The role of microglia and complement C5/C5a in the pathogenesis of rhegmatogenous retinal detachment with choroidal detachment

Abstract

Background

Rhegmatogenous retinal detachment with choroidal detachment (RRDCD) is an uncommon and sight-threatening disorder marked by fast development and significant inflammation. This study aimed to identify cellular and molecular signatures distinguishing RRDCD from typical rhegmatogenous retinal detachment (RRD) and to investigate the roles of microglia and the complement C5/C5a pathway in disease pathogenesis.

Methods

Single-cell RNA sequencing (scRNA-seq) was employed to analyze vitreous samples from patients with RRD and RRDCD to characterize the cellular composition and molecular pathways. In vitro co-culture experiments were performed to investigate the functional impact of complement C5 on primary retinal microglia, and their subsequent effects on RF/6 A endothelial cells and ARPE-19 epithelial cells.

Results

Our findings revealed a distinct cellular landscape in RRDCD, characterized by enhanced connectivity between microglia and dendritic cells, alongside a significant upregulation of the complement C5-C5AR1 interaction. In vitro experiments indicated that treatment with complement C5 enhanced microglial metabolic activity and activation, induced apoptosis in RF/6 A endothelial cells, and led to disruption of tight junction protein ZO-1 localization in ARPE-19 epithelial cells, suggesting a role in blood-retina barrier dysfunction.

Conclusion

The findings substantiate the inflammatory hypothesis regarding the pathogenesis of RRDCD, emphasizing the critical functions of microglia and the complement C5/C5a pathway in intensifying retinal inflammation and undermining vascular integrity.

Graphical Abstract

1. Introduction

Rhegmatogenous retinal detachment with choroid detachment (RRDCD) represents a severe variant of rhegmatogenous retinal detachment (RRD), characterized by rapid progression and marked inflammatory involvement. This condition typically necessitates vitrectomy and carries a high risk of postoperative recurrence, resulting in often profound visual loss. Global epidemiological data reveal notably higher RRDCD incidence rates in China compared to other regions [1], [2], [3]. Despite its clinical significance, the underlying mechanisms of RRDCD remain poorly understood, limiting the development of targeted therapies.

The distinct feature of RRDCD lies in choroidal involvement, which differentiates it from conventional RRD and reflects more extensive ocular tissue damage. Choroidal detachment in RRDCD suggests accelerated disease progression and correlates with intense intraocular inflammation, frequently triggering proliferative vitreoretinopathy (PVR) [4], [5]. This complication substantially compromises surgical outcomes and visual recovery.

Our previous studies [6], [7], [8], [9] demonstrated elevated expression of multiple cytokines and inflammatory-related proteins in the vitreous of RRDCD patients compared to RRD cases, particularly complement system including C1r, C3, C5, C6, and C8. As a key element of innate immunity, the complement system contributes substantially to immune surveillance and inflammatory responses. Dysregulation of this system has been established in several ocular pathologies, such as age-related macular degeneration and retinitis pigmentosa.

The pathogenic mechanisms of RRDCD, particularly the specific cellular contributors and their interplay with the complement system, remain incompletely understood. As the retina's resident immune cells, microglia participate in both inflammatory responses and tissue homeostasis. These cells dynamically adapt to retinal detachment-induced environmental shifts, modulating disease progression via phagocytic activity and cytokine secretion [10], [11], [12].

Given the complexity of RRDCD and the emerging evidence implicating microglia and the complement system in its pathogenesis, elucidating the underlying cellular and molecular mechanisms remains essential. Single-cell RNA sequencing (scRNA-seq) [13] enables precise characterization of cellular heterogeneity in vitreous samples, revealing distinct cell populations and molecular pathways that could inform targeted therapeutic strategies.

We used scRNA-seq to analyze the cellular composition of vitreous samples from patients with RRD and RRDCD, emphasizing microglia and the complement C5/C5a pathway. By identifying distinct cellular and molecular signatures between RRDCD and RRD, we sought to elucidate the contributions of microglia and complement C5/C5a in disease pathogenesis. These findings may inform novel therapeutic approaches for this severe disorder.

2. Materials and methods

2.1. Patient selection

Patients diagnosed with primary RRD or RRDCD, confirmed through clinical and laboratory examinations, were recruited from January 2021 to March 2023. We excluded individuals presenting with vitreous hemorrhage, ocular trauma, glaucoma, prior intraocular surgery, ocular tumors, or other clinically significant ocular or systemic diseases. Ethical approval and informed consent were obtained from all participants in accordance with the Declaration of Helsinki

2.2. Single-Cell RNA sequencing

2.2.1. Sample collection and preparation

Vitreous samples were collected during 25-gauge pars plana vitrectomy by a single experienced surgeon. Approximately 3 mL of vitreous fluid was aspirated near the vitrectomy entry site from each patient and immediately placed on ice for processing.

