Overactivation of Norepinephrine–β2-Adrenergic Receptor Axis Promotes Corneal Neovascularization

Abstract

Purpose: To investigate the role of the sympathetic nervous system in corneal neovascularization (CNV) and to identify the downstream pathway involved in this regulation.

Methods: Three types of CNV models were constructed with C57BL/6J mice, including the alkali burn model, suture model, and basic fibroblast growth factor (bFGF) corneal micropocket model. Subconjunctival injection of the sympathetic neurotransmitter norepinephrine (NE) was administered in these three models. Control mice received injections of water of the same volume. The corneal CNV was detected using slit-lamp microscopy and immunostaining with CD31, and the results were quantified by ImageJ. The expression of β2-adrenergic receptor (β2-AR) was stained with mouse corneas and human umbilical vein endothelial cells (HUVECs). Furthermore, the anti-CNV effects of β2-AR antagonist ICI-118,551 (ICI) were examined with HUVEC tube formation assay and with a bFGF micropocket model. Additionally, partial β2-AR knockdown mice (Adrb2^+/−) were used to establish the bFGF micropocket model, and the corneal CNV size was quantified based on the slit-lamp images and vessel staining.

Results: Sympathetic nerves invaded the cornea in the suture CNV model. The NE receptor β2-AR was highly expressed in corneal epithelium and blood vessels. The addition of NE significantly promoted corneal angiogenesis, whereas ICI effectively inhibited CNV invasion and HUVEC tube formation. Adrb2 knockdown significantly reduced the cornea area occupied by CNV.

Conclusions: Our study found that sympathetic nerves grow into the cornea in conjunction with newly formed vessels. The addition of the sympathetic neurotransmitter NE and activation of its downstream receptor β2-AR promoted CNV. Targeting β2-AR could potentially be used as an anti-CNV strategy.

Corneal neovascularization (CNV) is a sight-threatening condition that frequently arises from infection, hypoxia, trauma, corneal degeneration, or surgery.¹ Approximately 1.4 million people suffer from CNV each year, with 12% of them experiencing subsequent vision loss.² The cornea is characterized as transparent and avascular, with its avascularity maintained by a balance between angiogenic and antiangiogenic factors.^3,4 Disruption of this balance can lead to CNV.^5,6 At present, there is no effective drug treatment for CNV. Although steroids can reduce inflammation and limit CNV growth, they do not provide a complete cure,⁷ and they have side effects such as steroid-induced glaucoma, accelerated cataract formation, and increased infection risk. Anti-vascular endothelial growth factor (anti-VEGF) therapies are commonly used,^8–11 but they are less effective in treating corneal angiogenesis than retinal neovascularization.^12,13 Corneal transplantation is also an option, but the presence of CNV increases the risk of graft failure by 30%,^7,14 with higher CNV levels correlating with higher rejection risk.⁹ Therefore, the development of new targets and treatments to eliminate CNV prior to surgery is crucial. 

Various animal models have been developed to investigate CNV, with three models being commonly used: the suture model, the alkali burn model, and the basic fibroblast growth factor (bFGF) model. The suture model imitates CNV caused by sutures during corneal surgeries, particularly during corneal transplantation. The alkali burn model simulates chemical burn injuries on the cornea, resulting in severe corneal damage, corneal limbal stem cell deficiency, severe inflammatory reaction, and CNV.^15–17 Finally, the bFGF micropocket model involves implanting a bFGF pellet into the corneal stromal pocket, which can effectively induce CNV with a limited inflammatory reaction, thus making the model ideal for experimental observation.^17–20 

