Intraocular pressure elevation precedes a phagocytosis decline in a model of pigmentary glaucoma

Background: Outflow regulation and phagocytosis are key functions of the trabecular meshwork (TM), but it is not clear how the two are related in secondary open angle glaucomas characterized by an increased particle load. We hypothesized that diminished TM phagocytosis is not the primary cause of early ocular hypertension and recreated pigment dispersion in a porcine ex vivo model. Methods: Sixteen porcine anterior chamber cultures received a continuous infusion of pigment granules (Pg), while 16 additional anterior chambers served as controls (C). Pressure transducers recorded the intraocular pressure (IOP). The phagocytic capacity of the trabecular meshwork was determined by fluorescent microspheres. Results: The baseline IOPs in Pg and C were similar ( P=0.82). A significant IOP elevation occurred in Pg at 48, 120, and 180 hours (all P<0.01, compared to baseline). The pigment did not cause a reduction in TM phagocytosis at 48 hours when the earliest IOP elevation occurred, but at 120 hours onward ( P=0.001 compared to C). This reduction did not result in an additional IOP increase at 120 or 180 hours compared to the first IOP elevation at 48 hours ( P>0.05). Conclusions: In this porcine model of pigmentary glaucoma, an IOP elevation occurs much earlier than when phagocytosis fails, suggesting that two separate mechanisms might be at work.


Introduction
The conventional outflow is guarded by the trabecular meshwork (TM), a complex three dimensional, layered tissue that contains variable amounts of extracellular matrix (ECM) 1 . The aqueous passes from the anterior chamber into Schlemm's canal (SC) by entering first the uveal TM (UTM), the corneoscleral TM (CTM) and finally the juxtacanalicular TM (JCT) 2 . The JCT contains proteoglycans and hyaluronans and presents the aqueous humor with an increasingly tighter fluid passageway towards the SC in a process referred to as funneling 3 . The aqueous eventually passes the inner wall of SC endothelium mainly by two different mechanisms, a paracellular route in between endothelial cells and a transcellular route 4 consisting of intracellular pores and giant vacuoles that are time and pressure dependent 1,5,6 . Failure of the TM and the inner wall of SC endothelial cells to maintain homeostasis and a normal cytoskeleton and can cause ocular hypertension 7 . For instance, pigment dispersion 8 and corticosteroids can increase and contract actin stress fibers and result in an elevation of the intraocular pressure (IOP) 8,9 . Conversely, relaxing the cytoskeleton with Rho kinase inhibitors can reverse these effects 10,11 . Phagocytosis of debris is another function of TM cells 1 . A chronic phagocytosis demand in the form of pigment 12 , erythrocyte-derived ghost cells 13 , inflammatory cells 14 , photoreceptor outer segments 15 , lens and pseudoexfoliation material 16,17 can all lead to secondary glaucomas even though the amount of material itself is unlikely to cause a physical outflow obstruction. Although these glaucomas make for a sizable fraction of open angle glaucomas, it was difficult to study the cellular mechanism that leads to an IOP elevation.
We recently developed a porcine ex vivo pigmentary glaucoma (PG) model that recreates the IOP elevation, stress fiber formation, and phagocytosis reduction that are characteristic of human PG 8 . A gene expression analysis indicated an activation of the RhoA signaling pathway and a downstream effect of tight junction formation negatively regulated by RhoA-mediated actin cytoskeletal reorganization 8 . In the current study, we hypothesized that early ocular hypertension from pigment dispersion is the result of actin cytoskeletal changes and occurs before phagocytosis declines.