2.2.2. Library preparation and sequencing

Samples were transferred to 10 mL centrifuge tubes, diluted with an equal volume of pre-chilled 1 × PBS, and centrifuged at 350 g for 5 min at 4°C. The resulting cell pellet was resuspended in a precise volume (50 μL to 1 mL) of pre-chilled 0.01 mM EDTA and 0.04 % BSA in 1 × PBS. Cell concentration was measured using a cell counter, adjusted to 700–1200 cells/μL, and filtered through a 40 μm cell strainer to obtain a single-cell suspension. Single-cell gene expression libraries were prepared following the 10 × Genomics Chromium Single Cell 3’ protocol, which partitions cells into nanoliter-scale Gel Beads-In-Emulsions (GEMs) for cell lysis, reverse transcription, and cDNA barcoding. After PCR amplification to generate sequencing libraries, quality control was performed before paired-end sequencing on Illumina platforms.

2.2.3. Data processing and quality control

Raw sequencing data were aligned to the GRCh38.108 reference genome, with barcode assignment and UMI counting performed using Cell Ranger (10 × Genomics, version 7.1.0). Subsequent analysis was conducted using Seurat (R package, version 5.2.1). Cells with fewer than 300 or more than 7000 detected genes, UMI counts exceeding 50,000, or mitochondrial transcripts > 20 % of total counts were excluded. Ambient RNA contamination was removed using DecontX (contamination fraction threshold: 0.3).

2.2.4. Dimensionality reduction and clustering

Principal Component Analysis (PCA) reduced the high-dimensional scRNA-seq data to capture major sources of variation. Uniform Manifold Approximation and Projection (UMAP) further projected the data into two dimensions for clearer visualization. Louvain clustering groups cells with analogous expression profiles based on PCA-reduced data.

2.2.5. Cell-Cell communication analysis

CellPhoneDB identified ligand–receptor pairs and their cell-type-specific expression patterns, revealing active signaling pathways within the cellular microenvironment. This analysis elucidated potential intercellular communication networks.

2.3. Primary mouse retinal microglia isolation and identification

2.3.1. Isolation procedure

Primary retinal microglia were isolated from 1 to 3 day-old neonatal mice. Culture plates were precoated with poly-D-lysine (50 µg/mL) in DMEM/F12 the day prior to isolation. Mice were euthanized via intraperitoneal injection of 0.5 mL of 10 % chloral hydrate. Under aseptic conditions, eyes were carefully excised using sterile scissors and forceps and immediately placed in phosphate-buffered saline (PBS). A central window in the cornea was created using a syringe, and ophthalmic scissors were used to incise along the limbus. The choroid was gently detached, and structures such as the lens, vitreous body, and sclera were removed, retaining only the retina and choroid. The isolated tissues were transferred to a centrifuge tube containing trypsin digestion solution and mechanically dissociated into fragments. The mixture was incubated at 37°C for 25 min, after which DMEM/F12 supplemented with 10 % fetal bovine serum (FBS) was added to neutralize the trypsin. The cell suspension was passed through a 75 µm filter to remove undigested tissue and centrifuged. The poly-D-coated plates were washed three times with PBS to eliminate residual matrix. The cell pellet was resuspended in DMEM/F12 containing 10 % FBS and 1 % penicillin-streptomycin and seeded into culture flasks. Cells were incubated overnight at 37°C with 5 % CO₂ to allow attachment. The following day, cell morphology, density, adherence, and growth were assessed under a microscope, and the medium was refreshed as necessary.

2.3.2. Identification of microglia

Isolated microglia were identified using immunocytochemical staining for ionized calcium-binding adapter molecule 1 (IBA-1), a specific marker for microglia in the central nervous system. Cells were fixed with 4 % paraformaldehyde, permeabilized, and incubated with an anti-IBA-1 primary antibody, followed by a fluorescently labeled secondary antibody. Nuclei were counterstained with DAPI.

2.4. Complement C5 treatment and Co-culture systems

2.4.1. Treatment protocol

Microglia were divided into experimental and control groups. Experimental groups were treated with complement C5 (HY-P7695; MedChemExpress) at concentrations of 0.5 µg/mL, 1.0 µg/mL, and 2.5 µg/mL, while the control group received no complement C5.

