Purchase this article with an account.
L. E. Vazquez, B. Stevens, N. Nouri, G. R. Howell, S. W. M. John, B. A. Barres; The Role of the Complement Cascade in the Pathogenesis of Glaucoma. Invest. Ophthalmol. Vis. Sci. 2008;49(13):1648.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Glaucoma is the leading neurodegenerative cause of blindness worldwide. Blindness results from death of the retinal ganglion cells (RGCs). Although important risk factors that render RGCs susceptible to death have been identified, the cause of RGC death remains unknown. We hypothesize that RGCs degenerate because their synapses with the rest of the cells in the retina are destroyed by abnormal activation of the complement cascade. We took advantage of the DBA/2J mouse model of glaucoma to test this hypothesis.
Retinas from C57/B6, C1qKO, DBA/2J and DBA/2J-Gpnmb+ mice were used. Eyeballs were fixed, frozen and cryo-sectioned. Dye-trace and immunohistochemistry with anti-PSD95 and anti-C1q antibodies were analyzed by confocal microscopy and blind image analysis.
Here, we demonstrate that there is substantial synapse loss in the retina of DBA/2J mice. Moreover, within the retina, the area most severely affected is the inner plexiform layer (IPL), which hosts the RGC synaptic terminals. And more importantly, synapse loss coincides with, or precedes, RGC death, supporting the notion that synapse loss is not a result of RGC death. In fact, synapse loss was observed in young DBA/2J mice with no evidence of optic nerve degeneration (pre-Glaucoma). In addition, we found that C1q, a complement protein known to be upregulated in glaucoma, accumulates in the IPL around the time of synaptic elimination (preceding RGC death). Similar to the synaptic stain, the C1q stain is also punctate, suggesting that C1q may be opsonizing synapses and targeting them for destruction. This C1q accumulation was not observed in control mice. However, in developing wild-type mice (up to postnatal day 15), C1q is expressed in brain and in the retina in a similar pattern as glaucoma mice. This postnatal expression and synaptic localization coincides with the period of synaptic pruning in the developing brain. Indeed, we found that C1q-knockout mice were unable to prune-away unwanted synapses, leading to an overlap in LGN territories innervated by RGC axons from each eye, and confirming that C1q is necessary for synapse elimination.
Our data shows that C1q expression is activated in the retina during development, switched off in adulthood, and reactivated in glaucoma before RGC death. We also show that C1q is necessary for developmental synaptic pruning, and appears to drive pathological destruction of synapses in glaucoma. Our findings shed light on the pathophysiology of glaucoma in humans, and may open new avenues for treatment of this disease.
This PDF is available to Subscribers Only