Loss of X-box binding protein 1 in Müller cells augments retinal inflammation in a mouse model of diabetes.

ABSTRACT

AIMS/HYPOTHESIS: Müller glia (MG) are major sources of retinal cytokines, and their activation is closely linked to retinal inflammation and vascular leakage in diabetic retinopathy. Previously, we demonstrated that X-box binding protein 1 (XBP1), a transcription factor activated by endoplasmic reticulum (ER) stress in diabetic retinopathy, is involved in regulation of inflammation in retinal endothelial cells. Now, we have explored the role of XBP1 and ER stress in the regulation of MG-derived proinflammatory factors, and their influence on vascular permeability in diabetic retinopathy. METHODS: MG-specific conditional Xbp1 knockout (Xbp1Müller-/-) mice were generated by crossing Xbp1 flox/flox mice with Müller-Cre transgenic mice. Diabetes was modelled by induction with streptozotocin, and retinal vascular permeability was measured with FITC-conjugated dextran 2 months after induction. Primary Müller cells were isolated from Xbp1Müller-/- and Xbp1Müller+/+ mice and exposed to hypoxia and high levels of glucose. Levels of ER-stress and inflammatory factors were examined by real-time PCR, western blotting or immunohistochemistry. RESULTS: Xbp1Müller-/- mice exhibited normal retinal development and retinal function and expressed similar levels of ER-stress and inflammatory genes to Xbp1Müller+/+ littermates. In diabetes-inducing conditions, compared with Xbp1Müller+/+ mice, Xbp1Müller-/- mice had higher mRNA levels of retinal Vegf (also known as Vegfa) and Tnf-α (also known as Tnf) and ER-stress marker genes Grp78 (also known as Hspa5), Atf4, Chop (also known as Ddit3) and Atf6 and higher protein levels of vascular endothelial growth factor (VEGF), TNF-α, phospho-c-Jun N-terminal kinase (JNK), 78 kDa glucose-regulated protein (GRP78), phospho-eukaryotic translation initiation factor (eIF)2α and activating transcription factor (ATF)6. Retinal vascular permeability was significantly higher in diabetic Xbp1Müller-/- mice than in diabetic Xbp1Müller+/+ mice (p < 0.01). Results obtained in vitro with primary Müller cells isolated from Xbp1Müller-/- mice confirmed higher expression levels of inflammatory and ER-stress markers (but not GRP78) than in cells from Xbp1Müller+/+ mice. Moreover, XBP1-deficient Müller cells were more susceptible to high-glucose- or hypoxia-induced ER stress and inflammation than cells from Xbp1Müller+/+ mice. Inhibition of ER stress with chemical chaperones suppressed hypoxia-induced VEGF and TNF-α production in XBP1-deficient Müller cells. CONCLUSIONS/INTERPRETATION: Our results have revealed an important role of XBP1 and ER stress in MG-driven retinal inflammation, and suggest that targeting ER stress may represent a promising approach for the prevention and treatment of diabetic retinopathy.