Autophagy defect is associated with low glucose-induced apoptosis in 661W photoreceptor cells

Abstract

Glucose is an important metabolic substrate of the retina and diabetic patients have to maintain a strict normoglycemia to avoid diabetes secondary effects, including cardiovascular disease, nephropathy, neuropathy and retinopathy. Others and we recently demonstrated the potential role of hypoglycemia in diabetic retinopathy. We showed acute hypoglycemia to induce retinal cell death both in vivo during an hyperinsulinemic/hypoglycemic clamp and in vitro in 661W photoreceptor cells cultured at low glucose concentration. In the present study, we showed low glucose to induce a decrease of BCL2 and BCL-XL anti-apoptotic proteins expression, leading to an increase of free pro-apoptotic BAX. In parallel, we showed that, in retinal cells, low glucose-induced apoptosis is involved in the process of autophagosomes formation through the AMPK/RAPTOR/mTOR pathway. Moreover, the decrease of LAMP2a expression led to a defect in the autophagosome/lysosome fusion process. Specific inhibition of autophagy, either by 3-methyladenine or by down-regulation of ATG5 or ATG7 proteins expression, increased caspase 3 activation and 661W cell death. We show that low glucose modifies the delicate equilibrium between apoptosis and autophagy. Cells struggled against low nutrient condition-induced apoptosis by starting an autophagic process, which led to cell death when inhibited. We conclude that autophagy defect is associated with low glucose-induced 661W cells death that could play a role in diabetic retinopathy. These results could modify the way of addressing negative effects of hypoglycemia. Short-term modulation of autophagy could be envisioned to treat diabetic patients in order to avoid secondary complications of the disease.

Figure 1. Low glucose induces 661W cell death via the BCL2/BAX pathway.

661W cells were cultured as mentioned in material & methods and then incubated at 1 mM glucose for different time periods (2, 4, 8, 24 and 48 h) or for 48 h at high (25 mM) glucose concentrations. A) Graphic representation of Bcl2, Bcl-XL and Bax mRNA expression normalized by housekeeping genes. Results are expressed as mean ± SEM of 2 experiments performed in triplicate. B) Western blot of BCL2, BCL-XL and BAX proteins after exposure to 1 mM or 25 mM glucose concentration during 24 or 48 h. Protein amounts were normalized with TUBULIN and compared to 25 mM condition set to 100%. Results are expressed as mean ± SEM of 3 experiments, *p<0.040. C) Immunoprecipitation analysis of active BAX from cells exposed to 1 mM or 25 mM glucose during 48 h. Left panel showed the expression of total BAX in whole cell lysates, while right panel showed the expression of immunoprecipitated active BAX (black arrow). Immunofluorescence of active BAX (green) is observed in cells incubated at low glucose during 48 h but not at high glucose conditions.

Figure 2. Autophagosomes formation is induced by low glucose in 661W cells and retinal explants.

661W cells and retinal explants were cultured as mentioned in material & methods and then incubated at 1 mM glucose for different time periods (2, 4, 8, 24 and 48 h) or for 48 h at high (25 mM) glucose concentrations. A) Western blot of the LC3-II marker of autophagy is representative of 3 different experiments. B) Western blot and quantification of LC3-II expression in retinal explants, cultured at 1 mM or 25 mM glucose. Western blot is representative of 2–3 experiments and results are expressed as mean ± SEM of 5 different retinas for each condition, * p<0.03 and # p<0.0001. C) Electron micrographs showing the effect of low glucose-induced autophagosome accumulation in 661W photoreceptor cells treated 48 h with 1 mM glucose concentration (arrows) compared to normal glucose concentration (25 mM).

Figure 3. Low glucose induces autophagic flux through AMPK/RAPTOR/mTOR pathway.

661W cells were cultured as mentioned in material & methods, and then incubated at low (1 mM) or high (25 mM) glucose for 24 or 48 h. A) Western blot showed the phosphorylation and the expression of mTOR and is representative of three different experiments. B) Western blot showed the phosphorylation and the expression of AMPK. Quantifications of pAMPK/AMPK ratio are expressed as mean ± SEM of 3 experiments (n = 8), * p<0.035. C) Western blot showed the phosphorylation and the expression of RAPTOR. Oligomycin (Olig) was used at for positive control induction of pRAPTOR. Quantifications of pRAPTOR/TUBULIN ratio are expressed as mean ± SEM of 3 experiments (n = 8), # p<0.019.

Figure 4. Low glucose induces a decrease in LAMP2 expression.

661W cells were cultured as mentioned in material & methods, and then incubated at low (1 mM) or high (25 mM) glucose for diverse periods of time. A) Quantification of Lamp2 mRNA by qPCR analysis, RL8 gene was used to normalize gene expression. Results are expressed as mean ± SEM of 3 experiments, * p<0.02 and # p<0.0008. B) Western blot analysis of LAMP2a protein expression. Results are expressed as mean ± SEM of 3 experiments, * p<0.03 and #p<0.001. C) Immunofluorescence staining of LAMP2a (green) in 661W cells cultured at 1 mM and 25 mM glucose concentration. D) Western blot analysis and quantification of p62 in 661W cells, western blot is representative of three experiments performed in triplicate and results are expressed as mean ± SEM, ** p<0.0011. E) Western blot and quantification of p62 expression in retinal explants, cultured at 1 mM or 25 mM glucose. Western blot is representative of 2–3 experiments and results are expressed as mean ± SEM of 5 different retinas for each condition, * p<0.002.

