Tissue Type Plasminogen Activator Facilitates NMDA-Receptor–Mediated Retinal Apoptosis through an Independent Fibrinolytic Cascade

Abstract

purpose. To investigate the association between apoptosis and the fibrinolytic system in retinal cell damage.

methods. Tissue type plasminogen activator–deficient (tPA^−/−), urokinase type plasminogen activator–deficient (uPA^−/−), plasminogen activator inhibitor-1–deficient (PAI-1^−/−), α2 antiplasmin–deficient (α2 AP^−/−) mice, and their wild-type counterparts were used. Retinal cell damage was induced by intravitreal injection of the excitotoxin N-methyl-d-aspartate (NMDA). The TdT-dUTP terminal nick-end labeling (TUNEL) method was used to examine retinal cell damage.

results. tPA^−/− mice were resistant to retinal cell damage caused by administration of NMDA, and PAI-1^−/− mice were more injured than their wild-type. No significant difference was observed between uPA^−/− or α2 AP^−/− and their wild-type mice.

conclusions. The results strongly suggest that endogenous tPA, but not uPA acts as a facilitator in NMDA-induced retinal cell damage, and that its mechanism may not be associated with cleavage of plasminogen into plasmin in the fibrinolytic cascade.

Tissue-type plasminogen activator (tPA) is used for the clinical treatment of thromboembolic stroke to activate the fibrinolytic system. The major source of tPA is endothelial cells; however, tPA has also been detected in neurons. ¹ Therapeutic intervention with tPA in the nervous system may represent a two-edged sword, since it has been reported that tPA also promotes neurodegeneration after intracerebral injection of excitotoxins such as glutamate, ² and neuronal damage after a cerebral infarction is thought to be mediated by excitotoxins. ^{3 4 5 6}  

Retinal ganglion cell death is a common feature of many ophthalmic disorders, such as glaucoma and central artery or vein occlusion. Glaucoma in humans and monkeys is associated with a significant elevation in vitreal glutamate concentration. ⁷ The mechanism underlying retinal cell death in these diseases is not well understood. They are likely to involve, at least in part, ischemia–reperfusion injury, and the injury after ischemia may be due in part to the action of glutamate as an excitotoxin. ⁸  

Recently, we reported that tPA-deficient (tPA^−/−) mice are resistant to the retinal cell damage induced by excitotoxins, especially NMDA. ⁹ This result indicates that tPA facilitates NMDA-induced retinal cell damage, but the mechanism(s) by which tPA promotes retinal cell damage induced by NMDA, remains unclear. 

The blood fibrinolytic system, which degrades intravascular fibrin, is activated by tPA or urokinase-type plasminogen activator (uPA), which convert plasminogen into plasmin. ¹⁰ These plasminogen activators are antagonized by an endogenous factor, plasminogen activator inhibitor-1 (PAI-1), and plasmin is inhibited by α2 antiplasmin (α2 AP). 

In this study, to detect the association of excitotoxin-induced retinal cell death and the fibrinolytic system, tPA^−/−, uPA-deficient (uPA^−/−), PAI-1-deficient (PAI-1^−/−), and α2 AP-deficient (α2 AP^−/−) mice and their wild-type counterparts were used. According to a method we previously described, ⁹ insult to the retina was delivered by intravitreal injection of NMDA, and the degree of neuronal damage was estimated by the TdT-dUTP terminal nick-end labeling (TUNEL) method. 

Materials and Methods

Experiment Animals

Male tPA^−/−, uPA^−/−, PAI-1^−/−, α2 AP^−/−, and wild-type mice weighing 25 to 30 g (on C57BL/6 and SV129 backgrounds) were used in the present study. Deficient mice were generated by homologous recombination in embryonic stem cells, as described previously. ¹¹ All experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Intravitreal Injection

Mice were anesthetized with intraperitoneal injections of 50 mg/kg pentobarbital sodium (Nembutal; Dainippon Pharmaceutical Co., Ltd., Osaka, Japan). According to the method we reported, ⁹ all animals were intravitreally injected with a 30-gauge needle 0.5 mm behind the limbus in the temporal region of the globe, through the conjunctiva and sclera. In animals with dilated pupils, it was possible to view the needle entering the vitreous. Both eyes of each mouse (n = 6–9) were routinely injected with 3 μL of 10 mM NMDA (30 nanomoles). The mice that had postoperative complications such as retinal hemorrhage, vitreous hemorrhage, and retinal detachment were excluded from the analysis. 

