USF2 knockdown downregulates THBS1 to inhibit the TGF-β signaling pathway and reduce pyroptosis in sepsis-induced acute kidney injury.

OBJECTIVE
Acute kidney injury (AKI) is a serious complication of sepsis. This study was performed to explore the mechanism that THBS1 mediated pyroptosis by regulating the TGF-β signaling pathway in sepsis-induced AKI.


METHODS
Gene expression microarray related to sepsis-induced AKI was obtained from the GEO database, and the mechanism in sepsis-induced AKI was predicted by bioinformatics analysis. qRT-PCR and ELISA were performed to detect expressions of THBS1, USF2, TNF-α, IL-1β, and IL-18 in sepsis-induced AKI patients and healthy volunteers. The mouse model of sepsis-induced AKI was established, with serum creatinine, urea nitrogen, 24-h urine output measured, and renal tissue lesions observed by HE staining. The cell model of sepsis-induced AKI was cultured in vitro, with expressions of TNF-α, IL-1β, and IL-18, pyroptosis, Caspase-1 and GSDMD-N, and activation of TGF-β/Smad3 pathway detected. The upstream transcription factor USF2 was knocked down in cells to explore its effect on sepsis-induced AKI.


RESULTS
THBS1 and USF2 were highly expressed in patients with sepsis-induced AKI. Silencing THBS1 protected mice against sepsis-induced AKI, and significantly decreased the expressions of NLRP3, Caspase-1, GSDMD-N, IL-1β, and IL-18, increased cell viability, and decreased LDH activity, thus partially reversing the changes in cell morphology. Mechanistically, USF2 promoted oxidative stress responses by transcriptionally activating THBS1 to activate the TGF-β/Smad3/NLRP3/Caspase-1 signaling pathway and stimulate pyroptosis, and finally exacerbated sepsis-induced AKI.


CONCLUSION
USF2 knockdown downregulates THBS1 to inhibit the TGF-β/Smad3 signaling pathway and reduce pyroptosis and further ameliorate sepsis-induced AKI.


Background
As a common complication of critically ill patient, sepsis-induced acute kidney injury (sepsis-induced AKI) is related to high risk of chronic complication, drastic physiologic changes and high mortality [1,2].About 35% of sepsis patients develop renal failure, and nearly half of critically ill patients with AKI also developed sepsis [1,3].In the past, sepsis-induced AKI was considered to be caused by renal hypoperfusion, but studies have found in ammatory response and apoptosis may also be involved in the pathogenesis of sepsis-induced AKI [1,4].Oxidative stress has been reported to be related to sepsis and may contribute to AKI by generating reactive oxygen species (ROS) [5].Existing evidence also reports that pyroptosis of renal tubular epithelial cells induced by Caspase-11 is a crucial event in the development of sepsis-induced AKI [3].Current treatment of sepsis-induced AKI mainly focuses on renal replacement therapy, uid resuscitation, vasoactive drugs and antibiotics [5].Since the pathogenesis of sepsis-induced AKI involves multiple factors such as apoptosis, immune and in ammatory processes, molecular mechanism targeting signaling pathways may have potential therapeutic value for sepsis-induced AKI.Thrombospondin1 (THBS1), also called TSP1, is a homotrimeric glycoprotein released by activated platelets and involved in brin clots reacting to injury [6].Previous study demonstrates that THBS1 bound to CD47 to promote AKI and blocking of THBS1 signaling through CD47 improved renal interstitial brosis [7].Pyroptosis, a new programmed cell death mode discovered in recent years, is characterized by dependence on cysteine-aspartic proteases 1 (caspase-1), NLRP3 in ammasome activation and the release of a large number of proin ammatory factors [8].In human cells, TSP1 is an upstream inhibitor of IL-1β and caspase-1 mRNA, but not NLRP3 induced in the presence of LPS [9].The involvement of THBS1 binding to the CD36 receptor in apoptosis has been reported in endothelial cells [10].However, no study concerns the correlation between THBS1 and pyroptosis at present.
Previous study reports that activation of TGF-β can be attenuated by suppressing THBS1 with the LSKL peptide [6].TGF-β can induce the activation of NLRP3 in ammasome and the cleavage of Gasdermin D (GSDMD) in rat with chronic kidney disease [11].Those evidences support the association among THBS1, in ammatory response, pyroptosis and AKI.We further searched the upstream transcriptional regulators of THBS1 in the TTRUST database, including E2F1, NR1I2, SNAI1, TP53, USF2 and WT1.Overexpression of USF2 is found in previous study to result in renal injury [12].Hence, in this study, we aim to explore the mechanism that transcriptional regulator USF2 regulates THBS1 and pyroptosis via the TGF-β signaling pathway in sepsis-induced AKI.

