Urolithin A suppresses high glucose-induced neuronal amyloidogenesis by modulating TGM2-dependent ER-mitochondria contacts and calcium homeostasis

Abstract

Hyperglycemia in diabetes mellitus (DM) patients is a causative factor for amyloidogenesis and induces neuropathological changes, such as impaired neuronal integrity, neurodegeneration, and cognitive impairment. Regulation of mitochondrial calcium influx from the endoplasmic reticulum (ER) is considered a promising strategy for the prevention of mitochondrial ROS (mtROS) accumulation that occurs in the Alzheimer's disease (AD)-associated pathogenesis in DM patients. Among the metabolites of ellagitannins that are produced in the gut microbiome, urolithin A has received an increasing amount of attention as a novel candidate with anti-oxidative and neuroprotective effects in AD. Here, we investigated the effect of urolithin A on high glucose-induced amyloidogenesis caused by mitochondrial calcium dysregulation and mtROS accumulation resulting in neuronal degeneration. We also identified the mechanism related to mitochondria-associated ER membrane (MAM) formation. We found that urolithin A-lowered mitochondrial calcium influx significantly alleviated high glucose-induced mtROS accumulation and expression of amyloid beta (Aβ)-producing enzymes, such as amyloid precursor protein (APP) and β-secretase-1 (BACE1), as well as Aβ production. Urolithin A injections in a streptozotocin (STZ)-induced diabetic mouse model alleviated APP and BACE1 expressions, Tau phosphorylation, Aβ deposition, and cognitive impairment. In addition, high glucose stimulated MAM formation and transglutaminase type 2 (TGM2) expression. We first discovered that urolithin A significantly reduced high glucose-induced TGM2 expression. In addition, disruption of the AIP-AhR complex was involved in urolithin A-mediated suppression of high glucose-induced TGM2 expression. Markedly, TGM2 silencing inhibited inositol 1, 4, 5-trisphosphate receptor type 1 (IP3R1)-voltage-dependent anion-selective channel protein 1 (VDAC1) interactions and prevented high glucose-induced mitochondrial calcium influx and mtROS accumulation. We also found that urolithin A or TGM2 silencing prevented Aβ-induced mitochondrial calcium influx, mtROS accumulation, Tau phosphorylation, and cell death in neuronal cells. In conclusion, we suggest that urolithin A is a promising candidate for the development of therapies to prevent DM-associated AD pathogenesis by reducing TGM2-dependent MAM formation and maintaining mitochondrial calcium and ROS homeostasis.

Fig. 1. Effect of urolithins on mitochondrial calcium, ROS in neuronal cells under high glucose condition.

a SH-SY5Y cells were treated with 100 nM of ellagic acid, urolithin A, B, C, and D for 30 min prior to high glucose exposure (25 mM of D-glucose) for 72 h. LDH released from apoptotic cells in supernatant was analyzed with LDH assay kit, n = 5. b Cells were pretreated with ellagic acid (100 nM) or urolithin A (100 nM) or urolithin B (100 nM) or urolithin C (100 nM) or urolithin D (100 nM) for 30 min prior to high glucose (25 mM) treatment for 48 h. Antimycin A (10 μM) was used for inducing mtROS accumulation to make a positive control for analysis. Cells stained with MitoSOX (1 μM) for 20 min and MitoSOX-positive cells were analyzed with flow cytometer, n = 4. c Cells treated with high glucose at the timepoint (0–72 h). Mitochondrial calcium levels were measured by staining the cells with rhod-2 fluorescent dye (2 μM) for 20 min, and rhod-2-positive cells were analyzed with flow cytometer, n = 4. Unstained cells were used as a negative control. d, e Cells were exposed with urolithin A (100 nM) for 30 min prior to high glucose for 48 h. d rhod-2-positive cells are shown as percentages of the gated cells, n = 4. e Mitochondrial permeability transition pore (mPTP) assay was performed and analyzed by flow cytometer. Cells were pretreated with urolithin A (100 nM) or cyclosporin A (2 mM) for 30 min prior to high glucose (25 mM) treatment for 48 h. Cyclosporin A was used as a desensitization control, n = 6. f, g Cells were pretreated with Ru360 (1 μM) for 30 min prior to high glucose for 48 h. f DCF-DA staining was performed and the cells stained with DCF-DA was analyzed with flow cytometer, n = 4. g MitoSOX staining was achieved and the MitoSOX-positive cells were counted by using flow cytometer. Antimycin A (10 μM) was used for inducing mtROS accumulation to make a positive control for analysis, n = 4. Quantitative data are shown as a mean ± S.E.M. All flow cytometer figures are representative. *p < 0.05.

