Protective roles of peroxiporins AQP0 and AQP11 in human astrocyte and neuronal cell lines in response to oxidative and inflammatory stressors

Abstract

In addition to aquaporin (AQP) classes AQP1, AQP4 and AQP9 known to be expressed in mammalian brain, our recent transcriptomic analyses identified AQP0 and AQP11 in human cortex and hippocampus at levels correlated with age and Alzheimer's disease (AD) status; however, protein localization remained unknown. Roles of AQP0 and AQP11 in transporting hydrogen peroxide (H2O2) in lens and kidney prompted our hypothesis that up-regulation in brain might similarly be protective. Established cell lines for astroglia (1321N1) and neurons (SHSY5Y, differentiated with retinoic acid) were used to monitor changes in transcript levels for human AQPs (AQP0 to AQP12) in response to inflammation (simulated with 10-100 ng/ml lipopolysaccharide [LPS], 24 h), and hypoxia (5 min N2, followed by 0 to 24 h normoxia). AQP transcripts up-regulated in both 1321N1 and SHSY5Y included AQP0, AQP1 and AQP11. Immunocytochemistry in 1321N1 cells confirmed protein expression for AQP0 and AQP11 in plasma membrane and endoplasmic reticulum; AQP11 increased 10-fold after LPS and AQP0 increased 0.3-fold. In SHSY5Y cells, AQP0 expression increased 0.2-fold after 24 h LPS; AQP11 showed no appreciable change. Proposed peroxiporin roles were tested using melondialdehyde (MDA) assays to quantify lipid peroxidation levels after brief H2O2. Boosting peroxiporin expression by LPS pretreatment lowered subsequent H2O2-induced MDA responses (∼50%) compared with controls; conversely small interfering RNA knockdown of AQP0 in 1321N1 increased lipid peroxidation (∼17%) after H2O2, with a similar trend for AQP11 siRNA. Interventions that increase native brain peroxiporin activity are promising as new approaches to mitigate damage caused by aging and neurodegeneration.

Keywords: aquaporins; hydrogen peroxide; lipopolysaccharides; neurodegeneration; protein colocalization.

Figure 1. SHSY5Y neuronal differentiation induced by retinoic acid (RA) and effects on AQP expression

(A) Morphological differentiation induced by 10 μM RA at 1, 3, 5 and 7 days of treatment, compared with non-RA treated control. Scale bar is 0.5 mm; insets are 3× (167 µm). (B) Colorimetric assessment of increased acetylcholinesterase (AChE) enzymatic activity during retinoic acid (RA) induced differentiation, measured on a plate reader at OD410 (ODU). (C) Relative transcript levels of all classes of human AQPs (AQP0 to AQP12) in SHSY5Y cells at days 0, 3 and 7 of RA-induced differentiation, normalized to the reference gene, β-ACTIN, using the 2-ΔCT calculation method. **P<0.01, ****P<0.0001. Data are mean ± SEM; n = 6 per group.

Figure 2. Effects of LPS stress on aquaporin expression patterns in astrocyte and neuronal cell lines

Relative RNA levels of human AQP0 to AQP12 were assessed after 24 h in control or LPS treatment conditions for: (A) SHSY5Y RA-treated neurons at 7 days and (B) 1321N1 astrocytes. Levels calculated using the 2-ΔCT method were normalized to the housekeeping genes GAPDH for astrocytes, and ß-ACTIN for neurons and microglia; *P<0.05, **P<0.01, ***P<0.0001, ****P<0.00001. Data are mean ± SEM; n = 6 per group.

Figure 3. Effects of hypoxic stress on aquaporin expression patterns in astrocyte and neuronal cell lines

Relative RNA levels of human AQP0 to AQP12 were assessed after 1 h hypoxia followed by 0, 12 and 24 h normoxia in (A) SHSY5Y RA-treated neurons at day 7 and (B) 1321N1 astrocytes. Control cells received no hypoxic stress. Expression was normalized to the reference gene GAPDH for astrocytes and β-ACTIN for neurons and microglia using 2-ΔCT calculation method; **P<0.01, ***P<0.001, ****P<0.0001. Data are mean ± SEM; n = 6 per group.

Figure 4. LPS-induced expression of AQP0 and AQP11 in 1321N1 astrocytes and SHSY5Y neurons

Astrocytes (A) and neurons (C) immunolabeled for AQP0 (red), shown at 24 h without (control) or with 100 ng/ml LPS treatment; NPC, no primary antibody control. Immunolabeling of astrocytes (B) and neurons (D) for AQP11 (red), shown at 24 h without (control) or with 100 ng/ml LPS treatment. Hoechst nuclear staining is white for astrocytes and blue for neurons. Scale bars are 40 μM for astrocytes and 50 μM for neurons. Total fluorescence signal intensities for AQP0 in astrocytes (E), AQP11 in astrocytes (F), AQP0 in neurons (G) and AQP11 in neurons (H) were quantified from merged images in HALO using the Area-Quantification FL V.2.1.3 program (see Materials and methods for details), and plotted as histograms with mean ± SEM levels for two samples per treatment group; *P<0.05, ****P<0.0001.

