Structural change at the photoreceptor synapse occurs as an acute and destructive reaction to retinal detachment. Changes are most dramatic for the rod cells, which pull their axon terminals back toward their cell bodies. Electron microscopy has definitively demonstrated that retraction results in disruption of the rod-bipolar connection.
6 Unfortunately, reattachment does not reliably regenerate this synapse.
7,10 Thus, preserving the first synapse in the visual pathway by preventing the disconnection of rod terminals from bipolar dendrites is highly desirable and may prevent visual loss associated with detachment even after successful reattachment.
The in vitro model of porcine retinal detachment used in the present study mimics the morphologic changes, rounding, and rod axon retraction, seen in the rod and cone cell terminals of in vivo animal models and in human retinal detachment.
9,11,12 Isolated porcine retina also shows the characteristic increase in glial fibrillary acidic protein (GFAP; unpublished data, J.-F. Wang, E. Townes-Anderson, 2010) seen after in vivo detachment.
47
We have demonstrated that detachment of porcine neural retina from underlying RPE in vitro leads to RhoA activation. Treating detached retina with the RhoA and ROCK inhibitors C3 transferase and Y27632, respectively, had an inhibitory effect on axon terminal retraction from the OPL. Using nicardipine, an L-type Ca
2+ channel blocker, also significantly reduced rod cell axon retraction. Thus, RhoA and calcium are most likely involved in the retraction of mammalian rod axons and terminals after detachment. Calcium blockage and ROCK inhibition have also been shown to reduce rod axon retraction in isolated salamander photoreceptors.
17,44
It is possible that RhoA and Ca
2+ both affect myosin light chain (MLC) and, thus, actomyosin contraction (
Fig. 7). ROCK phosphorylates MLC and the myosin-binding site (MBS) of myosin phosphatase, thereby increasing levels of phosphorylated MLC and inducing actomyosin contraction and presumably retraction (see Ref.
19 for review). Influx of calcium has been shown to bind and activate calmodulin, which in turn activates MLC kinase (see Ref.
48 for review). MLC kinase phosphorylates MLC, leading to retraction. Thus, both ROCK and Ca
2+ can cause an increase in MLC phosphorylation. However, if one looks at the percentage decrease of SV2 labeling in the ONL, it appears that blocking Ca
2+ channels with nicardipine is less effective than blocking RhoA-ROCK activity (65% reduction of staining in the ONL with nicardipine vs. 75% and 79% reduction with CT-04 and Y27632, respectively). In addition, delayed treatment indicated that the timing of ROCK and Ca
2+ signaling is different, with calcium signaling more quickly stimulating actomyosin contraction than ROCK. It is possible that retinal detachment, which produces a spreading depression, opens calcium channels immediately after detachment,
49 accounting for the rapid timeline of Ca
2+ signaling.
Previous research on porcine retina showed that retraction due to detachment could also be reduced and modulated by the application of cAMP analogues or forskolin, which stimulates adenylyl cyclase.
32 Although cAMP has been shown to increase Ca
2+ influx in some neurons, in rod photoreceptors it is thought to reduce L-type channel activity.
50 Additionally, cAMP, through PKA, has been shown to phosphorylate RhoA at serine 188.
51 This phosphorylation blocks the ability of RhoA to bind its downstream effector, ROCK. Interestingly, this phosphorylation has been shown to interfere specifically with the interaction between RhoA and ROCK but not with other RhoA downstream effectors such as mDia and protein kinase N.
52 Thus, cAMP may reduce retraction by both reducing L-type channel activity and blocking the RhoA-ROCK interaction.
Sakai et al.
53 have shown that one of the causes of loss of outer segments and synaptic connectivity after detachment is hypoxia. These changes can be largely prevented with exposure to O
2 levels above normal room level immediately after detachment. Hyperoxia (O
2 70% of chamber air) produced a radical improvement in retinal morphology and reduced retraction of photoreceptor terminals. How might O
2 rescue photoreceptors and stop retraction? In pig pulmonary artery endothelial cells, RhoA activity has been linked to oxygen levels.
54 Endothelial cells responded to low and high levels, respectively, of O
2 by activating and inactivating RhoA. Low levels of O
2 led to a reduction of Rac1 activity but an increase of RhoA activity. When O
2 was increased, Rac1 activity increased and RhoA activity was inhibited. This may explain how increasing O
2 reduces axon retraction.
We observed a transient rise in RhoA activation immediately after detachment and a drop in activation after 2 and 24 hours. Retraction begins within minutes to hours after detachment
31 but is not maximal until approximately 24 hours after detachment. There appears to be a lag between RhoA activation and completion of retraction. Previous work in mouse neuroblasts and embryonic fibroblasts found that RhoA activation, stimulated by LPA, also occurs rapidly, within minutes after stimulation, and drops quickly.
55,56 In contrast, the activation of ROCK does not follow immediately after RhoA activation, and downstream myosin II motor activity is prolonged.
56 Shifting time lines as the cascade moves toward MLC phosphorylation may explain why the peak of retraction occurs as much as 24 hours after RhoA activation.
The apparent slow development of activity in the RhoA-ROCK pathway leading to retraction may allow for effective treatment of synaptic remodeling by rod, and perhaps also cone, photoreceptors in detached retinas and suggests that the RhoA pathway is a reasonable target for drug development. To our knowledge, this is one of the first demonstrations that delayed treatment can prevent a detachment-induced photoreceptor change, specifically rod terminal retraction. Most previous studies have used protocols in which therapies were applied concomitantly with or as much as 2 weeks ahead of the detachment.
53,57 Delayed treatment is clearly more clinically applicable. There are several drugs that target the RhoA pathway, some already approved for use in clinical trials for other CNS injuries.
58 –61 These approved drugs could be tested in animal models of retinal detachment. Further, increased understanding of the RhoA pathway may yield additional therapeutics for application to retinal injury and degeneration.