HUVECs were obtained from American Type Culture Collection (Manassas, VA, USA) and cultured in Dulbecco's Modified Eagle's Medium (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% heat-inactivated fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.), 80 U/ml penicillin, 0.08 mg/ml streptomycin and 2 mM glutamine at 37°C under humidified atmosphere with 5% CO₂.

All animal experiments were performed with the approval of the Cangzhou Central Hospital Ethics Committee for Animal Experimentation (Cangzhou, China). A total of 60 male C57BL/6 mice (age, 8 weeks; weight, 20–25 g) were purchased from the Medical Center of Yangzhou University (Yangzhou, China). Mice were housed and fed at 27°C, and 12-h light/dark cycle and had access to food/water. Sepsis was induced by CLP, following a previously published protocol (8). HuR-siRNA (10 nM; Biomics Biotechnology Inc., Nantong, China; Table I) was injected in the tail vein 3 days post-sepsis to detect the effects of HuR on CLP-induced HMGB1 release, IL-6 production and mortality, a scrambled-siRNA was used as a control, following the manufacture's siRNA injection procedure.

ELISA kits of HMGB1 (cat. no. H6-NBP1-42888), interleukin (IL)-6 (cat. no. NBP1-92668), tumor necrosis factor (TNF)-α (cat. no. NBP2-31085), IL-1β (cat. no. KA0357) and IL-10 (cat. no. KA1038) were purchased from Novus Biologicals, LLC, Littleton, CO, USA). Toll-like receptor (TLR)2 (cat. no. EK1130) ELISA kit was purchased from Wuhan Boster Biological Technology, Ltd., (Wuhan, China) TLR4 (cat. no. kt56334) and nuclear factor (NF)-κB (cat. no. kt21004) ELISA kits were purchased from MSK (Wuhan, China), which were used to detect the cytokines level in septic mice serum according to the manufacture's protocols.

OBiO Technology Corp., Ltd. (Shanghai, China) constructed the HuR overexpression (Ad-HuR), knockdown (Ad-HuR-shRNA; Table I) and HMGB1 overexpression (Ad-HMGB1) adenovirus vectors, which were verified by DNA sequencing by SprinGenBioTech (Nanjing, China).

Total RNA was extracted from 1×10⁶ HUVECs with TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.), following the manufacturer's protocol. First-strand cDNA was synthesized from the total RNA using a QuantiTect Rev. Transcription kit (cat. no. 205311; Qiagen China Co., Ltd., Shanghai, China). qRT-PCR was performed on the StepOnePlus PCR system with SYBR Green kit (Biomics Biotechnologies Co., Ltd., Nantong, China). β-actin was used as an internal control and gene expression levels were normalized to β-actin, and qRT-PCR reactions were performed at 58°C for 35 cycles in triplicate for each experiment. Primer sequences are listed in Table I. The relative gene expression levels were analyzed via 2^−ΔΔCq method (9).

HuR expression level was upregulated or downregulated by infecting with Ad-HuR or Ad-HuR-shRNA at 1×10⁶ multiplicity of infection and 70–80% of cell density, respectively, for at 37°C with DMEM medium for 48 h. Subsequently, de novo RNA synthesis was blocked with the addition of 5 µg/ml actinomycin D (ActD; Apexbio, Houston, TX USA) to the medium. Total RNA was harvested at the 0, 2, 4, 6, 8 and 10 h post-transfection, and HMGB1 mRNA expression levels were measured by RT-qPCR assay. mRNA half-life was examined by comparing with the mRNA expression level prior to the addition of ActD.

RIP assays were conducted as previously described (10). Briefly, HUVECs (1×10⁷) were lysed with 25 mM Tris-HCl buffer (pH 7.5) and 100 U/ml RNase inhibitor (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany), and subsequently incubated with protein-A sepharose beads (P001, 7Sea Biotech, Shanghai, China) precoated with 3 µg anti-HuR antibody or control rabbit immunoglobulin G for 1.5 h at 4°C. The RNA-protein complexes were pulled-down by protein A/G agarose beads, RNA was extracted with TRIzol and HMGB1 expression levels were detected with RT-qPCR.

The HMGB1 promoter sequence was inserted into the pGL3 vector (pGL3-HMGB1; Promega Corporation, Madison, WI, USA), which was transfected in HUVECs with Ad-HuR or Ad-HuR-shRNA co-infection. Transfection efficiency was normalized with β-gal activity, and the luciferase activity of pGL3-HMGB1 was measured for HMGB1 promoter transcriptional activity assays. Cells at a density of 80% were used for transfection with Lipofectamine^® 2000 (Thermo Fisher Scientific, Inc) at 37°C with DMEM medium for 72 h. The luciferase activity was measured by a Glomax 96 luminometer (Promega Corporation).

