Suppression of Pathological Ocular Neovascularization by a Small Molecular Multi-Targeting Kinase Inhibitor, DCZ19903

Abstract

Purpose: The administration of anti-vascular endothelial growth factor agents is the standard firs-line therapy for ocular vascular diseases, but some patients still have poor outcomes and drug resistance. This study investigated the role of DCZ19903, a small molecule multitarget kinase inhibitor, in ocular angiogenesis.

Methods: The toxicity of DCZ19903 was evaluated by 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide assays, flow cytometry, Calcein-AM/PI staining, and terminal uridine nick-end labeling staining. Oxygen-induced retinopathy and laser-induced choroidal neovascularization models were adopted to assess the antiangiogenic effects of DCZ19903 by Isolectin B4 (GS-IB4) and hematoxylin-eosin staining. EdU assays, transwell migration assays, tube formation, and choroid sprouting assays were performed to determine the antiangiogenic effects of DCZ19903. The antiangiogenic mechanism of DCZ19903 was determined using network pharmacology approach and western blots.

Results: There was no obvious cytotoxicity or tissue toxicity after DCZ19903 treatment. DCZ19903 exerted the antiangiogenic effects in OIR model and choroidal neovascularization model. DCZ19903 inhibited the proliferation, tube formation, migration ability of endothelial cells, and choroidal explant sprouting. DCZ19903 plus ranibizumab achieved greater antiangiogenetic effects than DCZ19903 or ranibizumab alone. DCZ19903 exerted its antiangiogenic effects via affecting the activation of ERK1/2 and p38 signaling.

Conclusions: DCZ19903 is a promising drug for antiangiogenic treatment in ocular vascular diseases.

Translational Relevance: These findings suggest that DCZ19903 possesses great antiangiogenic potential for treating ocular vascular diseases.

Introduction

Pathological angiogenesis is an important characteristic in a variety of ocular diseases, which can cause irreversible vision impairments.^1–4 It occurs in several ocular tissues, such as retina, iris, cornea, and choroid.⁵ Unlike the mature vessels, these new vessels have incomplete structures and hemorrhagic tendency.⁶ Hence, antiangiogenic treatment is a key strategy for ocular vascular diseases. 

Pathological ocular angiogenesis is usually caused by the imbalance between angiogenic factors and antiangiogenic factors.⁷ Previous studies have shown that vascular endothelial growth factor (VEGF) is a key driver of ocular angiogenesis.⁸ Neovascularization is activated by the binding between VEGF and VEGF receptors, which can contribute the proliferation, migration, and tube formation of ECs.^9,10 In addition to laser photocoagulation or vitrectomy surgery, anti-VEGF therapy has been verified as a major strategy for antiangiogenic treatment.^11,12 Currently, the US Food and Drug Administration has approved several anti-VEGF agents for intraocular injection, such as ranibizumab (Lucentis), aflibercept (Eylea), and pegaptanib (Macugen). They are VEGF antibodies or VEGF receptor fragments that can bind to VEGF molecules to suppress ocular neovascularization. However, frequent injections are still required to sustain the antiangiogenic efficacy. In addition, some patients have been reported to have no response to anti-VEGF treatment.^13,14 Thus, further studies are required to design novel drugs for antiangiogenic treatment. 

Angiogenesis is a complicated pathological process, requiring the participation of membrane receptors, adaptor proteins, and protein kinases.^15–19 To achieve an effective and durable treatment efficiency, it is vital to target several signaling targets simultaneously. The multitarget kinase inhibitors refer to a series of compounds that can suppress tumorigenesis and angiogenesis by regulating endothelial cell proliferation, motility, apoptosis, and differentiation. For instance, sorafenib plays an antitumor and antiangiogenic role via inhibiting receptor tyrosine kinases and serine and threonine kinases.^20–23 Even though great progress has been made, some issues remain to be resolved. For example, these multitarget kinase inhibitor drugs usually have low absorption efficiency owing to poor water solubility, rapid clearance, and a short half-life.²⁴ Hence, further studies are urgently needed to design novel drugs to enhance the treatment efficiency for current multitarget kinase inhibitor drugs. 

In this study, we designed and synthesized a novel small molecular agent, DCZ19903. The results showed that there was no obvious cytotoxicity and ocular toxicity after DCZ19903 treatment at the tested concentrations during the experimental period. DCZ19903 exerted its antiangiogenic effects by altering the activation of p38-MAPK and ERK1/2-MAPK signaling. Thus, DCZ19903 is a promising drug for antiangiogenic treatment in ocular vascular diseases. 

Methods

Synthesis of Tert-Butyl (3-Fluoro-4-Iodophenoxy) Dimethylsilane (Compound 2)

We dissolved 3-fluoro-4-iodophenol (compound 1) (10.0 g, 42 mmol), tert-Butyldimethylchlorosilane (7.6 g, 50 mmol), and triethylamine (5.0 g, 50 mmol) in dichloromethane followed by the agitation at room temperature for 12 hours. The mixture solution was extracted three times with dichloromethane. The extract was dried by a rotary evaporator. Silica gel column chromatography was conducted to obtain an oil 2 (9.2 g, yield 62%). 1H NMR (400 MHz, Chloroform-d) δ 7.56 (dd, J = 8.6, 7.7 Hz, 1H), 6.62 (dd, J = 9.5, 2.6 Hz, 1H), 6.48 (dd, J = 8.6, 2.6, 0.7 Hz, 1H), 0.99 (s, 9H), 0.23 (s, 6H). LRMS (EI) [M]+ found m/z 352. 

