STS Protects Diabetic Glomerular Vascular Endothelial Barrier by Ameliorating EPC Dysfunction: Targeting RAGE-TXNIP-NLRP3 In ammasome Pathway


 Background: Glomerular endothelial cell (GEC) injury is one of the crucial causes of diabetic kidney disease (DKD). Endothelial progenitor cell (EPC) is the essential mechanism of vascular endothelial repair, which damages by diabetic pathology. Sodium Tanshinone Sulfonate ⅡA (STS) is known to protect endothelium, but the mechanism and the role in DKD need to be studied. Methods: EPC was treated with high glucose (HG), and thioredoxin interacting protein (TXNIP), NLR family pyrin domain containing 3 (NLRP3) inflammasome, DNA damage, proliferation, differentiation and senescence were detected; STS and EPC were intravenous injected into diabetic nude mice, the urine protein quantitation and urine protein/creatinine were detected; the Dil-labeled EPC was traced and the expression of TXNIP, caspase-1 (p20), p21, Ki67, CD31 were detected by fluorescence co-location in glomerulus.Results: We found that STS inhibited HG-induced TXNIP expression and NLRP3 inflammasome activation, catalase (CAT) inactivation, DNA damage, senescence； STS restored EPC proliferation and differentiation functions; advanced glycation end products (AGEs) produced in HG treated EPC supernatant, the receptor of AGE (RAGE) blocking inhibited TXNIP expression and NLRP3 inflammasome activation, which mimicked by STS. STS protected EPC functions in diabetic glomerular and enhanced EPC renal function amelioration. Conclusions: We concluded that STS watched CAT activity to prevent HG-induced EPC DNA damage, proliferation, differentiation dysfunction, accelerated senescence by inhibiting the RAGE-TXNIP-NLRP3 inflammasome-caspase-1 pathway.


Background
Diabetic kidney disease (DKD) is one of the major microvascular complications of diabetes and the leading cause of the end-stage renal disease (ESRD) [1]. The main symptom of early DKD is persistent microalbuminuria [2]. Proteinuria may be associated with glomerular ltration barrier (GFB) injury. GFB simply divides into three layers: glomerular endothelial cell layer (GEC), glomerular basement membrane, podocyte layer [3]. As the rst GFB barrier directly contacts with blood, it has been proved that the glomerular endothelium (GE) cells secreted vasoactive substances and played an essential role in maintaining the glomerular homeostasis and the dysfunction of GE was a marker of the early progression of DKD, leading to the damage of GFB and the formation of proteinuria [4]. GEC fenestrations are trans-cytoplasmic holes, which accounts for 20%-50% of the surface area of the endothelial cell layer, and the diameter of the endothelial pores is about 60-100nm, which is 15 times more than the diameter of the albumin molecule (3.6nm) [5]. The negatively charged proteins synthesized and secreted by GEC combine with glycosaminoglycan in covalent bonds to form a networked layer of the glomerular endothelial surface layer (ESL) to ll the GEC hole to prevent leakage of albumin [6]. GEC loss may cause EPC induced by the high glucose environment. The present study demonstrated that excess glucose might form advanced glycation end products (AGEs). It activated the receptor of the advanced glycation end product (RAGE)-thioredoxin-interacting protein (TXNIP)-NLR family pyrin domain containing 3 (NLRP3) in ammasome pathway to eliminate catalase (CAT) activity and cause EPC dysfunction. STS restored the proliferation and differentiation function of EPC and slowed down its aging by reducing TXNIP expression to inhibit the NLRP3 in ammasome-caspase-1 pathway-dependent CAT inactivation in diabetes.
2. Methods 2.1 Isolation, culture, and characterization of mouse bone marrow-derived EPC EPCs were isolated from mouse bone marrow according to the previous study [29]. Bone marrow medulla was centrifuged by Density Gradient with a Human peripheral blood lymphocyte isolation uid (Ficoll Plus 1.077, Solarbio, P4350, Beijing China) to obtain the bone marrow Mononuclear Cells (MNCs). MNCs were plated in bronectin (Solarbio, F8180, Beijing, China) coated cell culture bottle with a ventilation lter and maintained in Endothelial Growth Medium-2 (EGM-2, Lonza, CC-4176, USA) with 10% fetal bovine serum (FBS, EPHRAIM, 26-500-FBS China) in a cell incubator with 37℃ and 5% CO 2 . The medium was replaced every 48h. Cell colonies those appeared about the rst 7days later, these cells had a fusiform morphology and were de ned as the early EPC [30]. After 28 days of culture, cells emerged abilities of proliferation and tube formation, had a pebble-like compact arrangement as endothelial cells (EC), and were called endothelial colony-forming cell (ECFC), outgrowth endothelial cells (OECs), or late-EPC [31].
High-fat diet (HFD, 10% saccharose, 10% lard, 10% sugar, 5% egg yolk powder, 0.5% cholesterol, 64.5% basal chow) and a low dose of streptozotocin (STZ) to develop diabetes according to the method described previously [33]. Brie y, after 4 weeks of feeding with HFD, mice were administered STZ (65mg/kg body weight, pH 4.5) or citrate buffer (vehicle) by intraperitoneal (i.p.) injections once a day for 5 days. After one week, the mice fasted for 12h for fasting blood glucose (FBG) testing. Plasma glucose concentration was detected using a commercial glucometer (GA-3, Sinocare, Changsha, China). Fasted mice with blood glucose levels higher than 11.1mmol/L were considered as diabetic mice. 24h urinary albumin quantitation and plasma and urinary creatinine quanti cation and urinary albumin-to-creatinine ratio were used to evaluate renal function in diabetic mice. Urinary albumin and creatinine were detected by a commercial kit purchased from Jiancheng bioengineering institute (C035-2-1 and C011-2-1 Nanjing, China)

