Epiretinal Amniotic Membrane Influences the Cellular Behavior of Profibrotic Dedifferentiated Cells of Proliferative Vitreoretinopathy In Vitro

Proliferative vitreoretinopathy (PVR) as a rare fbrotic ocular disease is the main reason for failure of retinal detachment surgery and a reduced prognosis following surgery. Amniotic membrane (AM) is a versatile surgical tool for tissue stabilization, antifbrotic properties, and regeneration. Initial clinical case studies now demonstrated intravitreal tolerance as well as good anatomical and functional results for degenerative retinal diseases. Due to its diverse wound healing properties, AM could have promoting, suppressive, or no efects on PVR. To illuminate the potential of epiretinal AM transplantation in complex retinal detachment cases, we investigated its infuence on human primary PVR (hPVR) cells in vitro . In our cell culture study, hPVR cells were isolated from surgically removed PVR membranes. Following incubation with AM for 48h, AM-incubated hPVR showed signifcantly reduced proliferation (BrdU-ELISA; p < 0 . 001), migration (Boyden chamber, scratch assay; p � 0.003 and p < 0 . 001), and cell adhesion ( p � 0.005). Collagen contraction was nearly unafected ( p � 0.04), and toxicity (histone-complexed DNA ELISA, WST-1 assay, and life/dead staining) was excluded. Next, immunofuorescence showed a myofbroblastic phenotype with reduced expression of fbrosis markers in AM-incubated cells, which was confrmed by Western blot analysis. In the proteomics assay, AM signifcantly regulated proteins by a more than 2-fold increase in expression which were related to the cytoskeleton, lipid metabolism, cell-matrix contraction, and protein folding. In conclusion, this in vitro work suggests no induction of fbrosis and other key steps in the pathogenesis of PVR through AM but rather inhibiting properties of profbrotic cell behavior, making it a possible candidate for suppression of PVR. Further clinical studies are necessary to evaluate the therapeutic relevance.


Introduction
Proliferative vitreoretinopathy (PVR) is a severe complication following rhegmatogenous retinal detachment and is associated with a poor visual prognosis [1].Currently, the only available therapeutic option is pars plana vitrectomy, an invasive microsurgical procedure that allows access to the posterior segment of the eye and the vitreoretinal interface through small incisions in the pars plana region of the eye [2].However, it is common for surgery to fail to provide lasting improvement and repeated procedures are often required.Terefore, in addition to surgery, a treatment method that inhibits the pathologic formation of PVR would be desirable [3].
Many diferent ocular tissues and cells are involved, such as retinal pigment epithelial (RPE) cells, Müller cells, and immune cells, which physiologically have critical functions in maintaining retinal homeostasis and immune responses.PVR is the result of retinal tears and detachments and a breakdown of the blood-retinal barrier, retinal hypoxia, and increased intraocular infammation with release of cytokines, growth factors, and chemokines [3].Te cells then undergo mesenchymal transformation to myofbroblasts and form tractional membranes that cause recurrent retinal detachment [4].Amniotic membrane (AM) transplantation is an established and popular treatment option for ocular surface reconstruction.In addition to its good biocompatibility, it has antineovascular, anti-infammatory, antiapoptotic, and immunomodulatory efects that inhibit scar formation and prevent pathologic wound healing [5] and thus promote the restoration, regeneration, and stabilization of a physiologic ocular surface [6].Tese properties make it attractive for intraocular implantation, and its use has been positively demonstrated in several clinical case series for various atrophic vitreoretinal diseases [7][8][9].Tese studies showed promising regenerative efects and stabilization of degenerative retinal tissue [10].
However, compared to atrophic retinal diseases, PVR is caused by a severe fbrotic cell response to retinal detachment [11,12].Despite the diferent pathogenesis, frst intraocular implantations of AM in case series with complex retinal detachments at high risk for PVR development also indicated a benefcial efect [13,14].Other studies showed that AM may promote wound healing through fbroblast proliferation and induction of migration and expression of the extracellular matrix [15].Tus, the exact efects of epiretinal AM application on PVR remain unclear at the cellular and molecular level.Hypothetically, this efect is concentration and interaction dependent; these growth factors may have a dediferentiating and proliferative efect on diferent cell types [5,16,17], and it is possible that they may also promote a fbrotic response depending on the dominant cell type and surrounding tissue environment [18,19].
As an in vitro study on this topic, we aim to determine if AM has the potential to inhibit and attenuate the pathological progression of primary isolated human PVR (hPVR) cells derived from surgically removed PVR membranes in terms of their profbrotic cell behavior, phenotypic changes, and protein expression pattern by immunofuorescence staining, protein expression by Western blotting and proteomics analysis, cell migration, and proliferation analysis.Ultimately, a better pathophysiological understanding of in vitro could lead to a potential clinical application.

