Genetic dissection of TAM receptor-ligand interaction in retinal pigment epithelial cell phagocytosis

Abstract

Although TAM receptor tyrosine kinases play key roles in immune regulation, cancer metastasis, and viral infection, the relative importance of the two TAM ligands-Gas6 and Protein S-has yet to be resolved in any setting in vivo. We have now performed a genetic dissection of ligand function in the retina, where the TAM receptor Mer is required for the circadian phagocytosis of photoreceptor outer segments by retinal pigment epithelial cells. This process is severely attenuated in Mer mutant mice, which leads to photoreceptor death. We find that retinal deletion of either Gas6 or Protein S alone yields retinae with a normal number of photoreceptors. However, concerted deletion of both ligands fully reproduces the photoreceptor death seen in Mer mutants. These results demonstrate that Protein S and Gas6 function as independent, bona fide Mer ligands, and are, to a first approximation, interchangeable with respect to Mer-driven phagocytosis in the retina.

Figure 1. Analysis of retinal degeneration in the retina of Gas6 and Protein S mutants

(A) Retinal sections. Left eyes were sectioned at 12 weeks along the dorsal-ventral (DV) plane for all genotypes analyzed, at the indicated position just nasal (N) to the optic disk (left, dotted lines), oriented with dorsal to the right, and H&E stained (middle, right). (B) Measurement of the thickness of the outer nuclear layer (white lines, numbers in microns), which is composed of PR nuclei, across 20 segments (5% increments) of the DV axis of a wild-type retinal section. This DV axis is plotted as the x-axis of Figure 2A–C. A section equivalent to the dashed box at center is enlarged in C. (C) The thickness of the outer nuclear layer (ONL; white line) can be measured reliably. This thickness is plotted as the y-axis of Figure 2A–C. Other abbreviations: GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; IS, inner segment; OS, outer segment; RPE, retinal pigment epithelium.

Figure 2. Relative roles of Protein S and Gas6 with respect to photoreceptor (PR) viability at 12 weeks

(A) Measurement of ONL thickness was performed on stained sections as illustrated in Figure 1B, and average values in microns plotted across the DV axis of the retina. ONL thickness across the Gas6^−/− retina (blue curve) is equivalent to wild-type (black curve). Arrowhead on y-axis represents the approximate thickness of a single PR nucleus. Error bars in this and all subsequent plots represent 1 S.D. from the mean, for independent measurements at the equivalent position in retinal sections from 3 separate mice. (B) Mertk^−/− mice (red curve) display a severe loss of PRs, with only 2–4 PR nuclei present in the ONL across the DV axis of the retina. Retinal sections from mice that have Protein S eliminated from cells of the RPE (Pros1^fl/fl/Trp1-Cre or Pros1^fl/−/Trp1-Cre) and are also either Gas6^+/+ or Gas6^+/− (light orange curves) have a normal distribution of ONL thickness across the retina. However, sections from mice in which Gas6 is completely knocked out and Protein S is also eliminated from the RPE display an intermediate phenotype (dark orange solid and dashed curves), with partial PR loss relative to wild-type (black curve). As highlighted in Supplemental Figure S1C, D, the reduction in ONL thickness is especially severe in these Pros1^fl/fl/Trp1-Cre/Gas6^−/− and Pros1^fl/−/Trp1-Cre/Gas6^−/− mice at the extreme ventral and dorsal ends of the retina. (C) Pros1^fl/−/Nes-Cre/Gas6^−/− mice (dark green solid curve) display the same severe reduction in ONL thickness seen in Mertk^−/− mice. Pros1^fl/fl/Nes-Cre/Gas6^−/− mice (dark green dashed curve) show an equivalent PR loss only in the central-most 30% of the retina, but this restriction in phenotype is due to reduced recombination activity of the Nestin-Cre driver in peripheral retinal locations (see Supplemental Figure S2). Adding a single wild-type Gas6 allele back to the Pros1^fl/−/Nes-Cre genotype restores the distribution of ONL thickness across the retina to wild-type (light green curves). Genotypes are ordered with respect to phenotype severity above the plots.

