Dear Editors,
In a transmission electron microscopy (TEM) study of the cornea involving 30 C57BL/6J mice, we made a fortunate discovery: a high‐resolution, two‐dimensional ultrastructural image capturing both ectosomes (ECTs) and exosomes (EXOs) within corneal endothelial cells (CEnCs). ECTs and EXOs represent two distinct classes of extracellular vesicles (ECVs), each playing a crucial role in intercellular communication.
ECVs were first reported before the 1990s and are released by various mammalian cells under both normal and pathological conditions. Nearly 30 years later, two distinct ECV subtypes, ECTs and EXOs, were identified. Both carry bioactive molecules like proteins, nucleic acids, and lipids, and facilitate intercellular communication via body fluids, influencing target cells through signaling, receptor transfer, or cargo delivery. ECTs (100–500 nm) and EXOs (50–150 nm, with the upper limit varying slightly between publications) differ in size due to their unique biogenesis. ECTs form by outward budding directly from the plasma membrane via specialized microdomains. EXOs, in contrast, originate from inward budding within endocytic membranes, forming multivesicular bodies (MVBs) that release EXOs through exocytosis. Unlike EXOs, ECT release bypasses the exocytic pathway. Both vesicle types offer advantages such as targeted delivery, protection of internal contents, and efficient transport of hydrophobic molecules.
Despite the growing recognition of their importance, ultrastructural imaging for budding of ECTs and EXOs in ocular surface cells had not been reported. Using TEM, we observed for the first time the simultaneous budding of these two vesicle types in CEnCs. In Figure 1A, we present ECTs in the midst of their outward budding, while Figure 1B reveals EXOs during their inward budding phase, underscoring the transient coexistence of both vesicle types within the single cell.
In Figure 1A, the isolated ECV measures approximately 275 nm in diameter, which exceeds the typical size range of EXOs (up to 150 nm). Its outward budding from the plasma membrane further indicates that it is most likely an ECT. The adjacent membrane‐like structure near the ECT suggests that the process of extracellular release is almost complete. A key characteristic distinguishing ECTs from EXOs is the potential accumulation of luminal proteins at high concentrations during ECT formation [
1]. These proteins can become concentrated through various mechanisms within the vesicles. When such protein‐enriched, dense areas are visualized using TEM, the denser regions absorb or scatter more electrons as the electron beam interacts with the sample. As a result, these areas appear darker in the TEM images, providing a clear visual cue of their composition and density.
In Figure 1B, we observed three vesicles at various stages of budding within a multivesicular body. Two of these vesicles are small (diameters < 150 nm), consistent with typical EXO precursors based on their size, location within endocytic cisternae, and inward budding characteristics. The third vesicle appears significantly larger, exceeding 150 nm in diameter, which is beyond the canonical size range for EXOs. Nonetheless, its position inside the MVB and morphological similarity to adjacent smaller vesicles suggest that it may represent an atypical EXO precursor. We acknowledge that this interpretation remains hypothetical, however, four factors may explain this deviation. First, TEM provides only a 2D cross‐sectional view of 3D structures, and this particular viewing angle may have exaggerated the apparent size. Second, ECVs are capable of membrane fusion, and accidental fusion of smaller precursors within the MVB cannot be ruled out. Third, aggregation of biomacromolecules such as proteins or lipids may cause localized density and swelling, further increasing its observed dimensions. Fourth, the interior of this atypical EXO precursor appears to have low electron density in the image, unlike ECTs, which typically exhibit high‐density contents. Given these possibilities and its context within the MVB, we still consider this vesicle most likely to be an EXO precursor despite its atypical size, though we must acknowledge the limitations of this assumption, as it requires further validation through molecular biological evidence.
Endocytosis, exocytosis, and intracellular trafficking are fundamental cellular processes mediated by membrane budding. While endocytosis‐related membrane budding has been extensively documented across various ocular cell types, including CEnCs, the study of exocytosis‐related budding has been less prevalent. For instance, Seow et al. [
2] demonstrated that certain peptide and DNA complexes are internalized by CEnCs via energy‐dependent endocytosis. In contrast, physiological functions represented by exocytosis, where cellular contents are transported to the extracellular space, including inward budding related to exocytosis and outward budding unrelated to exocytosis, have been reported much less frequently in ocular cells.
