Acrizanib as a Novel Therapeutic Agent for Fundus Neovascularization via Inhibitory Phosphorylation of VEGFR2

Abstract

Purpose: The present study aimed to evaluate the effect of acrizanib, a small molecule inhibitor targeting vascular endothelial growth factor receptor 2 (VEGFR2), on physiological angiogenesis and pathological neovascularization in the eye and to explore the underlying molecular mechanisms.

Methods: We investigated the potential role of acrizanib in physiological angiogenesis using C57BL/6J newborn mice, and pathological angiogenesis using the mouse oxygen-induced retinopathy (OIR) and laser-induced choroidal neovascularization (CNV) models. Moreover, vascular endothelial growth factor (VEGF)–treated human umbilical vein endothelial cells (HUVECs) were used as an in vitro model for studying the molecular mechanism underlying acrizanib's antiangiogenic effects.

Results: The intravitreal injection of acrizanib did not show a considerable impact on physiological angiogenesis and retinal thickness, indicating a potentially favorable safety profile. In the mouse models of OIR and CNV, acrizanib showed promising results in reducing pathological neovascularization, inflammation, and vascular leakage, indicating its potential efficacy against pathological angiogenesis. Consistent with in vivo results, acrizanib blunted angiogenic events in VEGF-treated HUVECs such as proliferation, migration, and tube formation. Furthermore, acrizanib inhibited the multisite phosphorylation of VEGFR2 to varying degrees and the activation of its downstream signal pathways in VEGF-treated HUVECs.

Conclusions: This study suggested the potential efficacy and safety of acrizanib in suppressing fundus neovascularization. Acrizanib functioned through inhibiting multiple phosphorylation sites of VEGFR2 in endothelial cells to different degrees.

Translational Relevance: These results indicated that acrizanib might hold promise as a potential candidate for the treatment of ocular vascular diseases.

Introduction

Pathological neovascularization occurs in a variety of fundus diseases, causing visual impairments of all ages across the world, including retinopathy of prematurity (ROP), proliferative diabetic retinopathy, and neovascular age-related macular degeneration (nAMD).^1–3 Fundus neovascularization is generally classified as retinal neovascularization (RNV) or choroidal neovascularization (CNV) according to the type of vascularized tissue. RNV is the pathological neovascularization that arises in the retinal capillary bed and can penetrate the internal limiting membrane (ILM) into the vitreous.⁴ CNV is the growth of abnormal choroidal blood vessels that breach Bruch's membrane into the subretinal space.⁵ Both angiogenic patterns can lead to complications such as leakage, vitreous hemorrhage, and retinal detachment, which are the leading causes of severe vision loss and irreversible blindness.⁶ Therefore antiangiogenic therapy is essential to treat ocular vascular diseases. 

Vascular endothelial growth factor (VEGF) is one of the most important regulators during physiological and pathologic angiogenesis.⁷ The VEGF protein family includes VEGF-A, -B, -C, and -D and placental growth factor.⁸ Among them, VEGF-A plays a critical role in regulating intraocular angiogenesis and vascular permeability,⁹ and it is currently the main therapeutic target of antiangiogenic drugs such as ranibizumab and aflibercept.¹⁰ However, its clinical efficacy has been limited by insensitivity and resistance to the drug.¹¹ Many patients require frequent injections to maintain a therapeutic effect, increasing the financial burden on patients and the risk of complications associated with intravitreal injections.^12,13 Furthermore, anti-VEGF treatment impaired physiological angiogenesis, whereas physiological angiogenesis could normalize blood vessels to repair tissue injury caused by pathological angiogenesis in disease progression.¹⁴ Inhibition of physiological angiogenesis is not only detrimental to therapeutic safety, but it also aggravates the disease severity, especially for neonates with ROP.¹⁵ With these issues in mind, there is an urgent need to identify novel effective therapies that can control the pathological ocular angiogenesis without hampering physiological blood vessel growth. 

VEGF receptors (VEGFR) belong to the receptor tyrosine kinase superfamily and comprise three subtypes: VEGFR1, 2, and 3. Among these receptors, VEGFR2, which is primarily expressed by vascular endothelial cells (ECs), is the main receptor involved in angiogenesis.¹⁶ It is worth noting that VEGFR2 can be activated by many factors and stimuli.¹⁷ Besides classical VEGF ligands (VEGF-A, -C, -D), non-VEGF ligands and mechanical forces can also induce VEGFR2 phosphorylation.¹⁸ Diverse phosphorylation sites recruit adaptor proteins to activate different downstream signaling pathways, thereby regulating vascular endothelial functions.¹⁷ In recent years, several studies have found that selective inactivation of VEGFR2 phosphorylation sites could achieve therapeutic objectives while leaving other VEGFR2-dependent functions intact.^19,20 Smith et al.¹⁹ elucidated that retinal vascular leakage was ameliorated in mice with induced retinopathy via inhibitory phosphorylation of VEGFR2-tyrosine949 (Tyr949, Tyr951 in human). Moreover, Kim et al.²⁰ found that inhibiting VEGFR2-Tyr1173 (Tyr1175 in human) phosphorylation could suppress pathological angiogenesis and vascular leakage in an ischemia/reperfusion mouse model. Overall, we believe that targeting VEGFR2 is a promising strategy for the treatment of ocular vascular diseases. 

Acrizanib, a small molecule tyrosine kinase inhibitor (TKI), specifically binds to the intracellular domain of VEGFR2, inhibiting its phosphorylation and blocking downstream signaling pathways.²¹ Initially, acrizanib was developed as a topically-administrated drug to treat nAMD. In preclinical studies, acrizanib exhibited the ability to strongly inhibit pathological neovascularization and had a favorable tolerability, safety, and metabolic profile.²¹ However, because of limitations of eye drop solutions, the acrizanib concentration was insufficient in fundus tissues, leading to the failure of clinical trials.²² Although the development of eye drops failed, acrizanib still has the potential to be used therapeutically against ocular vascular diseases. To date, no study has systematically assessed the effects of acrizanib on physiologic and pathologic fundus angiogenesis. Additionally, acrizanib's underlying cellular mechanisms and signaling pathways remain to be fully elucidated. 

This study aimed to investigate the effects of acrizanib on physiological and pathological angiogenesis in neonatal mice, the oxygen-induced retinopathy (OIR) model and the laser-induced CNV model and explore how it regulates downstream signaling pathways in both animal models and VEGF-treated human umbilical vein endothelial cells (HUVECs). 

