MD2 activation by direct AGE interaction drives inflammatory diabetic cardiomyopathy

Hyperglycemia activates toll-like receptor 4 (TLR4) to induce inflammation in diabetic cardiomyopathy (DCM). However, the mechanisms of TLR4 activation remain unclear. Here we examine the role of myeloid differentiation 2 (MD2), a co-receptor of TLR4, in high glucose (HG)- and diabetes-induced inflammatory cardiomyopathy. We show increased MD2 in heart tissues of diabetic mice and serum of human diabetic subjects. MD2 deficiency in mice inhibits TLR4 pathway activation, which correlates with reduced myocardial remodeling and improved cardiac function. Mechanistically, we show that HG induces extracellular advanced glycation end products (AGEs), which bind directly to MD2, leading to formation of AGEs-MD2-TLR4 complex and initiation of pro-inflammatory pathways. We further detect elevated AGE-MD2 complexes in heart tissues and serum of diabetic mice and human subjects with DCM. In summary, we uncover a new mechanism of HG-induced inflammatory responses and myocardial injury, in which AGE products directly bind MD2 to drive inflammatory DCM.

Supplementary Figure 8: MD2 inhibitor L6H21 protects against cardiac inflammatory injury in streptozotocin-induced diabetic mice. Male C57BL/6 mice were made diabetic by single intraperitoneal injection of streptozotocin (STZ). Non-diabetic controls received citrate buffer only. Upon confirmation of hyperglycemia, diabetic mice were divided into two groups: STZ mice and STZ mice treated with small molecule inhibitor of MD2, L6H21. L6H21 was administered at 20 mg/kg/2d for 16 weeks. Untreated STZ mice received 1% CMC-Na vehicle [Con = non-diabetic controls, n=5; STZ = streptozotocin-induced diabetic mice, n=8; STZ+L6H21 = diabetic mice treated with L6H21, n=8]. (a, b) Body weights (a) and fasting blood glucose levels (b) were measured weekly for 16 weeks (Values are reported as Mean ± SEM). . Primary cardiomyocytes, endothelial cells, and fibroblasts were isolated from heart tissues of C57BL/6 mice. Mouse peritoneal macrophages (MPMs) were also analyzed. Heart tissue RNA and protein lysates were used as control. mRNA levels were normalized to Actb [Mean ± SEM; n =4 examinations]. GAPDH was also used as loading control for immunoblotting. P-values by one-way ANOVA in a followed by Tukey's post hoc test are indicated.
Supplementary Figure 10: Bone marrow reconstitution reveals important contribution of cardiac cell-and macrophage-derived MD2 in diabetes-induced cardiac injury. Md2-/-mice (MD2KO) were given acidified water containing neomycin and polymyxin B sulphate. One day prior to transplantation, mice were subjected to total body irradiation (6 Gy). Bone marrow cells prepared from wildtype (WT) or MD2KO mice were prepared and injected intravenously at 5 x10 6 cells. Tail clip samples and peritoneal macrophages were used for genotyping and confirmation of reconstitution. Mice were then made diabetic by streptozotocin (STZ). Heart tissues were harvested 12 weeks following the onset of diabetes. Supplementary Figure 15: L6H21 inhibits HG-induced inflammatory signaling in rat primary cardiomyocytes. Primary cardiomyocytes isolated from Sprague Dawley rats were pretreated with L6H21 or DMSO control (Ctrl) for 1 h and then exposed to HG (33 mM glucose) for different time points. (a) Representative immunoblot showing MD2-TLR4 complex formation in primary cardiomyocytes in response to HG. Proteins were immunoprecipitated using MD2 antibody (IP) and TLR4 (IB) was detected. L6H21 pretreatment was carried out at 10 μM. Cells were then exposed to HG for 5 min [n=3 examinations]. (b, c) Representative Western blot analysis of IκB and phosphorylated mitogen-activated protein kinase proteins (p-ERK, p-JNK, and p-P38). L6H21 pretreatment was carried out at 10 μM. Cells were then exposed to HG for 30 min. (a) H9C2 cells were transfected with Md2 siRNA or negative control sequences (NC). Transfected cells were exposed to HG (33 mM glucose) for 24 h and RNA was isolated. Levels of Tgfb1 and Mmp2 were determined by qPCR Mean ± SEM; n = 3 examinations]. (b) H9C2 cells were pretreated with 10 μM L6H21 for 1 h and then exposed to HG (33 mM glucose) for 24 h. Cells were then stained for TGFβ1 (red), collagen I (red), and myosin heavy chain (MyHC) (red). DAPI (blue) was used to counterstain [n=3]. (c) Rat primary cardiomyocytes were pretreated with L6H21 at 2.5, 5.0, or 10 µM for 1 h and then exposed to HG (33 mM glucose) for 24 h. Protein levels of TGFβ1, MMP2, and collagen 1 were detected by immunoblotting. GAPDH was used as loading control [n=3 examinations]. P-values by one-way ANOVA in a followed by Tukey's post hoc test are indicated.
