Glaucoma is a serious and irreversible disease that causes blindness and is commonly termed the ‘silent thief of sight’, because loss of vision often occurs gradually. In the majority of cases, glaucoma is associated with intraocular pressure and low blood perfusion, and is characterized by progressive death of retinal ganglion cells (RGCs). As glaucoma progresses, it may lead to optic nerve degeneration and loss of retinal nerve fibers, and ultimately result in the loss of vision. So, facilitating the survival of RGCs may be an effective therapeutic target in the treatment of glaucoma. Within the retina, RGCs are the only projecting neurons that transmit visual information to the brain (1), representing the third-order neurons of the visual pathway (2). Following optic nerve injury, RGCs are unable to regenerate their axons and cell apoptosis is rapidly initiated (3).

Numerous studies have indicated that apoptosis may be the final common pathway for RGC death in glaucoma (4–7). However, it has become increasingly evident that apoptosis and necrosis alone do not adequately encompass cell death. Studies have revealed that although apoptosis is actively executed by cysteine-aspartic acid proteases (caspase), RGC death occurs independently of caspase-mediated pathways (8,9), which are a family of cysteine proteases that are also essential in necrosis and inflammation (10). These data imply that RGC death involves apoptosis as well as non-apoptotic programmed cell death (PCD). Autophagy is a type of cell death that is morphologically distinct from apoptosis and which usually mediates cytoprotection, avoiding the apoptotic or necrotic demise of cells (11,12). Activation of autophagy reduces or inhibits the apoptosis of the cells. Several studies have provided evidence that the survival of RGCs may be improved by autophagy following optic nerve axotomy in mice (13,14). These investigations demonstrated that besides from apoptosis and necrosis, other pathways may also be responsible for RGC death. The accumulating literature providing evidence of the mechanisms of non-apoptotic RGC death suggests that other possible pathways are involved.

Cell death is categorized into PCD and passive (necrotic) cell death (15,16). While apoptosis is the best characterized form of PCD, other non-apoptotic modes also exist, including autophagy, paraptosis and mitotic catastrophe (16). Autophagy, as a form of PCD, is a self-degradative process that involves the breakdown of intracellular proteins and organelles (17). Paraptosis is a novel type of PCD, that is caspase independent and is characterized by cytoplasmic vacuolization derived from the endoplasmic reticulum and/or mitochondria swelling. Paraptosis is distinguished from apoptosis by its non-response to caspase inhibitors (that block apoptosis) and lack of apoptotic morphology. This type of cell death is also insensitive to autophagic inhibitors and lacks DNA fragment and poly (ADP-ribose) polymerase (PARP) cleavage (15,18). Several studies have identified AIP-1/Alix and phosphatidylethanolamine binding protein 1 (PEBP-1) as inhibitors of paraptosis (15,19).

Paraptosis is fundamentally different from other forms of PCD, such as apoptosis, autophagy and oncosis. Apoptosis typically involves the activation of caspases, responds to caspase inhibitors and exhibits morphological features including chromatin condensation, nuclear fragmentation and the formation of apoptotic bodies. Autophagy exhibits a contrasting morphology during cytoplasmic vacuolization in comparison with paraptosis. The process of autophagy is the formation of autophagosomes, which are double-membrane vesicles responsible for delivering long-lived proteins and excess or damaged organelles, into the lysosome for degradation (20). In addition, paraptosis is not inhibited by autophagy inhibitors, including 3-mathyladenine (3-MA). Oncosis is another form of non-apoptotic PCD, which is distinguished from paraptosis and occurs in the destroyed cell membrane.

In the present review, it was hypothesized that the progressive death of the RGCs in glaucoma is associated with a distinguished form of cell death, aside from the well characterized apoptosis, necrosis and autophagy, termed paraptosis (Fig. 1). Paraptosis may be triggered in the early stages of glaucoma (such as the appearance of optic nerve abnormalities, consistent with glaucoma, but with no visual field abnormalities). Furthermore, it was hypothesized that paraptosis and apoptosis may, together with autophagy, occur alone or simultaneously in RGCs, depending not only on the clinical severity but also on the duration of glaucoma development.