Methods

Acrizanib Administration

Acrizanib powder was purchased from MedChemExpress (Monmouth Junction, NJ, USA) and dissolved in dimethyl sulfoxide (DMSO; Sigma-Aldrich Corp., St. Louis, MO, USA) as stock solutions. For in vitro experiments, the stock solutions were diluted with culture medium to 50 nM (DMSO < 0.1%). This concentration was determined in a preliminary experiment and did not exert toxic effects on HUVECs. For in vivo experiments, working solutions were prepared in PBS at a concentration of 10 µM, which was chosen on the results of our preliminary experiments (Supplementary Fig. S4). 

In the mouse OIR model, after transferred to room air (RA, 21% O₂) at postnatal day 12 (P12), mice were anesthetized using 1.5% isoflurane (Ruiwode Lifescience, Shenzhen, China), and the pupils were dilated with compound tropicamide eye drops (Xingqi Pharmaceutical Co. Ltd, Shenyang, China). Acrizanib solution (0.5 µL) was administered using a sterilized 33G needle attached to a microsyringe (Hamilton, Reno, NV). The needle was inserted through the sclera into the vitreous cavity of the eye. To avoid potential interference caused by the injection procedure and the low concentration of DMSO, mice in the control, OIR and CNV groups, which were not injected with acrizanib solution, were administered equi-volumes of DMSO solution as the vehicle. The same method was applied for testing the effects of acrizanib on physiological angiogenesis in P3 mice. In the mouse model of laser-induced CNV, 1 µL of acrizanib solution or DMSO solution was injected into the vitreous cavity using the same method described above after laser induction (Supplementary Fig. S1). 

Animals

The experiments were performed with male C57BL/6J mice (six to eight weeks of age) and pregnant female C57BL/6J mice, which were obtained from the Animal Laboratory of Zhongshan Ophthalmic Center (Guangzhou, China). All animal management procedures complied with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Animal Ethics Committee of Zhongshan Ophthalmic Center. In this study, the bilateral eyes of each mouse were used. For each experiment, six eyeballs were used from six distinct mice in each group. In the Western blot experiment, considering the relatively limited abundance of retinal and choroidal proteins in mice, tissue samples from two mice were combined into one sample. A total of three distinct samples were used for each group. 

Mouse OIR Model

The mouse OIR model was established as previously described.²³ In short, mother mice and their neonatal pups were transferred to hyperoxia environment (75% ± 5% O₂) at P7 and returned to RA at P12. At P17, pups were sacrificed for sample collection. Newborn mice for each group were obtained from at least three litters, and the number of pups per litter was limited to six to eight at birth. 

Laser-Induced CNV Model

We performed laser-induced CNV as described previously.²⁴ Mice were anesthetized with 1% pentobarbital sodium and their pupils dilated with compound tropicamide eye drops. Laser injury was performed on retinas using an 810 nm laser (Iridex Corporation, Mountain View, CA, USA; spot size 75 µm, power 120 mW, duration 75 ms) through a contact lens. Four laser burns were placed 2 disc diameters away from the optic disc at the 3, 6, 9, and 12 o'clock directions. Disruption of the Bruch's membrane was confirmed by the appearance of a cavitation bubble, indicating the success of modeling. On the seventh day after the modeling, mice were euthanized for sample collection. 

Immunofluorescence

The prepared samples were blocked and permeabilized in PBS for 12 hours at 4°C, which contained 0.2% bovine serum albumin, 0.3% Triton X-100, and 5% goat serum. Primary antibodies were then added and incubated overnight. The following day, samples were washed with PBS, then incubated with secondary antibodies and 4ʹ,6-diamidino-2-phenylindole (DAPI) for two hours. Images were captured with a Zeiss LSM880 confocal microscope (Carl Zeiss, Inc., White Plains, NY, USA). Diluted concentrations and supplier information for the antibodies are provided in Supplementary Table S1. 

Hematoxylin-Eosin (HE) Staining

The eyeballs were fixed in 4% paraformaldehyde, dehydrated in alcohol, and then embedded in paraffin. Paraffin-embedded samples were sliced into 3 µm sections. Then, to visualize tissue morphology, sections were dewaxed, rehydrated, HE-stained, and dehydrated. The slides were photographed with an optical microscope (Leica, Frankfurt, Germany). 

Evans Blue (EB) Assay

Evans blue dye (Sigma-Aldrich Corp., St. Louis, MO, USA) dissolved in saline solution(30 mg/mL) was injected intraperitoneally into P17 mice. Two hours after these injections, mice were anesthetized and sacrificed. For morphologic studies, retinal flat mounts were prepared and viewed as described above. For quantitative analysis, after cardiac 4% paraformaldehyde perfusion, retinas were harvested and fully dried at 4°C. Then, EB dye was extracted by incubating each retina in 60 µL formamide at 70°C for 18 hours. The extract was spun in a centrifuge at 10,000g for 20 minutes at 4 °C to obtain the supernatant, which was measured using spectrophotometry at 620 nm. The concentration of EB in the formamide was calculated from a standard curve. The formula was as follows: the concentration of EB in formamide (µg/µL) × 60 (µL)/dry weight of retina (mg). 

HUVEC Culture and Experiments

HUVECs were purchased from FuHeng Biology (Shanghai, China), confirmed by short tandem repeat profiling. The cells were cultured in DMEM/F12 (DF12) medium supplemented with 10% FBS and kept in a humidified atmosphere with 5% CO₂ at 37 °C. In VEGF-treated experiments, HUVECs were treated with complete medium containing 10 ng/mL VEGF (no. 293-VE-010/CF; R&D Systems, Minneapolis, MN, USA) referring to the previous study.²⁵ The methodological details of the cellular experiments (cell proliferation assay, scratch wound assay, transwell and tube formation assay) are given under Supplementary Methods. 

Western Blot

Protein extraction and western blot were performed as described previously.²⁶ The pools of tissue protein were prepared by mixing six retinas or RPE-choroid-sclera complexes from different mice. After separated by SDS-PAGE electrophoresis, the proteins were transferred to the polyvinylidene fluoride membranes, which were blocked and then incubated with the primary antibody and the secondary antibody. Finally, the blots were visualized by an enhanced chemiluminescence system (Millipore, Burlington, MA, USA). The list of antibodies used for the experiments is provided in Supplementary Table S1. 