Methods
2.1.Surgical Procedure.All cell culture experiments were performed with primary human proliferative vitreoretinopathy (hPVR) cells.Cell isolation was performed as previously described [20,21].Pars plana vitrectomy for rhegmatogenous retinal detachment with concomitant PVR was performed in 4 patients.Membranes were carefully removed using 0.25 mg/ml brilliant blue vital dye (Fluoron GmbH, Neu-Ulm, Germany) with terminal forceps (D.O.R.C, Zuidland, Te Netherlands).Removal of the tractive PVR membranes did not afect the surgical technique, as they must always be removed for the above diagnosis.All patients gave their informed written consent for PVR membrane collection and research.Furthermore, the study was approved by the ethics committees of the Universities of Ulm and Munich (Ethics Committee of the University of Ulm, approval ID: 26420, and Ludwig-Maximilians-University of Munich, approval ID: 47114), and guidelines of the Declaration of Helsinki were followed.

Cell Culture.
No later than three hours after surgical removal, PVR membranes were attached to cell culture plastic (Sarstedt, Nürmbrecht, Germany) with entomological pins (Entomoravia, Slavkov u Brna Czech Republic) under slight tension and suspended in Minimum Essential Medium (MEM, Gibco life technologies, Carlsbad, CA, USA) containing 10% fetal calf serum (Sigma Aldrich, St. Louis, USA).After the hPVR cells had grown from the membrane onto the cell culture plastic, the concentration of fetal calf serum was reduced to 2%, where it remained for all following experiments.Te medium was replaced every other day.

Experimental Setup.
For all experiments, cell culture inserts (Corning, Arizona, USA) with a porous membrane (0.4 µm) were suspended over the cells for 48 hours.Tey either contained cryopreserved fresh AM not suitable for transplantation purposes (Tissue bank of the University of Ulm, Ulm, Germany) or were empty for the control group.If an experiment had to be performed in a 96-well cell culture plate, the AM was added directly to the medium.

Proliferation.
A BrdU cell proliferation ELISA kit (Roche Diagnostics GmbH, Mannheim, Germany) was used.Briefy, cells were incubated in their experimental groups as described above, followed by another 48-hour incubation with medium containing BrdU labeling solution.All further steps followed the manufacturer's protocol exactly.Absorbance was measured by the photometer (Tecan infnite M200 pro, Tecan, Männedorf, Switzerland) at a wavelength of 450 nm and a reference of 690 nm.Te test was performed three times, with each reading recorded twelve times per group.
2.5.Scratch Assay.hPVR cells were cultured to complete confuence and scratched with a 100 µl pipette tip (Brand, Lippstadt, Germany).Subsequently, cell-free areas were documented at 0 and 24 hours using an inverted phasecontrast microscope (Zeiss Axiovert 35, Jena, Germany) and a digital camera (Nikon D31000, Tokio, Japan).ImageJ 1.53 k (National Institutes of Health, Bethesda, Maryland, USA) was used to evaluate the repopulated area as a percentage of the cell-free area at 0 hours.2.6.Boyden Chamber.Cell migration was observed using a Boyden chamber setup.Te upper chamber was a cell culture insert with a membrane (Corning) with pores of 8 µm in diameter.Te insert was suspended into the well of a cell culture plate which acted as the lower chamber.1.1 × 10 5 cells/cm 2 were flled into the top chamber, and medium was added to the setup.Chemotaxis was induced by placing AM in the lower chamber.Cells were able to migrate for 5 hours under standard cell culture conditions with and 2 Journal of Tissue Engineering and Regenerative Medicine without AM.After fxation with methanol at 4 °C for 20 min, a Giemsa Azur-Eosin-Methylene Blue solution (Merck Millipore) was used to make cells visible.Subsequently, all nonmigrated cells from inside the cell culture insert were removed by a cotton bud.Documentation was followed under an inverted phase-contrast microscope with a camera (Zeiss Axiovert 35 and Nikon D31000).Migrated hPVR cells were counted manually using ImageJ 1.53 k (NIH, Bethesda, MD, USA).

