Effects of O-GlcNAcylation on functional mitochondrial transfer from astrocytes

Abstract

Mitochondria may be transferred from cell to cell in the central nervous system and this process may help defend neurons against injury and disease. But how mitochondria maintain their functionality during the process of release into extracellular space remains unknown. Here, we report that mitochondrial protein O-GlcNAcylation is a critical process to support extracellular mitochondrial functionality. Activation of CD38-cADPR signaling in astrocytes robustly induced protein O-GlcNAcylation in mitochondria, while oxygen-glucose deprivation and reoxygenation showed transient and mild protein modification. Blocking the endoplasmic reticulum - Golgi trafficking with Brefeldin A or slc35B4 siRNA reduced O-GlcNAcylation, and resulted in the secretion of mitochondria with decreased membrane potential and mtDNA. Finally, loss-of-function studies verified that O-GlcNAc-modified mitochondria demonstrated higher levels of neuroprotection after astrocyte-to-neuron mitochondrial transfer. Collectively, these findings suggest that post-translational modification by O-GlcNAc may be required for supporting the functionality and neuroprotective properties of mitochondria released from astrocytes.

Keywords: CD38; O-GlcNAcylation; astrocytes; help-me signaling; neuroprotection; post-translational modification.

Figure 1.

CD38 stimulation increases mitochondrial O-GlcNAcylation in astrocytes: (a) Mitochondrial fraction was obtained from cultured rat cortical astrocytes over time after NAD (5 mM) treatment. (b) Western blot analysis demonstrated that O-GlcNAcylated proteins were steadily increased in isolated mitochondria (n = 4). (c) Fluorescent staining showed that some mitochondria were colocalized with O-GlcNAc. Scale: 20 µm. (d) Rat cortical astrocytes were subjected to oxygen-glucose deprivation for 2 hrs and reoxygenation for up to 24 hrs. Western blot analysis showed that O-GlcNAc modification in mitochondrial proteins was transiently increased during reoxygenation. All values are mean +/– SD.

Figure 2.

O-GlcNAcylation in extracellular mitochondria: (a) TEM analysis confirmed the presence of extracellular mitochondria in astrocyte-conditioned media (ACM). (b–c) Flow cytometry analysis showed that NAD (5 mM) treatment increased extracellular mitochondria (GFP-labeled mitochondrial pyruvate dehydrogenase E1 Alpha 1), and the effect was blocked by cADPR inhibitor (8-Br-cADPR, 10 µM) (Control; n = 7, NAD; n = 6, NAD + 8-Br-cADPR; n = 5). (d) A standard cell viability assay (WST) showed no cell death after treatment (n = 8). (e–f) Extracellular mitochondria in ACM were labeled by Mitotracker Deep Red (Mito DR, 50 nM) and O-GlcNAc-FITC antibody (2 µg/mL) for 30 min at 37 °C. Flow cytometry analysis showed that under normal condition, around 20% of extracellular mitochondria had O-GlcNAc modification. However, the number was increased up to 70–80% when astrocytes were stimulated by NAD (5 mM) or cADPR (100 µM) (Control; n = 7, NAD; n = 3, cADPR; n = 4). All values are mean +/– SD. *P < 0.05, **P < 0.01. Multiple comparisons were evaluated by one-way ANOVA followed by Tukey's test.

Figure 3.

The ER - Golgi traffic regulates mitochondrial protein O-GlcNAc modification: (a) Western blot demonstrated that Brefeldin A (BFA, 5 µg/mL, an inhibitor for the ER-Golgi traffic) inhibited protein O-GlcNAcylation in astrocytic mitochondria. (b–d) FACS demonstrated that NAD (5 mM) increased O-GlcNAc-positive extracellular mitochondria in ACM. Co-treatment with Brefeldin A (BFA, 5 µg/mL) significantly decreased O-GlcNAc-positive mitochondria, but mitochondrial secretion was not affected by BFA (n = 12). All values are mean +/– SD. *P < 0.05, **P < 0.01. One-way ANOVA followed by Tukey's test.

