In this study, we compared three inducible rodent glaucoma models against two different rat strains. Ocular hypertension was induced either by internal blockage of aqueous humor outflow (with magnetic bead or hydrogel injection) or by external compression with circumlimbal suture, which is thought to render aqueous humor overflow excessive of drainage capacity.
18 With both strains, the magnetic bead injection model resulted in a steady increase of IOP without spikes, whereas the hydrogel injection model and circumlimbal suture model required an immediate IOP spike to maintain increased IOP for more than 2 weeks. For the internal approach (magnetic bead or hydrogel injection), the BN rats were superior to the SD rats with regard to maintaining increased IOP. No such strain difference was observed for the external approach (i.e., the circumlimbal suture model). This implied that different strains might affect increases in IOP differently depending on the model mechanism and thus should be considered cautiously when establishing any rodent glaucoma model.
The use of magnetic beads has the merit of preventing reflux spillage during needle removal, specifically by means of a magnet that keeps the beads away from the needle track (
Supplementary Movie S1). Upon needle removal, the high-pressure gradient between the inside and outside of the eyeball and the patent needle track leads to reflux of the aqueous humor and beads. To prevent this, we adopted three strategies: (1) use of a glass capillary needle fabricated to have very narrow and sharp tip, (2) performing an incision parallel to the corneal limbus to make a longer track, and (3) use of magnetic beads and a magnet to hold the beads at the end of procedure. In this way, we could prevent spillage of magnetic beads, although aqueous humor reflux could not be completely blocked. The magnetic bead injection model showed a steady increase of IOP without any immediate IOP spike. This absence of IOP spike might be the result of the incomplete blocking of aqueous humor reflux, which presumably neutralized IOP in the immediately postoperative period.
The magnetic bead injection model produced two gross anatomical changes. First, the steady IOP increase induced apparent buphthalmos in all BN rats and in some of the SD rats; in fact, global eyeball expansion after magnetic bead injection in BN rats has been reported.
32 As occurs in congenital glaucoma, eyeball expansion might be possible under conditions of elastic sclera and steady IOP increase.
32,44,45 Although less evident, buphthalmos also was observed in the SD rats in the present study. Its lesser occurrence in SD rats might be associated with poorer maintenance of the IOP increase with the magnetic bead injection model. As for the second gross anatomical change produced by the magnetic bead injection model, it, unlike the other two models, induced neovascular complications including iris neovascularization and hyphema, although rarely. In our pilot study, the larger magnetic bead injection volume could enable better IOP maintenance from the immediately postoperative period, but it also led to greater chances of neovascular complications. Therefore, we should limit the injected magnetic bead volume for a steady IOP increase. In contrast, no neovascular complication was observed in the hydrogel injection or circumlimbal suture models, despite much higher IOP spikes in the immediately postoperative period. Thus, we speculated, as did Tribble et al.,
32 that magnetic bead injection itself might induce vascular compromise.
In the hydrogel injection model, aqueous humor outflow was blocked by the hydrogel, which had been transformed to the gel state after its injection. Because the premixed agent was injected in the liquid state, no needle blockage by the injected particles could occur, which allowed for the use of a very fine glass needle tip. Consequently, reflux was minimized, even to the extent that enabled immediately postoperative IOP spikes. Further, the anterior segment of the eye was unaffected by the particles, due to the transparency of the gel material (
Supplementary Movie S2). This certainly would be a great advantage in terms of postoperative imaging. Unfortunately, however, the IOP increase could not be maintained for more than 2 weeks without a booster injection. This was the reason for having to set a very high immediately postoperative IOP to maintain IOP elevation longer. Because eyeball expansion was rarely observed, we speculated that this model might simulate a subacute IOP elevation model better than chronic glaucoma. Insufficient IOP elevation in the hydrogel model, however, should be interpreted with caution. Our hydrogel model had less IOP elevation when compared to the reference work of Huang et al.
34 Later, Yu et al.
37 showed not only IOP elevation but also functional impairments, which were reversed after IOP-lowering treatments in the hydrogel injection model using SD rats. Therefore, our hydrogel injection method might have to be optimized further. Moreover, newer hydrogel agents
36,38,46 and post-injection modifications using ultraviolet lights
35,38 have been developed. Therefore, we speculated that the hydrogel injection model can be further improved in the future.
