Ultrastructural characteristics of the trabecular meshwork in primary open-angle glaucoma versus chronic angle-closure glaucoma: a comparative study

Abstract

Background

The trabecular meshwork (TM), the principal site of aqueous humor outflow resistance, plays a pivotal role in the regulation of intraocular pressure (IOP), the key modifiable risk factor for glaucoma. This study aimed to compare the ultrastructural characteristics of the TM in primary open-angle glaucoma (POAG) and chronic angle-closure glaucoma (CACG) using transmission electron microscopy (TEM).

Methods

TM specimens were obtained via Kahook Dual Blade goniotomy from 11 POAG and 9 CACG patients and processed for TEM analysis. We evaluated cellular morphology, extracellular matrix (ECM) organization, and collagen fibril architecture (using Feret’s diameter). A novel Organelle Severity Index (OSI) was introduced to semi-quantitatively evaluate the cumulative burden of subcellular stress, and clinico-pathological correlations were analyzed.

Results

Both POAG and CACG samples exhibited a markedly sparse TM cellularity. Residual cells demonstrated profound subcellular stress, including swollen mitochondria with loss of cristae and accumulation of secondary lysosomes. OSI scores were comparably high in POAG and CACG (1.69 vs. 1.63, P = 0.866), with no significant linear correlations to preoperative macroscopic clinical metrics (all P > 0.05). ECM alterations in both groups included trabecular beam fusion and TM cells abnormally migrating across fractured beams. Quantitatively, collagen fibrils were significantly larger in CACG compared to POAG (60.27 ± 12.97 nm vs. 43.61 ± 8.62 nm, P = 0.018).

Conclusion

Surgical-stage POAG and CACG share a “final common pathway” of severe TM cellular reduction and structural disruption. However, their distinct collagen fibril architectures suggest divergent ECM remodeling pathways that may contribute to subtype-specific outflow resistance and disease mechanisms.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12886-026-05112-0.

Background

The trabecular meshwork (TM) is the primary site of aqueous humor outflow resistance and plays a pivotal role in the regulation of intraocular pressure (IOP), the key modifiable risk factor for glaucoma [1, 2]. Located at the iridocorneal angle, the TM is a highly specialized, sieve-like structure composed of overlapping connective beams lined by trabecular cells. Anatomically, its multilayered structure includes the uveal meshwork, corneoscleral meshwork, and juxtacanalicular tissue (JCT), with the JCT and Schlemm’s canal inner wall contributing most to outflow resistance. The extracellular matrix (ECM) of the TM contains collagens (types I, III, IV, V, VI), elastin, glycosaminoglycans, and proteoglycans, which are continually remodeled in response to mechanical forces, IOP changes, and biochemical signals [3–6]. This dynamic remodeling is essential for maintaining normal aqueous humor drainage and IOP homeostasis [7].

In glaucoma, dysregulation of the TM impairs aqueous humor drainage and leads to elevated IOP, which in turn drives progressive optic nerve damage [8, 9]. In primary open-angle glaucoma (POAG), the pathological features include significant loss of TM cells, abnormal ECM accumulation, Schlemm’s canal endothelial dysfunction, and JCT obstruction, all of which contribute to increased outflow resistance [10–12]. In primary angle-closure glaucoma (PACG), a appositional or synechial closure of the angle—often due to peripheral anterior synechiae and prolonged iris-TM contact—results in structural injury to Schlemm’s canal [13], along with TM cell edema, trabecular beam fusion, and inflammatory cell infiltration, often leading to irreversible outflow blockage despite surgical intervention [11].

Although both POAG and chronic angle-closure glaucoma (CACG) involve TM dysfunction and share certain structural alterations, their precise ultrastructural differences remain poorly defined. Most prior studies have relied on cadaveric eyes or animal models, restricting direct insight into human disease. In this study, we employed Kahook Dual Blade (KDB, New World Medical, Inc, CA, USA) goniotomy, which offers a direct, minimally invasive, and efficient method for obtaining intact human TM specimens, to obtain samples from both POAG and CACG patients [14]. The objective was to investigate the ultrastructural characteristics of the TM in these two glaucoma subtypes and to compare the morphological differences between them.

