Distinct Functions of Photoreceptor Cell–Specific Nuclear Receptor, Thyroid Hormone Receptor β2 and CRX in Cone Photoreceptor Development

Abstract

purpose. To clarify the functions of a specific subtype of thyroid hormone receptor (TR), TRβ2, and photoreceptor cell–specific nuclear receptor (PNR) in the development of cone photoreceptors.

methods. The expression of short (S)- and medium (M)-wavelength cone opsins was analyzed by reverse transcription polymerase chain reaction (RT-PCR) and Northern blot analysis in mice without a functional PNR (rd7/rd7 mice), and levels of plasma thyroid hormones and expression of TRβ2 were also examined. Concomitantly, by means of reporter assays, the roles of PNR and TRβ2 in the S- and M-cone opsin expression were explored at the transcriptional level.

results. In rd7/rd7 mice, an abnormal increase in cone photoreceptors was observed immediately before retinal maturation normally occurs. Although an increase in S-cone opsin in the retina was observed during and after retinal development, the expression of M-cone opsin expression was not perturbed during retinal maturation. Plasma concentrations of thyroid hormone and levels of TRβ2 expression in the rd7/rd7 mouse retina over the developmental period were normal. Transcriptional studies demonstrated that TRβ2, but not PNR, activated the M-cone opsin gene promoter function, while suppressing the S-cone opsin promoter function enhanced by CRX in a thyroid hormone–dependent manner.

conclusions. The results indicate that PNR may suppress proliferation of cone photoreceptor progenitor cells and that the regulation of S- and M-cone opsin gene expression is mediated by TRβ2 and CRX, but not by PNR. Thus, our results partly disclosed the molecular mechanism of cone photoreceptor development, highlighting the distinct functions of PNR and TRβ2.

Color vision is mediated by cone photoreceptors. In rodents, cone photoreceptors express two types of cone pigments: short (S)- and medium (M)-wavelength cone opsins. Recent studies have disclosed that at least two factors are involved in the development of cone photoreceptors. One is a specific subtype of the thyroid hormone receptor (TR) β2, a member of the nuclear receptor superfamily acting as a ligand-inducible transcriptional factor. Experiments in vitro have demonstrated that thyroid hormone promotes the differentiation of cone photoreceptors of rat, human, and chick retinal progenitor cells. ¹ Recent study in vivo has demonstrated that TRβ2-deficient mice manifest a selective loss of M-cone opsin and concomitant increase in photoreceptors expressing S-cone opsin. ² These studies indicate that cone photoreceptors have the potential to follow a default S-cone pathway and that TRβ2 is essential in the commitment to an M-cone identity. Another molecule involved in cone photoreceptor development is photoreceptor cell–specific nuclear receptor (PNR). ³ PNR is an orphan nuclear receptor expressed exclusively in photoreceptors and belongs to the same gene superfamily as TR. PNR acts as a constitutive transcriptional repressor in vitro. ³ Deletion of the PNR gene is responsible for rd7/rd7 mice, ⁴ in which retinal degeneration develops with an increase in the number of total cone photoreceptors and photoreceptors expressing S-cone opsin. ⁵ Missense mutations in PNR gene in humans cause enhanced S-cone syndrome (ESCS) ⁶ characterized by retinal degeneration with increased sensitivity to short-wavelength light, presumably due to an increase in the number of cone photoreceptors that express S-cone opsin. ⁶ Individuals with ESCS also manifest varying degrees of long (L)-wavelength and M-cone photoreceptor sensitivity, and a previous study has suggested that mutations of the PNR gene may allow photoreceptor precursors to retain an ancestral S-cone default commitment rather than follow a normal L/M-cone sequence. ⁶ Indeed, the retinal phenotypes of rd7/rd7 mice and patients with ESCS appear similar to those of TRβ2-deficient mice, but with certain differences, suggesting that PNR and TRβ2 have distinct functions in cone photoreceptor development in vivo. However, thus far the factors involved in cone photoreceptor development have not been well studied. 

