Sema4D Knockout Attenuates Choroidal Neovascularization by Inhibiting M2 Macrophage Polarization Via Regulation of the RhoA/ROCK Pathway

Abstract

Purpose: The aim of this study was to elucidate the role of Sema4D in the pathogenesis of senescence-associated choroidal neovascularization (CNV) and to explore its underlying mechanisms.

Methods: In this study, we utilized a model of laser-induced CNV in both young (3 months old) and old (18 months old) mice, including those with or without Sema4D knockout. The expression and localization of Sema4D in CNV were assessed using PCR, Western blot, and immunostaining. Subsequently, the morphological and imaging examinations were used to evaluate the size of CNV and vascular leakage. Finally, the expression of M2 markers, senescence-related markers, and molecules involved in the RhoA/ROCK pathway was detected.

Results: We found that Sema4D was predominantly expressed in macrophages within CNV lesions, and both the mRNA and protein levels of Sema4D progressively increased following laser photocoagulation, a trend more pronounced in old mice. Moreover, Sema4D knockout markedly inhibited M2 polarization in senescent macrophages and reduced the size and leakage of CNV, particularly in aged mice. Mechanistically, aging was found to upregulate RhoA/ROCK signaling, and knockout of Sema4D effectively suppressed the activation of this pathway, with more significant effects observed in aged mice.

Conclusions: Our findings revealed that the deletion of Sema4D markedly inhibited M2 macrophage polarization through the suppression of the RhoA/ROCK pathway, ultimately leading to the attenuation of senescence-associated CNV. These data indicate that targeting Sema4D could offer a promising approach for gene editing therapy in patients with neovascular age-related macular degeneration.

Age-related macular degeneration (AMD) is the leading cause of irreversible blindness in industrialized nations, particularly in adults over 60 years of age.^1,2 The global prevalence of AMD is estimated at 8.7%, with the projected number of affected individuals worldwide expected to reach 288 million by 2040.³ Clinically, advanced AMD is classified into neovascular (wet) AMD or atrophic (dry) AMD.⁴ Neovascular AMD, accounting for almost 90% of vision loss associated with AMD, is characterized by choroidal neovascularization (CNV), which is the proliferation of abnormal blood vessels from the choroid through the Bruch's membrane toward the retina.^5,6 Currently, intravitreal injections of anti-vascular endothelial growth factor (VEGF) agents represent the first-line therapy for wet AMD.⁷ However, several challenges persist. Many patients experience an incomplete response to anti-VEGF treatment.⁸ Long-term repeated intravitreal injections of these agents can aggravate oxidative stress and hypoxia in the retino-choroidal complex, leading to retinal pigment epithelium (RPE) atrophy, photoreceptor apoptosis, and permanent vision loss.^9,10 In addition, this treatment also imposes substantial psychological and economic burdens on patients and their families.¹¹ Thus, continuing efforts to search for novel molecular targets for neovascular AMD are still needed. 

Although the exact pathogenesis underlying AMD remains unclear, it is well-established that the innate immune system, particularly macrophages, plays a pivotal role in regulating CNV.^12,13 Macrophages display remarkable plasticity and diversity, allowing for adaptation to the surrounding microenvironment. They can differentiate into classically activated (M1) or alternatively activated (M2) subtypes.¹⁴ M1 macrophages are activated by lipopolysaccharide or interferon-γ, expressing CD16/32, CD86, inducible nitric oxide synthase, interleukin (IL)-6 and IL-12. In contrast, M2 macrophages are induced by IL-4, IL-10, or IL-13, producing CD163, CD206, arginase-1 (Arg-1), chitinase 3-like 3 (Ym-1), and found in inflammatory zone 1 (Fizz-1).^15–17 Unlike their M1 counterparts, M2 macrophages are mainly involved in angiogenesis and tissue remodeling.¹⁸ We and others have previously reported that senescent macrophages polarize toward an M2 phenotype and promote CNV formation.^19,20 

