Extent of Detached Retina and Lens Status Influence Intravitreal Protein Expression in Rhegmatogenous Retinal Detachment

Abstract

Purpose: The aim of the study was to compare intravitreal cytokines and chemokines to clinical parameters in patients with rhegmatogenous retinal detachment (RRD).

Methods: In this prospective study vitreous samples were taken undiluted from 60 patients with RRD and 20 age-matched controls with idiopathic epiretinal membranes at the beginning of primary vitrectomy. The following clinical parameters were assessed from RRD patients prior to surgery: number of quadrants detached, RD height, lens status, symptom duration, and refractive power. Concentrations of 40 different proteins in the vitreous of RRD eyes were measured by multiplex protein array, compared with controls and correlated to clinical parameters.

Results: Ten cytokines and chemokines were significantly upregulated in the vitreous of RRD eyes compared with controls (tissue inhibitors of metalloproteinases [TIMP]-1 and -2, macrophage inflammatory protein [MIP]-1α, monocyte chemoattractant protein [MCP]-1, IL-6, and -8, inducible protein (IP)-10, brain-derived neurotrophic factor [BDNF], TGFβ-3, and platelet-derived growth factor [PDGF]-AB/BB). Linear regression analysis revealed that IL-8 and TGFβ-3 increased with the number of retinal quadrants detached, while TIMP-1 rose in eyes with greater RD heights. Concentrations of IP-10 and myeloperoxidase (MPO) peaked in eyes with two or more quadrants detached, while TIMP-2 was highest expressed in the vitreous of eyes with great RD height. In pseudophakic eyes with higher detachment height levels of vascular cell adhesion molecule (VCAM)-1 were significantly increased, while neural cell adhesion molecule (NCAM) was decreased in pseudophakic patients with shallow RD height.

Conclusions: Extent of RRD and lens status significantly influence intravitreal proinflammatory, profibrotic, and proapoptotic protein expression. These data contribute to the fundamental understanding of pathophysiological mechanisms in RRD and may serve as a basis for development of adjunct therapeutics to facilitate functional restoration.

Retinal detachment (RD) is a severe ocular disorder with an incidence of 12.4 per 100,000 people per year. It is characterized by the separation of the neurosensory retina, including photoreceptors from the underlying RPE.¹ Retinal detachment can be differentiated into a rhegmatogenous, tractional, and exudative form with the most common type being rhegmatogenous. In rhegmatogenous RD (RRD), liquefied vitreous reaches the space under the neurosensory retina via a retinal break resulting in splitting and detachment of the retinal layers from the underlying RPE. This exposes intraretinal cells including the RPE to liquefied vitreous rich in cytokines and chemokines. Detailed mechanisms leading to retinal dysfunction after RD are still not fully elucidated. The separation of the neurosensory retina from the RPE may result in severe visual loss due to degenerative responses in various retinal cell types ultimately resulting in “deconstruction” or death of photoreceptors by disrupted transport of nutrients and metabolic waste products between the two layers.² Animal studies have shown a degeneration of photoreceptor outer segments once detached from the underlying RPE with variable cone recovery rates after reattachment.^3,4 Moreover, retraction of photoreceptor axons and terminals toward the cell body occurs within hours after experimentally induced detachment of neural retina from the RPE.^5–7 

In the human retina, cell death after detachment of photoreceptors occurs mainly by apoptosis.⁸ Apoptotic processes within the retina are present for as long as 1 month after detachment was experimentally induced in cats.⁹ Retinal detachment stimulates the proliferation of Müller cells, astrocytes, microglia, and invading macrophages within different layers of the retina.¹⁰ Many of these pathological cellular processes can be halted or reversed once the retinal layers reattach again. 

It is hypothesized that the various cells affected during RD secrete factors involved in the destruction but also survival of retinal structures. Kaufman and coworkers¹¹ were among the first to report that macular involvement and duration of the detachment were decisive parameters for postoperative visual acuity. During the following decades, clinicians attempted to identify the impact of duration of macular detachment not affecting final visual acuity. According to clinical studies surgical repair of RD is best to be performed within 10 days after onset of symptoms for not negatively affecting postoperative final visual acuity.^12–18 Nevertheless, despite anatomically successful RD surgery resulting in reattached retinal layers visual acuity remains impaired in approximately 40% of cases, in particular when the macula was detached or proliferative vitreoretinopathy (PVR) developed after surgery.¹⁹ Two components found in the serum, fibronectin, and platelet-derived growth factor (PDGF), have been identified in rabbit eyes to cause tractional RD by stimulating cellular migration.²⁰ Clinical evidence evaluating intraocular cytokine/chemokine levels and clinical parameters in patients with RRD is sparse. Yet, this association is of uttermost clinical importance for understanding the pathophysiology of RRD, the potential of retinal recovery and complications following intervention. 