2.4.2. Co-culture with RF/6 A endothelial cells

For co-culture with RF/6 A endothelial cells (Ocutech Biotechnology Co., Ltd, China, Wuxi), microglia (Jiangsu Institute of Nuclear Medicine) were seeded in the upper chamber of Transwell inserts and allowed to adhere for 2 h. Non-adherent cells were gently removed with PBS. Different concentrations of C5 were added to the DMEM medium in the upper chamber, and RF/6 A cells were seeded in the lower chamber. The co-culture system was incubated at 37°C in 5 % CO₂ for 24 h. Cell viability of RF/6 A cells was assessed using the MTT assay, and apoptosis was evaluated using the TUNEL assay. Based on these results, a C5 concentration of 1.0 µg/mL was selected for further experiments.

2.4.3. Co-culture with ARPE-19 epithelial cells

For co-culture with ARPE-19 epithelial cells (China Center for Type Culture Collection), microglia treated with 1.0 µg/mL C5 were seeded in the upper chamber of Transwell inserts and allowed to adhere for 2 h. Non-adherent cells were removed with PBS, and ARPE-19 cells were seeded in the lower chamber. The co-culture was maintained for 24 h, after which the expression of the tight junction protein ZO-1 in ARPE-19 cells was examined using immunofluorescence staining.

2.5. Cellular assays

2.5.1. Cell viability assay

Microglia viability was quantified using the MTT assay. Cells were seeded in 96-well plates at 10,000 cells per well, with edge wells filled with sterile PBS to minimize evaporation. Following overnight incubation at 37°C under 5 % CO₂ to establish a confluent monolayer, cells were exposed to C5 at concentrations of 0 or 1.0 µg/mL (100 µl per well, six replicates per condition). After 24 h, morphology was examined by inverted microscopy. Cells were then incubated with 0.5 % MTT solution (20 µl/well) for 2 h at 37°C. The medium was aspirated, and formazan crystals were solubilized with 150 µl DMSO. Absorbance at 490 nm was measured to assess viability.

2.5.2. Immunofluorescence detection

Microglia cultured in 24-well plates were treated with 1.0 µg/mL C5 or left untreated as controls. After 24 h, cells were PBS-washed, fixed with 4 % paraformaldehyde for 15 min, and permeabilized before blocking for 1 h. Primary antibodies were applied overnight at 4°C, followed by fluorescent secondary antibodies (Cell Signaling Technology, 13663, 91882) (1 h at room temperature). Nuclei were counterstained with DAPI after three washes. Fluorescence microscopy captured staining patterns.

2.5.3. Western blot analysis

Cells were washed with cold PBS, trypsinized, and pelleted by centrifugation. Protein was extracted using RIPA buffer supplemented with PMSF, homogenized, and incubated on ice (30 min). Lysates were centrifuged (12,000 rpm, 15 min, 4°C). and supernatants were stored at −80°C. Denatured proteins (5–10 min, 100°C) were separated by SDS-PAGE (80 V for 20 min, then 100 V for 60 min) and transferred to PVDF membranes (400 mA, 40 min). Membranes were blocked with 5 % non-fat milk in TBST, probed with primary antibodies (overnight, 4°C), and incubated with fluorescent secondaries. Chemiluminescence imaging detected protein bands.

2.5.4. Enzyme-Linked Immunosorbent Assay (ELISA)

Microglia culture supernatants were harvested after 24-hours C5 treatment. IL-1β and TNF-αlevels were quantified using ELISA kits (Ocutech Biotechnology Co., Ltd, China, Wuxi) per manufacturer protocols. Samples (100 µl) were added to antibody-coated plates, incubated and washed before adding detection reagents. Absorbance at 450 nm was measured, with concentrations derived from standard curves.

2.5.5. TUNEL assay for apoptosis detection

Primary mouse retinal microglia were cultured in Transwell inserts and exposed to 1.0 µg/mL complement C5 in complete medium, with RF/6 A endothelial cells (1 × 10⁷ cells/well) plated in the lower chamber. Following 24 h of co-culture, RF/6 A endothelial cell apoptosis was evaluated via TUNEL assay. Cells fixed with 4 % neutral paraformaldehyde (10 min, room temperature) were treated with the TUNEL reaction mix, applied to glass slides, and dried. After PBS washes, TUNEL-positive nuclei were imaged and counted by fluorescence microscopy. Quantification was performed in Fiji (ImageJ) by isolating the red channel, converting to 8-bit, subtracting background (rolling ball radius = 50 px), and auto-thresholding with the Otsu method to generate a binary mask. Watershed segmentation was applied to separate adjacent signals and the Analyze Particles tool (particle size ≥ 120 px², circularity 0.3–1.0) was used to determine the total TUNEL-positive area per field. Results are expressed as the percentage of TUNEL-positive area relative to the total image area.