The development of CNV may be related to corneal nerve degeneration. In certain cases, such as herpes simplex keratitis (HSK), CNV can develop with decreased corneal sensation resulting from herpes simplex virus invasion of the trigeminal ganglion.^21,22 Neurotrophic keratopathy is another corneal disease that can present a CNV phenotype,^23,24 but the mechanism through which neuronal damage induces CNV is not yet fully understood. The cornea is a highly innervated tissue, with a density of nerve endings hundreds of times greater than that of the skin.²⁵ Most of the corneal nerve fibers are sensory nerves that originate from the ocular branch of the trigeminal nerve and the maxillary nerve,²⁶ but a small portion of corneal nerve fibers consists of sympathetic nerves originating from the superior cervical ganglion.^27,28 The inward growth of sympathetic nerves into the cornea has been observed during HSK infection.²⁹ 

In our previous study, we discovered the regulatory function of sympathetic nerves in intraocular inflammation after cataract surgery.³⁰ Considering that inflammation is a key contributor to CNV,^31,32 it is worthwhile to investigate the potential relationship between sympathetic innervation and CNV. A recent study suggested that the continuous activation of the sympathetic nerves could encourage angiogenesis in ovarian cancer mouse model.³³ Additionally, the sympathetic neurotransmitter norepinephrine (NE) could promote the growth of new vessels in tumors.³⁴ Adrenergic receptors (ARs) are G-protein-coupled receptors downstream of NE. Both β1- and β2-adrenergic receptors (β1/2-ARs) play crucial roles in retinal angiogenesis.³⁵ However, topical application of the nonselective β-AR blocker propranolol could not inhibit alkali-induced CNV in rat.³⁶ Concomitant systemic β-adrenergic blocking reduced the need for repeated intravitreal injections of bevacizumab in patients with choroidal neovascularization.³⁷ Blocking β2-AR could diminish choroidal neovascularization by inhibiting VEGF and interleukin 6 expression.^38,39 

In this study, sympathetic nerve ingrowth into the cornea was observed in mouse CNV corneas. NE promoted vessel formation both in vitro and in vivo. Medical and genetic blockage of corneal β2-AR effectively dampened CNV growth. As a result, this study revealed the regulatory role of sympathetic nerves in CNV and the potential of targeting β2-AR in treating CNV. 

Methods

Mouse Models

C57BL/6J mice, 6 to 8 weeks old, were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). Adrb2 knockdown mice (Adrb2^+/−) were developed by GemPharmatech (Nanjing, China). All animal procedures were conducted in compliance with the guidelines of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Animal Investigation Committee of the Eye Institute of Shandong First Medical University. 

Corneal Suture Model

Mice were randomly assigned to either the control group or the experimental group. Six hours before CNV induction, the experimental group received 1 mg/mL NE (Grand Pharma Co., Ltd., Wuhan, China) via subconjunctival injection; the control group was injected with an equal volume of sterile water. The mouse was anesthetized, and the cornea was marked by gently placing a 2-mm trephine at the central cornea. Three 11-0 nylon sutures were placed at the 4, 8, and 12 o'clock positions. From the second day after suture, 1 mg/mL NE or injection water was topically administered six times a day. Eye phenotypes were assessed with a slit lamp on days 1, 3, 5, and 7 after the operation. 

Corneal Alkali Burn Model

The mice were randomly assigned to either the control group or the experimental group. Six hours before CNV induction, the experimental group received 1 mg/mL NE via subconjunctival injection; the control group was injected with an equal volume of sterile water. The mice were anesthetized and placed beneath a stereoscopic microscope. A 2-mm-diameter round filter paper was soaked in 1-mol/L sodium hydroxide solution for 50 seconds and then placed on the central corneal surface of the right eye of each mouse for 50 seconds to induce alkali burn, followed by saline flushing for 50 seconds. Then, 1 mg/mL NE or injection water was topically applied six times a day beginning with the second day after the alkali burn. Eye phenotypes were recorded with a slit lamp on days 1, 3, 5, and 7 after the operation. 

