Quantitative proteomic analysis of aqueous humor from patients with drusen and reticular pseudodrusen in age-related macular degeneration

Abstract

Background: To identify novel biomarkers related to the pathogenesis of dry age-related macular degeneration (AMD), we adopted a human retinal pigment epithelial (RPE) cell culture model that mimics some features of dry AMD including the accumulation of intra- and sub-RPE deposits. Then, we investigated the aqueous humor (AH) proteome using a data-independent acquisition method (sequential window acquisition of all theoretical fragment ion mass spectrometry) for dry AMD patients and controls.

Methods: After uniformly pigmented polarized monolayers of human fetal primary RPE (hfRPE) cells were established, the cells were exposed to 4-hydroxy-2-nonenal (4-HNE), followed by Western blotting, immunofluorescence analysis and ELISA of cells or conditioned media for several proteins of interest. Data-dependent acquisition for identification of the AH proteome and SWATH-based mass spectrometry were performed for 11 dry AMD patients according to their phenotypes (including soft drusen and reticular pseudodrusen [RPD]) and 2 controls (3 groups).

Results: Increased intra- and sub-RPE deposits were observed in 4-HNE-treated hfRPE cells compared with control cultures based on APOA1, cathepsin D, and clusterin immunoreactivity. Additionally, the differential abundance of proteins in apical and basal chambers with or without 4-HNE treatment confirmed the polarized secretion of proteins from hfRPE cells. A total of 119 proteins were quantified in dry AMD patients and controls by SWATH-MS. Sixty-five proteins exhibited significantly altered abundance among the three groups. A two-dimensional principal component analysis plot was generated to identify typical proteins related to the pathogenesis of dry AMD. Among the identified proteins, eight proteins, including APOA1, CFHR2, and CLUS, were previously considered major components or regulators of drusen. Three proteins (SERPINA4, LUM, and KERA proteins) have not been previously described as components of drusen or as being related to dry AMD. Interestingly, the LUM and KERA proteins, which are related to extracellular matrix organization, were upregulated in both RPD and soft drusen.

Conclusions: Differential protein expression in the AH between patients with drusen and RPD was quantified using SWATH-MS in the present study. Detailed proteomic analyses of dry AMD patients might provide insights into the in vivo biology of drusen and RPD.

Keywords: Age-related macular degeneration; Complement; Drusen; Reticular pseudodrusen; SWATH-MS.

Fig. 1

Color fundus photos (left) and optical coherence tomography images (right) from patients with drusen or reticular pseudodrusen (RPD) (patients with dry age-related macular degeneration in Sample Set 2 in Table 1). (a) A 71-year-old woman with Drusen, (b) an 80-year-old woman with Drusen, (c) a 76-year-old woman with RPD, (d) a 67-year-old woman with RPD

Fig. 2

Flowchart of the AH proteome analysis. Pooled AH samples were further prepared via two processes (with or without depletion). A portion of the sample was separated with an ALB/IgG depletion column, and the fractions (flow-through and eluent) were subjected to SDS-PAGE and in-gel digestion (8 bands), followed by multistep fractionation methods (with a strong-anion exchanger and C18 reversed-phase fractionation) at the peptide level. The other portion was subjected directly to in-solution digestion or PLRP-S column chromatography (8 fractions) at the protein level. In-solution-digested samples were further subjected to the fractionation method of high-pH reverse fractionation (15 fractions) at the peptide level, and each PLRP-S fraction was individually digested with trypsin in solution. A total of 50 LC-MS/MS runs were completed, and the total LC-MS/MS running time was more than 100 h

Fig. 3

Polarized primary human fetal RPE (hfRPE) cell culture model. Cells were exposed to 4-hydroxy-2-nonenal (4-HNE) for 24 h. (a; upper) Phase-contrast microphotographs and microphotographs of the immunofluorescence (IF) staining of hfRPE cultures. (a; lower) The expression of RPE65, a RPE-specific protein, was confirmed through Western blot analysis of the lysate of hfRPE cells. (b) Quantification of transepithelial resistance (TER). (c) Expression of the epithelial marker E-cadherin and the mesenchymal marker vimentin in hfRPE cells exposed to 4-HNE. (d) Immunostaining of drusen-related proteins, APOA1, cathepsin D, clusterin and CFH in hfRPE cells. (e) Western blot analysis showing increased expression of APOA1, cathepsin D, and clusterin in 100 μM 4-HNE-treated hfRPE cells compared to that in controls

Fig. 4

Gene ontology analysis (localization) of the aqueous humor (AH) proteome of dry AMD patients with drusen and reticular pseudodrusen (RPD)

Fig. 5

(a) Principal component analysis (PCA) of quantified AH proteins from drusen, RPD, and healthy control samples. (b) 2D-plot of the relative abundance of AH proteins quantified from drusen and RPD patients versus healthy controls. Upregulated proteins are shown in the upper right quadrant and downregulated proteins in the lower left quadrant for both drusen and RPD samples. The dashed line indicates the boundary of changes in abundance according to the 2-fold criterion. Proteins indicated with boxes are new biomarker candidates in dry AMD patients

Fig. 6

Differential secretion of selected proteins in hfRPE cell cultures. Total protein concentrations in conditioned media from apical and basal baths (left end). To confirm the biological activity of VEGF, PEDF, and CFH, apical or basal media were collected at 24 h and then analyzed using ELISAs (the rest three graphs). The amount of VEGF protein in the basal media was higher than that in the apical media (2.86 ng/mL vs. 1.51 ng/mL) and was greatly decreased by 100 μM 4-HNE treatment (0.74 ng/mL vs. 0.51 ng/mL). PEDF secretion on the apical side was higher than that on the basal side (4.51 μg/mL vs. 1.74 μg/mL), and the level of PEDF was significantly decreased in a dose-dependent manner at 24 h (0.086 μg/mL and 0.32 μg/mL, in apical and basal baths treated with 100 μM 4-HNE). The ratio of the PEDF concentrations secreted in the apical and basal chambers was reversed in cultures treated with 100 μM 4-HNE. The apical/basal ratios were 2.59 and 1.83 in control cultures and cultures exposed to 50 μM 4-HNE for 24 h; the apical/basal ratio was 0.27 in cultures exposed to 100 μM 4-HNE for 24 h. The level of CFH showed similar changes: it decreased both apically and basally in cultures exposed to oxidative stress compared to control cultures. The apical/basal ratios were 4.19 and 2.64 in control cultures and cultures exposed to 50 μM 4-HNE for 24 h, respectively; the apical/basal ratio was 1.41 in cultures exposed to 100 μM 4-HNE for 24 h