Methods and materials

Study design

This was a non-interventional, comparative case series designed to analyze the ultrastructural characteristics of trabecular meshwork (TM) specimens using TEM. A total of 20 consecutive surgical specimens were analyzed, including 11 from patients with POAG and 9 from patients with CACG. This study involved analysis of TM tissues excised during glaucoma surgery and did not constitute a clinical trial; therefore, clinical trial number was not applicable. All patients provided written informed consent prior to participation. The study protocol was reviewed and approved by the Ethics Committee of Peking University People’s Hospital (Approval No. 2024PHB122-001).

Participants

Between August 2024 and June 2025, a total of 20 eyes from 20 POAG patients underwent KDB goniotomy at Peking University People’s Hospital. Among these, 11 eyes yielded complete TM specimens suitable for ultrastructural measurement and analysis. During the same period, 18 eyes from 18 CACG patients underwent KDB goniotomy, of which 9 eyes provided complete TM specimens for analysis. All participants received a comprehensive ocular examination, including best corrected visual acuity (BCVA), IOP measurement using Goldmann applanation tonometry (Haag-Streit, Koniz, Switzerland), detailed slit-lamp biomicroscopy, and stereoscopic evaluation of the optic disc utilizing a 90-diopter lens (Volk Optical, Inc., Mentor, OH, USA). When indicated, a glaucoma specialist (HJW) performed gonioscopy in a dimly lit room using a Goldmann-style four-mirror gonioscopy lens (Model G-4, Volk Optical, Inc., Mentor, OH) at 16× magnification, with and without indentation, to evaluate the presence of peripheral anterior synechiae. Optical coherence tomography (Spectralis HRA + OCT, Heidelberg Engineering GmbH, Heidelberg, Germany) was employed to detect retinal nerve fiber layer defects, and a visual field test (Humphrey Field Analyzer, Carl Zeiss Meditec, Inc., Dublin, CA, USA) was conducted to detect characteristic glaucomatous visual field defects.

POAG was defined as a pathological elevation of IOP with a 24-hour peak exceeding 21 mmHg, accompanied by characteristic glaucomatous optic disc and RNFL damage, with or without corresponding visual field loss. Diagnosis required the presence of an open anterior chamber angle and exclusion of secondary causes of elevated IOP. CACG was defined as the absence of clinical symptoms or signs indicative of prior acute angle-closure attacks (e.g., glaucomatous flecking, keratic precipitates, or iris atrophy), but with more than 3 cumulative hours of peripheral anterior synechia, chronically elevated IOP (> 21 mmHg), and evidence of glaucomatous optic neuropathy or visual field defects.

Exclusion criteria were: (1) inability to excise complete TM tissue; (2) intraoperative hemorrhage preventing TM retrieval from the anterior chamber; (3) failure to fully expose nasal TM during goniosynechialysis in CACG cases; (4) long-term use of agents potentially affecting TM structure, such as topical corticosteroids or miotics (e.g., pilocarpine); (5) history of uveitis, trabeculitis, or other diseases potentially affecting TM structure; and (6) unwillingness or inability to sign informed consent.

Surgical procedures

All patients received preoperative mydriasis with 0.5% tropicamide eye drops, administered six times at 5-minute intervals, beginning 30 to 40 min prior to surgery. All surgical procedures were performed by the same experienced surgeon (HJW) to ensure consistency.

Both the POAG and CACG groups underwent phacoemulsification combined with intraocular lens implantation, following standard surgical protocols. In cases of CACG, 360° goniosynechialysis was performed using a specialized spatula to reposition the iris root, and peripheral anterior synechiae were mechanically separated in two successive passes.

For goniotomy, the patient’s head was rotated 30° away from the surgical site to facilitate nasal angle visualization using an Ocular Hill Open Access Surgical Gonio lens (Ocular Instruments, WA, USA). A KDB was introduced under gonioscopic guidance, and a TM strip covering approximately four clock hours was excised circumferentially.

To minimize acquisition and processing artifacts, all samples were obtained by the same experienced surgeon using a standardized KDB technique. Immediately following excision (within 1–3 min), the TM specimens were transferred into Eppendorf tubes containing freshly prepared 2.5% glutaraldehyde fixative (Servicebio, Wuhan, China). This rapid fixation, along with uniform handling, helped reduce ischemia time and mechanical manipulation prior to fixation. The samples were fixed at room temperature for 2 h and stored at 4 °C. They were then rinsed three times (15 min each) in 0.1 M phosphate buffer (pH 7.4), followed by dehydration, resin infiltration, embedding, polymerization, ultrathin sectioning, and staining for TEM analysis.