Pig eye perfusion culture and pigmentary glaucoma model
This study was conducted in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research. Because no live vertebrate animals were used and pig eyes were acquired from a local abattoir (Thoma Meat Market, Saxonburg, PA), no Institutional Animal Care and Use approval was required.
Thirty-two porcine eyes were cultured within 2 hours of enucleation. Extraocular tissues were removed, and the eyes were decontaminated with 5% povidone-iodine solution (CAT# 3955-16, United States Pharmacopeia, Rockville, MD) for two minutes and washed three times in phosphate buffered saline (PBS). Posterior segments, lenses, and irises were removed and the anterior segments with intact TM mounted in the perfusion system as previously described 8,18,19 . We used the same method to generate pigment granules as recently described in a model of pigmentary glaucoma (PG) 8 . Briefly, pigment granules were produced by subjecting the iris to freezethaw and resuspension washing before dilution of the stock to a final concentration of 1.67×10 7 particles/ml. Eyes in the pigment dispersion group were continuously perfused with pigment added to the culture medium for up to 180 hours (Pg) and compared to controls (C). The perfusate consisted of Dulbecco's modified Eagle media (DMEM, SH30284, HyClone, GE Healthcare, UK) supplemented with 1% FBS and 1% antibiotics (15240062, Thermo Fisher Scientific, Waltham, MA) at a constant rate of 3 µl/min using a microinfusion pump (PHD 22/2000; Harvard Apparatus, Holliston, MA). IOP was measured intracamerally by a pressure transducer (SP844; MEMSCAP, Skoppum, Norway) and recorded at two-minute intervals (LabChart, ADInstruments, Colorado Springs, CO). Baseline IOPs were obtained after IOP stabilization for 48 hours.

TM phagocytosis
The in situ TM phagocytosis was measured using an epifluorescence microscope after microsphere perfusion. In brief, a suspension of 0.5 µm carboxylate-modified yellow-green fluorescent microspheres 20 (CAT# F8813, Thermo Fisher, Waltham, MA) at 5×10 8 particles/ml was added to the perfusate at 48, 120, and 180 hours and perfused for 24 hours. The eyes were removed from their perfusion dishes, washed three times with prewarmed PBS, secured again in the perfusion dishes, and placed upside down for imaging. The TM, visualized from the underside of the transparent perfusion dish, was photographed and measured by acquiring the images with a camera and epifluorescence equipped dissecting fluorescence microscope (SZX16, Olympus, Tokyo, Japan) at a 680×510 pixel resolution and a 200 ms exposure. The mean fluorescence intensity was quantified by ImageJ (Version 1.50i, NIH) as previously described 21 at 48, 120, and 180 hours by measuring the fluorescence intensity in the TM.
To validate that the microspheres were phagocytosed by TM cells, the TM was dissected and digested with collagenase type IA (C9891, Sigma Aldrich, St. Louis, MO) at 2mg/ml and 1% FBS for 30 min at room temperature. The cells were filtered with a 70-micron cell strainer and resuspended in 0.5 ml of PBS. The percentage of TM cells that had ingested fluorescent microspheres was determined using flow cytometry.

Amendments from Version 1
We have modified the Introduction to include more details about conventional outflow mechanisms. We emphasize more clearly in the Introduction and Discussion that we do not think that pigmentary debris causes a physical outflow obstruction.

REVISED
To get a more accurate visualization of the phagocytosed microbeads, we used confocal microscopy. TM cells were seeded into the wells of a six-well plate and fixed with 4% PFA. The cell membranes were labeled with Lycopersicon esculentum agglutinin (TL; Texas red-conjugated; #TL-1176, Vector Laboratories, Inc., Burlingame, CA) at room temperature for 1 hour. The cell nuclei were counterstained with DAPI (D1306, Thermo Fisher Scientific, Waltham, MA). Photos and 3D videos were taken using an upright laser scanning confocal at 400x magnification (BX61, Olympus, Tokyo, Japan).

Histology
After the TM phagocytosis assay, the anterior segments were fixed with 4% PFA for 24 hours, washed three times with PBS, dehydrated in 70% ethanol, and embedded in paraffin. Sections were cut to a thickness of 5 µm and stained with hematoxylin and eosin (H&E).

Statistics
Data were presented as the mean ± standard error and analyzed by PASW Statistics 18 (SPSS Inc., Chicago, IL). The baseline IOP was compared to the other time points of the same eye using a paired t-test. Other quantitative data were analyzed by one-way ANOVA. A p value ≤ 0.05 was considered statistically significant.