Extracellular Matrix Contraction.
A collagen-based cell contraction assay (Cell Biolabs, San Diego, CA, USA) was used according to the manufacturer's instructions.In brief, 500 µl of gel containing 2 × 10 6 hPVR cells was polymerized in a 24-well plate.
After 1 hour, medium was added and the collagen gels were exposed to AM or the control setting.After 2 days, the collagen gels were detached from the edge of the cell culture plate using a spatula.Photo documentation was performed daily starting after the detachment of the gel, and spatial dimensions were determined on the images measuring the surface area using ImageJ 1.53 k (NIH).
2.8.Adhesion.Cells were treated as described above and then subcultured.Successively, a cell concentration of 1 × 10 3 cells/cm 2 was resuspended in medium and photographed 30, 60, and 120 minutes later under the inverted phase-contrast microscope.Already attached cells were counted manually on the images at the diferent time points.
2.9.Immunofuorescence Staining.For immunofuorescence characterization, freshly extracted PVR membranes were fxed in 4% formaldehyde solution (Invitrogen, Waltham, MA, USA) overnight.Following three washes with 0.1 M phosphate-bufered saline (PBS, Termo Fisher Scientifc, Waltham, MA, USA), the membranes were fxed with entomological pins to a cell culture dish and incubated with blocking solution (3% bovine serum albumin, 0.1% triton X 100 in 0.1 M phosphate bufer) for blocking of unspecifc antigen staining.All further steps followed the same protocol as the one used for cell staining.

Liquid Chromatography with Tandem Mass
Spectrometry-Based Proteomic Analysis.A proteome analysis was performed as described previously [22] in collaboration with the Core Unit Mass Spectrometry and Proteomics of the University of Ulm under the direction of Dr. Sebastian Wiese.Samples were prepared as described above, frozen at −80 degrees, and then transferred to liquid nitrogen.After the transport to the Core Unit, cells were washed and then lysed in DIGE bufer (30 mM Tris base, 7 M urea, 2 M thiourea, and pH 8.5).For sample purifcation, all samples were precipitated by methanol/chloroform extraction according to known protocols [23].4 µg of each sample was reduced with 5 mM DTT (AppliChem, Darmstadt, Germany) for 20 min at RT, alkylated with iodoacetamide (Merck Millipore) for 20 min at 37 °C, and diluted with 50 mM ammonium bicarbonate.Trypsin was added at a 1 : 50 enzyme to protein ratio and digested overnight at 37 °C.Samples were measured using an LTQ Orbitrap Elite system (Termo Fisher Scientifc) coupled online to a U3000 RSLCnano (Termo Fisher Scientifc) as previously described [24] with the following modifcations: the column was frst equilibrated in 5% B for 5 min (solvent: A 0.1% FA; B 86% ACN, 0.1% FA).Tis was followed by various elution steps in which the amount of B was frst increased from 5% to 15% in 5 min and then from 15% to 40% B in 145 min.Te 20 most intense ions from the survey scan were selected for CID fragmentation.Singly charged ions were discarded, and the m/z of fragmented ions was excluded from fragmentation for 60 seconds.MS2 spectra were acquired with the LIT at fast scan speeds.Database searches were performed using MaxQuant ver.1.6.3.4 [25].For peptide identifcation, the Andromeda integrated search engine was used to Journal of Tissue Engineering and Regenerative Medicine correlate MS/MS spectra with the UniProt human reference proteome set.Carbamidomethylated cysteine was considered as a fxed modifcation along with oxidation (M) and acetylated protein N termini as variable modifcations.Te false discovery rate was set at 0.01 at both the peptide and protein levels.