We therefore proposed a prospective clinical study to correlate intravitreal cytokine and chemokine expression levels to various clinical parameters in eyes with RRD. 

Methods

The study was designed as a prospective, nonrandomized, open-clinical study and performed at the Department of Ophthalmology and Optometry at the Medical University of Vienna (Vienna, Austria). All research and measurements adhered to the tenets of the Helsinki agreement, the local ethics committee approved the study and informed consent was obtained from all individuals after a detailed discussion of the nature and possible consequences of the study procedures. 

Patients

Sixty patients with RRD and 20 control patients with idiopathic epiretinal membranes were included. The number of study patients was determined based on previously published levels of cytokines and chemokines in similar study populations. 

Patients with RRD scheduled for primary vitrectomy performed by one of three surgeons (MG, SS, USE) were recruited in a consecutive way. 

Inclusion and Exclusion Criteria

Main inclusion criteria for the RD group was detachment of the retina with early signs of proliferative vitreoretinopathy (PVR Grade A or B). Main exclusion criteria were previous intraocular surgery (including intravitreal injections) except cataract surgery (minimum 3 months ago), and the presence of any other confounding ocular disease except RD or cataract. 

Main inclusion criteria for the control group was planned surgery for the removal of idiopathic epiretinal membrane. Exclusion criteria were any previous intraocular surgery except phacoemulsification or concurrent ocular disease except cataract. 

Assessment of Clinical Parameters

The following clinical parameters were evaluated in RRD patients: number of quadrants detached based on the preoperative retinal evaluation; retinal detachment height (flat, high) as assessed by two experienced vitreoretinal surgeons (high detachment defined by extensive retinal convexity); lens status (phakic, pseudophakic); symptom duration (days) and refractive power (emmetropic, hyperopic, myopic <6 diopters [D], myopic ≥ 6 D). Details are listed in Table 1. 

Table 1

 

Clinical Parameters in Eyes With RRD

 

In the control group refractive and lens status was evaluated besides demographic data. 

Measurement of Cytokines and Chemokines in Vitreous Samples

Vitreous fluids were collected undiluted from patients undergoing primary pars plana vitrectomy at the beginning of the intervention. Vitreous samples were taken through a one-port system by 23-G vitrectomy and minimal suction directly into a 2-mL syringe. A minimum volume of 200 μL was obtained from each patient and stored immediately at −80°C. 

Samples were analyzed using a multiplex protein array (Luminex xMAP suspension array technology; Luminex Inc., Austin, TX, USA). Millipore bead kits (Upstate, Temecula, CA, USA) were used for the detection of 40 different proteins: IL-1α, -3, -6, -8, intercellular adhesion molecule-1 (ICAM-1), inducible protein-10 (IP-10), monocyte chemoattractant protein-1 (MCP-1), macrophage inflammatory protein-1a (MIP-1a), MIP-1b, Regulated on Activation, Normal T cell Expressed and Secreted (RANTES), platelet-derived growth factor-AA (PDGF-AA), PDGF-AB/BB, vascular endothelial growth factor (VEGF), VEGF-C, matrix metalloproteinases-1 (MMP-1), -2, -7, -9, -12, -13, monokine induced by gamma interferon macrophage inhibitory factor (MIG), macrophage migration inhibitory factor (MIF), myeloperoxidase (MPO), vascular cell adhesion molecule 1 (VCAM-1), hepatocyte growth factor (HGF) angiopoietin 2 (ANG-2), Fas Ligand, Plasminogen activator inhibitor-1 (PAI-1), receptor for advanced glycation endproducts (RAGE), Fas, tissue inhibitor of metalloproteinases–1 (TIMP-1), -2, -3, -4, transforming growth factor β–1 (TGFβ-1), -2, -3, brain-derived neurotrophic factor (BDNF), Cathepsin D, and neural cell adhesion molecule (NCAM). Twenty-five microliters of each sample were used undiluted, and kits were analyzed according to the instructions provided by the manufacturer. Samples were read on the Luminex xMAP system. Detection limits were 10 pg/mL for IL-6, MMP-9, IL-8, MMP-1, -2, -7, MIF, MCP-1, VEGF, MMP-13, 12, IL-1α, VEGF C, MIG, MPO, VCAM-1, ICAM-1, HGF, IP-10, ANG-2, Fas Ligand, PAI-1, RAGE, Fas, TIMP-1, -2, -3, -4, and 15 pg/mL for TGFβ-1, -2, -3, BDNF, Cathepsin D, PDGF-AA, -AB/BB, RANTES, NCAM, IL-3, MIP-1a, -1b (according to the instructions provided by the manufacturer). 

