In addition to laser photocoagulation, currently used therapeutic interventions for diabetic retinopathy (DR) include relatively short-lived anti-VEGF drugs targeting vascular endothelial growth factor (VEGF). The latter requires frequent administration via intravitreal injections to effect long-term VEGF suppression. However, due to the patient burden associated with this treatment modality, gene therapy may represent a preferable alternative, providing long-lasting yet patient-friendly effects. Here, we explore the therapeutic efficacy of rAAV2-sVEGFRv-1, a recombinant adeno-associated virus encoding a soluble variant of VEGF receptor-1, upon early DR processes. Bevacizumab, an anti-VEGF agent often prescribed off label to treat DR, was used as an experimental comparator. Administered by intravitreal injection to a streptozotocin-induced diabetic mouse model, rAAV2-sVEGFRv-1 was shown to effectively transduce the mouse retinas and express its transgene therein, leading to significant reductions in pericyte loss and retinal cell layer thinning, two processes that play major roles in DR progression. Acellular capillary formation, vascular permeability, and apoptotic activity, the latter being the cell death mechanism by which retinal neurodegeneration occurs, were also shown to be reduced by the therapeutic virus vector. Immunohistochemistry was used to visualize that rAAV2-sVEGFRv-1 has an effect on cell types important to DR pathophysiology, particularly the ganglion cell layer and glial cells. Combined with our previous work showing that the therapeutic virus vector reduces neovascularization, our current results reveal that rAAV2-sVEGFRv-1 addresses the early aspects of DR as well, thereby demonstrating its potential as a human gene therapeutic versus the condition as a whole.

1. Introduction

Diabetic retinopathy (DR) is a commonly occurring complication of diabetes mellitus (DM) and ranks among the leading causes of blindness worldwide [1]. A degenerative condition presents at pathogenesis as nonproliferative diabetic retinopathy (NPDR) before progressing to proliferative diabetic retinopathy (PDR), the form of the disease more closely associated with irreversible vision loss. PDR is characterized by neovascularization (NV), a process driven by vascular endothelial growth factor (VEGF), an angiogenic factor whose expression is upregulated in the diabetic retina [2]. As an estimated one quarter [1] of the 592 million DM patients projected by 2035 [3] are predicted to develop DR to sight-threatening levels, it can be seen that DR is a looming global health concern.

Increased VEGF secretion occurs during DR as a result of the hypoxic retinal environment that develops due to DM-associated hyperglycemia [4]. Among the binding targets of VEGF are two fms-like tyrosine kinase receptors, Flt-1 [5], and KDR/Flk-1 [6], with the interaction with the latter being responsible for driving the angiogenic process. VEGF is additionally implicated in the development of diabetic macular edema (DME), a complication of DR that can occur during either NPDR or PDR [3]. A swelling or thickening of the macula, DME is the most common source of DR-related blindness [7]. Due to the central role of VEGF in the pathology of PDR and DME, which directly threaten the vision of DR patients, the standard of care for both conditions includes anti-VEGF therapeutics [3], as well as laser photocoagulation [4]. Multiple safety and efficacy issues exist for the latter, including complications that contribute to DR progression [8], whereas the former possesses significant limitations inherent to the treatments themselves [9].

Generally, anti-VEGF therapeutics consist of protein-based drugs that are delivered intravitreally [3] and include bevacizumab, a monoclonal antibody commonly utilized off-label for DR due to the significant economic benefits it offers; ranibizumab, a Fab fragment; and aflibercept, a recombinant fusion protein [1, 7, 10]. Comparative studies and meta-analyses have demonstrated that the treatment efficacies of these drugs are not significantly dissimilar [1113], with the exception of aflibercept for more severe manifestations of the disease. Where advances have been made, however, can be seen in the increased intervals between treatments for the dosing schedule of each successive drug. This is because frequent administration of these relatively short-lived therapeutics is required for long-term VEGF suppression [3, 4], made necessary by the progressive [2] and degenerative nature of DR [4]. However, this may be both procedurally and economically burdensome to the patients, negatively affecting compliance and treatment outcomes. Alternative dosing schedules, such as treat and extend and pro re nata, are currently being explored for these therapeutics to further lengthen treatment intervals, particularly for DME, without significant reductions in efficacy [10]. Yet even these extended schedules require multiple intravitreal injections per year, meaning that a treatment strategy providing anti-VEGF activity in a long-lasting and patient-friendly manner, for which gene therapy is ideally suited, may prove to be highly advantageous.