Ethics statement
This study was performed with the approval of the Clinical Ethical Committee of Xinhua Hospital A liated to Shanghai Jiao Tong University School of Medicine.Written informed consents were obtained from all patients and healthy volunteers.The animal experiments were conducted based on minimized animal number and the least pains.

Clinical samples
Fifteen patients with sepsis-induced AKI were enrolled in this study.Inclusion criteria were: aged 18 ~ 80 years, male or female, and the diagnosis of all patients was in accordance with the International Sepsis Guidelines Diagnostic Criteria [13] and Acute Kidney Injury Diagnostic Criteria [14].Exclusion criteria were: chronic kidney disease or kidney transplantation or AKI caused by non-sepsis, or the presence of autoimmune diseases.Fifteen healthy volunteers (con rmed absence of renal impairment, autoimmune diseases, or malignant tumors) were also enrolled in this study.Peripheral blood was collected from all subjects, and the samples were centrifuged within 2 h and stored at -80 °C.

Establishment of animal model
A total of 120 C57BL/6J mice (aged 6 ~ 8 weeks) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd., Beijing, China with animal license number (SYXK (Jing) 2016-0011).Mice were housed in separate cages at 20-26 °C, with free access to food and water and a 12-hlight/dark cycle.After adaptive feeding for 6 days, mice were assigned into 5 groups randomly, with 24 mice in each group: sham group, AKI group, AKI + sh-NC group, AKI + sh-THBS1 group and AKI + sh-USF2 group.Mice in all groups (except the sham group) were injected intraperitoneally with lipopolysaccharide (LPS) 10 mg/kg.The mice in AKI + sh-THBS1 group and AKI + sh-NC group were injected with adenovirus low expression vector sh-THBS1 and its control adenovirus in the tail vein, respectively, and the mice in AKI + sh-USF2 group and AKI + sh-NC group were injected with adenovirus low expression vector sh-USF2 and its control adenovirus in the tail vein, respectively, with a dose of 0.1 mL.Adenovirus and control vectors were synthesized and packaged by Hanheng Biological Company (Shanghai, China), and the adenovirus virus was 10 10 /mL.After injection of lipopolysaccharide or phosphate buffered saline (PBS) for 24 h, 6 mice in each group were placed in sealed metabolic cages and fasted.The 24-h natural urine output of mice was collected and accurately weighed by an electronic balance.The remaining mice in each group were sacri ced after injection of LPS or PBS for 48 h, and serum samples were separated.Contents of serum creatinine and urea nitrogen were measured using an Automated Chemistry Analyzer (Shenzhen Rayto Life Science Co., Ltd, Shenzhen, Guangdong, China).

Hematoxylin-eosin (HE) staining
Mouse kidney tissues were collected and xed with 4% formaldehyde for 6 h and embedded in para n.
Para n-embedded tissue sections were depara nized in xylene and successively soaked in 100%, 95%, 80%, and 75% ethanol, and then washed with distilled water.Then sections were stained with hematoxylin staining solution for 10 min, rinsed with distilled water, differentiated tissues in differentiation solution for 30 s, soaked in distilled water for 15 min, stained with eosin staining solution for 2 min and then rinsed with distilled water, successively soaked in 95%, 95%, 100%, and 100% ethanol for 1 min.Next, sections were placed in xylene carbonic acid (3:1) for 1 min, soaked in xylene (I) for 1 min and soaked in xylene (II) for 1 min.Finally, sections were mounted with neutral resin and observed under a microscope (Olympus, Japan) with images randomly collected.