bFGF Pellet Preparation and Corneal Micropocket Assay

Pellets were prepared using poly(hydroxyethyl methacrylate) (poly-HEMA; Sigma-Aldrich, St. Louis, MO, USA), sucralfate (20 ng/pellet; Sigma-Aldrich), and mouse bFGF (80 ng/pellet; MedChemExpress, Monmouth, NJ, USA). A 12% (w/v) poly-HEMA was prepared with absolute ethanol, and a 10% (w/v) sucralfate solution was prepared with PBS. For 50 bFGF pellets, 5 µL 12% (w/v) poly-HEMA, 1 µL 10% (w/v) sucralfate solution, and 4 µL 1 µg/µL bFGF solution were mixed and vortexed thoroughly. For each pellet, 0.2 µL of the mixture was dropped onto the parafilm and dried at room temperature for 1 to 2 hours to form a pellet.⁴⁰ 

The mice were randomly assigned to either the control group or the experimental group. Six hours before CNV induction, the experimental group received 1 mg/mL NE via subconjunctival injection, and the control group was injected with an equal volume of sterile water. Corneal micropockets were created with a Von Graefe cataract knife on the cornea of the right eye, and the bFGF pellet was implanted into the corneal micropocket. Then, 1 mg/mL NE or infection water was applied as eye drops six times a day beginning on the second day after pocket formation. Eye phenotypes were recorded with a slit lamp on day 1, 3, 5, and 7 after the operation. 

ICI Treatment

Eight wild-type mice were used to establish the bFGF micropocket model and were randomly divided into an experimental group and a control group. The experimental group was given 2 mg/mL ICI-118,551 (ICI; MedChemExpress) as eye drops six times a day beginning with the second day after pocket formation, and the control group was given infection water. The eye phenotypes were recorded using a slit lamp on days 1, 3, 5, and 7 after the operation. 

Adrb2 Knockdown Mouse Model

Adrb2 knockout mice (Adrb2^+/−) was generated on a C57BL/6J background. Sequencing PCR primers for screening Adrb2 gene knockout areas were JS00580-Adrb2-5wt-tF2 (5′-ACTGCTCCAAGAAGCAGACTCTG-3′) and JS00580-Adrb2-5wt-tR2 (5′-AAGACGTGAGAGCACAGAAGAGC-3′). Sequencing PCR primers for detecting the wild-type Adrb2 were JS10580-Adrb2-wt-tF2 (5′-ATCACAGCCATTGCCAAGTTC-3′) and JS10580-Adrb2-wt-tR2 (5′-AAGTCTCCTCGGTGTAACAATCG-3′). Quantitative PCR primers for Adrb2 quantification were T007691-Adrb2-RT-qtF1 (5′-ACGGCTACTCTAGCAATAGCAACG-3′) and T007691-Adrb2-RT-qtR1 (5′-GCTAAGGCTAGGCACAGTACCTTG-3′). Four Adrb2^+/− mice were used to establish the bFGF micropocket model. Six hours before CNV induction, sterile water was applied as a subconjunctival injection. Sterile water was administered as eye drops six times a day in the following days. Eye phenotypes were recorded using a slit lamp on days 1, 3, 5, and 7 after the operation. 

Paraffin Sectioning and Hematoxylin and Eosin Staining

The corneas of mice were fixed in 4% paraformaldehyde (PFA; Solarbio Life Science, Beijing, China) overnight, embedded in paraffin, and sectioned into 5-µm-thick slices. The slides were deparaffinized in xylene, rehydrated, and stained with hematoxylin and eosin. 

Immunofluorescence Staining

Wild-type mouse corneas were collected and embedded in Tissue-Tek O.C.T. Compound (Sakura Finetek, Tokyo, Japan). Corneas were sectioned into 7-µm-thick slices and fixed. The sections were stained with an anti-Adrb2 antibody (1:100; Abcam, Cambridge, UK) overnight at 4°C and then with Alexa Fluor 594–conjugated Donkey anti-Rabbit IgG (H+L) (1:100; Abcam) as the secondary antibody. Negative control sections were only stained with the secondary antibody. The sections were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; Solarbio Life Science) and observed under a confocal microscope (ZEISS, Oberkochen, Germany). 