Results
In H&E stained tissue sections, normal TM ( Figure 1A) presented as a sparsely pigmented (red arrowheads), multilayered, porous tissue with Schlemm's canal-like segments within the aqueous plexus at the outer layer (black arrows). Pigment granules were seen phagocytosed by trabecular meshwork cells, particularly in the uveal TM, at 48, 120, and 180 hours ( Figure 1B, C and D) but were not dense enough to physically obstruct any part of the conventional outflow system. Baseline IOP in Pg was comparable to C (12.2±0.9 mmHg vs. 11.9±0.9 mmHg, P=0.82). Pigment dispersion caused a significant IOP elevation at 48, 120, and 180 hours ( The microsphere ingestion by TM cells was further assessed by flow cytometry and confocal microscopy. 28.1% of TM cells had phagocytosed microbeads in a normal perfusion eye ( Figure 3A) and the confocal microscopy confirmed them as being located within the cells ( Figure 3B) aided by tomato lectin-stained cell membranes and DAPI-stained nuclei. Confocal imaging showed clusters of green fluorescent microspheres within the intracellular space with no microspheres in the intercellular space. The 3D video also suggested the microspheres were in fact phagocytized and not merely on top of or below them since the microbeads were in the same z plane as the cells (Supplementary Video 1).   To further confirm that microspheres were phagocytosed, we digested a normal sample TM tissue into single cell suspension and sent for flow cytometry. The results suggested that 28.1% of the TM cells were actively phagocytic (A). We then seeded these cells into a six well plate to form monolayer. After labeling them with tomato lectin, the confocal imaging showed that clusters of green fluorescent microspheres were located in the intracellular but not in the intercellular space (B).

Discussion
Phagocytosis is a defining feature of TM cells 22 and plays a central but poorly understood role in the pathogenesis of secondary glaucomas caused by particulate matter such as erythrocytes and ghost cells, inflammatory cells, outer photoreceptor segments, lens and pseudoexfoliation material and pigmentary debris 8,[23][24][25] . These may interfere with various intracellular processes resulting in a gradual deterioration of outflow 23,26 . Although TM phagocytosis can remove particles from the aqueous humor 27 , the direct, short-term effects on outflow regulation remain insufficiently explained 7 . In this study, we measured IOP and TM phagocytic activity in the presence of pigment granules at different time points and found IOP was significantly elevated as early as 48 hours after exposure to pigment granules. This was contrasted with a phagocytic activity in Pg not different from C before the decrease at 120 and 180 hours. A worsening decline of TM phagocytosis at 120 and 180 hours did not result in a further increase of IOP. This suggests that reduction in phagocytosis is a downstream and secondary effect of actin cytoskeletal reorganization.
Past investigations suggested that the outflow obstruction was, in fact, physical 12,28 but newer studies indicated that ocular hypertension is in part caused indirectly by reorganization of the TM actin cytoskeleton 29 . This occurs at a concentration far lower (10,000-fold) 8 than that used in previous bolus experiments 12 . Consistent with Zhou et al. 29 , we recently reported that long, thick, and continuous TM actin bundles emerge as early as 24 hours after pigment exposure 8 and replicate this observation in the present study. Histological characteristics of pigment dispersion in porcine eyes matched those seen in samples from pigmentary glaucoma patients 28,30,31 showing that pigment particles were taken up by TM cells.
In summary, the results confirm that the IOP elevation caused by pigment dispersion is not the direct result of physical obstruction of outflow or a chronically overwhelmed phagocytosis. The reduction in phagocytosis considerably lags the evolving hypertension supporting the notion that these cytoskeletal changes occur early on and are separate from the impact of pigment on canonical phagocytosis pathways 8 .