Cell Counting Kit-8 (CCK-8, Vazyme Biotech Co., Ltd., Nanjing, China) was used to examine cell viability. Cells infected with Ad-HuR-shRNA or Ad-empty vector were seeded (4,000 cells/well) in 96-well plates and incubated for 24, 48 and 72 h at 37°C; growth rates were determined by measuring optical density with a microplate reader. Each experiment was performed in triplicate.

HUVECs were seeded (8×10⁴ cells/well) onto Transwell inserts (Millipore; Merck KGaA) for 6 h and subsequently treated with HMGB1 cytokine (1 µg/ml; cat. no. 34-8401-82; Thermo Fisher Scientific, Inc.) for 16 h followed by infection with or without Ad-HuR-shRNA for 24 h. Transwell inserts were then washed with PBS, followed by addition of 0.5 ml of Evans blue (0.67 mg/ml) diluted in growth medium containing 5% bovine serum albumin (BSA). Fresh growth medium was then added to the lower chamber, and the medium in the upper chamber was replaced with Evans blue/BSA. Following 10 min incubation, optical density was measured at 650 nm in the lower chamber.

CLP-operated 12 mice or control 12 mice were pretreated with HMGB1 (2 µg) intravenously for 16 h, after which Ad-HuR-siRNA was injected intravenously. A total of 12 mice treated with a scrambled siRNA were used as a control group. After 48 h, 1% Evans blue dye solution in normal saline was injected intravenously into each mouse, and the mice were sacrificed 30 min post-injection. The peritoneal exudates were collected following washes with 5 ml normal saline and centrifuged at 200 × g for 10 min at 4°C. The absorbance of the supernatant was read at 650 nm via a 96-well plate by GloMax^® Discover System (GM3000; Promega Corporation). Vascular permeability was expressed in terms of dye (µg)/mouse. To assess leukocyte and neutrophil migration, 20 µl of peritoneal fluid was mixed with 0.38 ml of Turk's solution (0.01% crystal violet in 3% acetic acid), and the leukocyte number was counted under a light microscope. The results were expressed as neutrophil ×10⁶ per peritoneal cavity.

All data are presented as mean ± standard deviation. The differences between the groups were analyzed with the Student's t-test or one-way analysis of variance of three independent experiments. The differences between the groups were analyzed using a one-way analysis of variance with the Tukey-Kramer post-hoc test. P<0.05 was considered to indicate a statistically significant difference.

The present study identified that HuR has 21 potential binding sites on HMGB1 in 11 different cancers and the binding ability ranked first by the starBase v2.0 (http://starbase.sysu.edu.cn/browseMrnaCeRNA.php) analysis (Table II). Subsequently, the ability of HuR to increase HMGB1 expression in HUVECs was examined. Transfection with Ad-HuR or Ad-HuR-shRNA resulted in the subsequent increase or decrease, respectively, of HuR mRNA and protein expression levels in HUVECs (Fig. 1A-D). In addition, overexpression of HuR significantly upregulated HMGB1 mRNA and protein expression levels (Fig. 1E and F, respectively). By contrast, HuR-knockdown led to a significant decrease in HMGB1 mRNA and protein expression (Fig. 1G and H, respectively). These results indicated that HuR was able to positively regulate HMGB1 expression.

Whether HuR was able to regulate HMGB1 mRNA stability in HUVECs was examined. HuR expression was induced or knocked down followed by the inhibition of de novo mRNA synthesis by ActD treatment. The decay rate of HMGB1 mRNA was faster in Ad-HuR-shRNA infected cells (t_1/2=3.8±0.4 h) compared with control cells (t_1/2=5.4±0.3 h; Fig. 2A), and slower in the Ad-HuR-treated group (t_1/2=7.9± 0.2 h) compared with the control cells (t_1/2=5.3± 0.4 h; Fig. 2B). However, the luciferase reporter assay demonstrated that the promoter transcriptional activity of HMGB1 was not affected by Ad-HuR or Ad-HuR-shRNA infection in HUVECs (Fig. 2C), which indicate that HuR could not regulate the transcriptional activity of HMGB1. Furthermore, RIP assays demonstrated that HuR interacted with HMGB1 more specifically than the control group (Fig. 2D). Our results demonstrate that HuR may interact with HMGB1 and enhances HMGB1 mRNA stability in HUVECs.

Previous studies have shown that LPS induced HMGB1 release in human endothelial cells (11,12). The release of HMGB1 was stimulated with LPS (100 ng/ml) treatment in HUVECs (Fig. 3A). HUVECs were transfected with Ad-HuR-shRNA and subsequently stimulated with LPS for 24 h. Knockdown of HuR with Ad-HuR-shRNA attenuated the inductive effects of LPS on HMGB1 release (Fig. 3B). However, HUVECs treated with Ad-HuR-shRNA alone did not affect the release of HMGB1 at 24 h, but decreased HMGB1 release at 48 and 72 h (Fig. 3C). As CLP-induced sepsis in mice closely resembles human sepsis (13), mice were subjected to sepsis by CLP treatment. Notably, injection of HuR-siRNA inhibited the CLP-mediated release of HMGB1 (Fig. 3D).