The therapeutic virus vector rAAV2-sVEGFRv-1 is a recombinant adeno-associated virus 2 (rAAV2) encoding a truncated form (sVEGFRv-1) of sFLT-1, a naturally occurring, alternatively spliced soluble variant of Flt-1 lacking the transmembrane domain found in the full-length VEGF receptor [14]. Nonpathogenic in nature and capable of transducing dividing and nondividing cells to elicit long-term transgene expression [15], rAAV2 vectors are particularly well-suited for delivering gene therapeutics targeting ocular conditions, as exhibited by the regulatory approval of voretigene neparvovec-rzyl, a treatment for Leber’s congenital amaurosis type 2 [16]. sFLT-1 has a high affinity towards VEGF and sequesters it from interacting with KDR/Flk-1 [14], which provides antiangiogenic activity, while also being able to directly associate with KDR/Flk-1, resulting in the formation of inactive heterodimers [17]. As a result, sFLT-1 is able to exert anti-VEGF effects via multiple mechanisms. The antiangiogenic activity of rAAV2-sVEGFRv-1 could be seen in our previous work [18], where it was demonstrated in a laser-induced mouse model of choroidal NV, among other effects.

Here, the therapeutic efficacy of rAAV2-sVEGFRv-1 was tested in comparison to bevacizumab in a streptozotocin- (STZ-) induced diabetic mouse model recapitulating the early processes of DR. After in vitro characterization of the therapeutic virus vector was conducted using human umbilical vein endothelial cells (HUVECs), rAAV2-sVEGFRv-1 was administered to the mice via intravitreal injection. There, it was shown to effectively transduce the mouse retinas and therein reduce pericyte loss, acellular capillary formation, vascular leakage, and the thinning of retinal cell layers while also having an antiapoptotic effect. Generally, the therapeutic virus vector compared favorably to bevacizumab, especially in its ability to reduce pericyte loss, an important early contributor to DR progression. Overall, our results show that rAAV2-sVEGFRv-1 demonstrates great promise as a potential human gene therapeutic for the treatment of DR.

2. Materials and Methods

2.1. Virus Vector Development

rAAV2-sVEGFRv-1 and rAAV2-GFP were prepared as previously described [18]. Briefly, the region spanning nucleotide positions 282 to 2,253 of the mRNA for human VEGF receptor-1 (XM_017020485.1; NCBI Reference Sequence, NIH) was inserted into a rAAV2 vector under the control of a CMV promoter along with a SV40 polyadenylation signal and both ITRs to produce rAAV2-sVEGFRv-1. rAAV2-GFP was generated as a negative control by an analogous insertion of an EGFP expression cassette. All virus vectors used in this study were sourced from CdmoGen Co., Ltd. (Cheongju, Korea).

2.2. Cell Culture and Infection

HUVECs (Lonza, Basel, Switzerland) were maintained in EGM-2 (Lonza) supplemented with EGM-2 Single Quots (Lonza). Cells were infected with their respective virus vectors at an MOI of 10,000.

2.3. Real-Time PCR

Total RNA was extracted from infected HUVECs using TRIzol (Thermo Fisher Scientific, Waltham, MA), per the manufacturer’s instructions, and the total RNA treated with DNase I (Thermo Fisher Scientific) to eliminate contamination by viral genomic DNA. cDNA was synthesized using M-MLV reverse transcriptase (Elpis Biotech, Daejeon, Korea) and real-time PCR performed using primers for the Flt-1 and β-actin genes, as follows: Flt-1 F: CGTGTAAGGAGTGGACCATC, Flt-1 R: TAAGACCGCTTGCCAGCTAC, β-actin F: TGAAGATCAAGATCATTGCTC, and β-actin R: TGCTTGCTGATCCACATCTG.