Establishment of cell model of sepsis-induced AKI
Human umbilical vein endothelial cell (HUVEC) line was purchased from Shanghai Cell Bank of Chinese Academy of Sciences (Shanghai, China).Cells were incubated in RPMI-1640 medium (Gibco, Grand Island, NY, USA) containing 10% fetal bovine serum, penicillin (1 × 10 5 U/L), and streptomycin (100 mg/L) in a cell incubator at 37 °C with 5% CO 2 .When the cell density reached 90%, HUVECs were detached with trypsin-ethylenediaminetetraacetic acid (EDTA) solution, then seeded in six-well plates, cultured again to reach a cell density of 90%, and continued to be cultured in serum-free medium for 24 h.Cells were randomly assigned into control group, LPS group, LPS + si-NC group, LPS + si-THBS1 group, LPS + si-USF2 + pc-NC, and LPS + si-USF2 + pc-THBS1 group.Cells in all groups (except control group) were added with LPS (Sigma-Aldrich, Merck KGaA, Darmstadt, Germany) at a concentration of 1 µg/mL and treated for 24 h.Control cells were treated with the same amount of PBS for 24 h.

Enzyme-linked immunosorbent assay (ELISA)
HUVECs and the supernatant after being centrifuged were collected to detect the expressions of TNF-α, IL-18, and IL-1β by ELISA.ELISA kits were purchased from Shanghai Bogoo Biotechnology Co., Ltd.(Shanghai, China).Kidney tissues were taken from mice and weighed, with cold PBS added (1:9 v/w) and tissues thoroughly ground.The homogenate was centrifuged at 300 g for 10 min at 4 °C and its supernatant was taken.The contents of TNF-α, IL-18, and IL-1β were detected using Elisa kits (Shanghai, China).

Transmission electron microscope
HUVECs were detached using trypsin containing disodium EDTA, centrifuged at 100 g for 5 min, with supernatant discarded, and the cells were xed by adding xative for electron microscopic.The xed cells were dehydrated through graded ethanol, stained with 70% ethanol and uranyl acetate, embedded, cut in section and observed under a transmission electron microscope (BX53, Olympus, Tokyo, Japan).

Co-immunoprecipitation (Co-IP)
The procedure was performed in strict accordance with the Co-IP Pierce® Direct IP Kit 26148.AMino Link Plus Resin (20 µL) and Pierce Control Agarose Resin (20 µL) (the negative control, without reacting with the antibody to remove non-speci c binding protein) were incubated with 4 µg of THBS1 monoclonal antibody for 90 min, followed by the addition of 500 µL of protein samples overnight at 4 °C and washed four times with IP lysis/wash buffer and once with solution buffer.USF2 protein complex was eluted with solution buffer.The expressions of THBS1 and USF2 in the above samples were detected by Western blot after electrophoresis with 10% SDS-PAGE.The above antibodies were purchased from Abcam (Cambridge, MA, USA).

Cell viability test
The viability of HUVECs in each group was detected using the CellTiter-Glo® Luminescent Cell Viability Assay kit (G7570, Promega, Beijing, China).HUVEC cells were adjusted to a density of 1 × 10 5 cells/mL, seeded into 6-well plates at 1 mL per well, and cells were collected for experiments when the cell con uence reached 80%~90%.Reagents were prepared per the manufacturer's instructions, and then each group of cells (50 µL) was seeded into a 96-well plate.After mixing and being incubated for 10 min, detection was performed on a plate-reading luminometer (Varioskan ™; Thermo Scienti c, Rockford, IL, USA).