After polymerization, ultrathin sections approximately 70–90 nm thick were cut using an ultramicrotome (Leica EM UC7, Leica Microsystems, Germany) and mounted on copper grids. Sections were subsequently stained with 2% aqueous uranyl acetate for 15 min and lead citrate for 8 min to enhance contrast of cellular and extracellular structures. The grids were examined using a Hitachi HT7800 transmission electron microscope (Tokyo, Japan)operating at an accelerating voltage of 80 kV. Digital micrographs were captured with a Gatan OneView CMOS camera (Gatan Inc., Pleasanton, CA, USA).

For each specimen, three non-overlapping regions were selected for imaging at low (×3,000–5,000) and high (×10,000–20,000) magnifications to ensure representative coverage of the uveal and corneoscleral TM. A minimum of five micrographs per region were analyzed, and images were obtained at least 2 μm away from the excision edge to minimize processing artifacts.

TEM imaging strategy and artifact control

Due to the inherent limitations of KDB goniotomy, clear visualization of the Schlemm’s canal inner wall was achieved in only 9 of the 20 specimens. As a result, all ultrastructural analyses in this study were restricted to the uveal and corneoscleral regions of the trabecular meshwork. To guarantee objectivity and eliminate potential processing artifacts, a layered examination strategy at designated magnifications was implemented. Low-magnification micrographs (×3,000 to ×5,000) were utilized to evaluate macro-structural tissue alterations and screen for cellular necrosis. Concurrently, high-magnification micrographs (×10,000 to ×20,000) were mandated for the precise identification of intracellular organelle stress and pigment density. To minimize surgical and mechanical artifacts, regions immediately adjacent to the cut edge were strictly excluded. Only fields located at least 2 μm away from the excision margin were analyzed. Furthermore, isolated membrane discontinuities were not considered pathologic; such alterations were interpreted as disease-related only when they occurred away from the cut edge, coexisted with other organellar alterations, and were reproducible across non-contiguous regions.

Evaluation of trabecular architecture and extracellular matrix

The ultrastructural evaluation of the TM was performed using TEM with focus on four principal parameters: (1) Trabecular Beam Morphology – This included the assessment of inter-beam fusion and quantification of ITS width; (2) ECM Characteristics – Specifically the presence of sheath-derived plaques, pigment granule deposition, and general matrix organization. The degree of pigment accumulation was semiquantitatively graded by two independent observers (ZQL and KL) using a predefined criterion. “Excessive pigment accumulation” was defined as the presence of dense pigment granules occupying > 30% of the trabecular beam area or observed within > 50% of TM cells in at least two non-overlapping fields at ×10,000 magnification. Mild pigment presence (< 10% beam area or < 20% of TM cells affected) was considered within the expected range for aging changes. The interobserver agreement was evaluated using Cohen’s kappa coefficient (κ > 0.85).

Analysis of cellular ultrastructure and Organelle Severity Index (OSI)

Cellular evaluation encompassed overall TM cell density, migratory behavior (e.g., cells traversing fractured trabecular beams), and detailed organelle morphology. To facilitate correlation analysis and quantify the cumulative burden of organelle dysfunction, we developed an ad-hoc semi-quantitative scoring system—termed the Organelle Severity Index (OSI) for this study. To prevent sampling bias, a maximum of five intact TM cells with discernible intracellular structures were randomly selected per patient. Briefly, each evaluated TM cell was assessed by blinded observers for three independent ultrastructural stress markers: (1) mitochondrial swelling/cristae loss, (2) severe ER dilation, and (3) the accumulation of secondary lysosomes or multivesicular bodies (MVBs). Each present marker was assigned 1 point, yielding a stress score of 0 to 3 per cell. The OSI for each patient was then calculated as the mean score of all evaluated cells within that specimen.

Quantitative morphometry of collagen fibrils

Within the connective tissue, collagen fibrils embedded in an irregular, interwoven ECM network were analyzed. To avoid processing artifacts and mitigate oblique sectioning bias, the analysis was strictly restricted to non-edge regions, and fibril diameters were quantified using the minimum Feret’s diameter on high-magnification TEM images, thereby ensuring a reliable comparative evaluation between the POAG and CACG specimens. To assess the dimensional distribution profiles, raw diameter measurements were categorized into discrete bins to generate overlapping frequency distribution histograms. Furthermore, to account for intra-patient correlation and accurately reflect authentic biological variance, these measurements were additionally visualized using a nested scatter plot, wherein individual technical replicates (single fibril measurements) were systematically grouped by their respective biological replicates (individual patients) within each diagnostic cohort.