Western Blot.
A Western blot was performed as described previously [26].In brief, whole cell extracts were prepared from the pretreated hPVR.For this purpose, the medium and AM were removed, and cells were washed, scrapped of the cell culture plastic, and afterwards centrifuged twice in cold PBS containing magnesium chloride.Te resulting cellular material was suspended in lysis bufer 17 (Bio-Techne, Minneapolis, MN, USA) containing protease inhibitor C 100X Halt cocktail (Termo Fisher Scientifc) and phosphatase inhibitor cocktail (Merck Millipore).Te hPVR cells were shaken on ice for 30 minutes before a 30minute centrifugation.Both the cell extract and the supernatant were frozen at −80 °C.Protein concentration detection was performed using a Pierce BCA Protein Assay Kit (Termo Fisher Scientifc) as detailed by the manufacturer.Following the results of the BCA protein detection, the samples were diluted using lysis bufer before performing gel electrophoresis to ensure that all samples fnally contained an equilibrated amount of total protein.Lysates were diluted in 4x Laemmli Sample Bufer (Bio-Rad, Hercules, CA, USA) containing 100 mM 1,4-dithiothreitol and then separated by electrophoresis in a 4-20% Mini Protean TGX Precast Gel (Bio-Rad) using Tris-Glycin-SDS bufer (Bio-Rad) for 41 min at 200 V and 0.4 A. Afterwards, proteins were transferred by electroblotting to a polyvinylidene difuoride membrane (Bio-Rad) using a PerfectBlue Semi-Dry-Blotter (PEQLAB, VWR, Darmstadt, Germany).A current of 0.08 A was set for 90 minutes for the transfer, and the voltage was limited to a maximum of 90 volts.Next, membranes were cut into pieces according to the desired molecular weights of the following antibodies.After blocking overnight in 0.05% blocking reagent (Roche) in 0.001% Tween-20 (Bio-Rad) in PBS without Mg 2+ and Ca 2+ (Termo Fisher Scientifc), the following primary antibodies were diluted in antibody dilution solution (1 : 4 blocking solution in PBS): alphasmooth muscle actin rabbit antihuman (1 : 2000, Merck Millipore A5228), cytokeratin-8 rabbit antihuman (1 : 10.000, Abcam AB53280), fbronectin rabbit antihuman (1 : 1000, Abcam AB268020), and vimentin rabbit antihuman (1 : 2500, Abcam AB92547).Subsequently, the membranes were incubated on a shaker with the primary antibody for 75 minutes and 30 minutes with the secondary antibodies that were horseradish peroxidase conjugated: mouse IgG solution (1 : 20.000 Biorad 170-5047) and rabbit IgG (1 : 30.000Biorad 170-5046).Pierce ECL Plus Western blotting substrate (Termo Fisher Scientifc) was used as detailed by the manufacturer, and chemiluminescence signals were detected by the imaging system Fusion Pulse TS (Vilbert Lourmat, VWR).To strip the former incubated membranes, membranes were washed several times and incubated with Restore Plus Western blotting stripping bufer (Termo Fisher Scientifc) for 45 minutes.After washing the membranes again, incubation with the primary antibodies actin mouse antihuman (1 : 5000, Novus NBP2-25142), HSP-70 mouse antihuman (1 : 1000, abcam ab5439), alpha tubulin IgG mouse antihuman (1 : 10.000 Abcam, ab7291) followed.All following steps are identical to the ones described above.
2.12.Histone-Complexed DNA ELISA.For the detection of apoptosis in hPVR, an ELISA to detect histone-DNA complexes (Cell Death Detection ELISA, Roche) was performed after 48 hours of incubation with AM in accordance with the manufacturer's recommendations using a cell density of 1,25 × 10 4 cells/cm 2 .Absorbance was measured at a wavelength of 405 nm including a reference at 490 nm by the Tecan infnite M200 pro (Tecan).
2.13.Life/Dead Staining.hPVR cells were seeded onto coverslips.Slides incubated with methanol for 10 min at 4 °C served as a positive control.Treatment was followed by incubation with Hoechst (Invitrogen, Waltham, MA, USA) 1 μg/ml and propidium iodide (Invitrogen) 2 μg/ml in medium for 15 min.After three washing steps, cells were fxed using 4% formaldehyde solution (Invitrogen) for 10 min.Following another three washings steps for 5 minutes each, the slides were mounted on slides using antifade mounting medium (Vectashield, Vector laboratories, Newark, CA, USA).Images were acquired using a fuorescence microscope equipped with a camera (DM4000B, Leica, Wetzlar, Germany).