Statistical Analysis

All experiments were independently repeated at least three times, and the data were presented as mean ± SEM. Statistical significance was determined using unpaired, two tailed t-test or one-way analysis of variance (ANOVA). P values <0.05 were considered to be significantly different. (*P < 0.05, **P < 0.01, and ***P < 0.001). Statistical analysis was carried out using Graphpad Prism 7.0 (GraphPad, San Diego, CA, USA). 

Results

Acrizanib had no Overt Toxic Effects on Retinal Vascular Development

To comprehensively assess the effect of acrizanib on physiological retinal vascular development, C57BL/6J mice were treated with an intravitreal injection of acrizanib (0.5 µL, 10 µM) at P3. Retinas were collected at P5, 7, 10, 12, 17, and 25 and processed for CD31-stained retinal whole mounts. The superficial retinal vessels of C57BL/6J mice began to grow from the center to the periphery at P0 and grew to the edge of the retina at approximately P7.²⁷ We thus chose P5 and P7 to observe superficial vessels and analyzed morphological features (such as tip cells) under different magnifications (Fig. 1A). At P5 and P7, acrizanib treatment had no effect on vascular area, vascular branches, tip cell number, and filopodia number (Fig. 1B). The effect of acrizanib on cell proliferation at P5 and P7 was also investigated (Supplementary Fig. S3A). The number of Ki67⁺ and CD31⁺ cells was not affected by acrizanib in either the vascular front and plexus (Supplementary Fig. S3B). The deep and middle vascular networks of C57BL/6J mice started to grow successively at P7.²⁷ Starting from P10, our research thus focused on the development of deep and intermediate vessels (Fig. 1C). We noted that a lower vascular density at P10 was detected in the acrizanib group compared to the control group, but this difference vanished after P10. Other than that, acrizanib treatment had no effect on the density of deep and intermediate vessels at P12, P14, P17, and P25 (Fig. 1D). Taken together, acrizanib had a mild effect on the development of deep retinal capillaries in P10 mice, which was very limited and did not cause any longer-term impact. Overall, acrizanib did not show negative effects on retinal vessel development. 

Figure 1.

 

The effect of acrizanib on the development of retinal vasculature. C57BL/6J mice were treated with an intravitreal injection of acrizanib (0.5 µL, 10 µM) at P3, and an equal volume of solvent was given in the fellow eye as a control. (A) Representative images of CD31-stained retinal whole mounts from P5 and P7 mice. From line 1 to line 4, images were presented at different magnifications to clearly show overall vascularization (line 1, scale bar: 1 mm), branches (line 2, scale bar: 300 µm), tip cells (line 3, scale bar: 100 µm), and filopodia of vessels (line 4, scale bar: 10 µm). White lines depict the size of retinas. (B) Quantification of vascular/total retinal area (%), branch points per field, tip cell number per field, and filopodia number per sprout in the Control group and the Control + Acrizanib group at P5 or P7 (n = 6 per group). (C) Representative images of retinal whole mounts stained with CD31 at P10, P12, P17, and P25 (scale bar: 50 µm). (D) Quantification of vascular/total retinal area (%) of superficial, intermediate, and deep vascular network in P10, P12, P17, and P25 retinas (n = 6 per group). Data are mean ± SEM. * P < 0.05; ns, no significance, using unpaired, two-tailed Student's t-test (B, D). GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer.

Acrizanib Attenuated Pathological RNV in the Mouse OIR Model

To investigate the effect of acrizanib on pathologic RNV, C57BL/6J mice received an intravitreal injection of acrizanib (0.5 µL, 10 µM) at P12 and were euthanized for sample collection at P17. Whole-mounted retinas stained with CD31 were used to visualize retinal vessels (Fig. 2A, upper image), and images at higher magnification were shown (Fig. 2A, lower image). Neovascular and central avascular areas were observed in retinas of mouse OIR models, which didn't exist in retinas of the control group. We found a significant decrease in retinal neovascular area (Fig. 2B) and central avascular area (Fig. 2C) after acrizanib intervention. In addition, we performed HE staining of eyeball paraffin sections (Fig. 2D) and CD31 immunofluorescence staining of eyeball frozen sections (Fig. 2F). In agreement with the preceding results, the numbers of neovascular cell nuclei anterior to the ILM (Fig. 2E) and CD31⁺ cells (Fig. 2G) decreased significantly in the OIR + Acrizanib group compared with the OIR group. To further confirm the antiangiogenic effect of acrizanib, we compared it with a widely used antiangiogenic agent in ophthalmic diseases, aflibercept. The results of RNV did not differ between the OIR + Acrizanib group and the OIR + Aflibercept group (Figs. 2B, 2E, 2G), and the central avascular area in the OIR + Aflibercept group was larger versus the OIR + Acrizanib group (Fig. 2C). In summary, acrizanib could significantly attenuate pathological RNV in the mouse OIR model, and its anti-angiogenic effect was comparable to that of aflibercept. 

Figure 2.

 

The effect of acrizanib on RNV in the mouse OIR model. C57BL/6J mice were put into a high oxygen environment (75% ± 5% O₂) at P7 and returned to room air at P12. At P12, the pups received an intravitreal injection of acrizanib (0.5 µL, 10 µM) or aflibercept (0.5 µL, 40 mg/mL) in one eye and solvent in the other eye as control. At P17, pups were euthanized for sample collection. (A) Upper image: Representative images of CD31-stained retinal whole mounts (scale bar: 1 mm). Lower image: The higher magnification images of areas indicated by white boxes (scale bar: 200 µm). Retinal neovascularization (B), and avascular area (C) were determined as described in the “Materials and Methods” using the retinal whole mounts (n = 6 per group). (D) Representative photomicrographs of HE-stained eyeball sections. Neovascular cell nuclei anterior to ILM represented extent of retinal neovascularization (scale bar: 50 µm). (E) Quantification of the neovascular cell nuclei anterior to the ILM per field in each group at P17 (n = 6 per group). (F) Representative immunofluorescent images for CD31 (red) and DAPI (blue) in each group (scale bar: 50 µm). (G) Quantification of CD31-positive cells nuclei per field in each group at P17 (n = 6 per group). Data are mean ± SEM. * P < 0.05; ** P < 0.01; *** P < 0.001; ns, no significance, using one-way ANOVA followed by Tukey's multiple comparisons test (B, C, E, G). OIR, oxygen induced retinopathy; ILM, internal limiting membrane; GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer.