Assessment of repeatability and reproducibility

To assess the reliability of our measurements of collagen fibril diameter and area, we conducted a detailed analysis of repeatability and reproducibility. For this purpose, ten patients from each group were randomly selected. The first observer (ZQL) conducted two separate measurements of each parameter within a two-week interval to assess intra-observer variability. Additionally, a second observer (KL) independently measured the same set of images on a different day to determine inter-observer variability. The intra-observer and inter-observer variabilities were quantified using the intra-class correlation coefficient (ICC).

Statistical analysis

All statistical analyses were conducted using IBM SPSS Statistics (version 24; IBM Corp., Armonk, NY, USA). Collagen fibril diameters were evaluated from high-resolution TEM micrographs acquired at 11,500× magnification. For each fibril, the minimum Feret’s diameter—defined as the shortest distance between two parallel tangents across the object—was measured using ImageJ software (version 2.14.0, National Institutes of Health, USA). To achieve sufficient statistical power, a minimum of 15 to 30 individual collagen fibrils was measured per sample. The Shapiro-Wilk test was used to assess the normality of continuous data, and Levene’s test was conducted to evaluate the homogeneity of variances between groups. Based on the distribution characteristics, group comparisons of collagen fibril diameters and OSI Scores between the POAG and CACG cohorts were conducted using the Mann–Whitney U test for non-normally distributed data. Spearman’s rank correlation analysis was utilized to assess the relationships between clinical parameters (e.g., age, IOP, MD) and ultrastructural quantitative indices, including the OSI and mean collagen fibril diameter. All statistical tests were two-tailed, and a p-value of < 0.05 was considered statistically significant.

Results

Demographic and clinical characteristics of POAG and CACG patients

This study analyzed TM specimens from a total of 20 eyes, including 11 eyes from patients with POAG and 9 eyes from patients with CACG (Table 1) (Supplementary Material S1). The two groups demonstrated comparable demographic profiles, with no significant differences in age (66.2 ± 10.5 years for POAG, 66.3 ± 13.3 years for CACG, P = 0.978) or gender distribution (P = 0.642). Both groups also exhibited comparable BCVA and IOP, with no significant difference between the POAG and CACG groups (P > 0.05). However, significant differences were observed in the visual field indices. The POAG group had a significantly higher visual field index (VFI, 61.50 ± 24.4) compared to the CACG group (30.9 ± 18.7, P = 0.01). The mean deviation (MD) was also significantly worse in the CACG group (-24.10 ± 5.93 dB) compared to the POAG group (-13.80 ± 7.16 dB, P = 0.005).

Table 1.

Comparison of demographic and clinical characteristics between the two groups

	POAG	CACG	P value
Eyes	11	9	-
Gender (M/F)	4/7	2/7	0.642
Age (years)	66.2 ± 10.5	66.3 ± 13.3	0.978
BCVA (decimal)	0.34 ± 0.29	0.42 ± 0.31	0.548
IOP (mmHg)	21.0 ± 4.9	21.0 ± 9.9	0.443
VFI	61.50 ± 24.4	30.9 ± 18.7	0.01*
MD (dB)	-13.80 ± 7.16	-24.10 ± 5.93	0.005*
Diameter of collagen fibrils (nm)	43.61 ± 8.62	62.22 ± 12.97	0.018*
Area of collagen fibrils (nm²)	1862.92 ± 794.19	3532.37 ± 1509.72	0.027*

*P < 0.05

M, male; F, female; BCVA, best-corrected visual acuity; IOP, intraocular pressure; VFI, visual field index; MD, mean deviation; PSD, pattern standard deviation

TM cellular alterations in POAG and CACG

Ultrastructural examination of TM tissues from patients with POAG and CACG demonstrated a consistent finding of markedly sparse TM cellularity across all specimens (100%), regardless of glaucoma subtype (Fig. 1C and D), when compared with the normal trabecular architecture described in previous histological studies of non-glaucomatous eyes [12, 15]. At higher magnification, a subset of the remaining TM cells in both groups exhibited profound subcellular stress and degenerative changes compatible with necrosis—in fields pre-specified to be strictly away from the excision edge. Specifically, localized plasma membrane discontinuities were prominently observed in POAG specimens (Fig. 1A). Furthermore, robust signs of organelle damage were present in both cohorts, characterized by swollen mitochondria exhibiting an intact double-membrane structure but a complete loss of cristae, alongside severely dilated mitochondria (Fig. 1A and B). Based on our semiquantitative grading criteria, excessive pigment accumulation was observed in 4 of 11 POAG specimens (36.4%) and in 6 of 9 CACG specimens (66.7%).