Synthesis of Ethyl2,2-Difluoro-2-(2-Fluoro-4-Hydroxyphenyl) Acetate (Compound 3)

Compound 2 (4.7 g, 10 mmol) and ethyl 2-bromo-2,2-difluoroacetate (4.1 g, 20 mmol) were dissolved in dry dimethyl sulfoxide. Then, the activated copper powder (1.6 g, 26 mmol) was added under inert gas environment. After stirring for 12 hours at 60°C, the reaction solution was poured into 1 M dilute hydrochloric acid to quench and extracted twice with ethyl acetate. After being evaporated by rotary evaporation, the combined organics were separated to obtain an oil 3 (1.0 g, yield 44%). 1 H-NMR (400 MHz, CDCl3) δ 7.46 (t, J = 8.5 Hz, 1H), 6.70 (dd, J = 8.6, 2.2 Hz, 1H), 6.62 (dd, J = 12.1, 2.1 Hz, 1H), 5.25 (s, 1H), 4.38 (q, J = 7.1 Hz, 2H), 1.34 (t, J = 7.1 Hz, 3H). HRMS (ESI) [M-H]- found m/z 233.0427 calcd for C10H8F3O3 233.0431. 

Synthesis of 2-(4-((6,7-Bis(2-Methoxyethoxy) Quinazolin-4-Yl) Oxy)-2-Fluorophenyl)-N-(4-Chloro-3-(Trifluoromethyl) Phenyl)-2,2-Difluoroacetamide (Compound 8, DCZ19903)

Ethyl 2,2-difluoro-2-(2-fluoro-4-hydroxyphenyl) acetate (compound 3), 4-chloro-6,7-bis(2-methoxyethoxy) quinazoline (compound 4), triethylenediamine, and triethylamine were dissolved in acetonitrile (5 mL) and reacted at 78°C for 2 hours. The reaction mixture was evaporated and separated by silica gel column chromatography to obtain compound 5. Compound 5 was dissolved in a mixed solution of tetrahydrofuran, methanol, and water (3:1:1). Then, 1 M sodium hydroxide solution was added and the mixture was stirred for 2 hours. The reaction solution was neutralized with 1 M hydrochloric acid solution to pH 1 and extracted with ethyl acetate. After the required treatment, the compound 6 was obtained. 

Compound 6, 4-chloro-3-(trifluoromethyl) aniline (compound 7), N,N-diisopropylethylamine, and 2-(7-azobenzotriazole)-N,N,Nʹ,Nʹ-tetramethylurea hexafluorophosphate were dissolved in acetonitrile and stirred for 12 hours. The solution was then quenched with water before being extracted with ethylacetate. Following the above-mentioned treatment, the residue was separated by silica gel column chromatography to obtain compound 8 (DCZ19903). 

Animals

We obtained C57BL/6J mice from the Model Animal Research Center of Nanjing University (Nanjing, China) and raised in the pathogen-free environment with specific temperature and humidity. The mice were raised on a 12-hour light:12-hour dark schedule. The food and water was provided available ad libitum. Anesthesia was administered by intraperitoneal injection of ketamine (80 mg/kg) and xylazine (10 mg/kg). All animal experiments were authorized by the Institutional Animal Care and Use Committee of Nanjing Medical University and conducted according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Cell Culture

Human umbilical vein endothelial cells (HUVECs) were purchased from American Type Culture Collection (ATCC, Manassas, VA) and cultured in Dulbecco's modified Eagle's medium (DMEM, Gibco, C11995, Grand Island, NY) comprising 10% inactivated fetal bovine serum and 1% penicillin–streptomycin under a normal condition (37°C, in a mixture of 5% CO₂ and 95% air). 

Cell Viability Assay

Cell viability was measured by 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) colorimetric assay. Briefly, HUVECs were seeded in 96-well plates and grown for 24 hours in a standard incubator. After the appropriate treatment, they were incubated with MTT solution (5 mg/mL, BioFroxx, 1334GR001, Einhausen, Germany) for 3 hours. Then, the supernatant was discarded and formazan crystals were dissolved using dimethyl sulfoxide (DMSO). The optical density value was measured using a microplate reader (Molecular Devices, FilterMax F5, San Jose, CA). 

Flow Cytometry

An Annexin V-FITC/PI Apoptosis Detection Kit I was used to evaluate cell apoptosis (Vazyme, A211-01, Beijing, China). After the treatment with DCZ19903 (1 nM to 100 µM), HUVECs were digested with trypsin without ethylene diamine tetra-acetic acid for 20 minutes. After being washed with phosphate-buffered saline (PBS), they were incubated with 100 µL of binding buffer containing Annexin-V-FITC (5 µL) and propidium iodide (PI, 5 µL) for 10 minutes. Finally, cell apoptosis was detected with the flow cytometer (Beckman Coulter, CytoFLEX, Brea, CA) and analyzed by CytoExpert 2.3 (Beckman Coulter). 

