Purple sweet potato color attenuated NLRP3 inflammasome by inducing autophagy to delay endothelial senescence

2.1 Cell culture and drug treatment

Human umbilical vein endothelial cells (HUVECs) were purchased from ScienCell Research Laboratories (ScienCell, Carlsbad, CA) and cultured in M199 medium with 10% fetal bovine serum (Thermo Fisher Scientific, Rockford, IL). In the following experiments, we adopted cells within a population doubling level 10, which were deemed nonsenescent cells as previously reported (Rogers, Zhang, Azhar, Luo, & Wei, 2013; Sun et al., 2010). Vascular endothelial cells obtained from mice transfected with green fluorescent protein (GFP) labeled microtubule-associated protein 1 light chain 3β (LC3) were isolated as previously reported (Marelli-Berg, Peek, Lidington, Stauss, & Lechler, 2000). Isolated cells were identified using 1,1′-dioctadecyl-3,3,3′,3′-tetramethyl-indocarbocyanine perchlorate-labeled acetylated LDL (Dil-Ac-LDL) uptake, and the positive rate was approximately 90%. These cells were named GFP-LC3 ECs in the following research.

Cells were divided into four groups: (a) cells incubated in culture medium with 5.5 mM d-glucose (normal group), (b) cells treated with 30 mM d-glucose to promote premature senescence (d-glucose group; Hayashi et al., 2014), (c) cells administered with 30 mM d-glucose and 50 μg/ml PSPC to disrupt the process of senescence according to our previous research (PSPC + d-glucose group; Sun et al., 2015), and (d) cells treated with 50 μg/ml PSPC in normal medium (PSPC group). PSPC extracted by microwave baking and acidified electrolyzed water was purchased from Tsingtao Pengyuan Natural Pigment Research Institute (Qingdao, China). All treatments were maintained for 72 hr.

To investigate the role of autophagy during endothelial senescence, HUVECs were exposed to 3-methyladenine (3-MA; 5 mM; autophagy inhibitor) for 2 hr or bafilomycin A1 (25 nM; inhibitor of autophagic flux; Sigma-Aldrich, St. Louis, MO) for 6 hr, respectively, before glucose or/and PSPC administration. Rapamycin at a concentration of 10 nM was applied to enhance autophagy. Cells challenged with an equal dose of dimethyl sulfoxide were regarded as the vehicle control. After pretreatment, all the cells were divided into four groups and treated as described above.

2.2 Detection of LC3 punctiform aggregation

HUVECs were incubated with LC3 primary antibody (Cat #2775; RRID:AB_915950; Cell Signaling Technology, Danvers, MA) overnight at 4°C and then incubated with the appropriate secondary antibody. Cells were mounted with ProLong Gold Anti-fade Reagent with DAPI (Life technology, Eugene, OR). Endothelial cells (ECs) isolated from GFP-LC3 mice were fixed and mounted with 4′,6-diamidino-2-phenylindole (DAPI). The ImageJ software (https://imagej.nih.gov/ij/) was utilized to quantify cells with LC3 dotted-distribution and determine the number of LC3 puncta.

2.3 Mouse model of diabetes

A diabetic mouse model was established as previously described (Sun et al., 2015). Briefly, 8-week-old male Institute of Cancer Research mice, which can be induced with type II diabetes mellitus by a high-fat diet (HFD; Li et al., 2012; Shan et al., 2014), were purchased from the Branch of National Breeder Center of Rodents (Beijing, China) and housed under conditions of constant temperature (23 ± 1°C) and humidity (60%), and a 12 hr light/dark schedule. After acclimatization, mice were randomly divided into the control group, HFD group, HFD + PSPC group, and PSPC group. Mice were fed with either normal diet containing 11.4% fat (control group) or HFD of 60% fat (HFD group). Half of the mice in the control group were allocated to PSPC in sterilized water at a dose of 700 mg/kg by oral gavage (PSPC group) every other day. Similarly, a population of HFD-treated mice was administered with PSPC (HFD + PSPC group). All the mice in these four groups were treated for 20 weeks. To confirm the establishment of type 2 diabetes mellitus, blood glucose values were detected using an Ascensia Elite glucose meter (Bayer Corporation, Mishawaka, IN), and a glucose tolerance test as well as an insulin tolerance test was performed as previously described (Shan et al., 2016).

All procedures were performed according to the policies established in the Guide to the Care and Use of Experimental Animals provided by the National Research Council. The protocol was approved by the Committee on the Ethics of Animal Experiments of the University of Jiangsu Normal University.