Semaphorins, a large family of transmembrane proteins originally characterized as axonal guidance factors, have since been shown to contribute to immune response regulation, angiogenesis, and bone metabolism, which influence a broad spectrum of processes related to health and disease.^21,22 Semaphorin 4D (Sema4D/CD100), a member of the class IV semaphorin family, is expressed on the surface of lymphocytes, monocytes, macrophages, microglia, and endothelial cells, and its expression typically increases upon activation.^23,24 Three distinct affinity receptors have been identified for Sema4D: PlexinB1, PlexinB2, and CD72.²⁵ By interacting with its high-affinity receptor plexinB1, Sema4D plays critical roles in immune cell activation and migration, angiogenesis, vascular permeability, and tumor progression.^26–28 Recent evidence has indicated a significant elevation in Sema4D levels within the aqueous humor of patients with diabetic retinopathy (DR), exhibiting an inverse correlation with the success of anti-VEGF therapy. Moreover, Sema4D/PlexinB1 induced endothelial cell dysfunction, pericyte loss, and retinal neovascularization in animal models of DR.²⁹ However, the functional roles of Sema4D in CNV development and its underlying mechanisms are not yet known. 

Ras homolog gene family member A (RhoA) is a small GTPase protein that belongs to the Ras superfamily, which is inactive when bound to GDP but becomes active upon binding to GTP via guanine nucleotide exchange factors (GEFs).³⁰ Rho-associated coiled-coil protein kinase (ROCK), a downstream effector of RhoA, comprises two isoforms: ROCK1 and ROCK2.³¹ The Rho/ROCK signaling pathway plays a key role in the modulation of several biological processes, including cell proliferation, adhesion, migration, and gene transcription.^32,33 It has been demonstrated that ligation with Sema4D prompts PlexinB1 to recruit both PDZ-RhoGEF and leukemia-associated RhoGEF (LARG), activating the RhoA/ROCK signaling, thereby leading to the phosphorylation of myosin phosphatase target subunit 1 (MYPT1) and myosin light chain (MLC).^34,35 Additionally, an earlier study has shown that ROCK signaling results in the pronounced increase of pro-angiogenic M2 macrophages, exacerbating CNV and vascular leakage.³⁶ 

In this study, wild-type (WT) and Sema4D knockout (Sema4D-KO) mice of different ages were used to explore the impact of Sema4D on macrophage polarization and senescence-associated CNV. Our findings demonstrated that knockout of Sema4D significantly inhibited macrophage polarization toward the pro-angiogenic M2 type, and effectively reduced the CNV size and associated leakage, predominantly by regulating the RhoA/ROCK signaling pathway. 

Materials and Methods

Animals

The young (3 months of age) and old (18 months of age) male WT C57BL/6J mice were purchased from GemPharmatech Co., Ltd (Nanjing, China). Sema4D-KO mice, on a C57BL/6J background and confirmed to be free of the rd8 mutation, were generated by Shanghai Model Organisms Center Inc. (Shanghai, China). The young (3 months of age) and old (18 months of age) male Sema4D-KO mice were used in this study. All experimental procedures in this study adhered to the ARVO Statement for the use of Animals in Ophthalmic and Vision Research, and were approved by the Institutional Animal Care and Use Committee of Zhongshan Ophthalmic Center. 

Laser-Induced CNV Mouse Model

CNV was induced in young and old mice with laser injury, as previously described.^20,37 Briefly, the animals were anesthetized using 1% pentobarbital sodium (50 mg/kg) injected intraperitoneally, and the pupils were dilated using tropicamide phenylephrine. Laser photocoagulation (OcuLight, IRIDEX) was performed at 810 nm wavelength, 120 mW power, 75 ms duration, and 75 µm spot size. Four laser spots were applied, approximately 2 disc diameters from the optic nerve at the 3, 6, 9, and 12 o'clock positions. Only burns that produced a cavitation bubble without hemorrhage, indicative of Bruch's membrane disruption, were deemed effective and included in this study. 

