All experiments were performed using male C57BL/6 mice, which were obtained from the Animal Center of Fudan University, Shanghai, China. MiR-21 knockout mice were generated at the Shanghai Model Organisms Center as described previously (Jia et al., 2017). Adult mice (8- to 10-week-old; 20–25 g) used were housed in temperature- and humidity-controlled cages, with free access to water and food, under a 12 h light/dark cycle. All animal experiments were approved by the Institutional Animal Care and Use Committee of Fudan University and were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Aristolochic acid I sodium salt (Sigma) was diluted in saline (0.5 mg/ml) and was administered to mice by a single intraperitoneal injection at a bolus dose of 10 mg/kg to establish an AKI-to-CKD model. The mice were sacrificed at days 1, 3, 7 14, or 28 after injected with AA.

To study the role of Wnt/β-catenin, two experiments were performed. In the first experiment, ICG-001-phosphate (5 mg/kg) was dissolved in dimethyl sulfoxide (DMSO) at a concentration of 4 mg/dl and was administered daily to mice via intraperitoneal injection for consecutive 7 days, starting 24 h after AA delivery or 7 days after AA delivery. In the second experiment, animals were injected intraperitoneally of LiCl (Sigma, 20 mg/kg/d, diluted in normal saline as 10 mg/ml), starting from 24 h post AA treatment, for 7 consecutive days. Mice were sacrificed and kidney tissues were collected for indicated analyses 14 and 28 days post AA administration, or at indicated time points.

For testing the effects of pharmacologic inhibition of miR-21, locked nucleic acid- (LNA-) modified anti-miR21 oligonucleotides or anti-scramble (Exiqon) was dissolved in saline (5 mg/ml) and administered via the tail vein (10 mg/kg) within 60 min prior to AA delivery (Xu et al., 2012). Additional injections (the same dose) were given on the 5th and 10th day after AA-I dosing. All the experiments were replicated at least twice.

MiR-21 knockout mice were generated at the Shanghai Model Organisms Center, Inc. (Jia et al., 2017). In brief, an frt-PGK-Neo-frt cassette followed by 5′ loxP site was introduced upstream of miR-21 gene and 3′ loxP site was inserted downstream of miR-21 gene. The targeting vector was introduced into SCR012 ES cells. Targeted ES clones were microinjected into ICR embryos and transferred into pseudopregnant ICR females. After bred with C57BL/6 mice, these heterozygous mice were mated with ACTFLPe transgenic mice to remove Neo cassette. After backcrossed with C57BL/6, heterozygous mice (miR-21 ^flox/+) were mated with EIIA-Cre transgenic mice to obtain miR-21 knockout mice (miR-21 ^+/−) which were mated to generate homozygous mutants (miR-21 ^−/−). Then, homozygous mutant mice mated with each other to obtain more miR-21 ^−/− mice. miR-21 knockout mice were identified by the following primer pairs: Forward, 5′ -CAG​AAT​TGC​CCA​GGC​TTT​TA -3′; Reverse, 5′- AAT​CCA​TGA​GGC​AAG​GTG​AC -3′.

HK-2 human proximal tubular epithelial cells were cultured in DMEM/F12 supplemented with 10% fetal bovine serum in an incubator with 5% CO₂ at 37°C.

For AA treatment, the cells were cultured in 6-well plates and randomly assigned to experimental groups. Cells were cultured in complete medium with AA (0, 1, 2, 5, 10 μg/ml) for 24, 48, and 72 h, for indicated in-vitro experiments. In order to suppress in-vitro miR-21 levels when HK-2 cells reached 60–70% confluency, 50 nM anti-miR21 or 50 nM anti-scramble (Exiqon) was respectively transfected into cells 6 h prior to AA treatment (2 μg/ml) by using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions, following which the medium was changed and cells were harvested 48 h after AA treatment. In order to inhibit β-catenin, HK-2 cells were exposed to different concentrations of DKK-1 (50, 100, 150 ng/ml) diluted in DMSO for 1 h, prior to 48 h AA exposure (2 μg/ml). All the experiments were replicated at least twice.

