Deletion of Thioredoxin-Interacting Protein (TXNIP) Abrogates High Fat Diet-Induced Retinal Leukostasis, Barrier Dysfunction and Microvascular Degeneration in a Mouse Obesity Model

Figures

• 

  Figure 1. (a) High fat diet (HFD) significantly increased weight in both wild type (WT) and TXNIP knockout (TKO) mice. Six-week-old age and gender matched C57Bl/6J WT and TKO mice were randomized for feeding with either standard chow (normal diet; WT-ND and TKO-ND groups) or high fat diet (WT-HFD and TKO-HFD groups) for 18 weeks. Mice were weighed weekly and clearly HFD resulted in similar increases in weight gain in both WT-HFD and TKO-HFD mice when compared to their ND-controls over the 18 weeks (* p < 0.05, n = 11). (b) Area under the curve (AUC) analysis indicated that WT-HFD gained overall more weight when compared to TKO-HFD (# p < 0.05, n = 11).

• 

  Figure 2. Deletion of TXNIP mitigates HFD-induced retinal leukostasis and BRB breakdown. (a) Representative pictures and (b) quantification of the number of adherent leukocytes occluding the poorly perfused retinal micro-vessels per field (indicated by white arrows) showed higher numbers in the WT-HFD group but had no significant effect on the TKO-HFD group compared with WT-ND group (n = 8–11 mice/group; * p < 0.05 vs. other groups). (c) Quantification of BSA-Fluorescence in the retina tissue was higher in the WT-HFD group whereas both TKO-ND and TKO-HFD groups had significantly lower levels of BSA-Fluorescence extravasation, respectively, compared with WT-ND group (n = 5–9 mice/group; * p < 0.05 vs. other groups).

• 

  Figure 3. Deletion of TXNIP mitigates HFD-induced retinal expression of adhesion molecules. (a,c) Representative WB blots and statistical analyses of (b) ICAM-1 and (d) VCAM-1 showed increased expression of ICAM-1 and VCAM-1 levels in WT-HFD group compared with WT-ND group. In contrast, TKO mice groups showed no changes in response to HFD. Two-way ANOVA showed significant interaction between the type of diet and genotype across both ICAM-1 and VCAM-1 expression (n = 4–6 mice/group; * p < 0.05 vs. other groups).

• 

  Figure 4. Deletion of TXNIP abrogates HFD-induced NLRP3-inflammasome activation in the retina. Representative WB blots (a) and statistical analyses of protein expression of retinal (b) cleaved caspase-1 and (c) cleaved IL-1β showed a significant increase in cleaved IL-1β expression, and a trend towards higher levels of cleaved caspase-1 in WT-HFD group compared with WT-ND. In contrast, TKO-ND and TKO-HFD groups showed either no changes or significantly lower levels of all target protein expression compared with WT-ND group, respectively. Two-way ANOVA showed significant interaction between the type of diet across cleaved IL-1β expression and between the genotype across expression of cleaved IL-1β and cleaved caspase-1 (n = 12 mice/group; * p < 0.05 vs. other groups).

• 

  Figure 5. Overexpression of TXNIP activates NLRP3-inflammasome and EC inflammation in an IL-1β-dependent fashion. Representative WB blots (a,e) and statistical analysis of protein expression detected from REC lysates, of endothelial (b) NLRP3, (c) cleaved caspase-1, (d) cleaved IL-1β, and (f) TNF-α, showing significant increases in NLRP3, cleaved IL-1β and TNF-α and a trend towards increased expression of cleaved caspase-1 in the TXNIP overexpressing (TXNIP ++ + vehicle) group compared with the EV-GFP + vehicle control group. Treatment of TXNIP-overexpressing REC with 100 ng/mL of IL-1RA (TXNIP ++ + IL-1RA) suppressed the effect of TXNIP overexpression on inflammasome activation and TNF-α expression. Two-way ANOVA showed significant interaction between IL-1RA and TXNIP overexpression across NLRP3 and cleaved IL-1β expression levels (n = 5–6, each measurement was performed in triplicate using two different cell preparations; * p < 0.05 vs. EV-GFP + vehicle group).

• 

  Figure 6. Silencing of TXNIP abolishes, while its overexpression induces, REC adhesion molecule upregulation in an IL-1β-dependent fashion. (a) Representative WB blots (boxes denote separate lanes from the same gel) and statistical analyses of (b) ICAM-1 and (c) PECAM-1 protein expression showed higher levels in cells incubated with PAL-BSA compared with BSA controls in the scrambled-siRNA group. Silencing TXNIP expression using siRNA abolished such responses. Two-way ANOVA showed significant interaction between Pal-BSA and silencing TXNIP expression across PECAM-1 levels (n = 3–4, each measurement was performed in duplicate using two different cell preparations; * p < 0.05 vs. SCsiRNA + BSA, # vs. SCsiRNA + Pal-BSA). (d) Representative blots and WB analysis of (e) ICAM-1 and (f) PECAM-1 protein expression showed higher levels in TXNIP overexpression (TXNIP ++ + vehicle) group compared with EV-GFP + vehicle control group, respectively. Treatment of TXNIP-overexpressing REC with 100 ng/mL of IL-1RA (TXNIP ++ + IL-1RA) suppressed the effect of TXNIP overexpression on expression of REC adhesion molecules (n = 5–6, each measurement was performed in triplicate using two different cell…

• 

  Figure 7. Deletion of TXNIP in leukocytes prevents leukostasis and REC death. (a) Representative images and (b) quantification of the number of adherent PBMCs (indicated by white arrows) to cultured mouse REC per field after 2 hours and (c) the number of TUNEL-positive nuclei per field after 24 h; showed significantly higher numbers of adherent PBMCs and a trend in increased numbers of TUNEL-positive nuclei in the WT-HFD group compared with WT-ND control group, respectively. In contrast, both TKO-ND and TKO-HFD groups showed no significant changes compared with WT-ND control group. Two-way ANOVA showed a significant interaction between the genotype across the number of adherent PBMCs and between the type of diet across the number of adherent PBMCs and the number of TUNEL-positive nuclei (n = 5–6, each measurement was performed in triplicate using two different cell preparations; * p < 0.05 vs. other groups).

• 

  Figure 8. Deletion of TXNIP abrogates HFD-induced retinal microvascular degeneration. (a) Representative pictures of degenerated acellular capillaries from each of the groups, (b) magnification of a degenerated acellular capillary showing loss of endothelial cell nucleus and the remaining thickened basement membrane and (c) quantification of the number of acellular capillaries per field showed significantly higher numbers in the WT-HFD group compared with WT-ND. Two-way ANOVA showed significant interaction between the type of diet and genotype across microvascular degeneration (n = 6 mice/group; * p < 0.05 vs. all groups). (d) Representative WB blots and (e) statistical analyses of cleaved PARP showed a trend towards increased expression in the WT-HFD group compared with the WT-ND control group. Two-way ANOVA showed significant interaction between genotype (TKO vs. WT) (n = 8 mice/group).

• 

  Figure 9. Deletion of TXNIP abrogates HFD-induced microvascular morphological changes. Representative images of (a) total retinal flat mounts stitched from individual lopes of the retina, (b) representative of a compressed Z-stack image of one of the retina loops and (c) skeletonized compressed images from each of the animal groups. (d) WT-HFD mice had decreased retinal vascular branching density compared with WT-ND control group. In contrast, TKO mice had a comparable morphological appearance compared to the WT-ND group that did not change in response to HFD. Two-way ANOVA showed significant interaction between the type of diet and genotype across microvascular density (n = 4–6 mice/group; * p < 0.05 vs. other groups).