2.4. Tube Formation Assay

Control and rAAV2-sVEGFRv-1-infected HUVECs were seeded in a 96-well μ-plate (Ibidi, Gräfelfing, Germany) at 20,000 cells/well and the manufacturer’s protocol followed for the tube formation assay. VEGF treatment was performed using Recombinant Human VEGF 165 Protein (293-VE-050; R&D Systems, Minneapolis, MN) at a concentration of 10 ng/mL and preincubated with conditioned media for 30 minutes. Five hours post cell seeding, results were obtained, and the percentage of tube areas was calculated using ImageJ (National Institutes of Health, Bethesda, MD).

2.5. Animal Care

All animal care and experiments were overseen by the Internal Review Board for Animal Experiments at the Asan Institute for Life Sciences (University of Ulsan, College of Medicine) and performed in accordance with the Association for Research in Vision and Ophthalmology Resolution on the Use of Animals in Ophthalmic and Vision Research. Mice were housed in cages containing 5 animals apiece in the same area of a 25°C temperature-controlled room with free access to water and chow and an alternating photoperiod of 12 hours of light and dark.

2.6. STZ-Induced Diabetic Mouse Model

A well-established protocol [19] for inducing diabetes mellitus in mice via the administration of a single high dose of STZ was followed using 7-week-old male C57/BL6 mice sourced from the Orient Bio (Sungnam, Korea), with minor modifications. To mitigate the possibility of death due to STZ-induced toxicity, mice were injected intraperitoneally with 150 mg/kg of STZ (Sigma-Aldrich, St. Louis, MO), rather than the 200 mg/kg specified. Additionally, 10% sucrose water was provided only to mice that experienced significant deterioration, determined via reductions in body weight (Supplementary Table 1). Successful generation of the diabetic model was confirmed in mice whose tail vein blood samples, taken 1 week post-STZ treatment, had blood glucose levels exceeding 300 mg/dL (Supplementary Table 2), measured using Accu-Chek (Roche Diagnostics, Indianapolis, IN).

2.7. Intravitreal Injections

Mice were anesthetized 1 month post-STZ treatment via intraperitoneal injection of a 4 : 1 mixture of 40 mg/kg Zoletil (zolazepam/tiletamine) from Virbac (Carros Cedex, France) and 5 mg/kg of Rompun (xylazine) from Bayer Healthcare (Leverkusen, Germany). The pupils were dilated with Mydrin-P (0.5% tropicamide and 2.5% phenylephrine) from Santen (Osaka, Japan), and then, both eyes were injected intravitreally with 1 μL of either rAAV2-sVEGFRv-1 or rAAV2-GFP at a concentration of viral genomes (v.g.)/mL. Bevacizumab was administered to anesthetized mice at 25 mg/kg.

2.8. Sacrifice and Tissue Preparation

Sacrifice occurred at either 2 or 5 months post intravitreal injection. Mice were deeply anesthetized using a 4 : 1 mixture of Zoletil (80 mg/kg) and Rompun (10 mg/kg), after which intracardial perfusion with 0.1 M PBS (7.4 pH) containing 150 U/mL heparin and infusion with 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB) were performed. To generate eyecups, the eyeballs were enucleated and fixed in 4% PBA in 0.1 M PB for 1 hour, and the anterior sections were removed, including the cornea and lens, before being placed overnight in 30% sucrose in PBS. Prior to preparing 5-10 μm thick frozen transverse retinal sections, eyecups were embedded in Tissue-Tek (Miles Scientific, Napierville, IL), an optimal cutting temperature compound.

2.9. ELISA

To determine human sVEGFRv-1 concentrations in vitro and in vivo, culture media were collected from infected HUVECs and vitreous humor sampled 10 days after the intravitreal injection of rAAV2-sVEGFRv-1, respectively. ELISA was then performed using the Human VEGFR1/Flt-1 Quantikine ELISA Kit (DVR100C; R&D Systems), according to the manufacturer’s protocol.