Acrizanib Attenuated CNV in the Laser-Induced CNV Mouse Model

To investigate the effect of acrizanib on CNV, C57BL/6J mice received an intravitreal injection of acrizanib (1 µL, 10 µM) immediately after laser induction and were euthanized for sample collection 1 week after modeling. Eyeball paraffin sections were stained with HE in the CNV group and the CNV + Acrizanib group (Fig. 3A, upper image), and images at higher magnification were shown (Fig. 3A, lower image). We found a significant decrease in CNV thickness (Fig. 3B) and length (Fig. 3C) after acrizanib intervention. In addition, CD31 immunostaining of choroidal flat mounts were prepared and captured by a confocal laser microscope which could reconstruct the stereo image of CNV lesions (Fig. 3D). Then, the CNV area, thickness, and volume were quantified with three-dimensional image software (Figs. 3E–G). Consistent with the trend of HE staining, the CNV area (Fig. 3E), thickness (Fig. 3F), and volume (Fig. 3G) decreased significantly in the CNV + Acrizanib group compared with the CNV group. As before, we compared the effects of acrizanib with those of aflibercept. In the experimental results mentioned above, no significant differences were observed between the CNV + Acrizanib group and CNV + Aflibercept group (Figs. 3B, 3C, 3E–G). In summary, acrizanib could significantly attenuate pathological CNV in the laser-induced CNV mouse model with an inhibitory effect on neovascularization comparable to that of aflibercept. 

Figure 3.

 

The effect of acrizanib on CNV in the mouse laser-induced CNV model. Choroidal neovascularization was laser-induced in C57BL/6J mice (six to eight weeks). After modeling, mice were immediately treated by intravitreal injection with acrizanib (1 µL, 10 µM) or aflibercept (1 µL, 40 mg/mL) in one eye and solvent in the other eye as control. (A) Upper image: Representative photomicrographs of HE-stained eyeball sections in each group (CNV, CNV + Acrizanib, CNV + Aflibercept) (scale bar: 100 µm). Lower image: The high power images from each group were shown (scale bar: 50 µm). The thickness (B) and length (C) of choroidal neovascularization were quantified (n = 6 per group). (D) Representative three-dimensional images of the area, thickness, and volume of CNV were scanned by the confocal laser microscope. The area (E), thickness (F), and volume (G) of choroidal neovascularization were quantified using Zeiss Zen software (n = 6 per group). Data are mean ± SEM. * P < 0.05; ** P < 0.01; *** P < 0.001; ns, no significance, using one-way ANOVA followed by Tukey's multiple comparisons test (B, C, E–G). CNV, choroidal neovascularization; GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer; CL, choroid layer; SL, sclera layer.

Acrizanib Suppressed Vascular Proliferation but had no Impact on Apoptosis in OIR and CNV Models

Next, we assessed the effect of acrizanib on cellular proliferation and apoptosis in the neovascularization areas. Flat-mounted retinas in the OIR group and OIR + Acrizanib group were stained with Ki67 (nuclear proliferation marker) and cleaved-caspase-3 (apoptotic marker) for immunofluorescence (Figs. 4A, 4B). Immunofluorescence results demonstrated that the number of Ki67⁺ and CD31⁺ cells in the OIR group was more than that in the OIR + Acrizanib group (Fig. 4C), whereas the number of cleaved-caspase-3⁺ and CD31⁺ cells did not differ between the two groups (Fig. 4D). Meanwhile, choroidal flat mounts in the CNV and CNV+Acrizanib groups were also stained with Ki67 and cleaved-caspase-3 for immunofluorescence (Figs. 4E, 4F). Consistent with the trend in OIR retinas, the number of Ki67⁺ and CD31⁺ cells in the CNV group was more than that in the CNV + Acrizanib group (Fig. 4G), whereas the number of cleaved-caspase-3⁺ and CD31⁺ cells did not differ between the two groups (Fig. 4H). Taking these results together, acrizanib suppressed cellular proliferation in the retinal and choroidal neovascular areas without affecting apoptosis. 

Figure 4.

 

The effect of acrizanib on cell proliferation and apoptosis of neovascularized areas. Immunofluorescence staining of retinal flatmounts from OIR and OIR + Acrizanib mice. Red represents CD31; green represents Ki67 (A) or C-Cas-3 (B); blue represents the nucleus stained by DAPI (scale bar: 50 µm). Quantitation of Ki67 (C) or C-Cas-3 (D) positive cells in the neovascularized areas (n = 6 per group). Immunofluorescence staining of choroidal flatmounts from CNV and CNV + Acrizanib mice. Red represents CD31; green represents Ki67 (E) or C-Cas-3 (F); blue represents the nucleus stained by DAPI (scale bar: 100 µm). Quantitation of Ki67 (G) or C-Cas-3 (H) positive cells in the neovascularized areas (n = 6 per group). Data are mean ± SEM. * P < 0.05; ** P < 0.01; ns, no significance, using unpaired, two-tailed Student t-test (C, D, G, H). OIR, oxygen induced retinopathy; CNV, choroidal neovascularization; C-Cas-3, cleaved caspase-3.

Acrizanib Reduced Vascular Leakage in OIR Mice by Inhibiting the Increase of VEGF- Induced Vascular Permeability

Pathological angiogenesis is usually accompanied by vascular hyperpermeability, leading to vascular leakage.²⁸ Therefore we investigated the role of acrizanib on vascular leakage in the mouse OIR model. EB assays were carried out to evaluate vascular permeability. EB dye was injected intraperitoneally, and retinas were harvested two hours later. Next, retinal flat mounts were prepared to visualize dye extravasation (Fig. 5A). It was observed that EB dye extravasation was significantly decreased in the neovascular tufts following acrizanib treatment compared with the OIR group (Fig. 5A, white box). Furthermore, spectrophotometric assays were used to quantify the residual EB dye in the retina following formamide extraction. The histogram showed that extravasated EB dye was significantly increased in the OIR group compared to the control group, and this trend was suppressed by acrizanib treatment (Fig. 5G). 

Figure 5.