For whole-mount staining, mouse CNV corneas were collected 7 days after the operation and fixed with 4% PFA for 30 minutes at 4°C. Corneas were then blocked with blocking buffer (PBS + 0.3% Triton X-100 + 5% donkey serum) at 4°C overnight. Anti-CD31 antibody (1:100; R&D Systems, Minneapolis, MN, USA), anti-tyrosine hydroxylase (TH) antibody (1:200; MilliporeSigma, Burlington, MA, USA), and anti-tubulin β3 antibody (1:400; BioLegend, San Diego, CA, USA) were administered to these corneas for 1 day at room and then overnight at 4°C. Corneas were then incubated with fluorescein-conjugated secondary antibodies, either Alexa Fluor 594–conjugated Donkey anti-Goat IgG (1:100; Abcam) or Alexa Fluor 488–conjugated Donkey anti-Rabbit IgG (1:100; Abcam), at room temperature for 2 hours. The resulting samples were flattened on a slide and imaged with a ZEISS laser-scanning confocal microscope. 

Cell Culture

To culture the human umbilical vein endothelial cells (HUVECs), cells were obtained from the Cell Line Library of the Chinese Academy of Sciences (Shanghai, China) and grown in endothelial cell basic medium-2 (EBM-2; Lonza, Basel, Switzerland) supplemented with 10% fetal bovine serum (FBS; Gibco, Waltham, MA, USA), 100 U/mL penicillin, and 100 mg/mL streptomycin (Gibco) at 37°C. 

CCK-8 Assay

HUVECs were cultured in a 96-well cell culture plate at a density of 10⁴ cells per well. Twenty-four hours after plating, 10 µL of different concentrations of ICI (10 µM, 30 µM, 50 µM, 100 µM, and 500 µM) or PBS as a control were added to the culture. Cell Counting Kit 8 (CCK-8) solution (Beyotime Biotechnology, Shanghai, China) was added to the culture after ICI incubation for 24 hours. The optical density value was measured at a wavelength of 450 nm using a microplate reader (Rayto Life and Analytical Sciences, Shenzhen, China). 

Tube Formation Assay

HUVECs were starved in EBM-2 supplemented with 0.1% FBS for 12 hours. Corning Matrigel Growth Factor Reduced Basement Membrane Matrix (50 µL/well; Gibco) was coated onto a 96-well plate and cultured in a 37°C incubator for 30 minutes. HUVECs at a density of 2 × 10⁴ cells per well were seeded on the Matrigel. NE (50 µM), NE combined with bevacizumab (100 ng/mL; Roche, Basel, Switzerland), or NE combined with ICI (60 µM) was added to the cells. The number of branch points was counted and they were photographed using a microscope at a magnification of 20× after 6 and 12 hours of incubation. 

Quantification of Corneal CNV

CD31-positive staining in the whole-mount stained corneas was analyzed using ImageJ (National Institutes of Health, Bethesda, MD, USA) with the VesselJ plug-in. Each whole-mount stained photograph was imported into ImageJ, and the “Custom border” function was used to encircle the cornea completely along the limbus. The cornea was manually marked as a region of interest, and “Threshold” was adjusted to select only the neovascularization area. The percentage of the neovascularization area was obtained by analyzing “vessels” in the VesselJ plug-in function. 

Statistical Analysis

ImageJ was used to analyze the results of the corneal whole-mount staining and tube formation assay. Prism 8 (GraphPad, San Diego, CA, USA) and SPSS Statistics 20.0 (IBM, Chicago, IL, USA) were used for statistical analysis. One-way ANOVA was used to compare differences among groups when the data met normal distribution, and a non-parametric test was applied otherwise. Results were reported as mean ± standard deviation (SD), and P < 0.05 was considered statistically significant (^*P < 0.05, ^**P < 0.01, ^***P < 0.001). 

