Methylglyoxal-induced AMPK activation leads to autophagic degradation of thioredoxin 1 and glyoxalase 2 in HT22 nerve cells

Abstract

Methylglyoxal (MGO) is a major glycating agent that reacts with basic residues of proteins and promotes the formation of advanced glycation end products which are believed to play key roles in a number of pathologies, such as diabetes, Alzheimer's disease, and inflammation. We previously showed that MGO treatment targets the thioredoxin and the glyoxalase systems, leading to a decrease in Trx1 and Glo2 proteins in immortalized mouse hippocampal HT22 nerve cells. Here, we propose that autophagy is the underlying mechanism leading to Glo2 and Trx1 loss induced by MGO. The autophagic markers p62, and the lipidated and active form of LC3, were increased by MGO (0.5mM). Autophagy inhibition with bafilomycin or chloroquine prevented the decrease in Trx1 and Glo2 at 6 and 18h after MGO treatment. Proteasome inhibition by MG132 exacerbated the effect of MGO on Trx1 and Glo2 degradation (18h), further suggesting a role for autophagy. ATG5 small interfering RNA protected Trx1 and Glo2 from MGO-induced degradation, confirming Trx1 and Glo2 loss is mediated by autophagy. In the search for the signals that control autophagy, we found that AMPK activation, a known autophagy inducer, was markedly increased by MGO treatment. AMPK activation was confirmed by increased acetyl coenzyme A carboxylase phosphorylation, a direct AMPK substrate and by decreased mTOR phosphorylation, an indirect marker of AMPK activation. To confirm that MGO-mediated Trx1 and Glo2 degradation was AMPK-dependent, AMPK-deficient mouse embryonic fibroblasts (MEFs) were treated with MGO. Wildtype MEFs presented the expected decrease in Trx1 and Glo2, while MGO was ineffective in decreasing these proteins in AMPK-deficient cells. Overall, the data indicate that MGO activates autophagy in an AMPK-dependent manner, and that autophagy was responsible for Trx1 and Glo2 degradation, confirming that Trx1 and Glo2 are molecular targets of MGO.

Keywords: AMPK, mTOR; Autophagy; Glyoxalase; HT22 cells; Methylglyoxal; Thioredoxin.

Fig. 1

Trx1 and Glo2 proteins in HT22 cells treated with MGO. (A and B) Trx1 and (C and D) Glo2. Representative images (A and C) and quantification (B and D) of Western blots. HT22 cells were treated for 6 or 18 h with 0.5 mM MGO. Statistical significance is presented relative to untreated control as ** p < 0.01 (N = 4–6)

Fig. 2

Protein adducts of MGO in HT22 cells. Treatment for 3 (A), 6 or 18 h (B) with MGO induced a dose dependent increase in MGO-adducts in proteins, as evaluated by Western blot (3 h) or slot blot (6 and 18 h). (C) Graph presenting data showed in A and B (N = 4–5).

Fig. 3

Protein ubiquitination in MGO-treated HT22 cells. (A) Cells were treated for the indicated times and concentrations of MGO, and protein ubiquitination evaluated by Western blot. (B) Graph presenting data showed in A. (C) Representative blot image of ubiquitin monomer and dimer. Position of molecular weight markers are indicated on the right. (D) Sum of α-ubiquitin monomer plus dimer. Statistical significance is presented relative to untreated control as * p < 0.05 or ** p < 0.01 (N = 4–7).

Fig. 4

The effect of autophagy and proteasome inhibitors on protein ubiquitination. After a pre-treatment with the autophagy inhibitor bafilomycin A (BAF, 1 μM) or the proteasome inhibitor MG132 (30 μM) for 1 h, HT22 cells were treated with 0.5 mM MGO for 1 (A), 6 (B) or 18 h (C). After electrophoresis (A) or slot blot (B and C), and transference to PVDF membranes, total protein was probed for α-ubiquitin and is presented relative to actin content (A) or Ponceau S (B and C) (N = 3–4).

Fig. 5

The effects of MGO on autophagy markers. HT22 cells were treated with MGO at the indicated concentrations for the indicated times. Membranes were probed with antibodies to (A) sequestosome (p62), (B) LC3 (LC3-I + LC3-II). Quantification of data relative to 0.5 mM MGO is presented as: (C) p62; (D) total LC3 and (E) the lipidated and active form of LC3 (LC3 II) relative to LC3 I. Statistical significance is presented relative to untreated control as ** p < 0.01 or *** p < 0.001 (N = 5–7).