Experimental Design

To examine the protein and mRNA levels of Sema4D in young and old mice post-laser photocoagulation at 0, 1, 3, and 7 days, 3 mice at each time point were used for Western blot, and 3 mice at each time point were used for quantitative real-time PCR (qPCR). Moreover, 7 days after laser injury, both young and old WT and Sema4D-KO mice were used in subsequent experiments: 6 mice in each group were used for Western blot, and 6 mice in each group were used for qPCR. Furthermore, 6 mice per group were used for optical coherence tomography (OCT) and hematoxylin and eosin (H&E), 6 mice per group were used for fundus fluorescein angiography (FFA) and immunostaining of CD31 on choroidal flat mounts, 3 mice per group were used for immunostaining of α-SMA on choroidal flat mounts, and 30 mice per group were used for other immunofluorescent staining of choroidal flat mounts. 

Histopathology

Eyeballs from mice were collected and immersed in FAS fixative solution (Servicebio, Wuhan, China). Following paraffin embedding, eyeball sections were sliced to a thickness of 3 µm and stained with H&E. Sections passing through the center of the lesions with the largest cross-sectional area were selected as representative samples for assessing the CNV size. Images were viewed and analyzed using CaseViewer (3DHISTECH, Budapest, Hungary). 

Choroidal Flat Mounts and Immunostaining

Seven days after CNV induction, the eyeballs were fixed in 4% paraformaldehyde for 1 hour. The anterior segments and neuroretinas were then excised from the eyecups. The isolated RPE-choroid-sclera complexes were blocked for 12 hours at 4°C, followed by an overnight incubation with primary antibodies (Supplementary Table S1). After several washes with PBS, the eyecups were incubated with the appropriate secondary antibodies for 2 hours. Subsequently, the eyecups were mounted onto glass slides as flat mounts with radial incisions and imaged using a LSM 880 confocal microscope. 

Optical Coherence Tomography

Seven days post-laser injury, the mice were secured onto an imaging platform, positioning their heads at an angle to allow the penetration of light vertical to the cornea. OCT images were acquired using the Phoenix Micron IV image-guided OCT system, using a live fundus image to navigate OCT scanning in the CNV lesions. The CNV size was quantified using ImageJ software. 

Fundus Fluorescein Angiography

Vascular leakage from CNV was assessed using FFA 7 days after laser injury in mice of different ages. After anesthesia and pupil dilatation, the mice received an intraperitoneal injection of 0.2 mL of 2% fluorescein sodium. Early phase (1–2 minutes) and late phase (6–8 minutes) FFA images were acquired using the Phoenix Micron IV retina imaging system. FFA images were independently evaluated by two masked retina specialists, using the following grading criteria: grade I indicates the absence of hyperfluorescence; grade II indicates the presence of hyperfluorescence without leakage; grade III indicates hyperfluorescence in the early or mid-transit images and late leakage, and grade IV indicates bright hyperfluorescence in the transit images and late leakage beyond the treated areas. 

Western Blot Analysis

Seven days after laser injury, the RPE-choroid-sclera complexes from mice were harvested and lysed in buffer supplemented with protease and phosphatase inhibitors. The isolated proteins were then subjected to SDS-PAGE and electrophoretically transferred to PVDF membranes. These membranes were blocked with 5% skim milk in TBST for 1 hour at room temperature before overnight incubation at 4°C with primary antibodies (see Supplementary Table S1). Following a thorough washing, the PVDF membranes were incubated with corresponding HRP-conjugated secondary antibodies for 2 hours. The membranes were developed with the ChemiDoc MP imaging system, and protein levels were quantified using ImageJ software. Raw images from Western blots are provided in Supplementary Fig. S1. 