To clarify the function of PNR in the development of cone photoreceptors, we analyzed the expression of S- and M-cone opsins in rd7/rd7 mice. Concomitantly, we explored the roles of PNR and TRβ2 in the expression of S- and M-cone opsin at the transcriptional level, with reporter assays. Our results indicate that PNR controls cone photoreceptor development, presumably affecting the proliferation of cone photoreceptors, but does not affect the TRβ2-mediated regulation of S- and M-cone opsin expression directly. 

Materials and Methods

Animals

Rd7/rd7 and C57BL/6J mice were obtained from The Jackson Laboratory (Portland, ME). All experiments were conducted in accordance with the Animal Care and Use Committee and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Histology

After the mice were killed by cervical dislocation, the eyes were immediately enucleated and prepared for light microscopic histology by immersing them in phosphate-buffered saline (PBS) containing 4% paraformaldehyde for 12 hours. They were then transferred into 70% ethanol and processed for paraffin embedding. Once embedded, 4.0-μm sections of tissue were prepared for staining with hematoxylin and eosin. 

RNA Isolation and Northern Blot Analysis

Total RNA was extracted from the eyes of mice with extraction reagent (Isogen; Nippon Gene, Tokyo, Japan) according to the manufacturer’s instructions. The RNA (100 μg) was separated by electrophoresis on 1% agarose, 1.1-M formaldehyde gels and transferred to nitrocellulose membranes by capillary action in 20× SSC (1× SSC contains 0.15 M sodium chloride and 0.015 M sodium citrate [pH 7.0]). Membranes were cross-linked under UV light and prehybridized at 42°C in 50% formamide, 5× SSPE (1× SSPE contains 0.1 M sodium chloride, 10 mM NaH₂PO₄, and 1 mM EDTA [pH 7.0]), 5× Denhardt reagent (1× Denhardt contains 0.02% polyvinylpyrrolidone, 0.02% BSA, and 0.02% single-density gradient medium [Ficoll 400; Pharmacia Upjohn, Uppsala, Sweden]), 1 mg/mL salmon sperm DNA, and 0.1% SDS for 4 hours. Thereafter, the membranes were hybridized at 42°C for 12 hours in 5× SSPE, 50% formamide, 0.2 mg denatured salmon sperm DNA/mL, 1× Denhardt reagent, and 1× 10⁶ counts per minute (cpm)/mL specific cDNA probe. Mouse S-cone opsin cDNA fragment (520 bp: 115-635), mouse M-cone opsin fragment (530 bp: 120–650), and TRβ2 isotype-specific fragment (350 bp: 3–353) were labeled with [³²P]deoxy-CTP by the random primer method and used as probes. The membranes were washed at room temperature for 15 minutes in 2× SSPE, 0.03% sodium pyrophosphate, and 0.1% SDS. The most stringent wash was then performed at 65°C in 0.1× SSPE containing 1.0% SDS and 0.03% sodium pyrophosphate. The membranes were dehybridized for 30 minutes in 0.1× SSPE and 0.1% SDS at 90°C and then hybridized with mouse glyceraldehyde-3-phosphate dehydrogenase (G3PDH) probes. 

Reverse Transcription–Polymerase Chain Reaction

Total RNA, extracted from the eyes as described earlier, was treated with 10 units of RNase-free DNase I (Toyobo, Osaka, Japan) in the presence of 10 U placental RNase inhibitor (Toyobo). The mRNA (2 μg) was then converted to cDNA in a 20-μL reaction mixture with reverse transcriptase (Superscript II; Gibco BRL, Rockville MD), as recommended by the manufacturer. An aliquot (1 μL) of each RT reaction mixture was then added to a standard 50-μL PCR mixture. After 5 minutes of preincubation at 95°C, amplification was performed for 30 cycles, consisting of 1 minute of denaturing at 95°C, 1 minute of annealing at 56°C, and 1 minute of extension at 72°C. PCR products were separated on 1% agarose gel. Preliminary experiments established that under these conditions of PCR, the amounts of each transcript are semiquantitative. 