HE and PAS staining of the glomerulus
Mouse kidney was perfused by PBS and xed in 4% PFA, dehydrated and embedded in para n. Para n sections were cut at 5µm. After depara nization and rehydration, glomerulus sections were dyed by hematoxylin solution (hematoxylin 1g, sodium iodate 0.2g, aluminum potassium sulfate, 50g, citric acid 1g, chloral hydrate 50g, distilled water added to 1 liter) for 10 minutes, alcoholic eosin (2.5g eosin, 500ml distilled water, hydrochloric acid 10ml, lter product dissolved in 1000ml 95% alcohol, double diluted before use) for 30 seconds respectively.
Periodic Acid-Schiff stain (PAS) of glomeruli was performed using the PAS stain kit (Shanghai Yuanye Bio-Technology Co., Ltd, Shanghai, China). After depara nization and rehydration, glomerulus sections were immersed in periodic acid for 5min, washed by running water and distilled water respectively, then in Schiff reagent for 10min in the dark. After being washed with running water for 10min, the nuclear was stained by the hematoxylin solution. The sections were washed with running water until the nuclear turning blue.
The sections were dehydrated by the gradient of ethanol, and graphs were captured by microscope (Carl Zeiss, Axio Scope A1, Germany) in the light eld mode.

Catalase activity detection
The activity of catalase was detected by in-gel catalase stain according to the previous study [35]. Cells were lysed by RIPA lysis buffer (Beyotime, P0013B, Shanghai, China) with PMSF (Beyotime, ST506, Shanghai, China) on ice, and the extract was centrifuged for 10min at 4℃ and 10000g. The supernatant was mixed with non-reducing 4×SDS loading buffer (50mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 12.5 mM EDTA, 0.02% bromophenol blue) and electrophoresed on a 10% PAGE gel in electrophoresis buffer (3.025g tris, 14.4glycine dissolve in 1L deionized water). The gel was washed with deionized water three times and then immersed in 0.003% H 2 O 2 (2µl 30% H 2 O 2 added to 20 ml deionized water) for 10min and stained with a freshly staining solution (2% potassium ferricyanide and 2% ferric chloride) until the bright band appeared. After three times washed by running water, the gels were photographed by an automatic chemiluminescence/ uorescence image analysis system (5200 Multi, Tanon, Shanghai, China). The αtubulin detected by western blot was used to normalize the total protein.