WST Assay.
A colorimetric dye reduction assay was conducted as detailed by the manufacturer (WST-1, Roche Diagnostics, Mannheim, Germany).Te same experimental settings were used as described above with a cell density of 2.5 × 10 4 cells/cm 2 .WST-1 was added to the cell culture medium for 60 minutes, and subsequently, absorbance was measured at a wavelength of 450 nm and a reference of 690 nm.

Statistical Analysis. Data collection and calculations
were done in EXCEL 365 (Microsoft, Redmond, Washington).Statistical analysis was performed with SPSS 28 (IBM, Armonk, United States) and GraphPad PRISM 9 (GraphPad Software, Inc.San Diego, California).All quantitative assays consisted of either 3 or 4 biological replicates with at least 3 or more technical replicates each.Te results from the technical replicates were used for statistical testing.A two-tailed t-test was used for all assays comparing two groups (BRDU cell proliferation, scratch assay, adhesion, Boyden chamber, extracellular matrix contraction, Western blot, and WST).A t-test was also performed for proteomics.For more than two groups (histone-complexed DNA ELISA), ANOVA with post hoc LSD test was used.Data are presented as the mean and error bars as the standard deviation.P values <0.05 were considered statistically signifcant.

Proliferation.
In PVR formation, stimulation of proliferation of uncontrolled transdiferentiated myofbroblastic cells (hPVR) leads to rapid disease progression.Cell proliferation of hPVR, as measured by the BrdU assay, was statistically signifcantly reduced (p < 0.001; n � 48 technical repeats of 4 independent biological experiments) when AM was added to the culture (Figure 1).

Cell Migration.
PVR causes a pathologic cell migration into the vitreous and other retinal areas after retinal detachment.To quantify migration, two diferent experiments were performed.In the scratch wound healing assay, the area repopulated by AM-exposed cells 24 hours after the induced scratch was signifcantly reduced compared to the control (p < 0.001; n � 40 technical repeats of 4 independent biological experiments).In addition, migration of AMexposed cells was attenuated in a Boyden chamber assay (p � 0.003; n � 28 technical repeats of 4 independent biological experiments) (Figure 2).Tus, both assays are consistent that migration activity is reduced by AM.

Extracellular Matrix Contraction.
Trough contraction and shortening of retinal and fbrotic tissue, PVR leads to the recurrent retinal detachments.Te efect of AM on cell contraction did not play a major role in our experiments.At most time points, AM was unable to afect the contraction of collagen discs populated with hPVR cells compared to the untreated control.A slight statistical signifcance was found at day 4, when the contraction in the AM group was lower than that in the control group (p � 0.04; n � 9 technical repeats of 3 independent biological experiments) (Figures 3(a) and 3(c)).