Interfering RNAs (siRNAs) targeting Wnt1, Wnt4 and NF-κB subunit p65 (RelA) or an RNA duplex with random sequence as negative control (NC) were transfected into HK2 cells at 50 nM 6 h (for Wnt1/Wnt4 siRNA) or at 100 nM12 h (for NF-κB p65 siRNA) prior to AA treatment (2 μg/ml) by using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. The RNA transfection efficiency was 65–80% for at least 48 h. RNA oligoribonucleotides were obtained from GenePharma (Shanghai, China) and Ribo Life Science (Suzhou, China) and the sequences are listed in Table 1.

Changes in renal morphology were examined in paraffin-embedded tissue sections (4 μm) stained with hematoxylin. Immunohistochemical staining was performed as previously described (Xu et al., 2012). The following primary antibodies were used: anti-Wnt1 (1:100, Abcam), anti-Wnt4 (1:150, Santa Cruz Biotechnology), anti-c-caspase3 (1:500, Cell Signaling Technology), anti-F4/80 (1:200, Abcam). The secondary antibody was horseradish peroxidase-conjugated anti-mouse IgG (1:1,000; Jackson ImmnoResearch), which was used according to the manufacturer’s instructions. The peroxidase was visualized with diaminobenzidine. To determine the number of F4/80-positive cells, 8–10 fields for each mouse were examined under a microscope (magnification, 200×), and the average number of positive cells was calculated.

The relative protein abundances in kidneys and HK-2 cells were analyzed using western blotting, as described previously (Xu et al., 2012). The following primary antibodies were used: anti-α-SMA (1:1,000, Sigma), anti-E-cadherin (1:1,000, Cell Signaling Technology), anti-vimentin (1:1,000, Santa Cruz Biotechnology), anti-collagen I (1:1,000, Abcam), anti-collagen IV (1:1,000, Abcam), anti-Wnt1 (1:100, Abcam), anti-Wnt4 (1:150, Santa Cruz Biotechnology), anti-Wnt3 (1:500, Abcam), anti-Wnt2b (1:700, Abcam), anti-Wnt7a (1:300, Abcam), anti-β-catenin (1:1,000, Cell Signaling Technology), anti-Snail 1 (1:1,000, Santa Cruz Biotechnology), anti-PAI-1 (1:1,000, Santa Cruz Biotechnology), and anti-NF-κB p65 (1:1,000, Cell Signal). Horseradish peroxidase-conjugated secondary antibodies (anti-rabbit or anti-mouse or anti-goat IgG, 1:5,000, Jackson ImmunoResearch) were used according to the manufacturer’s instructions. Image analysis software (Image J, National Institutes of Health) was used to determine the gray value of all bands.

Total RNA was isolated from kidney tissues and cells using TRIzol RNA Isolation System (Invitrogen). MiR-21 expression was quantified by RT-qPCR using Taqman chemistry (Applied Biosystems), as described previously, and was normalized to U6 small nuclear RNA expression (Jia et al., 2017). qRT-PCR was performed on an ABI PRISM 7000 Sequence Detection System (Applied Biosystems, Foster City, CA, United States). mRNA levels of SOD2, PGC-1α, and Mpv17 were quantified using SYBR Green PCR Master Mix (Applied Biosystems), as previously described (Xiao et al., 2016), and were normalized to the level of 18S rRNA. The 2^–ΔΔCt method was used to determine relative changes in mRNA and miR-21 expression. Relative gene levels were expressed as ratios to the control.

Kidneys were fixed in 10% formalin overnight, dehydrated with an ethanol gradient, cleared in xylene, embedded in paraffin, and cut into 4 μm-thick sections. Histopathological changes in the corticomedullary junction were assessed by H&E and Masson’s trichrome staining. Tissue damage was scored according to the following scale: no injury, “0”; mild (<25% staining) “1”; moderate (<50%) “2”; extensive (>50%) “3”; very severe (>75%) “4.” An average of 10 fields for each mouse were examined under a microscope at a magnification of 200× or 400× (Xu et al., 2012).

The 500 bp 3′UTR of human Wnt2b containing potential miR-21 binding site and its mutant were amplified by PCR and cloned into pMIR-REPORT vector, respectively. For luciferase reporter assays, cells were transiently transfected with pMIR-REPORT vectors together with indicated miR mimics using Lipofectamine 2000 for 24 h. Reporter activity was measured by the dual-luciferase assay-system (Promega). Renilla luciferase activity was used to normalize for transfection efficiency. The data were presented as fold change relative to the control group.