2.10. Immunohistochemistry

An antibody for human VEGFR1/Flt-1 (AF321; R&D Systems) was used to immunostain section samples, and anti-NeuN (MAB377; Millipore, Burlington, MA) and anti-GFAP (12389; Cell Signaling Technology, Danvers, MA) were used to visualize retinal cells. The samples were incubated overnight with diluted primary antibodies at 4°C, followed by washing in PBST 3 times for 10 minutes apiece, incubation for 2 hours with the secondary antibodies Alexa Fluor 568 or 488 (Thermo Fisher Scientific) at room temperature, and staining with DAPI (D9542; Sigma-Aldrich). A LSM 710 fluorescence confocal microscope (Carl Zeiss Microscopy, Jena, Germany) was used to examine the section samples, and images were captured using the black edition of the Zeiss Zen software (Carl Zeiss Microscopy), followed by analysis using ImageJ.

2.11. Retinal Trypsin Digest

To observe the retinal vasculature, enucleated eyeballs were fixed for 24 hours in 10% formalin solution and washed in PBS before the retina was isolated and trypsin digestion performed [20], with minor modifications to the protocol. The retina was incubated in a 3% trypsin solution (15090046; Thermo Fisher Scientific) in 0.1 M Tris buffer (pH 7.8) for 1 hour at 37°C with gentle shaking, washed with water, and stained using conventional hematoxylin and eosin (H&E) methods. Light microscopy was used to determine the numbers of retinal pericytes and acellular capillaries in the center of the retinal area, with the former identified via its morphology (bound to capillaries and rounded nuclei). The number of pericytes and acellular capillaries was counted in 5 randomly selected areas of the retinal capillaries, standardized using the retinas of normal control mice, and the data were expressed as a ratio.

2.12. FITC-Dextran Staining

Vascular leakage was visualized via tail vein injection of 50 mg/mL of FITC-dextran (FD2000S; Sigma-Aldrich) 30 minutes prior to sacrifice and the fixing of enucleated eyeballs for 1 hour in 10% formalin solution. After washing with PBS, the retinas were isolated from the RPE-choroid complex and flat mounts generated via four equidistant cuts prior to observation via fluorescence microscopy (Eclipse Ti-U; Nikon, Tokyo, Japan), with the detailed images taken at 100x magnification.

2.13. Determination of Retinal Cell Layer Thinning

Frozen transverse sections, including the optic nerve head, were stained using H&E methods, and the cell layers were visualized using ImageJ. By determining, as a ratio, the changes of the inner retina from the nerve fiber layer to the inner nuclear layer, the extent of retinal cell layer thinning was calculated using the following equation: .

2.14. TUNEL Assay

TUNEL assay (12156792910; Roche Diagnostics) was performed following the manufacturer’s instructions, and the frozen sections were washed in PBST 3 times for 10 minutes apiece before being stained with DAPI to visualize the cell nuclei.

2.15. Statistical Analysis

Statistical analysis was performed using paired -testing, with significant difference determined versus sham at or , whereas significant difference versus bevacizumab was signified using double crosses (††). Data was visualized using dot plot graphs and includes significance and mean standard error of mean values.

3. Results

3.1. rAAV2-sVEGFRv-1 Is Able to Effectively Infect Endothelial Cells as well as Transduce Mouse Retinas to Express Its Transgene

72 hours post infection, total RNA and culture media were collected from infected HUVECs in order to determine the in vitro characteristics of the therapeutic virus vector. Real-time PCR (Figure 1(a)) showed that relative to uninfected control cells (), rAAV2-sVEGFRv-1 infection led to an increase in sVEGFRv-1 mRNA levels () while rAAV2-GFP treatment did not (), with a similar pattern of results obtained by ELISA (Figure 1(b)). The activity of the expressed sVEGFRv-1 was then examined via tube formation assay (Figure 1(c)), as endothelial cells under VEGF stimulation form tube-like structures prior to the development of NV [21]. Compared to uninfected and untreated control HUVECs (), tube formation was significantly elevated in VEGF-stimulated cells infected with rAAV2-GFP (), as well as uninfected HUVECs treated with VEGF (). On the other hand, cells stimulated with VEGF and infected with rAAV2-sVEGFRv-1 () exhibited levels of tube formation that were below even that of the control cells (Figure 1(d)).