Lactate dehydrogenase (LDH) activity test
HUVECs were seeded into 96-well plates (5 × 10 3 cells/well).After appropriate treatment, LDH activity released into the culture medium was measured using the Cytotoxicity Assay Kit (Roche Applied Science, Mannheim, Germany).The release of LDH from cells was obtained by measuring the absorbance at 490 nm with a microplate reader.

Oxidative stress-related test
HUVECs in logarithmic growth phase were taken with the density adjusted to 2 × 10 5 cells/mL, and seeded to 6-well plates with 2 mL per well; after the supernatant was discarded, cells were washed with PBS twice, added with serum-free RPMI-1640 culture medium containing 10 µmol/L DCFH-DA, and incubated in an incubator containing 5% CO 2 at 37 °C for 30 min.The cells were shaken well to make the probe in full contact with the cells.Next, culture medium was discarded and cells were washed with PBS 3 times.Intracellular ROS content was measured by microplate reader (excitation wavelength 493 nm, emission wavelength 525 nm).
HUVECs were taken with the density adjust to 2 × 10 5 cells/mL and seeded into 6-well plates at 2 mL per well.At the end of the culture, cells were detached with trypsin, washed twice with PBS, and centrifuged.The supernatant was discharged before the cells were collected.Total glutathione (GSH) content was calculated by measuring the absorbance at 410 nm with a microplate reader (Nanjing Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China).The absorbance at 560 nm was measured to calculate superoxide dismutase (SOD) activity and the absorbance at 532 nm was measured to calculate malondialdehyde (MDA) content.

Quantitative reverse transcription polymerase chain reaction (qRT-PCR)
Total RNA was extracted from HUVECs by TRIzol one-step method, and cDNA was synthesized per the instructions of the reverse transcription kit and placed in a uorescence quantitative PCR instrument for ampli cation with an ampli cation system of 20 µL.qRT-PCR was performed on an ABI7900HT fast PCR real-time system (Applied Biosystems, Foster city, CA, USA) using cDNA as a template and GAPDH as an internal reference per the instructions of SYBR Premix ExTaqII (Takara, Dalian, China), and relative expression was calculated by the 2-ΔΔCt method.All primers (Table 1) were designed and synthesized by Sangon Biotech (Shanghai) Co., Ltd.

Statistical analysis
Data analysis was performed with SPSS 21.0 (IBM Corp. Armonk, NY, USA) software.The data were normally distributed as tested by Kolmogorov-SmiRnov, and the data were expressed as mean ± standard deviation (SD).Comparisons between two groups were analyzed by independent sample t-test.
Comparisons among multiple groups were analyzed by one-way analysis of variance (ANOVA) or twoway ANOVA.Post-hoc tests were performed by Tukey multiple comparison test.All tests of signi cance were 2-tailed, and P < 0.05 (P < 0.01) was considered statistically signi cant.