Plasmid Construction

The original constructions of full-length human PNR, mouse TRβ2, and mouse CRX have been described elsewhere (see Refs. ³ , ⁷ , ⁸ , respectively). DNA fragments containing the human S-cone opsin enhancer element ⁹ and human M-cone opsin locus-controlling region (LCR) ¹⁰ were amplified from human genomic DNA by PCR and inserted between the KpnI and SmaI sites of pGL3-basic vector (Promega, Madison, WI) harboring the herpes simplex thymidine kinase (tk) promoter at the BamHI site (S-cone enhancer and M-cone LCR). The sequences of cDNA primers are available from the authors on request. 

Transfection and Luciferase Assay

293T cells were maintained in Dulbecco’s modified Eagle’s medium without phenol red, supplemented with 5% dextran-coated charcoal-stripped fetal bovine serum. The cells were transfected at 70% to 80% confluence in 6-well dishes with lipofectin (Gibco BRL) by following the manufacturer’s instructions as described. ¹¹ One microgram of reporter vector was transfected with 100 ng PNR, TRβ2, and/or CRX, as indicated. In all assays, 200 ng pRL-tk vector (Promega) was transfected as an internal control. Cognate ligands were added to the medium 1 hour after transfection and at each change of medium. Six to 8 hours after transfection, cells were washed with fresh medium, ligands were added to the medium, and cells were incubated for an additional 24 to 36 hours. Cell extract preparations and dual luciferase assays were performed according to the manufacturer’s protocol (Promega). 

Results

Temporal Expression Pattern of PNR and Histologic Analysis of rd7/rd7 Mice

To clarify the temporal expression pattern of the PNR gene in relation to TRβ2 in mice, we chose a semiquantitative RT-PCR analysis. mRNA of PNR was detected as early as embryonic day 18 (E18) and became abundant in adult retina. This expression pattern of the PNR gene is different from that of TRβ2, with expression that was detected as early as E16, peaked at E18, and decreased thereafter (Fig. 1A) . Extensive RT-PCR analysis detected no mRNA other than in the retina (Fig 1. A and data not shown) as previously described. ³  

Previous studies have shown that in rd7/rd7 mice, formations of whorls and rosettes in the outer nuclear layer are observed as early as postnatal day 12.5 (P12.5), ⁵ become most evident at 1 month, and vanish at 5 months, to be followed by retinal degeneration at 16 months. ⁴ Whorls and rosettes are surrounded by cone photoreceptors that express S-cone opsin, and their formation is a consequence of an excess proliferation of the cone photoreceptors. ⁵ To determine the exact time point of the onset of this retinal phenotype, we analyzed serial sections of rd7/rd7 mouse retina systematically. Consistent with the previous studies, we found whorls and rosettes at P14 (Figs. 1E 1F) . Close examination revealed that the formation of whorls and rosettes started earlier than previously documented ⁵ —as early as P5 (Figs. 1D 1E) , but not before P3 (Figs. 1B 1C) . 

Expression of S- and M-Cone Opsin in rd7/rd7 Mouse Retina

Previous studies have hypothesized abnormal S- and L/M-cone development in ESCS. ⁶ Both the S-cone opsin gene promoter and the M-cone LCR, located upstream of the tandem L- and M-cone opsin genes, are conserved between humans and mice and are indispensable to directing gene expression in S- and M-cone photoreceptors in the mouse retina. ⁹ Thus, the transcriptional regulatory mechanisms of the cone opsin genes are similar in humans and mice. To clarify the developmental abnormalities of S- and M-cone opsin gene expression in rd7/rd7 mouse retina, we studied the developing and adult retinas of rd7/rd7 mice. Northern blot analysis and RT-PCR disclosed an increase in S-cone opsin in all periods examined (Figs. 2A 2B) . However, unexpectedly, the expression of M-cone opsin was not perturbed at P14 (Fig. 2A , lanes 1, 2), when retinal morphogenesis was completed and at P7 (Fig. 2B) , immediately after photoreceptors began expressing S- and M-cone opsins. In adulthood, when retinal degeneration occurs, ⁴ a decrease in expression of M-cone opsin was observed (Fig. 2A , lanes 3, 4). 