Catalase mutant and HEK293T cell transfection
Catalase (CAT) point mutants were performed by transfecting pcDNA3.1 plasmid which was purchased from GENEWIZ (Suzhou, China). By calculating the molecular weight of the CAT fragment, aspartic acid sites in D298, D307, and D226 were replaced with alanine, generated D298A, D307A, and D226A CAT plasmids, and the wild type CAT was also constructed into pcDNA3.1 plasmid. These mutant plasmids were delivered into EPCs by Lipofectamine™ 2000 Transfection Reagent (Invitrogen, 11668019, USA) and HEK293T cells (Fenghbio, Changsha, China) by Calcium Phosphate Cell Transfection Kit (Beyotime, C0508, Shanghai, China). EPCs were treated by LG, HG, HG with LPS+ATP and WEHD, HEK293T cells were transfected by NLRP3, ASC, and caspase-1 plasmids those purchased from Fenghbio (Changsha, China), and the NLRP3 in ammasome was activated by LPS+ATP treatment. The protein samples extracted from EPCs and HEK293T cells were detected by western blot and co-IP.

BrdU incorporation assay
BrdU incorporation assay was performed as described previously [37]. EPCs were seeded on coverslips in six-well plates with LG and HG medium to 60-70% con uences and then incubated with serum-deprived LG and HG medium for 12 h. EPCs were labeled with BrdU (10 µΜ, Solarbio, Beijing, China) for 2h. After xed by 4% PFA, perforated by 0.3% Triton X-100 and denatured by 2M HCl, EPCs were incubated with mouse anti-BrdU primary antibody (1:100, Proteintech™, Wuhan, China) overnight, followed by Cy3conjugated Goat Anti-mouse IgG (H + L) (1:100, Proteintech TM , Wuhan, China), nuclei were labeled by DAPI. The immuno-uorescence images were taken by a Fluorescence microscope (Carl Zeiss, scope A1, Germany). The percentage of BrdU positive cells to the total amount of cells was calculated by Image J software (NIH, Littleton, CO, USA).

AGEs detection
The abundance of AGEs in EPC supernatant and mice serum was detected using a commercial kit

Hydrogen Peroxide detection
Hydrogen Peroxide content in EPC was detected using a commercial kit (Solarbio, BC3595, Beijing, China). 2×10 7 treated cells were collected in reagent , and the cell membrane was broken by intermittent ultrasonic, cell debris was centrifuged at 8000g for 10min, the supernatant was collected. According to speci cation, the absorbance at 415nm was detected by a microplate reader (Molecular Devices SpectraMax M4, USA). The sample hydrogen peroxide content was calculated according to the absorbance of the standard substance with known concentration.

EPC SA-β-gal staining
Accumulation of Senescence-Associated β-galactosidase (SA-β-gal) is one of the hallmaker of senescence. EPCs (1×10 5 /ml) were seeded on sterilized coverslips, synchronized growth by FBS-free medium, and incubated by LG, HG for 48h; the NLRP3 in ammasome positive control was induced by LPS treatment for 4h and ATP treatment for 1h. EPCs were washed twice by precooled PBS and xed by an immobilized reagent in SA-β-gal kit (C0602, Beyotime, Shanghai, China) for 15min. Staining solution (Formulated as per kit instructions, 1ml containing A, B, C and X-Gal solutions 10µl, 10µl, 930µl, and 50µl) was added and incubated the cells for 36h at 37℃, avoiding light. The coverslips were moisturized with glycerine PBS (1:1) solution, observed, and photographed by a microscope (Carl Zeiss, Germany, Axio Scope A1) in the light eld. The images were processed using ZEN Blue 2.3 software (Carl Zeiss, Germany). Image J software was used to analyze the percentage of SA-β-gal positive area (green staining).