Quantitative Real-Time PCR

Quantitative real-time PCR (qPCR) was conducted as described previously.^37,38 RNA was extracted using TRIzol following the manufacturer's protocol. The cDNA synthesis was performed using the cDNA First‐Strand Synthesis Kit, and qPCR analysis was carried out with SYBR Green Master Mix. Gene expression levels were normalized to β-actin and calculated using the 2^−ΔΔCq method. The primers utilized in this study are listed in Supplementary Table S2. 

Statistical Analysis

All experiments were repeated at least three times. Data are presented as the mean ± standard deviation (SD). Statistical analyses were carried out using one-way analysis of variance (ANOVA) followed by Tukey's multiple comparison test, using GraphPad Prism version 8.0. P values less than 0.05 were considered statistically significant. 

Results

Sema4D Expression was Significantly Increased in Laser-Induced CNV Lesions in Old Mice

We initially performed an experimental CNV mouse model through laser photocoagulation, and then assessed whether Sema4D expression was altered in CNV lesions in both young (3 months of age) and old (18 months of age) mice. Western blot (Fig. 1A) revealed that the protein level of Sema4D was progressively increased with time after CNV, which became more evident in aged mice (Fig. 1B). Likewise, a similar trend of Sema4D mRNA expression in mice at different ages was observed with qPCR assay (Fig. 1C). To further explore the role of Sema4D in CNV pathogenesis, we generated Sema4D-KO mice. Seven days after laser injury, choroid flat mounts were stained with α-smooth muscle actin (α-SMA) to demonstrate successful CNV formation in WT and Sema4D-KO mice of different ages (Fig. 1D). Immunofluorescent staining of choroidal flat mounts (Fig. 1E) verified successful target gene knockout of Sema4D. Moreover, the immunofluorescent results also showed that Sema4D was highly expressed in macrophages, and confirmed that Sema4D expression was upregulated in the CNV lesions of aged mice (Fig. 1F). 

Figure 1.

 

The expression and localization of Sema4D in CNV lesions in mice at different ages. (A) The protein level of Sema4D in the RPE-choroid-sclera tissue containing CNV lesions was assessed by Western blot. GAPDH served as the loading control. (B) Quantification of the Western blot is shown in the bar graph. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 3). (C) The mRNA level of Sema4D in the RPE-choroid-sclera tissue containing CNV lesions was evaluated by PCR. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 3). (D) Representative images of choroidal flat mounts stained with α-SMA (green) and DAPI (blue) are shown. Scale bar represents 100 µm. (E) Representative images of choroidal flat mounts stained with Sema4D (red), Iba1 (green), and DAPI (blue) are shown. Scale bar represents 100 µm. (F) The fluorescence intensity of Sema4D was quantified and is displayed in the bar graph. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 6). NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Knockout of Sema4D Dramatically Reduced the Size of CNV Lesions in Old Mice

To determine the role of Sema4D in CNV formation, histological and imaging examinations were used to assess CNV lesion size seven days post-laser injury in both young and old WT and Sema4D knockout mice. H&E staining of histopathology sections (Fig. 2A) showed that CNV lesions in old WT mice had greater thickness and length compared to those in young WT mice. Interestingly, Sema4D knockout led to a reduction in both thickness and length of CNV lesions across different age groups, with more marked effects in aged mice (Figs. 2B, 2C). Additionally, choroidal flat mounts stained with CD31 (Fig. 3A) also validated that knockout of Sema4D led to a reduction in the CNV area, and this decrease was more robust in older mice compared to younger mice (Fig. 3B). Concordant with these findings, additional imaging analysis using OCT (Fig. 4A) revealed that CNV lesions from aged WT mice exhibited significantly greater thickness and length compared to those from young WT mice. Furthermore, Sema4D knockout markedly reduced the CNV size in both young and aged mice, with these alterations being especially pronounced in the latter group (Figs. 4B, 4C). Overall, these results suggest a crucial role for Sema4D in CNV formation, and demonstrate that Sema4D knockout can effectively suppress laser-induced CNV, particularly in aged mice. 

