Table 4.

Summary of in vivo studies utilizing an AAV vector encoding different genes as gene therapy.

Reference/Study	Country	Injury Model and Animal	Groups and Sample Size	Parameters for Efficacy	Main Result	Remarks
4. Viral vector encoding other genes
a. scAAV2-C3 gene therapy Tan J. et al., 2020 [21]	China	I/R injury induced by elevated IOP by saline injection, rats	4 negative control, 4 injury only, 4 GFP only, 4 injury + GFP, 4 treatment only, 4 injury + treatment	RhoA inhibition Apoptosis measured via TUNEL positive and cleaved caspase-3 positive cells Retinal neuron and RGC counts via NeuN and Brn3a staining Retinal thickness	Injury + treated rats had significantly lower protein levels of RhoA compared to negative control and GFP only rats Injury + treated rats had significantly lower TUNEL positive cells compared to injury + GFP rats and injury only rats. Injury only and injury + GFP groups had a significantly higher number of caspase-3 positive cells compared to negative control, and injury + treated rats had significantly reduced caspase-3 positive cells compared to injury only and injury + GFP groups. NeuN positive cells were significantly rescued in the injury + treatment group compared to both injury only and injury + GFP rats. Brn3a positive cells were significantly rescued in the injury + treatment group compared to injury + GFP rats. Injury + treated mice were able to significantly attenuate retinal thickness loss compared to injury only and injury + GFP rats.	RhoA has involvement in apoptosis and C3 has been shown to inhibit Rho and protect RGCs from NMDA induced damage. There was an absence of TUNEL positive cells in the groups with no injury. Injury only and negative control rats were not used in the Brn3a experiments. There was still a significant difference between injury + treated rats and the negative control and GFP only rats, indicating only a partial attenuation.
b. AAV-mSncg promoter-driven CRISPR/Cas9 gene therapy Wang Q. et al., 2020 [22]	United States	ON crush injury model, mice	8 control gRNA in LSL-Flag-SpCas9 mice, 9 target gRNA in LSL-Flag-SpCas9 mice, 8 control gRNA in WT mice, 9 target gRNA in WT mice	AAV-mSncg promoter and GFP expression Ddit3 and Sarm1 expression Ganglion cell complex (GCC) thickness, RGC somata, and RGC axons	The mouse Scng promoter (AAV-mSncg-EGFP) binds and targets hPSC-derived human RGCs expressing tdTomato based on merge labelling (of GFP and tdTomato) significantly more than the human Scng promoter (AAV-hSncg-EGFP). Mice treated with AAV-mSncg-CRISPR/Cas9 with Ddit3 and Sarm1 guide RNA had reduced expression of Ddit3-mCherry and Sarm1-mCherry. Mice treated with AAV-mSncg-CRISPR/Cas9 with Ddit3 and Sarm1 guide RNA after ON injury had significantly thicker GCC, increased RGC somata, and RGC axons compared to non-treated mice.	Yellow fluorescence was indicated as proper targeting of inherent tdTomato expression in the hPSC-derived human RGCs and appropriate AAV-mSncg-EGFP targeting. Only the expression of Ddit3 mCherry was significantly decreased with AAV-mSncg-CRISPR/Cas9 with Ddit3 and Sarm1 guide RNA treatment. Sarm1 mCherry expression was not assessed in the same way. There were significantly more RGC somata in treated WT mice compared to treated LSL-Cas9 mice with ON injury.
c. AAV2-pigment epithelium-derived factor (PEDF) + human mesenchymal stem cell (hMSC) gene therapy Nascimento-dos-Santos G. et al., 2020 [23]	Brazil	ON crush injury model, rats	9 control, 5 GFP only, 7 PEDF only, 9 PEDF + hMSC, 3 GFP + hMSC	RGC neuroprotection assessed via Tuj1+ cells Neuroprotective factor FGF-2 and IL-1β Microglial response and Müller glia activation Axonal outgrowth	Injured mice that were treated with PEDF had a significantly higher number of Tuj1+ cells compared to GFP only. Injured mice treated with PEDF had significantly higher levels of FGF-2 and significantly lower IL-1β compared to GFP only. Injured mice treated with PEDF had a significantly lower number of Iba1+ cells and Müller GFAP+ cells compared to GFP only. PEDF treatment after injury did not stimulate RGC axonal regrowth. However, the combination of PEDF + hMSC was able to significantly increase the number of Tuj+ cells and CTB+ axons.	Significance was not reported between PEDF treated and control mice, although it seems that PEDF treatment was not able to increase Tuj+ cells to control levels. There was no significant difference between control and PEDF treatment groups, suggesting levels comparable to normal state. Despite a significant reduction in Iba+ and Müller GFAP+ cells, PEDF treated mice still had a significantly higher number compared to controls. CTB was used as an anterograde tracer of RGC axons and this increase in CTB+ axons was only significant at 0.25 mm from the injury site and comparable to GFP only at further distances.
