The orphan nuclear hormone receptor ERRbeta controls rod photoreceptor survival

Abstract

Mutation of rod photoreceptor-enriched transcription factors is a major cause of inherited blindness. We identified the orphan nuclear hormone receptor estrogen-related receptor beta (ERRbeta) as selectively expressed in rod photoreceptors. Overexpression of ERRbeta induces expression of rod-specific genes in retinas of wild-type as well as Nrl(-/-) mice, which lack rod photoreceptors. Mutation of ERRbeta results in dysfunction and degeneration of rods, whereas inverse agonists of ERRbeta trigger rapid rod degeneration, which is rescued by constitutively active mutants of ERRbeta. ERRbeta coordinates expression of multiple genes that are rate-limiting regulators of ATP generation and consumption in photoreceptors. Furthermore, enhancing ERRbeta activity rescues photoreceptor defects that result from loss of the photoreceptor-specific transcription factor Crx. Our findings demonstrate that ERRbeta is a critical regulator of rod photoreceptor function and survival, and suggest that ERRbeta agonists may be useful in the treatment of certain retinal dystrophies.

Fig. 1.

ERRβ is expressed in rod photoreceptors and promotes rod-specific gene expression. (A) Developmental expression of ERRβ (green) coimmunostained with the horizontal cell marker Lhx1 (red) at P0, 4, 7, 10, 14, and 35. Expression overlaps in immature horizontal cells at P7 (white arrowheads). ONBL, outer neuroblastic layer; INBL, inner neuroblastic layer; ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. (B) Double immunostaining of ERRβ (green) and S-opsin (red) with DAPI (blue) on a P40 retinal section, demonstrating nonoverlapping expression of the two markers. OONL, outer outer nuclear layer. (C–E) Section immunohistochemistry of P14 retinas electroporated in vivo at P0 with CAG-GFP or CAG-ERRβ colabeled with Rho4D2 (C and D) or calbindin (E) antibodies. (F) Retinal cell-type composition of P14 retinas electroporated in vivo at P0 with CAG-GFP and CAG-ERRβ estimated by dissociated immunohistochemistry using cell-specific markers (Rho4D2 for rods, calbindin for horizontal cells, PKCα for rod bipolar cells, syntaxin for amacrine cells, and glutamine synthetase for Muller glia). All data are represented as mean ± SD. *P < 0.05, **P < 0.005 by Student's t test (n = 3 electroporated retinas). (G–J) ERRβ overexpression is sufficient to induce rhodopsin expression in photoreceptors of Nrl^−/− mice. Section (G) and flat-mount (H) immunohistochemistry of P14 retinas electroporated in vivo at P0 with CAG-GFP and CAG-ERRβ immunostained with S-opsin (blue) and Rho4D2 (red) antibodies. Asterisks indicate background staining of blood vessels. (I) Quantification of the fraction of Rho4D2/GFP double-positive photoreceptor cells from (H). All data are represented as mean ± SD (n = 3 electroporated retinas). (J) Quantification of rhodopsin and GAPDH mRNA in P7.5 Nrl^−/− retinas electroporated with CAG-ERRβ as measured by quantitative RT-PCR (n = 4 RNA preparations). Fold change in mRNA levels is expressed relative to the contralateral retina, which was not electroporated. Levels of mRNA are normalized relative to the fraction of electroporated cells (≈10% of the retinal surface, and 50% of photoreceptors within that area, were electroporated in the samples used). (K) ERRβ directly regulates rhodopsin −2174 - - +70 bp promoter construct activity. Black and white bars represent luciferase levels relative to untransfected 293T cells expressed from wild-type and a mutant promoter construct, in which the ERRβ-binding motif (AGGTCA) at the proximal promoter (-136 bp) has been mutated. Pink and green triangles represent dose treatment of DY131 and 4-OHT, respectively. + and ++ represent 15 and 30 ng, respectively. All data are represented as mean ± SD (n = 3 independent transfections). *P < 0.05 by Student's t test (n = 3 independent transfections, repeated in two separate trials). (Scale bars, 20 μm.)

Fig. 2.

ERRβ controls rod photoreceptor survival. (A and B) H&E staining (A) and section immunohistochemistry (B) labeled with M-opsin and S-opsin (green) and Rho4D2 (red) antibodies of aged (P21, 3, 6, 10, 18, and 24 mo) mouse retinas. (C and D) High-power images of photoreceptor outer and inner segments (C) and cone outer segments (D) from A and B, respectively. (E) Quantification of ONL thickness normalized relative to the INL from A. (F) Quantification of rod and cone outer segment length from C and D, respectively. All data are represented as mean ± SD (n = 20 from three different individuals). (G) Density of cone photoreceptors estimated from B. All data are represented as mean ± SD (n = 3 retinas from different individuals). (H) Scotopic (rod) and photopic (cone) ERG responses from 10-mo-old wild-type, ERRβ^+/−, and ERRβ^−/− mice. The amplitude of the scotopic ERG response of ERRβ^−/− mice was significantly reduced, whereas the wave forms of WT and ERRβ^+/− were similar. There were no significant differences in the amplitude of the photopic ERG response. (I) Scotopic a- and b-wave amplitudes of 11 flash intensities, and photopic b-wave amplitudes of 3 flash intensities. All data are represented as mean ± SD (n = 4 animals). (J) TUNEL staining (red) with DAPI (blue) of WT, ERRβ^+/−, and ERRβ^−/− retinas that were harvested at P14 and explanted for 2 d in vitro (div 2) or 4 d in vitro (div 4) d exposed to 0.1% EtOH (carrier), 10 μM DES, and 10 μM 4-OHT. (K) The fraction of TUNEL-positive cells in ONL in J normalized by the area of ONL. All data are represented as mean ± SD (n = 3 independent explants). *P < 0.05 by Student's t test. [Scale bars, 20 μm (A, B, and J) and 10 μm (C and D).]

Fig. 3.

Quantitative RT-PCR analysis of transcripts down-regulated in P21 ERRβ^−/− mice. RNA from whole dissected retinas was used, and the results were normalized to β-actin levels. Signal levels represent mRNA levels in mutant retinas normalized to those of age-matched wild-type retinas (n = 3 RNA preparations). Sequences for primers used for are shown in Table S2.

Fig. 4.

ERRβ overexpression and the ERR-specific agonist DY131 rescue defects in rod function in Crx^−/− mice. (A) Retinal sections of P33 Crx^−/− retinas electroporated in vivo at P0 with CAG-GFP and CAG-ERRβ that were TUNEL-positive (A) and immunostained with Rho4D2 (B). The fractions of TUNEL-positive cells and Rho4D2-positive electroporated cells in ONL are shown in the lower panel. (C) TUNEL staining of P7 + 5 d in vitro (div5) Crx^−/− retinal explants exposed to 0.1% EtOH (carrier) and 10 μM DY131. The fraction of TUNEL-positive cells in ONL was normalized by the area of ONL. All data are represented as mean ± SD (n = 3). (D) Dissociated cell immunohistochemistry using P7 + div5 Crx^−/− retinal explants. The fraction of Rho4D2-positive cells is shown. All data are represented as mean ± SD (n = 3 retinas). *P < 0.05 by Student's t test. (Scale bars, 20 μm.)