2.6. Statistical analysis and visualization

All experiments were conducted in triplicate with at least three independent replicates. Results are expressed as mean ± standard deviation (SD). We assessed statistical significance using appropriate tests: one-way analysis of variance (ANOVA) with Tukey’s post hoc test multi-group comparisons and unpaired Student’s t-tests for pairwise analyses. Statistical significance was defined as p < 0.05. Analyses were conducted using R software (version 4.4.2) or GraphPad Prism (version 10.0). UMAP projections, feature plots, and dot plots were visualized using scRNAtoolVis, ggSCvis, scCustomize, and the SCP platform for scRNA-seq analysis.

3. Results

3.1. Patient demographics and clinical characteristics

Single-cell transcriptome analysis was performed on samples from four patients with RRD and four patients with RRDCD, with demographics and clinical characteristics detailed in Table 1.

Table 1.

Baseline data for the patients with RRD and RRDCD in the study.

Group	Sex	Age	Course (days)	Visual acuity	IOP (mmHg)	Severity
RRD1	Female	52	6	OD0.4 OS0.2	OD13.5 OS14.4	OS PVRB, RD in two quadrants
RRD2	Male	64	21	OD0.06 OS0.5	OD10.3 OS10.9	OD PVRB, RD in two quadrants
RRD3	Male	51	5	OD0.4 OS0.25	OD12.3 OS8.3	OS PVRB, RD in two quadrants
RRD4	Male	62	5	OD0.8 OS0.4	OD9 OS6.6	OS PVRB, RD in one quadrant
RRCD1	Male	44	10	OD0.3 OS0.4	OD7 OS13.2	OD PVRB, RD+CD in one quadrant
RRCD2	Male	48	7	OD CF OS0.5	OD6.8 OS13.3	OD PVRB, RD in two quadrants, ciliary choroid detachment in 360 degrees
RRCD3	Female	49	5	OD HM OS NLP	OD5.2 OS26.1	OD PVRB, RD+CD in two quadrants
RRCD4	Female	63	7	OD hand move OS1.0	OD6.8 OS16.4	OD PVRB, RD+CD in two quadrants

CD: Choroidal Detachment; IOP: Intraocular Pressure; NLP: No Light Perception; OD: Oculus Dexter; OS: Oculus Sinister; PVRB: Proliferative Vitreoretinopathy, Grade B; RD: Retinal Detachment.

B-scan ultrasound and ultrasound biomicroscopy (UBM) findings are presented in Fig. 1. Retinal detachment was observed in RRD patients, while the ciliary body and choroid remained intact on UBM. In RRDCD, B-scan ultrasound detected retinal detachment, often accompanied by choroidal detachment, whereas UBM confirmed concurrent detachment of the ciliary body and choroid.

Fig. 1.

Ultrasound findings in retinal detachment patients. (A) B-scan images of patients with RRD, indicating retinal separation without ciliary body or choroid detachment. (B) Images of patients with RRDCD, depicting both retinal and choroidal detachment along with ciliary body involvement.

3.2. Clustering and Cellular Composition

scRNA-seq analysis revealed seven distinct cell populations in the retinal microenvironment of RRD and RRDCD patients. Initial classification separated immune (PTPRC⁺) from non-immune (PTPRC⁻) populations using PTPRC (CD45), with AIF1 further identifying myeloid lineage cells (Fig. 2 A). Subsequent UMAP (Fig. 2 C) clustering resolved specific cell types: microglia, dendritic cells, T cells, γδT cells, photoreceptors, Müller cells, and proliferating cells, with marker gene expression patterns illustrated in Figs. 2B and 2D.

Fig. 2.

scRNA-seq analysis of cellular heterogeneity in RRD and RRDCD patients. (A) Feature plots showing the expression of PTPRC (CD45) and AIF1, along with immune cell classification distinguishing immune and non-immune populations. (B) Marker dot plot displaying expression levels of key cell type markers, with color intensity indicating expression level and size reflecting relative abundance. (C) Dimensionality reduction plot (DimPlot) showing annotated cell clusters including microglia, dendritic cells (DCs), photoreceptors, T cells, Müller cells, proliferating cells, and γδT cells. (D) Feature plots of all cell markers used for cell type annotation, illustrating spatial expression patterns across the UMAP embedding. DCs: dendritic cells.