Normal Plasma Concentration of Thyroid Hormone and Expression of TRβ2 in the Developing rd7/rd7 Mouse Retina

Because thyroid hormone and TRβ2 are involved in cone photoreceptor development, ^{1 2} we examined levels of plasma thyroid hormones and TRβ2 expression in rd7/rd7 mice retina. By Northern blot analysis, the expression of TRβ2 was normal at E18 (Fig. 2C , lanes 1, 2) when it usually reaches a peak, and was normal at P14 (Fig. 2C , lanes 3, 4) as in wild-type mice, even when retinal morphologic abnormalities became evident in rd7/rd7 mice. Moreover, plasma concentrations of thyroxin and tri-iodothyronine were normal (Fig. 2D) . Thus, the thyroid hormone–TRβ2 system in the retina was unaffected in rd7/rd7 mice. 

Roles of TRβ2 and PNR in Expression of S- and M-Cone Opsin

M-cone LCR, which is located 3.1 to 3.7 kb upstream from the transcription start site of the L-cone gene, ¹⁰ contains DNA elements that match the half-site recognition motif of the nuclear receptor superfamily ¹² conserved between humans and mice (data not shown), raising the possibility that M-cone is transcriptionally regulated, either by PNR or TRβ2. In addition, both the S-cone enhancer, which is located 0.47 kb upstream of the S cone gene promoter, ¹³ and the M-cone LCR contain putative binding sites for the photoreceptor-specific homeobox transcription factor CRX. ^{14 15} Thus, we performed luciferase assays and examined possible effects of PNR, TRβ2, and CRX on the putative regulatory elements in their promoters. In the S-cone enhancer, CRX robustly increased reporter activity as expected (Fig. 3A , lanes 1–4). Neither overexpressed PNR nor TRβ2 altered reporter activity by itself (Fig. 3A , compare lanes 1–5 with 5–8). TRβ2, however, abrogated the potentiation of reporter activity mediated by CRX in a thyroid hormone–dependent manner (Fig. 3A , lanes 11, 12). In the M-cone LCR, TRβ2 robustly increased reporter activity in a thyroid hormone–dependent manner (Fig. 3B , lanes 7, 8). PNR and CRX had no effect on the reporter activity by themselves or the reporter activity elicited by TRβ2 (Fig. 3B , lanes 1–6, 9–16). 

Discussion

Factors regulating the proliferation and differentiation of retinal progenitor cells are under intensive study. Retinal progenitor cells are considered multipotent—namely, one retinal progenitor cell proliferates to produce several types of cells in the retina, although they are biased to produce specific types of cells by intrinsic and extrinsic factors. ¹⁶ We and others ⁵ observed that overproduction of cone photoreceptor became apparent postnatally in rd7/rd7 mice. Normally, cone photoreceptors are generated from cells that exit the cell cycle in the embryonic period. ¹⁶ In contrast, cells that exit the cell cycle after the postnatal period produce bipolar cells and Müller cells in normal development. ¹⁶ However, even though additional retinal cell proliferation became apparent only after the postnatal period in rd7/rd7 mice, we found no substantial change in bipolar cells and Müller cells by immunohistochemical analysis using antibody against protein kinase C, Chx10, and glutamine synthetase, together with semiquantitative RT-PCR analysis of Chx10 and cellular retinaldehyde binding protein (CRALBP; data not shown), although a previous study found that PNR regulates the expression of the CRALBP gene. ¹⁷ These observations suggest that retinal progenitor cells, the progeny of which give rise exclusively to cone photoreceptors, may exist in the neonatal retina, and the proliferation of such progenitor cells is suppressed by PNR. In contrast to the retinal phenotype of rd7/rd7 mice, a normal number of total cone cell photoreceptors has been documented in TRβ2-deficient mouse retina. ² In addition, we found normal plasma concentrations of thyroid hormone and a normal expression level of TRβ2 in rd7/rd7 mouse retina. These observations indicate that the proliferation of cone photoreceptor progenitor cells is regulated by PNR, not by TRβ2. 

