2.1 Increased IL-33 and ST2 levels in DN subjects

To investigate how IL-33 signaling is involved in DN pathogenesis, we examined their transcript and protein levels in three kinds of DN models.^{15, 22, 23} Western blots showed significantly higher IL-33 (full length) and ST2 in the kidney of DN models compared with the counterparts (Figure 1A). In type 2 diabetic (T2D) mice induced by a combination of a high-fat diet (HFD) and streptozocin (STZ), both circulating and mRNA levels of IL-33 and ST2 were also elevated (Figure 1B,C). Histologically, IL-33 and ST2 positive spots were more expressed in the kidneys of DN patients than in paracancerous kidneys and mainly concentrated in the renal tubules (Figure 1D). To further confirm our findings, we analyzed 15 DN-related datasets from the Gene Expression Omnibus (GEO) database. IL-33 mRNA was upregulated in 14 datasets (Figure 1E). Notably, we observed the evolving increase of IL-33 with DN progression in OVE26 mice (GSE20636) and STZ-treated rats (GSE7253) (Figure 1F,G). Similarly, there was an evident increase of IL-33 in advanced DN patients (urinary albumin-to-creatinine ratio (ACR) > 300 mg/g, glomerular filtration rate (eGFR) < 90 mL/min/1.73 m²), but not early DN (ACR between 30 and 300 mg/g, eGFR > 90 mL/min/1.73 m²) (Figure 1H).

2.2 IL-33 or ST2 knockout exacerbates renal injury in DN mice

Given the remarkable and common increase of IL-33 and ST2 in DN, we constructed IL-33 and ST2 knockout mice respectively (IL-33^−/− or ST2^−/−) (Figure S1A–C). In control mice, knockout of IL-33 and ST2 had no significant impact on renal function (Figure S2A–H). Both IL-33^−/− and ST2^−/− DN mice had a higher kidney-body ratio (KW/BW), but there was no significant change in fasting blood glucose (FBG) compared with wild-type (WT) DN mice (Figure 2A,B). Contrary to expectations, knockout resulted in worse renal outcomes, including higher levels of serum creatinine (Scr), blood urea nitrogen (BUN), and ACR (Figure 2C–E). Pathological examinations revealed that DN mice with IL-33^−/− or ST2^−/− exhibited significantly disrupted renal morphology, characterized by epithelial cell desquamation, vacuolar degeneration, enlargement of glomerular Bowman's space, and expansion of the mesangium (Figure 2F–H). Furthermore, examination of the glomerular ultrastructure revealed thickening of the basement membrane and effacement of podocyte foot processes (Figure 2I,J). Besides glomerular damage, the tubular injury score and kidney injury molecular 1 (KIM-1) were further increased in IL-33^−/− and ST2^−/− DN mice (Figure 2K,L). The damage of ST2^−/− DN mice was similar to that of IL-33^−/− DN mice, indicating that IL-33 acted mainly through its receptor in DN.

2.3 RIL-33 aggravates renal dysfunction in DN mice

The present contradictory results encouraged us to revisit the role of IL-33. We administrated rIL-33 to DN mice, which caused a 76% increase in serumal IL-33 at the end of the experiment (Figure 3A). Administering rIL-33 to DN mice worsened renal function (increased Scr, BUN, and ACR) (Figure 3B–F). Histomorphologically, periodic acid-Schiff (PAS) staining showed mesangium expansion, and transmission electron microscope (TEM) images revealed the abnormal basement membrane and podocyte foot process in rIL-33-treated DN mice (Figure 3G,H). The increased renal tubular injury score and KIM-1 expression indicated the negative effects of rIL-33 on renal tubular in DN mice (Figure 3I,J).

To further determine the clinical relevance of the findings, we examined the expression of IL-33, ST2, and KIM-1, eGFR, and glomerular area from renal biopsies in DN patients (Table S2). IL-33 showed a negative correlation with eGFR (p = 0.0333), and a positive correlation with the KIM-1 levels (p = 0.0070) and glomerular area (p = 0.0001). ST2 also showed a negative correlation with eGFR (p = 0.0038) and a positive correlation with the glomerular area (p = 0.0001) (Figure 3K).