Statistics
Statistical analysis was performed with IBM SPSS Statistics 20 software. Data are presented as means ± SE. Signi cant differences between and within multiple groups were examined using ANOVA for repeated measures, followed by Duncan's multiple range test. The Independent-Samples t-test was used to detect signi cant differences between the two groups. p< 0.05 was considered statistically signi cant.

STS inhibited EPC NLRP3 in ammasome activation and dysfunction induced by HG
To observe the activation of the NLRP3 in ammasome by HG on EPC, we used western blot to detect the changes of relevant indicators. We found that the NLRP3 expression and the degree of procaspase-1 converted to caspase-1 and proIL-1β to IL-1β were increased by HG, which was similar to NLRP3 in ammasome positive control, LPS+ATP treatment ( Figure 1A-1B). Further, immuno uorescence results showed that HG and LPS+ATP treatment induced signi cant merging of NLRP3 (red uorescence dots) and ASC (green uorescence dots) to form yellow dots, which represented NLRP3 in ammasome assembly, STS inhibited the assembly ( Figure 1E). These changes indicated that HG caused signi cant activation of NLRP3 in ammasome in EPC. STS inhibited the activation.
Similar to LPS+ATP treatment, HG signi cantly caused the increase of DNA damage index γ-H2AX expression and positive immuno uorescence in the nucleus ( Figure 1C-1D, 1F), cell senescence-related index p21 expression ( Figure 1C-1D), and senescence-associate beta-galactosidase (SA-β-gal) positive area ( Figure 1G). At the same time, HG caused the decrease of cell proliferation index Ki67 ( Figure 1C-1D) and EPC differentiation function index CD31 and VWF expression on cytomembrane ( Figure 1C-1D). STS treatment alleviated the above EPC damage changes caused by both HG and LPS+ATP.
These results indicated that STS alleviated HG-induced EPC DNA damage, down-regulated proliferation, differentiation, and accelerated senescence while inhibits HG-induced activation of EPC NLRP3 in ammasome.

STS ameliorated HG impaired EPC functions by Inhibiting NLRP3 in ammasome activation
To investigate the possible role of in ammasome activation in HG-induced EPC function impairment, we used NLRP3 in ammasome-speci c inhibitor MCC950 to intervene [36]. The results showed that similar to MCC950, STS signi cantly inhibited the production of in ammasome activation markers active caspase-1 and mature IL-1β (Figure 2A-2B), inhibited the expression of γ-H2AX ( Figure 2C-2D) and restored the expression of cell proliferation indexes Ki67 expression ( Figure 2C-2D) and BrdU incorporation positive cell area ( Figure 2F), EPC differentiation indexes CD31 expression ( Figure 2C-2D) and immuno uorescence positive area on the cell membrane ( Figure 2E) and the membrane expression of VWF ( Figure 2C-2D).
These results indicated that STS improved HG-induced EPC dysfunction by inhibiting the activity of the NLRP3 in ammasome.

