich) under the manufacturer’s specifications.

Frozen sections (3 μm) or cells cultured on coverslips were assessed by senescence β-galactosidase activity (#9860; Cell Signaling Technology), MitoTracker deep red (M22426; Thermo Fisher), mitoSOX (M36008; Thermo Fisher) staining according to the specifications.

HKC-8 cells were fixed in 1.25% glutaraldehyde/0.1 mol/L phosphate buffer and harvested for making ultrathin sections (60 nm). The mitochondrial morphology was viewed using an electron microscope (JEOL JEM-1010, Tokyo, Japan).

All data were calculated as mean ± SEM. Statistical analyses of the data were performed by SPSS 25.0 (SPSS Inc., Chicago, IL, United States). One-Way ANOVA was used for comparison between groups followed by the Least Significant Difference or Dunnett’s T3 procedure. Unpaired t test was applied to analyze the differences between two groups. Bivariate correlation analysis was performed using Pearson correlation analysis. p < 0.05 represents statistical significance.

To identify the role of CXCR family in the pathogenesis of CKD, we first examined the expression of all members of CXCR family by qRT-PCR in mice at 7 days after UUO surgery. As shown, the mRNA levels of CXCR1-6 were unregulated, while CXCR7 was downregulated. Notably, among CXCR family, CXCR2 increased most (Figure 1A), and was time-dependently upregulated in UUO mice (Figure 1B). We next assessed the expression of CXCR2 in other CKD models. As shown, CXCR2 was also increased in the kidney of UIRI mice (Figure 1C). Consistently, immunohistochemical staining showed that CXCR2 protein was also highly increased in renal tubular epithelium in UUO and UIRI mice (Figures 1D,E).

To determine the exact location of CXCR2, the co-staining of CXCR2 with specific tubular cell markers of renal different segments were performed in UUO mice. Notably, CXCR2 was largely co-localized with lotus tetragonolobus lectin (LTL), a proximal tubule marker, and peanut agglutinin (PNA), a distal tubules marker, but less co-expressed with dolichos biflorus agglutinin (DBA), a collecting duct marker (Figure 1F).

To clarify the correlation between CXCR2 and mitochondrial function, we assessed the expression of the translocase of outer mitochondrial membrane 20 (TOMM20, a mitochondrial related marker) and CXCR2 by immunofluorescence. Interestingly, as shown in Figure 1G, compared with low expression of CXCR2 in sham kidneys, CXCR2 was greatly increased in UUO-affected kidneys. On the contrary, the expression of TOMM20 showed the opposite expressional fashion, suggesting the intimate relationship between CXCR2 and mitochondrial dysfunction. Quantification analysis confirmed the upregulation of CXCR2 and the downregulation of TOMM20 in UUO (Figures 1H,I). We next assessed cellular senescence and β-catenin. As shown in Figure 1J, CXCR2 was nearly completely co-localized with tubules with positive expression of p16^INK4A, a senescence marker. Furthermore, CXCR2 was also co-localized with β-catenin and active β-catenin in tubules (Figures 1K,L). The correlation analysis indicated that CXCR2 was positively correlated with p16^INK4A and β-catenin (Figures 1M,N). These results suggest that CXCR2 is highly involved in renal tubular cell senescence and mitochondrial dysfunction, and is associated with β-catenin activation.

To explore the potential role of CXCR2 in tubular senescence, UUO mice were injected with SB225002, a specific inhibitor of CXCR2 (Figure 2A). Groups of mice were sacrificed under anesthesia at 7 days after UUO. As shown in Figures 2B–D, the protein expressional levels of p19^ARF and fibronectin (FN) were significantly downregulated after administration of SB225002 in UUO mice. We also examined the expression of CXCR2 in SB225002-treated mice model. As shown in Figure 2E, CXCR2 was obviously suppressed according to the immunohistochemical staining analyses. We then performed Masson trichrome and Sirus Red staining. As shown in Figures 2E–G, treatment with SB225002 significantly inhibited the progression of renal fibrosis in UUO mice. Similar results were observed when FN was stained by immunostaining (Figures 2E,H). Then we tested cellular senescence by the staining of senescence-associated β-galactosidase (SA-β-gal) activity and γ-H2AX (Figures 2E,I,J) (Luo et al., 2018; Miao et al., 2019). Quantification analysis showed that inhibition of CXCR2 by SB225002 reduced renal fibrosis and tubular senescence. Similar results were observed when the mRNA level of p16 ^INK4A was analyzed by qRT-PCR (Figure 2K).