2.4 Tissue preparation

Mice were deeply anesthetized and killed. The hearts and aortas were rapidly removed after perfusion with ice-cold phosphate-buffered saline (PBS). The adventitia was thoroughly stripped.

The aortic roots were fixed in 4% paraformaldehyde and then immersed in O.C.T. Compound (Sakura Finetek, Torrance, CA). Serial cross-sections with a 7-μm-thickness were collected for immunofluorescence staining.

Both heart and artery samples were homogenized in lysis buffer (KeyGen Biotech, Nanjing, China) containing phenylmethylsulfonyl fluoride (PMSF) and protease inhibitor. The cold tissues were homogenized two times for 30 s intervals using MM400 (Retsch GmbH, Haan, Germany) and centrifuged at 12,000g for 10 min at 4°C. The supernatants were collected for western blot analysis. The concentration of samples was determined using a BCA protein assay kit (Thermo Fisher Scientific).

2.5 Western blot analysis

HUVECs with different treatments were lysed in lysis buffer (pH 8.0 Tris-HCl 20 mM, NaCl 137 mM, 1% NP40, 10% glycerol, and Na3VO4 1 mM). Protein extracts (20 μg of cells or 50 μg of tissue sample) were separated using 8% (mammalian target of rapamycin [mTOR] and NLRP3) or 15% (LC3, beclin-1, p62, etc.) sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto a polyvinylidene difluoride membrane (Millipore, Billerica, MA) by electrophoretic transfer (Bio-Rad Laboratories, Inc., Hercules, CA). The membranes were incubated with rabbit anti-LC3 antibody, rabbit anti-p62 antibody (Cat #5114; RRID:AB_10624872; Cell Signaling Technology), mouse anti-beclin-1 antibody (Cat #4122; RRID:AB_11178656; Cell Signaling Technology), rabbit anti-mTOR antibody (Cat #2972; RRID:AB_330978; Cell Signaling Technology), rabbit anti-p-mTOR (Ser2448) antibody (Cat #2971S; Cell Signaling Technology), rabbit anti-p-mTOR (Ser2481) antibody (Cat #2974S; RRID:AB_2231885; Cell Signaling Technology), mouse anti-NLRP3 (Cat #NBP1–97601; RRID:AB_11189153; Novus Biologicals, Littleton, CO), rabbit anti-caspase 1 (p10 + p45/42; Cat #ab179515; Abcam, Cambridge, MA), rabbit anti-IL 1β (35 + 17kD; Cat #16806-1-AP; Proteintech Group, Rosemont, IL), mouse anti-β-actin antibody (Cat #sc-47778 HRP; RRID:AB_2714189; Santa Cruz Biotechnology, Dallas, TX), and rabbit anti-p21 antibody (Cat #sc-469; RRID:AB_632121; Santa Cruz Biotechnology) at 4°C overnight. After incubation with horseradish peroxidase-conjugated secondary antibody, proteins were detected with enhanced chemiluminescence reagents (Bio-Rad Laboratories) using Amersham Imager 600 (GE Healthcare, Fairfield, CT). Quantification of the detected bands was analyzed using the ImageJ software. The data were normalized using β actin as an internal control and were standardized with the vehicle control as 1.0.

2.6 Analysis of cell cycle

Cell cycle was analyzed using a flow cytometer (ACEA Biosciences, San Diego, CA) according to the manufacturer's instructions. HUVECs were seeded onto a six-well plate at a density of 1.5 × 10⁵ cells/well. Cells treated for 3 days as described above were collected using centrifugation, washed with PBS, and then fixed with 70% ethanol at 4°C overnight. Subsequently, we washed and centrifuged the cells, and then resuspended pellets in PBS containing RNase (50 μg/ml) and propidium iodide (50 μg/ml). After incubation for 30 min, the mixture was detected using flow cytometer at the excitation wavelength of 488 nm.

2.7 Senescence-associated β-galactosidase (SA-β-gal) staining

As a widely used marker for senescence, SA-β-gal staining was performed as previously described (Sun et al., 2015). Cells were fixed using formaldehyde and then rinsed in staining solution at pH 6.0. The staining solution consisted of 5 mM of K₄[Fe(CN)₆]3H₂O, 5 mM of K₃[Fe(CN)₆], 150 mM of sodium chloride, 2 mM of magnesium chloride, and 1 mg/ml of X-gal. After overnight incubation at 37°C, the cells were washed in PBS and viewed using phase contrast microscopy. Cells stained blue were senescent cells. More than 2,000 cells were quantified in each sample. The equation for determining the proportion of senescence was as follows: (number of blue cells/number of total cells) × 100%.

