Abstract
Purpose.:
To examine the functional significance of EphB/ephrin-B upregulation in mouse experimental glaucoma.
Methods.:
In a loss-of-function approach, mouse mutants lacking EphB2 (EphB2 −/− ) or EphB3 (EphB3 −/−) protein, and mutants expressing EphB2 truncated in the C-terminus (EphB2lacZ/lacZ ) were subjected to laser-induced ocular hypertension (LIOH), an experimental mouse model of glaucoma. The number of optic nerve axons was counted in paraphenylenediamine (PPD)-stained sections and compared between EphB mutants and wild type littermates. In a gain-of-function approach, retina/optic nerve explants obtained from LIOH-treated animals were exposed to EphB2-Fc recombinant proteins or Fc control proteins. Tissue sections through the optic nerve head (ONH) were labeled with neuron-specific anti-tubulin β-III antibody to determine axonal integrity.
Results.:
Both EphB2 and EphB3 null mutant mice exhibited more severe axonal degeneration than wild type littermates after treatment with LIOH. Mutant mice in which the C-terminal portion of EphB2 is truncated had an intermediate phenotype. Application of EphB2-Fc recombinant protein to LIOH-treated optic nerve explants resulted in greater sparing of tubulin β-III–containing retinal ganglion cell (RGC) axons.
Conclusions.:
These results provide genetic evidence in mice that both EphB/ephrin-B forward and reverse signaling feed into an endogenous pathway to moderate the effects of glaucomatous insult on RGC axons. LIOH-induced axon loss is maintained in retina/optic nerve explants after removal from an ocular hypertensive environment. Exogenous application of EphB2 protein enhances RGC axon survival in explants, suggesting that modulation of Eph/ephrin signaling may be of therapeutic interest.
As glaucoma is a leading cause of blindness worldwide, the underlying disease mechanisms that impact visual function are of substantial interest. From a consideration of the early patterns of visual field loss in patients, the optic nerve head (ONH) is generally regarded as an important site of pathogenesis resulting in retinal ganglion cell (RGC) axon injury. Support for this view also comes from experimental findings of significant structural remodeling
1 –7 and obstruction of axoplasmic transport
4,8 –11 at the ONH early in the course of disease. In addition, genetic evidence indicates that when RGC soma apoptosis is prevented, intraretinal axons are largely preserved until they reach the ONH, where massive degeneration ensues.
12 Efforts have been undertaken to identify genes differentially expressed at the glaucomatous ONH to shed some light on the pathways involved in axon injury and survival.
13 –17
In this study, we focused on the potential role of the Eph/ephrin family of cell surface signaling molecules in optic axon degeneration after glaucomatous injury. The Eph receptor tyrosine kinases and their ephrin ligands function not only in developmental axonal guidance, cell migration, and morphogenesis, but also in adult synaptic plasticity, homeostasis, and cancer.
18 –22 In addition, altered Eph and ephrin expression has been reported in many central nervous system pathologies
23 –25 and genetic studies have demonstrated a functional role for Eph/ephrin signaling in modulating axon survival and regrowth after spinal cord
26 and optic nerve injury.
27 A key feature of the Eph/ephrin system is that signal transduction occurs bidirectionally
22,28 and involves both forward and reverse signaling. In forward signaling, ephrin binding to Eph molecules triggers autophosphorylation of the Eph tyrosine kinase domain and a subsequent signaling cascade. In reverse signaling, the binding of an Eph with an ephrin molecule activates signaling pathways within the ephrin-expressing cell.
Among glaucoma-related changes at the ONH, a finding that is consistently observed across multiple animal models and in glaucoma patients is the upregulation of
EphB/ephrin-B gene expression and protein signaling. The expression of a number of Eph and ephrin family members is upregulated in cultured ONH astrocytes derived from human patients,
13,29 and in several different animal models of glaucoma including monkey,
29 the DBA/2J mouse model of pigmentary glaucoma,
17,30 and the laser-induced ocular hypertension (LIOH) model in CD-1 mice.
31 In mice, where this process has been examined in the greatest detail,
EphB/ephrin-B upregulation is tightly correlated with axon loss,
30 and occurs early in disease, preceding or coinciding with the initial morphologic signs of axon damage.
31 This upregulation of
EphB/ephrin-B gene expression has been found to be associated with increased active protein signaling in both axons and glia at the ONH.
31 Furthermore, morphologically normal axons exhibit higher levels of ephrin-B reverse signaling, whereas this signaling pathway is downregulated in aberrant axons.
31 Despite these correlational findings, whether Eph-ephrin signaling plays a functional role in disease remains unknown.
In the present study, we subjected mouse mutants lacking EphB2 (
EphB2 −/−) or EphB3 (
EphB3 −/−) protein or mutants with engineered alleles of
EphB2 (
EphB2lacZ/lacZ ) to glaucomatous optic nerve damage induced by LIOH.
EphB2 and
EphB3 were chosen as the genes of interest because their mRNAs were shown to be upregulated at the ONH as early as 1 to 2 days after LIOH treatment.