Figure 4.

The Golgi-resident slc35b4 may transport O-GlcNAc for mitochondrial protein modification: (a) UDP-GlcNAc antiporter slc35b4 was highly expressed in astrocytes and siRNA treatment selectively inhibited slc35b4 (n = 3). (b) Representative plots from FACS analysis. (c–d) FACS revealed that suppression of slc35b4 decreased O-GlcNAc positive extracellular mitochondria (c) without affecting mitochondrial secretion (d) (n = 3). All values are mean +/– SD. *P < 0.05, **P < 0.01, unpaired t-test (two-tailed).

Figure 5.

Inflammation and longer period of cell culture may interfere ER-Golgi traffic in astrocytes: (a) Rat cortical astrocytes were pre-exposed to combination of cytokines (IL-1α: 3 ng/mL, TNF-α: 30 ng/mL, C1q: 400 ng/mL) for 24 hrs or cultured for longer period of time (3 weeks or 12 weeks). (b) qPCR analysis demonstrated that slc35b4 was downregulated (Control: n = 18, A1 astrocyte induction: n = 9). (c) Culturing for 12 weeks downregulated slc35b4 (3w: n = 12, 12w: n = 6). (d) NAD (5 mM) was treated to A1 astrocytes for 24 hrs. Western blot analysis confirmed that mitochondrial protein O-GlcNAcylation was decreased in astrocytic mitochondria. (e) Concomitantly, mitochondrial protein O-GlcNAcylation was reduced by 12 week-cell culture compared to 3 week-cell culture. All values are mean +/– SD. *P < 0.05, unpaired t-test (two-tailed).

Figure 6.

O-GlcNAcylated mitochondria show higher neuroprotection in vitro. (a) JC1 mitochondrial membrane potential in extracellular mitochondria was decreased when O-GlcNAcylation was suppressed by BFA (5 µg/mL) (n = 15). (b) Extracellular vesicles were collected from ACM, then mtDNA content (ND-1) was assessed. BFA significantly decreased mtDNA in ACM (n = 3). (c) We assessed neuroprotective effect of pure mitochondrial subset isolated from astrocytes instead of using astrocyte-conditioned media that potentially contain multi-beneficial soluble factors besides extracellular mitochondria. O-GlcNAc-mitochondria (NAD mito, 2 µg/well) or glycosylation-reduced mitochondria (NAD+BFA mito, 2 µg/well) were treated in rat cortical neurons following oxygen-glucose deprivation (OGD) for 2 hrs. (d) After 24 hrs, NAD mito were more remained in the treated neurons than NAD+BFA mito. (e) Neuronal viability following 2-h OGD was about 60%. Treatment with NAD mito increased neuronal viability, but Brefeldin A treatment canceled mitochondria-mediated neuroprotection (n = 8). (f–g) Immunostaining (f) and western blot (g) demonstrated that a neuroplasticity marker GAP43 was recovered by treatment with O-GlcNAcylated mitochondria, while blocking O-GlcNAcylation diminished the effect (n = 4). All values are mean +/– SD. *P < 0.05, **P < 0.01. A,B,D: unpaired t-test (two-tailed), (e,g) One-way ANOVA followed by Tukey's test.

Figure 7.

Schematic of O-GlcNAc modification-mediated secretion of functionally beneficial mitochondria. NAD-CD38 pathway activates the ER-Golgi traffic-mediated O-GlcNAcylation in mitochondrial protein. Oxygen-glucose deprivation (OGD) and reoxygenation may be partly involved in the regulation of mitochondrial protein O-GlcNAcylation in astrocytes, while excessive inflammation or aging pathology may retard mitochondrial protein modification with O-GlcNAc through downregulating SLC35B4. Dotted line from the OGD stimulus indicates that multiple wide-ranging and complex signaling pathways in astrocytes so any change in O-GlcNAc response may include CD38-dependent and CD38-independent pathways.