2.8 RNA interference (RNAi)

HUVECs were transfected with specific human NLRP3 small interfering RNA (siRNA) oligonucleotides at a dose of 20 nM (Santa Cruz Biotechnology) for 6 hr using the RNAi Fect™ Transfection Reagent (Qiagen, Duesseldorf, Germany) in serum-free medium. Next, normal culture medium with serum was added to cells, and the silent efficiency was maintained for 3 days as previously observed (Sun et al., 2015). Control cells were treated with scrambled siRNA (Santa Cruz Biotechnology). To test the effect of high glucose or PSPC on endothelial senescence after NLRP3 silencing, glucose and/or PSPC was administered to cells during the RNAi process.

2.9 Double immunofluorescence staining

Analysis of colocalization between NLRP3 and LC3 or p62 in HUVECs was performed using double immunofluorescence staining. Briefly, cells in different groups were fixed and incubated with primary antibody rabbit anti-LC3 antibody or rabbit anti-p62 antibody, and then with the monoclonal antimouse NLRP3 antibody at 4°C overnight. Appropriate fluorescein isothiocyanate-conjugated or tetramethylrhodamine (TRITC)-conjugated secondary antibody was applied for 30 min at 37°C. After mounting with DAPI, the cells were monitored using a fluorescence microscope.

2.10 Coimmunoprecipitation

HUVECs were lysed in radioimmunoprecipitation assay lysis buffer. Equal amounts of total proteins (450 μg) were divided into three equal aliquots. The first aliquot was supplemented with 15 μg/ml purified nonimmune IgG (IgG group; Santa Cruz Biotechnology); the second one was mixed with a 1:100 dilution of monoclonal anti-NLRP3 antibody overnight at 4°C (NLRP3 group). The antibody-antigen complexes were then incubated with fresh protein A-Sepharose beads (Thermo Fisher Scientific) for 4 hr with gentle rotation. The third aliquot served as whole cell input without any operation.

The proteins-bead complexes were centrifuged and washed three times in lysis buffer. Lysis buffer was added to the beads to release proteins by boiling. Proteins from the input group, immunoglobulin G (IgG) group, and NLRP3 group were subjected to SDS-PAGE and immunoblotted with rabbit anti-p62 antibody or normal rabbit IgG antibody.

2.11 Immunofluorescence staining for histologic section

After fixation in 4% paraformaldehyde, antigen in the tissues was unmasked in boiling sodium citrate (pH 6.0) for 10 min. Nonspecific binding sites were blocked with 5% normal goat serum. Next, the sections were incubated respectively with rat anti-CD31 (Cat #sc-18916; RRID:AB_627028; Santa Cruz Biotechnology), rabbit anti-LC3 or rabbit anti-p62, as well as sheep anti-NLRP3 (Cat #AF6789; Novus Biologicals) at 4°C overnight. After washing in PBS, a mixture of appropriate goat antirabbit IgG (Alexa Fluor® 488) (Thermo Fisher Scientific), goat anti-rat IgG (TRITC), or donkey anti-goat IgG (Alexa Fluor® 647; Abcam) was added. Subsequently, DAPI was applied, and the stained sections were captured using a laser scanning confocal microscope (Leica, Wetzlar, Germany).

2.12 Statistics

All experiments were duplicated and repeated at least three times. Data were expressed as the mean ± standard error of the mean. The Student t test was used to make a comparison between the two groups. Differences were considered significant at p < 0.05.

3.1 PSPC promoted autophagy in senescent endothelial cells

Punctiform distribution of LC3β is widely used to monitor autophagy induction. We have demonstrated in a previous study that high glucose induced the premature senescence of endothelial cells, while 50 μg/ml PSPC could delay cellular aging. To understand whether autophagy was involved in cell senescence, we examined LC3 aggregation in cells using immunofluorescence staining. The aggregation of LC3 indicates the formation of autophagosomes or autolysosomes (Klionsky et al., 2008). As shown in Figure 1a,c a few of the normal cultured cells were positive with LC3 puncta. However, 30 mM d-glucose treatment decreased both the percentage of LC3-puncta-positive cells and the number of dots in the positive cells. In contrast, the number of reduced dots was reversed by PSPC, suggesting that PSPC might inhibit endothelial senescence by increasing autophagy. To further verify the enhanced autophagy by PSPC, we isolated the endothelium from the arterial arch and vascular vessel of transgenic mice with GFP-labeled LC3, a well-known model for monitoring autophagy, and tested the aggregation of LC3. The results in Figure 1b show that PSPC augmented the number of LC3 dots, which was impaired by d-glucose, and this was consistent with the results in HUVECs.