One immune cell cluster was identified as microglia due to its robust co-expression of TMEM119 (>90 % microglia-specific) and P2RY12 (essential for microglial homeostasis), alongside the macrophage marker CD68 and M2-associated CD163 (Figs. 2B, 2D), indicative of an activated immunoregulatory phenotype. While TMEM119 and P2RY12 expression appeared moderately reduced in this activated state, their consistent presence across the microglia cluster in feature plots confirmed their widespread distribution. The persistent detection of these markers despite their typical downregulation during activation, particularly when observed with CD68/CD163 co-expression, further validated the microglial classification[14], [15].

Other clusters demonstrated distinct transcriptional profiles (Figs. 2B, 2D): photoreceptors expressed RHO and GNAT1; T cells exhibited CD3D/E and CD2 alongside CD8A, GZMM, and IL7R (suggesting an effector-memory CD8⁺ phenotype), complemented by CXCR3 and CD27; γδT cells co-expressed TRDC, NKG7, and KLRD1 (revealing a unique TCR-δ⁺/NK receptor⁺ hybrid phenotype distinct from conventional NK or αβ T cells); dendritic cells showed CD1C and FCER1A; Müller cells contained GLUL and CRABP1; and proliferating cells were marked by MKI67 and TOP2A. These molecular signatures enabled clear discrimination of each cell type within the retinal microenvironment.

3.3. Intercellular communication networks in RRD and RRDCD

To examine differences in retinal cellular interactions between RRD and RRDCD groups, we employed CellPhoneDB to analyze their molecular communication networks. The heatmaps in Fig. 3 A (RRD) and 3B (RRDCD) reveal distinct communication patterns based on the total number of predicted interactions. Cellular interactions were more densely distributed in the RRDCD group (Fig. 3B) than in the RRD group (Fig. 3 A). Notably, microglia-dendritic cell connectivity was substantially stronger in RRDCD, with 170 interactions compared to 73 in RRD, as reflected in the heatmap intensities. The chord diagrams (Fig. 3 A and 3B) further illustrate the specific ligand-receptor pairs underlying these networks. These findings demonstrate a pronounced shift in intercellular communication in RRDCD, particularly through enhanced microglia-dendritic cell crosstalk.

Fig. 3.

(A) The heatmap (left) displays the absolute counts of statistically significant intercellular interactions among cell types in RRD patients. The accompanying chord diagram (right) highlights key ligand-receptor pairs specific to this cohort. (B) Similarly, the heatmap (left) quantifies the absolute counts of statistically significant intercellular interactions in RRDCD cases, with the accompanying chord diagram (right) visualizing prominent ligand-receptor relationships. For both heatmaps, absolute counts represent the number of statistically significant interactions (p < 0.05). The chord diagrams visualize individual ligand-receptor connections. From the statistically significant pairs (p < 0.05), the top 40 pairs, sorted by their mean interaction score, are selected and displayed for visualization.

3.4. Differential expression of complement C5 components and altered intercellular communication in retinal detachment

We investigated microglial complement signaling by analyzing baseline ligand and receptor expression using UMAP and violin plots (Fig. 4 A). Microglial clusters exhibited distinct expression patterns of complement components, with C5AR1 showing relatively high abundance in both RRD and RRDCD, while C5 and C5AR2 were less prevalent. Significant differences in microglial expression of C5, C5AR1, and C5AR2 were observed between RRD and RRDCD (p < 0.0001 for each). CellPhoneDB analysis of the bubble plot (Fig. 4B) revealed that microglia-to-microglia C5-C5AR1 and C5-C5AR2 interactions in RRDCD exhibited lower p-values and higher mean interaction scores than in RRD, suggesting stronger signaling potential.

Fig. 4.

scRNA-seq Analysis of Complement C5 Components and Intercellular Communication in Retinal Detachment. (A) UMAP feature plots (left) show C5, C5AR1, and C5AR2 expression across retinal cell clusters in RRD and RRDCD, with color intensity indicating expression levels. Violin plots (right) compare cell type-specific expression distributions between conditions. (B) CellPhoneDB-derived dot plot displays ligand-receptor interactions, with bubble size reflecting interaction significance (-log₁₀ p-value) and color denoting mean expression. Comparisons include C5-C5AR1, C3-C5AR2, TNF-TNFRSF1A/B, and VEGF (VEGFA-FLT1, VEGFB-NRP1) pathways between RRD and RRDCD.