 

The effect of acrizanib on vascular endothelial permeability in vivo and in vitro. (A) Upper image: Representative image of Evans blue leakage in retinas of OIR and OIR + Acrizanib mice (scale bar: 50 µm). Lower image: The high power images from each group were shown (scale bar: 25 µm). (G) Quantitation of Evans blue extravasation in retinas of mice at P17. The content of EB in retina (µg/mg) = the concentration of EB in formamide (µg/µL) × 60 (µL)/dry weight of retina (mg). (n = 6 per group). (B) Immunofluorescence staining of retinal flatmounts from OIR and OIR + Acrizanib mice. Red represents CD31; green represents VE-cadherin; blue represents the nucleus stained by DAPI (scale bar: 50 µm). (H) Mean fluorescence intensity of VE-cadherin in the neovascularized areas was quantified (n = 6 per group). Immunofluorescence images of HUVECs in different treatment groups (control, control + Acrizanib, VEGF, VEGF + Acrizanib). Red represents Claudin-1 (C) or ZO-1 (D); blue represents the nucleus stained by DAPI (scale bar: 30 µm). Mean fluorescence intensity of Claudin-1 (E) or ZO-1 (F) at cell junctions was quantified (n = 6 per group). (I) Western blot of HUVEC lysate showed the expression of Claudin-1 and ZO-1 in each group. GAPDH served as the loading control. The histogram showed the densitometric analysis of the average level of Claudin-1/GAPDH (J) and ZO-1/GAPDH (K). Relative protein levels were presented by taking control as 100% (n = 3 per group). Data are mean ± SEM. * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001; ns, no significance, using unpaired, two-tailed Student's t-test (H) and one-way ANOVA followed by Tukey's multiple comparisons test (E–G, J, K). OIR, oxygen-induced retinopathy.

The tight and adherens junctions between adjacent ECs are important factors that affect the permeability of blood vessels.²⁹ We next analyzed the levels or distribution of tight junction proteins (Claudin-1 and ZO-1) and adherens junction protein (VE-cadherin) in ECs. Flat-mounted retinas were stained with VE-cadherin in the OIR and OIR + Acrizanib groups for immunofluorescence (Fig. 5B). Compared with the OIR group, the fluorescence intensity of VE-cadherin was significantly higher in the OIR + Acrizanib group, which exhibited a more intact and continuous morphological structure (Fig. 5H). HUVECs were stained for Claudin-1 and ZO-1 in different groups for immunofluorescence (Figs. 5C, 5D). The fluorescence signal of Claudin-1 and ZO-1 was low and discontinuous in the VEGF group compared to the control group, and this trend was reversed by acrizanib treatment (Figs. 5E, 5F). Western blot of Claudin-1 and ZO-1 also showed the similar trend (Figs. 5I–K). The evidence above suggested that acrizanib reduced vascular leakage by inhibiting VEGF-induced disruption of tight and adherens junctions in ECs. 

Acrizanib Reduced Inflammation Associated With Neovascularization in OIR and CNV Models

Because inflammation has a strong association with angiogenesis,³⁰ we examined the inflammatory reaction in the neovascular areas in OIR and CNV models. Retinal flat mounts were stained with Iba1 (microglia/macrophage marker), CD68 (activated microglia/macrophage marker) and CD45 (leukocyte marker) in the OIR and OIR + Acrizanib groups for immunofluorescence (Figs. 6A–C). We observed large numbers of Iba1⁺ cells, CD68⁺ cells and CD45⁺ cells that accumulated at the neovascular areas of retinas in the OIR group, and the numbers of these cells decreased after acrizanib intervention (Figs. 6D–F). Choroidal flat mounts were also stained with Iba1, CD68, and CD45 in the CNV and CNV + Acrizanib groups for immunofluorescence (Figs. 6G–I). Consistent with the trend in OIR retinas, the numbers of Iba1⁺ cells, CD68⁺ cells and CD45⁺ cells in the CNV group were higher than those in the CNV + Acrizanib group (Figs. 6J–L). These results indicated that acrizanib reduced the infiltration of microglia/macrophages and leukocytes in the neovascular areas in OIR and CNV models and suppressed the activation of microglia/macrophages. In addition, we examined the protein expression of the inflammatory factors TNF-α and IL-1β in the tissue (Figs. 6M, 6P). We observed an increase in both TNF-α and IL-1β expression in the retinas of the OIR model (Figs. 6N–O) and the choroids of the CNV model (Figs. 6Q–R), a trend that acrizanib was able to inhibit. Overall, our results demonstrated acrizanib's ability to inhibit inflammation associated with neovascularization. 

Figure 6.

 

The effect of acrizanib on inflammation associated with neovascularization. Immunofluorescence staining of retinal flatmounts from OIR and OIR + Acrizanib mice. Red represents CD31; green represents Iba1 (A), CD68 (B) or CD45, (C); blue represents the nucleus stained by DAPI (scale bar: 50 µm). Quantitation of Iba1 (D), CD68 (E), and CD45 (F) positive cells in the neovascularized areas (n = 6 per group). Immunofluorescence staining of choroidal flatmounts from CNV and CNV + Acrizanib mice. Red represents CD31; green represents Iba1 (G), CD68 (H), or CD45 (I); blue represents the nucleus stained by DAPI (scale bar: 100 µm). Quantitation of Iba1 (J), CD68 (K), CD45 (L) positive cells in the neovascularized areas (n = 6 per group). (M) Western blot of retinal lysate showed the expression of TNF-α and IL-1β in each group. GAPDH served as the loading control. (N, O) The histogram showed the densitometric analysis of the average levels of TNF-α/GAPDH (N) and IL-1β/GAPDH (O). Relative protein levels were presented by taking control as 100% (n = 3 per group). (P) Western blot of RPE-choroid-sclera lysate showed the expression of TNF-α and IL-1β in each group. GAPDH served as the loading control. The histogram showed the densitometric analysis of the average levels of TNF-α/GAPDH (Q) and IL-1β/GAPDH (R). Relative protein levels were presented by taking control as 100% (n = 3 per group). Data are mean ± SEM. * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001; ns, no significance, using unpaired, two-tailed Student's t-test (D–F, J–L), one-way ANOVA followed by Tukey's multiple comparisons test (N, O, Q, R). OIR, oxygen induced retinopathy; CNV, choroidal neovascularization.

Acrizanib Inhibited the Proliferation, Migration, and Tube Formation Ability of VEGF-Treated HUVECs

To gain a better understanding of the effect of acrizanib on ECs, we performed in vitro experiments using HUVECs stimulated with or without VEGF. We first measured the cytotoxicity of acrizanib using CCK8 assays to determine the non-cytotoxic concentration (50 nM) for subsequent experiments (Fig. 7A). In CCK8 cell proliferation assays, acrizanib inhibited cell proliferation in VEGF-treated HUVECs and had no impact on cells in the control group (Fig. 7B). Western blot quantification for the cell proliferation marker PCNA (Figs. 7C, 7D) and EdU incorporation assays (Figs. 7E, 7I) also demonstrated similar results. 

