The diabetic microenvironment causes mitochondrial oxidative stress in glomerular endothelial cells and pathological crosstalk with podocytes

Abstract

Background: In the setting of diabetes mellitus, mitochondrial dysfunction and oxidative stress are important pathogenic mechanisms causing end organ damage, including diabetic kidney disease (DKD), but mechanistic understanding at a cellular level remains obscure. In mouse models of DKD, glomerular endothelial cell (GEC) dysfunction precedes albuminuria and contributes to neighboring podocyte dysfunction, implicating GECs in breakdown of the glomerular filtration barrier. In the following studies we wished to explore the cellular mechanisms by which GECs become dysfunctional in the diabetic milieu, and the impact to neighboring podocytes.

Methods: Mouse GECs were exposed to high glucose media (HG) or 2.5% v/v serum from diabetic mice or serum from non-diabetic controls, and evaluated for mitochondrial function (oxygen consumption), structure (electron microscopy), morphology (mitotracker), mitochondrial superoxide (mitoSOX), as well as accumulation of oxidized products (DNA lesion frequency (8-oxoG, endo-G), double strand breaks (γ-H2AX), endothelial function (NOS activity), autophagy (LC3) and apoptotic cell death (Annexin/PI; caspase 3). Supernatant transfer experiments from GECs to podocytes were performed to establish the effects on podocyte survival and transwell experiments were performed to determine the effects in co-culture.

Results: Diabetic serum specifically causes mitochondrial dysfunction and mitochondrial superoxide release in GECs. There is a rapid oxidation of mitochondrial DNA and loss of mitochondrial biogenesis without cell death. Many of these effects are blocked by mitoTEMPO a selective mitochondrial anti-oxidant. Secreted factors from dysfunctional GECs were sufficient to cause podocyte apoptosis in supernatant transfer experiments, or in co-culture but this did not occur when GECs had been previously treated with mitoTEMPO.

Conclusion: Dissecting the impact of the diabetic environment on individual cell-types from the kidney glomerulus indicates that GECs become dysfunctional and pathological to neighboring podocytes by increased levels of mitochondrial superoxide in GEC. These studies indicate that GEC-signaling to podocytes contributes to the loss of the glomerular filtration barrier in DKD. Video abstract.

Keywords: Crosstalk; Diabetes; Endothelial cells; Mitochondria; Podocytes; ROS.

Fig. 1

Diabetic milieu induces mitochondrial oxidative stress and decreases mitochondrial function of glomerular endothelial cells. a Representative histograms of MitoSOX low and bright fluorescence in mGEC control cells in normal glucose (NG; 5.5 mM glucose) or after culture with either control serum (CS), or diabetic serum (DS) for 24 h without or with mitoTEMPO (mTEMPO; 5μg/ml). b Representative images of mitochondrial networks detected by mitoTracker Red in NG, CS, DS treated mGECs without or with mTEMPO. Scale = 50 μm. c Mitochondrial respiratory reserve capacity was measured by oxygen consumption rate (OCR; Uncoupled respiration (FCCP) over basal respiration) and glycolysis as determined by measurements of d extracellular acidification rate (ECAR) in NG, high glucose (HG; 30 mM glucose), CS, DS treated mGECs. Bars represent mean ± s.e.m. of 3–4 independent experiments; *P < 0.05 versus NG controls

Fig. 2

Diabetic milieu induced increased mtDNA oxidative lesions but not nuclear (nuc) DNA lesions. a Representative immunofluorescence images detecting DNA 8-oxoG (green) and mitochondrial transcription factor A (mTFA; red) and DAPI in mGECs treated with NG, HG, CS, DS without or with mTEMPO for 24 h. Open arrows in A depict clusters and colocalization of 8-oxoG and TFAM. Arrows in B show mGEC nucleus with positive γ-H2AX foci. Scale = 100 μm. b Representative immunofluorescence images detecting γ-H2AX foci (green) and nuclei (DAPI blue) in mGECs treated with NG, HG, CS and DS for 24 h. Scale = 100 μm. Quantification of lesion frequencies in mtDNA and in nuclear (nuc) DNA by QPCR assay in mGECs treated 24 h with NG, HG, CS and DS (n = 3 ± s.e.m. relative amplification normalized to non-damaged NG controls; * P < 0.05 or ** P < 0.01 versus mtDNA control, #P < 0.05 versus nucDNA control). d RT-PCR quantification of transcripts for mitochondrial encoded genes; ND1 and ND4 in mGECs cultured in NG, HG, CS and DS for 24 h. Bars represent mean ± s.e.m. of 4 independent experiments; ND4 *P < 0.05 versus NG or CS controls, ND1 #P < 0.05 versus NG or CS controls