Fig. 1.

Ultrastructural alterations of the trabecular meshwork (TM) in POAG and CACG specimens. (A) Representative transmission electron microscopy (TEM) micrograph of a POAG specimen demonstrating a nearly necrotic TM cell. The solid arrow and adjacent line indicate a discontinuous plasma membrane. The open arrow points to a swollen mitochondrion characterized by a double-membrane structure and loss of cristae, and the open star denotes a dilated mitochondrion. The square highlights a secondary lysosome. Scale bar: 1 μm. (B) Representative TEM micrograph of a CACG specimen showing comparable subcellular stress. The open arrow points to a normal mitochondrion with intact double membranes and visible cristae; the open star denotes a dilated mitochondrion. Solid arrows indicate two secondary lysosomes. Scale bar = 1 μm. (C, D) Lower magnification TEM micrographs of POAG (C) and CACG (D) specimens illustrating extensive tissue remodeling. Both panels display partially fused trabecular beams and cells (asterisks) abnormally migrating across a fractured beam. Arrows indicate compressed nuclei, while solid arrows highlight the disrupted beam structure. Scale bars: 5 μm

At lower magnification, macroscopic tissue remodeling was evident in both groups, featuring partially fused trabecular beams and disrupted structural integrity (Fig. 1C and D). Occasionally, TM cells were observed abnormally migrating across these fractured trabecular beams (18.2% in POAG; 33.3% in CACG). During this translocation, their nuclei appeared markedly compressed and deformed (Fig. 1C and D), suggesting a highly challenging mechanical microenvironment within the diseased extracellular matrix.

To substantiate our qualitative ultrastructural observations, we performed a semi-quantitative cellular analysis (Table 2). Implementing our balanced sampling threshold, a total of 31 intact TM cells from the 11 POAG specimens and 26 intact TM cells from the 9 CACG specimens were successfully identified and pooled for evaluation. The analysis revealed that a substantial proportion of residual TM cells in both glaucoma subtypes suffered from profound metabolic and autophagic stress (P > 0.05). High frequencies of specific stress markers, including swollen mitochondria, dilated ER, and the accumulation of secondary lysosomes/MVBs, were documented across the pooled cells in both cohorts.

Table 2.

Semi-quantitative analysis of ultrastructural alterations in trabecular meshwork cells from POAG and CACG Specimens

Ultrastructural Alteration	POAG	CACG	P value
Organelle Alteration (in intact cells)	31	39
Mitochondrial swelling / loss of cristae	48.4% (15/31)	64.1% (25/39)	0.189
Endoplasmic reticulum dilation	80.6% (25/31)	82.1% (32/39)	0.880
Presence of heterolysosomes / MVBs	45.2% (14/31)	43.6% (17/39)	0.895
Organelle Severity Index	1.69	1.63	0.866

MVB, multivesicular body

TM ECM alterations in POAG and CACG

In TM tissues from both groups (100% of specimens), the extracellular matrix (ECM) exhibited distinct structural abnormalities characterized by trabecular beam thickening, inter-beam fusion, and narrowed intertrabecular spaces (Fig. 1C-D). Quantitative ultrastructural analysis demonstrated that collagen fibrils in the CACG group were significantly thicker than those in the POAG group, displaying greater mean diameters (60.27 ± 12.97 nm vs. 43.61 ± 8.62 nm) and cross-sectional areas (3532.37 ± 1509.72 nm² vs. 1862.92 ± 794.19 nm²) with high measurement reproducibility (intra- and inter-observer ICCs > 0.927) (Table 1; Fig. 2). This morphometric divergence was further corroborated by frequency distribution analysis (Fig. 3A), which revealed a prominent rightward shift in the CACG cohort, indicating a higher proportion of abnormally thickened fibrils. Furthermore, nested scatter analysis at the individual patient level (Fig. 3B) confirmed that this increased fibril dimension in CACG is a generalized, cohort-wide pathological remodeling process rather than an artifact driven by individual extreme variations.

Fig. 2.