Results

Concomitant Growth of Sympathetic Nerve and CNV in the Cornea of the Suture Model

To investigate the relationship between corneal nerves and CNV, we analyzed the corneas from CNV mice generated by three frequently used methods: suture-induced CNV, alkali-induced CNV, and bFGF-induced CNV. We first evaluated the CNV characteristics of each model and found they displayed similar ocular phenotypes. One day after the construction of the model, vascular congestion was evident in the limbus region (Supplementary Fig. S1A). Three days after surgery, new blood vessels emerged through the limbus barrier and entered the peripheral cornea (Supplementary Fig. S1A). Vessels became more pronounced by day 5 (Supplementary Fig. S1A). Corneal edema and decreased transparency were also observed (Supplementary Fig. S1A). The suture model exhibited the slowest growth of blood vessels, with lower vessel length and density. The area covered by CNV was primarily around each surgical knot (Supplementary Figs. S1A, S2A). In contrast, the alkali model showed CNV invading the cornea from all directions of the limbus, reaching the pupillary area at 7 days after alkali treatment. Corneal thinning, edema, perforation, and pathological cataract were also observed in the alkali-treated mice (Supplementary Fig. S1A). It is the most severe CNV model with dramatic inflammatory responses (Supplementary Fig. S1B). The bFGF micropocket model is an experimental CNV model with limited corneal inflammation (Supplementary Fig. S1B). As the CNV induction factor bFGF was quantified, all mice in this group displayed uniform corneal phenotypes (Supplementary Fig. S1A, S2A), making it an ideal CNV model for drug efficacy testing. The CNV mainly invaded the micropocket point, accounting for about one-third of the total corneal area (Supplementary Fig. S2B). In conclusion, the suture model and the alkali model are more relevant to clinical CNV, whereas the bFGF model is more suitable for reducing experimental variance. 

We selected the suture model for sympathetic nerve staining as this model has limited inflammatory response (Supplementary Fig. S1B) and is closely related to clinical CNV. In this model, we observed the simultaneous growth of sympathetic nerves and CNV in the cornea. Sympathetic nerves were found to invade the limbus region and extend toward the central cornea along the newly formed vessels (Fig. 1A). To confirm the colocalization of these sympathetic nerves with CNV, we conducted three-dimensional (3D) confocal microscopy scans of the cornea, which revealed that the sympathetic nerves were located in the same scanning layer as the new vessels (Fig. 1B). The entanglement of the nerves and vessels in the cornea suggests a potential regulatory relationship between them. 

Figure 1.

 

The invasion of sympathetic nerves in the mouse suture cornea. (A) Immunostaining of the day 3 suture model cornea with vessel marker CD31 and sympathetic nerve marker TH. Intertwined vessels and sympathetic nerves were observed within the corneal rim. Upper panel, 20× magnification; lower panel, 40× magnification. (B) A 3D confocal scanning of the stained cornea. CD31-positive vessels were in the same scanning layer with the TH-positive sympathetic nerves (n = 3).

NE Promotes the Growth of CNV

To investigate the potential contribution of sympathetic nerves to the growth of CNV, we applied NE to the three CNV models. For all three models, significant promotion of CNV growth was observed after NE treatment, as evidenced by slit-lamp microscopy (Figs. 2A, 3A, 4A) and with immunostaining of vessel marker CD31 (Figs. 2B, 3B, 4B). The length and density of CNV were dramatically increased after NE treatment (Figs. 2C, 3C, 4C). The area of the cornea covered by CNV increased from 16.2% to 30.9% in the suture model (Fig. 2C), from 23.9% to 38.9% in the alkali burn model (Fig. 3C), and from 26.3% to 37.3% in the bFGF group (Fig. 4C). These results suggest that sympathetic neurotransmitter NE promotes CNV growth. 

Figure 2.