Calcein-AM/PI Staining

Calcein-AM/PI staining assays were carried out to measure cell apoptosis. After incubation with DCZ19903 (1 nM to 100 µM), HUVECs were immediately stained with Calcein-AM (2 µM) and PI (3 µM) for 30 minutes. The photographs were captured by a fluorescence microscope (Olympus, IX73, Japan). 

Transwell Migration Assay

Transwell migration assays were used to detect cell migration. Following the required treatment, HUVECs were digested and re-suspended in the serum-free DMEM. Then, the upper chamber of 8-µm transwell (Millipore, MCEP24H48, Burlington, MA) was filled with cell suspension. The medium containing 10% fetal bovine serum was added to the lower chamber. After 12 hours in culture, the migrated cells were fixed in 100% methanol for 20 minutes, washed with PBS, and stained with 0.5% crystal violet. The photographs were captured by an Olympus IX73 microscope and the migrated cells were calculated. 

EdU Assay

Cell proliferation was determined by EdU assays using the BeyoClick EdU Cell Proliferation Kit with Alexa Fluor 488 (Beyotime, C0071S, Jiangsu, China). After the required treatment, HUVECs were incubated with EdU (10 mΜ) reagents for 4 hours. Then, they were fixed and rinsed with 3% bovine serum albumin (BSA) in PBS three times. Subsequently, they were permeabilized with 0.3% Triton X-100 for 15 minutes. After washing with PBS containing 3% BSA, HUVECs were incubated with the reaction solution for 30 minutes. DAPI (Beyotime, C1002) was used to label cell nuclei. The images were captured by a microscope (Olympus, IX73, Tokyo, Japan). 

Tube Formation Assay

The angiogenic ability of HUVECs was measured by tube formation assay. The 24-well plate was frozen and the Matrigel (Corning, 354234, Corning, NY) was added to the bottom of each well. HUVECs were seeded on the 24-well plate with Matrigel addition. After culture for the required time, the images were obtained by a light microscope and the tube length was analyzed by Image J software. 

Laser-Induced Choroidal Neovascularization (CNV) Model

C57BL/6J mice (8-week-old, male) were used for building laser-induced CNV model. Briefly, 0.5% phenylephrine and 0.5% tropicamide were used to dilate the pupils completely. Using a coverslip containing 1 drop of levofloxacin gel, the cornea was flattened and the fundus was visualized. Then, four laser-induced foci were induced by green argon laser pulses (OcuLight GL, Iridex, Mountain View, CA) at the 3, 6, 9, and 12 o'clock positions around the optic disc. The appropriate laser settings were selected: 532 nm wavelength, 50 µm spot size, 120 mW power, and 100 ms duration. The formation of a subretinal bubble was recognized as a successful laser injury. After laser injury, the mice received an intravitreal injection of DCZ19903 (2 µL, 1 µg/µL), ranibizumab (2 µL, 10 mg/mL), or 2 µL of mixture solution of DCZ19903 (1 µg/µL) plus ranibizumab (10 mg/mL). The contralateral eye of mice received an intravitreal injection of 10% DMSO (2 µL) as the control. At day 7 after treatment, the eyeballs were removed. Immunofluorescence staining and hematoxylin-eosin (HE) staining was conducted to detect CNV lesions. 

Oxygen-Induced Retinopathy (OIR) Model

C57BL/6J mouse pups and their nursing mother were placed in 75% oxygen from postnatal 7 day to 12 day (P7–P12). At P12, the mouse pups received an intravitreal injection of DCZ19903 (1 µL, 1 µg/µL), ranibizumab (1 µL, 10 mg/mL), or 1 µL of mixture solution of DCZ19903 (1 µg/µL) plus ranibizumab (10 mg/mL). The control group received an intravitreal injection of 10% DMSO (1 µL). Then, the mouse pups were exposed to room air for 5 additional days. At P17, the mice were anesthetized and the eyeballs were collected. The whole-mount retinas were visualized by Isolectin B4 staining (GS-IB4, Thermo Fisher Scientific, I21413, Waltham, MA). 

Immunofluorescent Staining

The frozen retinal sections were permeabilized and blocked with 5% BSA at 37°C for 1 hour, followed by incubation with intercellular adhesion molecule 1 (ICAM-1) antibody (Santa Cruz Biotechnology, sc-107, Santa Cruz, CA) at 4°C overnight. After washing with PBS containing 0.1% Tween 20, the frozen sections were incubated with the secondary antibody for 2 hours in darkness. DAPI was used to label nuclei. The image were captured by an Olympus IX73 microscope. 

HE Staining

HE staining was used to detect histopathological changes. After routine dewaxing and hydration, the sections were stained with hematoxylin and eosin. The structural changes of retinas and choroids were observed under a microscope (Olympus, IX73). Image J software was used to quantify the area and the maximum thickness of lesions. 