ECM Comparison of Trabecular Meshwork Collagen Fibrils in POAG and CACG Patients. Representative TEM micrographs showing the collagen fibril architecture within the trabecular meshwork of POAG (A) and CACG (B). Quantitative analysis reveals that the diameter of collagen fibrils in POAG patients is significantly smaller than those observed in CACG patients. Scale bar = 0.4 μm

Fig. 3.

Quantitative ultrastructural analysis of trabecular meshwork collagen fibrils. (A) Frequency distribution histogram comparing the diameters of collagen fibrils between the POAG (blue) and CACG (red) groups. The CACG cohort exhibits a wider, right-skewed distribution, demonstrating a prominent shift toward larger fibril diameters. (B) Nested scatter plot detailing the individual fibril diameters for each patient. Each semi-transparent dot represents a single collagen fibril measurement, with data clustered by individual patients along the x-axis to illustrate intra- and inter-patient variability. The consistent elevation of measurement clusters in the CACG group confirms the generalized nature of fibril thickening. POAG, primary open-angle glaucoma; CACG, chronic angle-closure glaucoma

Furthermore, a Spearman’s rank correlation analysis was performed to explore potential linear relationships between ultrastructural alterations (fibril diameter and OSI) and a comprehensive panel of macroscopic clinical metrics, including demographics, preoperative IOP, and visual field functional indices (VFI, MD, pattern standard deviation). However, no statistically significant correlations were observed within the current cohort (all P > 0.05; Supplementary Table S2).

Discussion

In this study, we conducted a comparative ultrastructural analysis of TM tissues from patients with POAG and CACG. The TM specimens were collected using the KDB excisional goniotomy, a technique previously utilized by Junk et al., [14] although their analysis was limited to histological examination. In contrast, we employed TEM to investigate cellular morphology, ECM organization, and collagen fibril architecture at high resolution. Our findings reveal distinct ultrastructural alterations between POAG and CACG specimens, underscoring the potential role of ECM remodeling in the pathogenesis of glaucoma. This study is among the first to provide quantitative measurements of collagen fibrils in human glaucomatous TM using freshly excised surgical tissue, thereby ensuring superior preservation of fine ultrastructural features and mitigating the artifacts commonly associated with postmortem specimens.

The TM plays a crucial role in regulating aqueous humor outflow and IOP. This function relies on its intricate extracellular matrix (ECM) and specialized resident cells that preserve its structural and physiological integrity [2, 16]. Both POAG and CACG samples exhibited a marked reduction in TM cellularity, accompanied by pronounced ultrastructural abnormalities including plasma membrane rupture, cytoplasmic content leakage, chromatin aggregation, karyopyknosis, swollen mitochondria with disrupted cristae, and dilation of the endoplasmic reticulum. Additionally, in both groups, the ECM of the TM showed trabecular beam thickening and reduced intertrabecular spaces. These changes align with previously reported glaucomatous degeneration of TM cells and reflect tissue compromise. Although a healthy control group was not available for direct comparison due to ethical restrictions, the pathological alterations observed in our study were evaluated against the established ultrastructural baseline of normal human TM reported by Lütjen-Drecoll (1989), Gong (2002) and Tamm (2009). In contrast to the abundant cellular coverage and regularly aligned intertrabecular spaces characteristic of healthy tissues, our specimens exhibited prominent glaucomatous remodeling [2, 8, 10–12, 17, 18]. Intracellular pigment granules were frequently observed within residual TM cells and in the surrounding ECM. These granules may represent phagocytosed cellular debris or pigment originating from the iris or ciliary epithelium. Their accumulation suggests an increased phagocytic burden and impaired lysosomal clearance capacity [19], indicative of TM cellular senescence and declining functional competence. This dysfunction may further impair ECM remodeling and exacerbate resistance to aqueous humor outflow.

To move beyond purely qualitative observations, we introduced the OSI to systematically quantify this subcellular stress. Our semi-quantitative analysis revealed comparably high OSI scores in both the POAG and CACG cohorts, indicating that once the disease progresses to a surgical stage, resident TM cells suffer from a uniform, severe degree of metabolic and autophagic exhaustion, regardless of the primary anatomical trigger (open versus closed angle). This shared “final common pathway” of cellular decompensation provides a critical pathophysiological context for the lack of significant linear correlations between OSI scores and preoperative clinical metrics (e.g., maximum IOP, VFI, or MD) observed in our cohort. Tissues harvested during incisional surgery inherently represent advanced, end-stage glaucomatous remodeling, where intracellular organelle damage has likely reached a pathophysiological saturation point (a ceiling effect). Furthermore, the long-term application of multiple topical hypotensive agents prior to surgery significantly artificially modulates clinical IOP, thereby decoupling the current macroscopic functional indices from the cumulative, irreversible microscopic structural damage within the TM.

