High‐mobility group box‐1 promotes vascular calcification in diabetic mice via endoplasmic reticulum stress

Abstract Several studies reported the role of endoplasmic reticulum stress (ERS) in vascular calcification. High‐mobility group box‐1 (HMGB‐1) plays a substantial role in diabetes and its complications. However, relatively little information is available regarding the association between HMGB‐1 and calcification, and the underlying mechanism has still remained elusive. Therefore, in the present study, we attempted to indicate whether HMGB‐1 could promote vascular calcification via ERS in diabetes. After induction of diabetes by Streptozotocin (STZ), mice were treated with glycyrrhizin (Gly) or 4‐phenylbutyrate (4‐PBA). Mineral deposition was confirmed by reverse transcription‐polymerase chain reaction (RT‐PCR) and calcium assay. In cell experiments, calcification of vascular smooth muscle cells (VSMCs) was performed with Alizarin Red staining, alkaline phosphatase (ALP) activity and RT‐PCR. Expression and location of HMGB‐1 in aortic tissue were detected by Western blotting, immunocytochemistry (ICC) and immunohistochemistry (IHC). Diabetic mice demonstrated increased HMGB‐1 expression, ERS and vascular calcification. However, inhibition of HMGB‐1 with Gly or inhibition of ERS with 4‐PBA ameliorated the enhanced vascular calcification and ERS in diabetic mice. In vitro experiments unveiled that inhibition of HMGB‐1 attenuated advanced glycation end products (AGEs)‐induced ERS in VSMCs. In addition, AGEs promoted translocation and secretion of HMGB‐1 in VSMCs, which was reversed by 4‐PBA. Moreover, VSMCs exhibited increased mineralization and osteogenic gene expressions in response to HMGB‐1 and AGEs. However, inhibition of ERS with 4‐PBA partially, although noticeably, attenuated VSMC calcification induced by HMGB‐1. Thus, diabetes induced translocation and secretion of HMGB‐1 via ERS, which resulted in calcification in diabetic mice and in AGEs‐treated VSMCs.


| INTRODUC TI ON
Cardiovascular complications have been identified as principal cause of morbidity and mortality in patients with diabetes mellitus (DM). 1 Increasing evidence demonstrates that DM is associated with coronary artery calcification (CAC), 2,3 which is an independent risk factor for cardiovascular events. 4,5 DM is deemed as a coronary artery disease (CAD) equivalent, which doubles or even triples the CAD incidence. Despite the fact that the increased prevalence of DM may result in a higher incidence of CAC, therapeutic strategies aiming to effectively prevent or reduce CAC in patients with diabetes have been still rarely reported because of the incomplete understanding of the underlying mechanisms.
High-mobility group box-1 (HMGB-1), a highly conserved nuclear protein, can be translocated into the cytoplasma and released into extracellular space under particular conditions, such as diabetes and inflammation. 6 HMGB-1 can bind to at least three distinct receptors, including toll-like receptor 4 (TLR4), toll-like receptor 2 (TLR2) and receptor for advanced glycation end products (RAGEs), thereby activating downstream signalling cascades. With regulating inflammation, fibrosis, migration, oxidative stress and apoptosis, HMGB-1 showed to play a critical role in diabetic cardiovascular diseases. 7 Advanced glycation end products (AGEs), a diverse group of highly oxidant compounds with pathogenic significance in diabetes and in several other chronic diseases, was associated with CAC in patients with DM. 8 Our previous in vitro experiments demonstrated that AGEs induced secretion of HMGB-1 from various cells. 9,10 Recent studies in patients and animal models reported association between HMGB-1 and calcific aortic valve disease (CAVD), because tissue and plasma levels of HMGB-1 were increased in patients with CAVD. 11,12 Taken together with the previous finding that HMGB-1 mediates high-glucose-induced calcification in vascular smooth muscle cells (VSMCs) of saphenous veins, 13,14 it has still remained elusive whether HMGB-1 could link diabetes with vascular calcification.
Endoplasmic reticulum (ER) is a critical organelle that responds to changes in homeostasis, including calcium homeostasis, modification and protein synthesis. When the structure and function of ER is disrupted, a series of cellular responses are made to avoid the accumulation of unfolded proteins in the ER, which is known as endoplasmic reticulum stress (ERS). Several studies have demonstrated that ERS promoted development and progression of chronic kidney disease (CKD)-induced vascular calcification. 15,16 It was previously found that HMGB-1 could promote inflammatory response via RAGE-mediated stimulation of ERS in endothelial cells. 17 Recently, HMGB-1 was shown to be involved in clostridium difficile toxin A-induced ERS in murine colon adenocarcinoma cells. 18 Furthermore, ERS induced higher cytoplasmic expression and secretion of HMGB-1 in tumour-infiltrating lymphocytes in triple-negative breast cancer. 19 Thus, the present study aimed to indicate whether HMGB-1 could promote vascular calcification via ERS in diabetic mice and VSMCs.