31 As substantial data indicate axon dysfunction and degeneration precede retinal ganglion cell body loss,
32 –34 we focused our analysis on the integrity of axons in the optic nerve. Mice totally deficient in EphB2 or EphB3 both exhibited more severe axon degeneration compared with wild type littermates, suggesting that the EphB/ephrin-B pathway normally operates to moderate axon loss in LIOH-induced experimental glaucoma. Exogenous application of EphB2 recombinant protein attenuated axon degeneration in LIOH-treated optic nerve explants, further supporting the involvement of EphB/ephrin-B signaling in glaucomatous optic nerve pathophysiology.
Mice were anesthetized with an overdose of pentobarbital and perfused transcardially with 4% PFA in 0.1 M phosphate buffered saline (PBS, pH 7.4). Enucleated eyes with a segment of retrobulbar optic nerve attached were harvested, and the anterior segment, lens, and vitreous were removed. The eye cups were immersion-fixed in the same fixative at room temperature for 2 hours, cryoprotected with 30% sucrose in PBS, and embedded in optimal cutting temperature (OCT) compound (Tissue-Tek; Sakura Finetek, Torrance, CA). Tissues were sectioned at 12 μm thickness along the longitudinal axis through the ONH, and mounted on slides (Super Frost Plus; Fisher Scientific, Springfield, NJ).
Cryosections were washed in 0.1 M PBS and blocked with 10% normal donkey serum (NDS) for 1 hour at room temperature. Triton X-100 (0.1%) was included in the blocking solution to achieve tissue permeabilization. Primary antibodies were incubated overnight at 4°C, followed by PBS washes and 1 hour of secondary antibody incubation at room temperature. Slides were mounted in mounting medium (Vectashield; Vector Laboratories, Burlingame, CA).
The following primary antibodies were used: mouse anti-tubulin β-III (TUJ1; 1:500; Covance Research Products, Denver, PA), and mouse anti-glial fibrillary acidic protein (GFAP; 1:400; Sigma, St. Louis, MO). Colabeling experiments were performed with secondary antibodies conjugated to cy3 and cy5 (1:200; Jackson ImmunoResearch, West Grove, PA) to ensure maximal spectral separation. Confocal images were acquired on a microscope (LSM5 Pascal; Carl Zeiss Meditec, Inc., Thornwood, NY), and pseudocolored in red and green to aid visualization.
Optic nerve explants were cultured with retinas attached at the interface of media and a 5% CO2/air mixture. RGC axons approximately 5 mm long were in continuity with their cell bodies, prolonging survival in vitro. In addition, the cellular architecture and potential interactions within the optic nerve were preserved by this approach, potentially providing a more physiologically-relevant environment compared with dissociated cell culture. Mouse eyes with connected optic nerves were quickly removed, and the anterior segment, sclera, and meninges were dissected away in cold HEPES-buffered aCSF. Complete removal of meninges around the ONH required great care so as not to damage the nerve. The retina/optic nerve explants were rinsed briefly with pre-equilibrated growth media (containing Neurobasal-A supplemented with B-27 and Penicillin-Streptomycin; Invitrogen, Carlsbad, CA), and transferred onto an organotypic culture insert (Millicell; Millipore, Bedford, MA). One primary cut was made from the retinal periphery to the optic disc region, and the retina (ganglion cell layer facing up) was spread out flat on the organotypic culture insert (Millicell; Millipore) with a custom-made ball-headed instrument. To ensure sufficient gas exchange, this tissue was arranged so that the optic nerve was not covered by the retina, while maintained in a moist condition during the entire procedure. After removal of excessive growth media, the insert was placed into a 6-well plate filled with media. Cultures were inspected every day to make sure a thin film of media remained over the explant, and media change was performed every 2 days.
To compare axonal degeneration between LIOH-treated optic nerves and control unoperated optic nerves in vitro, CD-1 mice (Charles River) were subjected to LIOH in one eye, while the contralateral eye was left untouched. Two days after treatment, optic nerves were harvested and cultured as described above for an additional 4 days. The explants were then fixed in 4% PFA and embedded in OCT. The frozen tissues were cut into 12 μm longitudinal sections, and every third section was collected onto the same slide. Only sections through the ONH were analyzed, yielding five to six samples per slide. Neuron-specific anti-tubulin β-III antibody
41,42 was used to stain the optic nerve sections as a measure of axonal integrity.
To examine the effect of exogenous EphB2 application, we used a recombinant protein of mouse EphB2 fused with a human Fc tag (EphB2-Fc). The Fc tag allows for the purification, detection, and preclustering of EphB2 protein. Functional Eph-ephrin signaling requires ligand-induced receptor clustering. Soluble ephrin- or Eph-Fc fusion proteins therefore need to be preclustered with anti-Fc antibodies to induce receptor autophosphorylation and activation, while nonclustered fusion proteins may act as antagonists. CD-1 mice were treated with LIOH bilaterally. Two days after treatment, the pair of optic nerves from each mouse was cultured in growth media supplemented with EphB2-Fc (10 μg/mL; R&D Systems, Minneapolis, MN) or control Fc (10 μg/mL; Jackson ImmunoResearch) respectively. EphB2-Fc and Fc recombinant proteins were both preclustered with goat anti-human Fc antibodies (20 μg/mL; Jackson ImmunoResearch) before their addition to the media. Explants were harvested and analyzed after 4 days in vitro.
