Quantifying the number of axons in the optic nerve is of interest in many research questions. Here, we show that a stereological method allows simple, efficient, precise and unbiased determination of the total axon number in the murine optic nerve. Axons in semi-thin optic nerve cross sections from untreated eyes (n = 21) and eyes subjected to retinal damage by intravitreous NMDA injections (n = 32) or PBS controls (n = 5) were manually identified, counted and digitally labeled by hand. A stereological procedure was empirically tested with systematic combinations of different sampling methods (simple random sampling without replacement, systematic uniform random sampling, stratified random sampling) and sampling parameters. Extensive numerical Monte Carlo experiments were performed to evaluate their large-sample properties. Our results demonstrate reliable determination of total axon number and superior performance compared to other methods at a small fraction of the time required for a full manual count. We specify suitable sampling parameters for the adoption of an efficient stereological sampling scheme, give empirical estimates of the additionally introduced sampling variance to facilitate experimental planning, and offer AxonCounter, an easy-to-use plugin implementing these stereological methods for the multi-platform image processing application NIH ImageJ.

Olfactomedin 1 (OLFM1) is a secreted glycoprotein and member of the olfactomedin protein family, which is preferentially expressed in various areas throughout the central nervous system. To learn about the functional properties of OLFM1 in the eye, we investigated its localization in the mouse and pig eye. In addition, we analyzed the ocular phenotype of Olfm1 mutant mice in which 52 amino acids were deleted in the central part (M2 region) of OLFM1. OLFM1 was detected in cornea, sclera, retina, and optic nerve of both wild-type and Olfm1 mutant littermates. By immunohistochemistry and double labeling with the lectin peanut agglutinin, OLFM1 was found in the interphotoreceptor matrix (IPM) of mouse and pig retina where it was directly localized to the inner segments of photoreceptors. Western blotting confirmed the presence of the OLFM1 isoforms pancortin 1 (BMY) and pancortin 2 (BMZ) in the IPM. The retinal phenotype of Olfm1 mutant mice did not obviously differ from that of wild-type littermates. In addition, outer nuclear layer (ONL) and total retinal thickness were not different, and the same was true for the area of the optic nerve in cross sections. Functional changes were observed though by electroretinography, which showed significantly lower a- and b-wave amplitudes in Olfm1 mutant mice when compared to age-matched wild-type mice. When light damage experiments were performed as an experimental paradigm of photoreceptor apoptosis, significantly more TUNEL-positive cells were observed in Olfm1 mutant mice 30 h after light exposure. One week after light exposure, the ONL was significantly thinner in Olfm1 mutant mice than in wild-type littermates indicating increased photoreceptor loss. No differences were observed when rhodopsin turnover or ERK1/2 signaling was investigated. We conclude that OLFM1 is a newly identified IPM molecule that serves an important role for photoreceptor homeostasis, which is significantly compromised in the eyes of Olfm1 mutant mice.

Epithelial-to-mesenchymal transition (EMT) of the retinal pigment epithelium (RPE) is commonly observed at sites of choroidal neovascularization in patients suffering from age-related macular degeneration. To learn in an experimental model how RPE EMT affects the biology of the choroidal vasculature, we studied transgenic mice (βB1-TGF-β1) with ocular overexpression of transforming growth factor-β1 (TGF-β1). RPE EMT was detectable at postnatal day (P)1 and included marked structural and functional alterations such as loss of the outer blood–retina barrier and reduced mRNA expression of the RPE-characteristic molecules Rlbp1, Rpe65, Rbp1 and Vegfa. Moreover, vascular endothelial growth factor (VEGF) was not detectable by immunohistochemistry at the RPE/choroid interface, while RPE cells stained intensely for α-smooth muscle actin. The choriocapillaris, the characteristic choroidal capillary network adjacent to the RPE, developed normally and was not obviously changed in embryonic transgenic eyes but was absent at P1 indicating its atrophy. At around the same time, photoreceptors stopped to differentiate and photoreceptor apoptosis was abundant in the second week of life. Structural changes were also seen in the retinal vasculature of transgenic animals, which did not form intraretinal vessels, and the hyaloid vasculature, which did not regress. In addition, the amounts of retinal HIF-1α and its mRNA were markedly reduced. We conclude that high amounts of active TGF-β1 in the mouse eye cause transdifferentiation of the RPE to a mesenchymal phenotype. The loss of epithelial differentiation leads to the diminished synthesis of RPE-characteristic molecules including that of VEGF. Lack of RPE-derived VEGF causes atrophy of the choriocapillaris, a scenario that disrupts photoreceptor differentiation and finally results in photoreceptor apoptosis. Lack of retinal vessel formation and of hyaloid vessel regression might be caused by the decrease in the metabolic requirements of the neuroretina leading to low amounts of retinal HIF-1α. In summary, our data indicate that failure of RPE differentiation may well precede and cause atrophy of the choriocapillaris. In contrast, RPE EMT is not sufficient to cause choroidal neovascularization.

