Metformin differentially activates ER stress signaling pathways without inducing apoptosis

Abstract

Endoplasmic reticulum stress signaling (ERSS) plays an important role in the pathogenesis of diabetes and heart disease. The latter is a common comorbidity of diabetes and worsens patient outcome. Results from clinical studies suggest beneficial effects of metformin - a widely used oral drug for the treatment of type 2 diabetes - on the heart of diabetic patients with heart failure. We therefore analyzed the effect of metformin on ERSS in primary rat cardiomyocytes. We found that metformin activates the PERK-ATF4 but not the ATF6 or IRE1-XBP1 branch in ERSS and leads to a strong upregulation of CHOP mRNA and protein. Surprisingly, long-term induction of CHOP by metformin is not accompanied by apoptosis even though CHOP is regarded to be a mediator of ER-stress-induced apoptosis. In conclusion, metformin induces distinct ER stress pathways in cardiomyocytes and our results indicate that CHOP is not necessarily a mediator of apoptosis. Metformin might exert its cardioprotective effect through selective activation of ERSS pathways in the cardiomyocyte.

Fig. 1.

Metformin leads to the induction of CHOP in rat cardiomyocytes. (A) Induction of CHOP mRNA and protein. Neonatal rat cardiomyocytes were cultured for 48 hours in the presence or absence of 2.5 mM metformin. mRNA was reverse transcribed and analyzed by quantitative PCR (qPCR). Measurements were normalized against the expression of the adiponectin receptor. CHOP levels in the absence of metformin were set to 100%. n=8, P>0.0001. Total protein were prepared from the same cultures and analyzed by SDS-PAGE and western blotting. (B) Concentration dependency of CHOP induction. Samples were treated for 48 hours with the indicated concentration of metformin and processed for qPCR as in A. n=2–4. (C) Time course of CHOP induction. Samples were treated for the indicated time span with 2.5 mM metformin and processed as in A. Co, 48 hour treatment in the absence of metformin. n=2. (D) Time course of CHOP protein induction in response to metformin. Total protein was harvested at the indicated time points and analyzed by SDS-PAGE and western blotting.

Fig. 2.

Induction of the PERK-ATF4 pathway by metformin. (A) Phospho-PERK (PERK-P) is stimulated by metformin. Western blot analysis. Tu, Tunicamycin treatment for 6 hours; Co, 48-hour mock treatment; Metformin, 48-hour 2.5 mM treatment. (B) Time course of PERK phosphorylation in response to metformin. Samples were treated as in Fig. 1A. (C) ATF4 mRNA induction by metformin. Conditions as in Fig. 1A, n=8. (D) Metformin-dependent induction of ATF4: concentration dependency after treatment for 48 hours. n=2–4. (E) Time course of ATF4 induction during incubation with 2.5 mM metformin. co, 48-hour incubation in the absence of metformin. n=2. (F) Knockdown of ATF4. Cells were treated with siRNA directed against ATF4 or with control siRNA. In both cases, ATF4 mRNA levels were tested in the absence or presence of metformin. ATF4 knockdown also blocks metformin-dependent upregulation. (G) Metformin-dependent CHOP induction is dependent on ATF4. In cells treated with ATF4 siRNA, CHOP induction by metformin is suppressed. (H) Dependence of CHOP inducibility on ATF4 knockdown. Degree of CHOP induction by metformin (y-axis: relative CHOP induction by metformin in the absence and presence of ATF4 siRNA) is negatively correlated with ATF4 knockdown. n=3.

Fig. 3.

The IRE1-XPB1 and the ATF6 pathways are not induced by metformin. (A) Metformin does not activate XBP1 splicing. Splicing of XBP1 mRNA was analyzed after restriction with PstI. The unspliced form of XBP1 can be cut by PstI, but XBP1s (spliced) is not a PstI substrate. XBP1 splicing is strongly induced by tunicamycin (Tu). (B) Quantification of A. XBP1 splicing is expressed as the relative proportion of the intensities of XBP1s and the sum of XBP1s and the larger form of the PstI digest. (C,D) ATF6 mRNA is not upregulated by metformin treatment. Concentration (C)- and time (D)-dependent analysis of mRNA expression. n=2. (E) ATF6 is not activated by metformin. Rat cardiomyocytes were transfected with FLAG-ATF6. ATF6 and activated N-terminal domain of ATF6 were detected by western blotting. 5 mM DTT treatment was used as a positive control. The band marked by one asterisk is the only band detected in the absence of transfection (not shown). Two asterisks: Poinceau-S-stained western blot as loading control.

Fig. 4.

Effect of metformin on the ER luminal chaperone Bip. (A) Induction of Bip (not significant: P=0.1, n=8). (B) Metformin-dependent induction of Bip: concentration dependency after treatment for 48 hours. n=2–4. (C) Time course of ATF4 induction during incubation with 2.5 mM metformin. co, 48-hour incubation in the absence of metformin. n=2 (40 hours: n=1).

Fig. 5.

Effect of metformin treatment on ATF3 and UPR-to-apoptosis signaling. (A,B) ATF3 expression is not upregulated upon metformin treatment. Concentration (A)- and time (B)-dependent analysis of mRNA expression. n=2. (C) Puma and Bim mRNA levels are not changed by metformin. Bcl-2 expression is reduced by 40% (P<0.02, n=4). (D) Induction of Puma and Bim by tunicamycin.

Fig. 6.

Metformin treatment does not induce apoptosis. (A,B) Caspase 12 is not induced by metformin. (B) Caspase 12 is specifically induced by tunicamycin (Tu), but not by staurosporine. Co, control. (C) Induction of caspase 3 by staurosporine, but not by metformin. (D) Tunicamycin, but not metformin, causes apoptosis in neonatal rat cardiomyocytes: TUNEL (48 hours of 2.5 mM metformin; 6 hours of tunicamycin; mock treatment of 48 hours). (E) Phase-contrast bright-field microscopy: metformin does not induce apoptotic stress granules in cardiomyocytes. Cells were treated as in D.

Fig. 7.

Model: metformin triggers specific ERSS pathways without inducing apoptosis. Schematic integration of our results into the ERSS and apoptosis pathways. Factors investigated in this study are highlighted in green and symbols nearby indicate the effect of metformin on these factors: red and green arrows indicate upregulation and downregulation, respectively; gene products marked with a blue circle remain unchanged. Metformin leads to activation and phosphorylation of PERK, and, in turn, the upregulation of the transcription factor ATF4, which is one of the activators of the transcription factor CHOP. Chemical activation of ERSS (e.g. by tunicamycin) shifts the stoichiometric relation of Bcl-2 family members towards apoptosis by upregulation of proapoptotic BH3-only proteins (including Bim and Puma) and by downregulation of anti-apoptotic members such as Bcl-2. These changes are believed to be mediated by CHOP. By contrast, metformin leads to downregulation of Bcl-2, but the expression levels of Bim and Puma remain unchanged. Specific chemical activation of ERSS also leads to the activation of the ER-located caspase 12, and to splicing and activation of XBP1. By contrast, metformin neither activates ATF6 proteolytic processing nor the IRE1-XBP1 arm of ERSS. In consequence, metformin does not activate the mitochondrial- and caspase-12-dependent ER stress pathways that eventually lead to the activation of caspase 3. G, Golgi compartment.