STS alleviated EPC DNA damage and dysfunction by inhibiting Caspase-1 dependent catalase activity downregulation
The above results concluded that inhibition of NLRP3 in ammasome activity might improve HG-induced EPC damage. So, we speculated that caspase-1, the core active product activated by the in ammasome, might play a key role in EPC dysfunction. We used a caspase-1 inhibitor, Z-WEHD-FMK [37], treated EPC, and the results showed that similar to WEHD, STS restored HG damaged CAT activity, and reduced the cleaved CAT production (a fragment with a molecular weight of 34kDa) ( Figure  In addition, we used exogenous CAT intervention to observe the EPC function changes. Results showed that exogenous CAT enhanced the protective effect of STS on the HG-induced CAT inactivation, EPC DNA damage, Ki67, CD31 and VWF down-regulation, the p21up-regulation, and hydrogen peroxide accumulation ( Figure 3F-3G, 3I).
These results indicated that NLRP3 in ammasome mediated caspase-1 activation weakened CAT activity and led to EPC DNA damage and dysfunction; STS reversed the damage by protecting CAT activity from caspase-1 activation induced by HG.
3.4 STS protected CAT activity by inhibiting NLRP3 in ammasome dependent caspase-1 activation To investigate the potential mechanisms of STS protecting HG impaired EPC function, we observed CAT activity following inhibition of NLRP3 in ammasome activation. Results showed that similar to MCC950 treatment STS reduced the active caspase-1 (p20) production, restored the impaired CAT activity, and reduced the Hydrogen peroxide accumulation induced by HG or LPS+ATP ( Figure 4A-4B), STS inhibited the HG or LPS+ATP treatment-induced CAT cleavage ( Figure 4A-4B). More importantly, we further investigated the possible interaction between caspase-1 and CAT by Co-immunoprecipitation (Co-IP) and immuno uorescence co-localization. Results showed that MCC950 reduced the degree of Co-IP of caspase-1 and CAT ( Figure 4C-4D); meanwhile, MCC950 inhibited the formation of yellow spots, which represented co-localization of CAT (red spots) and caspase-1 (green spots) ( Figure 4E-4F), STS mimicked the MCC950 effect on inhibiting the interaction between caspase-1 and CAT ( Figure 4C-4F). These results indicated that STS inhibited NLRP3 in ammasome dependent active caspase-1 production, and STS protected CAT activity by inhibiting NLRP3 in ammasome dependent caspase-1 activation.
We super cially explored the possible cleavage sites of CAT by caspase-1 using the point mutation method. According to molecular weight estimation and software (Editseq 7.1.0) calculation, we mutated aspartic acid 298, 307, and 226 sites in CAT amino acid sequence into alanine. The point mutant plasmids were transfected into HEK293T cells, which transfected with NLRP3, ASC, and caspase-1 plasmids simultaneously. The NLRP3 in ammasome was activated by LPS+ATP, and then the splice of CAT after the mutant site was observed. The results showed that the D298A mutation prevented CAT splicing following activation of NLRP3 in ammasome ( Figure 4G-4H), and D298A mutation also prevented CAT and p20 coprecipitation ( Figure 4I-4L). These results indicated that active caspase-1 deactivated CAT by cleaving CAT at D298.
These results indicated that STS protected CAT activity by inhibiting NLRP3 in ammasome-dependent caspase-1 activation and CAT cleavage at the D298 site.

STS improved CAT activity and EPC function by inhibiting the RAGE-TXNIP-caspase-1 pathway
TXNIP is a recognized protein that activates NLRP3 in ammasomes [38]. TXNIP is highly expressed in the HG environment and causes extensive intracellular oxidative stress and upregulation of in ammatory status. EPCs were treated with TXNIP siRNA for observation. Results showed that TXNIP gene silencing signi cantly inhibited the activation of caspase-1, CAT activity damage, hydrogen peroxide accumulation, γ-H2AX, and p21 expression, and restored the expression of Ki67, CD31, and VWF, and STS had a similar effect to TXNIP siRNA ( Figure 5A-5C). These results indicated that STS restored CAT activity and EPC proliferation and differentiation functions, inhibited DNA damage and cell senescence of EPC by inhibiting the TXNIP-caspase-1 pathway.
We further attempted to explore the possible mechanism by which HG impaired EPC function. The content of AGEs in EPC culture supernatant was detected by a commercial AGEs Assay Kit. The results showed that with the increase of HG treated time and concentration, many AGEs were positive in the culture medium ( Figure 5D-5E). These results meant that the excess glucose in the culture medium induced AGEs formation. Therefore, we speculated that the AGE-RAGE-TXNIP pathway is likely to play an essential role in impairing EPC function by HG. EPCs were treated with BSA-AGE, and the RAGE pathway was blocked by TTP488, a small-molecule inhibitor of RAGE [39]. The results showed that TTP488 inhibited the TXNIP expression, activation of caspase-1, CAT activity damage, hydrogen peroxide accumulation, γ-H2AX, and p21 expression, and restore the expression of Ki67, CD31, and STS had a similar effect to TTP488 ( Figure 5F-5H). These results indicated that excess glucose was likely to activate the RAGE-TXNIP-caspase-1 pathway through the formation of AGEs analogs to damage CAT activity, resulting in DNA damage, and ultimately impaired the proliferation and differentiation of EPC functions and accelerated EPC senescence.

