Retinal inflammation is well known to contribute to the pathogenesis of retinal diseases, such as AMD, DR, or RP [138,139,140,141]. Inflammation is one of the important functions that are regulated by GSK3 since it is well established that GSK3 acts as a modulator of inflammatory components [38,142,143]. For instance, GSK3 negatively regulates anti-inflammatory cytokine production such as IL-2, IL-10, IL-22, or IL-33 [144]. Conversely, this kinase acts as a positive regulator of pro-inflammatory cytokines and chemokines, such as TNF-α, interleukin (IL-)1β, IL-6, IL-17, IL-18, IL-23, IL-12, IFN-γ, IL-8, C-C motif chemokine ligand (CCL) 2, 3, 4, and 12, C-X-C motif chemokine ligand (CXCL) 1, 2, 5, and 10 [144]. Together, this raised the possibility that inhibitors of GSK3 may prove to be beneficial for inflammatory conditions. One signaling molecule of interest targeted by GSK3 is NFκB, a pivotal mediator of inflammatory responses that has long been proposed as a potential target for the therapy of inflammatory diseases [145,146]. It was shown that NF-κB is activated in rd mice and light-induced retinal degeneration [147,148] Interestingly, GSK3β was shown to facilitate NF-κB transactivation by TNF-α since GSK3β deficient mouse embryonic fibroblasts exhibit defective NF-κB activation in response to TNFα [66]. Mechanistically, direct phosphorylation of NF-κB subunits p65 by GSK3 was reported in hepatocytes [149]. Along the same line, GSK3 inhibition in microglial cells decreases LPS-induced inflammation through the decrease of the activation of p65 [150]. GSK3 also activates non-canonical NF-κB signaling through the phosphorylation of p100, an inhibitor of NF-κB, targeting it to the proteasome [151]. Moreover, the increased β-catenin levels following GSK3 inactivation may further enhance NF-κB inhibition, since β-catenin was shown to inhibit NF-κB activity through physical interaction [152]. It is however noteworthy that GSK3 could differentially regulate NF-κB activity depending on the physiological state of the cell. Indeed, although it is required for the activation of NF-κB in response to cytokine stimulation [153], it may inhibit NF-κB in resting cultured cells [153,154,155]. Moreover, NF-κB can exhibit anti-apoptotic effects. Indeed, mice with inactivated GSK3β die from hepatocyte apoptosis during development due to a defect in NF-κB activation [66]. Finally, although GSK3 inhibition in different models predominantly contributes to the amelioration of inflammation, it may alternatively lead to the opposite effect as it could also prohibit the termination of inflammation [144]. As a whole, it is clear that more knowledge is needed to evaluate the net outcome of GSK3 inhibitors on NF-κB activity and on neuroinflammation in general in different models of retinal degeneration.

Another interesting factor linked to inflammation and GSK3 is the P2X7 receptor (P2X7-R) which is known to promote chronic neuroinflammation and neurodegenerative brain diseases [156]. In AD mouse models, P2X7-R inhibition was shown to have a protective effect through GSK3 inhibition [157,158]. P2X7-R is expressed in the retina and the RPE, and during retinal degeneration its expression increases [159]. P2X7-R induces the expression of inflammatory factors in the retina [160,161]. Interestingly, the pharmacological inhibition of P2X7-R prevents the increased inflammation and neovascularization induced by oxidative stress in the mouse eye in vivo [162]. It would be interesting to know whether some of these effects on inflammation and neuroprotection involve GSK3 inhibition.

Retinal neovascularization is observed in some retinal degenerative diseases, such as DR or wet AMD, where there is a disruption of the blood-retinal barrier (BRB). This phenotype is correlated with the increased expression of Vascular Endothelial Growth Factor (VEGF), a target gene of Wnt signaling [163,164]. A commonly used treatment strategy for wet-AMD and DR relies on reducing angiogenesis via anti-VEGF agents [165,166,167,168,169,170,171].

Wnt signaling activation promotes retinal vascularization [172,173,174], BRB development, and maintenance [175]. Accordingly, the reduction of Wnt signaling by the loss of LRP5, a canonical Wnt co-receptor, suppresses pathologic neovessel formation in a mouse model of oxygen-induced retinopathy [176]. Even if the loss of LRP5 negatively regulates retinal neovascularization in development and adulthood, there are still some vessels formed, with great disorganization and blood leakage [177]. Acting on downstream effectors of Wnt signaling could be another strategy to prevent neovascularization. In this pathological context, promoting GSK3 activity might thus be an interesting alternative approach to inhibit Wnt signaling and subsequently diminish VEGF expression.