Having thus examined the in vitro efficacy of rAAV2-sVEGFRv-1, an STZ-induced diabetic mouse model was established (Figure 2(a)) to explore the in vivo effects of the therapeutic virus vector. Retinal transduction was shown to occur from a single intravitreal administration of rAAV2-sVEGFRv-1 and was visualized by immunostaining transverse section samples with an antibody for the secreted sVEGFRv-1 (Figure 2(b)). ELISA showed that sVEGFRv-1 expression was significantly elevated (Figure 2(c)) in mice injected with rAAV2-sVEGFRv-1 ( pg/mL) when compared to the normal control group ( pg/mL).

3.2. rAAV2-sVEGFRv-1 Reduces Pericyte Loss, the Formation of Acellular Capillaries, and Leaky Vessel Development

Relative to normal control mice (), rAAV2-sVEGFRv-1 administration resulted in a marked reduction in the loss of pericytes (), which is among the earliest observed manifestations of DR [7] and demonstrated here via trypsin digestion assay (Figure 3(a), white arrowheads). The extent of this reduction, measured 3 months after establishing the model system and 2 months post intravitreal injection (Figure 3(b)), was greater than of mice treated with bevacizumab (), whereas pericyte loss readily occurred in both control rAAV2-GFP-treated mice () and the STZ-induced diabetic mouse model ().

Among other effects, pericyte loss may lead to retinal nonperfusion [7, 22], which may be observed in diabetic animal models in the form of acellular capillaries [22]. Acellular capillary formation (Figure 3(a), black arrows) was virtually absent in the normal control group () while being elevated in mice injected with the control virus vector () and the sham-treated diabetic mouse model (). In contrast, treatment with the therapeutic virus vector () and bevacizumab () both reduced the formation of these structures, with this effect being more pronounced in mice injected with rAAV2-sVEGFRv-1 (Figure 3(c)).

In turn, retinal nonperfusion is linked to the development of retinal vasodegeneration and the subsequent formation of vascular leakage, another key feature of the earlier stages of DR [23] and a process that may result in DR-related vision loss. Leaky vessels were visualized 2 months post intravitreal injection via FITC-dextran staining of retinal flat mounts (Figure 4(a)). In normalized to untreated control mice (), vascular leakage was observed in the sham- () and rAAV2-GFP-treated () control groups, but this process was significantly reduced when rAAV2-sVEGFRv-1 () and bevacizumab () were administered (Figure 4(b)).

3.3. Decreased Retinal Cell Layer Thinning and Antiapoptotic Activity Suggest That rAAV2-sVEGFRv-1 Is Neuroprotective

In addition to pericytes, other vascular and neuronal cells of the retina are involved in DR progression [2, 24], and H&E staining was used to visualize the thinning of the retinal cell layers (Figure 5(a)). Relative to normal control mice (), cell layer thinning was observed in the diabetic mouse model () and the rAAV2-GFP-treated control group (). However, this process was markedly reduced (Figure 5(b)) by rAAV2-sVEGFRv-1 treatment (), with this effect being especially pronounced in the inner nuclear layer, suggesting that the therapeutic virus vector may have neuroprotective qualities. However, comparisons to bevacizumab were not made here, as retinal cell layer thinning was determined 5 months post intravitreal injection in order to examine the long-lasting effects of rAAV2-sVEGFRv-1.

Meanwhile, a TUNEL assay performed 2 months after virus vector treatment showed that apoptotic cells were almost wholly absent (Figure 6(a)) in normal control mice (), whereas TUNEL-positive cells were found speckled throughout the retinal cell layers of the STZ-induced diabetic mouse model () and mice treated with rAAV2-GFP (). On the other hand, the administration of rAAV2-sVEGFRv-1 () and bevacizumab () both resulted in a reduction in observed apoptotic activity (Figure 6(b)).