Data availability
All the raw data generated or analyzed in this study are included in following datasets. I have no major criticisms and suggest the paper be indexed as is.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
Referee Expertise: Management of complex glaucoma, optic nerve injury in glaucoma I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. The title of the article is very stimulating but the content has several et al. things that should be corrected. The authors state that in the conventional way of outflow the aqueous passes into Schlemm's canal by paracytosis or giant vacuoles. This is a partial view that does not correspond to the truth. I invite the authors to read Saccà et al. The Outflow Pathway: A Tissue With Morphological and Functional Unity . Then the effects of the inhibitors of the Rho kinase inhibitor may have the ability to modify the metabolism of endothelial cells in the node in which the authors think it is open to question. These drugs have the ability to block TM motility by exposing more cells to aqueous humor. This is the reason why we initially witness a decline in IOP then this worsens because the exposed cells are blocked and are more exposed to apoptosis (to read Saccà from DNA damage to et al. functional changes of the trabecular meshwork in aging and glaucoma ). Also the hypothesis formulated by the authors to explain the pigmentary glaucoma is partial in the Trabecolato the so-called pores have a very low meaning and above all the pigment does not obstruct anything. We do not work like sinks -I am amazed that today we can still believe that TM is a porous tissue, and above all, how it is possible to think that outflow is a passive phenomenon. In pigmentary glaucoma there are 2 fundamental pathogenic moments. The first concerns the back sheet of the iris, where the pigment comes from (a defect that involves the loss of the pigment itself) the second point concerns autophagy which is not able to ensure adequate cell homeostasis. So these cells first enter suffering and then die and hence glaucoma is why the cells do not work, not because it obstructs something. There are cells, not tubes. The cytoskeleton in all this has little to do with it: pigment granules are only a metabolic burden that in the case of pigment dispersal syndrome is absorbed by a functioning autophagy, while in pigment glaucoma, autophagy does 1 2 dispersal syndrome is absorbed by a functioning autophagy, while in pigment glaucoma, autophagy does not work. Finally, I remember that the changes in the cytoskeleton occur in all types of glaucoma. I do not believe that this article can be indexed in this form and needs a remake of both the introduction and the discussion and the results should be reviewed in the light of a more modern interpretation of the physio-pathogenical events.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes We thank the reviewer, Professor Saccà, for having taken the time to carefully review our Authors: manuscript. We hope that we are able to address all of them with extensive manuscript changes as requested. requested.

Reviewer 1:
The authors state that in the conventional way of outflow the aqueous passes into Schlemm's 1) canal by paracytosis or giant vacuoles. This is a partial view that does not correspond to the truth. I invite the authors to read Saccà et al. The Outflow Pathway: A Tissue With Morphological and Functional Unity .
1 Thank you for bringing this review article to our attention. We are pleased to use this Authors: publication and your other article as additional references. We have modified the Introduction to include more comprehensive details about conventional outflow better aligned with Saccà et al. We would like to change the less common phrase "paracytosis" with "paracellular passage through pores" to distinguish it from the active process of pathogen intrusion through epithelia e.g. by haemophilus. Overall, it appears our manuscript matches Saccà et al well who state that the main 1 mechanisms of aqueous outflow are "a paracellular route [...] and a transcellular pathway". We have added to the Introduction the description of a transcellular pathway through intracellular pores with inducible pore density that depends on outflow demand . . Conversely, relaxing the cytoskeleton with Rho 11,12 kinase inhibitors can reverse these effects . Phagocytosis of debris is another function of TM 13,14 cells . A chronic phagocytosis demand in the form of pigment , erythrocyte-derived ghost cells 8 15 16 , inflammatory cells , photoreceptor outer segments , lens and pseudoexfoliation material 17 18 19,20 can all lead to secondary glaucomas even though the amount of material itself is unlikely to cause a physical outflow obstruction. Although these glaucomas make for a sizable fraction of open angle glaucomas, it was difficult to study the cellular mechanism that leads to an IOP elevation." Then the effects of the inhibitors of the Rho kinase inhibitor may have the ability to modify the 2) metabolism of endothelial cells in the node in which the authors think it is open to question. These drugs have the ability to block TM motility by exposing more cells to aqueous humor. This is the reason why we initially witness a decline in IOP then this worsens because the exposed cells are blocked and are more exposed to apoptosis (to read Saccà et al. from DNA damage to functional changes of the trabecular meshwork in aging and glaucoma ).
2 We completely agree with the Reviewer. We would like to point out that -consistent with Authors: Sacca et al -we wrote that "we hypothesized that ocular hypertension is the result of a reorganization of the actin cytoskeleton and occurs before phagocytosis declines." We did not intend to describe outflow through the TM as sink-like. We hope the expanded Introduction makes it more obvious that this is a paracellular and transcellular process. In addition, we specifically stated before that "material itself is unlikely to cause a physical outflow obstruction" (Introduction), that granules "were not dense enough to physically obstruct any part of the conventional outflow system" (Results) and discuss that this had already been shown before: "Pigment treatment has previously been shown to cause ocular hypertension in part by reorganizing the TM actin cytoskeleton and not by physical obstruction of the outflow tract ." 15,21 (Discussion).
(Discussion). We now clarify in the Discussion that "Past investigations suggested that the outflow obstruction was, in fact, physical but newer studies indicated that ocular hypertension is in part caused 15,22 indirectly by reorganization of the TM actin cytoskeleton . Consistent with Zhou et al. , we 23 23 recently reported that long, thick, and continuous TM actin bundles emerge as early as 24 hours after pigment exposure and replicate this observation in the present study. " 11 Also the hypothesis formulated by the authors to explain the pigmentary glaucoma is partial in 3) the Trabecolato the so-called pores have a very low meaning and above all the pigment does not obstruct anything. We do not work like sinks -I am amazed that today we can still believe that TM is a porous tissue, and above all, how it is possible to think that outflow is a passive phenomenon.
We did not state anything about an outflow obstruction or a location in our hypothesis.