Fig. 2. Effect of urolithin A on amyloidogenesis in neuronal cells under high glucose condition.

a SH-SY5Y cells were treated with 100 nM of ellagic acid, urolithin A, B, C, and D for 30 min prior to high glucose exposure (25 mM of D-glucose) for 48 h. Amyloid beta (Aβ) concentration in culture medium was calculated by using high-sensitive Aβ ELISA, n = 4. b iPSC-ND cells were treated with urolithin A (100 nM) for 30 min prior to high glucose exposure for 48 h. Aβ concentration in media supernatant was determined by high-sensitive Aβ ELISA kit. Aβ concentration levels of each group were presented as percentage of control, n = 5. c, d APP and BACE1 protein expressions were detected by western blot analysis. c SH-SY5Y and d iPSC-ND cells were exposed to urolithin A for 30 min prior to high glucose treatment for 24 h, n = 4–5. e, f SH-SY5Y cells were treated with Ru360 (1 μM) or MitoTEMPO™ (1 μM) for 30 min prior to high glucose treatment for 24 h. APP and BACE1 protein expressions were detected by western blot analysis. β-Actin was used as a loading control, n = 4–5. g–i Normal and STZ-induced diabetic mice were delivered vehicle (0.5 mM NaOH) or urolithin A solution (2.5 mg/kg) via intraperitoneal injection for 8 weeks, n = 6–10. g Body weight of the mice was measured at the timepoint of initial urolithin A injection (0 week) and final injection (8 weeks). h Blood glucose levels of the mice were evaluated in each experimental group. i The memory function of each group of mice was assessed with spontaneous alternation test (Y-maze). The percentage of spontaneous alternation of entering the other side of the arm and total duration time was analyzed. j, k Brain tissue of mice were divided into the two part as prefrontal cortex (left) and hippocampus (right) at the termination of experiment. j The expression levels of APP, BACE1, and p-Tau (S262 and S396) were analyzed by western blot analysis, n = 4. k Aβ concentration levels of the mice brain tissue were measured with high-sensitive Aβ ELISA kit. Left panel is the result from the prefrontal cortex tissue, and right panel is the result from the hippocampus, n = 5. Quantitative data are shown as a mean ± S.E.M. *p < 0.05.

Fig. 3. Effect of urolithin A in MAM-regulated mitochondrial calcium under high glucose condition.

