Increased Angiopoietin-1 and -2 levels in human vitreous are associated with proliferative diabetic retinopathy

Abstract

Background: Diabetic retinopathy is a frequent complication of diabetes mellitus and a leading cause of blindness in adults. The objective of this study was to elucidate the diabetic retinopathy pathophysiology in more detail by comparing protein alterations in human vitreous of different diabetic retinopathy stages.

Methods: Vitreous samples were obtained from 116 patients undergoing pars plana vitrectomy. Quantitative immunoassays were performed of angiogenic factors (VEGF-A, PIGF, Angiopoietin-1, Angiopoietin-2, Galectin-1) as well as cytokines (IL-1β, IL-8, IFN-γ, TNF-α, CCL3) in samples from control patients (patients who don't suffer from diabetes; n = 58) as well as diabetes mellitus patients without retinopathy (n = 25), non-proliferative diabetic retinopathy (n = 12), and proliferative diabetic retinopathy patients (n = 21). In addition, correlation analysis of protein levels in vitreous samples and fasting glucose values of these patients as well as correlation analyses of protein levels and VEGF-A were performed.

Results: We detected up-regulated levels of VEGF-A (p = 0.001), PIGF (p<0.001), Angiopoietin-1 (p = 0.005), Angiopoietin-2 (p<0.001), IL-1β (p = 0.012), and IL-8 (p = 0.018) in proliferative diabetic retinopathy samples. Interestingly, we found a strong positive correlation between Angiopoietin-2 and VEGF-A levels as well as a positive correlation between Angiopoietin-1 and VEGF-A.

Conclusion: This indicated that further angiogenic factors, besides VEGF, but also pro-inflammatory cytokines are involved in disease progression and development of proliferative diabetic retinopathy. In contrast, factors other than angiogenic factors seem to play a crucial role in non-proliferative diabetic retinopathy development. A detailed breakdown of the pathophysiology contributes to future detection and treatment of the disease.

Fig 1. Up-regulation of different angiogenic factors in vitreous samples.

A) VEGF-A was significantly up-regulated in vitreous samples of PDR patients in comparison to all other analyzed groups (all: p<0.001). B) Additionally, PIGF was elevated in the PDR group in contrast to the control (p<0.001), the DM (p<0.001), and the NPDR group (p = 0.001). C) Angiopoietin-1 was increased in PDR samples compared to control (p = 0.005) and DM samples (p = 0.013). D) Angiopoietin-2 was also elevated in the PDR group in contrast to the other study groups (all: p<0.001). DM: diabetes mellitus; NPDR: non-proliferative diabetic retinopathy; PDR: proliferative diabetic retinopathy (PDR); values are mean ± SEM ± SD; *p<0.05; **p<0.01; ***p<0.001.

Fig 2. Increased pro-inflammatory cytokine expression in PDR.

A) An increased IL-1β expression level was noted in PDR vitreous in contrast to control (p = 0.012) and DM (p = 0.043) samples. B) In addition, the IL-8 concentration was significantly higher in PDR patients than in control (p = 0.018) and NPDR patients (p = 0.034). C) Galectin-1 displayed a trend towards an up-regulation in PDR samples when compared to control samples (p = 0.053). DM: diabetes mellitus; NPDR: non-proliferative diabetic retinopathy; PDR: proliferative diabetic retinopathy; values are mean ± SEM ± SD; *p<0.05.

Fig 3. Correlation of protein concentration and fasting glucose level.

A) Correlation analysis of VEGF-A concentration in vitreous humour and fasting glucose level revealed a significant correlation for all samples (r = 0.366; p<0.001; r² = 0.134). B) Correlation analysis of PIGF concentration and fasting glucose also showed a significant correlation (r = 0.186; p = 0.046; r² = 0.035). C) Scatterplot of Angiopoietin-1 versus fasting glucose. A significant correlation between Angiopoietin-1 level and fasting glucose level could be observed (r = 0.3275; p<0.001; r² = 0.076). D) Also, a significant correlation was found between Angiopoietin-2 and fasting glucose level (r = 0.348; p<0.001; r² = 0.121). E) A correlation between levels of IL-1β and fasting glucose could not be observed (r = 0.023; p = 0.813; r² = 0.001). F) In contrast, a significant correlation between IL-8 and fasting glucose level was detectable (r = 0.374; p<0.001; r² = 0,140). G) Scatterplot of Galectin-1 levels and fasting glucose levels revealed no correlation (r = 0.099; p = 0.386; r² = 0.0098). H) A significant correlation between Angiopoietin-1 and Angiopoietin-2 levels were detectable, excluding one outlier patient (patient #52; r = 0.245; p = 0.011; r² = 0.060). Each blue dot represents one patient, linear regressions are displayed as solid grey lines.

Fig 4. Correlation of VEGF-A levels with other protein concentrations.

A) A significant correlation of Angiopoietin-1 and VEGF-A was noted for all samples (r = 0.228; p = 0.023; r² = 0.052). B) Scatterplot of Angiopoietin-2 versus VEGF-A. A significant correlation between Angiopoietin-2 level and VEGF-A could be observed (r = 0.348; p<0.001; r² = 0.121). C) Correlation analysis of VEGF-A and PIGF concentration also showed a significant correlation (r = 0.479; p < .001; r² = 0.229). D) Also, between Galectin-1 and VEGF-A a significant correlation was detectable (r = 0.271; p = 0.019; r² = 0.074). E) A significant correlation between IL-8 and VEGF-A was detectable (r = 0.371; p<0.001; r² = 0.138). F) In contrast, a correlation between levels of IL-1β and VEGF-A could not be observed (r = 0.196; p = 0.051; r² = 0.039). Each blue dot represents one patient, linear regressions are displayed as solid grey lines.