d. AAV2-CR2-Crry gene therapy Bosco A. et al., 2018 [24]	United States	DBA/2J mouse model of glaucoma	25 control, 26 GFP only, 34 treatment	IOP measurement RGC degeneration rates measured via Brn3a Axon fasciculations Optic nerve pathology via light microscopy images	Mice that were treated did not have any difference in expected IOP elevation compared to aged control and GFP only mice. Mice that were treated had significant differences in Brn3a+ densities compared to control and GFP only mice. Mice that were treated had significantly increased axon fasciculations compared to control and GFP only mice. Mice that were treated had significantly attenuated optic nerve pathology compared to control and GFP only mice.	DBA/2J mice develop progressive RGC degeneration as they age. These differences are based on Brn3a+ density and categorized into pre-degenerative, declining, and degenerative. Treated mice seemed to have an increased proportion and number of cells in the pre-degenerative and declining category, but these absolute numbers were not separately assessed. There was no significant difference between control and GFP only mice. Optic nerve pathology was assessed through light microscopy sections and may have subjective measuring as an established marker for ON pathology was not used.
e. AAV-STC-1 gene therapy Roddy GW. et al., 2017 [25]	United States	Transgenic mice models of retinal degeneration (P23H-1 and S334ter-4)	7 negative control P23H-1 mice, 8 negative control S334ter-4 mice, 7 treatment + P23H-1 mice, 8 treatment + S334ter-4	Optic nerve length (ONL) thickness and ERG responses Gene expression changes	Both P23H-1 and S334ter-4 treated mice had significantly thicker ONL compared to untreated transgenic mice. However, only P23H-1 treated mice had significant improvement in scotopic and photopic B-waves in ERG responses compared to untreated P23H-1 mice. S334ter-4 treated mice had significant gene expression changes with many genes involved in neuroprotection and anti-ROS.	Only P23H-1 treated mice showed improvement in ERG responses. These same genes were not upregulated in P23H-1 mice.
f. AAV2-Neuroglobin (NGB) gene therapy Cwerman-Thibault H. et al., 2017 [26]	France	DBA/2J mouse model of glaucoma	31 control (untreated), 18 early treatment (2 months), 12 late treatment (8 months)	RGC count via Brn3a+ cells Retinal stress measured via GFAP Morphological changes in neurons measured via Brn3a and β3-tubulin	Early treatment with AAV2-NGB significantly increased RGC counts compared to control. GFAP fluorescence intensities were reduced in both early and late treatments compared to untreated control. Early treatment appeared to maintain RGC density and dendrite branching, comparable to young untreated mice. In late treatment, dendritic profiles were better preserved than untreated and age matched mice.	Late treatment showed no increase in RGC counts compared to control. GFAP is used as a marker for retinal stress. Despite using actual markers to potentially assess fluorescence intensity or physical counts, these results were qualitative in nature.
g. AAV2-serum soluble Fas Ligand (sFasL) gene therapy Krishnan A. et al., 2016 [27]	United States	DBA/2J mouse model of glaucoma or microbead induction of elevated IOP in mice	10 negative control, 10 positive, control 10 GFP only, 10 treatment	IOP measurement RGC density and axon density Pro-apoptotic and anti-apoptotic mediators RGC density and axon density	There was no significant difference in IOP at months 3, 5, 7 or 9 (all measured time points) between positive control, GFP only, and treatment groups. RGC density and axon density were significantly increased with treatment compared to positive control and GFP only mice. Treated mice had significantly reduced mRNA levels of pro-apoptotic mediators, Fas, FADD, and BAX compared to GFP only. Treated mice had significantly increased mRNA levels of anti-apoptotic mediators cFLIP, Bcl2, and cIAP2 compared to GFP only. RGC density and axon density were significantly increased with treatment compared to positive control and GFP only mice.	Treatment had no improvement on IOP, and qualitatively, there was no improvement in pigment dispersion and iris atrophy. This increase in RGC and axon density was not significantly different from the negative control, suggesting protection comparable to normal levels. GFAP and TNFα levels were also reduced, and along with qualitative confocal microscopy of Müller glial cells, suggest reduced Müller glial cell activation in treated mice compared to GFP only. RGC and axon density was assessed in a microbead induced elevation of IOP model.