| Induction of diabetes
All animal protocols were approved by Animal Care and Use Committee of the Affiliated Drum Tower Hospital of Nanjing University Medical School. C57bl/6 mice were obtained from Model Animal Research Center of Nanjing University. Male mice were randomly assigned into four groups and treated as follows: (a) control group (non-diabetic mice without Streptozotocin (STZ)/ glycyrrhizin (Gly)/4-phenylbutyrate (4-PBA)), (b) STZ group (diabetic mice with STZ injection), (c) STZ + Gly group (STZ-induced diabetic mice with Gly (Sigma-Aldrich)) and (d) STZ + 4-PBA group (STZ-induced diabetic mice with 4-PBA (Sigma-Aldrich)). A model of STZ-induced diabetic mice was established as previously described. 20 Briefly, mice were injected with STZ (40 mg/kg; Sigma-Aldrich) for five consecutive days.
When mice were considered as diabetes with blood glucose level > 13.9 mmol/L, some mice were treated with Gly (10 mg/ kg/d) or 4-PBA (50 mg/kg/d) one week after injection of STZ.
Bodyweight and blood glucose level were measured every 4 weeks until the animals were killed.

| Enzyme-linked immunosorbent assay (ELISA)
Serum level of HMGB-1 was detected using ELISA according to the manufacturer's protocol (Keygen Biotech Co., Ltd.). Signaling Technology). All immunoblots were detected by an enhanced chemiluminescent reagent kit and exposed to X-ray films.

| Western blot analysis
β-actin was used as a loading control (Abcam). The band intensities were quantified using Quantity One analysis software (Bio-Rad Laboratories Inc).

| Reverse transcription-polymerase chain reaction (RT-PCR)
Total RNA was extracted using mRNA isolation kit according to the manufacturer's instructions (Takara). In RT-PCR, an RNA population was converted into cDNA by reverse transcription (RT), and then, the cDNA was amplified by PCR. The sequences of the primers used were as follows: runt-related transcription factor 2 (Runx2) (mice), CCAGGCAGGTGCTTCAGAACTG Laboratories Inc). β-actin, a housekeeping gene, was used for internal normalization. Gene expressions were analysed using the 2 -ΔΔCt method.

F I G U R E 1
Increased HMGB-1 expression in aortas of type 1 diabetic mice. A, Protocol of induction of diabetes and experimental design. B, Bodyweight was measured in non-diabetic and diabetic mice. **P < .01 vs Control group; n = 6. C, Blood glucose level was measured in non-diabetic and diabetic mice. *P < .05 vs Control group; **P < .01 vs Control group; n = 6. D, Serum HMGB-1 level in control and STZ groups. **P < .01 between two groups; n = 6. E, Expression of HMGB-1 in diabetic aortas of two groups. **P < .01 between two groups; n = 8. F, Immunohistochemistry of HMGB-1 in aortas of non-diabetic and diabetic mice. Scale bar = 50 um; n = 6. STZ: 40 mg/kg

| Immunocytochemistry (ICC)
Cells were washed with PBS and fixed with 4% paraformaldehyde (PFA). After incubation with 0.1% Triton X-100 for 2 minutes on ice, the cells were washed twice with cold PBS. Then, the cells were blocked with bovine serum albumin (BSA) and incubated with HMGB-1 at 37°C for 30 minutes. After washing with PBS, the cells were incubated with appropriate fluorescently labelled secondary antibodies (Invitrogen) at 30°C in the dark. Finally, Sigma-Aldrich) was used to stain the cell nuclear. The cells were visualized under a fluorescence microscope (Olympus).

