Making the gradient: thyroid hormone regulates cone opsin expression in the developing mouse retina

Abstract

Most mammals have two types of cone photoreceptors, which contain either medium wavelength (M) or short wavelength (S) opsin. The number and spatial organization of cone types varies dramatically among species, presumably to fine-tune the retina for different visual environments. In the mouse, S- and M-opsin are expressed in an opposing dorsal-ventral gradient. We previously reported that cone opsin patterning requires thyroid hormone beta2, a nuclear hormone receptor that regulates transcription in conjunction with its ligand, thyroid hormone (TH). Here we show that exogenous TH inhibits S-opsin expression, but activates M-opsin expression. Binding of endogenous TH to TRbeta2 is required to inhibit S-opsin and to activate M-opsin. TH is symmetrically distributed in the retina at birth as S-opsin expression begins, but becomes elevated in the dorsal retina at the time of M-opsin onset (postnatal day 10). Our results show that TH is a critical regulator of both S-opsin and M-opsin, and suggest that a TH gradient may play a role in establishing the gradient of M-opsin. These results also suggest that the ratio and patterning of cone types may be determined by TH availability during retinal development.

Fig. 1.

Exogenous T3 inhibits S-opsin. (A) Whole retinas from embryonic day 17 mice cultured for up to 2 weeks. S-opsin immunostaining (green) shows a normal ventral-to-dorsal gradient. (B) The number of S-opsin immunolabeled cones were counted in untreated (CTRL) and T3-treated explants after 11 days in vitro (∗, P < 0.01; ∗∗, P < 0.001). (C–F) Newborn mice (P0) were injected s.c. with either saline or T3 (1.5 μg) for 3 days. The animals were killed, and the retinas were flat-mounted and labeled with S-opsin (red) and PNA (green) to label the total population of cones. Images were collected with a Zeiss Pascal confocal microscope at ×40, and both S-opsin and total cones (PNA⁺) were quantified from four 900-μm² regions in each of three saline and three T3-treated retinas. A total of 881 cones were counted in the saline group and 914 for the T3-treated group. (C and D) There was a significant decline in the number of S-opsin-expressing cones in the T3-treated retinas (∗, P < 0.015 Student’s t test; mean and SEM shown in graph). (E) There was no significant difference in the total number of PNA⁺ cones between the groups (saline, 73.4 ± 6.45 SD; T3-treated, 76.16 ± 4.06 SD). (F) The T3-treated retinas still demonstrate a very shallow gradient in the number of S-opsin cones per field, but in all regions there were many fewer S-opsin-expressing cones than in the saline-injected animals.

Fig. 2.

T3 inhibits S-opsin in vivo. A 1.5-μg dose of T3 was injected into mice once daily, from P0 to P3. (A) Northern blots show that S-opsin transcript is reduced in wild-type mice that were treated with T3. (B) Digital quantification of band intensity shows a 75% decrease in S-opsin transcript in mice injected with T3. (C) T3 reduces the number of S-opsin-expressing cones in vivo. Transgenic mice expressing LacZ under the control of the S-opsin promoter show a typical S-opsin gradient, with β-galactosidase expression highest in the ventral retina (Upper Left). Injecting T3 into reporter mice from P0 to P3 dramatically reduces the number of cones expressing S-opsin (Upper, arrow). Injecting T3 into S-opsin reporter mice lacking the TRβ2 receptor did not effect the number of S-opsin-expressing cones (Lower). ONL, outer nuclear layer; GCL, ganglion cell layer; RPE, pigment epithelium.

Fig. 3.

T3 binding is required for opsin patterning. (A) Whole-mount retinas from adult WT, TRβ^PV/+ (not shown), and TRβ^PV/PV were double-labeled with antibodies to M-opsin (green) and S-opsin (blue). Wild-type retinas show almost exclusive expression of M-opsin in the dorsal retina and S-opsin in the ventral retina. TRβ^PV retinas, which have a targeted mutation that blocks T3 binding to TRβ, have no M-opsin expression, and S-opsin is ectopically expressed in dorsal cones. (B) Quantification of the percentage of S-opsin immunoreactive cones shows statistically significant changes in S-opsin for both the homozygous mutant as well as the TRβPV^+/− retinas. ∗, P < 0.005; ∗∗, P = 9.3E. (C) T3 up-regulates M-opsin expression. P7 pups were injected s.c. with T3 for 4 days and then killed, and their retinas were flat-mounted for immunolabeling with M-opsin and PNA (n = 4 for saline and n = 5 for T3-treated). The percentage of M-opsin⁺ cones was increased in each region of the retina, but most significantly in the ventral retina (∗, regions compared with ANOVA P < 0.01; pairwise comparison, P < 0.05 for D). (D) Total cone number is not affected by T3 treatment in P7–P10 mice; total number of PNA⁺ cones per field shown.

Fig. 4.

Changes in thyroid hormone during mouse retinal development between P0 and P10. Retinas were bisected into dorsal and ventral halves and pooled for measurement of T3 and T4 by RIA. The concentrations are expressed per retina (Left) and per gram of retinal tissue (Right). T3 and T4 increase in the dorsal retina between P0 and P10.

Fig. 5.

Ligand gradients in the developing retina. (A) There is no difference in the ratio of dorsal–ventral T3 or T4 at P0 or P4. At P10, both T3 and T4 were elevated in the dorsal retina. (B) Model of opsin regulation by thyroid hormone. At S-opsin onset (schematized on the left), TRβ2, RXRγ, and T3 are required to inhibit S-opsin expression in dorsal cones. However, TH is not graded in the retina at this time, indicating that another factor (X) is required to establish the S-opsin gradient. TRβ2 is required for M-opsin onset around P10 (schematized on the right). T3 is highest in the dorsal retina at this time, which likely promotes M-opsin expression in the dorsal retina.