Results
THBS1 was highly expressed in patients with sepsisinduced AKI Totals of 213 differentially expressed genes were obtained by differential expression analysis on microarray GSE60088 (related to sepsis-induced AKI), containing 154 signi cantly upregulated genes and 59 downregulated genes (Fig. 1A).Totals of 241 genes related to sepsis-induced AKI were obtained from the GeneCards database.The intersection between differentially expressed genes and the genes related to sepsis-induced AKI were performed and 23 candidate genes obtained (Fig. 1B).The expression heatmap of candidate genes was plotted (Fig. 1C).We focused on the 4 genes BMP6, MYC, TGIF1 and THBS1.Considering the difference value and innovation and the searched literature, we found few reports on the study about the relationship between THBS1 and sepsis-induced AKI.THBS1 was highly expressed in diseased samples in the microarray GSE60088 (Fig. 1D).THBS1 expression was detected by qRT-PCR in patients with sepsis-induced AKI, and the results showed that the mRNA expression of THBS1 was signi cantly up-regulated (P < 0.01) (Fig. 1E).Previous study reported that excessive release of in ammatory mediators was a major trigger for sepsis-induced AKI [15].ELISA was performed to detect the expressions of TNF-α, IL-1β, and IL-18 in the serum of patients with sepsis-induced AKI and normal subjects, and the results showed that the expressions of TNF-α, IL-1β, and IL-18 in patients with sepsisinduced AKI were signi cantly higher than those in healthy subjects (P < 0.01) (Fig. 1F).Correlation analysis between THBS1 and in ammatory factors showed that the expression of THBS1 was positively correlated with the expression of in ammatory factors (all P < 0.05) (Fig. 1G).
Protective effect of silencing THBS1 on sepsis-induced AKI mice Mice in each group were intraperitoneally injected with 10 mg/kg LPS to construct a mouse model of sepsis-induced AKI, and qRT-PCR showed that the expression of THBS1 in mice was signi cantly increased (P < 0.01) (Fig. 2A).sh-NC and sh-THBS1 were transfected into the mouse model and the results of qRT-PCR indicated successful transfection (P < 0.01) (Fig. 2A).Compared with the shamoperated mice, serum creatinine and urea nitrogen were signi cantly elevated and 24-h urine output was reduced in mice in the AKI group (P < 0.01) (Fig. 2B-D), while the opposite results were presented in the AKI + sh-THBS1 group.HE staining showed that the renal sections of sham-operated mice had normal morphology, clear histological structure, no degeneration, atrophy, swelling and necrosis or in ammatory in ltration of renal tubular cells, and no dilatation of lumen of renal tubulars (Fig. 2E).Renal tubular epithelial cell edema, vacuolar degeneration, disappeared brush border and dilatation of lumen of renal tubulars were observed in the renal tissue in mice of AKI group and AKI + sh-NC group, while silencing THBS1 could signi cantly improve the renal histopathological damage in AKI mice (Fig. 2E).ELISA was s showed that the expressions of TNF-α, IL-1β, and IL-18 in serum in the AKI mice were signi cantly increased, while silencing THBS1 could signi cantly reduce the expression of in ammatory factors in AKI mice (all P < 0.01) (Fig. 2F).

Silencing THBS1 protected mice against sepsis-induced AKI by inhibiting pyroptosis
Pyroptosis might result in cell death, renal in ammation and renal injury, while inhibition of pyroptosis ameliorates pathological injury [3,16].Hence, we speculated that the protective effect of si-THBS1 on mice with sepsis-induced AKI might be achieved by regulating pyroptosis.FAM-FLICA Caspase-1 Detection Kit showed that the number of pyroptotic cells was signi cantly increased in the AKI group, but decreased in the AKI + sh-THBS1 group (all P < 0.05) (Fig. 3A).Western blot showed that the levels of caspase-1 and GSDMD-N were signi cantly increased in the AKI group, while the levels were decreased in the AKI + sh-THBS1 group (all P < 0.01) (Fig. 3B).The above results con rmed that the protective effect of si-THBS1 on sepsis-induced AKI might be achieved by inhibiting pyroptosis.

THBS1 knockdown inhibited the activation of NLRP3 in ammasome and reduced in ammation
The above in vivo experiments demonstrated that silencing THBS1 signi cantly reduced the expressions of in ammatory factors in a mouse model of sepsis-induced AKI.In vitro cell experiment was performed to further con rm those ndings.LPS at a concentration of 1 µg/mL was used to induce HUVECs for 24 h for establishing a sepsis-induced AKI cell model.qRT-PCR results showed that THBS1 in the LPS group was signi cantly elevated (P < 0.01) (Fig. 4A).si-NC and si-THBS1 were transfected to HUVECs to reduce THBS1 expression (P < 0.01) (Fig. 4A).qRT-PCR and ELISA showed that TNF-α, IL-1β, and IL-18 in the LPS group were signi cantly increased, and THBS1 knockdown signi cantly reduced the expression of in ammatory factors in cells (all P < 0.01) (Fig. 4B-C).Western blot demonstrated that the expression of NLRP3 was signi cantly increased in the LPS group, while clearly decreased after down-regulating THBS1 expression (all P < 0.01) (Fig. 4D).Brie y, THBS1 knockdown inhibited the activation of NLRP3 in ammasome and decreased in ammatory factors.