Figure 7.

 

The effect of acrizanib on the proliferation, migration, and tube formation of HUVECs. (A) The toxicity of different concentrations of acrizanib on HUVECs was detected by the CCK8 assay. HUVECs were treated with acrizanib (0, 25, 50, 100, 200, and 400 nM) for 24 hours (n = 3 per group). The control group was set at 100%. (B) The effect of 50 nM acrizanib on the proliferation of 10 ng/mL VEGF-treated HUVECs was measured by CCK8 assays (n = 3 per group). The control group was set at 100%. (C) Western blot of HUVEC lysate showed the expression of PCNA, a cellular proliferation marker, in each group. GAPDH served as the loading control. (D) The histogram showed the densitometric analysis of the average level of PCNA/GAPDH. Relative protein levels were presented by taking control as 100% (n = 3 per group). (E) Representative images of EdU incorporation assay in each group. EdU staining (green) showed the effects of 50 nM acrizanib on the proliferation of VEGF-treated HUVECs (scale bar: 100 µm). (F) Representative photomicrographs of scratch assay in each group for 0, 12 hours (scale bar: 1 mm). (G) Representative photomicrographs of transwell assay captured after 18 hours in each group (scale bar: 1 mm). (H) Representative photomicrographs of HUVEC tube formation assay captured after six hours in each group (scale bar: 1 mm). (I) Ratio of EdU-positive cells/total cells was quantified (n = 6 per group). (J) Wound closure (%) was quantified as (wound closure area/initial wound area) × 100% (n = 6 per group). (K) Number of HUVECs was counted on the lower surface of the transwell membrane in each group (n = 6 per group). (L) Number of branches in each group was quantified (n = 6 per group). Data are mean ± SEM. * P < 0.05; ** P < 0.01; *** P < 0.001; ns, no significance, using one-way ANOVA followed by Tukey's multiple comparisons test (B, D, I–L).

Migration and tube formation of ECs are important steps of angiogenesis.³¹ Scratch and transwell assays were performed to assess the effect of acrizanib on the migration of ECs (Figs. 7F, 7G). Histograms of the healing speed of scratches and the number of invading cells suggested that acrizanib significantly inhibited VEGF-induced HUVECs migration (Figs. 7J, 7K). Matrigel tube formation assays were conducted to evaluate endothelial tube formation capacity (Fig. 7H). Based on the number of branch points, VEGF-induced tube formation in HUVECs was reduced in the presence of acrizanib by counting the branch points (Fig. 7L). In summary, acrizanib treatment suppressed the proliferative, migratory and tube-forming capacity of VEGF-treated HUVECs. 

VEGFR2 Phosphorylation Associated With Neovascularization Was Strongly Inhibited by Acrizanib in OIR and CNV Models

We next examined whether total VEGFR2 (t-VEGFR2) and phosphorylated VEGFR2 (p-VEGFR2) are affected by acrizanib in OIR and CNV mice. Immunofluorescence (Figs. 8A, 8B) and Western blot (Fig. 8G) were used to detect the expression of t-VEGFR2 and p-VEGFR2 in different groups of retinas. Immunofluorescence results showed that t-VEGFR2 and p-VEGFR2^Tyr1173 were highly colocalized with CD31⁺ ECs. Compared with the control group, t-VEGFR2 was significantly upregulated in the OIR group, and this trend was down-regulated after the intervention of acrizanib (Figs. 8C, 8I). However, p-VEGFR2^Tyr1173 was highly expressed in retinas of the OIR group but rarely expressed in the control group and the OIR + Acrizanib group (Figs. 8E, H). We repeated the experiments described above on the choroids (Figs. 8J, 8K, 8P). The results were similar to those in the retinas (Figs. 8L, 8N, 8Q, 8R). 

Figure 8.

 

The effect of acrizanib on total/phosphorylated VEGFR2 in vivo. Immunofluorescence staining in sections of eye balls from the different groups (Control, Control + Acrizanib, OIR, and OIR + Acrizanib). Red represents CD31; green represents VEGFR2 (A), p-VEGFR2^Tyr1173; (B) blue represents the nucleus stained by DAPI (scale bar: 50 µm). Quantitation of VEGFR2- (C, D) and p-VEGFR2^Tyr1173– (E, F) positive cells in the neovascularized areas (n = 6 per group). (G) Western blot of retinal lysate showed the expression of p-VEGFR2^Tyr1173 and VEGFR2 in each group. GAPDH served as the loading control. The histogram showed the densitometric analysis of the average levels of p-VEGFR2^Tyr1173/GAPDH (H) and VEGFR2/GAPDH (I). Relative protein levels were presented by taking control as 100% (n = 3 per group). (J, K) Immunofluorescence staining of choroidal flatmounts from CNV and CNV + Acrizanib mice. Red represents CD31; green represents VEGFR2 (J), p-VEGFR2^Tyr1173 (K); blue represents the nucleus stained by DAPI (scale bar: 100 µm). Quantitation of VEGFR2 (L, M), p-VEGFR2^Tyr1173 (N, O) positive cells in the neovascularized areas (n = 6 per group). (P) Western blot of RPE-choroid-sclera lysate showed the expression of p-VEGFR2^Tyr1173 and VEGFR2 in each group. GAPDH served as the loading control. The histogram showed the densitometric analysis of the average levels of VEGFR2^Tyr1173/GAPDH (Q) and VEGFR2/GAPDH (R). Relative protein levels were presented by taking control as 100% (n = 3 per group). Data are mean ± SEM. * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001; ns, no significance, using unpaired, two-tailed Student's t-test (L–O), one-way ANOVA followed by Tukey's multiple comparisons test (C–F, H, I, Q, R). OIR, oxygen induced retinopathy; CNV, choroidal neovascularization; GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer.

It is noteworthy that the number of vascular ECs varied greatly between different groups, so it seems more meaningful to evaluate the effect of acrizanib on VEGFR2 expression in vascular ECs. We assessed VEGFR2 expression by quantifying the proportion of vascular ECs (CD31⁺) expressing VEGFR2 (VEGFR2⁺ and CD31⁺) (Figs. 8D, 8F, 8M, 8O). Statistical results showed that acrizanib could reduce the number of p-VEGFR2^Tyr1173+ ECs in the OIR model and CNV model (Figs. 8F, 8O) but had no effect on t-VEGFR2 (Figs. 8D, 8M). 