Mouse renal tubular epithelial cells (mTECs) with indicated treatments were cross-linked with 1% formaldehyde for 10 min at 37°C and chromatin immunoprecipitation (ChIP) assay was performed using the Pierce Agarose ChIP Kit from Thermo (26156). Five micrograms of NF-κB p65 (Cell Signal) antibody was added for each assay and normal IgG were used as negative control. The miR-21 promoter region covering the potential NF-κB binding site (−900 ∼ −1,200 bp relative to its start codon in human miR-21 gene, −1 ∼ −300 bp in mouse miR-21 gene) was assayed for NF-κB binding by PCR using Premix Taq™ from Takara (RR900Q) and following primers. ATA​CTT​TTT​CCT​TCT​GTG​CTA​AGG​T (forward) and ACA​TGA​TAA​ACA​TGC​AAG​ACT​GTT​A (reverse) for human; CAG​AGG​ACA​CTA​GCA​AGA​AAG​GCT​T (forward) and GCC​ATG​CGA​TGT​CAC​GAC​CAC​GAC​A (reverse) for mouse. The PCR products were analyzed by 2% agarose electrophoresis.

Blood was sampled from mouse eyes at the indicated times. Serum creatinine was detected by the improved Jaffe method using a Quantichrom™ creatinine Assay Kit (BioAssay Systems).

GSH-Px activity in the mouse kidney was detected using a glutathione peroxidase (GSH-PX) assay kit (colorimetric method; Nanjing Jiancheng Bioengineering Institute) per the manufacturer’s instructions. The change in absorbance at 412 nm during the conversion of GSH to oxidized glutathione was measured.

The suspension cytokine arrays were used to verify the concentration of cytokines in mice samples. The mouse cytokines of IL-10, IL-13, Eotaxin, chemokine (C-X-C motif) ligand 1 (CXCl-1), monocyte chemotactic protein 1 (MCP-1), macrophage inflammatory protein-1a (MIP-1a), and macrophage inflammatory protein-1b (MIP-1b) were quantified using Mouse 23-plex Multi-Analyte Kit (Bio-Plex Suspension Array System; Bio-Rad, Hercules, CA, United States) according to the manufacturer’s instructions. In brief, 50 μL of sample was added to each well, and incubated with capture antibody-coupled beads in the dark at room temperature with shaking at 850 rpm for 2 h. After washing with 3 times, 50 μL biotinylated antibody was added and then incubated for 1 h. The captured cytokines were visualized with streptavidin-phycoerythrin (PE). Finally, the plate was read using a Bio-Plex MAGPIX Multiplex Reader (Bio-Rad, Hercules, CA, United States). Bio-Plex Manager 6.0 software was used for data acquisition and analysis.

The statistical software SPSS version 20.0 (SPSS Inc.) was used for statistical analysis. For continuous variables, data are presented as the mean ± SD. Multigroup differences were compared by one-way analysis of variance followed by Bonferroni’s post-hoc test. p < 0.05 was considered significant.

After a single intraperitoneal administration of aristolochic acid (10 mg/kg), the mice were sacrificed at different time points (days 1, 3, 7, 14, 28 after AA delivery), and a drug-related kidney injury model was successfully established. The mouse serum creatinine and the miR-21 level began to rise at 3 days post AA administration, peaked at day 14, and maintained a high level thereafter (Supplementary Figures S1A,B). Kidney histological examination revealed comparable results: severe renal tubular damage occurred on day 3, renal tubular atrophy along with interstitial fibrosis were observed on day 14, and more obvious on day 28 (Supplementary Figures S1C–E).

Next, we investigated the expression of miR-21 in cultured tubular epithelial (HK-2) cells treated with different doses of AA to delineate the mechanisms driving AKI-to-CKD progression. Even at the lowest doses of AA tested (1 or 2 μg/ml), miR-21 abundance in HK2 cells started to increase within 48 h after treatment, and then returned to baseline (Supplementary Figure S1F). Moreover, low-dose AA could induce cell phenotypic changes on account of the activation of some markers of epithelial-mesenchymal transition (EMT), such as α-smooth muscle actin (α-SMA), collagen I, and collagen IV. However, there was no change in the protein abundance of vimentin (Supplementary Figures S1G,H).