THBS1 knockdown inhibited pyroptosis and attenuated cell injury
FAM-FLICA Caspase-1 Detection Kit showed that the number of pyroptotic cells was signi cantly reduced in the LPS + si-THBS1 group compared with the LPS group (Fig. 5A).Western blot revealed that THBS1 knockdown signi cantly decreased the levels of caspase-1 and GSDMD-N (all P < 0.01) (Fig. 5B).Cell viability and LDH activity in each group were detected by different kits.The results showed that cell viability was signi cantly increased and LDH activity was signi cantly decreased after down-regulating THBS1 expression (all P < 0.01) (Fig. 5C-D).Transmission electron microscopy showed that cells in the LPS group were characterized with cell membrane rupture, release of cytoplasmic contents, and chromatin condensation; while THBS1 knockdown partially reversed the morphological changes of the cells in the LPS group (Fig. 5E).
THBS1 mediated pyroptosis and aggravated cell injury by activating the TGF-β/NLRP3/Caspase-1 pathway Enrichment analysis of candidate genes was performed by WebGestalt tool (Fig. 6A) and the results indicated that THBS1 gene was mainly enriched in the AGE-RAGE and TGF-βpathways.Previous study reported that THBS1 could activate the TGF-β pathway [6].TGF-β could induce the activation of NLRP3 in ammasome and the cleavage of GSDMD in rat with chronic kidney disease [11].Thus we speculated that THBS1 could regulate pyroptosis through the TGF-β/NLRP3/Caspase-1 pathway.Western blot revealed that THBS1 was able to regulate TGF-βexpression (Fig. 6B).LPS-induced cells were treated with the TGF-β activator SRI-011381 hydrochloride and THBS1 knockdown.Western blot showed that the levels of TGF-β, NLRP3, caspase-1, and GSDMD-N were signi cantly increased in the LPS + si-THBS1 + SRI group compared with the LPS + si-THBS1 + DMSO group (all P < 0.01) (Fig. 6C), suggesting that THBS1 was able to activate the NLRP3/caspase-1 pathway through TGF-β.Results of ELISA showed that the expressions of TNF-α, IL-1β, and IL-18 were signi cantly increased in the LPS + si-THBS1 + SRI group (all P < 0.01) (Fig. 6D).Cell viability was notably decreased in the LPS + si-THBS1 + SRI group compared with the LPS + si-THBS1 + DMSO group, while LDH activity was signi cantly increased (all P < 0.01) (Fig. 6EF).All those ndings suggested that THBS1 accelerated pyroptosis and cell injury by activating the TGF-β/NLRP3/caspase-1 pathway.
Transcription factor USF2 exacerbated cellular oxidative stress response by up-regulating THBS1 to activate the TGF-β/NLRP3/Caspase-1 axis The upstream transcriptional regulators of THBS1 were searched through the TTRUST database (Table 2), and overexpression of USF2 was found in previous study to result in renal injury [12].Oxidative stress is involved in the pathogenesis of sepsis-induced AKI [17].We speculated that the transcription factor USF2 might exacerbate the cellular oxidative