Choroid Sprouting Assay

Choroid sprouting assays were conducted to evaluate the angiogenic potential. The eyes were removed and immersed in cold PBS. The choroid–sclera explants were then cut into 1 mm × 1 mm pieces. The pieces were seeded on the 24-well plates precoated with Matrigel. After the required treatment, the image of each explant was taken under a light microscope on days 4, 5, and 6. 

Terminal Uridine Nick-End Labeling (TUNEL) Staining

TUNEL staining with the One Step TUNEL Apoptosis Assay Kit (Beyotime, C1088) was used to assess cell apoptosis. After the required treatment, the sections were incubated with proteinase K solution (20 µg/mL) for 15 minutes. Then, the sections were washed twice with PBS and incubated with the TUNEL reaction mixture for 1 hour. The sections of positive control were penetrated with recombinant DNase I (Beyotime, C1082) for 10 minutes. Finally, the nuclei were labeled with DAPI for 10 minutes and observed by a microscope (Olympus, IX73). 

Cell Permeability Assay

Evans Blue powder was dissolved in PBS solution to obtain a stock solution, filtered, and stored at 4°C in the dark. Evans Blue stock solution (20 mg/mL) was diluted with BSA solution to the final concentration of 0.67 mg/mL, which was referred to as EBA. HUVECs were incubated with VEGF (10 ng/mL) for 12 hours, then treated with DCZ19903 (50 nM), ranibizumab (100 µg/mL), or DCZ19903 (50 nM) plus ranibizumab (100 µg/mL) for 12 hours. The control group received no treatment. After the required treatment, 100 µL of cell suspension was placed in the upper chamber of a 0.4 µm transwell (Corning, 3413) insert and 600 µL of DMEM containing 10% fetal bovine serum was added to the lower chamber. After 8 hours culture, the liquid in the upper and lower layers of the chamber was removed. The upper chamber was filled with 100 µL of EBA and the lower chamber was filled with 600 µL of 4% BSA solution. After incubating for 1 hour in the dark, the transwell insert was removed. Finally, the absorbance of lower layer solution was measured by a microplate reader (Molecular Devices, FilterMax F5). 

Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR)

Total RNAs were extracted using TRIzol (Invitrogen, 15596018, USA). qPCR assays were performed with the PowerUp SYBR Green Master Mix (Thermo Fisher Scientific, A25742, Waltham, MA) in PikoReal 96 Real-Time PCR System (Thermo Fisher Scientific) according to the manufacturer's instructions. The reaction conditions were conducted as follows: predenaturation at 94°C for 5 minutes, after 40 cycles of 94°C for 30 seconds, 54°C for 30 seconds, 72°C for 1 minute, and 72°C for 3 minutes. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was detected as the internal control. Gene expression was calculated by relative 2^−ΔΔCt method. 

Potential Targets Intersection of DCZ19903

The species was confined to “Homo sapiens” in the Swiss Target Prediction database, and the candidate targets with prediction scores of greater than 0 were selected as the probable targets of DCZ19903. The Disgenet database was used to find disease targets by using the keywords “ocular neovascularization.” Venny 2.1.0 was used to extract the common targets of DCZ19903 and ocular neovascularization. The drug-bioactive ingredient target-disease network of DCZ19903 acting on ocular neovascularization was established using Cytoscape (version 3.2.1). 

Protein–Protein Interaction Network Construction

The targets of DCZ19903 and disease were imported to STRING database to determine the network connection of target interaction, with the minimum required interaction score greater than 0.4. All data were uploaded into Cytoscape for generating PPI network, in which the color and size indicated the degree of nodes. 

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway Enrichment Analysis

GO and KEGG enrichment analyses were performed on the common targets of DCZ19903 and disease. The items with a corrected P value of less than 0.05 were screened by the STRING database to provide the pathway enrichment data. The Cluster Profiler package was installed and the bubble charts were created using the R 3.6.3 software. 

Western Blot

Total proteins were extracted by the radioimmunoprecipitation assay lysis buffer (Beyotime, P0013B) containing the protease inhibitors (Roche, 04693132001, USA). The Bio-Rad protein assay (Bio-Rad, 23227, Basel, Switzerland) was used to measure the concentrations of total proteins. Then, the separated proteins using SDS-PAGE gels were transferred onto PVDF membranes (Millipore, IPVH00010). These membranes were blocked with 5% free-fat milk for 1 hour, followed by incubation overnight at 4°C with the primary antibodies: p38 (Cell Signaling Technology, 9212, Danvers, MA), ERK1/2 (Cell Signaling Technology, 9102), JNK (Cell Signaling Technology, 9252), p-p38 (Cell Signaling Technology, 9215), p-ERK1/2 (Cell Signaling Technology, 4370), p-JNK (Cell Signaling Technology, 9251), ICAM-1 (Abcam, ab171123, Waltham, MA), and GAPDH (Cell Signaling Technology, 2118). After being washed with PBS containing 0.1% Tween 20, the membranes were incubated with the secondary antibody (Beyotime, A0208 and A0216) for 2 hours. Finally, the signaling was detected by the ECL detection kit (Beyotime, P0018S). 