The circumlimbal suture model also required a very high immediately postoperative IOP. This IOP elevation had dropped by the next day but remained above the normal range for a month. IOP spikes immediately after suturing have been reported consistently for the circumlimbal suture models.
18–20 This finding contrasted with the gradual decrease or gradual increase in the hydrogel injection model or magnetic bead injection model, respectively. Interestingly, no inter-strain difference was observed in this model. Because the SD rats are less expensive than the BN type, the circumlimbal suture model would be beneficial with regard to cost relative to the other models. On the other hand, this model resulted in less RGC death, which is probably related to the poorer maintenance of increased IOP. Considering the IOP profile and the absence of eyeball expansion, we speculated that the circumlimbal suture model might simulate acute IOP elevation better than chronic glaucoma.
The internal aqueous outflow blocking approach showed inter-strain differences, as increased IOP was maintained better in the BN rats than in the SD type. Similar differential susceptibility according to strains has been reported for mice glaucoma models aiming to obstruct the aqueous humor outflow pathway with microbeads. Cone et al.
22,26 induced an experimental glaucoma model by injecting microbeads and viscoelastics into C57/BL6, DBA/2J, and CD1 mice, and the smallest extent of IOP elevation was observed in the CD1 mice. Because both SD rats and CD1 mice have white hairs, unlike other strains, we speculated that the amount of pigmentation might affect obstruction at the level of the trabecular meshwork, as observed in pigmentary glaucoma.
47 This might in fact be one clue to ethnic differences in glaucoma manifestation.
48 Further study is necessary to confirm this speculation.
Host immunologic reaction might be another reason for the difference in susceptibility between the internal and external approaches. In contrast to the external approach, intracameral injection of foreign bodies could aggravate inflammation.
27 Further, Kezic et al.
49 showed that anterior chamber cannulation alone could induce microglial activation, whereas the concomitant IOP elevation led to additional Müller cell activation. Therefore, IOP elevation through the internal aqueous outflow blocking approach may be due in part to inflammatory trabeculitis.
10,21 Interestingly, the retinal pigment epithelium has been reported to contribute to the immune and inflammatory response of the eye not only as part of the blood–eye barrier preserving the immune-privileged status but also by being the source and target of inflammatory cytokines.
50 Thus, strains with different pigmentation status may show different immunologic reactions. Although we did not observe any gross anatomical difference in inflammation, the differences among the glaucoma models observed in this study could become a cornerstone of a novel immunologic evaluation of glaucoma pathogenesis.
4
This study has several limitations. First, this study evaluated only RGC density as the result of elevated IOP in each glaucoma animal model. Evaluation of other retinal cells (such as bipolar cells or photoreceptors) would have been helpful in order to quantify which model led to more RGC-specific injury and thus would be more useful as a glaucoma model. Second, despite our success in demonstrating the different clinical profiles of each glaucoma animal model according to strains, we could not determine the exact reason for such differences. Further study would be required to elucidate the downstream molecular pathways of each model. Third, variation of RGC density existed among the glaucoma models. This was somewhat inevitable, as we could not control for the exact IOP status of every subject. Further, cases with neovascular complications had to be excluded from the magnetic bead injection model due to the fact that flat mounting of the retina was not possible. And, because those cases were generally associated with higher IOP, we might have excluded severe cases selectively, thus leading to underestimation of RGC deaths in our models. RGC density would be more informative if, in future work, IOP status also could be incorporated into models for comparative purposes. Fourth, our comparison of RGC density was based on cross-sectional data, not longitudinal data, because we had to sacrifice the rats in order to count RGCs on the retinal flat mounts. In vivo real-time evaluation of RGCs and their function would be necessary to elucidate the individual effect of IOP change over time in each glaucoma model in the future. Fifth and finally, we did not compare the IOP-normalizing treatment outcomes according to the models and strains. Given that the ultimate goal of these models is to develop novel human glaucoma treatment strategies, further study on the treatment outcomes of different models and strains based on the conventional treatment would be helpful.
In conclusion, the magnetic bead and hydrogel injection models were affected by animal strain but the circumlimbal suture model was not. Strains should be considered as an important factor when establishing rodent glaucoma animal models. Our recommendations are as follows: (1) use the magnetic bead injection model with BN rats if a steady increase in IOP is required, as occurs in chronic glaucoma; (2) use the hydrogel injection model with BN rats if ocular imaging is planned; and (3) use the circumlimbal suture model with either BN or SD rats if acute increases of IOP are required.