Amelotin is expressed in retinal pigment epithelium and localizes to hydroxyapatite deposits in dry age-related macular degeneration

Abstract

Deposition of hydroxyapatite (HAP) basal to the retinal pigment epithelium (RPE) is linked to the progression of age-related macular degeneration (AMD). Serum-deprivation of RPE cells in culture mimics some features of AMD. We now show that serum-deprivation also leads to the induction of amelotin (AMTN), a protein involved in hydroxyapatite mineralization in enamel. HAP is formed in our culture model and is blocked by siRNA inhibition of AMTN expression. In situ hybridization and immunofluorescence imaging of human eye tissue show that AMTN is expressed in RPE of donor eyes with geographic atrophy ("dry" AMD) in regions with soft drusen containing HAP spherules or nodules. AMTN is not found in hard drusen, normal RPE, or donor eyes diagnosed with wet AMD. These findings suggest that AMTN is involved in formation of HAP spherules or nodules in AMD, and as such provides a new therapeutic target for slowing disease progression.

Fig. 1.

Serum deprived ARPE-19 monolayers express Amelotin (AMTN) mRNA and protein at day nine. A. Diagram of the transwell cell culture system of ARPE-19 cell monolayers to mimic RPE in vivo. B. Confluent ARPE-19 monolayers showing polygonal cells with organized packing geometry, and localized immunofluorescence for known RPE-expressed proteins. C. Mean counts of total AMTN mRNA expression from RNA sequencing at day 0 (10% serum) and days 1, 3, 4, 5, 6 and 9 of serum deprivation. Data displayed are normalized mean counts for the 3 biological replicates from each time and n = 3 technical replicates. D. Representative western blot of AMTN (22 kDa) expression from cell lysate from ARPE-19 cultured in 6-well plates at day 0 (10% serum) and days 3, 5 and 9 of serum deprivation with 2 biological replicates from each time. Blots are representative of n = 3. E. In situ hybridization image of an ARPE-19 monolayers cultured in chambers at days 0 with 10% serum showing sparse AMTN mRNA detection, n = 3. F. In situ hybridization image of an ARPE-19 monolayers cultured in chambers at day 9 in serum free media showing abundant AMTN mRNA detection, n = 3. G. Changes in gene expression for genes annotated with bone and mineral associated genes with P value < 0.05 during serum deprivation time course of ARPE-19 cells as determined by RNA-Seq. Each tile represents a gene. The columns are ordered by days of serum deprivation. Genes with increased expression appears above the 0-reference line in Orange. Genes with decreased expression appears below the 0-reference line in Blue. The saturation of the color represents the magnitude of the change.

Fig. 2.

Serum deprivation increases calcium phosphate deposition in ARPE-19 cells. A. Images showing changes in Alizarin red (calcium phosphate) staining in ARPE-19 cells cultured in 6 well plates at day 0 in 10% serum and days 6, and 9 with serum deprivation, n = 3. B. Calcium phosphate quantification assay measuring the Alizarin red absorbance correlated to calcium deposition in cells at day 0 in 10% serum and serum deprived cells at days 6, and 9. One-way ANOVA with post-hoc Dunnett’s test. Mean ± SD. Day 0 vs Day 6 adjusted P = 0.0008, Day 0 vs. Day 9 adjusted P < 0.0001; n = 3. C. Alizarin red staining of cells showing higher levels of positive nodules for nine days serum free media (SFM) and SFM + 50 mg/ml rh-AMTN compared to cells cultured in 10% serum. D. Quantification of Alizarin red staining positive nodules after nine days SFM or SFM + 50 mg/ml rh-AMTN compared to cells cultured in 10% serum. One-way ANOVA with post hoc Tukey’s multiple comparisons test. Mean ± SD. 10% Serum vs. Day 9 SFM adjusted P = 0.0007, 10% Serum vs. Day 9 SFM + 50 mg/ml rh-AMTN adjusted P = 0.0003, Day 9 SFM vs. Day 9 SFM + 50 mg/ml rh-AMTN adjusted P = 0.0186; n = 3. E. Light-scatter plot of mineralization buffers containing 0 to 100 μg/ml rh-AMTN. Buffer vs Days shows a correlation of +0.6938 and P = 0.0179; Days vs 5μg/ml shows a correlation of +0.9821 and P <0.0001; Days vs 20mg/ml shows a correlation of +0.8766 and P = 0.0010; Days vs. 40 μg/ml shows a correlation of +0.9518 and P < 0.0001; Days vs 100 μg/ml shows a correlation of +0.9379 and P < 0.0001. The mineral-promoting effect of rh-AMTN is dose-dependent. F-G. Mineral precipitates from the mineralization buffer containing 100 mg/ml rh-AMTN after 4 days of incubation at 37°C. H. Mineral precipitates from the mineralization buffer containing 100 mg/ml rh-AMTN after 4 days of incubation at 37°C stained with Alizarin red. I-J. Buffer only wells after 4 days of incubation at 37°C. High-resolution light microscopy imaging of HAP mineral structure consisting of needle-like crystallites. K. Confocal images showing AMTN protein expression (green) in cross-sections of ARPE-19 transwell culture monolayers for 6 weeks in 10% serum and nine days SFM, n = 3. L. Confocal images showing accumulation of HAP (bone-tag680RD-magenta) positive sub-RPE deposits in cross-sections of ARPE-19 transwell culture monolayers for 6 weeks in 10% serum and nine days SFM, n = 3. (Color version of figure is available online.)

Fig. 3.