Histology and TdT-dUTP Nick-End Labeling

The mice were killed with an overdose of pentobarbital sodium 12 hours after intravitreal injection. The eyes were enucleated and postfixed overnight in phosphate-buffered 10% formalin and then embedded in paraffin. Sections 3 μm thick were cut along the vertical meridian through the optic nerve. To detect the retinal cells undergoing DNA fragmentation in the course of apoptosis, TUNEL staining was performed according to a method previously described. ¹² The number of labeled cells in the ganglion cell layer (GCL), inner nuclear layer (INL), and outer nuclear layer (ONL) was counted in two central areas of the retina, approximately 250 μm long each, chosen from both sides of the optic nerve head. ¹³ Data were analyzed independently by two coauthors (MK, MN) in a blinded fashion. The number of TUNEL-positive cells per 250-μm length of the area in each retinal layer were averaged and plotted as the number of TUNEL-positive cells. The experimental results are expressed as the mean ± SD. Statistical analyses were performed by analysis of variance (ANOVA) with the Fisher protected least significant difference (Fisher’s PLSD) test. 

Results

On the basis of preliminary experiments, we injected 30 nanomoles of NMDA intravitreally and dissected the eyes 12 hours later. To determine the association of apoptosis and the fibrinolytic system in retinal cell damage, we first compared NMDA-induced retinal damage in tPA^−/− and wild-type mice. TUNEL-positive cells in both the GCL and INL in tPA^−/− mice after intravitreal injection of NMDA were significantly fewer than in wild-type mice (Figs. 1 2A) . This result strongly indicates that endogenous tPA acts as a facilitator in NMDA-induced retinal cell damage. 

To confirm the specificity of tPA, next we examined the contribution of another type of endogenous plasminogen activator, uPA. No significant difference in TUNEL-positive cells was observed between uPA^−/− and wild-type mice in the GCL and INL after intravitreal injection of NMDA (Fig. 2B) . 

Endogenous tPA and uPA activity is negatively regulated by the endogenous inhibitory factor PAI-1. NMDA was injected intravitreally in PAI-1^−/− mice to determine the role of endogenous PAI-1 in retinal damage. The number of TUNEL-positive cells in the GCL and INL after intravitreal injection of NMDA was significantly greater in PAI^−/− mice than in wild-type mice (Fig. 2C) . 

tPA and uPA are serine proteases that convert plasminogen into plasmin, and α2 AP is an inhibitor of plasmin. To clarify whether plasmin is the key factor in the facilitative effect of tPA against the retinal cell damage induced by intravitreal injection of NMDA, we determined the contribution of endogenous α2 AP. After administration of NMDA, no significant difference was observed between α2 AP^−/− and wild-type mice (Fig. 2D) . 

Discussion

Using tPA^−/− mice we recently reported that tPA facilitates NMDA-induced retinal cell death. ⁹ In the present study, to investigate the association of retinal cell damage and the fibrinolytic system, we used tPA^−/−, uPA^−/−, PAI-1^−/−, α2 AP^−/− mice, and their wild types. TUNEL-positive cells in both the GCL and INL in tPA^−/− mice, but not in uPA^−/− mice, after intravitreal injection of NMDA were significantly fewer than those in the wild type. Because endogenous tPA activity is negatively regulated by the endogenous inhibitory factor PAI-1, to determine the role of endogenous PAI-1 in retinal damage, we injected NMDA intravitreally into PAI-1^−/− mice. TUNEL-positive cells in the GCL and INL after intravitreal injection of NMDA were significantly greater in PAI^−/− mice than in wild-type mice. These results strongly suggest that tPA acts as a facilitator in NMDA-induced retinal cell damage, and that its mechanism may not be associated with cleavage of plasminogen into plasmin in the fibrinolytic cascade. 

It has been reported that tPA and uPA are present in the retina. Tripathi et al. ¹⁴ examined various structures of human and monkey eyes for the presence of tPA by using the peroxidase–antiperoxidase immunohistochemical technique with a monoclonal antibody specific for human tPA. As a result, the anterior layers of the retina were weakly stained. In many of the tissues examined, uPA appeared to coexist with tPA. Tripathi et al. ¹⁵ investigated the presence of uPA in various structures of the human eye by using an immunohistochemical technique. A moderately intense to intermediate reaction product was seen in the anterior layers of the retina, a weak reaction product appeared in the posterior layers of the retina, and the retinal pigment epithelium contained both tPA and uPA. Therefore, the defect of PAI-1 would enhance endogenous tPA activity in the inner retina and lead to retinal cell death. 

tPA is synthesized in basal conditions and is stored in vesicles. ^{16 17 18 19} However, in hippocampal CA1 neurons, tPA is undetectable in basal conditions, but is transiently induced after excitotoxic injury, ²⁰ suggesting that induced tPA facilitates NMDA-induced CA1 damage. Although the precise role of constitutive or induced tPA in excitotoxic injury has not yet been determined, our results in tPA-deficient mice support the hypothesis that endogenous tPA is an essential factor in NMDA-mediated neuronal degeneration. 