To examine the role of CXCR2 in inflammation, we analyzed the expression of TNF-α and MCP-1 by qRT-PCR. Results showed that inhibition of CXCR2 via SB225002 alleviated renal inflammation (Figures 2L,M). We also assessed the expression of CXCR1. As shown in Supplementary Figures S1A,B, although CXCR1 was increased in UUO mice, the protein level of CXCR1 was not affected after SB225002 treatment.

To further explore the underlying mechanisms of CXCR2 in renal fibrosis, we investigated β-catenin signaling and mitochondrial function. As shown in Figures 3A and B, administration of SB225002 significantly inhibited the expression of active β-catenin. Furthermore, the protein expressional levels of TOMM20 and the fatty acid shuttling enzyme carnitine palmitoyl-transferase 1A (CPT1A), a marker for mitochondrial function, were significantly preserved by SB225002 (Figures 3A, C,D). MitoTracker staining showed that SB225002 could inhibit the loss of mitochondrial mass (Figures 3E,F). Furthermore, consistent results were found when β-catenin and TOMM20 were tested by immunostaining (Figure 3F).

To find out a causal link between CXCR2 and tubular senescence, we assessed shRNA-mediated silencing of p16^INK4A in UIRI mice model injected with IL-8, a ligand of CXCR2 (Figure 4A). As shown in Figures 4B and C, serum creatinine (Scr) and blood urea nitrogen (BUN) were increased in UIRI mice, and further induced after administration of IL-8. However, these effects were suppressed by knockdown of p16^INK4A. Furthermore, we examined the expression of CXCR2. Obviously, immumohistochemical staining and qRT-PCR results showed that CXCR2 was further activated when administration of IL-8 in UIRI mice model (Figures 4D,E). We then performed periodic acid-Schiff (PAS) staining to distinguish damaged tubules by detecting glycogen content. As shown in Figure 4F, IL-8 treatment further exacerbated tubular injury in UIRI mice, but this was inhibited by knockdown of p16^INK4A. In addition, knockdown of p16^INK4A also significantly attenuated IL-8-induced upregulation of FN mRNA and protein expression (Figures 4G–I). Furthermore, staining of Sirus Red and immunofluorescence for FN disclosed that knockdown of p16^INK4A reduced renal fibrotic lesions in IL-8-treated UIRI mice (Figure 4J). As CXCR1 is also a receptor of IL-8, we examined the expression of CXCR1 in this mice model. Western blot analysis showed that CXCR1 was upregulated after UIRI surgery, however, administration of IL-8 had no effects on CXCR1 induction in UIRI mice (Supplementary Figures S1C,D).

We then assessed β-catenin signaling. As shown in Figures 5A–C, the immunostaining of β-catenin and western blot analyses for active β-catenin showed that IL-8 activated β-catenin singling, but inhibited by knockdown of p16^INK4A. We then investigated the protein expression of senescent-related marker γ-H2AX and mitochondria-related marker TOMM20. As shown in Figures 5D–F, IL-8 treatment further triggered the upregulation of γ-H2AX and downregulation of TOMM20 in UIRI mice, while knockdown of p16^INK4A significantly inhibited these effects. Consistent results were observed when SA-β-gal activity and TOMM20 were examined by staining (Figure 5G). Furthermore, mitochondrial mass analysis by MitoTracker staining showed that IL-8 treatment significantly aggravated mitochondrial dysfunction, but blocked by knockdown of p16^INK4A (Figures 5G, H). These data further suggest that CXCR2 could induce tubular senescence and mitochondrial dysfunction through β-catenin activation.

We further explored the role of CXCR2 in mediating cellular senescence and mitochondrial dysfunction in vitro. We adopted TGF-β1, a major fibrogenic factor, to treat human proximal tubular cell line (HKC-8) (Xi Liu et al., 2020). HKC-8 cells were pretreated with SB225002 for 1 h, then incubated with TGF-β1 for 24 h. As shown in Figures 6A and B, SB225002 could largely inhibit the upregulation of active β-catenin in TGF-β1-treated cells. We then explored mitochondrial dysfunction and cellular senescence. As shown, the protein levels of CPT1A and TOMM20 were restored by SB225002 in TGF-β1-treated HKC-8 cells (Figures 6A,C,D). Furthermore, the expressional levels of senescent-related marker p16^INK4A and fibrotic-related marker FN were reduced by SB225002 in TGF-β1-induced HKC-8 cells (Figures 6A,E,F). The staining also showed that SB225002 markedly reduced the expression of FN and γ-H2AX, and restored the expression of TOMM20, as shown in Figure 6G. Similar results were observed when CXCR2, p14, FN and P21 mRNA levels were tested by qRT-PCR (Figures 6H–K). To further study the role of CXCR2 in renal fibrosis, we transfected HKC-8 cells with siRNA to CXCR2 or negative control (siNC). The interference efficiency of CXCR2 was firstly confirmed (Figure 6L). As shown in Figures 6M,N, interference of CXCR2 blocked the expression of FN in TGF-β1 treated HKC-8 cells.