Serum-deprived ARPE-19 cells form HAP basal deposits which are blocked by siRNA knock-down AMTN (AMTN-KD). A. Quantitative PCR analysis of AMTN gene expression from 9 days SFM, 9 days SFM + AMTN siRNA, 9 days SFM + GAPDH siRNA, and 9 days SFM + nontargeting siRNAs. Bars represent the ratio of the gene expression level compared to the levels in cells cultured in nine days SFM. All measures normalized to the average values of ABCF3 genes. Error bars represent standard deviation. n = 3. B. Representative western blots and quantification of AMTN expression from ARPE-19 cell lysates from cells cultured in 6-well plates with 10% serum, 9 days SFM and both 10% serum and 9 days SFM cells transfected with AMTN siRNA or GAPDH siRNA. Bands were detected for AMTN at 22 KDa and b-actin at 40 KDa. Western blot signals were quantified using ImageJ software (version 1.45; National Institutes of Health, Bethesda, MD). *, P < 0.05; **, P < 0.01. One-way ANOVA followed by Tukey’s multiple comparison test, n = 3. C. In situ hybridization images showing change in AMTN mRNA (green) and protein (blue) expression in 10% serum, nine days SFM, nine days SFM + AMTN-KD cells and 9 days SFM + GAPDH siRNA. D. Cross-section image showing AMTN protein (green) expression in of ARPE-19 cells grown in 9 days SFM and AMTN-KD ARPE-19 cells grown in 9 days SFM with diminished AMTN protein expression (green), n = 3. E. Cross-section images showing accumulated HAP deposit (bone-tag680RD-magenta) in of ARPE-19 cells grown in 9 days SFM and AMTN-KD ARPE-19 cells grown in 9 days SFM with diminished HAP accumulation, n = 3. F. Cross-section images showing EFEMP-1/Fibulin3 protein (green) expression in ARPE-19 cells in serum free media and in SFM + AMTN-KD. G. Cross-section images showing ApoB protein (red) expression in of ARPE-19 cells in serum free media and in SFM + AMTN-KD. (Color version of figure is available online.)

Fig. 4.

AMTN is expressed in RPE near soft drusen in AMD eyes. A-B. Images showing AMTN mRNA (red) in regions of RPE surrounding drusen. Blue * indicate drusen. C. Image showing AMTN mRNA (red) in a region of RPE without drusen in an AMD eye. D. Image of a human cadaver eye diagnosed with wet AMD negative for AMTN mRNA staining. E. Image of a normal human cadaver eye negative for AMTN mRNA staining. F. Positive control image of a human cadaver AMD eye hybridized with EFEMP1 mRNA (red) in the RPE. Black arrows indicate regions with positive in situ mRNA labeling. G. Image of an AMD eye with drusen and RPE positive for AMTN protein (green) staining. H. Image of another AMD eye with drusen and RPE positive for AMTN protein (green) staining. White * indicates drusen. I. Image of a normal eye negative for AMTN protein (green) staining in RPE. White arrows indicate regions of positive AMTN protein labeling. Cryosections were treated with TrueBlack Lipofuscin Autofluorescence Quencher (Biotium, Inc, Fermont CA) prior to immunofluorescence staining to obtain true-positive fluorescence signal in RPE. Cryosections stained with; nuclei: DAPIblue; photoreceptors: PAN-magenta; AMTN-green; autofluorescence-assigned cyan. PR, photoreceptors; RPE, Retinal pigment epithelium. (Color version of figure is available online.)

Fig. 5.

Colocalization of HAP spherules and nodules and AMTN. A, E. Autofluorescence images of individual druse (due to overlapping emission spectra, and autofluorescence was assigned cyan for clarity of images). B, F. Xylenol Orange labeling. C, G. Alizarin Red S labeling. D, H. Combined images. I, J. AMTN labeling of HAP spherules of another druse in 2 different slices of a Z-image series with black background. Autofluorescence was assigned blue. K, L. AMTN labeling of HAP spherules of a druse in 2 different slices of a Z-image series with phase-contrast background. M. Bone-Tag 680 RD labeling of a large drusen with HAP spherules and nodules. N. AMTN labeling of the large drusen. O. Combined image of AMTN and Bone-Tag HAP labeling. P-R. Magnified images of AMTN and Bone-Tag 680 RD co-labeling of HAP spherules and nodules isolated from soft drusen. RPE, Retinal pigment epithelium; BrM, bruch’s membrane. (Color version of figure is available online.)

Fig. 6.

OCT images of heterogeneous internal reflectivity in drusen and 3-dimensional confocal images of AMTN labelling of nodules in the calcified druse. OCT imaging of an eye from an 80-year-old Causcasian female with advanced AMD imaged in vivo2 months prior to death and super-resolution confocal imaging of the druse corresponding to the OCT postmortem. A. Near-infrared (NIR) image shows hyper-reflective large drusen. The Yellow line showing the level of OCT scan. B-E. OCT images corresponding to 1st, 2nd, 4th and 5th Yellow lines from the NIR image. F. OCT image showing heterogeneous internal reflectivity in drusen. Calcified lesions in OCT are highlighted with red frames. G. Magnified area of the calcified druse appear hyperreflective on NIR, loss of hyper-reflective RPE and surrounding hyper-reflective dots on OCT. H-I. Druse corresponding to the OCT. Three-dimensional volume rendered super-resolution confocal images of AMTN positive labeling of the calcified druse. Calcified lesions in OCT are highlighted with red frames. J. OCT of the eye of a 79-year old Caucasian female with advanced AMD imaged in vivo1 year and 10-months prior death. K. AMTN labeling (green) of the large calcified drusen from the same eye postmortem. L. OCT of the eye of an 89-yearold Caucasian male with advanced AMD imaged in vivo 5 years prior death. M. AMTN labeling (green) of the large calcified drusen from the same eye postmortem. (Color version of figure is available online.)