Figure 2.

 

The effect of Sema4D knockout on the size of CNV lesions in mice at different ages by examining H&E staining. (A) Representative H&E staining images of young and old WT and Sema4D-KO mice are shown. INL = inner nuclear layer; ONL = outer nuclear layer; CL = choroid layer; SL = sclera layer. The red dashed line indicates the range of laser-induced CNV lesions. Right panels present magnified views of the corresponding left panels. Scale bar represents 100 µm. (B, C) Quantification of the thickness and length of CNV lesions in each group is shown in the bar graphs. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 6). NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Figure 3.

 

The effect of Sema4D knockout on CNV area in mice at different ages by examining choroidal flat mounts. (A) Representative immunofluorescent images of CNV lesions stained with CD31 (red) and DAPI (blue) in choroidal flat mounts are shown. Scale bar represents 100 µm. (B) The area of CNV lesions, as quantified based on CD31 staining, is displayed in the bar graph. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 6). NS, not significant, **P < 0.01, and ****P < 0.0001.

Figure 4.

 

The effect of Sema4D knockout on the size of CNV lesions in mice at different ages by examining OCT. (A) Representative fundus photographs (left) and OCT images (right) from the WT group and Sema4D-KO group in young and old mice are shown. GCL = ganglion cell layer; INL = inner nuclear layer; ONL = outer nuclear layer; RPE = retinal pigment epithelium. (B, C) Quantification of the thickness and length of CNV lesions in each group is shown in the bar graphs. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 6). NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Knockout of Sema4D Remarkably Decreased Vascular Leakage From CNV Lesions in Old Mice

To further investigate the impact of Sema4D on vascular leakage within CNV lesions, we induced CNV via laser photocoagulation in WT and Sema4D-KO mice of different ages and subsequently evaluated leakage with FFA (Fig. 5A), a technique commonly used in the examination of patients with neovascular AMD. The results indicated that, at 7 days post-laser, both early and late-phase FFA showed an increase in fluorescein leakage from CNV in old WT mice compared to young WT mice. Conversely, Sema4D knockout led to a significant reduction in the clinically relevant grade IV vascular leakage in both young and aged mice, with this repressive effect being more pronounced in the older group. 

Figure 5.

 

The effect of Sema4D knockout on vascular leakage of CNV lesions in mice at different ages. (A) Vascular leakage from CNV lesions in different groups was visualized by FFA during the early phase (1–2 minutes) and late phase (6–8 minutes) following intraperitoneal injection of fluorescein. The lower panels display magnified views of the areas highlighted by white dotted boxes in the upper panels. (B) The extent of CNV leakage was quantified in late-phase images, based on the following grading criteria for FFA images: grade I indicates absence of hyperfluorescence; grade II indicates presence of hyperfluorescence without leakage; grade III indicates hyperfluorescence in the early or mid-transit images and late leakage, and grade IV indicates bright hyperfluorescence in the transit images and late leakage beyond the treated areas. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 6). NS, not significant, *P < 0.05, ***P < 0.001, and ****P < 0.0001.