We knocked down miR-21 by using locked nucleic acid (LNA) modification technology. Compared with the anti-scramble + AA group at each time point, the abundance of miR-21 was significantly reduced in the anti-miR21 + AA group (Figure 1A). However, the decrease in serum creatinine level was more significant in the late stage of AA intervention (day 14 and day 28) (Figure 1B). Both H&E staining and Masson’s trichrome staining revealed a similar pattern to that of serum creatinine on day 14 and 28 (Figures 1C–E). The pathological changes in the anti-scramble + AA and AA groups were manifested as larger area of renal tubular necrosis, massive monocytes infiltration, and severe renal tubulointerstitial fibrosis (Figure 1C). And inhibition of miR-21 partially reduced kidney injury and renal fibrosis (Figures 1D–F).

We first verified the dynamic expression patterns of miR-21 and Wnt/β-catenin signaling in mouse kidneys at different time points after high-dose AA (10 mg/kg) administration. The expression of Wnt ligands was evaluated on days 1, 3, 7, 14, and 28 after AA administration (Figure 2). The western blot analyses revealed that protein abundances of Wnt 1 and Wnt 4 displayed similar patterns to that of miR-21 (Figures 2A–C; Supplementary Figure S1B). Comparable results were obtained by immunohistochemical staining (Figures 2H–J). Furthermore, the two ligands were predominantly distributed in the renal tubule and interstitium, suggesting that they might be involved in the development of renal tubulointerstitial fibrosis. In contrast, other Wnt ligands, including Wnt 2b, Wnt 3, and Wnt 7a, exhibited in an opposite trend, reaching minimum expression values on day 28 (Figures 2A–F). Notably, western blotting also revealed that the renal β-catenin protein level gradually increased, lagging behind Wnt1 or Wnt4, peaking on day 28 (Figure 2G). Together, these data indicated that both miR-21 and Wnt/β-catenin were upregulated and likely involved in kidney fate after acute insult of AA exposure.

We also investigated the expression of Wnt1/4-mediated β-catenin pathways in HK-2 cells treated with different doses of AA. A trend similar to miR-21 was observed. The protein levels of Wnt1, Wnt4, and β-catenin in HK2 cells started to increase within 48 h after treatment, and then returned to baseline (Supplementary Figure S1G).

We further examined the potential connection between activation of miR-21 and Wnt/β-catenin signaling in a mouse model of AA-induced nephropathy (AAN). As shown in Figures 3A–D, protein levels of Wnt1, Wnt4, and β-catenin declined in the later stage (i.e., day 14 and 28) after anti-miR21 intervention, which was confirmed by immunohistochemistry (Figures 3H–J). Wnt2b and Wnt7a showed expression trends opposite to that of miR-21, and Wnt3 protein expression varied (Figures 3E–G), which suggested that miR-21 may have a direct negative effect on Wnt2b and Wnt7a.

Wnt2b is the only Wnt ligand predicted as a target of miR-21 by TargetScan v7.2 (http://www.targetscan.org/), but this remaines to be experimentally confirmed. A 3′-UTR luciferase reporter assay was performed to directly examine the 3′-UTR-mediated interaction between miR-21 and Wnt2b. The miR-21 mimic, compared to control, significantly reduced Wnt2b 3′-UTR activity by 95%, but have no significant effects on mutated Wnt2b 3′-UTR (Supplementary Figure S2A). The reciprocal repression relationship between miR-21 and Wnt2b in our AAN model further confirmed Wnt2b as a target of miR-21.

In line with the results of the in-vivo study, the protein levels of Wnt1, Wnt4, and β-catenin in HK-2 cells were attenuated following miR-21 inhibition. Meanwhile, protein abundances of α-SMA, collagen I, and collagen IV in HK2 cells were evidently decreased, suggesting that suppression of miR-21 could partially reverse the phenotypic changes induced by AA (Figures 3K,L).

