Retinal pigment epithelial cell multinucleation in the aging eye - a mechanism to repair damage and maintain homoeostasis

Abstract

Retinal pigment epithelial (RPE) cells are central to retinal health and homoeostasis. Dysfunction or death of RPE cells underlies many age-related retinal degenerative disorders particularly age-related macular degeneration. During aging RPE cells decline in number, suggesting an age-dependent cell loss. RPE cells are considered to be postmitotic, and how they repair damage during aging remains poorly defined. We show that RPE cells increase in size and become multinucleate during aging in C57BL/6J mice. Multinucleation appeared not to be due to cell fusion, but to incomplete cell division, that is failure of cytokinesis. Interestingly, the phagocytic activity of multinucleate RPE cells was not different from that of mononuclear RPE cells. Furthermore, exposure of RPE cells in vitro to photoreceptor outer segment (POS), particularly oxidized POS, dose-dependently promoted multinucleation and suppressed cell proliferation. Both failure of cytokinesis and suppression of proliferation required contact with POS. Exposure to POS also induced reactive oxygen species and DNA oxidation in RPE cells. We propose that RPE cells have the potential to proliferate in vivo and to repair defects in the monolayer. We further propose that the conventionally accepted 'postmitotic' status of RPE cells is due to a modified form of contact inhibition mediated by POS and that RPE cells are released from this state when contact with POS is lost. This is seen in long-standing rhegmatogenous retinal detachment as overtly proliferating RPE cells (proliferative vitreoretinopathy) and more subtly as multinucleation during normal aging. Age-related oxidative stress may promote failure of cytokinesis and multinucleation in RPE cells.

Keywords: aging; cytokinesis; multinucleation; phagocytosis; photoreceptor outer segments; retinal pigment epithelium.

Figure 1

RPE cells in mice of different ages. RPE/choroid/sclera flat mounts were stained with phalloidin (for F‐actin, green) and PI (red) and imaged by confocal microscopy. (A) a schematic graph showing different geographic locations of RPE flat mounts used in image analysis. (B–D) typical confocal images of RPE flat mounts from a 6‐month‐old mouse showing RPE cells in the peripheral (B), equatorial (C) and central (D) regions. (E–G) the number of RPE cells and the number of RPE nuclei in different regions of the eye from different ages of mice. *, P < 0.05; **, P < 0.01, ***, P < 0.001 compared to cell number at the same age time point. †, P < 0.05; ††, P < 0.01 ††††, P < 0.001 compared to the cell number of the 3 m age group. N ≥ 8.

Figure 2

RPE cell morphological changes in aged mice. RPE/choroid/sclera flat mounts from 18‐month (A, B)‐ and 24‐month(C–D)‐old mice were stained with phalloidin (green) and PI (red) and imaged by confocal microscopy. (A) a F‐actin^hi lesion (arrow) is surrounded by 8 RPE cells, and a few of the cells have an appearance of cell body extension. This type of lesion was frequently observed in mice older than 6 months. (B) a confocal image from a 18‐month‐old mouse shown the transition between areas with normal RPE cells (lower left) and giant RPE cells (upper right). A few pigmented cells (arrowheads) were seen on RPE surface. (C–D) high‐magnification images showing giant RPE cells with diffused PI staining and multiple small F‐actin⁺ intracellular vacuoles (small arrows). The vacuoles appear to be connected with cytoplasmic membrane in z‐stack images (D). Many pigmented cells were observed on the RPE surface (arrowheads in C). E, a giant RPE cell with multiple nuclei in a 24‐month‐old mouse.

Figure 3

Detection of α‐tubulin⁺ RPE cells and BrdU⁺ RPE cells in mice of different ages. (A–D) RPE/choroid/sclera flat mounts from 1‐week(A)‐ and 12‐week‐old (B, C) mice were stained with phalloidin (red) and α‐tubulin (green) and imaged by confocal microscopy. An α‐tubulin‐expressing RPE cell is shown in A (asterisk), and α‐tubulin‐expressing tubular structures was detected in the peripheral (B) and central (C) RPE flat mounts in 12‐week‐old mice. (D) histogram shown the number of α‐tubulin expressing RPE cells at different ages. **, P < 0.01 compared to 1‐week‐old mice, #, value not detected. N = 6 mice. E‐H, 3‐ and 12‐month‐old mice were injected with BrdU for 7 days. RPE flat mounts were stained for BrdU (green) and PI (red) and examined by confocal microscopy. (E–G) confocal images from 12‐month‐old mice showing BrdU⁺ cells in the peripheral (E), equatorial (F) and central (G) regions. H, histogram showing the total number of BrdU⁺ cells in each eye. **, P < 0.01 compared to 3‐month‐old mice, unpaired Student's t‐test. N = 6 mice.