Cell Adhesion.
Cell adhesion was examined as a measure of the cells ability to adhere to other tissues after proliferation and migration.Te amount of cells adhering to culture plastic after subculture was statistically signifcantly reduced at all time points when exposed to AM compared to the untreated control (p < 0.001 at 30 and 60 min; p � 0.005 at 120 min; n � 32 technical repeats of 4 independent biological experiments) (Figure 3(b)).

Immunofuorescence Staining.
To correlate and characterize cells on PVR membranes and cultured hPVR, immunofuorescence staining for epithelial, glial, fbroblastic, and immunocellular markers was performed.Cells showed staining and morphologic correlation for F-actin, vimentin, fbronectin, and α-SMA in PVR membranes similar to hPVR.Primary PVR membrane cells as well as hPVR showed partial staining but no morphologic correlation for the glial specifc proteins GFAP and CRALBP.Immune cell involvement was examined by detection of CD45 as a marker for leukocytes and Iba1 and CD68 as markers for microglia and macrophage activation.Proteinspecifc staining was detected only for Iba1, which was slightly more abundant in PVR membrane cells than in hPVR (n � 3 biologicals experiments) (Figure 4).
Transdiferentiation of various cell types into myofbroblastic contractile cells has been implicated in the development of PVR.Te hPVR cells used in our study showed specifc immunofuorescence staining for several proteins involved in this process.Fibronectin appeared as foci located around the plasma membrane.Subjectively, these foci were reduced after exposure to AM. Cytokeratin-8, vimentin, and F-actin are proteins of the cytoskeleton.All were detected as intracytoplasmic flaments, suggesting a myofbroblast-like phenotype possibly originating from the retinal pigment epithelium.Vimentin and F-actin were subjectively reduced after AM exposure compared with the controls.Cytokeratin-8, a marker of an epithelial cell phenotype, was equally abundant in controls and AM.All stained markers correlated well with the expected fbrotic response in PVR cells.AM did not lead to fbrosis progression, but rather a slight loss of typical cellular pathological PVR characteristics was observed (n � 3 biologicals experiments) (Figure 5).

Western Blot.
In PVR, cells undergo myofbroblastic transdiferentiation.To determine whether AM has an efect on the myofbroblastic phenotype and to verify the observed changes in immunofuorescence staining, typical protein levels expressed by the hPVR were evaluated by a Western blot assay.Compared to the control, F-actin was signifcantly reduced by 3.3-fold (p � 0.03), α-SMA by 3-fold (p � 0.03), cytokeratin-8 by 3.7-fold (p � 0.03), and vimentin by 1.8fold (p � 0.03) in AM-exposed cells.No statistically significant changes were observed for fbronectin (p � 0.3) and HSP70 (p > 0.99).Tese data suggest that hPVR cells express less myofbroblastic markers when exposed to AM (n � 4 biological independent experiments) (Figure 6 and Supplementary Figure 1).

Proteomics.
To investigate the efect of AM on the hPVR cells, a detailed analysis of phenotypic changes and proteome expression was performed.Tis was achieved by a large-scale proteomic analysis in which changes in protein levels were determined both in the culture medium and in the cells.An untreated control was compared with AM-exposed hPVR cells.Protein expression analysis by label-free/liquid chromatography-tandem mass spectrometry identifed 2,657 proteins.Based on more than 2-fold increase in expression and statistical signifcance, 21 proteins were upregulated in AM-treated cells and 17 were observed exclusively in this group (Figure 7).Te signifcantly regulated proteins are mainly related to cytoskeleton, lipid metabolism, cell-matrix contraction, and protein folding (n � 4 biological independent experiments) (Table 1).