Interestingly, the observed phenomenon of TM cells abnormally migrating across fractured trabecular beams may represent a maladaptive response to chronic structural disruption within the glaucomatous outflow pathway. In POAG, sustained elevation of IOP and cumulative mechanical stress can lead to ECM stiffening, trabecular beam fragmentation, and narrowing of intertrabecular spaces [5]. Under such mechanical challenges, TM cells are known to exhibit mechanosensitive behavior, responding to stiffness gradients within the ECM through processes such as durotaxis and mechanotaxis [20]. These gradients may direct TM cells toward sites of structural compromise in an attempt to restore homeostasis. Furthermore, TM cells possess inherent reparative capacities and have been observed to migrate into regions of ECM damage, akin to wound healing responses in other connective tissues. However, in glaucomatous conditions, impaired ECM turnover and excessive matrix cross-linking may hinder effective repair, resulting in persistent architectural disruption [21]. Consequently, the migration of TM cells through fractured beams, coupled with their compressed morphology and cytoskeletal distortion, reflects both an effort at structural adaptation and a consequence of mechanical overload. Taken together, this pathological remodeling may further compromise outflow resistance and contribute to disease progression.

Our study suggests that ECM remodeling within the TM may differ between POAG and CACG, with distinct structural alterations observed in collagen fibril architecture. An important observation is a notable difference in collagen fibril diameter between POAG and CACG. Quantitative analysis suggests that fibrils in POAG may be thinner compared to those in CACG. In POAG, disease progression may be associated with increased transforming growth factor-beta 2 (TGF-β2) expression, which could potentially promote collagen fibril crosslinking via lysyl oxidase (LOX) and its homologs (LOXL). This process could lead to a denser ECM structure and might result in a reduction in the collagen fibril diameter, as previously reported by Wordinger et al. [22] They suggested that such ECM changes may lead to increased stiffness and impaired outflow regulation, which appears consistent with our findings in POAG. This could lead to an overall increase in resistance to aqueous humor drainage, contributing to sustained elevation in IOP. Supporting this hypothesis, Fleenor et al. found that TGF-β2 induces changes in the human trabecular meshwork ECM, exacerbating the structural alterations associated with glaucoma [23].

In contrast, in CACG, prolonged mechanical compression due to elevated IOP and the potential formation of peripheral anterior synechiae may contribute to aberrant ECM remodeling characterized by increased collagen synthesis but with disrupted fibrillar organization. This pathological remodeling results in abnormally thickened collagen fibrils and disordered ECM architecture. Powell and colleagues have attributed such alterations to mechanical stress–induced ECM changes, which contribute to trabecular meshwork (TM) dysfunction [7]. Supporting evidence from studies on connective tissues such as tendons and ligaments suggests that reduced mechanical tension may disrupt collagen fiber alignment, potentially leading to fibril thickening and compromised biomechanical properties. These effects are most evident in models of weightlessness or disuse, where the absence of physiological mechanical loading may contribute to significant ECM disorganization [24]. In the context of CACG, such ECM disruption and fusion of trabecular beams could potentially reduce the intertrabecular space, possibly collapsing aqueous outflow channels or misaligning key structural boundaries. These biomechanical barriers may further exacerbate angle closure by limiting aqueous access to the conventional outflow pathway. This interpretation is further supported by our previous findings demonstrating that surgical excision of thickened and disorganized ECM using trabectome or KDB goniotomy significantly improves IOP control and maintains long-term outflow patency in CACG patients [25, 26].

While our study primarily focused on collagen fibril remodeling, it is important to acknowledge that ECM in the TM is a complex and dynamic structure composed of not only collagens but also microfibrils, elastin fibers, and proteoglycans, which contribute to tissue stiffness and aqueous humor outflow resistance. Previous studies have identified abnormalities in these ECM components in glaucomatous eyes, with elastin fibers and microfibrils exhibiting altered organization and density, further exacerbating outflow dysfunction. Given the interdependence of these ECM components, future research should focus on a more comprehensive analysis of the ECM, including the characterization of elastin, fibronectin, and proteoglycans, using advanced imaging techniques such as immunohistochemistry and proteomics. These approaches would provide a deeper understanding of the molecular mechanisms driving ECM remodeling in glaucoma and help identify potential therapeutic targets for improving aqueous outflow.