Statistical Analysis

Single t-tests were calculated to determine differences between the concentration of proteins in patients with retinal detachment and the control group. The resulting P values were Benjamini-Hochberg corrected for multiple hypothesis testing. 

Multiple linear regression models were fitted to identify predictors of variation in expression of intravitreal cytokines and chemokines. The covariates in the full regression model were symptom duration, detachment height, refractive power, lens status, number of quadrants, interaction of quadrants, and lens status and interaction of detachment height and lens status. Best subset selection was applied to choose the subset of clinical parameters with the optimal Bayesian Information Criterion (BIC). Variables, which were included in the model after best subset selection were considered to be statistically relevant predictors. 

In addition, a nonparametric sensitivity analysis was performed to investigate, if the observed differences remain significant if the assumptions underlying the parametric model do not hold. For each cytokine and chemokine we computed Spearman correlation coefficients with the variables duration, height (control, low RD height, high RD height), refraction, quadrants (control, <2 quadrants, ≥2 quadrants) and Wilcoxon tests for the variable lens, which has only two levels. Additionally we computed Spearman correlation coefficients for the variables height and quadrants in the pseudophakic subgroup. These tests reflect all independent variables that were used in the regression analysis. To adjust for multiple testing (280 tests) Benjamini-Hochberg adjusted P values were computed to control the False Discovery Rate at level 0.05. For all parameters (except duration and lens) that showed a significant correlation, pairwise group comparisons were performed. These were again adjusted for multiple testing with the Benjamini Hochberg procedure for the three comparisons. 

Results were only considered statistically significant when the sensitivity analysis did confirm the results of the parametric analysis. 

All calculations were performed in R 3.0.2 (R Foundation for Statistical Computing, Vienna, Austria). 

Results

Sixty patients with RRD were included in the study. Thirty-eight were males and 22 females with a mean age of 60.1 ± 12.0 (SD) years. Thirty patients reported visual symptoms typical for RD such as flashes, shadows, visual field defects or unspecified decrease in visual acuity of less than 10 days before surgery, while 28 individuals had symptoms present for a longer period of time. Two study patients were unsure about symptom duration and were therefore not included in this subanalysis. The mean symptom duration in the study population was 19.4 ± 29.5 days (range, 1–180 days). In 34 eyes less than two retinal quadrants and in 26 eyes two or more retinal quadrants were affected by detachment. Thirty-four patients presented with a flat and 26 with a high retinal detachment height. Out of the 60 study eyes examined, 17 were emmetropic, 3 hyperopic, 26 myopic with less than 6 D and 13 myopic with 6 D or more. Information from one patient regarding refractive status is missing. 

Twenty-eight individuals had undergone previous cataract surgery with an intraocular lens implanted in the capsular bag. Table 1 lists a summary of clinical parameters recorded from RRD patients. 

Twenty vitreous samples obtained from patients with idiopathic epiretinal membrane during routine surgery for the removal of the membrane served as controls. In this group 13 patients were female and 7 male with a mean age of 64.3 ± 9.0 years. Of the 20 control eyes 6 were emmetropic, 9 myopic, and 5 hyperopic. In 8 patients, cataract surgery had been performed in the study eye. 

Intravitreal Cytokine/Chemokine Patterns

A total of 40 different cytokines and chemokines involved in a variety of pathways such as inflammation, angiogenesis, or cellular growth were analyzed in the vitreous samples and compared between the two groups. Of these 40 factors investigated, 25% were significantly differentially regulated: IL-6, -8, MCP-1, IP-10, TIMP-1, -2, TGFβ-3, BDNF, PDGF-AB/BB, and MIP-1α were significantly increased in the vitreous of RRD eyes compared with controls (Table 2). Supplementary Table 1 lists all concentrations of remaining cytokines and chemokines in the RRD and control group. 

Table 2

 

Statistically Significantly Differentially Regulated Cytokines and Chemokines Between RRD and Control Group Measured by Multiplex Bead Array

 

Correlation of Intravitreal Protein Expression and Clinical Characteristics

For subsequent analyses, cytokines and chemokines were compared with different clinical parameters (Table 1) and a best subset selection was performed for every factor. The selected variables are statistically significant predictors of cytokine and chemokine concentrations. Linear regression analysis revealed IL-8 and TGFβ-3 as factors increasing with the number of retinal quadrants detached (Figs. 1A, 1B), while TIMP-1 serves as a parameter increased in eyes with high RD heights (Fig. 1C). 

