Bevacizumab Injection in Patients with Neovascular Age-Related Macular Degeneration Increases Angiogenic Biomarkers

Abstract

Purpose: To evaluate the expression of 19 angiogenic biomarkers in the aqueous humor before and after intravitreal bevacizumab injection (IVB) in eyes with neovascular age-related macular degeneration (AMD).

Design: Prospective, noncomparative, interventional case series.

Participants: Twenty-three eyes of 23 treatment-naïve patients with choroidal neovascularization (CNV) secondary to neovascular AMD.

Methods: Eyes were diagnosed with CNV secondary to neovascular AMD and were treated with 3 monthly IVBs. Aqueous humor samples were obtained by anterior chamber paracentesis at baseline and immediately before each intravitreal bevacizumab injection.

Main outcome measures: Aqueous humor levels of 19 angiogenic biomarkers (angiopoietin 2, bone morphogenetic protein 9 [BMP-9], epidermal growth factor [EGF], endoglin, endothelin 1, fibroblast growth factor [FGF]-1 and FGF-2, follistatin, granulocyte colony-stimulating factor [GCSF], heparin-binding EGF-like growth factor [HB-EGF], hepatocyte growth factor [HGF], interleukin 8, leptin, placental growth factor [PLGF], vascular endothelial growth factor [VEGF]-A, VEGF-C, VEGF-D, and tissue inhibitor of metalloproteinases [TIMP]-1 and TIMP-2) were measured. Best-corrected visual acuity (BCVA), spectral-domain OCT parameters, and intraocular pressure also were evaluated.

Results: Baseline aqueous VEGF-A expression was elevated in all study eyes before treatment initiation. A statistically significant decrease of VEGF-A was observed at the 1- and 2-month follow-ups. A statistically significant increased concentration was observed in 7 biomarkers: VEGF-C, angiopoietin 2, endothelin 1, follistatin, HB-EGF, HGF, and interleukin 8. The other 11 study biomarker levels (VEGF-D, BMP-9, EGF, endoglin, FGF-1, FGF-2, GCSF, leptin, PLGF, TIMP-1, and TIMP-2) did not show any significant difference during follow-up. The BCVA statistically improved significantly at 2 months. Spectral-domain OCT parameters improved significantly at all follow-ups. Mean intraocular pressure values were not statistically different during the study period.

Conclusions: Despite a decrease in VEGF-A, the aqueous levels of VEGF-C, angiopoietin 2, endothelin 1, follistatin, HB-EGF, HGF, and interleukin 8 increased significantly after intravitreal injection of bevacizumab. These upregulated angiogenic biomarkers may represent new therapeutic targets in exudative AMD.

Figure 1

A–H, Graphs showing angiogenesis-related biomarker aqueous humor levels (in picograms per milliliter) in patients with neovascular age-related macular degeneration (AMD) before (0, 1, and 2 months) each bevacizumab intravitreal injection. A, Vascular endothelial growth factor (VEGF)-A concentration was reduced statistically significantly at the first and second months when compared with baseline. B–H, Seven of 19 studied biomarkers showed a statistically significant increase in their levels at the second month when compared with baseline. I, Percentage variation of studied angiogenesis-related biomarker aqueous humor levels in patients with neovascular AMD before (0, 1, and 2 months) each bevacizumab intravitreal injection. Baseline (month 0) levels were considered as 100%. Biomarkers listed in the legend at the right of panel I were ordered according to final concentration. Dark gray lines correspond to statistically significant biomarkers (VEGF-A was reduced significantly at the first and second months, whereas interleukin 8 [IL-8], hepatocyte growth factor [HGF], VEGF-C, follistatin [FST], angiopoietin 2 [ANG-2], endothelin 1 [ET-1], and heparin-binding epidermal growth factor–like growth factor [HB-EGF] concentrations increased significantly at the second month when compared with baseline), whereas light gray lines correspond to those without statistical significance (VEGF-D, bone morphogenetic protein 9 [BMP-9], epidermal growth factor [EGF], endoglin [ENG], fibroblast growth factor [FGF]-1, FGF-2, granulocyte colony-stimulating factor [GCSF], leptin, placental growth factor [PLGF], tissue inhibitor of metalloproteinases [TIMP]-1, and TIMP-2). P < 0.05 when baseline aqueous humor values were compared with 1-month (Φ) and 2-month (*) values.

Figure 2

Diagram showing the protein–protein interaction network of statistically significant biomarkers investigated in the present study. Each circle represents a significant pathway, and each arrow symbolizes a pathway crosstalk between proteins. Orange and red arrows represent, respectively, the influence of hepatocyte growth factor (HGF) and heparin-binding epidermal growth factor–like growth factor (HB-EGF; shown as light blue circles) on 5 angiogenic factors (dark blue circles) and on the vascular endothelial growth factor (VEGF)-A pathway (red circle). It is important to highlight that HGF does not present significant connection with follistatin (FST), nor does epidermal growth factor (EGF) with angiopoietin 2 (ANG-2). Black arrows represent the influence of VEGF-A on 3 protein pathways (ANG-2, interleukin 8 [IL-8], and endothelin 1 [ET-1]). The VEGF-A blockage represents the upregulation of the 7 represented biomarkers directly (HGF and EGF) and indirectly (ANG-2, IL-8, VEGF-C, ET-1, and FST). The network was adapted from Qiagen’s Ingenuity Pathway Analysis (Qiagen, Redwood City, CA), generated in October 2016.