Statistical Analyses

Statistical analysis was performed by GraphPad Prism 9 software and the continuous data were expressed as mean ± SD. The significant differences were assessed by one-way analysis of variance or repeated measures analysis of variance followed by the post hoc Bonferroni test as appropriate. A P value of less than 0.05 was considered statistically significant difference. 

Results

Synthesis of DCZ19903

We designed and synthesized a novel small molecular compound, DCZ19903 (molecular weight of 660.0) using the computer-aided drug design. The synthetic route and molecular structure of DCZ19903 was shown in Figure 1A. We then used the Swiss Target Prediction database (http://www.swisstargetprediction.ch/index.php) to predict the potential targets of DCZ19903 (Table 1). 

Figure 1.

 

Synthesis of DCZ19903.

Table 1.

 

Swiss Target Prediction Analysis of the Targets of DCZ19903 (Top 15)

 

DCZ19903 has no Obvious Cytotoxicity or Tissue Toxicity

To determine the cytotoxicity of DCZ19903, we performed MTT assays to detect cell viability after the administration of DCZ19903. The results indicated that DCZ19903 exhibited no observable cytotoxicity on HUVECs at concentrations ranging from 1 nM to 10 µM (Fig. 2A). Flow cytometry analysis and Calcein-AM/PI staining further confirmed that DCZ19903 had no obvious cytotoxicity at the tested concentrations (1 nM to 10 µM) (Figs. 2B and 2C). To examine the tissue toxicity of DCZ19903, mice received an intravitreal injection of PBS, 10% DMSO, or DCZ19903. HE staining assays were used to observe the changes of retinal structures. The results showed that intravitreal injection of DMSO or DCZ19903 did not induce the obvious histopathological changes in the retinas compared with the control group (Fig. 2D). TUNEL staining was used to quantify retinal apoptosis. The results revealed that DMSO or DCZ19903 treatment did not cause a detectable apoptosis in the retinas, whereas DNase I treatment (positive control group) led to a significant increase in retinal apoptosis (Fig. 2E). These results demonstrated that DCZ19903 had no obvious toxicity at the working concentrations in vitro or in vivo. 

Figure 2.

 

DCZ19903 administration has no obvious cytotoxicity and tissue toxicity. (A–C) HUVECs were treated with DCZ19903 (1 nM to 100 µΜ), or left untreated (Ctrl) for 24 hours. MTT assays were used to evaluate cell viability (A; n = 4). (B) Annexin V-FITC/PI assays were used to quantify the apoptotic percentage of HUVECs (B, n = 4). Calcein-AM/PI staining was used to detect cell apoptosis (C) (n = 4, scale bar, 20 µm). ^*P < 0.05 versus Ctrl group. (D and E) C57BL/6J mice received intravitreal injections of PBS (Ctrl), DMSO, or DCZ19903 (1 µg/µL). At day 7 after the injection, the histological changes and cell apoptosis in the retinas were evaluated by HE staining and TUNEL staining assays. In TUNEL staining experiment, DNase I was detected as the positive control (n = 4, scale bar: 50 µm).

DCZ19903 Inhibits Ocular Neovascularization In Vivo

Next, we constructed two animal models of ocular neovascularization to assess the antiangiogenic effects of DCZ19903. In a laser-induced CNV model, HE staining assays were performed to detect CNV lesions by calculating the areas and thickness of CNV lesions. Compared with the laser-injured group, administration of DCZ19903 with or without ranibizumab significantly attenuated the size of CNV lesions. Notably, the DCZ19903 plus ranibizumab group had the best antiangiogenic effects on laser-induced CNV as shown by the smallest area and thickness of CNV lesions (Fig. 3A). In addition, CNV formation was observed by GS-IB4 staining on day 7 after laser injury. The area of CNV lesions was markedly decreased in the DCZ19903-treated group and the ranibizumab-treated group compared with the laser-injured group. The area of CNV lesions was smaller in the DCZ19903 plus ranibizumab group than that in other experimental groups, including the DCZ19903-treated group and the ranibizumab-treated group (Fig. 3B). In the OIR model, compared with the control group, the avascular area was significantly decreased in the DCZ19903-treated group or the ranibizumab-treated group. Moreover, the area of neovascular tufts was significantly decreased in the DCZ19903-treated group or the ranibizumab-treated group. Notably, DCZ19903 plus ranibizumab showed greater antiangiogenic effects than DCZ19903 or ranibizumab alone (Fig. 3C). These results confirm that DCZ19903 exerted an antiangiogenic effect on ocular neovascularization in vivo. 

Figure 3.