 

Impact of NE on CNV in the suture model. (A) Slit-lamp examinations were conducted on days 1, 3, 5, and 7 after the model was established (n = 4). (B) Corneal whole-mount staining with the vessel marker CD31 on day 7. The yellow dotted line represents the boundary between the limbus and the cornea (n = 4). (C) The proportion of cornea areas covered by CNV on day 7 (n = 4; ^***P < 0.001).

Figure 3.

 

Effect of NE on CNV in the alkali burn model. (A) Slit-lamp examinations were conducted on days 1, 3, 5, and 7 after CNV induction (n = 4). (B) Corneal whole-mount staining with the vessel marker CD31 on day 7. The yellow dotted line represents the boundary between the limbus and the cornea (n = 4). (C) The proportion of cornea areas covered by CNV on day 7 (n = 4; ^***P < 0.001).

Figure 4.

 

Effect of NE on CNV in the bFGF micropocket model. (A) Slit-lamp examinations were conducted on days 1, 3, 5, and 7 after CNV induction (n = 4). (B) Corneal whole-mount staining with the vessel marker CD31 on day 7. The yellow dotted line represents the limbus and cornea boundary, and the white dotted circle indicates the location of the micropocket (n = 4). (C) The percentage of cornea area occupied by CNV on day 7 (n = 4; ^***P < 0.001).

NE Promotes CNV Via the β2-AR Receptor

As NE exerts its function through binding to the corresponding receptors, to explore the action of NE on the cornea we investigated the distribution of its corresponding receptor, β2-AR. Immunostaining revealed that β2-AR was present in both the central cornea (Supplementary Fig. S3A) and the peripheral cornea (Supplementary Fig. S3B), with prominent expression in the epithelial cell layer. This suggests that NE can exert its effects by binding to β2-AR in the cornea. 

To further confirm the involvement of β2-AR, we conducted an in vitro tube formation assay using HUVEC cells, which are commonly used in studies on vessel formation and are known to express β2-AR as confirmed by immunostaining (Fig. 5A). β2-AR agonist NE and antagonist ICI were applied to HUVECs. The cytotoxicity of ICI was first tested using a CCK-8 assay. HUVECs showed good proliferation even when treated with up to 60 µM ICI (Supplementary Fig. S4), so we selected this concentration for subsequent experiments. HUVECs were treated with NE, NE supplemented with the classical anti-CNV drug bevacizumab, and NE supplemented with 60-µM ICI. As expected, ICI was able to effectively inhibit tube formation even in the presence of NE (Fig. 5B), and the number of HUVEC branch points was significantly lower in the ICI group than in the other two groups (Fig. 5C). Interestingly, bevacizumab did not show a significant effect in suppressing tube formation when NE was present (Figs. 5B, 5C), suggesting that it may not be an ideal treatment for corneal angiogenesis. 

Figure 5.

 

ICI inhibited angiogenesis in HUVECs. (A) Immunostaining of β2-AR (green) and DAPI (blue) on HUVECs. Upper panel, staining with primary antibody; lower panel, negative control without primary antibody. (B) Representative images of HUVEC tube formation. (C) Number of HUVEC branch points (red circles in B) in the three groups at 6 hours and 12 hours after drug treatment (n = 4; ^**P < 0.01).

In addition to the corneal epithelium, β2-AR was found to be highly expressed in CD31-positive vessels in the cornea (Fig. 6A), indicating a direct regulatory role of the sympathetic system to CNV. To further verify this, we used Adrb2 knockdown mice (Adrb2^+/−) for in vivo experiments, as β2-AR is encoded by Adrb2. bFGF CNV was induced in both wild-type and Adrb2^+/− mice. Adrb2 knockdown significantly blocked the progression of CNV according to the corneal slit-lamp photography and CD31 staining (Figs. 6B–6D). Furthermore, ICI treatment in the bFGF model constructed with wild-type mice effectively inhibited CNV formation compared to the control group (Fig. 6B). Vessel density and length were significantly reduced after ICI treatment (Figs. 6C, 6D). Both genetic and pharmaceutical blockage of Adrb2 eliminated CNV. 