Thus far, little is known about the factors that determine the cell fate of cone photoreceptors through regulation of the expression of cone opsins at the transcriptional level. The results of reporter assays clearly demonstrated that the expression of S-cone opsin was induced by CRX and this induction was suppressed by TRβ2 in a ligand-dependent manner. The expression of M-cone opsin was induced by TRβ2 in a ligand-dependent manner, but not by CRX, although a putative binding site for CRX exists in the M-cone LCR. ¹⁵ We found unexpectedly that PNR has no regulatory role in the S-cone enhancer and M-cone LCR. These results support that the expression of S-cone opsin is a default status, and the conversion from S-cone to M-cone photoreceptors requires additional events. We propose the default status of S-cone opsin expression to be defined by factors including CRX and liganded TRβ2 to work as a switching molecule that inhibits expression of S-cone opsin while inducing expression of M-cone opsin (Fig. 4A) . 

Our observations showed that the increased number of cone photoreceptors and the increased expression in S-cone opsin are the result of developmental abnormalities, and that the decrease in M-cone opsin in adult rd7/rd7 mice is a result of degeneration. In contrast, previous work demonstrated that TRβ2-deficient mice overexpress S-cone opsin accompanied by a marked reduction in M-cone opsin due to developmental defects. ² Thus, the retinal phenotype of rd7/rd7 mice is completely different from that of TRβ-deficient mice. Our results disclosed different molecular mechanisms of overexpression of S-cone opsin between TRβ2-deficient and rd7/rd7 mice, at least in part. In TRβ2-deficient mice, the increase in S-cone and decrease in M-cone opsins was due to the absence of conversion of S-cone opsin–expressing photoreceptors to M-cone opsin–expressing photoreceptors (Fig. 4B) . In rd7/rd7 mice, because the plasma concentration of thyroid hormone and expression level of TRβ2 are normal in the retina, a normal degree of conversion from S-cone opsin to M-cone opsin occurred despite the increase in the number of cone photoreceptors, resulting in overproduction of S-cone opsin with normal expression of M-cone opsin (Fig. 4C) . 

Patients with ESCS show increased S-cone photoreceptor sensitivity, and a histopathologic study of an ESCS-affected retina revealed an increase in the total number of cone photoreceptors, with an increase in the number of S-cone opsin immunoreactive photoreceptors and a decrease in the number of L/M-cone photoreceptors. ¹⁸ L/M-cone function in ESCS retinas was demonstrated to decline progressively and to vary from subnormal to nonrecordable. ^{6 18} It is impossible to know whether the disruption in L/M cone function in patients with ESCS is due to degeneration, similar to rd7/rd7 mice, or a developmental perturbation of the expression of L/M-cone opsin. However, the remarkable resemblance in retinal phenotype between patients with ESCS and rd7/rd7 mice may support the idea that the same mechanism underlies the S- and (L-/)M-cone abnormality in both cases and that PNR shares a similar function in the development of cone photoreceptors in humans and mice. 

Our results partly disclosed the molecular mechanism of cone photoreceptor development, highlighting the distinct functions of PNR and TRβ2. The other type of photoreceptor in the mammalian retina is the rod photoreceptor, functioning under dim light and characterized by rhodopsin gene expression. A very recent study demonstrated that the retina of mice without a basic leucine zipper-motif protein, neural retina leucine zipper (Nrl), shows a complete loss of rod photoreceptors and increased number of cone photoreceptors, most of which express S-cone opsin, and shows no PNR gene expression. ¹⁹ The retinal morphology of Nrl-deficient mice is strikingly similar to that of rd7/rd7 mice. However, by Northern blot analysis, we found the expression of the rhodopsin gene to be normal in rd7/rd7 mice (data not shown), and consistent with this observation, an electrophysiological study has demonstrated that the rod photoreceptor function was normal in rd7/rd7 mice before retinal degeneration became evident. ⁴ Thus, PNR is involved in cone photoreceptor development but is likely to be dispensable to rod photoreceptor development. Identification of ligands, target genes, other factors regulating the expression of PNR, and transcriptional cofactors of PNR would clarify the close molecular mechanism of cone photoreceptor development. 

 Supported by a Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Sports, Science and Culture of Japan (SK).