Figure 5. Low glucose-induced LC3-II and p62 accumulation is principally due to a defect of autophagosome/lysosome fusion.

661W cells were cultured as mentioned in material & methods, and then incubated at low (1 mM) or high (25 mM) glucose for 48 h in absence or in presence of fusion inhibitors (10 µg/ml Pepstatin/E64 and 75 µM Chloroquine) during the last 4 h. A) Representative western blot analysis of LC3-II protein expression and quantification. Results are expressed as mean ± SEM of 4 experiments, *p<0.03 (1 mM vs. 25 mM without inhibitors). B) Transfection of 661W cells with a lentivirus expressing the GFP-LC3 chimeric protein and incubation for 48 hrs at 1 mM (a-d) and 25 mM (e-h) glucose in absence (a and e) or in presence of 50 nM Bafilomycin (b and f), 10 µg/ml Pepstatine/E64 (c and g) and 75 µM Chloroquine (d and h) for the last 4 h. C) Transfection of 661W cells with a lentivirus vector expressing the mRFP-GFP-LC3 chimeric protein and incubation for 48 h at 1 mM (a–f) and 25 mM (g–l) glucose in absence (a–c and g–i) or in presence (d–f and j–l) of 75 µM chloroquine during the last 4 h.

Figure 6. 3-MA chemical inhibition of autophagy increases low glucose-induced cell death and caspase 3 activity.

661W cells were cultured as mentioned in material & methods, and then incubated at low (1 mM) or high (25 mM) glucose for different periods of time (8 or 48 h). A) TUNEL assay was performed in absence (a, d) or in presence (b, c and e, f) of 600 µM 3-MA, at 1 mM (a, b, c) or 25 mM (d, e, f) glucose concentrations, for different periods of time as indicated to the left. White arrows show TUNEL positive cells. Quantification of TUNEL positive cells was performed in three different experiments, *p<0.05; #p<0.0001 and **p<0.002 B) Measures of Caspase 3 activity, results are expressed as mean ± SEM of 3 experiments (n = 13), *p<0.04 and #p<0.0002, and immunostaining of cleaved Caspase 3 in 661W cells incubated at 1 mM (a and b) or 25 mM (c and d) glucose concentrations in absence (a and c) or presence of 600 µM 3-MA (b and d).

Figure 7. Specific inhibition of ATG5 and ATG7 decrease low glucose-induced LC3-II accumulation.

Lentiviruses expressing specific shRNA were used to decrease the expression of ATG5 and ATG7 as described in material & methods, then each clonal cell colony was treated and cultured as mentioned in material & methods. A) Representative western blot analysis of ATG7, and quantification. Results are expressed as mean ± SEM of 3 experiments performed in triplicate, *p<0.0001. B) Quantification of Atg5 mRNA expression was performed by PCR and results expressed as mean ± SEM of 2 experiments performed in duplicate, *p<0.002. C) Representative western blot analysis of LC3-II expression in specific ATG7 and ATG5 knockdown cells cultured at 1 mM and 25 mM glucose during 48 h and quantification. Results are expressed as mean ± SEM of 3–4 experiments performed in triplicate, #p<0.0001.

Figure 8. Specific ATG5 or ATG7 inhibition induces cell death and caspase 3 activity.

661W cells were cultured as mentioned in material & methods. A) Each clonal cell colony was cultured at 1 mM (g, h, i) or 25 mM (j, k, l) glucose for 48 h prior cell death analysis by TUNEL assay. White arrows show TUNEL positive dying cells and quantification is representative of three distinct experiments, *p<0.005; and #p<0.0001. B) Representative western blot and quantification showing cleaved Caspase 3 expression in ATG5 and ATG7 down-regulated clonal cell cultured at 1 mM and 25 mM glucose concentrations. Western blot is representative of three distinct experiments and quantification expressed as 100% of control (25 mM), **p<0.005 and #p<0.0001.

Figure 9. Diagrammatic pathways involved in low glucose-induced autophagy defect.

In high (25 mM) glucose concentration, the BCL-2/BAX ratio is not modified, this lead to very low apoptosis; Moreover, AMPK is inactivated and the mTOR complex is able to inhibit autophagosomes formation. On the opposite, in low glucose (1 mM) condition and after 24 h, the autophagic flux is activated via the AMPK/RAPTOR/mTOR pathway. Moreover, low glucose decreases BCL-XL protein expression, which freeing BAX and leads to apoptosis. Direct implication of BCL-2 family proteins in autophagy was not described in this study (dotted lines), but has been described in literature as modulator of free BECLIN1. Longer exposition to low glucose (48 h) induced a decrease in BCL-2 protein expression, which enhanced the effect on BAX (and possibly on BECLIN1). In parallel low glucose induces a decrease in LAMP2 expression, which impaired the autophagosomes/lysosomes fusion process that normally ends autophagy. Inhibition of elongation/maturation process, either chemically (3-MA) or genetically (ATG5/ATG7 KD) led to apoptosis via an increase of Caspase 3. Low glucose. Lysosomal degradation inhibitors (Chloroquine, Bafilomycin and Pepstatin/E64) induce similar pattern than low glucose. Adapted and modified from , and .