Epithelial-Mesenchymal Transition and Senescence in the Retinal Pigment Epithelium of NFE2L2/PGC-1α Double Knock-Out Mice

Abstract

Age-related macular degeneration (AMD) is the most prevalent form of irreversible blindness worldwide in the elderly population. In our previous studies, we found that deficiencies in the nuclear factor, erythroid 2 like 2 (NFE2L2) and peroxisome proliferator-activated receptor gamma coactivator 1-α (PGC-1α) genes caused AMD-like pathological phenotypes in mice. In the present work, we show hijacked epithelial-mesenchymal transition (EMT) due to the common loss of PGC-1α and NFE2L2 (double knock-out, dKO) genes in aged animals. The implanted area was assessed by histology, immunohistochemistry and transmission electron microscopy. Confocal microscopy revealed altered regions in the filamentous actin ring. This contrasted with hexagonal RPE morphology in wild-type mice. The ultrastructural RPE features here illustrated loss of apical microvilli, alteration of cell-cell contact, loss of basal in-folding with deposits on Bruch's membrane, and excessive lipofuscin deposition in dKO samples. We also found the expression of epithelial-mesenchymal transition transcription factors, such as Snail, Slug, collagen 1, vimentin and OB-cadherin, to be significantly different in dKO RPEs. An increased immunoreactivity of senescence markers p16, DEC1 and HMGB1 was also noted. These findings suggest that EMT and senescence pathways may intersect in the retinas of dKO mice. Both processes can be activated by damage to the RPE, which may be caused by increased oxidative stress resulting from the absence of NFE2L2 and PGC-1α genes, important for antioxidant defense. This dKO model may provide useful tools for studying AMD pathogenesis and evaluating novel therapies for this disease.

Keywords: AMD; NFE2L2/PGC-1α dKO mice; age-related macular degeneration; epithelial-mesenchymal transition; senescence.

Figure 1

Light micrographs of toluidine stained semithin retinal sections derived from one-year-old wild-type (WT) and NFE2L2/PGC-1α dKO mice. Retinal cross sections were prepared from one-year-old WT control mice (A,C) and dKO samples (B,D) and plastic sections were stained with toluidine blue. Boxed areas are shown with high power in (C,D). The RPE cell layer from WT animals is thicker, with a higher number of nuclei (arrowheads in (C)) compared to age-matched dKO littermates (arrowheads in (D)). In the NFE2L2/PGC-1α dKO mice RPE layer is flattened. Melanin granules tend to occupy more of their cytoplasm. (E) illustrates the decrease in the density of RPE nuclei for double knockout animals, relative to the age-matched WT littermates (n = 30). ** p < 0.001 (Mann-Whitney U-test). Abbreviations: GCL: ganglion cell layer; IPL: inner plexiform layer; INL: inner nuclear layer; OPL: outer plexiform layer; ONL: outer nuclear layer; PIS: photoreceptor inner segment; POS: photoreceptor outer segment; RPE: retinal pigment epithelium. Scale bars: 5 µm.

Figure 2

Structural changes of the apicolateral retinal pigment epithelium (RPE) cell–cell junctions in 1-year-old NFE2L2/PGC-1α dKO mice as compared with their wild-type (WT) counterparts. The pictures shown in (A,C) are from phalloidin-based immunofluorescent flat mount confocal microscopy, while transmission electron microscopy (TEM) images are presented in (B,D). Phalloidin was used to visualize filamentous actin rings on flat-mount samples and indirectly examine the cellular density and arrangement. The WT retinal epithelium consists of typical hexagonal cells (A), while dKO RPE cells are, by contrast, less regular in their outline. While cellular density can vary somewhat, regions of the WT RPE display higher cellular density. In 1-μm-thick confocal sections of phalloidin-labeled tissues, apical cell boundaries are marked by rings of actin filaments associated with intercellular junctions. Yellow arrowheads indicate multiple locations of the altered phalloidin staining for the apical aspect of the RPE complex. Inserts in (B,D) illustrate the RPE apicolateral zone and the high power TEM images show the profile of lateral RPE cell–cell contact (black (B) and red (D) arrows). Scale bars: 20 µm (A,C) and 1 µm (B,D).