stress response by activating the TGF-βpathway through up-regulation of THBS1.The presence of USF2 and THBS1 binding was revealed by Co-IP (Fig. 7A).USF2 was elevated in LPS-induced cells, and si-USF2 successfully decreased USF2 expression (Fig. 7B) and signi cantly reduced THBS1 in LPS-induced cells (Fig. 7B) (all P < 0.01).PcDNA-THBS1 and its control were transfected to LPS-induced cells that were previously transfected with si-USF2 (Fig. 7C).
The ROS content in the cells was detected using uorescent probe DCFH-DA, and the results showed that the uorescence was signi cantly increased in the LPS + si-USF2 + pc-THBS1 group compared with the LPS + si-USF2 + pc-NC group (P < 0.01) (Fig. 7D).Compared with the LPS + si-USF2 + pc-NC group, the contents of GSH and SOD in the LPS + si-USF2 + pc-THBS1 group were reduced, while the content of MDA was signi cantly increased (all P < 0.01) (Fig. 7E-G).Western blot showed that the protein levels of THBS1, TGF-β, NLRP3, Caspase-1, and GSDMD-N were signi cantly elevated in the LPS + si-USF2 + pc-THBS1 group (P < 0.01) (Fig. 7H).The results of ELISA also con rmed that the expressions of TNF-α, IL-1β, and IL-18 were signi cantly increased in the LPS + si-USF2 + pc-THBS1 group (all P < 0.01) (Fig. 7I).In summary, transcription factor USF2 upregulated THBS1 to activate the TGF-β/NLRP3/caspase-1 pathway and nally exacerbate the cellular oxidative stress response.USF2 downregulation inhibited THBS1 to block the TGFβ/NLRP3/caspase-1 pathway ameliorate sepsisinduced AKI In vivo experiments in mice were performed to con rm the nding that USF2 downregulation inhibited the TGF-βpathway by down-regulating THBS1 to ameliorate sepsis-induced AKI.qRT-PCR revealed that USF2 was signi cantly increased in AKI mice (P < 0.01) (Fig. 8A).sh-NC and sh-USF2 were transfected into the AKI mouse model (P < 0.01).The effect of USF2 downregulation on the levels of THBS1, TGF-β, NLRP3 and Caspase-1 in mice was detected.Western blot showed that the levels of THBS1, TGF-β, NLRP3 and Caspase-1 in the AKI + sh-USF2 group were signi cantly decreased (all P < 0.01) (Fig. 8B).Compared with the sham group, mice in the AKI group had signi cantly enhanced serum creatinine, urea nitrogen, and lowered 24-h urine output (all P < 0.01) (Fig. 8C-E), while the AKI + sh-USF2 group showed the opposite trend.ELISA showed that the expressions of TNF-α, IL-1β, and IL-18 were upregulated in the AKI group, while USF2 downregulation could signi cantly reduce the expression of in ammatory factors in the AKI mice (all P < 0.01) (Fig. 8F).GSDMD-N was signi cantly increased in the AKI group, but decreased in the AKI + sh-THBS1 (all P < 0.01) (Fig. 8B).HE staining also demonstrated that USF2 downregulation had protective effects on mice with sepsis-induced AKI (Fig. 8G).The above results suggested that the protective effect of USF2 downregulation on mice with sepsis-induced AKI.