It is widely accepted that increased LC3 and decreased SQSTM1/p62 levels are associated with autophagy flux (Kliosnky, 2016). The levels of LC3 and p62 were detected using western blot analysis to investigate the roles of PSPC in autophagic flux. Compared with normal endothelial cells, premature senescent cells (d-glu group) exhibited lower levels of LC3 and beclin-1 (Figure 1d) but more p62 expression (Figure 1f). All alterations were reversed by PSPC (PSPC + d-glu group). In addition, changes in the p21 level, which was increased by high glucose but decreased by PSPC, confirmed the inhibitory efficiency of PSPC on cellular senescence (Figure 1d).

To further investigate the roles of autophagy in the endothelium related to cardiovascular diseases, immunofluorescence staining was performed to detect LC3 and p62 levels in endothelium using diabetic mice, in which endothelial senescence has been reported (Sun et al., 2015). Diabetic symptoms in mice were confirmed by the glucose and insulin tolerance test (Shan et al., 2016). The endothelium was specifically labeled by CD31. As shown in Figure 2a,b the expression of LC3 in the arterial endothelium was decreased after the mice were fed with a HFD (HFD group) as indicated by the arrow, compared with normal diet mice (CTR group). However, PSPC restored LC3 levels in the endothelium (indicated by an arrowhead). In addition, HFD promoted the accumulation of p62 but p62 levels were decreased by PSPC (Figure 2c,d). Furthermore, similar variation tendencies of LC3, p62, and beclin-1 were observed in homogenates of the heart and artery by western blot analysis (Figure 2e,f). Taken together, these results confirmed that PSPC enhanced autophagy to perturb endothelium senescence.

As an autophagy receptor, p62 is degraded during autophagy. In this study, two potential explanations accounted for the high levels of p62 in the d-glucose group: impaired autophagy or blockage of degradation. To elucidate the mechanism, we pretreated HUVECs with bafilomycin A1, which is often used to measure autophagic flux (Yamamoto et al., 1998), to disturb fusion between lysosomes and autophagosomes. Western blot analysis showed that p62 was increased in normal culture medium after pretreatment with bafilomycin A1. There was also a dramatic accumulation of p62 in d-glucose or PSPC + d-glucose treatment after bafilomycin A1 preincubation (Figure 1f). These results excluded the possibility of blockade of p62-degradation by d-glucose and suggested that PSPC suppressed endothelial senescence by promoting cellular autophagy.

3.2 Autophagy was responsible for PSPC-mediated endothelial senescence

Is autophagy a vital factor during cell senescence? To answer this question, we detected the phosphorylation of both ser2448 and ser2481 of mTOR in HUVECs. The level of phosphorylated mTOR (p-mTOR; ser2448) was elevated by high-glucose administration. However, high glucose failed to increase the phosphorylation of mTOR (ser2448) in the presence of PSPC. No significant difference in p-mTOR (ser2481) was observed after glucose or PSPC treatment (Figure 3a,b). In addition, eukaryotic initiation factor 4E-binding protein (4EBP1), a substrate of the mTOR complex 1 (mTORC1), was phosphorylated by glucose, while the phosphorylation of 4EBP1 was dramatically inhibited after PSPC administration (Figure 3a). In addition, increased p62 levels in d-glucose but decreased levels in the PSPC group confirmed the role of autophagy in this process. Taken together, these data revealed that PSPC directly deactivated the mTOR signaling pathway.

Subsequently, we interfered with endothelial autophagy to further understand the antagonism between autophagy and senescence. Rapamycin, which targets to the mTORC1 signaling pathway and is a well-known pharmacological promoter of autophagy, or 3-MA (autophagy inhibitor) was administered to HUVECs before glucose or PSPC treatment. The induced or inhibited efficiency was confirmed by changes in LC3-II, beclin-1, and p62 in Figure 5a. Next, SA β-gal staining was performed to determine the effect of autophagy dysfunction on endothelial senescence. The proportion of senescent cells in the high-glucose group was decreased once pretreated with rapamycin. In contrast, there was accelerated senescence in glucose-treated cells after 3-MA administration, and even PSPC could not prevent senescent progression (Figure 4c).