Cell Viability.
To exclude the possibility that the results of this study were caused by toxic efects of AM on hPVR, three assays were performed.In the WST assay, the viability of AM-exposed cells was not signifcantly reduced compared to the control (p � 0. independent biological experiments).Tere was also no increase in dead cells in a dye exclusion assay.As a positive control, methanol incubation signifcantly increased the number of dead cells (n � 3 technical repeats of 3 independent biological experiments).In contrast, however, histone-complexed DNA fragments were signifcantly higher in AM-treated cells (p � 0.01).In this experiment, we were unable to determine whether this increase was due to the AM possibly undergoing apoptosis during preparation and freezing, as apoptosis was not increased in AM alone compared to medium alone (p � 0.7).Tere was also no signifcant increase in apoptosis between medium alone and hPVR alone (p � 0.2) and no signifcant increase between AM and AM plus hPVR cells (p � 0.2; n � 24 technical repeats for hPVR and AM + hPVR; n � 6 technical repeats for AM only; and n � 3 for medium only, all of 3 independent biological experiments) (Figure 8).

Discussion
In PVR, fbrotic contractile cells invade at various sites, migrate, proliferate, and produce extracellular matrix to form membranes that can shorten and detach the retina [27].Several cell phenotypes have been identifed as precursors of these transdiferentiated myofbroblasts.Tese include, for example, retinal pigment epithelial cells, glial cells, and immune cells.Tey are activated and modulated by various growth factors and cytokines [28].Although AM has been shown to be a highly versatile surgical tool, its mode of action on the proposed pathologic basis of PVR in potential intravitreal implantation has been unclear, as it may promote either an activating, suppressive, or no efect on PVR progression.In the primary patient-derived PVR myofbroblasts (hPVR) used for this study, we demonstrated an inhibitory and regenerative efect of AM on PVR.After incubation with AM, hPVR showed signifcantly reduced fbrotic properties as determined by migration, proliferation, and cell adhesion, which may be attributed to the stabilizing efect of AM.Immunofuorescence showed a reduced expression of the fbrosis markers, which was confrmed by our Western blot.Proteomics also revealed signifcant changes in potentially fbrosis-inducing mediators in hPVR, mainly related to the cytoskeleton, lipid metabolism, extracellular matrix contraction, and protein folding.Tis in vitro work suggests PVR-inhibiting properties of AM or at least negates the induction of fbrosis progression that would make its use contraindicated in other vitreoretinal diseases.
Previous clinical fndings from several case series are consistent with our laboratory results.Treatment of persistent macular hole by an AM patch transplanted during pars plana vitrectomy is well known [29].However, few clinical case reports have been published on the use of AM for retinal detachment.Te use of intraocular AM in complicated retinal detachment cases has been reported in small numbers by various groups [7,30].It seems that AM was well tolerated and the cases demonstrated an acceptable clinical course [7,[30][31][32].
Although early clinical data show theoretical benefts and our results suggest no contraindication for a clinical application of AM in retinal diseases, the intraocular use of AM is still far from clinical routine.For PVR, pars plana vitrectomy with careful membrane removal, retinal reattachment, and vitreous tamponade will continue to be the treatment of choice [3].However, epiretinal AM may be a tool to improve the rate of postoperative PVR in selected cases.Te mode of application is still unclear; whole membrane implantation [7][8][9] or AM secretomes [33] have been described.In the case of secretomes, application in the form of a single or repeated intravitreal injection would be conceivable.Mechanical properties of epiretinal AM may also play a role in the PVR suppression [34,35].Terefore, we would most likely opt for this option.
As from our perspective, it is not yet clear whether explantation of M is necessary, and it is unclear at what time explantation should be initiated.After the release of the biologically active factors, AM is probably no longer useful or could be visually disturbing and have adverse efects if it foats in the eye.A possible suggestion would be to remove the membrane at the time of a possible silicone-oil tamponade removal, which is often used in these severe PVR cases that could also beneft from AM transplantation.