To demonstrate EphB2-Fc protein binding within optic nerve explants, we cultured non-LIOH treated nerves for 1 day in the presence of Fc or EphB2-Fc preclustered with goat anti-human Fc antibodies. Cryostat tissue sections of the optic nerves from these explants were labeled with additional anti-Fc antibodies to boost the signal, followed by donkey anti-goat secondary antibodies.
Confocal images were analyzed with software developed by Wayne Rasband (ImageJ; National Institutes of Health, Bethesda, MD; available at
http://rsb.info.nih.gov/ij/index.html). Thresholds for tubulin fluorescence were set using identical parameters for each pair of optic nerves from the same animal. A 150 × 200 μm box was selected at the ONH from each section, in which the extent (pixel area) of the tubulin signal was determined and averaged as a measure of axon survival. Data from the treatment groups were normalized against the contralateral control sample from the same animal.
EphB2 and EphB3 Deficiencies Independently Result in Increased Pressure-Related Axon Degeneration In Vivo
The in vivo analysis of axon integrity and damage in mutant animals provided evidence that EphB/ephrin-B signaling influences axonal survival after LIOH treatment. As the loss-of-function data implicated both EphB forward signaling and ephrin-B-mediated reverse signaling, we wondered whether manipulation of EphB/ephrin-B signaling via a gain-of-function approach can modulate the severity of axon damage after LIOH.
Explants consisting of optic nerves with retinas attached were cultured and stained with tubulin β-III antibodies as a measure of axon integrity in the ONH region. Untreated control optic nerves from wild type CD-1 mice exhibited robust tubulin β-III staining (
Fig. 4A), indicating that this explant culture method adequately supports RGC axon survival for at least 4 days in vitro. In contrast, retina/optic nerve explants cultured 2 days after LIOH treatment (
Fig. 4B) and examined after 4 days in vitro displayed evidence of axonal degeneration, with readily apparent axon swelling, aberrant axon trajectories, and large regions of axon drop-out at the ONH. Quantitative analysis showed that the level of tubulin β-III immunoreactivity was significantly reduced by 50 ± 9% (
P = 0.003) in LIOH treated retina/optic nerve explants compared with untreated controls (
Fig. 4C). Although some degeneration inevitably occurs in any culture system, LIOH treatment demonstrably resulted in more damage than that seen in explants from untreated animals prepared and cultured in the same manner.
Examination of LIOH-treated optic nerve explants revealed that the axon damage and loss observed after 4 days in vitro occurred after placement in culture. Optic nerves harvested 2 days after LIOH and examined immediately with immunohistochemistry exhibited some local axon swellings and defasciculation, but widespread axon loss had not yet become pronounced (data not shown). In contrast, the extent of axon loss was more severe in optic nerves treated with LIOH in parallel but subjected to the additional 4 days of culture. These findings suggest that although IOP elevation is no longer present in vitro after tissue removal from the animal and placement in culture, axonal degeneration in cultured LIOH-treated retina/optic nerve explants continues to progress further.
Using LIOH-treated optic nerve explants, we next investigated the effect of exogenous EphB2 recombinant protein application on axonal survival in vitro. Preclustered EphB2-Fc was added to the culture media, while Fc protein alone was used as control. Fc-treated optic nerves (
Fig. 4D) were morphologically similar to those cultured in media alone (
Fig. 4B). In EphB2-Fc-treated optic nerves, axonal swelling and meandering trajectories could still be observed, but a larger number of tubulin β-III-positive axons were preserved (
Fig. 4E). Quantification of tubulin β-III staining indicated that the level of immunoreactivity in LIOH-treated optic nerves was significantly improved by 80 ± 11% (
P < 0.0003) with EphB2-Fc application compared with Fc alone (
Fig. 4F).
To demonstrate EphB2-Fc protein binding within optic nerve explants, we cultured non-LIOH treated nerves for 1 day in the presence of preclustered Fc or EphB2-Fc. Significant binding of recombinant EphB-Fc protein was observed in EphB2-Fc treated nerves (
Fig. 5B), compared with the background level in Fc treated controls (
Fig. 5A). EphB2-Fc colocalized with both the axonal marker tubulin β-III (
Figs. 5C–E) and the astrocytic marker GFAP (
Figs. 5F–H), consistent with the expression of ephrin-B in both axons and astrocytes.
31 These results are also reminiscent of previous studies showing that EphB2-Fc and ephrin-B2-Fc proteins bind to RGC axons after in vivo application in the optic nerve.
30
Bidirectional EphB/Ephrin-B Signaling Promotes Axonal Survival in Mouse Experimental Glaucoma