Acrizanib Suppressed Different VEGFR2 Phosphorylation Sites to Varying Degrees and Their Downstream Pathways in VEGF-Treated HUVECs

Several key phosphorylated sites (Try951, Try996, Try1059, Try1175 and Try1214) in the cytoplasmic domain of VEGFR2 mediate different important signaling pathways.¹⁷ Western blot was used to explore the effect of acrizanib on the different phosphorylation sites of VEGFR2 in HUVECs stimulated with or without VEGF (Fig. 9A). The histogram showed that VEGF could stimulate the phosphorylation of VEGFR2 at multiple sites in HUVECs, while acrizanib could inhibit phosphorylation of VEGFR2-Tyr951, 996, 1059, 1175 and 1214 in VEGF-treated HUVECs (Figs. 9B–F). The p-VEGFR2^Tyr1175 immunofluorescence results showed the same trend as that of the Western blot (Figs. 9H, 9I). We also compared the extent of phosphorylation inhibition at different phosphorylation sites of VEGFR2. The histogram suggested that the degree of inhibition was significantly different between different VEGFR2 phosphorylation sites, with a roughly 2.5-fold difference between the highest (Tyr1214) and lowest (Tyr996) values (Fig. 9G). Next, we explored the role of acrizanib on VEGFR2 downstream signal pathways activated by VEGF (Fig. 9J). The phosphorylation levels of AKT, eNOS, PLC-γ1, ERK1/2, p38-MAPK and FAK were increased by stimulation of VEGF, and this stimulative effect was inhibited by acrizanib (Figs. 9K–P). In conclusion, acrizanib could suppress phosphorylation of VEGFR2- Tyr951, 996, 1059, 1175, and 1214 in VEGF-treated HUVECs to varying degrees, thereby inhibiting the activation of key molecules in the signaling pathways downstream of VEGFR2 (Supplementary Fig. S5). 

Figure 9.

 

The effect of acrizanib on different phosphorylation sites of VEGFR2 and their downstream pathways in vitro. (A) Western blot of HUVEC lysate showed the expression of p-VEGFR2^Tyr951, p-VEGFR2^Tyr996, p-VEGFR2^Tyr1059, p-VEGFR2^Tyr1175 and p-VEGFR2^Tyr1214 in each group (Control, Control+Acrizanib, VEGF and VEGF + Acrizanib). The histogram showed the densitometric analysis of the average levels of p-VEGFR2^Tyr951/t-VEGFR2 (B), p-VEGFR2^Tyr996/t-VEGFR2 (C), p-VEGFR2^Tyr1059/t-VEGFR2 (D), p-VEGFR2^Tyr1175/t-VEGFR2 (E), and p-VEGFR2^Tyr1214/t-VEGFR2 (F). Relative protein levels were presented by taking control as 100% (n = 3 per group). (G) The histogram showed the extent of inhibition of different VEGFR2 phosphorylation sites. Inhibition of phosphorylation (%) = (difference between relative protein levels in VEGF group and VEGF + Acrizanib group) /relative protein level in VEGF group × 100%. (H) Immunofluorescence images of HUVECs in different treatment groups. Red represents p-VEGFR2^Tyr1175; Green represents VEGFR2; Blue represents the nucleus stained by DAPI (scale bar: 50 µm). (I) Mean fluorescence intensity of VEGFR2^Tyr1175 was quantified (n = 6 per group). (J) Western blot of HUVEC lysate showed the expression of p-AKT, t-AKT, p-eNOS, t-eNOS, p-PLC-γ1, t-PLC-γ1, p-ERK1/2, t-ERK1/2, p-p38-MAPK, t-p38-MAPK, p-FAK and t-FAK in each group. GAPDH served as the loading control. The histogram showed the densitometric analysis of the average levels of p-AKT/t-AKT (K), p-eNOS/t-eNOS (L), p-PLC-γ1/t-PLC-γ1 (M), p-ERK1/2/t-ERK1/2 (N), p-p38-MAPK/t-p38-MAPK (O), and p-FAK/t-FAK (P). Relative protein levels were presented by taking control as 100% (n = 3 per group). Data are mean ± SEM. * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001; ns, no significance, using one-way ANOVA followed by Tukey's multiple comparisons test (B–G, I, K–P).

Discussion

Anti-VEGF therapy has now become the mainstay for treating fundus neovascularization.¹⁰ However, previous research revealed that phosphorylation of VEGFR2 would be restored in vascular ECs with longer durations of anti-VEGF treatment,³² suggesting that anti-VEGF could not continuously prevent the activation of VEGFR2. One reason is that intravitreal anti-VEGF agents only target part of the VEGF family.³³ VEGF-C and -D are not blocked by anti-VEGF and can induce angiogenesis by activating VEGFR2.^34,35 Also, the VEGF family is not the only factor that activates VGEFR2. It has been reported that non-VEGF ligands and shear stress can also induce VEGFR2 phosphorylation in the absence of VEGF.¹⁸ The interaction between anti-VEGF therapy and VEGFR2 was found in ophthalmic clinical practice. For example, VEGFR2 gene polymorphisms affected the clinical outcomes of anti-VEGF therapy in nAMD patients.³⁶ Additionally, anti-VEGF was correlated with VEGFR2 levels in the vitreous fluid.³⁷ All of this suggested that directly blocking VEGFR2 may be an attractive therapeutic strategy to treat fundus neovascularization. 