Figure 5. AQP0 and AQP11 channels up-regulated by LPS in 1321N1 astrocytes are located in plasma membrane

Double-immunolabeling results show 1321N1 astrocytes with anti-Na⁺-K⁺-ATPase (‘Na pump’, green; plasma membrane marker), and anti-AQP (red) for AQP0 (A–C) or AQP11 (D–F). Top rows in (A) and (D) show non-LPS controls; bottom rows show treatments with 100 ng/ml LPS. Left columns in A and D show nuclear stain with Hoechst (white). Merged images analyzed by HALO (B,E) illustrate double labeling. Signal intensities for Na pump colocalization with AQP0 (C) or AQP11 (F) were measured with IMARIS in 100 ng/ml LPS-stimulated 1321N1 astrocytes using Z-stack images (‘colocaliz’; left images in C and F); colocalization scores are depicted in white (right images, C and F) as measured using IMARIS. Scale bars are 50 μm.

Figure 6. Increased levels of AQP0 and AQP11 channels in 1321N1 astrocytes at 24 h after LPS as compared with non-LPS controls

Double labeling of 1321N1 astrocytes used the endoplasmic reticulum marker, cytopainter (green; ER marker) and anti-AQP antibodies visualized with anti-rabbit secondary (red) for AQP0 (A–C) or AQP11 (D–F). Top rows in (A) and (D) show non-LPS controls; bottom rows show treatments with 100 ng/ml LPS. Left columns in A and D show nuclear staining with Hoechst (white). Merged images analyzed by HALO (B,E) illustrate double labeling. Signal intensities for cytopainter colocalization with AQP0 (C) or AQP11 (F) were measured with IMARIS in 100 ng/ml LPS-stimulated 1321N1 astrocytes using Z-stack images (‘colocaliz’; left images in C and F); colocalization scores are depicted in white (‘IMARIS’; right images, C and F) as measured using IMARIS software. Scale bars are 50 μm.

Figure 7. Increased levels of AQP0 but not AQP11 channels in SHSY5Y neurons at 24 h after LPS as compared with non-LPS controls

Images show double labeling of SHSY5Y neurons with cytopainter (green; ER marker), and anti-AQP (red) for AQP0 (A–C) or AQP11 (D–F). Top rows in (A) and (D) are non-LPS controls; bottom rows are treatments with 100 ng/ml LPS. Left columns in A and D show nuclear stain with Hoechst (white). Merged images analyzed by HALO (B,E) illustrate double labeling. Signal intensities for cytopainter colocalization with AQP0 (C) or AQP11 (F) were measured with IMARIS in LPS-stimulated SHSY5Y neurons using Z-stack images (‘colocaliz’; left images in C and F); colocalization scores depicted in white (right images, C and F) were measured using IMARIS. Scale bars are 50 μm.

Figure 8. Novel peroxiporin function of AQP0 in 1321N1 astrocyte-like cells

MDA levels (nmol/ml) in (A) 1321N1 astrocytes and (B) BV2 microglia in control and LPS stimulation with and without 5 μM H2O2; n=8. (C) Immunocytochemistry of 1321N1 astrocytes incubated with 2 siRNA kits for AQP0 (top row) and AQP11 (bottom row) for 72 h; scrambled siRNA kits served as an experimental control; LPS stimulation served as a positive control. (D) Expression of AQP0 and AQP11 relative to GAPDH in astrocytes after siRNA knockdown, using two kits each with scrambled siRNA kit as a control. RNA extraction was done at 72 h of siRNA treatment, after a time course that included replacement of transfection media at 48 h and 100 ng/ml LPS stimulation for 24 h. Approximately 100 ng/ml LPS-stimulated cells served as a positive control; no treatment was a negative control; n=6. (E) MDA levels in astrocytes following scrambled, AQP0 and 11 siRNA treatments (using kits 2) at 72 h, after 10 ng/ml LPS stimulation with and without H₂O₂ as indicated; *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001; Mean ± SEM; n=8.

Figure 9. Peroxiporin pathway enabling the efflux of the oxidative stressor hydrogen peroxide from cells

Schematic illustration of the role of the peroxiporin AQP11 in conferring protection from elevated hydrogen peroxide (H₂O₂) levels, enabled by high levels of expression of AQP11 channels in endoplasmic reticulum, with additional expression in plasma membrane. Export of H₂O₂ from intracellular organelles into the cytoplasm followed by removal from the cytoplasm to extracellular fluid compartments is proposed to offset consequences of oxidative stressor accumulation associated with metabolic activity. Graphic generated using Biorender, Publication Agreement # AU26IJ5C1E.