 Submitted for publication April 3, 2002; accepted May 21, 2002.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Temporal expression pattern of PNR and phenotype of rd7/rd7 mice. (A) RT-PCR analysis of PNR and TRβ2 transcripts at various stages of development. Poly(A)⁺-RNA (2 μg) from the eyes was reverse-transcribed and amplified by PCR using mouse PNR and TRβ2 isotype-specific primers. Amplified fragments were detected by agarose gel electrophoresis. The sequences of these fragments were identical with those of PNR and TRβ2 transcripts. Representative data are shown. (B–G) Histology on serial 4-μm paraffin sections of P3 to P14 wild-type (WT) and rd7/rd7 (rd7) mice. In the wild-type mice, at P3, retinal cells were composed of two layers: a pre–ganglion cell layer (PGCL) and a neuronoblastic cell layer (NBL) (B). The formation of three cell layers began at P5 (D) and ended at P14 (F). In rd7/rd7 mice, retinal morphology was normal at P3 (C). The formation of rosette in the outer nuclear layer (ONL) starts as early as P5 (E) and becomes evident at P14 (G). INL, inner nuclear layer.

Figure 2.

 

Expression of S-cone opsin, M-cone opsin, thyroid hormone β2 and plasma concentrations of thyroid hormones in rd7/rd7 mice. (A) Northern blot analysis of S- and M-cone opsin transcripts of wild-type (WT) and rd7/rd7 (rd7) mice at P14 and adulthood. Total RNA (100 μg) was electrophoresed and 520- and 530-bp cDNA fragments corresponding to the N-terminal region of mouse S- and M-cone opsin, respectively, were used as a probe. (B) RT-PCR analysis of S- and M-cone opsin transcripts at P7. Poly(A)⁺-RNA (2 μg) from the eyes was reverse-transcribed and amplified by PCR using mouse S- and M-cone opsin–specific primers. Amplified fragments were detected by agarose gel electrophoresis. The sequences of these fragments were identical with those of S- and M-cone opsin transcripts. Representative data are shown. (C) Northern blot analysis of thyroid hormone β2 transcript of wild-type (WT) and rd7/rd7 (rd7) mice at E18 and P14. Total RNA (100 μg) was electrophoresed, and the 350-bp cDNA fragment corresponding to the N-terminal mouse TRβ2 isotype-specific region was used as a probe. The hybridized membranes were exposed to x-ray film for 10 days. Similar results were obtained in three independent sets of experiments. As an internal control, the G3PDH transcript was used to confirm the presence of intact mRNAs from the tissues. (D) Plasma concentrations of thyroxin (T₄) and tri-iodothyronine (T₃). Graphs show the average for wild-type (WT; n = 6) and rd7/rd7 (rd7) mice (n = 6); error bars indicate SD.

Figure 3.

 

Effects of PNR, CRX, and TRβ2 on induction of S- and M-cone opsin. (A) 293T cells were transfected with 1 μg of luciferase reporter bearing the S-cone opsin enhancer element and the (B) M-cone opsin LCR, with various combinations of PNR, CRX, and/or TRβ2, as indicated, in the presence or absence of 10⁻⁶ M tri-iodothyronine (T₃). Graphs show the multiples of change (average results in at least three independent experiments; error bars indicate SD) in luciferase activity relative to that in the absence of exogenous expression vectors and ligands.

Figure 4.

 

Model of S- and M-cone photoreceptor development. (A) In normal development, the default status of S-cone opsin expression is defined by factors including CRX, and liganded TRβ2 works as a switching molecule, to inhibit S-cone opsin expression induced by CRX and induce M-cone expression. (B) The absence of TRβ2 causes an increase in S-cone opsin–expressing cells with decreased M-cone opsin expression because there is no conversion from S-cone opsin–expressing photoreceptors to M-cone opsin–expressing photoreceptors. (C) In rd7/rd7 mice, a normal degree of conversion from S-cone opsin to M-cone opsin occurs despite an increase in the number of cone photoreceptors, which results in overproduction of S-cone opsin with normal expression of M-cone opsin.

The authors thank Constance L. Cepko for the CRX plasmid, Takeshi Nakajima for 293T cells, and Mikiro Mori and Mime Kobayashi for helpful discussions.