Fig. 6

The effect of autophagy and proteasome inhibitors on Trx1 and Glo2 protein levels. After a pre-treatment with the autophagy (BAF, 1 μM and CQ, 10 μM), or the proteasome (MG132, 30 μM) inhibitors, HT22 cells were treated with 0.5 mM MGO for 6 h (A, C, E and G), or 18 h (B, D, F and H) and probed for Trx1 (A – D) and Glo2 (E – H). Statistical significance is presented relative to untreated control (0, CTL) as * p < 0.05, ** p < 0.01, *** p < 0.001, or as compared to MGO-treated control (0.5, CTL) ^# p < 0.05 or ^### p < 0.001 (N = 3–4 at 6h, and N = 5–8 at 18h).

Fig. 7

Effect of ATG5 siRNA on Trx1 and Glo2 expression in HT22 cells treated with MGO. (A) The effectiveness of the ATG5 siRNA in decreasing ATG5 protein levels; Trx1 (B and C) and Glo2 (B and D) protein levels were decreased by treatment with 0.5 mM MGO for 18 h in control siRNA treated cells (CTL), but not in the ATG5 siRNA treated cells (ATG5 si). Statistical significance is presented relative to untreated control as ** p < 0.01 relative to CTL; or ^### p < 0.001 relative to the difference between control siRNA (CTL) and ATG5 siRNA (ATG5 si) treated with MGO (N = 3).

Fig. 8

Effects of MGO on the AMPK/mTOR pathway. HT22 cells were treated with MGO for the indicated time points and concentrations. Membranes were probed with antibodies to p-AMPK (A and B); p-ACC (C and D); or p-mTOR (E and F) and expressed relative to α-AMPK, ACC, and actin, respectively. Statistical significance is presented relative to untreated control as * p < 0.05, ** p < 0.01 or *** p < 0.001 (N = 4–7; p-mTOR, N = 2).

Fig. 9

Trx1 and Glo2 expression in wild type and AMPK-deficient MEFs treated with MGO. Trx1 (A and B) and Glo2 (C and D) protein levels were decreased by treatment with 0.5 mM MGO for 18 h in the wild type MEFs (CTL), but not in the AMPK-deficient (AMPK-KO) MEFs. Statistical significance is presented relative to untreated control as * p < 0.05 relative to control MEFs; or ^# p < 0.05 relative to the difference between AMPK-wild type (CTL) and AMPK-deficient (AMPK-KO) MEFs treated with MGO (N = 3–4).

Fig. 10

MGO-dependent mechanisms targeting Trx1 and Glo2 for degradation by autophagy. Activated LC3 (LC3 II) mediates membrane elongation by associating with autophagosomal membrane. Membrane-attached LC3-II allows the anchoring of p62 adaptor protein carrying Ub-tagged proteins, possibly Trx1 and Glo2. MGO putatively induces autophagy by increasing the abundance of LC3-II and p62. By binding to LC3-II, p62 facilitates autophagy by localizing in autophagic compartments, and allowing the transport of Ub-proteins and organelles for degradation. The complex ATG5-ATG12-ATG16 is necessary for autophagosome formation. Knockdown of ATG5 impairs the autophagosome formation and Trx1 and Glo2 degradation. The proteasome inhibitor MG132, which is known to activate autophagy [26,27,45], promotes Trx1 and Glo2 loss. Inset: MGO leads to AMPK activation by phosphorylation (p-AMPK), and as indicated by acetyl coenzyme A phosphorylation and mTOR inhibition, thus releasing mTOR’s negative effect on autophagy. AMPK-KO or autophagy inhibitors BAF and CQ prevent MGO-induced Trx1 and Glo2 degradation. Legend: AMPK, 5′ AMP-activated protein kinase; ATG, autophagy related; BAF, bafilomycin; CQ, chloroquine; Glo2, glyoxalase 2; LC3, autophagy marker light chain 3; MGO, methylglyoxal; mTOR mechanistic target of rapamycin; Trx1, thioredoxin 1; Ub, ubiquitin;