Reviewer Figure 17: Intracellular HG increases the expression of TLR4 via ROS but is not involved in MD2-TLR4 complex formation and TLR4 activation. (a) MD2-TLR4 interaction in mouse primary macrophages (MPM) exposed to HG with or without ROS scavenger N-acetyl cysteine (NAC; 5 mM) pretreatment. Cells were exposed to 33 mM HG for 15 minutes following a 1-hour pretreatment with 5 mM NAC. MD2 was immunoprecipitated (IP) and interaction with TLR4 was determined by immunoblotting (IB). Representative blots are from 3 independent experiments. (b) MD2-TLR4 interaction in H9C2 cells. Cells were treated as indicated for Panel a. Representative blots are from 3 independent experiments. (c) Mouse primary macrophages were exposed to 33 mM glucose (HG) for 12 hours, with or without 1h pretreatment with NAC (5 mM). mRNA levels of Tnfa and Il6 were measured by real-time qPCR [mRNA data normalized to Actb; Mean ± SEM; n=3 examinations]. (d) Tnfa and Il6 mRNA levels in H9C2 cells. Cells were treated as indicated for Panel c [Mean ± SEM; n=3 examinations]. (e) Tlr4 expression in mouse primary macrophages and H9C2 cells following exposure to HG. Cells were treated as indicated for Panel c [mRNA data normalized to Actb; Mean ± SEM; n=3 examinations]. P-values by one-way ANOVA in c, d, e followed by Tukey's post hoc test are indicated.
Supplementary Figure 18: HG-induced cytokine production is dependent on the presence of serum. (a, b) H9C2 cells were exposed to HG (33 mM glucose) for 6 h, in the presence or absence of heat-inactivated serum. Levels of Il6 and Tnfa mRNA were determined [Mean ± SEM; n=3 examinations]. (c, d) H9C2 cells were exposed to 0.5 μg/mL lipopolysaccharide (LPS) for 6 h, in presence or absence of heat-inactivated serum. Il6 and Tnfa mRNA levels were then determined [Mean ± SEM; n=3 examinations]. P-values by one-way ANOVA in a, b, c, d followed by Tukey's post hoc test are indicated.
Supplementary Figure 19: Effect of heat-inactivation on HG-induced inflammatory cytokine production in H9C2 cells. (a, b) H9C2 cells were cultured in media containing 10% heatinactivated fetal bovine serum or serum that did not undergo heat-inactivation for 10 days. Cells were then exposed to HG (33mM glucose) for 24 h. mRNA levels of Tnfa (a) and Il6 (b) were determined [Mean ± SEM; n = 3 examinations]. P-values by one-way ANOVA in a, b followed by Tukey's post hoc test are indicated.
Supplementary Figure 20: Effect of high glucose on glycolytic function of cardiomyocytes and macrophages. Glycolytic function of cells was determined by Agilent Seahorse XF Glycolysis Stress Test Kit. Extracellular acidification rate (ECAR; mpH/min/µg protein) in H9C2 (a, 88000 cells per group) and MPMs (b, 22000 cells per group) exposed to HG (33 mM glucose). Sequential addition of glucose (Glu) and 2-deoxy-glucose (2-DG) is indicated by arrows. No changes in glycosis was observed in H9C2 cells exposed to HG compared to cells cultured in 5 mM glucose. Glycosis was increased in MPMs exposed to HG [n=3 repeats in Con and HG groups; n=4 repeats in DG+HG group; Values are reported as Mean ± SD].