The second hypothesis is that paraptosis in glaucomatous RGCs may be derived from damage to the mitochondria. Mitochondria destruction plays a key role in inducing paraptosis in glaucomatous RGCs, and this may be associated with mitochondria-derived reactive oxygen species (ROS) reactions.

Our results indicated that swelling mitochondria and vacuoles were detected within 5 and 7 days, following optic nerve crush injury by clamping, in RGCs of adult rats and that these vacuoles originated from the mitochondria. In this investigation, the optic nerve of adult wistar rats underwent crushing as described previously (31). Male wistar rats weighing between 200 and 250 g were purchased from the Laboratory Animal Center, School of Medicine of Xi’an Jiaotong University (Xi’an, Shaanxi, China). All animals received human care in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The present study was approved by the ethics committee of Shaanxi Institute of Ophthalmology (Xi’an, China). Briefly, the rats were anesthetized and the optic nerves were gently crushed with blunt forceps for 50 sec, sparing the retinal and optic disc circulations. The rats were then returned to their cages and sacrificed at the time points of day 3, 5 or 7 and retinal ultrastructure impairment was analyzed using a transmission electron microscope (H-7650; Hitachi, Tokyo, Japan). All procedures were performed aseptically and on the left eye, with the right eye serving as a sham-operated control. The optic nerve crush injury by clamping induced ultrastructure changes in the RGCs, as demonstrated in Fig. 2.

All the images of RGCs had well-defined plasma membranes and uniformly distributed chromatin in the nucleus, suggesting apoptosis had not evolved in the RGCs within 5 and 7 days following optic nerve crush injury. However, the images in Fig. 2C and 2D demonstrate swollen mitochondria with disintegrated cristae and decreased matrix density. At day 5 and 7, the rate of mitochondrial fusion increased until the cells were almost fully occupied by several large mega mitochondria. However, autophagosomes containing cytoplasmic organelles were not detected, demonstrating that the cell death was not due to autophagy. All the ultrastructural analysis confirmed that the mitochondria were targeted in the RGCs within 5 and 7 days following optic nerve crush injury and the type of RGC death was neither apoptosis nor autophagy, but instead fitted the criteria of paraptosis (15,18).

Glaucoma is a multi-factorial group of eye diseases that result in optic nerve damage, particularly axonal damage, which triggers RGC cell death and the subsequent loss of vision. The optic nerve crush injury rat model is an important experimental disease model for glaucoma and in our study, this model was utilized to examine the injury and morphological changes of RGCs.

Taken together, the above study confirmed that optic nerve crush injury triggered paraptosis in RGCs, suggesting that paraptosis may be involved in glaucomatous RGC death. Further investigations are required to determine whether the excessive production of ROS is triggered in glaucomatous RGCs. However, ultrastructural analysis of swollen mitochondria with disintegrated cristae and intact endoplasmic reticulum further confirmed the hypothesis that mitochondria are important in triggering paraptosis in glaucomatous RGCs.

In the present review, the evidence suggesting that progressive death of RGCs in glaucoma involves another novel non-apoptotic PCD termed ‘paraptosis’ in the early stages of glaucoma has been summarized. We conceive this to be a two step process: (i) an excessive production of ROS impairs the mitochondria in glaucomatous RGCs and (ii) the damage to the mitochondria leads to RGC paraptosis. In addition, paraptosis and apoptosis, possibly even together with autophagy, may occur simultaneously in RGCs in the moderate and/or severe stages of glaucoma. Further studies are under way to confirm this theory. These hypotheses stand to enhance the understanding of the mechanisms of glaucomatous RGC damage and provide a novel strategy for protecting RGC by inhibiting paraptosis. Furthermore, we expect this will facilitate the identification of a number of chemical compounds/proteins that directly delay or prevent paraprosis-induced RGC death and thus improve visual damage caused by glaucoma.