We further examined the effect of CXCR2 in tubular senescence and mitochondrial dysfunction in vitro. HKC-8 cells were transfected with CXCR2 expression plasmid. Overexpression of CXCR2 was first confirmed by western blot (Figures 7A, B). As shown in Figures 7A,C, overexpression of CXCR2 significantly promoted the upregulation of active β-catenin. Next, we examined the expressional levels of mitochondria-related proteins. As shown, the expressional levels of PGC-1α, the key transcription factor in mitochondria biogenesis, cytochrome c oxidase 1 (COX1), TFAM and TOMM20 were greatly decreased by transfection with CXCR2 expression plasmid (Figures 7A, D–G). In addition, CXCR2 overexpression aroused the expression of two senescent markers p16^INK4A and γ-H2AX (Figures 7A,H,I). CXCR2 also induced the upregulation of FN in HKC-8 cells (Figures 7A,J). Similar results were found when CXCR2 and p16 ^INK4A mRNA levels were assessed by qRT-PCR (Figures 7K,L). Moreover, as shown in Figure 7M, CXCR2 induced a significant increase in SA-β-gal activity and γ-H2AX nuclear foci in HKC-8 cells. In addition, we also test the SASP component TGF-β1 in CXCR2-overexpressed HKC-8 cells. Transfection with CXCR2 expression plasmid was significantly induced the expression of TGF-β1 (Supplementary Figure S2). These data further confirm that tubular senescence has an important role in the CXCR2-induced fibrogenic response.

HKC-8 cells were also stimulated with IL-8, a CXCR2 ligand. Treatment with recombinant protein IL-8 significantly upregulated the expression of CXCR2 (Figures 7N, O). Furthermore, IL-8 treatment downregulated the protein level of TOMM20 and upregulated the expression of p16^INK4A and FN (Figures 7N,P–R), suggesting the meditative role of CXCR2 in tubular senescence and mitochondrial dysfunction. IL-8 treatment also increased the activity of SA-β-gal (Figure 7S). These results suggest that CXCR2 is a strong inducer to cell senescence and mitochondrial dysfunction in renal tubular cell.

Our previous study found that β-catenin plays a central role in mitochondrial dysfunction and renal tubular senescence (Luo et al., 2018; Miao et al., 2019). To further study the downstream signals of CXCR2, we pretreated HKC-8 cells with ICG-001 (Yahong Liu et al., 2020; Zhou et al., 2021), an inhibitor of β-catenin. HKC-8 cells were pretreated with ICG-001, and then transfected with CXCR2 expression plasmid. As shown in Figures 8A–C, ICG-001 pretreatment greatly restrained CXCR2-induced increase in β-catenin and decrease in PGC-1α. In addition, the expression of p16^INK4A and γ-H2AX were also tested by western blot. As shown in Figures 8A,D,E, CXCR2-induced increase in p16^INK4A and γ-H2AX were greatly blocked by ICG-001 treatment. ICG-001 also repressed the expression of FN in CXCR2-overexpressed HKC-8 cells (Figures 8A,F). In addition, the expression of PGC-1α was also tested by qRT-PCR. As shown in Figure 8G, CXCR2-induced decrease in PGC-1α was greatly inhibited by ICG-001 treatment. Immunostaining of β-catenin showed that CXCR2 aggravated β-catenin activation in HKC-8 cells, however, treatment with ICG-001 largely inhibited it (Figure 8H). Treatment with ICG-001 markedly induced the expression of γ-H2AX, as assessed by immunofluorescence (Figure 8H). We further analyzed TOMM20, the production of mitochondrial ROS by mitoSOX staining and the ultrastructure of mitochondria using transmission electron microscopy (TEM). As shown in Figure 8H, ICG-001 greatly protected TOMM20 expression and inhibited mitochondrial ROS production in CXCR2-overexpressed tubular cells. Compared with normal mitochondria morphology in control group, CXCR2 overexpression induced the destruction of mitochondrial ultrastructure, resulting in the swelling, fragmentation and disorder of most mitochondria, while treatment with ICG-001 largely retained normal morphology of mitochondria (Figure 8H). These results further clarify that CXCR2 plays a key role in tubular senescence and mitochondrial dysfunction through activation of β-catenin.