Figure 6.

 

Effects of Adrb2 knockdown and ICI on CNV in the bFGF micropocket model. (A) Whole-mount staining with the vessel marker CD31 and the receptor β2-AR in suture model mouse cornea on day 7. (B) Slit-lamp examinations were conducted on days 1, 3, 5, and 7 after the model was established (n = 4). (C) Corneal whole-mount staining with the vessel marker CD31 on day 7. The yellow dotted line indicates the boundary between the limbus and the cornea, and the white dotted circle indicates the position of the micropocket (n = 4). (D) The percentage of cornea areas occupied by CNV on day 7 (n = 4; ^**P < 0.01).

In conclusion, our in vitro and in vivo experiments confirmed the role of the sympathetic nerve–NE–Adrb2 axis in regulating corneal angiogenesis. This regulatory axis presents a potential target for treating CNV. 

Discussion

CNV is a major risk factor for both corneal transplantation rejection and progressive vision loss. Unfortunately, drugs that show efficacy in treating retinal neovascularization may not be as effective in treating CNV, leading to an emergency in exploring alternative treatments for CNV. The cornea is known to have the most abundant sensory nerve endings in the body, and damage to these nerves often leads to CNV. However, the underlying mechanism behind this is not yet clear. A previous study has reported inhibitory interactions between sensory nerves and sympathetic nerves.²⁹ Sympathetic nerves begin to invade the cornea after the depletion of sensory nerves, at around the same time as CNV onset. We observed the intriguing phenomenon that sympathetic nerves were colocalized with the newly formed vessels, which suggests a potential role of sympathetic nerves in regulating CNV. Similar observations have been reported where sympathetic nerves grew near the pericytes of choroidal microvessels in rats⁴¹ and endothelial cells of brain capillaries in rats.⁴² In our study, we not only confirmed the pro-angiogenesis effect of the sympathetic neurotransmitter NE but also identified its downstream receptor in the cornea, which is β2-AR. NE promotes the proliferation of uterine artery endothelial cells through β2-AR and β3-AR.⁴³ Our results indicate that NE promoted angiogenesis in the cornea through β2-AR (Adrb2). In mice with hypoxia-induced retinopathy, β2-AR blockers were able to reduce VEGF expression and pathological retinal neovascularization.^44,45 

In this study, we utilized three different CNV models that are commonly used in corneal angiogenesis research; however, limited knowledge existed regarding the similarities and differences between these models. To examine the CNV progression in these models, we monitored their status after day 3, when CNV invasion was evident in all three models. It is not clear whether the vessel progression was the same in these three models before day 3; more confounding factors might be introduced for comparison if using models before day 3. We will continue to evaluate the progression of neurons and CNV before the vessels break the limbus. 

In this study, it was demonstrated that NE could promote angiogenesis in all three CNV models, regardless of the corneal inflammatory status or damage severity, indicating the universal regulatory role of the sympathetic system in CNV. Despite these findings, there were several limitations in the study, such as the use of only heterozygous Adrb2 knockdown mice for the inhibitory study. Further experiments involving homozygous Adrb2 knockout mice would be necessary to determine whether NE–Adrb2 is the sole factor or one of the major factors involved in the regulation of CNV. The results of the study suggest that the sympathetic circuit could be an important factor contributing to various types of CNV, thereby serving as a potential target for anti-CNV drugs. 

Acknowledgments

The authors thank Qi Xia for her advice on animal model construction and histopathological preparations. 

Supported by grants from the Shandong Provincial Natural Science Foundation (ZR2020QH140 to BNZ), National Natural Science Foundation of China (82101091 to BNZ; 82271059 to SL), and Youth Specialist Program of Taishan Scholars (tsqn202211343 to SL). 

Disclosure: Q. Dong, None; B. Qi, None; B. Zhang, None; X. Zhuang, None; S. Chen, None; Q. Zhou, None; B.N. Zhang, None; S. Li, None