Discussion
AKI is a serious complication of sepsis with high morbidity and mortality [3].Current studies and reviews mainly concern the pathogenesis, prevention and supportive treatment of sepsis-induced AKI [1,2].
THBS1 is signi cantly involved in the progression of renal diseases [7].By bioinformatics analysis, THBS1 is found as one of the candidate genes related to sepsis-induced AKI.In our study, USF2 knockdown downregulated THBS1 and inhibited TGF-β pathway and reduced pyroptosis, thus ameliorating sepsis-induced AKI.
After bioinformatics analysis and detection in clinical samples, we found that THBS1 was highly expressed in patients and mouse model with sepsis-induced AKI.Additionally, expression of THBS1 was positively related with the expressions of in ammatory factors.Previous study reports that THBS1 is in human renal brotic conditions and affects renal brosis through regulating in ammation [18].After transfecting sh-THBS1 into the AKI mice, we found that silencing THBS1 using sh-THBS1 signi cantly reduced the expression of in ammatory factors.A review reports that THBS1 is involved in the in ammatory response after hemorrhagic stroke [19].Interestingly, Maimaitiyiming H et al. previously reported that THBS1 was a signi cant contributor to the progression of renal disease and THBS1 de ciency reduced renal macrophage in ltration and in ammation [20].The ndings in the previous studies are consistent with our results.
Pyroptosis might result in cell death, renal in ammation and renal injury, while inhibition of pyroptosis ameliorates pathological injury [3,16].Central to pyroptosis is the activation of the NLRP3 in ammasome and mediation of GSDMD, rapidly causing cell membrane rupture and release of cellular contents and nally resulting in an in ammatory response [16].In our study, the pyroptotic cells and Caspase-1 and GSDMD-N levels were clearly increased in AKI mice, but decreased after sh-THBS1 treatment.Cell experiment also showed that THBS1 knockdown inhibited pyroptosis, the activation of NLRP3 in ammasome and in ammatory factors, and attenuated cell injury.THBS1 is involved in the pathophysiology of kidney ischemia and considered as a new mediator of cell injury [21].However, no study has been performed on the correlation between THBS1 and pyroptosis.All those results initially con rmed that the association between of THBS1 and pyroptosis and concluded that the protective effect of THBS1 knockdown on sepsis-induced AKI might be achieved by inhibiting pyroptosis.
Enrichment analysis of candidate genes indicated that THBS1 was mainly enriched in the AGE-RAGE and TGF-β pathways.Suppression of THBS1 can attenuate the activation of TGF-β [6].Additionally, TGF-β can induce the cleavage of GSDMD and the activation of NLRP3 in ammasome in rat with chronic kidney disease [11].Our study showed that the levels of TGF-β, NLRP3, Caspase-1, and GSDMD-N were signi cantly increased in the LPS + si-THBS1 + SRI group, suggesting that THBS1 was able to activate the NLRP3/caspase-1 pathway through TGF-β.It's noted that the constitutive activation of NLRP3 in ammasome leads to serious liver in ammation and pyroptotic cell death in hepatocytes [22].Dai et al. found that interaction of C/EBPβ and TFAM promoted pyroptosis through the activation of NLRP3/caspase-1, and further boosted the progression of AKI [23].Taken together, all those evidences support our ndings that THBS1 knockdown ameliorated sepsis-induced AKI by inhibiting pyroptosis and alleviating cell injury by inactivating the TGF-β/NLRP3/Caspase-1 pathway.
USF2 is an upstream transcriptional regulator of THBS1 searched in TTRUST database and the presence of USF2 and THBS1 binding was con rmed by Co-IP.Elevated expression of USF2 suppresses the transcriptional activity of the Smurf and results in the enhanced activity of TGF-β [24].Transgenic USF-2 mice overexpress TGF-β1 to modulate the progression of kidney disease and up-regulated TGF-β1 level is related to the activation of pro brotic pathways triggered by oxidative stress [25].Oxidative stress has been proven to be related to AKI development [17].High expression of THBS1-CD47 signaling is responsible for the enhanced oxidative stress in diabetes [26].Consistently, our nding revealed that USF2 upregulated THBS1 to activate the TGF-β/NLRP3/caspase-1 pathway and nally exacerbate the cellular oxidative stress response.Serum creatinine concentration and urine output are two early indicators for the diagnosis of AKI [5].Our study found that silencing USF2 decreased serum creatinine, urea nitrogen and increased urine output.IUSF2 downregulation could signi cantly reduce in ammatory factors in AKI mouse model by inhibiting THBS1.Overexpression of USF2 is found in a previous report to result in renal injury [12].Overexpression of USF2 contributes to the elevation of THBS1 and activation of TGF-β, leading to the development of diabetic nephropathy [27].Taken together, USF2 downregulation inhibited THBS1 to inactivate the TGF-β/NLRP3/caspase-1 pathway and ameliorated sepsis-induced AKI.

Conclusion
In conclusion, USF2 activates the TGF-β pathway by up-regulating THBS1 to promote pyroptosis and aggravate sepsis-induced AKI, which may provide therapeutic strategy for patients with sepsis-induced AKI.However, our study only simply revealed that USF2 had a protective mechanism against sepsisinduced AKI by regulating THBS1 to affect the TGF-β pathway, the direct role in the treatment of clinical septic patients has not yet been deeply identi ed, which requires further study in future.

Table 1
Primers Sequence of qRT-PCR