VEGF regulates local inhibitory complement proteins in the eye and kidney
Lindsay S. Keir, … , Moin A. Saleem, Martin Friedlander
Lindsay S. Keir, … , Moin A. Saleem, Martin Friedlander
Published December 5, 2016
Citation Information: J Clin Invest. 2017;127(1):199-214. https://doi.org/10.1172/JCI86418.
View: Text | PDF
Research Article Nephrology Ophthalmology

VEGF regulates local inhibitory complement proteins in the eye and kidney

Abstract

Outer retinal and renal glomerular functions rely on specialized vasculature maintained by VEGF that is produced by neighboring epithelial cells, the retinal pigment epithelium (RPE) and podocytes, respectively. Dysregulation of RPE- and podocyte-derived VEGF is associated with neovascularization in wet age-related macular degeneration (ARMD), choriocapillaris degeneration, and glomerular thrombotic microangiopathy (TMA). Since complement activation and genetic variants in inhibitory complement factor H (CFH) are also features of both ARMD and TMA, we hypothesized that VEGF and CFH interact. Here, we demonstrated that VEGF inhibition decreases local CFH and other complement regulators in the eye and kidney through reduced VEGFR2/PKC-α/CREB signaling. Patient podocytes and RPE cells carrying disease-associated CFH genetic variants had more alternative complement pathway deposits than controls. These deposits were increased by VEGF antagonism, a common wet ARMD treatment, suggesting that VEGF inhibition could reduce cellular complement regulatory capacity. VEGF antagonism also increased markers of endothelial cell activation, which was partially reduced by genetic complement inhibition. Together, these results suggest that VEGF protects the retinal and glomerular microvasculature, not only through VEGFR2-mediated vasculotrophism, but also through modulation of local complement proteins that could protect against complement-mediated damage. Though further study is warranted, these findings could be relevant for patients receiving VEGF antagonists.

Authors

Lindsay S. Keir, Rachel Firth, Lyndsey Aponik, Daniel Feitelberg, Susumu Sakimoto, Edith Aguilar, Gavin I. Welsh, Anna Richards, Yoshihiko Usui, Simon C. Satchell, Valeryia Kuzmuk, Richard J. Coward, Jonathan Goult, Katherine R. Bull, Ruchi Sharma, Kapil Bharti, Peter D. Westenskow, Iacovos P. Michael, Moin A. Saleem, Martin Friedlander

×

Figure 4

Intravitreal anti-VEGF results in reduced CFH and complement activation.

Options: View larger image (or click on image) Download as PowerPoint
Intravitreal anti-VEGF results in reduced CFH and complement activation....
The retinas of mice injected with anti-mouse VEGF showed a significant reduction in VEGF protein after 48 hours (A, n = 4). This was associated with a reduction in CFH RNA, as shown by in situ hybridization (B, CFH red, DAPI blue, n = 8) and confirmed by qPCR (C, n = 6), CFH protein in the retina (D, n = 3), and choroid/RPE (E, 150 kDa band, n = 6). VEGF and CFH concentrations measured in the same lysates showed significant correlation with a Pearson coefficient of 0.94 (95th CI 0.87–0.97) (F, n = 30). There was increased C3 RNA after a single anti-VEGF injection (G, n = 8). A linear C3 staining pattern in the in situ hybridization suggests that Müller cells may be a C3 source (C3 red, DAPI blue). A 200-fold increase in retinal C3 RNA was confirmed by qPCR (H, n = 6). WT mice injected with anti-VEGF also showed increased C5b-9, indicating complement activation (I, n = 3). C3 knockout mice did not show this effect. The changes in RNA were detected after 24 hours, while protein changes were detected after 48 hours. Aqueous humor from 10 ARMD patient eyes was sampled before and 48 hours after a single intravitreal bevacizumab. The samples obtained after injection showed reduced VEGF (J) and increased C3a (K), C4a (L), and C5a levels (M). Patients who had a recurrence of wet ARMD within 100 days of treatment (range 47–86 days, n = 6) had a greater rise in C3a 48 hours after intravitreal bevacizumab injection than patients who relapsed more than 100 days later (range 152–449days, n = 4) (N). There were no significant differences in VEGF (O). Human samples were analyzed by CBA or ELISA carried out in triplicate and analyzed using Mann-Whitney U test. Images represent 3 independent experiments. Scale bars: 100 μm. Unpaired, 2-tailed t test (A, C, D, E, H, J–O); 1-way ANOVA (I). *P < 0.05; **P < 0.01; ***P < 0.001.