In line with our results, the main efects of this application were further investigated in vitro.RPE cells undergo epithelial-mesenchymal transformation as part of PVR, which could be prevented by the phenotype-stabilizing efect of AM [36].In addition, RPE cells have been shown to form dense colonies on AM that maintain their epithelial phenotype [37,38].In vivo studies in mice and rats [39] suggest a neuroprotective efect against damage to the peripheral and central nervous system, which would include the retina.In rats sufering from uveitis, subretinal AM transplantation attenuated the infammatory response and contributed to the preservation of retinal structure [40].
Immunofuorescence, proteomics, and Western blot analysis suggest that the main fndings observed may be due to extensive changes in the cytoskeleton.Te cytoskeleton is a dynamic network that is constantly remodeled by many diferent stimuli and is involved in a variety of cellular functions.Alterations in the cytoskeleton have been implicated in the development of PVR.For example, a complete proteomic analysis of vitreous samples obtained during pars plana vitrectomy in PVR eyes revealed alterations of several cytoskeleton-associated proteins [41].Epithelialmesenchymal transformation has been identifed as a hallmark of PVR [42], and alterations in the actin cytoskeleton, particularly upregulation of the protein α-SMA, have been widely used as markers of epithelial-mesenchymal transformation [43].
Consistent with other fndings in the literature and as seen in the proteomic analysis, in addition to cytoskeletonrelated proteins, proteins involved in lipid metabolism and other metabolic pathways were also afected by AM.Simvastatin, a 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitor, belongs to the statin group and is commonly used to treat hypercholesterolemia.It reduced PVR formation in the rabbit model [44] and in human RPE Journal of Tissue Engineering and Regenerative Medicine cells [45,46].Clinical evidence of a benefcial efect against PVR after retinal detachment has also been found.Concentrations of several growth factors in the vitreous of statin-treated retinal detachment patients were lower after vitrectomy, and an improvement in visual acuity was observed [47].In a Finnish population-based cohort study of vitrectomized patients with rhegmatogenous retinal detachment, the risk of repeat vitrectomy was 28% lower when systemic statins were used [48].Te clinical fndings and observations of protein regulation in lipid metabolism and other metabolic pathways associated with a PVR reduction should be investigated in further studies.
It should be noted that the experimental data of this study are limited to the nature of an in vitro setting.Te hPVR cells used are a very simplifed model of PVR and do not fully represent the pathologic basis of PVR, as the retina is a complex tissue with glial cells, neuronal tissue, immune responses, and blood vessels.Te cell culture environment also does not fully represent the conditions in the eye after retinal detachment and breakdown of the aqueous humor barrier.Terefore, the hPVR model does not guarantee complete translation to the clinical situation.Other limitations include the lack of ocular pharmacokinetics, which may result in lower or higher concentrations of cytokines, growth factors, and extracellular matrix components.All of these limitations may have led to an overestimation or underestimation of the observed efects compared to the clinical situation, and future studies need to verify our results.
In conclusion, this in vitro work suggests possible PVRinhibitory properties of AM, while it does not show induction of fbrosis progression that would make its use contraindicated in the applications discussed.Further studies are needed to assess the clinical therapeutic relevance.

Figure 1 :Figure 2 Figure 3 Figure 2 :
Figure 1: Addition of AM resulted in a statistically signifcant decrease in cell proliferation compared to the unexposed control in the BrdU assay ( * * * p < 0.001; n � 48 technical repeats of 4 independent biological experiments).

Figure 3 :
Figure 3: While collagen contraction was largely unafected, (a) cell adhesion of hPVR was signifcantly reduced by AM (b), with a statistically signifcant reduction in adhesion being observed at all time points (p < 0.001 at 30 and 60 min; p � 0.005 at 120 min).At most time points, AM was unable to prevent contraction of collagen discs populated by hPVR cells compared to the untreated control.Shown are representative examples of the collagen slices, including size as measured by circular area projection.(c) * p � 0.04; * * p � 0.005; * * * p < 0.001; collagen contraction: n � 9 technical repeats for AM and n � 6 for co of 3 independent biological experiments; cell adhesion: n � 32 technical repeats of 4 independent biological experiments.

Table 1 :
Classifcation of x-fold regulation and signifcances of regulated proteins.