Figure 3

Structural changes of the basal compartment of the retinal pigment epithelium (RPE) in NFE2L2/PGC-1α double knockout (dKO) mice as compared to their age-matched wild-type (WT) counterparts. Representative high power transmission electron microscopy (TEM) images of the samples of three-months- (A,C) and one-year-old (B,D) WT and dKO animals are shown. RPE of WT mice at both time points (A,B) and the three-months-old dKO samples (C) show normal basal arrangement: well-organized infoldings above Bruch’s membrane, supporting intracellular organelles (infoldings indicated by arrowheads) and typical thickness of Bruch’s membrane (BM). In contrast, many of the one-year-old NFE2L2/PGC-1α dKO mice RPEs show ultrastructural signs of loss of the basal infoldings with basal linear deposits (D; arrow), resulting in unsupported intracellular organelles above Bruch’s membrane. Scale bars: 1 µm.

Figure 4

Loss of apical microvilli and flattened RPE in one-year-old NFE2L2/PGC-1α dKO mice as compared with their age-matched wild-type (WT) littermates. Panel 4 A shows the apical portion of WT RPE, where microvilli (arrowhead) are arranged normally, surrounding the photoreceptor outer segments (POS). In contrast, many of the apical portion of RPEs from dKO mice showed loss of apical microvilli: only a few microvilli on the apical surface were evident (B; arrow), with an increased density of melanin granules. Total RPE height from TEM graphs was measured in the region located 300 µm from the optic nerve head (C). Combined results from n = 6 animals, ** p < 0.001 (Mann-Whitney U-test). Scale bars: 1 µm (A,B).

Figure 5

Confocal microscopy analysis of the immunoreactivity of mesenchymal markers in the retinal pigment epithelium (RPE) of one-year-old NFE2L2/PGC-1α dKO mice as compared with their wild-type (WT) counterparts. Semiquantitative comparative densitometric analysis done by ImageJ of the immunoreactivity of Slug, Snail, vimentin, OB-cadherin (OB-CAD) and collagen-1 (COL1) in 1-μm parasagittal sections of RPE. Alexa Fluor secondary antibodies were used to detect specific mesenchymal markers. DAPI was used to stain the nuclei of RPE cells. Scale bar: 5 μm. The number of puncta specific to a particular marker is presented in plots in the lower panels. Twenty-five pictures were analyzed for each animal. * p < 0.05, ** p < 0.001 (Mann-Whitney U-test).

Figure 6

Confocal microscopy of lipofuscin (LF) autofluorescence and quantitative TEM analysis of lipofuscin granules in wild-type (WT) and dKO RPE. Sections from WT and dKO mice were analyzed by fluorescence microscopy. Arrowheads point to LF granules within RPE cytoplasm of one-year-old WT (A) and dKO mice (B). In contrast to WT, dKO RPE showed increased levels of LF fluorophores. Transmission electron microscopy was used for quantitative analysis of polymorphic lipofuscin granules in WT (C) and dKO (D) samples. Considerable accumulation of the number of LF granules was detected in dKO samples. The lipofuscin granules in the cytoplasm of RPE and decrease of organelles were observed; most LF granules resulted to be completely occupied by strongly electron-dense substance (E). The average area of LF granules increased within the RPE of dKO samples per 100 μm² area measured from electron micrographs (F). ** p < 0.001 (Mann-Whitney U-test). Scale bars: 5 μm (A,B), 0.5 µm (C,D).

Figure 7

Confocal microscopy analysis of the immunoreactivity of the senescence markers p16, DEC1 and HMGB1 in the retinal pigment epithelium (RPE) of one-year-old NFE2L2/PGC-1α dKO mice as compared with their WT counterparts. Semi-quantitative comparative densitometric analysis performed by ImageJ of the immunoreactivity of the markers in 1-μm parasagittal sections of RPE. Alexa Fluor secondary antibodies were used to detect senescence markers. DAPI was used to stain the nuclei of RPE cells. Scale bar: 5 μm. Arrows indicate staining of senescence markers in the RPE. ** p < 0.001 (Mann-Whitney U-test).

Figure 8

ROS-induced EMT-triggered signal propagation and transduction: an intricate interplay in the RPE of NFE2L2/PGC-1α double knockout model enhances epithelial-mesenchymal transition (EMT). Simplified representation of the oxidative stress (OS)-dependent major molecular mechanisms of EMT. Reprogramming of gene expression during EMT, and non-transcriptional changes, are initiated and controlled by signaling pathways that respond to extracellular cues. Here, the arrows pointing from the signaling routes to the EMT transcription factors in the nucleus are grouped. EMT involves a functional transition of highly polarized retinal epithelial cells into ECM component-secreting mesenchymal-type flattened cells. Identification of all mesenchymal cells originating from the RPE via EMT is hardly possible as many mesenchymal cells display some epithelial markers while a transition is not fully completed.