To search for genetic evidence in the population, we screened IL-33 expression-related single nucleotide polymorphisms (SNPs) by conducting the expression quantitative trait loci (eQTL) analyses in whole blood from the Genotype-Tissue Expression (GTEx) (Figure 3L). Among 28 SNPs with p < 0.01, we obtained relatively independent 7 SNPs based on linkage disequilibrium (LD) (r² < 0.8) after excluding 11 SNPs negatively correlated with IL-33 expression (Figure S3A,B). On account of the attractive role of polygenic risk score (PRS) in measuring the cumulative effect of multiple risk-associated variants, we calculated IL-33 expression-related PRS among 6693 DN and 463,084 non-DN participants from the UK biobank (Table S1). We found that PRS was positively associated with DN risk after adjusting age, sex, race, the first 10 primary components of ancestry, and genotype measurement batches (OR = 1.135, 95% confidence intervals (CIs): 1.041–1.236) (Figure 3M). Compared with participants with low genetic risk, those with a high genetic risk had a higher DN risk (OR = 1.096, 95% CIs: 1.029–1.167) (Figure 3M). In addition, similar results were obtained after excluding the SNPs associated with DN and then recalculating PRS (Figure S3C,D). These analyses indicated the high level of IL-33 as a risk factor for DN.

2.4 Partial reduction of IL-33 signaling attenuates renal damage in DN mice

Previous research reported that the action of IL-33 was sometimes contradictory within a similar pathology.²⁴ Thus, we speculated the promotion role of IL-33 dysregulation in this illness. We used neutralizing IL-33 antibodies (αIL-33) and heterozygous IL-33 and ST2 mice (IL-33^+/− and ST2^+/−) to exogenously or endogenously reduce IL-33 signaling in DN mice (Figure 4A). Compared with DN mice, αIL-33-treated and heterozygous DN mice exhibited decreased KW/BW, Scr, BUN, and ACR (Figure 4B–F). The improvement of the pathological indicators was also present in IL-33^+/− and ST2^+/− DN mice or mice treated with αIL-33, manifested as reduced glomerular volume and mesangial dilation, narrowed basement membrane, and wide foot process space (Figure 4G,H). The renal tubule-related injury was effectively relieved (Figure 4I,J). Overall, partial reduction of IL-33 signaling ameliorated kidney injury in DN, but not complete elimination.

2.5 Cellular senescence is involved in the role of IL-33 in DN

Senescence is increasingly considered an important cause of DN. ²⁵ In T2D mice, there is the greatest accumulation of senescent cells in the kidneys compared with the brain, heart, liver, and pancreas (Figure S4A). In DN patients, p16 expression showed a negative correlation with eGFR, and a positive correlation with the glomerular area (Figure S4B). In vitro, the injury of renal tubular epithelial cells was ameliorated by silencing cellular senescence regulators (p16 and p21) (Figure S4E–H), indicating the importance of cellular senescence in renal cell injury. Moreover, IL-33 and ST2 showed well colocalization with senescent renal tubular epithelial cells in DN mice and patients (Figure S4C,D). We therefore investigated the effects of IL-33 on renal senescence.

RIL-33 administration increased the senescence-associated β-galactosidase⁺ (SA-β-gal⁺) area, while αIL-33-treated DN mice had an occasional positive area (Figure 5A). The treatment of αIL-33 also upregulated Ki67, and decreased senescence-related proteins (p53, p21, and p16) and DNA damage response-associated proteins (ATM and γ-H2AX) (Figure 5A,B). The SASP, which represents the unique secreted factors from senescent cells, has been documented to disturb the microenvironment and promote disease advancement.²⁶ Thus, we colocalized the IL-6, a component of SASP, with senescent cells and found a significant decrease in merged fluorescent spots in αIL-33-treated DN mice (Figure 5C). Likewise, the colocalization of NF-κB, a regulator of SASP,²⁷ with SA-β-gal showed occasional positivity, in contrast to the evident colocalized positive spots in IgG-treated DN mice (Figure 5D). Inversely, rIL-33 significantly increased the production of SASP. Since resistance to apoptosis is responsible for the accumulation of senescent cells,²⁸ we examined the apoptosis-related pathways in the kidney. The protein levels of p-AKT/AKT and Bcl-2/Bax were increased and caspase-3 was decreased in αIL-33-treated DN mice (Figure 5E). Importantly, the prosurvival protein Bcl-2 was profoundly impaired in senescent cells (Figure 5F). However, the apoptotic pathway was not significantly changed in DN mice after adding rIL-33, although Bcl-2 was mildly increased in senescent cells. Moreover, αIL-33 treatment also alleviated markers of renal aging, including renal fibrosis and oxidative stress (Figure 5G,H). For clinical relevance, we found positive correlations of IL-33 and ST2 with p16 in DN patients (Figure 5I).