Despite its strengths, this study has several limitations. First, the use of two-dimensional TEM may introduce measurement bias due to oblique sectioning of fibrils; this was mitigated by applying minimum Feret’s diameter for improved accuracy. Second, collagen subtypes were not immunohistochemically distinguished, limiting insight into specific ECM contributions. Third, variations in tissue processing could have affected ultrastructural preservation. While KDB goniotomy provides fresh intraoperative tissue, it inherently limits our analysis to the proximal outflow pathway—specifically the uveal and corneoscleral trabecular meshwork. Consequently, our sampling does not reflect global outflow tissue pathology. Critical distal structures, including Schlemm’s canal, collector channels, and the episcleral venous system, could not be evaluated. Because glaucomatous eyes exhibit significant distal pathological changes that contribute heavily to overall outflow resistance, our ultrastructural findings represent only a localized segment of the outflow apparatus. To achieve a comprehensive understanding of global outflow pathology, future studies should combine histological analysis of proximal TM tissues with advanced in vivo imaging, such as anterior segment optical coherence tomography or aqueous angiography, to evaluate distal outflow pathways. Future studies using advanced techniques such as micro-ultrasound or 3D electron tomography may enable more comprehensive analysis of both the trabecular meshwork and Schlemm’s canal. Furthermore, the fixation protocol used 2.5% glutaraldehyde without osmium tetroxide postfixation, which may have contributed to suboptimal preservation of cellular membranes. However, rapid fixation, prolonged buffer washing, and exclusion of edge regions were implemented to mitigate these limitations. Several images showing minor structural disruptions were interpreted with caution and corroborated by adjacent regions to ensure biological relevance. Lastly, the lack of an age-matched normal control group is a significant limitation due to ethical constraints and the unavailability of adequately preserved normal human TM from eye banks. Compared with autopsy eyes, surgical samples provide fresh, intraoperative tissue but come with inherent limitations. Our findings should therefore be viewed as complementary to, and largely consistent with, prior autopsy-based studies.

Conclusions

In conclusion, our study reveals that POAG and CACG exhibit distinct ultrastructural alterations within TM, particularly in terms of collagen fibril architecture. These structural differences may reflect divergent pathological processes and contribute to outflow resistance through different mechanisms. Further studies incorporating advanced imaging and molecular analyses are warranted to identify novel targets and optimize patient management.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

BCVA

Best-corrected visual acuity

CACG

Chronic angle-closure glaucoma

ECM

Extracellular matrix

ICC

Intra-class correlation coefficient

IOP

Intraocular pressure

JCT

Juxtacanalicular tissue

KDB

Kahook Dual Blade

MD

Mean deviation

PACG

Primary angle-closure glaucoma

POAG

Primary open-angle glaucoma

TEM

Transmission electron microscopy

TM

Trabecular meshwork

VFI

Visual field index

Author contributions

ZQL was responsible for conceptualization, original draft preparation, and manuscript review. KKW contributed to methodology, investigation, and data curation. YL performed methodology and formal analysis. FRY and KL were involved in investigation, data curation, and formal analysis. HCY was responsible for visualization and data curation. YNH was involved in the production of trabecular tissue specimens. HJW provided supervision and validation, contributed to conceptualization, and was responsible for writing, review, and editing. All authors have read and approved the final manuscript.

Funding

This work was supported by the Capital’s Funds for Health Improvement and Research (2024-2-4087), the Peking University People’s Hospital Research and Development Funds (RDGS2024-06, RDE2025-08), and the Natural Science Foundation of Beijing (L258058). The funders had no role in the design and conduct of the study; the collection, management, analysis, and interpretation of the data; the preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

Data availability

The datasets generated and analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study protocol was reviewed and approved by the Ethics Committee of Peking University People’s Hospital (Approval No. 2024PHB122-001). All procedures involving human participants were conducted in accordance with the ethical standards of the institutional research committee and the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Furthermore, all patients provided written informed consent prior to participation in the study.

Consent for publication

Written informed consent was obtained from all individual participants for the publication of their clinical data and the accompanying ultrastructural images included in this study.

Competing interests

The authors declare no competing interests.