3.4. rAAV2-sVEGFRv-1 Protects against Glial Cell Activation and the Loss of Ganglion Cells in Diabetic Mouse Retinas

The effects of rAAV2-sVEGFRv-1 upon the diabetic retina were additionally visualized using anti-NeuN (Figure 7(a)) and anti-GFAP (Figure 7(b)), which stain ganglion cells and glial cells, respectively. While NeuN-positive cells were found in both sham-treated mice () and those injected with the control virus vector (), the greatest number was observed (Figure 7(c)) in the retinas of mice treated with rAAV2-sVEGFRv-1 (), bevacizumab (), or the normal control group ().

Conversely, GFAP expression was elevated in rAAV2-GFP-treated mice () and the sham-treated diabetic mouse model (), indicating higher amounts of active glial cells. Relative to the latter, GFAP expression was markedly reduced (Figure 7(d)) in mice injected with bevacizumab () and rAAV2-sVEGFRv-1 (), as well as normal control mice ().

4. Discussion

Due to the limitations of laser photocoagulation and/or anti-VEGF therapeutics, currently the most commonly employed treatment options for DR and DME, the development of a gene therapy alternative [25] may provide the answer in effecting safe and long-term VEGF suppression while reducing patient burden. Laser-based treatment, in the form of panretinal photocoagulation, has been linked to the formation of subretinal fibrosis [26], vascular leakage, and choroidal NV [8], all of which are implicated in DR pathoprogression, whereas anti-VEGF DR drugs, in addition to risks associated with the injection procedure [4], are short-lived while requiring multiple loading injections to begin exerting a therapeutic effect [27]. Yet despite newer DR therapeutics continuing to be and having been developed and approved, directly targeting VEGF remains unchanged as the treatment strategy.

As such, we initially explored the therapeutic potential of intravitreally injected rAAV2-sVEGFRv-1, its antiangiogenic activity in particular, in a laser-induced mouse model of choroidal NV, an animal model system for the wet subtype of age-related macular degeneration (wAMD) [18]. There, it was shown to reduce the extent to which choroidal NV occurred at levels comparable to bevacizumab, in addition to being antiapoptotic and reducing inflammatory cell infiltration. The relevance of this study can be linked to the developmental trajectory of anti-VEGF DR drugs, the most common of which were originally used to treat wAMD before being indicated for PDR and DME, as wAMD, like PDR, is characterized by VEGF-driven NV leading to vision loss [10]. A number these therapeutics subsequently received regulatory approval for all stages of DR [28, 29], and here, we explored the effects of rAAV2-sVEGFRv-1 in a STZ-induced diabetic mouse model. While STZ-induced diabetic mice do not recapitulate the vascular aspects of DR [30], they accurately portray early processes of DR pathophysiology that work to influence disease progression, including the loss of pericytes and resulting retinal vessel dysfunction [31], as well as neurodegeneration [23]. Bevacizumab was used as an experimental comparator in both studies as it was among the most widely used anti-VEGF agents up to the very recent past [10].

As a key histopathological characteristic of DR [32], pericyte loss is linked to the hyperglycemic conditions present in the diabetic retina and is observed early on in the pathophysiology of DR [7]. Due to the important role played by pericytes in normal retinal function [32], their loss is an influential factor in DR pathoprogression [22]. With pericytes playing a crucial role in maintaining the structural integrity of retinal vessels [22], their loss may lead to other vessel abnormalities, including capillary occlusion, ischemia [7], and/or nonperfusion [22]. As a result, the retinal environment may become hypoxic, leading to hypoxia inducible factor 1 (HIF-1) activation and the upregulation of its downstream target, VEGF [7]. In addition to its role driving NV during PDR, VEGF contributes to DR progression in the early stages of the condition [33]. Pericyte loss and VEGF are also linked to the formation of leaky vessels as a result of retinal vasodegeneration [22, 34]. This may be due to the role pericytes play in proper tight junction formation [22], as several studies have suggested that VEGF induces conformational changes upon tight junctions [4], leading to vascular leakage and potential DR-related vision loss. In addition to showing that rAAV2-sVEGFRv-1 has anti-VEGF activity (Figure 1), the therapeutic virus vector was additionally demonstrated here to reduce pericyte loss (Figure 3) and the subsequent development of leaky vessels (Figure 4).