STS enhanced EPC therapy of ameliorating renal function in diabetic mice
To observe the effect of EPC transplantation on the kidney function in diabetes, we transplanted EPC into diabetic nude mice and observed the changes in kidney functions and the pathological change of glomeruli. The results showed that EPC transplantation did not signi cantly reduce the blood and urine sugar ( Figure 6C-6D), but decreased 24h urine protein level ( Figure 6E), the plasma creatinine ( Figure 6F), restored the urine creatinine excretion ( Figure 6G), and suppressed the ratio of urinary albumin/creatinine ( Figure 6H) in diabetic mice. STS enhanced the above-improved changes in kidney functions by EPC transplantation (Figure 6E-6H). Pathological examination showed that EPC transplantation reduced glomerular cell loss and amaranth staining ( Figure 6A-6B) in STZ treated mice kidneys. STS treatment enhanced the glomerular changes ( Figure 6A-6B).
These results suggested that STS prior intravenous infusion enhanced the therapeutic effects of EPC transplantation on glomerular cell well-organizing and glycogen deposition inhibition and kidney function improvement in diabetic mice.

STS protected EPC function in diabetic mice glomeruli
To observe the EPC function changes and the STS protection in diabetic mice glomeruli, CM-Tracker labeled EPCs were transplanted, and immuno uorescence in the frozen glomerular section was detected.
Results showed that EPC (red uorescence) gathered in STZ treated mice glomeruli, but the control glomeruli had very little EPC ( Figure  These results suggested that glomerular vascular endothelium was repaired and integrated by EPC homing, STS protected EPC from TXNIP expression, in ammasome activation, and senescence, restored EPC proliferation and differentiation functions in STZ-induced diabetic pathological environment.

Discussion
Based on the activation of the NLRP3 in ammasome, this study discussed that the decrease of EPC proliferation and differentiation functions, accelerated senescence induced by HG, the kidney function bene t from EPC transplantation were weakened in diabetic mice, STS protected EPC functions from the diabetic pathological environment. In terms of mechanism, excess glucose activated the RAGE-TXNIP-NLRP3 in ammasome pathway through the production of AGEs analogs, leading to caspase-1-mediated CAT inactivation, resulting in the accumulation of hydrogen peroxide, DNA damage, and eventually EPC dysfunction. STS blocked the RAGE-TXNIP-NLRP3 in ammasome pathway to protect the CAT activity, block DNA damage and restore the proliferation and differentiation functions of EPC, slow down its senescence.