TNF and VEGF signaling displayed condition-specific variations. In RRD, TNF interactions were predominantly confined to TNF-TNFRSF1A, whereas RRDCD involved broader engagement with microglia, dendritic cells (DCs), and T cells, accompanied by intensified TNF signaling. VEGF interactions also differed, with VEGFB-NRP1 and VEGFA-FLT1 extending to additional cell types in RRDCD compared to their narrower distribution in RRD.

3.5. Complement C5-treated microglia reduce endothelial cell viability and promote microglial metabolic activity

To evaluate how complement C5 influences microglial function and its downstream effects on ocular cells, primary mouse retinal microglia were exposed to C5 at concentrations of 0, 0.5, 1.0, or 2.5 µg/mL for 24 h. These microglia were then co-cultured with RF/6 A choroidal vascular endothelial cells, and endothelial viability was measured via MTT assay. Fig. 5A shows that all tested C5 concentrations (0.5–2.5 µg/mL) significantly reduced RF/6 A viability compared to untreated microglia. Although 2.5 µg/mL C5 elicited the strongest effect, no significant differences were observed between 0.5 and 1.0 µg/mL or between 1.0 and 2.5 µg/mL. For subsequent experiments, 1.0 µg/mL C5 was selected as the standard treatment concentration.

Fig. 5.

Characterization of complement C5 effects on cellular responses and molecular markers. (A) MTT assay assessing the viability/metabolic activity of RF/6 A choroidal vascular endothelial cells after 24-hour co-culture with primary mouse retinal microglia that were pre-treated with varying concentrations of complement C5 (0, 0.5, 1.0, 2.5 µg/mL)， with results expressed as ODsample/ODblank ratios. (B) Apoptotic effects of C5-treated (1 μg/mL, 24 h) microglia on RF/6 A choroidal vascular endothelial cells, as evidenced by TUNEL staining (red) and quantification of TUNEL-positive area. (C) C5 exposure (1 μg/mL, 24 h) altered microglial cell viability (MTT assay, Student's t-test) and morphology (brightfield images). (D) ZO-1 immunofluorescence (red) in ARPE-19 cells co-cultured with C5-treated retinal microglia (1 μg/mL, 24 h) showed tight junction integrity (nuclei: DAPI, blue). (E) CD86 expression (red) increased in retinal primary mouse microglia following C5 treatment (1 μg/mL, 24 h). (F) Western blot confirmed elevated CD86 protein levels in C5-treated microglia (1 μg/mL, 24 h). (G) ELISA detected significantly higher TNF-α and IL-1β secretion in C5-treated microglia (1 μg/mL, 24 h). Data represent mean ± SD of ≥ 3 independent experiments.

Direct exposure of primary microglia to 1.0 µg/mL C5 for 24 h increased their metabolic activity, as evidenced by MTT assay (Fig. 5C, left). Brightfield microscopy revealed higher cell density and morphological changes in C5-treated microglia (Fig. 5C, right), suggesting enhanced growth and activation.

3.6. C5-activated microglia trigger endothelial apoptosis and disrupt RPE tight junctions

Given the reduced viability of RF/6 A cells, we assessed whether apoptosis contributed to this effect using the TUNEL assay. Microglia treated with 1.0 µg/mL C5 for 24 h were co-cultured with RF/6 A cells, resulting in a significant increase in TUNEL-positive area compared to controls (Fig. 5B), confirming C5-activated microglia induce endothelial apoptosis. To examine retinal barrier integrity, ARPE-19 cells were co-cultured with C5-treated microglia. Immunofluorescence staining of ZO-1 showed disrupted localization and reduced intensity in ARPE-19 cells exposed to activated microglia (Fig. 5D), indicating compromised tight junction organization. These experiments used confluent ARPE-19 monolayers to ensure proper junction formation.

3.7. Complement C5 promotes a pro-inflammatory microglial phenotype

To assess C5-induced inflammatory responses, we analyzed the expression of CD86, a marker of M1 polarization. Immunofluorescence demonstrated elevated CD86 levels in microglia treated with 1.0 µg/mL C5 (Fig. 5E), corroborated by Western blot (Fig. 5F). ELISA assays further revealed increased secretion of TNF-α and IL-1β in C5-treated microglia(Fig. 5G). Together, these data show that C5 drives a pro-inflammatory microglial state characterized by CD86 upregulation and heightened cytokine production.

4. Discussion

Our scRNA-seq analysis revealed microglia as the predominant immune cell type in vitreous samples from both RRD and RRDCD patients. This finding highlights the significant inflammatory component in both conditions, with microglia likely playing a central role. Furthermore, the vitreous sampling technique itself (vitrectomy) may favor the collection of free-floating immune cells over fragile, degenerating retinal cells. Consequently, the cellular landscape observed in the vitreous largely represents active immune populations, particularly microglia, rather than all resident retinal cells.