However, retinal pericytes are not the only cell type affected by DR, with the condition also influencing endothelial cells, glial cells, and ganglion cells, which collectively form the neurovascular unit (NVU) [2, 24]. In some respects, DR can be considered a disease of the NVU [2], with NVU neurodegeneration constituting a major component of DR progression [35]. Retinal neurodegeneration is characterized by the loss of various amacrine cells [2] and ganglion cells, inner nerve fiber layer thinning [23], and glial cell activation and dysfunction [36]. By determining the changes to the thickness of the inner retina, which includes the ganglion cell layer and inner nerve fiber layer, rAAV2-sVEGFRv-1 was demonstrated to reduce retinal cell layer thinning (Figure 5) while specifically protecting against the loss of retinal ganglion cells, which are particularly susceptible to neurodegenerative effects [23]. The latter was visualized via anti-NeuN staining, whereas the low levels of GFAP expression resulting from rAAV2-sVEGFRv-1 administration (Figure 7) showed that the therapeutic virus vector reduced glial cell activation and activity [36]. rAAV2-sVEGFRv-1, which previously demonstrated its antiapoptotic activity in a laser-induced mouse model of choroidal NV [18], was confirmed here to suppress apoptosis (Figure 6), an important contributor to DR pathophysiology [35] and the cell death mechanism by which neurodegeneration occurs [32]. Apoptosis has also been implicated in the development of vascular leakage [7, 32]. Taken together, these results strongly suggest that the therapeutic virus vector has neuroprotective qualities, thereby addressing another major contributor to DR pathophysiology and progression.

5. Conclusions

Therefore, in combination with our previous work demonstrating the antiangiogenic properties of rAAV2-sVEGFRv-1, the ability of the therapeutic virus vector to address early aspects of DR pathophysiology exhibits its great promise as a potential human gene therapeutic versus DR. rAAV2-sVEGFRv-1 generally compared positively to currently used treatments, particularly in reducing pericyte loss, a consequential process in DR progression. Future studies for the continuing development of rAAV2-sVEGFRv-1 as a human gene therapeutic include long-term safety and efficacy studies in nonhuman primates and are currently being planned.

Data Availability

Data is available upon request from the authors.

Conflicts of Interest

S.H.S.L., J.-S.C., H.J.K., J.K., S.C., and K.P. are employees of CdmoGen Co., Ltd., in which S.H.S.L. and K.P. have personal financial interests. No other potential conflicts of interest relevant to this article were reported.

Authors’ Contributions

H.L. and K.P. were responsible for the conceptualization. H.L., J.Y.L., S.H.S.L., I.K.K., H.J.K., J.K., and S.C. were responsible for the methodology. H.L., J.Y.L., S.H.S.L., I.K.K., J.-S.C., H.-N.W., K.J.L., and K.P. were responsible for the investigation. H.L., J.Y.L., S.H.S.L., I.K.K., J.-S.C., and K.P. were responsible for the writing—original draft. H.L., J.Y.L., S.H.S.L., I.K.K., H.-N.W., K.J.L., and K.P. were responsible for the writing—review and editing. Steven Hyun Seung Lee, Joo Yong Lee, and Im Kyeung Kang are co-first authors.


We would like to acknowledge S.H.S.L.’s dissertation, “Inhibiting the mechanistic target of rapamycin (mTOR) via RNA interference as a novel human gene therapeutic strategy for the treatment of diabetic retinopathy,” which also discusses the development of a gene therapeutic for diabetic retinopathy. This work was supported by grants from TIPS (Korea Tech Incubator Program for Startup, administered by the Republic of Korea Ministry of SMEs and Startups) to CuroGene Life Sciences Co., Ltd, which was recently acquired by CdmoGen Co., Ltd., and from the National Research Foundation of Korea (NRF-2021R1A2C1093614) to H.L.

Supplementary Materials

Supplementary Table 1: body weights (g) of the STZ-induced diabetic mouse model. Body weight data of the mice from which the STZ-induced diabetic mouse model was generated. Supplementary Table 2: blood glucose levels (mg/dL) of the STZ-induced diabetic mouse model. Blood glucose data of the mice from which the STZ-induced diabetic mouse model was generated. (Supplementary Materials)