In addition, the effects of HuR on HMGB1 receptors TLR2 and TLR4 were detected in HUVECs. The results indicated that infection with Ad-HMGB1 significantly induced TLR2 and TLR4 protein expression, which was attenuated by co-infection with Ad-HuR-shRNA (Fig. 3E). To exclude the possibility that the decline of HMGB1 release was due to cytotoxicity caused by Ad-HuR-shRNA infection, cell viability was examined following Ad-HuR-shRNA infection by CCK-8 assay. Ad-HuR-shRNA did not affect cell viability at 24, 48 or 72 h post-transfection (Fig. 3F). Therefore, these results suggested that HuR-knockdown may attenuate the LPS- and CLP-mediated HMGB1 release.

The effects of HuR on the barrier integrity of HUVECs were also examined. No significant difference was identified in barrier integrity in cells transfected with Ad-HuR-shRNA alone compared with control transfected cells (Fig. 4A). As endothelial membrane barriers were previously demonstrated to be disrupted by LPS induction (14), HUVECs were transfected with Ad-HuR-shRNA for 24 h following LPS treatment. Knockdown of HuR attenuated the LPS-mediated membrane disruption (Fig. 4B). In addition, the barrier integrity has been reported to be perturbed by HMGB1 induction (15). Therefore, HUVECs were transfected with HMGB1 followed by Ad-HuR-shRNA co-transfection. Cells co-transfected with Ad-HuR-shRNA exhibited an decrease in HMGB1-mediated membrane disruption (Fig. 4C). In addition, these effects were investigated in vivo by examining HMGB1- or CLP-induced vascular permeability in mice. Treatment with HuR-shRNA suppressed HMGB1- or CLP-induced peritoneal leakage of dye (Fig. 4D and E, respectively). These results demonstrated that the knockdown of HuR expression may reduce HMGB1-mediated endothelial disruption and maintain the human endothelial cell barrier integrity.

A previous study reported that the inflammatory effects mediated by HMGB1 were caused by increasing the cell surface expression of intracellular adhesion molecule (ICAM)-1, vascular (V)CAM-1 and E-selectin, which promote leukocyte adhesion and migration from the endothelium to inflammation sites and hence of the proinflammatory response (16). Transfection with Ad-HuR-shRNA suppressed the mRNA expression levels of ICAM-1, VCAM-1 and E-selectin (Fig. 5A). The binding ability of leukocytes to HMGB1-stimulated HUVECs was significantly inhibited by HuR knockdown (Fig. 5B). In addition, HuR-knockdown inhibited the association of leukocytes to HUVEC binding with the subsequent transendothelial migration of leukocytes (Fig. 5C). Notably, mice injected with HuR-siRNA exhibited an inhibition of HMGB1- or CLP-induced migration of neutrophils (Fig. 5D and E). The results demonstrated that knockdown of HuR inhibits HMGB1-induced proinflammatory signaling and thereby downregulates the activation of inflammatory pathways, such as the upregulation of CAMs, leukocyte adhesion and migration induced by HMGB1.

To investigate the potential effects of HuR-knockdown on the induction of inflammatory signaling molecules, the production levels of TNF-α and IL-1β were detected in HMGB1-activated HUVECs. The production of TNF-α and IL-1β was increased by HMGB1 treatment (Fig. 6A and B, respectively), which was attenuated by HuR knockdown with Ad-HuR-shRNA. In addition, the effects of HuR knockdown on the NF-κB pathway were evaluated. Knockdown of HuR inhibited the promotive effects of HMGB1-mediated NF-κB activity (Fig. 6C). IL-10 is an anti-inflammatory cytokine (17); the production of IL-10 was detected in HMGB1-activated HUVECs. The production of IL-10 was decreased by HMGB1 treatment (Fig. 6D), and this effect was reversed by HuR knockdown. These results indicated that HuR may regulate the NF-κB pathway that is involved in the induction of proinflammatory responses in human endothelial cells.

Sepsis has a poor prognosis, which is often characterized as the production of IL-6 (18,19). Therefore, the present study hypothesized that treatment with HuR-siRNA would be able to reduce the production of IL-6 and mortality rates in HMGB1- or CLP-induced sepsis mouse model. Production of IL-6 in HMGB1- or CLP-induced mice was reduced by HuR-siRNA injection (Fig. 7A and B, respectively). In addition, HuR-siRNA injection resulted in an increase in the survival rate, analyzed by Kaplan-Meier survival plot (Fig. 7C), which suggests that HuR might be a potential therapeutic target for sepsis.