We further explore whether endogenous partial reduction of IL-33 had the same antisenescence effects. After partial knockdown of IL-33 and ST2, SA-β-gal was significantly reduced in DN mice, even though the Ki67 expression was not different (Figure S5A). Other senescence-related and aging-related markers were also normalized in IL-33^+/− and ST2^+/− DN mice (Figure S5A–I). Conversely, complete knockout of IL-33 and ST2 increased senescence markers in DN mice (Figure S6A–H). Strikingly, Bcl-2 was heavily accumulated in senescent cells in IL-33^−/− and ST2^−/− DN mice, suggesting IL-33 signaling deletion resulted in severe resistance to apoptosis in the senescent kidney (Figure S6E,F).

2.6 Cellular senescence mediates the role of partial reduction of IL-33 in cell injury in vitro

Our results so far indicated that senescence may be involved in the role of IL-33 in DN. We further explored the underlying mechanism in vitro. Because both senescence and the increased expression of IL-33 and ST2 were located in the renal tubule (Figure S4D,E), human tubular epithelial cells (HK-2) were exposed to high glucose (HG, 35 mmol/L) for 15 days to establish a model that mimicked in vivo. Treatments of αIL-33 (50 ng/mL) or rIL-33 (50 ng/mL) did not cause damage in cells exposed to normal glucose (NG, 5.5 mmol/L) (Figure S7A). The KIM-1 expression and LDH release were increased with the increase of rIL-33 concentration (Figure 6A,B). Similarly, cell damage was aggravated by 50 ng/mL of αIL-33 but effectively reversed by 10 ng/mL dosage (Figure 6A,B). Besides, the antisenescence role of 10 ng/mL αIL-33 was supported by the increased EdU⁺ cells, decreased SA-β-gal⁺ cells, γ-H2AX⁺ cells, and nuclear area (Figure 6C,D), and down-regulated p53, p21, and p16 (Figure 6E). However, both 50 ng/mL αIL-33 and 10 or 50 ng/mL rIL-33 accelerated the senescence of HK-2 cells challenged by HG. Consistently, by endogenous regulation of IL-33 expression, we found that both 20% and 80% reduction of IL-33 improved cell damage and senescence, and these effects were enhanced with the decrease of IL-33 (Figure S6B–E).

Further, we investigated whether IL-33 caused renal cell damage through cellular senescence. In the process of cellular senescence, p21 and p16 are considered to be vital regulators.²⁹ Therefore, we used RNA interference to silence both p16 and p21 in HK-2 cells simultaneously, which blocked HG-induced cellular senescence and injury (Figure S4G,H). More importantly, the increase in KIM-1 expression and LDH release in HK-2 cells induced by 50 ng/mL rIL-33 was repressed by the simultaneous silence (Figure 6F,G), indicating cellular senescence mediated the role of IL-33 in renal cell damage. Moreover, p16 silencing almost completely reversed cellular senescence and damage induced by rIL-33 (Figure 6H), suggesting IL-33 caused cellular senescence mainly through upregulating p16.

In addition, we explored the possible mechanisms by which the deletion of IL-33 signaling aggravated cellular senescence. In senescent HK-2 cells cultured for 17 days, we used ABT-263, a senolytic, to clear the senescent cells. Consistent with the above in vivo experiments, ABT-263 was more difficult to decrease the senescent cell damage and induce death after continuous treatment with 50 ng/mL αIL-33, indicating IL-33 signaling deletion resulted in resistance to apoptosis (Figure 6I–K).