In contrast to the occurrence of neovascularization in some retinal dystrophies, other eye diseases are characterized by the poor formation of intraocular vasculature, such as the familial exudative vitreoretinopathy (FEVR) [178]. Consistent with the importance of Wnt signaling in retinal vascular development, Wnt inhibition in Lrp5^−/− mice produces eye vascular pathologies that model FEVR in humans [179]. Importantly, inhibition of GSK3 by LiCl treatment in this mouse model was shown to rescue defective retinal vasculature through restoring Wnt signaling, providing a potential treatment approach for FEVR [179].

Cell-based therapy is an appealing approach in late-stage retinal degeneration when most cells are already dead. One approach relies on the stimulation of endogenous repair processes. Some species have a high regenerative capacity, such as zebrafish or Xenopus, in which MGCs exhibit stemness properties. After retinal damage, dormant Müller cells can exit quiescence, proliferate and differentiate into different retinal cell types [180,181,182]. In contrast with fish and amphibians, these stemness and neurogenic capacities are highly limited in the mammalian retina [183,184]. The goal is to identify cellular pathways able in mammals to trigger Müller glia reprogramming and differentiation of Müller-derived progenitors under pathological conditions.

Several results gathered in different animal models converge on the idea that regeneration of retinal neurons can be promoted by applying GSK3 inhibitors to the retina. Zebrafish regenerative capacity observed after retinal damage is mediated by Ascl1 [185], which induces the Wnt signaling pathway [186]. Remarkably, in zebrafish undamaged retinas, Wnt activation, via GSK3 inhibition with LiCl, is sufficient to stimulate MGCs proliferation and induce retinal regeneration [186]. In the chick retina, which has a low regenerative potential, GSK3 drug-inhibition associated with FGF2 treatment promotes MGCs proliferation and dedifferentiation [187]. In the rat retina, although some MGCs are stimulated to proliferate and produce retinal cells in retinal explants, this remains very limited [188]. Activation of Wnt signaling, either by Wnt3a treatments or using SB216763 or AR-A014418 as inhibitors of GSK3β, was shown to promote the proliferation of Müller glia-derived retinal progenitors and neural regeneration in the wild-type retina [189] (Figure 1e). However, in rd mice, similar approaches induce MGC proliferation only at P12 but not at a later stage suggesting that the retinal microenvironment variation under pathological conditions might contribute to the lack of retinal repair observed in mammals [189]. Along the same line, the proliferative response of MGCs in mouse retinal explants following the addition of the GSK3 inhibitor Chir99021 varies between mouse strains, highlighting the importance of the genetic background [190]. Nevertheless, altogether these data in different models and species suggest that targeting GSK3 for retinal cell regeneration deserves further attention.

Given the critical function of GSK3 in regulating axon growth, modulation of GSK3 activity may also represent an interesting strategy to trigger axon regeneration following injuries [7,191,192,193,194]. Interestingly enough, antagonist effects were observed on GSK3 inhibition and axon growth. GSK3 activation promotes peripheral nerve axon growth, whereas GSK3 inhibition promotes CNS axon growth [195]. Lack of phosphatase and tensin homolog expression (Pten^−/−) is neuroprotective and enhances RGC axon regeneration. Indeed, in this model, RGC axonal regeneration in mature neurons is promoted by activation of mTOR [196]. Such a mechanism relies essentially on GSK3 inactivation supporting that GSK3 inhibitors could serve as a regenerative stimulus.

Axon regeneration is an appealing approach for preserving the optic nerve in glaucoma [197]. Activation of Wnt signaling through intravitreal administration of Wnt3a after an optic nerve crush (ONC), a classical axon injury paradigm, contributes to axonal regeneration [198]. Consistent with this, GSK3β inhibition also enhances optic nerve regeneration after ONC [195]. This effect is mediated by the Collapsin response mediator protein 2 (CRMP2), a microtubule-binding protein involved in neuronal polarization, migration, and differentiation. Therefore, GSK3/CRMP2 axis is a pathway of interest to treat glaucoma by promoting axon regeneration [196]. Another interesting factor involved in regeneration is mTOR. mTOR positively regulates cell growth, proliferation, and survival, and this is mediated by activation of Wnt pathway, thus GSK3 inhibition [199]. Similarly, mTOR promotes axon regeneration in the CNS through GSK3 inhibition [200]. In the retina, after an ONC and inflammatory stimulation, mTOR signaling enhances optic nerve regeneration as well as RGCs neuroprotection [201]. Thus, inhibiting GSK3 is an appealing strategy for optic nerve regeneration as it should enhance the activity of key downstream effectors, in particular CRMP2 and mTOR.

4. Precautions and Advantages of Inhibiting GSK3 or GSK3 Targets as a Therapeutic Strategy for Eye Diseases