Supplementary Figure 21: Effect of glycolysis inhibitor on HG-induced MD2-TLR4 complex formation and inflammation. (a) H9C2 cells were pretreated with 50 mM 2-DG for 30 min and then exposed to HG (33mM glucose) for 15 min. Proteins were immunoprecipitated with TLR4 antibody (IP) and levels of MD2 were determined by immunoblotting (IB) [n = 3 examinations]. (b, c) H9C2 cells were pretreated with 50 mM 2-DG for 30 min and then exposed to HG (33mM glucose) for 12h. mRNA levels of Tnfa (b) and Il6 (c) were determined by real-time qPCR assay. Glycolysis inhibitor 2-DG dose not suppress HG-increased mRNA levels of Tnfa and Il6 in H9C2 cells. In contrast, glycolysis inhibitor 2-DG aggravates HG-induced Tnfa and Il6 expression, possibly due to the decreased glucose consuming and increased extracellular glucose concentration.
[Mean ± SEM; n=3 examinations]. P-values by one-way ANOVA in b, c followed by Tukey's post hoc test are indicated.
Supplementary Figure 22: Levels of methylglyoxal and Nε-carboxymethyl-lysine in H9C2 cells exposed to high levels of glucose. (a, b) H9C2 cells were exposed to 33 mM glucose (HG) for different time periods, either in the absence (a) or presence of 10% heat-inactivated serum (b). Levels of methylglyoxal (MGO) were measured by two-photon fluorophore conjugated to ophenylenediamine which contains a MGO recognition site [n=3 examinations; Values are reported as Mean ± SEM]. (c) H9C2 cells were exposed to 33 mM glucose for 15 minutes in the presence of 10% heat-inactivated serum and levels of CML were measured by ELISA [n=3; Values are reported as Mean ± SEM]. (d) HG incubation in cell-free medium do not produce MGO and CML. Complete growth media containing serum was incubated with 33 mM glucose (HG) for 15 minutes. Levels of AGE-, MGO-, and CML-modified proteins were measured by the methods described in Methods section, respectively [n=3 examinations; Values are reported as Mean ± SEM]. P-values by unpaired t test are indicated in d.
Supplementary Figure 23: AGE-BSA but not MGO-BSA or CML-BSA activates MD2/TLR4 inflammation. (a) Mouse primary macrophages were exposed to 30 μg/mL AGE-, MGO-, or CML-modified BSA for 15 minutes. Proteins were isolated, immunoprecipitated by MD2 antibody, and interaction between MD2-TLR4-MyD88 were assessed by immunoblotting. Cells in control group were exposed to 30 μg/mL BSA protein. Representative blots were shown from three independent experiments. (b) MD2-TLR4-MyD88 interaction in H9C2 cells. H9C2 cells were treated as indicated for Panel a. (c) Primary macrophages or H9C2 cellls were exposed to 30 μg/mL AGE-BSA, MGO-BSA, or CML-BSA for 1 hour. Cell lysates were probed for level of inhibitor of κB (IκBα). GAPDH was used as loading control. Cells in control group were exposed to 30 μg/mL BSA protein. Representative blots were shown from three independent experiments. (d) Real-time qPCR assay shows the mRNA levels of Tnfa and Il6 in primary macrophages exposed to 30 μg/mL AGE-BSA, MGO-BSA, or CML-BSA for 12 hours [Con = 30 μg/mL BSA; n=9 repeats; Values are reported as Mean ± SEM]. P-values by one-way ANOVA in d followed by Tukey's post hoc test are indicated.
Supplementary Figure 24: MD2-binding polymyxin B does not alter AGE-induced inflammatory responses in H9C2 cells. (a) H9C2 cells were pretreated with 30 μg/mL polymyxin B for 1 hour and then exposed to 30 μg/mL AGE-BSA (AGE). AGE-BSA (with or without polymyxin B pretreatment) were carried out for 15 minutes. Protein samples were immunoprecipitated using MD2 antibody (IP) and levels of TLR4 were determined by immunoblotting (IB). Representative blots are from 3 independent experiments. (b, c) H9C2 cells were exposed to 30 μg/mL AGE-BSA for 24 hours, with or without pretreatment with 30 μg/mL polymyxin B for 1 hour. Levels of Tnfa (b) and Il6 (c) mRNA were measured. Data was normalized to Actb