a The mRNA expression levels of VDAC1, MCU1, MICU1, MICU2, MCUR1, and MCUB were analyzed in the SH-SY5Y cells treated with vehicle or high glucose for 24 h by using qPCR, n = 4. b–i SH-SY5Y cells pretreated with urolithin A (100 nM) for 30 min prior to high glucose exposure for 24 h. b The mRNA expression levels of VDAC1 and MCU1 was quantified by qPCR, n = 3. c Protein expression levels were analyzed by western blot by using anti-VDAC1 and anti-MCU1 antibodies, n = 4. d Physical association of mitochondria and ER was visualized by the staining of MitoTracker (green), ER-tracker (red), and Hoechst 33342 (blue). Merged images are shown and representative. n = 5. Magnification ×1,000. Scale bars are 8 μm. e Co-immunoprecipitation of IP3R1 and IP3R3 with IgG or VDAC1 antibodies are shown in left panel. Total protein expression of IP3R1, IP3R3, VDAC1 and β-Actin are shown in right panel, n = 4. f Interaction between VDAC1 and IP3R1 (VDAC1-IP3R1, red) in SH-SY5Y cells was assessed by PLA assay. n = 6. Magnification ×1,000. Scale bars are 8 μm. g The mRNA expressions of BAX, BCL2L1, BCL2, GRP75, and TGM2 in SH-SY5Y cells were analyzed by qPCR. Normalization was achieved by 18s rRNA expression levels, n = 5. h TGM2 protein expression in SH-SY5Y cells was analyzed by western blot in cells treated urolithin A for 30 min prior to high glucose exposure for 24 h, n = 4. i TGM2 protein expression was analyzed in iPSC-ND treated with urolithin A (100 nM) for 30 min prior to high glucose exposure for 24 h, n = 4. j TGM2 activity was evaluated in the SH-SY5Y cells treated urolithin A (100 nM) for 30 min prior to high glucose exposure for 24 h with TGM2 activity assay kit, n = 5. k Prefrontal cortex and hippocampus tissues were dissected from the STZ-induced diabetic mice treated with vehicle or urolithin A. TGM2 expression was analyzed by western blot analysis, n = 4. Quantitative data are shown as a mean ± S.E.M. *p < 0.05.

Fig. 4. Role of TGM2 in high glucose-induced MAM formation and amyloidogenesis.

a–f SH-SY5Y cells were transfected with TGM2 siRNA (25 nM) or NT siRNA (25 nM) prior to high glucose exposure for 24 h. a Physical association of mitochondria and ER was visualized by the staining of MitoTracker (green), ER-tracker (red), and Hoechst 33342 (blue). Merged images are shown and representative, n = 5. Magnification ×1,000. Scale bars are 8 μm. b Co-immunoprecipitation of IP3R1 and TGM2 with anti-IgG or -VDAC1 antibodies are shown in left panel. Total protein expression of IP3R1, VDAC1, TGM2, and β-Actin are shown in right panel, n = 4. c Interaction between VDAC1 and IP3R1 (VDAC1-IP3R1, red) in SH-SY5Y cells was assessed by PLA assay, n = 5. Magnification ×1,000. Scale bars are 8 μm. d Mitochondrial calcium levels were measured by staining the cells with rhod-2 fluorescent dye (2 μM) for 20 min, and rhod-2-positive cells were analyzed with flow cytometer. Unstained cells were used as a negative control, n = 6. e The mtROS levels were evaluated with MitoSOX fluorescent dye (1 μM) for 20 min MitoSOX-positive cells were analyzed with flow cytometer. Antimycin A (10 μM) was used for inducing mtROS accumulation to make a positive control for analysis. All flow cytometer figures are representative, n = 4. f APP, BACE1, and TGM2 protein expressions in SH-SY5Y cells were analyzed by western blot, n = 5. g Aβ concentration levels of the cell culture media originated from the cells transfected with TGM2 siRNA or NT siRNA prior to high glucose exposure for 48 h were measured with high-sensitive Aβ ELISA kit, n = 5. Quantitative data are shown as a mean ± S.E.M. *p < 0.05.

Fig. 5. Effect of urolithin A on inhibition of TGM2 expression via formation of AIP–AhR transcriptomal complex.

a SH-SY5Y cells were treated with high glucose for 0, 24, and 48 h. AhR and AIP expressions were analyzed by western blot analysis, n = 4. b Cells treated with urolithin A (100 nM) for 30 min prior to high glucose (25 mM) exposure for 24 h. Subcellular fractions of cytosol and nucleus were analyzed with western blot analysis. AhR expression level was analyzed with anti-AhR antibody. The anti-α-tubulin (cytosol) and -Lamin A/C (nucleus) antibodies were used for confirmation of fractionization, n = 4. c Nuclear localization of AhR was confirmed in the cells treated with urolithin A (100 nM) for 30 min prior to high glucose (25 mM) exposure for 24 h. Cells stained with AhR (green) antibody and DAPI (blue) were visualized and merged images are shown and representative, n = 5. Magnification ×1,000. Scale bars are 8 μm. d, e Cells treated with CH-223191 (10 μM) for 30 min prior to high glucose exposure for 24 h. d TGM2 mRNA expression level was analyzed by using qPCR, n = 5. e TGM2 protein expression level was analyzed by western blot. β-Actin was used as loading control, n = 4. f, g Cells treated with urolithin A (100 nM) for 30 min prior to high glucose (25 mM) exposure for 24 h. f Co-immunoprecipitation of AIP with anti-IgG or -AhR antibodies are shown in left panel. Total protein expression of AIP, AhR, and β-Actin are shown in right panel, n = 4. g Interaction between AIP and AhR (AIP–AhR, red) in SH-SY5Y cells was assessed by PLA assay, n = 5. Magnification ×1,000. Scale bars are 8 μm. Quantitative data are shown as a mean ± S.E.M. *p < 0.05.