The investigation into membrane budding related to ECTs and EXOs in corneal cells, particularly CEnCs, is even more limited. McKay et al. discovered that ECVs secreted by corneal epithelial cells (30–100 nm) possess bioactivity and may promote myofibroblast differentiation [
3]. Desjardins et al. isolated and characterized EXOs (30–150 nm) through ultracentrifugation from human corneal epithelial cells, human corneal fibroblasts, and human CEnCs. They demonstrated that each of these ECV subtypes significantly accelerated wound closure in human corneal epithelial cells in vitro. This effect was hypothesized to be associated with the activation of signal transduction mediators involved in the HSP27, STAT, β‐catenin, GSK‐3β, and p38 pathways [
4]. In an earlier study, Han et al. used EXOs derived from mouse corneal epithelial cells (47.97 d.nm) and showed that they could enhance corneal fibroblast proliferation in vitro. They further demonstrated that EXOs derived from mouse corneal fibroblasts (34.22 d.nm) had a comparatively weaker effect on angiogenesis than those from epithelial cells [
5]. Shojaati et al. reported that EXOs (130–150 nm) derived from corneal stromal mesenchymal stem cells can reduce corneal fibrosis and inflammation [
6]. Similarly, Samaeekia et al. found that EXOs derived from human corneal mesenchymal stromal cells (40–100 nm) can accelerate corneal epithelial wound healing [
7]. To date, only 4 studies [
4,
8–
10] have provided functional and phenotypic characterization of EXOs derived from CEnCs. Notably, Verta et al. [
8] isolated EXOs (about 100 nm) from human aqueous humor and identified a subset originating from CEnCs, offering compelling evidence that these cells can secrete EXOs under physiological conditions. Gordon et al. [
9] further demonstrated selective endocytosis of proteins in organ‐cultured rat corneal endothelia, with internalized tracers localized in vesicles, endosomes, and MVBs, indicating active intracellular vesicle trafficking. Raphael et al. [
10] complemented these findings by showing adsorptive endocytosis dominates transcellular transport in rabbit CEnCs, with vesicles undergoing diverse intracellular fates. These studies collectively support the presence of complex vesicular activity in CEnCs and other corneal cells. Conversely, reports of ECTs originating from ocular surface cells remain absent and warrant further investigation to elucidate their specific mechanisms, despite the understanding that both types of vesicles and their associated physiological mechanisms are expected to be present in all cell types [
1].
The challenges in studying ECTs and EXOs are substantial, largely due to difficulties in purifying ECVs from single‐cell sources within tissue samples and the technical complexities of accurately imaging the nanoscopic ECVs. Most observations of ECVs are derived from specific in vitro experimental settings, making it particularly challenging to capture either ECTs or EXOs in tissues from living animals. Our ability to capture this unique moment is largely attributed to the characteristics and special anatomical structure of corneal endothelium. The corneal endothelium is composed of a single layer of CEnCs, with the inner surface of this layer closely adjacent to the aqueous humor but without direct contact with other cells. This means that the cell membrane on the aqueous humor side, including the ECVs forming on the cell membrane, can be observed very clearly via TEM. Additionally, CEnCs continuously exchange nutrients and metabolites with the aqueous humor, which increases the likelihood of observing ECVs. We believe the latter may be the primary reason, as ECTs and EXOs typically rely on diffusion through body fluids to mediate communication between cells. This fundamental physiological and pathological mechanism is expected to apply equally to the corneal endothelium. However, we are unable to conduct a quantitative study on this budding phenomenon using the TEM method to further elucidate its mechanism, as TEM is typically a morphological observation technique. Generally, nanoparticle tracking analysis, dynamic light scattering, and flow cytometry are more commonly used methods for quantitative ECV study. Further studies employing conventional molecular and biochemical approaches are also required to confirm and characterize these two types of vesicles under physiological conditions.
In summary, our study fortuitously captured a rare instance of membrane budding in mouse CEnCs, showcasing the formation of both ECTs and EXOs. This serendipitous observation provided exceptionally clear ultrastructural images. However, we must acknowledge the limitation of this work, as it is based on a single representative image (Supporting Information S1: Sample methods for TEM). To our knowledge, this study not only offers the first morphological evidence supporting the existence of ECTs originating from corneal tissue, but also provides a unique side‐by‐side comparison of the budding processes of both ECTs and EXOs within the same clearly visualized field, an observation that has not been previously reported in such detail. Its primary contribution lies in providing foundational ultrastructural observations with pedagogical and methodological value for ocular surface ECV research, as observed by TEM. Ultimately, these morphologic insights will guide future studies to selectively perturb each biogenesis pathway and decipher their specific functional roles in the cornea.
2026 The Author(s). Eye & ENT Research published by John Wiley & Sons Australia, Ltd on behalf of Higher Education Press.
PDF
(1361KB)