Sema4D Knockout Markedly Inhibited M2 Polarization of Macrophage in Old Mice

It is widely acknowledged that macrophages are the primary type of infiltrating inflammatory cells in neovascular AMD eyes, and a series of studies have proved the essential role of M2 polarization in the promotion of CNV.^37,39 To assess the impact of Sema4D knockout on macrophage polarization during CNV development in mice at different ages, PCR technology was used to quantify the expression of genes associated with M2 polarization. We observed elevated mRNA levels of M2 macrophage markers (Arg-1, Ym-1, Fizz-1, CD163, and CD206) in aged mice. However, Sema4D knockout resulted in reduced mRNA levels of these M2 markers in both age groups, with a more pronounced effect in old mice (Figs. 6A–E). Notably, there were no significant changes in the CD163 mRNA levels between WT and Sema4D-KO young mice (see Fig. 6D). After genetic verification, we performed immunofluorescent staining of choroidal flat mounts (Fig. 6F) to detect changes in the expression of CD206 (a distinct marker for M2 polarization), and to further ascertain whether Sema4D gene deletion impacts macrophage polarization in CNV. The results revealed that Sema4D knockout significantly diminished the percentage of CD206⁺ Iba1⁺ M2 macrophages in CNV lesions on choroidal flat mounts in aged mice compared with the younger group (Fig. 6G), indicating that Sema4D facilitates macrophage polarization toward the M2 type, particularly in age-related CNV. 

Figure 6.

 

The effect of Sema4D knockout on M2 macrophage polarization in mice at different ages. (A–E) The mRNA levels of M2 polarization markers (Arg-1, Ym-1, Fizz-1, CD163, and CD206) in the RPE-choroid-sclera tissue containing CNV lesions were examined by PCR. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 3). (F) The choroidal flat mounts were immunostained with CD206 (red), Iba1 (green), and DAPI (blue) in CNV lesions. Representative immunofluorescent images from each group of mice at different ages are shown. Scale bar represents 100 µm. (G) The fluorescence intensity of CD206 was quantified and is displayed in the bar graph. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 6). NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Sema4D Knockout Suppressed Macrophage Senescence in Old Mice

Age is generally recognized as an important independent risk factor for AMD.⁴⁰ Consequently, to assess the impact of Sema4D knockout on macrophage senescence, the levels of senescence markers were detected. Western blot (Fig. 7A) demonstrated the augmented levels of senescence-associated biomarkers (p16 and p21) in old WT mice, a phenomenon that was effectively mitigated by Sema4D knockout. Moreover, the young mice showed a relatively low expression of senescence markers with or without knockout of Sema4D (Figs. 7B, 7C). Immunostaining of choroidal flat mounts (Fig. 7D) demonstrated that p16 colocalized with Iba1 in old WT mice, indicating the presence of senescence in macrophages. However, the percentage of p16⁺ Iba1⁺ macrophages was reduced in CNV lesions of old Sema4D-KO mice. Furthermore, quantitative analysis of fluorescence showed lower expression of p16 in young mice with or without Sema4D knockout (Fig. 7E), which was in line with the results of Western blot. These findings suggested that Sema4D knockout inhibited senescence in macrophages of old mice. 

Figure 7.

 

The effect of Sema4D knockout on macrophage senescence in mice at different ages. (A) The protein levels of senescence markers (p16 and p21) in the RPE-choroid-sclera tissue containing CNV lesions were examined by Western blot. GAPDH served as the loading control. (B, C) Quantification of the Western blot is shown in the bar graphs. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 3). (D) The choroidal flat mounts were immunostained with p16 (red), Iba1 (green), and DAPI (blue) in CNV lesions. Representative immunofluorescent images from each group of mice at different ages are shown. Scale bar represents 100 µm. (E) The fluorescence intensity of p16 was quantified and is displayed in the bar graph. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 6). NS, not significant, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

RhoA/ROCK Signaling Pathway was Implicated in the Inhibitory Effect of Sema4D Knockout on Senescence-Associated CNV