 

DCZ19903 administration inhibits ocular angiogenesis in vivo. (A and B) Laser-induced CNV models were used to determine the antiangiogenic effects of DCZ19903. After the laser injury, the mice received intravitreal injections of 10% DMSO (Ctrl), DCZ19903 (1 µg/µL), ranibizumab (10 mg/mL), or DCZ19903 (1 µg/µL) plus ranibizumab (10 mg/mL), respectively. HE staining was used to measure the area of neovascular lesions after 7 days. The thickness was calculated from the bottom of the choroid to the top of the lesion, as indicated by the yellow line. The area of the lesion was measured by Image J (A) (n = 4, scale bar, 50 µm). CNV formation was observed by GS-IB4 staining (B) (n = 4, scale bar, 100 µm). (C) P7 mouse pups were exposed to hyperoxia (75% O₂) for 5 days with their nursing mothers. Subsequently, they were returned to normoxic condition and injected intravitreally with 10% DMSO (Ctrl), DCZ19903 (1 µg/µL), ranibizumab (10 mg/mL), or DCZ19903 (1 µg/µL) plus ranibizumab (10 mg/mL) at P12. To evaluate retinal vasculature, the retinas were extracted at P17 and stained with GS-IB4. Avascular regions were highlighted by white dashed lines. Angiogenic regions were highlighted by yellow markers (n = 4) (scale bar, 200 µm). ^*P < 0.05 versus Ctrl group. ^△P < 0.05 versus DCZ19903 + Ran group.

DCZ19903 Inhibits Endothelial Angiogenic Function In Vitro

Endothelial cells play a crucial role in ocular angiogenesis.²⁵ To determine whether DCZ19903 exerted an antiangiogenic effects in vitro, HUVECs were pretreated with VEGF (10 ng/mL) to induce endothelial angiogenic activities. After treatment with VEGF for 12 hours, HUVECs were incubated with DCZ19903 (10 nM to 1 µM) for another 24 hours. MTT assays showed that VEGF treatment led to enhanced cell viability, which was inhibited by the administration of DCZ19903 ranging from 50 nM to 1 µM (Fig. 4A). EdU assays showed that administration of DCZ19903 could cause a marked reduction of VEGF-induced cell proliferation (Fig. 4B). Transwell migration assays revealed that administration of DCZ19903 significantly decreased the migration ability of HUVECs induced by VEGF (Fig. 4C). Tube formation assays revealed that DCZ19903 obviously suppressed VEGF-mediated tube formation of HUVECs (Fig. 4D). We also conducted choroid sprouting assays to evaluate the antiangiogenic effects of DCZ19903. The result showed that the administration of DCZ19903 significantly decreased choroid sprouting ability at the tested concentrations (Fig. 4E). Moreover, the inhibitory effect of DCZ19903 plus ranibizumab on choroid sprouting was greater than that of monotherapy group as shown by decreased tube formation, decreased proliferation ability, and decreased migration ability. 

Figure 4.

 

DCZ19903 inhibits endothelial angiogenic function in vitro. (A–D) HUVECs were incubated with VEGF (10 ng/mL), VEGF (10 ng/mL) plus DCZ19903 (50 nM), VEGF (10 ng/mL) plus ranibizumab (100 µg/mL), VEGF (10 ng/mL) plus DCZ19903 (50 nM) and ranibizumab (100 µg/mL), or left untreated (Ctrl). Cell viability was detected by MTT assays (A, n = 4). Cell proliferation was measured by EdU assays (B) (n = 4; scale bar, 20 µm). Cell migration was detected by transwell assays (C) (n = 4; scale bar, 20 µm). Tube formation was observed under a light microscope (D) (n = 4; scale bar, 100 µm). (E) The RPE/choroid complexes of C57BL/6J mice were prepared and sliced into 1 mm × 1 mm pieces and then placed in 24-well plates precoated with Matrigel. The sprouting potency of choroidal explants were observed on day 4, day 5, and day 6 after seeding. Quantification of sprouting area and representative images of choroidal sprouting were shown (E) (n = 4; scale bar, 200 µm). ^* P < 0.05 versus Ctrl group; ^#P < 0.05 versus VEGF group; ^△P < 0.05 DCZ19903 or Ran versus DCZ19903 + Ran group.

DCZ19903 Inhibits Vascular Permeability

Previous studies have shown that increased vascular permeability contributes to the development of ocular neovascular diseases.²⁶ ICAM-1 is a member of immunoglobulin gene superfamily of adhesion molecules. VEGF directly activates endothelial signaling, promoting ICAM-1 expression to increase endothelial cell permeability.²⁷ We thus explored the effect of DCZ19903 administration on vascular permeability. qRT-PCR assays revealed that DCZ19903 administration led to a reduced levels of ICAM-1 expression (Fig. 5A). In the monolayer cell permeability assay, the optical density values of BSA-Evans Blue in the lower layer of transwell inserts were used to assess the changes in endothelial monolayer cell permeability. After the required treatment, DCZ19903 administration increased the permeability of HUVECs. The group treated with DCZ19903 plus ranibizumab had a greater inhibitory effect on endothelial permeability, compared with the group treated with DCZ19903 or ranibizumab alone (Fig. 5B). In addition, we investigated the effects of DCZ19903 on vascular permeability in vivo. We constructed an OIR model and performed immunofluorescent staining of ICAM-1. The expression of ICAM-1 in the retina was decreased after the administration of DCZ19903 or ranibizumab. Administration of DCZ19903 plus ranibizumab had a greater inhibitory effect on ICAM-1 expression (Fig. 5C). We also investigated the effect of DCZ19903 administration on vascular permeability in CNV model. Western blots and qRT-PCR assays showed that administration of DCZ19903 led to a decreased expression of ICAM-1 in the CNV model (Figs. 5D and 5E), suggesting that administration of DCZ19903 could decrease the expression of ICAM-1 and inhibit vascular permeability. 