2.7 SASP derived from senescent cells contributes to renal cell damage

We next investigated how cellular senescence caused renal injury. Conditioned medium (CM) was collected from IL-33-induced senescent cells to incubate young HK-2 cells. Figure 7A,B presented the increase of LDH release and KIM-1 expression in young HK-2 cells exposed to the CM-(HG+rIL-33) compared with CM-HG exposed cells, indicating SASP caused young cell injury. Recent research reported that senescence entailed significant changes in lipid metabolism, especially in oxylipin. In the current study, lipid profiles showed that some oxylipin in the kidney of DN mice increased significantly after administration with rIL-33 (Table S3), including prostaglandin D2, prostaglandin E2, prostaglandin E1, thromboxane B2, and 6-Keto-prostaglandin F1a (Figure 7C). Conversely, αIL-33 treatment decreased the levels of these oxylipins (Figure 7D). Further, some prostaglandins in the supernatant of senescent HK-2 cells were also elevated, indicating prostaglandins were SASP secreted by senescent renal tubular epithelial cells (Figure 7E). Also, these prostaglandins were regulated by prostaglandin synthases (cyclooxygenase-2, COX-2). The COX-2 inhibitor NS-398 relieved senescence and damage of HK-2 cells caused by HG and rIL-33 (Figure 7F–H). Compared with the exposure to CM-(HG+rIL-33), the damage of young HK-2 cells was reduced when exposed to CM-(HG+rIL-33+NS-398) and CM-(HG+rIL-33+Si-IL-33) (Figure 7I,J). Moreover, when IL-33 and COX-2 were inhibited simultaneously, the damage of young HK-2 cells was further reduced (Figure 7I,J).

4.1 Animal experimental design

Eight-week-old male wild-type C57BL/6 mice were procured from Charles River. IL-33^−/− and ST2^−/− C57BL/6 mice were kindly provided by the Department of Immunology at Tongji Medical College. Mice were divided into two groups and fed either a 60% HFD or a normal diet (ND) for 12 weeks. Mice on HFD were subsequently injected intraperitoneally with 50 mg/kg of STZ (Sigma–Aldrich; S0130) dissolved in citrate buffer (pH 4.5) daily for 5 consecutive days, while control mice received a vehicle solution.¹⁵ One week postinjection, mice with a glucose concentration of over 11.1 mmol/L were regarded as diabetic. After the diabetic model was produced, all mice were maintained for an additional 12 weeks of continued HFD or ND feeding. Meanwhile, WT diabetic mice were administered with αIL-33 (37 µg/kg) (R&D Systems; AF3626) or rIL-33 (12.5 µg/kg) (MedChemExpress; HY-P7218), and control mice were injected with control IgG or PBS, twice a week.

Additional diabetes models involving Zucker lean (ZL) rats, Zucker diabetic fatty (ZDF) rats, and ApoE^−/− mice followed established protocols.^{22, 23} We obtained male ZL rats and ZDF rats from Vital River Laboratories at the age of 6 weeks. Following an 8-week feeding period, the animals were sacrificed for further analysis. Additionally, 6-week-old male ApoE^−/− mice and their littermate control ApoE^+/+ mice were fed an HFD (21% fat and 1.25% cholesterol) or ND, respectively, for 16 weeks.⁴⁴ The Tongji Medical College Council on Animals Care Committee approved all animal procedures carried out in this study.

4.2 Human kidney tissues and analysis

Renal biopsies from 14 DN and 8 non-diabetic patients were obtained from the Tongji Hospital, Huazhong University of Science. Informed consent was obtained from each patient, and the patient's clinical biochemical data were extracted from medical records. The non-diabetic samples consisted of peritumoral, normal kidney tissues. These tissues were fixed in formalin, embedded in paraffin, and then sectioned at a thickness of 2 µm for immunohistochemistry (IHC) staining (the dilution ratio: 1:100) of IL-33 (Proteintech; 12372-1-AP), ST2 (Proteintech; 11920-1-AP), p16 (Proteintech; 10883-1-AP), and KIM-1 (ABclonal; A2831), or immunofluorescence colocation between IL-33 or ST2 with C₁₂FDG (Invitrogen; I2904, 33 µM). Quantitative analysis of protein expression was performed using Image-Pro Plus software, and the percentage positive staining of 5–10 images (×200 magnification) in each specimen was calculated.

Linear regression models were applied to examine the associations among IL-33, ST2, p16, eGFR, KIM-1, and glomerular area. The determination coefficient (R²) was used to assess the power to which the independent variable explains variations in the dependent variable.

4.3 Cell culture

HK-2 cells were purchased from Procell and maintained in DMEM/F12 (Gibco) supplemented with 10% fetal bovine serum, 100 U mL⁻¹ penicillin/streptomycin (Gibco) at 37°C in a humidified atmosphere containing 95% air and 5% CO₂. The cells were seeded in 6-well plates with NG (5.5 mM glucose) or HG (35 mM glucose) medium for 5 days and then treated with αIL-33 (10 or 50 ng/mL) or rIL-33 (10 or 50 ng/mL) for 10 days.