The origin and identi cation typing of EPC remains controversial
At present, there are some controversies about the source of EPC. For example, some studies proved that EPC does not derive from bone marrow [40]. However, most studies believe that EPC is a bone marrowderived stem cell [41]. Consequently, these controversies indicate that EPC has multiple sources, and this issue needs further research. At present, the typing characteristics of EPC have not been completely determined. Some studies believe that the following functions of the cells are different according to the different antigen markers those emerge on the cell membrane. For example, CD14, CD16, CD34, and VEGFR2 positive cells play the role of angiogenesis, but early EPC also expresses surface markers of bone marrow-derived stem cells such as CD14 and CD45 [30]. The most recognized EPC identi cation includes that cells express positive cell surface markers CD34, CD133, and VEGFR2 (Flk1/KDR); endothelial markers represented by VWF and CD31 are gradually expressed in the later stage; cells can be double labeled with Dil-Ac-LDL and FITC-UEA-I; and cells have the physiological function of repairing damaged vascular endothelium and participating in angiogenesis [32,42].
Our results showed that compared with mature endothelial cells MAVEC, the isolated and cultured mouse bone marrow-derived cells had all three makers, CD34, CD133, and VEGFR2 positive (supplemental Figure   1B), and were double uorescence-labeled by Dil-Ac-LDL and FITC-UEA-I (supplemental Figure 1C). The early fusiform growth cell was observed under a light microscope compared with cobble-like colony growth cell and tubule formation ability at the later stage (supplemental Figure 1A). These characteristics were similar to most current studies' conclusions, so the bone marrow-derived cells we isolated and cultured had the recognized EPC characteristics.
EPC transplantation alleviated the kidney dysfunction in diabetic mice The protective effect of EPC on glomerular injury has been reported. In vitro studies, EPC secreted: vascular growth factor, insulin-like growth factor 1, broblast growth factor 2, and Hepatocyte growth factor to restore the damage of human glomerular endothelial cells caused by Shigellin [43]. In vitro tissue engineering studies have shown that CD133 positive cells had a potential role in forming arti cial glomeruli [44]. EPC transferred encode Factor H, CD55, and CD59 mRNA into mesangial cells through extracellular vesicles inhibited mesangial cell apoptosis induced by anti-Thy1.1 antibody/complement and C5B-9 /C3 mesangial cell deposition in anti-Thy1.1 glomerulonephritis [45]. It has been summarized that the endothelial repair mechanism was impaired in the diabetic state, and diabetic patients with proteinuria were accompanied by a decrease in circulating EPC count [46]. In vivo studies have also shown that transplantation of CM-Dil labeled bone marrow-derived stem cells restored the number of glomerular endothelial cells and reduced the area of α-SMA positive regions in the mesangial region and the degree of macrophage in ltration in a rat model of anti-Thy1.1 glomerulonephritis [47]. Conclusions of clinical studies also indicated that EPC had a mitigating effect on DKD[48]; moreover, the impairment of EPC function in the process of chronic renal failure might be associated with in ammation [49]. These ndings suggested that EPC transplantation very likely had the potential repairment of the damaged glomerular vascular endothelial cell layer, and the mechanism might be related to EPC differentiate into mature vascular endothelial cells and the EPC paracrine effect. Our conclusion is similar to those in previous studies, EPC transplantation improved glomerular dysfunction. EPC transplantation signi cantly inhibited proteinuria production, reduced plasma creatinine, increased urinary creatinine excretion and limited the ratio of urinary albumin/creatinine in diabetic mice. More, EPC transplantation provided better organized cell arrangement and less glycogen deposition in diabetic glomerulus. However, the effect of EPC paracrine action on glomerular function has not been involved in this study. Our results focused more on the EPC functions of proliferation and differentiation, as well as the role of cell senescence in DKD, because these results are still lacking.

Caspase-1mediates CAT inactivation
As the core product of the assembly of the NLRP3 in ammasome, caspase-1 has the speci c aspartic acid cysteine proteinase activity and the most widely known role in mediating the maturation of IL-1β and IL-18, and the activation of GSDMD, which plays the crucial role in pyroptosis [50]. Furthermore, according to studies, there are more than 1000 substrates of caspase-1, those are affected cytoskeleton structure, programmed cell death, glucose, and lipid metabolism, autophagy ow, and even cell differentiation and other cellular functions [51][52][53][54][55]. The results presented in this study showed that caspase-1 might interact with CAT. The active caspase-1 cleaved CAT to produce a fragment of about 34kDa, and then CAT enzyme activity decreased, which further led to hydrogen peroxide accumulation, DNA damage, and eventually EPC dysfunction. Previous ndings seemed to provide some potential correlation between the NLRP3 in ammasome-caspase-1 pathway and CAT activity. CAT activity tends to decrease with activation of NLRP3 in ammasomes[56-58],but inhibition of NLRP3 in ammasome played a role in restoring CAT activity [59]. However, it is still unclear whether the reason for the decrease in CAT activity is related to the activity of caspase-1. A study showed that CAT might be one of the substrates of caspase-  [71,72], and the RAGE-TXNIP pathway has also been shown to play a crucial role in a variety of cellular functional impairments [73][74][75]. Therefore, our results concluded that HG activated the RAGE-TXNIP-NLRP3 in ammasome pathway through the formation of AGEs. The pathway was rstly observed in EPC and revealed one of the underlying mechanisms of EPC dysfunction in diabetes.