To clarify the effect of miR-21 on Wnt/β-catenin signaling, we next studied the potential mechanism in miR-21^−/− mice. As shown in Supplementary Figure S2B, all our experimental mice were genetically identified as miR-21^−/− mice. Compared to wild-type mice, miR-21 knockout mice presented decreased serum creatinine levels on day 7 and 14 after AA administration (Figures 4A,B). Both H&E staining and Masson’s trichrome staining revealed a histological alleviation of kidney injury, with less renal tubular cast formation, tubulointerstitial fibrosis, and inflammatory cells infiltration in the miR-21^−/− mice (Figures 4C–E). Correspondingly, the upregulation of protein levels of wnt1, wnt4, and β-catenin, as well as of fibrotic indicators such as α-SMA and collagen I following AA exposure was attenuated in miR-21 KO mice (Figures 4F–L). Without AA exposure, there were no significant differences between the two groups in these parameters. Genetic deletion of microRNA-21 in vivo thus protected against AA-induced kidney injury and mitigated the detrimental profibrotic effects via downregulation of the canonical Wnt pathway.

Considering the imperative role of β-catenin activation in driving AA-associated AKI-to-CKD transition, we speculated that targeted inhibition of this signaling may impede renal fibrogenic action followed by AKI. To this end, a small-molecule inhibitor (ICG-001) was applied to selectively inhibit gene transcription mediated by β-catenin (Eguchi et al., 2005; Henderson et al., 2010). ICG-001 (5 mg/kg) was administered daily for consecutive 7 days to mice, starting from 24 h after AA administration (early ICG-001 treatment group, ICG-e group), or starting from 7 days after AA administration (late ICG-001 treatment group, ICG-l group). The mice were sacrificed on day 14 or 28 after AA administration (Figure 5A). Interestingly, ICG-001 treatment markedly downregulated protein expression of β-catenin and α-SMA at each time point in both groups (Figures 5B–D). However, a decrease in serum creatinine was observed only in the ICG-e group on day 28 (Figure 5I), not on day 14, suggesting that timely inhibition of β-catenin was critical to effectively improve kidney recovery upon AA administration. Correspondingly, mice that received early intervention with ICG-001 showed less tubular damage and less interstitial fibrosis on day 28 than on day 14 (Figure 5H, Supplementary Figures S2C,D). As there was no significant difference in miR-21 abundance among the groups (Figure 5J), the Wnt/β-catenin pathway may have no reverse regulatory effect on miR-21.

To evaluate whether the antifibrotic effect of ICG-001 was independent of the severity of AKI, we observed dynamic changes in serum creatinine and renal histology on days 1, 3, and 7 after ICG-001 treatment. We observed no improvement in renal function or attenuation of renal fibrosis before day 7 (Supplementary Figures S2E–H). Thus, early treatment with ICG-001 protected against AA-induced renal fibrosis irrespective to the severity of AKI. Considering the serum creatinine levels, it seemed that early inhibition of β-catenin had limited effect on mouse kidney function in the acute phase post AA injury.

Treatment with either Wnt1 siRNA or Wnt4 siRNA 6h prior to AA treatment significantly suppressed collagen IV protein expression (Figures 6A–D), but there was no significant change in the level of miR-21 (Figures 6E,F). When HK-2 cells were incubated with different concentrations of Dickkopf-related protein 1 (DKK-1), another potent β-catenin signaling blocker (Davidson et al., 2002; Mao et al., 2002), β-catenin activation secondary to AA exposure was remarkably hindered by DKK-1 at 150 ng/ml (Figures 6G,H). In addition, DKK-1 inhibited the upregulation of α-SMA and collagen IV protein expression following AA exposure (Figures 6I,J). The vimentin protein level and miR-21 abundance still remained unchanged after DKK-1 treatment (Figures 6K,L). These findings suggested that inhibition of β-catenin signaling by DKK-1 could suppress AA-mediated fibrogenic action in HK-2 cells.

To explore the role of the Wnt/β-catenin pathway in renal fibrosis further, mice were treated with LiCl (20 mg/kg/d), a β-catenin signaling activator (Narcisi et al., 2016), for seven consecutive days. The efficiency of LiCl in enhancing β-catenin abundance was verified by western blot analysis, both on day 14 and day 28 post AA treatment. The abundance of β-catenin was up-regulated much more on day 28 compared to that on day 14 (Supplementary Figures S3A,B). In comparison with untreated mice, renal expression of collagen I, α-SMA, and Snail1 was significantly increased in the LiCl-treated group, in line with the increase in serum creatinine (Supplementary Figures S3C–E). Further, pathological changes in the kidney tissue revealed that the overexpression of β-catenin concomitant with LiCl treatment exaggerated inflammatory cell infiltration, and promoted extracellular matrix deposition and the development of renal fibrosis (Supplementary Figures S3H–J). Of interest, miR-21 abundance in kidney tissue was not altered when β-catenin was activated (Supplementary Figure 3G). Given the consistent results on miR-21 in vivo and in vitro, we inferred that miR-21 might not be directly regulated by the Wnt/β-catenin pathway. In short, exogenous induction of β-catenin promoted renal fibrosis secondary to AA-induced AKI.