Our preliminary results showed that intravitreal injection of NMDA induces a dose-dependent loss of inner retinal elements, and there was a time-related appearance of TUNEL-positive nuclei in the inner retina. Lam et al. ²¹ showed intense labeling of nuclei between 12 and 24 hours after injection of NMDA. In the inner retina, retinal ganglion cells are particularly affected by extracellular glutamate, but a small percentage of cells in the INL are also stimulated. Although several different cell types in the INL express NMDA receptor subunits, only amacrine cells appear to express the same subunits as those detected in retinal ganglion cells. Amacrine cells may be adversely affected by NMDA. ^{22 23} The neuronal damage by NMDA is caused by calcium entry through the NMDA receptor, and elevation of intracellular calcium concentrations activate calcium-dependent protease, leading to neuronal death. ^{24 25}  

tPA promotes NMDA-induced neuronal degeneration in brain hippocampal CA1 neurons. ²⁶ Together with our present results, we can say that endogenous tPA is a common and important factor in NMDA-mediated neuronal degeneration. However, although it has been reported that tPA promotes not only NMDA-, but also transient ischemia-induced neuronal degeneration in the brain, ³ tPA^−/− mice showed resistance to NMDA- but not transient ischemia-induced neuronal damage in the retina. ⁹ We therefore speculate that in addition to NMDA receptor activation, another mechanism is involved in transient ischemia-induced retinal damage. 

The mechanism by which tPA modulates NMDA-receptor–mediated signaling is unknown, but Nicole et al. ²⁷ reported that tPA potentiates signaling mediated by glutamatergic receptors by interacting with and cleaving the NR1 subunit of the NMDA receptor in the cerebral cortical neuron cultures. At the same time, they report that this interaction between tPA and NR1 is prevented by pretreatment with recombinant PAI-1, a protein that blocks the tPA catalytic site. ²⁸ It has been suggested that tPA interacts with the NR1 subunit of the NMDA receptor through its catalytic site. ²⁷ However, Matys and Strickland ²⁹ questioned the data of Nicole et al., ²⁷ by suggesting that the anti-NR1 antibody used in their experiments was not specific for NR1 and may cross-react with plasminogen. They additionally indicated that Nicole et al. ²⁷ used cultures maintained in serum-supplemented medium to coimmunoprecipitate and identify the NR1 subunit as a substrate for tPA. This method could have led to misidentification of plasminogen or plasmin bands as the NR1 subunit in its native or cleaved form. In response, Nicole et al. stated that the excitotoxic injury and cleavage experiments were all conducted in serum-free solutions. A casein gel zymography assay did not detect the presence of active plasmin, thereby excluding a possible contamination of their samples. Our results show that tPA increased NMDA-induced retinal cell damage, not associated with another function of tPA, cleavage of plasminogen into plasmin. Our data are consistent with the results of Nicole et al., but whether this effect is due to cleavage of the NR1 subunit by tPA is a subject of future studies. 

In summary, tPA increased NMDA-induced retinal cell damage, and its mechanism is probably not associated with cleavage of plasminogen into plasmin in the fibrinolytic cascade. Retinal ganglion cell death is a common feature of many ophthalmic disorders, such as glaucoma and central artery or vein occlusion. Although the mechanism underlying retinal cell death in these diseases is not well understood, glaucoma in humans and monkeys is associated with a significant elevation in vitreal glutamate concentration. ⁷ Therefore, it is reasonable to hypothesize that retinal damage in ophthalmic diseases involves ischemia–reperfusion injury and the action of glutamate as an excitotoxin. Our study has provided key information on the mechanisms underlying retinal cell death and provides a basis for further investigation to identify fully all the mechanisms involved and novel therapeutic avenues for the treatment of various ophthalmic disorders. 

 Supported in part by a grant-in-aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan.

 Submitted for publication May 26, 2004; revised September 29, 2004; accepted October 28, 2004.

 Disclosure: M. Kumada, None; M. Niwa None; A. Hara, None; H. Matsuno, None; H. Mori, None; S. Ueshima, None; O. Matsuo, None; T. Yamamoto, None; O. Kozawa, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

TUNEL staining of retinas at 12 hours after intravitreal administration of NMDA (30 nanomoles) or vehicle. (A) Vehicle in wild-type; (B) NMDA in wild-type; and (C) NMDA in tPA-deficient mice. TUNEL-positive cells appeared in the GCL and INL. Scale bar, 25 μm.

Figure 2.

 

TUNEL-positive cells counted in the GCL and INL 12 hours after intravitreal injection of NMDA (30 nanomoles) in wild-type (□) and in (▪) tPA^−/− (A), uPA^−/− (B), PAI-1^−/− (C) or α2 AP^−/− (D) mice. Results are expressed as the mean ± SD (n = 12–18 eyes; six to nine mice). *Significant difference from the wild-type group at P < 0.05. Numbers in parenthesis indicate the percentage of TUNEL-positive cells versus each strain of wild-type mice.

The authors thank Kyoko Takahashi for her technical assistance.