A previous study proved that the RhoA/ROCK pathway was involved in age-induced M2 macrophage polarization and pathological CNV.³⁶ To further explore the underlying mechanisms of Sema4D, a key upstream regulator of the RhoA/ROCK signaling, in AMD, we examined the expression of related pathways in both young and aged mice, with and without the deletion of Sema4D. The results from Western blot (Fig. 8A) showed that the expression levels of RhoA, ROCK1, ROCK2, p-MYPT1, and p-MLC were higher in aged WT mice compared to young WT mice. Furthermore, Sema4D knockout led to a reduction of RhoA/ROCK signaling proteins and their downstream mediators in both age groups, which was more pronounced in aged mice. However, Sema4D deletion did not lead to a decrease in ROCK1 expression in young mice. Additionally, t-MYPT1 and t-MLC expression remained consistent across all four groups (Figs. 8B–F). Moreover, immunofluorescence staining of choroidal flat mounts (Figs. 8G, 8I) demonstrated that RhoA and p-MLC were predominantly expressed in macrophages within CNV lesions. Furthermore, the proportion of RhoA⁺ Iba1⁺ and p-MLC⁺ Iba1⁺ macrophage was significantly reduced in aged Sema4D-KO mice (Figs. 8H, 8J), aligning with the Western blot findings. Overall, the data suggested that Sema4D knockout efficiently inhibited the activation of RhoA/ROCK pathway in aged mice. 

Figure 8.

 

The effect of the RhoA/ROCK signaling pathway on the anti-CNV efficacy of Sema4D KO. (A) The protein levels of RhoA, ROCK1, ROCK2, p-MYPT1, t-MYPT1, p-MLC, and t-MLC in the RPE-choroid-sclera tissue containing CNV lesions were examined by Western blot. GAPDH served as the loading control. (B–F) Quantification of the Western blot is shown in the bar graphs. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 3). (G, I) The choroidal flat mounts were immunostained with RhoA/p-MLC (red), Iba1 (green), and DAPI (blue) in CNV lesions. Representative immunofluorescent images from each group of mice at different ages are shown. Scale bar represents 100 µm. (H, J) The fluorescence intensity of RhoA and p-MLC was quantified and is displayed in the bar graph. Data are analyzed using one-way ANOVA and presented as mean ± SD (n = 6). NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Discussion

Accumulating evidence has highlighted the importance of Sema4D and its downstream RhoA/ROCK pathway in the occurrence and progression of multiple neovascular diseases, including DR,^29,41 whereas their roles in neovascular AMD have not yet been elucidated. Herein, we reported that Sema4D mRNA and protein expression were significantly elevated over time in a laser-induced CNV model, especially in aged mice. Additionally, we revealed that genetic disruption of Sema4D alleviated pathologic neovascularization and vascular leakage in CNV lesions. Functional analyses further validated Sema4D and its mediated RhoA/ROCK pathway as pivotal factors to M2 polarization in the pathogenesis of CNV formation. A schematic diagram summarizing our results is provided in Figure 9. These data indicate that targeting Sema4D may represent a promising therapeutic approach for patients with CNV, offering a valuable reference for the clinical translation of gene editing therapies in wet AMD. 

Figure 9.

 

A schematic diagram illustrating the molecular mechanism of Sema4D and its downstream RhoA/ROCK signaling pathway in pathological CNV.

It is well established that Sema4D and its receptors are expressed in various cell types, including macrophages, and Sema4D has been suggested to be associated with macrophage polarization.⁴² A recent study found that Sema4D not only promoted endothelial cell adhesion and migration but also enhanced lactic acid release and induced M2 polarization of macrophages.⁴³ Another study demonstrated correlations between Sema4D expression and M2 macrophage counts in primary tumors, as well as M2 macrophage percentages in ascites. Additionally, peripheral blood monocytes tended to differentiate into M2-type macrophages in vitro after Sema4D stimulation.⁴⁴ In our research, we observed that Sema4D KO mice showed a reduction in the levels of M2-related surface markers, suggesting a shift in macrophage polarization toward an alternatively activated phenotype. Particularly, this effect was more obvious in aged mice. 