To demonstrate the role of the canonical Wnt pathway in the pathogenesis of AA-induced AKI further, we treated mice with both anti-miR21 and LiCl. The combined use of AA and LiCl (20 mg/kg, i.p.) exacerbated renal injury (serum creatinine: AA vs. AA + LiCl, 140.2 ± 15.3 μmol/L vs. 175.4 ± 9.6 μmol/L, p < 0.05) and renal fibrosis (Figures 7A,H–J) on day 14 post AA treatment. At 20 mg/kg/d for 7 consecutive days, LiCl did not abolish the anti-fibrotic renoprotective effects bestowed by anti-miR21, as indicated by quantitative western blot data and RT-qPCR results for miR-21 abundance showed that overactivation of β-catenin by LiCl treatment was substantially suppressed upon anti-miR21 treatment (Figures 7B–G).

Immune inflammation, extracellular matrix deposition, and oxidative stress, which are triggered by AKI, are all involved in the development of CKD (Ko et al., 2008; Rodríguez-Iturbe and García García, 2010; Dounousi et al., 2012; Meng et al., 2014). We found that either inhibition of miR-21 or early blockage of β-catenin signaling downregulated the protein abundance of NF-κB, a key transcription factor responsible for inflammatory and immune responses (Figures 8A,B). On the other hand, treating HK2 cells with NF-kappa B p65 siRNA 12 h before AA exposure resulted in significant decreases in the levels of pre-miR-21 and miR-21 (Figures 8C,D). Bioinformatics analysis (PROMO database: alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3; TESS database: www.cbil.upenn.edu/tess) showed that the miR-21 promoter region has potential NF-κB binding sites located at approximately −3,000 bp upstream of the pre-miR21 5′-end. To verify the direct binding of NF-κB to the miR-21 promoter, we carried out a ChIP assay using mTECs. As expected, mTECs showed considerable NF-κB binding to the predicted regions of miR-21 (Figure 8E), suggesting that NF-κB was the transcription factor responsible for miR-21 transcriptional activation in HK2 cells. However, when mice were treated with ICG-001 at a late stage during the course of AKI, this beneficial effect disappeared (Figures 8F,G). Additionally, immunohistochemistry revealed that the cleaved (c-)caspase-3 positively stained area and the number of F4/80-positive cells were reduced after inhibition of miR-21/Wnt signaling, which was in accordance with the expression of NF-κB (Figure 8J, Supplementary Figures S4A–D). Similar results were found for plasminogen activator inhibitor-1 (PAI-1) and Snail1, two Wnt/β-catenin downstream profibrotic genes (Figures 8H,I) (Xiao et al., 2016) implying that delayed delivery of ICG-001 was incapable of hindering β-catenin-mediated gene transcription implicated in renal fibrogenesis. As shown in Supplementary Figure S4E, in the ICG-e groups, a battery of proinflammatory cytokines, including eotaxin, chemokine (C-X-C motif) ligand 1 (CXCl-1), and macrophage inflammatory protein-1β (MIP-1b), were markedly reduced when compared with the respective levels in the ICG-l and AA groups. Glutathione (GSH) levels and relative mRNA levels of manganese superoxide dismutase (SOD2), peroxisome proliferator-activated receptor γ (PPARγ) coactivator 1α (PGC-1α), and mitochondrial inner membrane protein (Mpv17) in the kidneys were all attenuated on day 14 and 28 after early ICG-001 delivery or inhibition of miR-21 (Supplementary Figures S4F–M). Taken together, the data indicated that miR-21 inhibition and early β-catenin blockade ameliorated apoptosis, macrophage infiltration, and inflammatory cytokine storm.