Sriganesh Ramachandra Rao, Steven J. Fliesler

The retinal pigment epithelium (RPE) provides vital metabolic support for retinal photoreceptor cells and is an important player in numerous retinal diseases. Gene manipulation in mice using the Cre- LoxP system is an invaluable tool for studying the genetic basis of these retinal diseases. However, existing RPE-targeted Cre mouse lines have critical limitations that restrict their reliability for studies of disease pathogenesis and treatment, including mosaic Cre expression, inducer-independent activity, off-target Cre expression, and intrinsic toxicity. Here, we report the generation and characterization of a knockin mouse line in which a P2A-CreER ^T2 coding sequence is fused with the native RPE-specific 65 kDa protein ( Rpe65 ) gene for cotranslational expression of CreER ^T2 . Cre ^+/– mice were able to recombine a stringent Cre reporter allele with more than 99% efficiency and absolute RPE specificity upon tamoxifen induction at both postnatal days (PD) 21 and 50. Tamoxifen-independent Cre activity was negligible at PD64. Moreover, tamoxifen-treated Cre ^+/– mice displayed no signs of structural or functional retinal pathology up to 4 months of age. Despite weak RPE65 expression from the knockin allele, visual cycle function was normal in Cre ^+/– mice. These data indicate that Rpe65 ^CreERT2 mice are well suited for studies of gene function and pathophysiology in the RPE.

We describe the utility of a tandem-tagged autophagy reporter mouse model (CAG-RFP-EGFP-MAP1LC3B) in investigating basal macroautophagic/autophagic flux in the neural retina. Western blot, in situ hybridization, immunohistochemistry, and confocal microscopy showed that CAG promoter-driven expression of RFP-EGFP-MAP1LC3B increased “cytosolic” RFP-EGFP-LC3B-I levels, whereas RFP-EGFP-LC3B-II decorates true phagosomes. We verified that the electroretinographic (ERG) responses of tandem-tagged LC3B mice were comparable to those of age-matched controls. Optimized microscope settings detected lipofuscin autofluorescence in retinas of abca4^−/- mice. The majority of retinal phagosomes in the reporter mice exhibited only RFP (not EGFP) fluorescence, suggesting rapid maturation of phagosomes. Only ~1.5% of the total phagosome population was EGFP-labeled; RFP-labeled (mature) phagosomes colocalized with lysosomal markers LAMP2 and CTSD. In the outer retina, phagosome sizes were as follows (in µm², ave ± SEM): RPE, 0.309 ± 0.015; photoreceptor inner segment-myoid, 0.544 ± 0.031; and outer nuclear layer, 0.429 ± 0.011. Detection of RPE phagosomes by fluorescence microscopy is challenging, due to the presence of melanin. Increased lipofuscin autofluorescence, such as observed in the abca4^−/- mouse model of Stargardt disease, is a strong confounding factor when attempting to study autophagy in the RPE. In addition to RPE and photoreceptor cells, phagosomes were detected in inner retinal cell types, microglia, astrocytes, and endothelial cells. We conclude that the tandem-tagged LC3B mouse model serves as a useful system for studying autophagy in the retina. This utility, however, is dependent upon several technical and biological factors, including microscope settings, transgene expression, choice of fluorophores, and lipofuscin autofluorescence. Abbreviations: ACTB: actin, beta; AIF1: allograft inflammatory factor 1; ATG: autophagy related; CTSD: cathepsin D; DAPI: (4’,6-diamido-2-phenylindole); DIC: differential interference contrast; EGFP: enhanced green fluorescent protein; ELM: external limiting membrane; ERG: electroretinography; GCL: ganglion cell layer; GLUL: glutamine-ammonia ligase (glutamine synthetase); INL: inner nuclear layer; IS-E/M: inner segment – ellipsoid/myoid; ISH: in situ hybridization; LAMP2: lysosomal-associated membrane protein 2; L.I.: laser Intensity; MAP1LC3B: microtubule-associated protein 1 light chain 3 beta; MTOR: mechanistic target of rapamycin kinase; O.C.T.: optimal cutting temperature; OS: outer segment; ONL: outer nuclear layer; PE: phosphatidylethanolamine; RFP: red fluorescent protein; R.O.I.: region of interest; RPE: retinal pigment epithelium

Dehydrodolichyl diphosphate synthase (DHDDS) is required for protein N-glycosylation in eukaryotic cells. A K42E point mutation in the DHDDS gene causes an autosomal recessive form of retinitis pigmentosa (RP59), which has been classified as a congenital disease of glycosylation (CDG). We generated K42E Dhdds knock-in mice as a potential model for RP59. Mice heterozygous for the Dhdds K42E mutation were generated using CRISPR/Cas9 technology and crossed to generate Dhdds^K42E/K42E homozygous mice. Spectral domain-optical coherence tomography (SD-OCT) was performed to assess retinal structure, relative to age-matched wild type (WT) controls. Immunohistochemistry against glial fibrillary acidic protein (GFAP) and opsin (1D4 epitope) was performed on retinal frozen sections to monitor gliosis and opsin localization, respectively, while lectin cytochemistry, plus and minus PNGase-F treatment, was performed to assess protein glycosylation status. Retinas of Dhdds^K42E/K42E mice exhibited grossly normal histological organization from 1 to 12 months of age. Anti-GFAP immunoreactivity was markedly increased in Dhdds^K42E/K42E mice, relative to controls. However, opsin immunolocalization, ConA labeling and PNGase-F sensitivity were comparable in mutant and control retinas. Hence, retinas of Dhdds^K42E/K42E mice exhibited no overt signs of degeneration, yet were markedly gliotic, but without evidence of compromised protein N-glycosylation. These results challenge the notion of RP59 as a DHDDS loss-of-function CDG and highlight the need to investigate unexplored RP59 disease mechanisms.

Patients with certain defects in the dehydrodolichyl diphosphate synthase (DHDDS) gene (RP59; OMIM #613861) exhibit classic symptoms of retinitis pigmentosa, as well as macular changes, suggestive of retinal pigment epithelium (RPE) involvement. The DHDDS enzyme is ubiquitously required for several pathways of protein glycosylation. We wish to understand the basis for selective ocular pathology associated with certain DHDDS mutations and the contribution of specific ocular cell types to the pathology of mutant Dhdds-mediated retinal degeneration. To circumvent embryonic lethality associated with Dhdds knockout, we generated a Cre-dependent knockout allele of murine Dhdds (Dhdds^flx/flx). We used targeted Cre expression to study the importance of the enzyme in the RPE. Structural alterations of the RPE and retina including reduction in outer retinal thickness, cell layer disruption, and increased RPE hyper-reflectivity were apparent at one postnatal month. At three months, RPE and photoreceptor disruption was observed non-uniformly across the retina as well as RPE transmigration into the photoreceptor layer, external limiting membrane descent towards the RPE, and patchy loss of photoreceptors. Functional loss measured by electroretinography was consistent with structural loss showing scotopic a- and b-wave reductions of 83% and 77%, respectively, at three months. These results indicate that RPE dysfunction contributes to DHDDS mutation-mediated pathology and suggests a more complicated disease mechanism than simply disruption of glycosylation.