Safety and toxicity issues are essential considerations for the clinical translation of a therapeutic intervention. Because the physiological functions of ECs are almost entirely mediated through activation of VEGFR2,¹⁷ blocking VEGFR2 usually affects physiological angiogenesis. How to inhibit pathological angiogenesis while reducing the damage to physiological angiogenesis is the focus and difficulty of developing anti-VEGFR2 therapies. To date, it seems that many studies aiming to use VEGFR2 as a therapeutic target for retinal neovascular diseases have not thoroughly examined this potential risk.^38–40 Budd et al.⁴¹ reported that the VEGFR2 inhibitor SU5416 could significantly reduce the number or length of filopodia on endothelial tip cells in rats. In our study, the effects of acrizanib on physiological angiogenesis were comprehensively evaluated in neonatal mice. The results showed that acrizanib had a mild effect on the development of deep retinal capillaries in P10 mice, which was very limited and did not cause any long-term impact. Overall, acrizanib did not show negative effects on retinal vessel development, most likely because the activation of VEGFR2 cannot be completely blocked by acrizanib. The study by Zarkada et al.⁴² confirmed that small amounts of VEGFR2 can support physiological retinal angiogenesis to a certain extent. Another important point to note is that VEGFR2 is expressed in the neural retina and can regulate the development and function of neurons.⁴³ To evaluate the neurotoxicity of acrizanib in the retina, we additionally examined the retinal morphology of the acrizanib therapeutic control groups. The thicknesses of the INL and ONL had no noteworthy change (Supplementary Fig. S2). This was in line with the literature that demonstrated that the suppression of VEGFR2 did not damage retinal neurons.⁴⁴ Although acrizanib did not show significant toxicity in this study, considering the differences between mouse and human eyes and the transient decrease in vascular density of P10 mice, the safety of acrizanib still needs to be closely monitored. 

Acrizanib is a novel receptor TKI targeting VEGFR2. It was reported that acrizanib eye drop treatment had better antiangiogenic efficacy than intravitreal anti-VEGF injections in animal experiments.²¹ This prompted our interest in further research despite the failure of the clinical trial for acrizanib eye drops.²² Therefore, we used mouse OIR and laser-induced CNV models to comprehensively access the inhibitory capacity of acrizanib on fundus neovascularization. The OIR and CNV models are most widely used to induce angiogenesis and stimulate pathological RNV and CNV, respectively.⁴⁵ Our results showed that a single intravitreal injection of acrizanib can significantly inhibit RNV and CNV with an inhibitory effect on neovascularization comparable to that of aflibercept, a drug heavily used in common clinical practice. Furthermore, reduction of avascular area during hypoxia is a key characteristic feature of physiological retinal revascularization.⁴⁶ Acrizanib decreased the avascular area of the retina, suggesting its capability of promoting the physiological revascularization of the central retinas of OIR mice. Consistently, Simmons et al.⁴⁷ reported that endothelial-specific VEGFR2 knockout led to a decreased avascular area in OIR rats. Next, we further evaluated the antiangiogenic effect of acrizanib in vitro. EC proliferation, migration and tube formation are critical steps in angiogenesis.³¹ Our findings showed that acrizanib abrogated the VEGF's induction of cell proliferation, migration and tube formation without affecting the biological function of ECs not stimulated by VEGF. Taken together with the Ki67 staining of retinal mounts, we speculate that acrizanib principally targets activated ECs in neovascular tufts. 

The mechanism of VEGFR2 activation is very well characterized.¹⁸ In short, VEGFR2 forms dimers after being activated by the ligand, with subsequent trans-phosphorylation of tyrosine residues within the intracellular domain.¹⁸ Its phosphorylation process can be prevented by acrizanib.²¹ By using computer simulations, Modi et al. inferred that acrizanib interacted specifically with the inactive conformation (called DFG out) of VEGFR2.⁴⁸ This interaction prevents the intracellular kinase domain from binding adenosine triphosphate, making tyrosine residues unable to be phosphorylated.⁴⁸ In this study, we validated that acrizanib could inhibit VEGFR2 phosphorylation of activated vascular ECs in vitro and in vivo. Remarkably, TKI-mediated suppression of phosphorylation is site-selective. Zhang et al.⁴⁹ found that the phosphorylation of epidermal growth factor receptor (EGFR)-Tyr1197 could be inhibited by afatinib but not erlotinib in human lung adenocarcinoma cells, both of which were TKIs against EGFR. However, different phosphorylation sites of the same protein mediate diverse signaling pathways, which result in distinct phenotypic effects.¹⁷ Taking VEGFR2 for example, VEGFR2-Tyr951 affects EC permeability through the PI3K-AKT pathway⁵⁰; VEGFR2-Tyr1175 promotes EC proliferation and migration via PLCγ-ERK1/2 and FAK, respectively^51,52; VEGFR2-Tyr1214 facilitates EC migration through p38 MAPK.⁵³ Therefore we further explored the effect of acrizanib on the different phosphorylation sites of VEGFR2 and downstream pathways. Our study found that acrizanib could inhibit the phosphorylation of VEGFR2- Tyr951, 996, 1059, 1175 and 1214 in VEGF-treated ECs but to different extents. Previous studies highlighted that different phosphorylation sites of VEGFR2 have varying importance in physiological and pathological conditions.⁵⁴ For instance, VEGFR2- Tyr1175 mutant mice are lethal but VEGFR2- Tyr951, 1214 mutant mice are viable.^54,55 We speculated that by modulating the degree of VEGFR2 phosphorylation at different sites, acrizanib can inhibit pathological neovascularization without affecting physiological vascular development. Certainly, further research is required to confirm this hypothesis. 

Limitations of this study should be noted. The determination of drug dosages in the vitreous cavity is important for obtaining optimum therapeutic results and for clinical tests in the drug development process.⁵⁶ This study evaluated the effects of different acrizanib doses on pathological and physiological angiogenesis in the retina, using mice as experimental subjects. The results showed that intravitreal injection of acrizanib at concentrations of 1 to 10 µM (0.23-2.3 ng/eye) was safe and effective in mice (Supplementary Fig. S4). It should be noted that although many experiments use mice for intravitreal drug injection, to date, there is no literature demonstrating the translation of intravitreal pharmacokinetics from mice to humans.⁵⁶ Currently, research on ocular drug delivery and pharmacokinetics typically involves larger animals with larger eyes, such as rabbits and monkeys.⁵⁷ Thus our subsequent studies will primarily focus on rabbits and monkeys to further evaluate the safe dosage range of acrizanib. 

In summary, our current study suggested the potential efficacy and safety of acrizanib for suppressing fundus neovascularization, which functioned by inhibiting multiple phosphorylation sites of VEGFR2 in ECs to varying degrees. These results indicated that acrizanib might hold promise as a potential candidate for the treatment of ocular vascular diseases. 

Acknowledgments

Supported by the National Natural Science Foundation of China (82271099, 81970807), the Research Grant from Guangzhou Municipal Science and Technology Bureau in China (202201020507), and the Natural Science Foundation of Guangdong Province in China (2022A1515010355). 

Disclosure: X. Tang, None; K. Cui, None; P. Wu, None; A. Hu, None; M. Fan, None; X. Lu, None; F. Yang, None; J. Lin, None; S. Yu, None; Y. Xu, None; X. Liang, None