Figure 5.

 

DCZ19903 inhibits vascular permeability. (A and B) HUVECs were incubated with VEGF (10 ng/mL), VEGF plus DCZ19903 (50 nM), VEGF plus ranibizumab (100 µg/mL), VEGF plus DCZ19903 (50 nM) and ranibizumab (100 µg/mL), or left untreated (Ctrl). qRT-PCR assays were performed to assess the effects of DCZ19903 administration on ICAM-1 expression (A, n = 4). Evans Blue-transwell experiments were used to detect the role of DCZ19903 treatment on the permeability of HUVECs induced by VEGF (B) (n = 4). (C) P7 pups were exposed to hyperoxia (75% O₂) for 5 days with their nursing mothers. Subsequently, they were returned to normoxic condition and injected intravitreally with 10% DMSO (Ctrl), DCZ19903 (1 µg/µL), ranibizumab (10 mg/mL), or DCZ19903 (1 µg/µL) plus ranibizumab (10 mg/mL) at P12. WT indicates the wild-type group without treatment. Immunofluorescence analysis of ICAM-1 was performed at P17 to detect the expression of ICAM-1 in OIR retinas. Representative images and quantitative data were presented (n = 4; scale bar, 50 µm; nuclei, blue; ICAM-1–positive cells, green). (D and E) Laser-induced CNV models were used to determine the antiangiogenic effects of DCZ19903. After laser injury, the mice received intravitreal injections of 10% DMSO (Ctrl), DCZ19903 (1 µg/µL), ranibizumab (10 mg/mL), or DCZ19903 (1 µg/µL) plus ranibizumab (10 mg/mL). The wild-type mice received no treatment (WT). qRT-PCRs and western blots revealed that DCZ19903 administration reduced the expression of ICAM-1 in CNV model. GAPDH was detected as the internal control (n = 4). ^*P < 0.05 versus Ctrl group; ^#P < 0.05 versus VEGF group; ^△P < 0.05 versus DCZ19903 + Ran group.

DCZ19903 Exerts Its Antiangiogenic Effect via Regulating MAPK Signaling

There were 2461, 59, and 200 ocular neovascularization-related target genes identified from GeneCards database, NCBI database, and OMIM database, respectively. After merging and removing the target genes identified from these databases, 2563 intersected disease targets were derived. Subsequently, a total of 65 common target genes of DCZ19903 were identified by the Venn diagram, which intersected with ocular neovascularization. Then, a network diagram of drug–component–target–disease interaction was drawn (Fig. 6A). To visualize the relationship between the active components of DCZ19903 and disease targets, we obtained the PPI data using the STRING online database and imported the results from STRING database into Cytoscape to establish the PPI relationship network. In the network, the top 10 key targets were AKT1, mTOR, PI3KCA, MAPK1, MDM2, KDR, MAPK8, HGF, GSK3B, and PRKCB (Fig. 6B). 

Figure 6.

 

DCZ19903 exerts antiangiogenic effects via regulating MAPK signaling. (A) Network diagram of interaction between DCZ19903 and ocular neovascularization. (B) PPI network analysis. (C) GO enrichment analysis. (D) KEGG pathway enrichment analysis. (E) HUVECs were treated with DCZ19903, ranibizumab, or DCZ19903 plus ranibizumab for 24 hours before being stimulated with VEGF (50 ng/mL). The proteins were electrophoresed, transferred to the membranes, and probed with the specified antibodies. (F) HUVECs were pretreated with or without SB203580, U0126, and then cultured with VEGF (50 ng/mL), VEGF plus DCZ19903 (50 nM), or left untreated (Ctrl). GAPDH was detected as the internal control for protein loading (n = 4). ^*P < 0.05 versus VEGF group; ^△P < 0.05 versus DCZ19903 + Ran group.

To explore the biological functions of predicted targets, the 65 target genes were used for GO and KEGG pathway enrichment analysis. GO enrichment analysis showed that a total of 128 biological processes were enriched, such as peptidyl-serine phosphorylation and peptidyl-tyrosine phosphorylation. A total of 56 molecular functions were enriched, such as protein serine/threonine kinase activity, protein tyrosine kinase activity, and transmembrane receptor protein tyrosine kinase activity. A total of 29 cellular components were enriched, such as membrane raft, membrane microdomain, membrane region (Fig. 6C) (P < 0.01). We a performed KEGG pathway enrichment analysis on these 65 common target genes. As shown in Figure 6D, MAPK signaling pathway was ranked as the most enriched pathway involved in antiangiogenic effects of DCZ19903. 