Delving deeper into the vitreous microenvironment, our cell-cell communication analysis demonstrated a markedly different and more extensive intercellular communication network in RRDCD compared to RRD. Specifically, microglia-dendritic cell connectivity was substantially stronger in the RRDCD group, with a significantly higher number of predicted interactions. Crucially, among these amplified interactions, the complement C5-C5AR1 pathway emerged as a prominent feature, significantly enhanced in both microglia and dendritic cells of RRDCD patients (Fig. 4B). This finding suggests that the complement C5/C5a signaling axis plays a critical role in modulating inflammatory responses during the progression of RRDCD. Additionally, interactions involving VEGFA with NRP1 and FTL1 showed variability, reflecting the complexity of signaling pathways underlying the disease's inflammatory microenvironment.

Our in vitro experiments revealed how complement C5 influences microglial activity and subsequently affects retinal vascular and epithelial cells. Exposure of primary mouse retinal microglia to complement C5 elevated their metabolic activity and promoted activation (Fig. 5C). When these activated microglia were co-cultured with RF/6 A endothelial cells, endothelial viability declined dose-dependently while apoptosis rates rose significantly (Figs. 5A, 5B), confirming microglial activation damages vascular integrity. Co-culture experiments with ARPE-19 retinal pigment epithelial cells showed complement C5-treated microglia disrupted ZO-1 tight junction protein distribution, with staining intensity diminishing and localization shifting from membrane junctions to perinuclear regions (Fig. 5D). This ZO-1 disorganization indicates blood-retina barrier impairment, a critical structure for retinal homeostasis. Tight junction degradation likely enables choroidal vessel leakage, explaining the choroidal detachment characteristic of RRDCD. Similar complement-mediated vascular barrier breakdown occurs in other inflammatory conditions, including blood-brain barrier dysfunction[16].

Our findings indicate that complement C5 markedly increases the expression of pro-inflammatory cytokines TNF-α and IL-1β in microglia. The cytokine surge highlights the contribution of complement C5/C5a to amplifying retinal inflammation. The resulting inflammatory cascade not only worsens retinal damage but also creates reinforces complement system activation, creating a self-perpetuating cycle of tissue injury. Although we identified increased expression of critical inflammatory mediators, further mechanistic studies should investigate the precise intracellular signaling pathways triggered by C5/C5a in microglia, particularly NF-κB or MAPK cascades. Such investigations could employ transcriptomic approaches like gene set enrichment analysis of scRNA-seq data or targeted biochemical assays to elucidate these pathways.

Previous studies has shown that complement signaling pathway, particularly C3-C3AR and C5-C5AR, critically modulate inflammatory processes, and are implicated in various inflammatory diseases [17], [18]. We detected elevated levels of complement C5-related receptor-ligand pairs in the vitreous of both RRD and RRDCD patients, with greater abundance in RRDCD, consistent with its more severe inflammatory profile.

Regarding the pathogenesis of RRDCD, two main hypotheses have been proposed: one emphasizes a hypotonous origin [19], while the other stresses an inflammatory origin [2]. Our scRNA-seq analysis demonstrated that microglia constitute the primary cell population in RRD and RRDCD vitreous samples. Cell-cell communication analysis revealed that ligand-receptor interactions in both conditions predominantly involve the complement C5/C5a, with microglia serving as the main interactors. These results suggests RRD already exhibits baseline inflammation, which becomes exacerbated in RRDCD. In vitro experiments confirmed that complement C5/C5a activation in microglia compromises the blood-retina barrier, supporting the inflammatory hypothesis of RRDCD progression.

Retinal microglia homeostasis depends on complex interactions with the retinal microenvironment. Fate mapping studies confirm that under physiological conditions, retinal microglia formed a self-renewing population with tightly controlled numbers, distribution, and functional states [20], [21], [22]. Their homeostasis is regulated by neuronal and retinal cell-derived signals, including cytokines, complement factors, and damage-associated molecules [23]. For example, RPE-secreted TGFβ stimulates microglial IL-10 production, which suppresses antigen-presenting molecules like MHC-II, CD80, and CD86 to maintain an anti-inflammatory phenotype [24]. Physical interactions mediated by CD200-CD200R and CX3CL1-CX3CR1 further restrain pro-inflammatory activation while preserving microglial surveillance functions [25], [26], [27], [28], [29].