Figure 3. Retinal histology of mutant phenotypes

All images are H&E stained sections from 30% of the retinal DV axis, as defined in Figs. 1 and 2. (A, F) Wild-type. Abbreviations are as for Figure 1C. (B) Mertk^−/−. ONL is reduced to a thickness of only 2–4 PR nuclei, and PR OS are almost entirely absent. RPE and other retinal laminae are intact. (C) Gas6^−/−. ONL is of normal thickness, and OS are longer and more densely packed than wild-type. See also Figure 4. (D, G) Pros1^fl/−/Nes-Cre/Gas6^−/−, in which both Protein S and Gas6 are eliminated from the retina. ONL is nearly obliterated, as in Mertk^−/− mice (compare with B). (E) Pros1^fl/−/Nes-Cre/Gas6^+/−. Adding a single wild-type Gas6 allele to the genotype in (D, G) completely restores the ONL. OS are longer than wild-type; see also Figure 4. (H, I) Pros1^fl/−/Trp1-Cre/Gas6^−/− and Pros1^fl/fl/Trp1-Cre/Gas6^−/−, respectively, in which Gas6 is eliminated from the entire retina and Protein S from the RPE. ONL thickness is reduced by ~30% relative to wild-type. (See also Figure 2B.) (J) Adding a single wild-type Gas6 allele to the genotype in (H) completely restores the ONL. OS are again longer than wild-type; see also Figure 4. Measurements across the complete DV axis of multiple mice indicate that a statistically significant diminution in INL thickness in Gas6^−/− mutants relative to wild-type is not a consistently observed phenotype. Scale bar is 100 μm.

Figure 4. Outer segments (OS) in TAM ligand single mutants are longer than those of wild-type

(a) A representative side-by-side comparison of central retinal sections, stained with H&E, taken from wild-type (left) versus Gas6^−/− (right) mice at 12 weeks. The sections are entirely comparable, except for (i) a longer average OS length in the Gas6^−/− mice (bars in left and right panels, IS length is unchanged); and (ii) a greater frequency of gaps in the OS layer (asterisks in left panel) in wild-type section. Scale bar is 50 μm. (B) Quantitation of the central retinal ratio of the length of the outer segment to the length of the inner segment (OS:IS ratio) performed on two retinal sections from two separate mice of each indicated genotype. (n is the total number of measurements performed.) Variation is 1 standard deviation from the mean. *** indicates highly significant difference (p value <0.0001) between the indicated genotypes and wild-type upon analysis by two-tailed t test. (C) Quantitation of opsin-positive phagosomes, as described in Experimental procedures, per 100μm length of RPE in sections of WT (white bar) versus Gas6^−/− (gray bar) retinae, at 30 min after subjective dawn. (n is the total number of sections counted, from two eyes per genotype). Variation is 1 standard error of the mean. * indicates significant difference (p value =0.03) between Gas6^−/− and wild-type upon analysis by two-tailed t test.

Figure 5. Gas6 expression in the retina

(A, B) Anti-Gas6 IHC with HRP-conjugated secondary antibodies across all retinal layers (labels as for Figure 1C) in wild-type (A) and Gas6^−/− (negative control) mice (B). Arrowheads in A denote Gas6-positive cells (brown signal) in the inner nuclear layer (INL), the cytoplasm-rich inner segments (IS) of PRs, and the apical microvilli of RPE cells, which penetrate into the PR OS layer. [These same apical microvilli are positive for both Mer and Tyro 3 (Prasad et al., 2006).] Images equivalent to (but distinct from) the indicated boxed areas are enlarged in panels C, D, and E. (C, D) Anti-Gas6 IHC in PR IS layer and region populated by RPE microvilli in wild-type (C) and Gas6^−/− (D) mice. (E) High power view of anti-Gas6 IHC in the INL, equivalent to the box indicated in A. The positions of retinal laminae are indicated by the dotted lines. (F) Co-staining with anti-Gas6 (green) and anti-PKCα (red); nuclei are counter-stained with a Hoechst dye (blue). (G) Co-staining with anti-Gas6 (red) and anti-glutamine synthetase (GS; green); nuclei are blue. (H) Co-staining with anti-Gas6 (green) and anti-calbindin (Calb; red); nuclei are blue. Scale bars are 25 μm.

Figure 6. TAM receptor and ligand mRNA expression in the eye

(A) TAM receptor mRNAs in the RPE. The RPE cell layer was dissected from the retina, and the absolute level of Tyro3, Axl, and Mertk mRNAs determined by qRT-PCR (see Experimental Procedures). Mertk mRNA is 4.2-fold more abundant than Tyro3 mRNA; Axl mRNA is not detectable. (B, C) The eye cup (primarily choroid and sclera remaining after removal of neural retina and RPE), the single-cell RPE layer, the pigmented ciliary body at the retinal periphery that is contiguous with the RPE (CB), and the neural retina were each dissected from the retina, and the absolute level of Gas6 (B) and Pros1 (C) mRNAs determined by qRT-PCR. Gas6 mRNAs were also measured in retina obtained from Pros1^fl/−/Nes-Cre and Pros1^fl/−/Trp1-Cre mice (B, white and gray-striped bars, respectively); and Pros1 mRNAs were measured in all eye tissues obtained from Gas6^−/− mice (C, white bars). Both TAM ligand mRNAs are expressed broadly in ocular tissues.