Western blots revealed that administration of DCZ19903 could reverse VEGF-induced increase in ERK1/2 phosphorylation. Likewise, the levels of phosphorylated p38 were also reduced after the administration of DCZ19903 (Fig. 6E). To further confirm whether MAPK signaling pathway contributed to DCZ19903-mediated antiangiogenic effects, HUVECs were treated with SB203580 (p38 inhibitor) or U0126 (ERK1/2 inhibitor) and then exposed to VEGF. DCZ19903 had a similar effect as SB203580 or U0126 on the inhibition of MAPK signaling (Fig. 6F). These results suggest that DCZ19903 exerts its antiangiogenic effects via inhibiting the phosphorylation of ERK1/2 and p38. 

Discussion

Ocular neovascularization is a primary cause of visual loss in many ocular disorders.²⁸ It can occur in several eye regions, such as the retina, choroid, iris, and cornea.⁵ VEGF is regarded as the key regulator during ocular neovascularization.²⁹ As a result, blocking VEGF/VEGF receptor-2 signaling is efficacious in antiangiogenic therapy. Currently, intravitreal injection of anti-VEGF drug is the mainstay strategy for ocular neovascularization.^30,31 However, these anti-VEGF drugs are often expensive and some patients have no response to current drugs.^32,33 In this study, we designed a novel small molecule compound, DCZ19903, which exerted great antiangiogenic effects on ocular neovascularization. DCZ19903 had no obvious cytotoxicity on HUVECs, and had no obvious ocular toxicity at the tested concentrations. 

We further explored whether DCZ19903 could inhibit ocular neovascularization in vitro and in vivo. DCZ19903 played an inhibitory role in VEGF-induced migration, proliferation, viability, and tube formation in endothelial cells. Moreover, the inhibitory effects of DCZ19903 on ocular neovascularization have been verified in OIR model and laser-induced CNV model. In laser-induced CNV model, the area of CNV lesion was decreased in the DCZ19903-treated group and the ranibizumab-treated group compared with the control group. In OIR model, DCZ19903 attenuated avascular area and neovascular area on P17. Moreover, combination therapy with DCZ19903 and ranibizumab showed a greater antiangiogenic effect than single treatment. 

Angiogenesis is a complex pathological process that is regulated by multiple proteins and signaling pathways.³⁴ We explored the potential mechanism of DCZ19903 in ocular neovascularization using network pharmacology. MAPK signaling pathway was predicted as the enriched pathway involved in DCZ19903-mediated antiangiogenic effects. MAPK signaling pathways are composed of ERK1/2, JNK, and p38 signaling pathways. Previous studies have reported that p38 is a proangiogenic factor for endothelial migration. ERK1/2 signaling is tightly associated with the proliferation ability of endothelial cells.³⁵ DCZ19903 could decrease the phosphorylated levels of ERK1/2 and p38 induced by VEGF. Moreover, DCZ19903 could mimic the effects of SB203580 (p38 inhibitor) and U0126 (ERK1/2 inhibitor) on endothelial cell functions. Thus, DCZ19903 could exert antiangiogenic effects via inactivation of p38 MAPK and ERK1/2 MAPK signaling. 

Increased vascular permeability has been reported to contribute to the progression of ocular vascular diseases. ICAM-1 is highly expressed in endothelial cells and immune cells, which can mediate cell–cell adhesion.³⁶ ICAM-1 can mediate vascular permeability by recruiting these inflammatory cells to accelerate the development of the inflammation response.^37,38 Under the noninflammatory condition, ICAM-1 is expressed at a low level in endothelial cells.³⁹ However, ICAM-1 expression is significantly increased in response to the stimulation of inflammatory factors, such as interleukin-1 β and tumor necrosis factor.^36,40 VEGF is also recognized as a vascular permeability factor.⁴¹ VEGF up-regulation also contributes to increased vascular permeability and enhanced expression of ICAM-1.⁴² VEGF treatment led to increased endothelial permeability, which was interrupted by DCZ19903, ranibizumab, and DCZ19903 plus ranibizumab. Moreover, DCZ19903 plus ranibizumab treatment has greater inhibitory effects on endothelial permeability DCZ19903 or ranibizumab alone. This result provides more evidence for the antiangiogenic effects of DCZ19903 in ocular vascular diseases. 

In conclusion, this study demonstrates that administration of DCZ19903 can inhibit ocular angiogenesis by blocking the MAPK signaling pathway and decrease vascular permeability. DCZ19903 plus ranibizumab can achieve better antiangiogenic effects than DCZ19903 or ranibizumab alone. Overall, DCZ19903 is a promising antiangiogenic drug and can be used as an adjunct therapy for ocular vascular diseases. 

Acknowledgments

Supported by the National Natural Science Foundation of China grants 81570859 and 82070983 (to Q.J), 81872797 (to WL.Z.), and 22077131 (to B.L.). The authors declare no conflict of interest. 

Disclosure: J. Ding, None; B. Li, None; H. Zhang, None; Z. Xu, None; Q. Zhang, None; R. Ye, None; S. Feng, None; Q. Jiang, None; W. Zhu, None; B. Yan, None