The relationship between microglia and the complement system is intricate. Microglia express complement receptors, ligands, and activation factors, and are involved in processes such as synaptic pruning through the complement receptor 3 (CR3) complex [30], [31]. In addition to being effectors of complement-mediated phagocytosis, microglia themselves are significant sources of complement components in the retina. Our scRNA-seq data corroborate these findings, showing a close association between microglia and complement C5/C5a in RRDCD. In vitro experiments demonstrated that complement C5/C5a interaction with microglia enhances their pro-inflammatory state, leading to blood–retina barrier disruption. Specifically, the observed impairment of RPE tight junctions provides a mechanistic basis for the choroidal leakage observed in RRDCD.

In summary, our study provides compelling evidence for an inflammatory-driven model of RRDCD pathogenesis, in which the complement C5/C5a-microglia axis is a central pathogenic driver. This mechanistic insight pinpoints this axis as a potential therapeutic target. Specifically, targeting the complement pathway—either by blocking the C5a receptor with antagonists like avacopan to directly prevent microglial activation, or by inhibiting C5 cleavage with therapies like eculizumab to neutralize both the pro-inflammatory C5a and the lytic Membrane Attack Complex—presents a direct strategy to interrupt the inflammatory cascade, leveraging approaches validated in other systemic inflammatory disorders. Furthermore, modulating the effector cell itself through approaches like CSF1R inhibition could offer a powerful adjunct therapy. This could dampen the overall inflammatory potential of microglia by limiting their proliferation and survival, potentially working in synergy with complement blockade to restore immune homeostasis. Such targeted interventions warrant investigation as they hold the promise of protecting the blood-retina barrier, mitigating severe inflammation, and ultimately improving surgical outcomes in RRDCD.

The study has several notable limitations. First, the sample size, while meaningful for investigating rare conditions like RRDCD, was constrained by practical challenges including the COVID-19 pandemic; larger cohorts would enhance statistical robustness and generalizability. Second, our in vitro experiments employed primary neonatal mouse microglia, which differ functional and phenotypically from adult microglia and lack the complex interactions of the native retinal microenvironment. Although these experiments revealed C5-mediated effects, future work should validate these findings through direct functional assays, including TEER measurements for barrier integrity and Ki67 staining for proliferation, alongside C5a receptor antagonist rescue experiments to establish mechanistic specificity. Additionally, since our conclusions primarily derive from transcriptomic data, immunohistochemical confirmation of microglial C5AR1 expression in RRDCD patient tissues would strengthen the clinical relevance of these observations. Third, while the in vitro results suggest a potential mechanism, confirming the causal role of C5-activated microglia in blood-retina barrier disruption and choroidal detachment requires validation through physiological in vivo models. Future studies should incorporate larger clinical cohorts, employ adult primary microglia or iPSC-derived human microglia for in vitro work, and establish advanced in vivo systems to better define how microglia and complement components drive RRDCD pathogenesis. These improvements will facilitate the development of targeted therapies to reduce inflammation, protect the blood–retina barrier, and enhance surgical outcomes in this severe disorder.

CRediT authorship contribution statement

Zhifeng Wu: Validation, Supervision, Funding acquisition, Conceptualization. xu huiyan: Writing – review & editing, Data curation. chen xuan: Investigation, Data curation. wang qiuhong: Writing – review & editing, Writing – original draft, Methodology. xu shasha: Methodology. huang qingyu: Investigation.

Ethical Approval

The study was approved by the Ethics Committee of Wuxi Second Hospital (Ethics Approval No: 2019Y-30, 2022Y-114).

Data Acquisition

Data used in this study can be accessed from the National Genomics Data Center (NGDC) at the following link: https://ngdc.cncb.ac.cn/search/, with dataset identifier PRJCA036783.

Synopsis/Precis

This study elucidates the distinct cellular landscape of rhegmatogenous retinal detachment with choroidal detachment (RRDCD), highlighting enhanced microglial activation and upregulated C5/C5a complement signaling. Our findings demonstrate that C5/C5a-driven microglial activity contributes to blood-retina barrier disruption, offering novel therapeutic targets for RRDCD.

Funding

This work was supported by the High Level Talent Project of the Wuxi TaiHu Committee (Wuxi Medical 2020–71).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

Figure S1: Quality control metrics for scRNA-seq data. Figure S2: UMAP plots with single-sample cell type annotations for RRD and RRDCD patients. Table S1: All statistically significant ligand-receptor interactions in RRD patients. Table S2: All statistically significant ligand-receptor interactions in RRDCD patients