Fig. 3

Diabetic milieu induces increased endonuclease G accumulation in mitochondrial intermembrane space (IMS) and disrupts mitochondrial architecture. a Immunofluorescence detection of endoG-GFP located in the IMS of mGECs cultured in NG, b with CS, c MG132 and with d DS for 48 h. Arrows show endoG clustering. Scale = 20 μm. e Western blot detecting anti-GFP from endoG-GFP whole cell extracts in a-d with β-actin shown as a loading control (n = 3). f Representative transmission electron microscopy images (magnification 7000x, Scale = 1 μm) showing mitochondria (M) in mGEC cultured in NG control shown as elongated mitochondria with dense matrix and densely aligned cristae, N represents nucleus. Below is a higher power images (12,000x, Scale = 0.5 μm), showing clear mitochondrial inner and outer membranes. g Mitochondrial matrix in DS treated mGECs is largely reduced and few cristae remain, outer membrane rupture is seen (open arrow). There are triple and quadruple membrane rings indicative of autophagy and/or mitophagy (*), and a large number of vacuoles were evident (V) and mitochondria with electron dense material (white *). H) DS + mTEMPO treated mGECs show improved mitochondrial architecture with intact outer membrane and inner membrane cristae

Fig. 4

Diabetic milieu induces endothelial cell dysfunction without cell death. a Nitric oxide synthase (NOS) activity detected by DAF-FM fluorescence in mGEC cultured in NG control conditions without or with L-NAME (100 μM), or cultured with HG, CS, DS and DS + mTEMPO for 48 h. b Percentage viable (Annexin and PI negative) and apoptotic (Annexin and PI positive) mGECs cultured in the same conditions as a for 48 h. Bar graphs represent n = 4–5 ± s.e.m. averaged independent experiments; *P < 0.05 versus NG controls

Fig. 5

Diabetic milieu induces autophagy flux in mGECs. a Representative immunofluorescence images detecting 8-oxoG (green), LC3 (red) and nucleai with DAPI (blue) in mGEC cultured with CS, or cultured with DS for 6, 24 and 48 h. Arrows depict clustering and colocalization of 8-oxoG and LC3 puncta. Scale = 100 μm. b western blot show the ratio of LC3-II/LC3-I protein levels in lysates from mGECs cultured in CS or DS for 6 and 48 h in absence (black and white bars) or presence of 6 h bafilomycin A1 (Baf, 100 nM; black bars). c Percentage viable (Annexin and PI negative) and apoptotic (Annexin and PI positive) mGECs cultured in DS without or with 3-MA for 48 h. d Representative contrast images of mGECs treated with 3-MA, or DS without or with 3-MA for 48 h. Bar graphs represents n = 4 ± s.e.m. independent experiments; *P < 0.05 versus DS-Baf in b, viability #P < 0.05 versus DS and apoptotic +P < 0.05 versus DS

Fig. 6

mGEC exposed to diabetic serum secrete factors that cause podocyte cell death. a Percentage podocyte cell death (AnnexinV-FITC and PI positive cells) 24 h after co-culture with mGEC in transwell inserts; without or with HG, or HG in the presence of mitoTEMPO (5μg/ml), or CS, or DS, or DS + mTEMPO for 24 h (n = 4). b Percentage apoptotic podocytes after supernatant transfer (SN) from mGECs; NG-SN, or HG-SN, or HG + mTEMPO-SN, or CS-SN, or DS-SN, or DS + mTEMPO-SN. MitoTEMPO dose used was 5 μg/ml (n = 3). c Representative immunofluorescence images detecting cleaved caspase 3 (red), and F-Actin (phalloidin; green) and nuclei (DAPI; blue) in podocytes cultured with NG-SN, or HG-SN, or CS-SN, or DS-SN, or DS + mTEMPO-SN for 24 h. Scale = 50 μm. Cultures were performed in 10% FCS. Bars in a and b represent averaged independent experiments ± s.e.m.; *P < 0.05, **P < 0.01 versus NG or CS controls, +P < 0.05, ++P < 0.01 versus HG, #P < 0.05 versus DS