4.4 Plasma biochemical and histological analysis

FBG was measured from mouse tail veins using a blood glucose meter (Roche Diagnostics). Scr, BUN, and ACR were measured with commercially available assay kits (Jiancheng Bioengineering Institute). Renal samples were subjected to standard staining including hematoxylin–eosin (H&E), PAS, and Masson's trichrome stain. The glomerular area and the extent of the mesangial matrix were assessed by the PAS satin. Tubular damage including epithelial cell desquamation, vacuolar degeneration, swelling, and necrosis was scored: 0: tubular damage area < 10%; 0−1: 10−25% of the tubular area was affected; 1−2: 25−50% of the tubular area was affected; 2−3: 50−75% of the tubular area was affected. 3−4: 75−90% of the tubular area was affected. 4−5: >90% of the tubular area was affected.

For transmission electron microscopy, small fragments of the renal cortex were fixed in 2.5% glutaraldehyde and embedded in Araldite. Ultrathin sections (80-100 nm) were cut from the tissue using an ultramicrotome (Leica EM UC7). The thin slices were placed on a copper mesh grid, stained, and examined under a Tecnai G220 TWIN TEM.

4.5 SA-β-gal staining

Frozen kidney sections and cells were utilized to assess SA-β-gal activity using a commercially available kit (Beyotime Biotechnology), following the instructions provided by the manufacturer. The samples were incubated with a detection solution at 37°C overnight, and the presence of blue staining was indicative of SA-β-gal positive cells. Quantification was performed via light microscopy and Image J.

4.6 Immunofluorescence and IHC

The renal tissue or cell was fixed with 4% paraformaldehyde for 15 min. After washing three times with PBS, cells were permeabilized with 0.1% Triton X-100 for 15 min and then washed again with PBS 3 times. After being blocked with 5% normal goat serum, the samples were incubated overnight with primary antibodies at 4°C (the dilution ratio: 1:100–200), including IL-33, ST2, Bcl-2 (Proteintech; 66799-1-lg), NF-κB (Cell Signaling Technology; 8242s), IL-6 (Abcam; ab259341), and γH2AX (ABclonal; AP0099). The following day, the tissue was labeled with secondary antibodies (Thermo Scientific; 1:800) for 1 h and then washed three times with PBS. DAPI was applied for 10 min and subsequently photographed under a fluorescent microscope (Olympus).

Paraffin-embedded sections of kidney tissue from mice or patients with diabetes were used for IHC. After rehydration and blocking, tissue sections were exposed to the following primary antibodies overnight, followed by incubation with an HRP-conjugated secondary antibody. After signaling amplification using the TSA Biotin System (PerkinElmer), the slides were developed with a VECTASTAIN Elite ABC Kit and ImmPACT DAB Peroxidase Substrate (Vector Laboratories).

4.7 Immunoblots

The cultured cells and renal cortex samples were homogenized using an ice-cold lysis buffer and subsequently quantified using a BCA protein assay kit (Beyotime Biotechnology). The protein was then separated and transferred onto PVDF membranes (Millipore). After the membranes were blocked in 5% skim milk for 1 h at room temperature, they were incubated with primary antibodies overnight at 4°C (1:1000), including p53 (Cell Signaling Technology; 2524), p21 (Proteintech; 27296-1-AP), p16 (Proteintech; 10883-1-AP), ATM (ABclonal; A19650), γH2AX (ABclonal; AP0099), p-AKT (Cell Signaling Technology; 4060S), AKT (Cell Signaling Technology; 4691S), Bcl-2 (Proteintech; 66799-1-lg), Bax (Proteintech; A11931), and caspase-3 (Cell Signaling Technology; 9662s), followed by incubation with the corresponding secondary antibodies (Cell Signaling Technology; 1:10000) for 1 h at room temperature. The density of each target band was quantified by Image Pro-Plus 6.0 software and normalized to GAPDH (Proteintech; 60004-1-Ig) as optical density. All sample sizes of animals or cells were greater than or equal to 3.