| Immunohistochemistry
Aortic tissues were immediately fixed for 2 hours in 4% methanolfree formaldehyde. The fixed samples were then embedded into paraffin and were cut into slices. Heat-mediated antigen retrieval was performed with citrate buffer. After blocking, paraffin-embedded sections were stained overnight at 4°C with HMGB-1 antibody and then with corresponding secondary antibodies for 1 hour. Next, 3,3′-diaminobenzidine (DAB) was used as chromogen, and images were taken using a microscope.  Besides, ALP activity and protein concentrations were detected with ALP activity kit and BCA assay kit, respectively. ALP activity was evaluated with normalization to the protein.

| Statistical analysis
All experiments were replicated at least three times independently.
All data were expressed as mean ± standard error of the mean (SEM). Statistical analysis was conducted with SPSS 21.0 software (IBM). Comparing the results between two groups was carried out using the Student's t test, and multiple groups were compared by analysis of variance (ANOVA). P < .05 was considered statistically significant.

| Up-regulated expression of HMGB-1 in aortas of STZ-induced diabetic mice
One week after injection of STZ, diabetic mice were treated with Gly or 4-PBA for another 16 weeks ( Figure 1A). We observed a sustained increase in bodyweight in control group and poor body gain in STZ-induced diabetic mice ( Figure 1B). Blood glucose level was significantly higher in STZ-induced diabetic mice than non-diabetic  Figure 1C). As shown in Figure 1D, diabetic mice exhibited noticeably increased serum HMGB-1 level.

| Diabetes induced translocation and secretion of HMGB-1 via ERS in VSMCs
Herein

| Diabetes-induced ERS and vascular calcification were attenuated by Gly or 4-PBA
The next aim of the investigation was to ascertain the roles of HMGB-1 and ERS in diabetes-induced vascular calcification in diabetic mice.
Firstly, we detected the expression levels of ERS markers in mice models of diabetes treated with Gly or 4-PBA. As illustrated in Figure 3A

| HMGB-1 and tunicamycin (TM) induced calcification in VSMCs
As in vivo experiment showed that inhibition of HMGB-1 or ERS prevented diabetes-induced vascular calcification, in vitro experiments on VSMCs were performed to confirm the roles of HMGB-1  Figure 4A,B). In order to indicate whether ERS could contribute to calcification in VSMCs, TM (an ERS inducer) was added into β-GP-induced VSMCs. As depicted in Figure 4C,D, TM induced calcification in VSMCs in a dose-dependent manner, suggesting that induction of ERS promoted calcification in VSMCs.

| HMGB-1 mediated AGEs-induced ERS and calcification in VSMCs
In line with the results of in vivo experimentz, VSMCs showed increased ERS in response to AGEs challenge, which was reversed by pre-treatment with Gly ( Figure 5A-D). To confirm the effect of HMGB-1 in diabetes-induced calcification, HMGB-1 and Gly were pre-treated in AGEs-induced VSMCs. As shown in Figure 5E

| D ISCUSS I ON
In summary, the findings of the current study suggested that inhibition of HMGB-1 and ERS attenuated vascular calcification in STZinduced diabetic mice. Additionally, our finding indicated that AGEs induced translocation and secretion of HMGB-1 in VSMCs via ERS.
Eventually, we found that ERS contributed to HMGB-1-promoted osteoblastic differentiation in VSMCs. Thus, we assessed the presence of a positive association between HMGB-1 and ERS in diabetes ( Figure 7).
Epidemiological studies show that DM is widely recognized as a modern-day disease, and vascular calcification is an important complication in diabetic patients. 21,22 However, the pathogenesis of diabetic vascular calcification needs to be further elucidated.
HMGB-1, a typical damage-associated molecular pattern protein, exhibited a broad range of biological properties in diabetes and its complications, as discussed in detail in our previous review. 7 Clinical studies demonstrated that serum HMGB-1 level was increased in both of type 1 23 and type 2 24,25 diabetes mellitus, which reflected pro-inflammatory state of diabetes. In line with our previous studies, 9,10 animal experiments in the present study disclosed that serum HMGB-1 level was noticeably elevated in STZ-induced diabetic mice. A clinical investigation demonstrated that serum HMGB-1 level was associated with peripheral artery disease in DM patients, 32 whereas the underlying mechanism has still remained elusive.

ACK N OWLED G EM ENTS
The present study was financially supported by the Natural

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data used during the study appear in the submitted article.