Figure 4

The effect of photoreceptor outer segment (POS) on RPE cell proliferation and multinucleation. B6‐RPE07 mouse RPE cells were treated with different concentrations of POS or oxidized POS (oxPOS) or latex beads for 48 h. (A) cell proliferation was detected by MTT assay. *, P < 0.05 compared to control group, One‐way ANOVA followed by Dunnett's multiple comparison test. N = 3. (B–F) following treatment, RPE cells were fixed with ethanol and stained for α‐tubulin and PI. The cells were imaged by confocal microscopy. (B) confocal image from control nontreated cells. (C) confocal image from oxPOS (5:1)‐treated cells. (D) confocal image from latex beads (5:1)‐treated cells. (E) a high‐magnification image shown a RPE cell with 6 nuclei following oxPOS treatment. (F) histogram shown the percentage of binucleated and multinucleated RPE cells following different treatments. **, P < 0.01, ***, P < 0.001 compared to controls. One‐way ANOVA followed by Dunnett's multiple comparison test. N = 3. (G) Growth curve of B6‐RPE07 mouse RPE cells with and without ox‐POS treatment. B6‐RPE07 cells were cultured in 24‐well plates at 1 × 10⁴ ml^‐1. Twenty‐four hours later (day 1), one group of cells was treated with 5 × 10⁴ ml⁻¹ ox‐POS for 24 h. At day 2, ox‐POS was removed from the culture. Cell numbers from each group were counted from three wells each day.

Figure 5

Mechanism of RPE multinucleation in vitro. (A) ARPE19 cells labelled with MitoTrackerRed or CFSE were mixed (1:1) and then treated with oxPOS (5:1) for 48 h, the cells were then fixed and stained for DAPI. Arrows indicate MitoTracker⁺ multinucleate RPE cells, and arrowhead indicates CFSE ⁺ multinucleate RPE cells. (B) ARPE19 cells were treated with oxPOS (5:1) for 48 h and stained for phalloidin, γ‐tubulin and DAPI. Arrows indicate a multinucleate cell with three centrosomes. (C) Control and blebbistatin‐treated ARPE19 cells were stained for phalloidin (green) and PI (red) and observed by confocal microscopy. 48% of blebbistatin‐treated cell were multinucleated. **, P < 0.01, unpaired Student's t‐test. N = 3. (F–H) CellRox Green staining in control (F), POS (G)‐ and oxPOS (H)‐treated ARPE19 cells. Green – CellRox Green, Red – MitoTracker, Blue – Hoechst 33342. I, Fluorescence intensity of CellRox Green in different groups of cells. (J–L) 8‐OHdG expression in control (J), oxPOS (K)‐ and UV light (L)‐ treated ARPE19 cells. Green – 8‐OHdG, red – phalloidin(F‐actin), blue – DAPI. M, Fluorescence intensity of 8‐OHdG in different groups of cells. *, P < 0.05; **, P < 0.01; ***, P < 0.001, n = 50 cells, Tukey's multiple comparison test.

Figure 6

A Model of RPE cell repair during aging. (A) when a single RPE cell is damaged, adjacent cells expand in size and migrate towards the lesion site to repair the damage. (B) when many more cells are damaged over a sustained period of time, the remaining cells need to expand in size more extensively, sometimes up to 2–5 times of their original size to repair damage. The workload (e.g. phagocytize POS, transport nutrients and oxygen) of these enlarged hypertrophic RPE cells is 2–5 times more than their original workload. Multinucleated cell may cope with the substantial increase in cell volume and to maintain homoeostasis, but it is likely that this response is less efficient than cell replication. The stressed multinucleate cells or hypertrophic single nuclear cells may be at greater risk of cell death lead to patches of RPE loss (geographic atrophy).