The RhoA/ROCK pathway has been consistently identified as a crucial regulator of macrophage polarization. Zandi et al.³⁶ discovered a master switch for macrophage polarization through ROCK signaling in the aging eye, demonstrating that selective inhibition of ROCK2 reduced M2-like macrophages and pathologic angiogenesis. Similarly, Li and colleagues⁴⁵ reported that inhibition of ROCK significantly attenuated pulmonary fibrosis by dampening M2 macrophage polarization via phosphorylation of STAT3. Consistent with these studies, our data revealed a clear increase in the expression of RhoA/ROCK signaling related molecules in old mice versus young mice, and showed that KO of Sema4D efficiently inhibited the activation of RhoA/ROCK pathway in aged mice. This indicates that the RhoA/ROCK signaling may be involved in Sema4D promoting polarization of the M2-type macrophage. 

In fact, the functional significance of Sema4D/plexinB1 signaling in regulating angiogenesis has been extensively studied. Sierra et al.⁴⁶ discovered that tumor-associated macrophages, characterized by M2-type polarization markers, were the primary producers of Sema4D within tumor stroma and that their capacity to produce Sema4D was identified as crucial for promoting tumor angiogenesis. Moreover, Chen et al.⁴⁷ established that Sema4D and VEGF exhibited synergistic effects in promoting angiogenesis in epithelial ovarian cancer. Furthermore, other studies have demonstrated that either genetic disruption of Sema4D/PlexinB1 or pharmacological inhibition of Sema4D significantly mitigated pathological retinal angiogenesis and vascular leakage in the oxygen-induced retinopathy model, as well as reduced pericyte loss in the streptozotocin-induced diabetes model.^29,41 Meanwhile, they also observed no significant difference in retinal vascular development between Sema4D-KO mice and their littermate WT mice.²⁹ Our current study revealed that Sema4D is highly expressed in macrophages within CNV lesions and plays a significant role in regulating pathological neovascularization and leakage. Additionally, targeted deletion of the Sema4D gene diminished the CNV area and leakage in mice of varying ages, with the effects being more pronounced in aged mice. These data indicate that blocking Sema4D could serve as a potential alternative anti-angiogenic therapy for age-associated eye diseases. 

Similarly, there is increasing focus on the effect of the RhoA/ROCK pathway on angiogenesis. Prior studies have established the RhoA/ROCK pathway as a crucial mediator in various angiogenic processes, including the migration, survival, and permeability of endothelial cells, thereby underscoring its fundamental role in VEGF-dependent angiogenesis.^48,49 Takata et al.⁵⁰ demonstrated that hypoxia stimulated RhoA activity and revealed that fasudil, a specific ROCK inhibitor, impeded the hypoxia-induced VEGF/VEGFR2 autocrine loop by promoting the degradation of HIF-1α. Additionally, another study reported that ROCK inhibitor not only reduced VEGF-induced proliferation and migration of human umbilical vein endothelial cells, but also inhibited laser-induced CNV and vessel leakage.⁵¹ Our previous study also proved that the RhoA/ROCK signaling was activated in CNV, and that inhibiting this pathway could attenuate CNV and vascular leakage by shifting macrophage polarization from the M2 type to the M1 type.³⁷ In this study, our findings indicated that RhoA and p-MLC were predominantly expressed in macrophages within CNV lesions. Furthermore, we observed that blocking the RhoA/ROCK signaling pathway through Sema4D knockout led to a reduction in M2 macrophages polarization, ultimately resulting in the suppression of CNV development. 

In summary, we confirmed the critical role of Sema4D and its downstream RhoA/ROCK pathway in the pathogenesis of CNV. KO of Sema4D gene markedly inhibited M2-type macrophages through the suppression of the RhoA/ROCK signaling, ultimately leading to the attenuation of senescence-associated CNV. These data indicate that targeting Sema4D could represent a promising therapy for neovascular AMD patients and potentially for other diseases where alternatively activated macrophages contribute to disease pathophysiology. 

Acknowledgments

Supported by the National Natural Science Foundation of China (82271099, 82371068), the Young Scientists Fund of the National Natural Science Foundation of China (82101133), and the Guangdong Basic and Applied Basic Research Foundation (2022A1515010355, 2022A1515110213, and 2023A1515012622). 

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