Fig. 6. Protective effects of urolithin A on Aβ-induced mitochondrial calcium influx, mtROS accumulation, and neuronal cell death.

a LDH release assay was performed in the SH-SY5Y cells treated with Amyloid beta (Aβ) (1–42) for 24, 48, and 72 h, n = 6. b SH-SY5Y cells treated with urolithin A (100 nM) for 30 min prior to high glucose exposure for 24 h. Mitochondrial calcium levels were measured by staining the cells with rhod-2 fluorescent dye (2 μM) for 20 min, and rhod-2-positive cells were analyzed with flow cytometer. Unstained cells were used as a negative control and ionomycin (2 μM) treated cells were used for a positive control, n = 4. c iPSC-ND treated with urolithin A (100 nM) for 30 min prior to high glucose exposure for 24 h. rhod-2 stained cells were measured by flow cytometer, n = 4. d, e MitoSOX-positive SH-SY5Y and iPSC-ND cells were analyzed with flow cytometer. Antimycin A (10 μM) was used for inducing mtROS accumulation to make a positive control for analysis, n = 4. All flow cytometer figures are representative. f LDH release assay was performed in the SH-SY5Y cells treated with urolithin A for 30 min prior to Aβ (1–42) for 72 h, n = 5. Quantitative data are shown as a mean ± S.E.M. *p < 0.05.

Fig. 7. Role of TGM2 in Aβ-induced mitochondrial calcium overload, mtROS accumulation, tau phosphorylation, and neuronal cell death.

a–d SH-SY5Y cells were transfected with TGM2 siRNA or NT siRNA prior to Aβ exposure for 48 h. APPSwe cells were transfected with TGM2 siRNA or NT siRNA and then cells were treated urolithin A for 48 h. a, b rhod-2-positive cells were analyzed by flow cytometer, n = 4. c, d MitoSOX-positive cells were quantified by flow cytometer, n = 4–5. e Tau phosphorylation levels at S262 and S396 were investigated in the SH-SY5Y cells transfected with TGM2 siRNA or NT siRNA prior to Aβ exposure for 24 h, n = 3. f Tau phosphorylation levels at S262 and S396 in the APPSwe cells transfected with TGM2 siRNA or NT siRNA prior to exposure urolithin A for 24 h, n = 3. g LDH release in the cells transfected with TGM2 siRNA or NT siRNA prior to Aβ (1–42) for 72 h was investigated, n = 6. Quantitative data are shown as a mean ± S.E.M. *p < 0.05.

Fig. 8. The schematic model for action mechanism of urolithin A on high glucose-induced neuronal amyloidogenesis by modulating TGM2-dependent ER-mitochondria contacts and calcium influx.

High glucose induces neuronal degeneration and amyloidogenesis via mitochondrial calcium influx and mtROS accumulation. High glucose disrupts AhR–AIP complex and MAM formation. Urolithin A inhibits AhR nuclear translocation under high glucose exposure. Urolithin A reduces high glucose-induced AhR-mediated transcription of TGM2, which is critical for IP3R1–VDAC1 interactions. Conclusively, urolithin A-mediated suppression of TGM2 expression prevents the neuronal degeneration and Aβ production under high glucose conditions.