4.8 Enzyme-linked immunosorbent assay

Enzyme-linked immunosorbent assay (ELISA) kit (Elabscience) was used to determine the level of IL-33 and ST2 protein in the sera. The experimental procedures were performed according to the manufacturer's instructions.

4.9 Gene silencing

Silencing of IL-33, p16, and p21 in HK-2 cells was achieved by using a reverse siRNA transfection procedure performed in six-well plates. Once grown to 70% confluence, cells were transfected with siRNA or scrambled siRNA (RiboBio) using Lipofectamine®RNAiMAX (Invitrogen) according to the manufacturers.

4.10 Generation of CM

CM was generated by culturing HK-2 cells for 14 days and then in serum-free and drug-free media for 24 h before harvest, followed by clarification through centrifugation. The resulting CM was mixed with fresh medium in a proportion of 1 to 1 to generate CM containing 10% FBS. Young HK-2 cells were then treated with different groups of CM for 48 h.

4.11 Liquid chromatography–tandem mass spectrometry

The extraction of oxylipins from kidney tissue was performed as described previously with some modifications.⁴⁵ Briefly, frozen tissue was mixed with an antioxidant solution and 5 µL of deuterated internal standards before tissue homogenization and protein precipitation. The ultrasound-assisted extraction was performed under 4°C for 30 min. Following centrifugation, the supernatant was loaded onto Phenomenex Strata-X polymeric sorbent columns. After loading the samples, the columns were washed with 90:10 water/methanol to remove impurities, and the oxylipins were then eluted with 1 mL of 100% methanol. The eluant was dried under a nitrogen stream and reconstituted in 100 µL 50:50 water/methanol before liquid chromatography–tandem mass spectrometry (LC–MS) analysis.

LC–MS analyses of oxylipins were performed on a Shimadzu LC-20AD series HPLC system coupled to an AbSciex 4500 QTRAP mass spectrometer fitted with a TurboV ion source. Oxylipins were separated on a reversed-phase Zorbax Eclipse Plus C18 column. The solvent system consisted of ACN/water/acetic acid (60:40:0.02, v/v/v) (mobile phase A) and ACN/IPA (50:50, v/v) (mobile phase B) with a flow rate of 0.4 mL/min. The following gradient program was used: 0−4.0 min, 0.1–55% B; 4.0–15.0 min, 55−99% B; 15.0-20.0 min, 99% B; 20.1–25.0 min, 55–0.1%B. Samples were kept at 4°C and the injection volume was 10 µL. Mass spectrometric analyses were conducted using the following parameters: curtain gas, nebulizer gas, and turbo-gas (GS2) were set at 30, 40, and 40 psi, respectively. Oxylipins were analyzed using scheduled multiple reaction monitoring in the scanning range of m/z 50−640 and the scan rate was 1000 µ/s. Data acquisition and analysis were performed using Analyst software, version 1.6.3. Oxylipins were quantified by the stable isotope dilution method. For each class of oxylipins to be quantified, one internal standard was chosen and calibration curves were established.

4.12 Acquisition of microarray data

The datasets of transcriptome profiling by microarray were searched on GEO profiles (https://www.ncbi.nlm.nih.gov/geoprofiles/) with the keywords “diabetic kidney disease” or “diabetic nephropathy.” All microarray datasets with IL-33 were downloaded for analyses, including GSE96804, GSE33744, GSE118089, GSE51205, GSE142025, GSE20636, GSE30528, GSE107942, GSE30529, GSE86300, GSE87359, GSE90842, GSE134327, and GSE131266.

4.13 Study population

Our study was based on the data from the UK Biobank, a large-scale prospective cohort study.⁴⁶ The cohort`’s procedures and baseline characteristics are available on the UK Biobank website (www.ukbiobank.ac.uk). The definition was described according to the 9th and 10th revisions of the International Classification of Diseases (ICD-9 and ICD-10) and self-reported data fielded with the choice-, disease-, or procedure-specific codes. Excluding individuals without relevant SNPs data, we eventually included 6693 DN and 463,084 non-DN participants.

4.14 PRS calculation

4.15 Data analysis

The population data were analyzed using SAS (version 9.4) or R (version 4.2.2). The experiment data were analyzed using Graph Pad Prism 8.3 and showed as the mean ± SEM. Differences among the groups were determined by the Student's t-test or one-way analysis of variance. Differences were identified as statistically significant at 3 levels: p < 0.05, p < 0.01, and p < 0.001.