There is an ongoing controversy regarding the role of WDR36 sequence variants in the pathogenesis of primary open-angle glaucoma (POAG). WDR36 is a nucleolar protein involved in the maturation of 18S rRNA. The function of WDR36 is essential as homozygous Wdr36-deficient mouse embryos die before reaching the blastocyst stage. Here we provide a detailed analysis of the phenotype of heterozygous Wdr36-deficient mice. Loss of one Wdr36 allele causes a substantial reduction in the expression of Wdr36 mRNA. In the eyes of Wdr36+/− animals, the structure of the tissues involved in aqueous humor circulation and of the optic nerve head are not different from that of control littermates. In addition, one-year-old Wdr36+/− animals do not differ from wild-type animals with regards to intraocular pressure and number of optic nerve axons. The susceptibility of retinal ganglion cells to excitotoxic damage induced by NMDA is similar in Wdr36+/− and wild-type animals. Moreover, the amount of optic nerve axonal damage induced by high IOP is not different between Wdr36+/− and wild-type mice. Transgenic overexpression of mutated Del605-607 Wdr36 in Wdr36+/− animals does not cause changes in the number of optic nerve axons or susceptibility to excitotoxic damage. In addition, analysis of 18S rRNA maturation in Del605-607 Wdr36+/− or Wdr36+/− mice does not show obvious differences in rRNA processing or in the amounts of precursor forms when compared to wild-type animals. Our data obtained in Wdr36+/− mice do not support the assumption of a causative role for WDR36 in the pathogenesis of POAG.

Norrin is a retinal signaling molecule which is expressed in Müller glia and binds to Frizzled-4 to activate canonical Wnt/β-catenin signaling. Norrin is part of an essential signaling system that controls the formation of retinal capillaries during development. To evaluate neuroprotective properties of Norrin independently from its function during retinal angiogenesis, we generated transgenic mice (Rpe65–Norrin) that constitutively express Norrin in the retinal pigmented epithelium. Substantial amounts of Norrin were secreted into the outer retina, which triggered retinal Wnt/β-catenin signaling in conjunction with an increase in the expression of endothelin-2 (EDN2), endothelin receptor B (EDNRB), and glial fibrillary acidic protein (GFAP). Photoreceptors of Norrin-overexpressing mice were significantly less vulnerable to light-induced damage compared to their wild-type littermates. Following light damage, we observed less apoptotic death of photoreceptors and a better retinal function than in controls. The protective effects were abolished if either Wnt/β-catenin or EDN2 signaling was blocked by intravitreal injection of Dickkopf-1 or BQ788, respectively. Light-damaged retinae from transgenic mice contained higher amounts of brain-derived neurotrophic factor (BDNF) and pAkt than those of wild-type littermates. We conclude that constitutive overexpression of Norrin protects photoreceptors from light damage, an effect that is mediated by Wnt/β-catenin and EDN2 signaling and involves neurotrophic activities of BDNF. The findings suggest that Norrin and its associated signaling pathways have strong potentials to attenuate photoreceptor death following injury.

Mutations in the myocilin gene (MYOC) are causative for 10% of cases with juvenile open-angle glaucoma and 3-4% of those with primary open-angle glaucoma. Myocilin is a secreted protein with relatively ill-defined matricellular properties. Despite its high expression in the eye, myocilin-deficient mice have originally been reported to have no obvious ocular phenotype. Here we revisited the ocular phenotype of myocilin-deficient mice and detected a higher number of neurons in their inner (INL) and outer (ONL) nuclear layers, as well as a higher number of retinal ganglion cells (RGC) and their axons. The increase in retinal neurons appears to be caused by a decrease in programmed developmental cell death, as apoptosis of retinal neurons between postnatal days 4 and 10 was found to be attenuated when compared to that of wildtype littermates. In contrast, when Myoc(-/-) mice were crossed with βB1-crystallin-MYOC mice with ectopic overexpression of myocilin in the eye, no differences in developmental apoptosis, RGC number and INL thickness were observed when compared to wildtype littermates. The amounts of the anti-apoptotic Bcl-2-like protein 1 (BCL2L1, Bcl-xL) and its mRNA were increased in retinae of Myoc(-/-) mice, while lower amounts of BCL2L1 and its mRNA were detected in mixed Myoc(-/-)/βB1-crystallin-MYOC mice. The structural differences between Myoc(-/-) mice and wildtype littermates did not result in functional differences as measured by electroretinography. Noteworthy though mixed Myoc(-/-)/βB1-crystallin-MYOC mice with ocular overexpression of myocilin had significant cone function deficits. Myocilin appears to modulate apoptotic death of retinal neurons likely by interacting with the intrinsic apoptotic pathway.