STS protected EPC function in diabetic pathology
A few studies have shown that tanshinone A has a protective effect on EPC function impairment, and the mechanism was related to the inhibition of the production and release of in ammatory factors associated with TNF-α-induced activation of NFκB signal. For example, Tanshinone A reversed TNF-αdamaged proliferation, migration, adhesion, and angiogenesis function of EPC, and inhibited EPC secretion of in ammatory cytokines [27], also, inhibition of NF-κB activation and phosphorylation of IκB-α down-regulated EPC vascular cell adhesion molecule-1 (VCAM)/intracellular adhesion molecule-1 (ICAM-1) expression, those play a key role in the recruitment of in ammatory cell in ltration in early atherosclerotic lesions[28]. Our results showed a similar conclusion that STS inhibited NLRP3 in ammasome, improved EPC proliferation, and differentiation, and slowed down its senescence. Our results suggested that STS inhibited the activity of NLRP3 in ammasome by inhibiting TXNIP expression, which is similar to the conclusion that STS protects against ischemia/reperfusion myocardial injury [76]. Our ndings demonstrated for the rst time that STS protected EPC function by inhibiting the RAGE-TXNIP-NLRP3 in ammasome pathway and improved diabetic glomerulopathy, enriched the protective effect of STS on EPC function, and con rmed its mechanism related to in ammation. It has been reported that STS enhanced CAT activity and protein level of CAT in Human Neuroblastoma Cells [77]; HG inhibited CAT expression and induced oxidative stress in EPC [29]. But the CAT impairing mechanism is poorly studied now. Our results suggested that HG promoted CAT cleavage and inactivation by forming AGEs to activate the RAGE-TXNIP-NLRP3 in ammasome-caspase-1 pathway, which inhibited by STS to protect CAT activity. The renal protective effects of STS have been reported. Tanshinone IIA blocked renal brogenesis and in ammation by inhibiting the GSK3β activation in folic acid-induced acute kidney injury, and the following progressive chronic kidney disease [78], and also Tanshinone IIA inhibited the thickening of glomerular basement membrane and the collagen deposition by attenuating PERK signaling activities to reduce endoplasmic reticulum stress in STZ-induced diabetic nephropathy [23]. But the protective effect of STS on glomerular capillary is lack of study. Our results suggested that STS protected the glomerular vascular endothelium repairment of EPC and enhanced the renal function recovery by EPC transplantation in STZ induced diabetic mice. However, our conclusion about glomerular vascular endothelium repairment of EPC was inferred mainly based on the recovery of CD31 membrane expression after EPC transplantation and STS injection, the glomerular vascular endothelium repairment of EPC needs further study in a more microscopic way. The VWF and CD31 membrane expression was insu ciently studied in EPC differentiating into mature endothelial cells, and the mechanism of VWF and CD31 coming to the surface of cytomembrane also needs to explore in further study.

Conclusions
Our results conclude that under diabetic conditions, HG promoted CAT cleavage and inactivation through activation of the RAGE-TXNIP-NLRP3 in ammasome-caspase-1 pathway, resulting in the accumulation of hydrogen peroxide and DNA damage and EPC function impairment. STS protected CAT activity and EPC function from RAGE-TXNIP-NLRP3 in ammasome pathway activation. EPC transplantation effectively restored the renal function in diabetic mice, which was enhanced by STS treatment. Our

Availability of data and materials
The datasets used and analyzed in current study are available from the corresponding author based on reasonable request Author contributions: HYY and WW: laboratory work, data analysis and interpretation, manuscript writing, and nancial support; WW: conception and study design and nal approval of the manuscript; ZXY, WFF, ZPF: manuscript proofreading and administrative support. All authors have read and approved the manuscript.     Representative PAGE gel and summarized data show the CAT activity, representative Western blot gel and summarized data show and the degree of CAT cleavage and p20 production (the full-length gels were showed in supplemental gure 2-D); (C-D): Representative Co-immunoprecipitation gel and summarized data show the p20-CAT and CAT-p20 coprecipitation; (E-F): Representative immuno uorescence images