Furthermore, we examined the distribution of cell cycle by flow cytometric analysis. In the vehicle control group, high-glucose-exposed HUVECs were arrested at the G1 phase of the cell cycle and this arrest was recovered by the addition of PSPC (Figure 4a,b). Increased autophagy by rapamycin hindered glucose-induced G1 phase arrest and synergistically facilitated the inhibitory effect of PSPC on cell cycle arrest. However, HUVECs with impaired autophagic ability (3-MA group) exhibited a dramatic acceleration of senescence (higher proportion of G1/G0 cells) after d-glucose or/and PSPC treatment. Altogether, attenuated autophagy deteriorated endothelial senescence but reinforced autophagy alleviated senescence, indicating a crucial role of autophagy in PSPC-mediated endothelial senescence.

3.3 Inflammatory inhibition mediated by PSPC was autophagy-dependent

In our previous study, PSPC inhibited endothelial senescence by blocking the NLRP3 inflammasome (Sun et al., 2015). It has been reported that the NLRP3 inflammasome can be degraded by autophagy (Shi et al., 2012). Therefore, we questioned whether the anti-inflammatory effect of PSPC on endothelial senescence was mediated by increased autophagy. We unbalanced cellular autophagy using an enhancer or inhibitor to test the effect on glucose/PSPC-mediated senescence. We found that in the vehicle control group, glucose promoted the activation of the NLRP3 inflammasome, whereas administering PSPC in senescent cells (induced by high glucose) significantly reduced the activation of caspase-1 and release of IL-1β (Figure 5). These results were consistent with previous studies (Sun et al., 2015). However, once endothelial autophagy was enhanced by rapamycin, high glucose failed to activate the NLRP3 inflammasome, including a failure of increased NLRP3 expression, damaged segmentation of procaspase-1 (p45/42) into caspase-1 (p10), and less secretion of IL-1β from progenitors (pro-IL-1β; Figure 5). In contrast, the reduction in NLRP3 triggered by PSPC was completely reversed when the cells were pretreated with the autophagy inhibitor 3-MA (Figure 5). Moreover, few precursor proteins of p10 and IL-1β (p45/42 and pro-IL-1β) but increased p10 and IL-1β were observed following a challenge with 3-MA. Taken together, these results suggested an antagonistic correlation between autophagy and the NLRP3 inflammasome.

Several studies have demonstrated that activation of the NLRP3 inflammasome initiates autophagosome formation and induces autophagy (Hayrabedyan et al., 2016). Based on this finding, we examined whether NLRP3 activation by high glucose in this study may function as a feedback mechanism to affect autophagy. As shown in Figure 6, compared with the control siRNA group, no obvious alterations in LC3 transformation and beclin-1 accumulation were observed after NLRP3 interference. These results indicated that endothelial autophagy regulated by glucose or PSPC was an upstream signal of the NLRP3 inflammasome.

3.4 NLRP3 inflammasome colocalized with the autophagy receptor

To further elucidate the mechanism underlying NLRP3 inflammasome deactivation by autophagy, double immunofluorescence staining was performed to analyze the localization of NLRP3 and autophagy receptor. Colocalization of NLRP3 and LC3 was observed in endothelial cells (indicated by an arrow in Figure 7a). This phenomenon disappeared in cells treated with glucose or PSPC due to scarce LC3 conversion in the d-glucose group or low levels of NLRP3 in d-glucose combined PSPC treatment. All these results suggested that the NLRP3 inflammasome might overlap with the autophagosome.

To further support this assumption, we tested the distribution of another autophagy receptor p62, which serves as a molecular bridge for the delivery of substrates to the autophagosome (Seibenhener et al., 2004). The immunofluorescence staining in Figure 7b shows similar variation tendencies of p62 and NLRP3, suggesting that cells exposed to high glucose exhibited strong expression of p62 and NLRP3. Interestingly, the distributions of p62 and NLRP3 in glucose-treated cells were superposed (indicated by an arrow in Figure 7b). In addition, overlapping distribution of p62 and NLRP3 was detected in the senescent endothelium of diabetic mice (HFD group in Figure 7c), indicating the colocalization of the autophagosome and NLRP3 inflammasome.

Furthermore, we performed coimmunoprecipitation assays and found that p62 was detected in all three groups (normal, d-glucose, PSPC + d-glucose) in the immunoprecipitations obtained with anti-NLRP3 antibody but not with IgG (Figure 7d). A reciprocal coimmunoprecipitation assay showed similar results (data not shown), suggesting that the two proteins interacted in endothelial cells. Interestingly, compared with the other two groups, much greater p62 was pulled-down by the same amount of NLRP3 antibody after high glucose administration (Figure 7d), suggesting that glucose promoted the accumulated p62 to bond with NLRP3. These results confirmed that autophagy targeted inflammasome components to degrade NLRP3 inflammasomes during PSPC-modulated endothelial senescence.