Authors:
We wrote, "In the current study, we hypothesized that ocular hypertension is the result of a reorganization of the actin cytoskeleton and occurs before phagocytosis declines." As we detail in our Response to 3), we never wrote that the TM functions as a sink. On the contrary, we elaborate multiple times that there is no physical obstruction as detailed in our Response to 2). Please see our response above where we hope to clarify this point in the rewritten manuscript.
In pigmentary glaucoma there are 2 fundamental pathogenic moments. The first concerns the 4) back sheet of the iris, where the pigment comes from (a defect that involves the loss of the pigment itself) the second point concerns autophagy which is not able to ensure adequate cell homeostasis. So these cells first enter suffering and then die and hence glaucoma is why the cells do not work, not because it obstructs something. There are cells, not tubes.
We completely agree as detailed in 1)-3) and have expanded our manuscript to make Authors: this more obvious.
The cytoskeleton in all this has little to do with it: pigment granules are only a metabolic burden 5) that in the case of pigment dispersal syndrome is absorbed by a functioning autophagy, while in pigment glaucoma, autophagy does not work. Finally, I remember that the changes in the cytoskeleton occur in all types of glaucoma. I do not believe that this article can be indexed in this form and needs a remake of both the introduction and the discussion and the results should be reviewed in the light of a more modern interpretation of the physio-pathogenical events.
We are unsure why the Reviewer states that cytoskeletal changes have nothing to do Authors: with an IOP elevation after a phagocytosis challenge. This is rather well established: Zhou et al found that 4 h after phagocytosis, the cytoskeletal structure in trabecular meshwork cells was disrupted . As we detail in the Introduction "pigment dispersion and corticosteroids can 23 11 increase and contract actin stress fibers and result in an elevation of the intraocular pressure (IOP) . Conversely, relaxing the cytoskeleton with Rho kinase inhibitors can reverse these effects 11,12 ." 13,14 The new finding described in this manuscript is that the IOP rises when the cytoskeleton changes occur as a result form pigment exposure but before phagocytosis starts to fail. This has not been observed before because an appropriate ex vivo anterior segment model of pigment dispersion did not exist. Our recent studies show that ex vivo perfused pig eyes experience reduced outflow in 11,24,25 response to continuous exposure to pigment at a concentration far lower (10,000-fold) than that used in previous bolus experiments and that severe cytoskeletal changes are associated with 26 this. Conversely, we show that rho-kinase inhibitors normalize the cytoskeleton and the phagocytosis . TM regulates aqueous outflow by changing its cytoskeleton, stiffness, cell 24 adhesion, migration, contraction, and phagocytosis , in which ROCK signaling plays a central 10,27 role . There is a close interaction between actin cytoskeleton reorganization, cell or 28,29 role . There is a close interaction between actin cytoskeleton reorganization, cell or 28,29 extracellular matrix stiffness and aqueous outflow facility . Only a small proportion of stress 30,31 fiber formation was found in the normal TM culture while pigment exposure increased it by 2.06-fold . 11 Lastly, we caution not to confuse TM cell phagocytosis of pigmented debris with autophagy. To address the Reviewer's concern we have added to the Discussion his own reference and state 32 that "An increased pigmentary debris may interfere with many intracellular functions, including important autophagy functions that can cause a gradual deterioration of TM cell function ." 32,33