Figure 1

 

Median and mean (cross) concentrations of IL-8 (A), TGF-β3 (B), and TIMP-1 (C) in picograms per milliliter in eyes with RRD and control patients. Rhegmatogenous retinal detachment patients are divided into two groups based on the number of retinal quadrants detached (A, B) or height of RD (C). Interleukin-8 and TGFβ-3 significantly increase with the number of retinal quadrants detached (A, B), while TIMP-1 significantly rises with the height of RD. Statistically significant differences between groups are marked by an asterisk.

Additional regression analyses revealed significantly increased levels of IP-10 and MPO in eyes with two or more retinal quadrants detached compared with eyes with fewer quadrants involved (Figs. 2A, 2B). TIMP-2 was significantly increased in eyes with higher compared with more shallow detachment heights (Fig. 2C). For all three latter factors, concentrations in control eyes were not significantly altered compared with eyes with less than two quadrants/flat detachment heights. 

Figure 2

 

Median and mean (cross) intravitreal levels of IP-10 (A) and MPO (B) in picograms per milliliter in RRD eyes grouped by number of quadrants detached and controls. Median values of TIMP-2 (C) in picograms per milliliter in the RRD group divided by detachment height and controls. IP-10 and MPO levels are highest in eyes with two or more quadrants detached, while TIMP-2 concentrations peak in eyes with high detachment height. Levels in control eyes are not significantly altered compared with eyes with less than two quadrants/flat detachment heights. Statistically significant differences between groups are marked by an asterisk.

In further subgroup analyses, we compared protein concentrations and morphologic characteristics between phakic and pseudophakic patients. The concentration of VCAM-1 was significantly increased in the vitreous of pseudophakic patients with higher detachment height (Fig. 3A), whereas levels of NCAM were significantly reduced in pseudophakic eyes with shallow detachment height (Fig. 3B). 

Figure 3

 

Median and mean (cross) levels of VCAM-1 (A), NCAM (B), TGF-β2 (C), and Cathepsin D (D) in picograms per milliliter in pseudophakic eyes with RRD grouped by detachment height and in phakic RD/control patients. VCAM-1 concentrations significantly increase in pseudophakic eyes with higher detachment height compared with shallower RD or phakic RD eyes/controls (A). Mean levels of NCAM (B), TGF-β2 (C), and cathepsin D (D) are lowest in eyes with shallow RD, medium in eyes with high RD and highest in phakic RD/controls. Levels below detection limit are not depicted by graphs (1, 1, 3, 4 value(s) not shown in [A–D], respectively). Statistically significant differences between groups are marked by an asterisk.

Additionally, we observed lowest mean concentrations of TGFβ-2 in pseudophakic RD eyes with shallow detachment height, yet not statistically significant (Fig. 3C). 

Lowest concentrations for cathepsin D were observed in pseudophakic eyes with low RD height compared with eyes with high RD height or phakic RD/control eyes (Fig. 3D). 

Levels of TIMP-4 and IL-1α showed variation between subgroups (Figs. 4A, 4B). 

Figure 4

 

Median and mean (cross) concentrations of TIMP-4 (A) and IL-1α (B) in picograms per milliliter in pseudophakic RD eyes subdivided by the number of quadrants involved and in phakic RD/control patients. Mean levels of TIMP-4 (A) and IL-1α (B) are higher in pseudophakic patients with two or more quadrants involved and lower in pseudophakic patients with less than two quadrants affected compared with phakic RD eyes/controls.

Mean concentrations were lowest in eyes with less than two quadrants, medium in phakic RD/control eyes, and largest in pseudophakic eyes with two or more quadrants detached (Figs. 4A, 4B). 

Increased concentrations of IL-6, MCP-1, TIMP-2, TGFβ-3, BDNF, PDGF-AB/BB, and MIP-1α in the vitreous of RD eyes did not significantly depend on the extent of retinal detachment. 

The comparison of duration in days of visual symptoms before surgery and intraocular protein levels revealed lower values of TIMP-1 in patients with longer symptom duration (Fig. 5). However, the significance is mainly caused by a single highly influential observation from a patient reporting RD symptoms for 180 days. 

Figure 5

 

Retinal detachment patients with longer symptom duration show decreasing levels of TIMP-1.

No statistically significant association of intraocular proteins with the refractive power in RRD eyes could be observed. 

Discussion

We performed a prospective clinical study in 60 patients with RRD to determine quantitatively the intravitreal cytokine/chemokine profile and to examine the correlation with relevant clinical parameters such as the extent of RD, lens status, or the duration of clinical symptoms. Our study results showed that the vitreous of eyes with RRD contained a distinct pattern of proteins involved in different pathways, such as inflammation, apoptosis, autophagy, or matrix remodeling. Further, we identified proteins specifically linked to height of RD, number of detached retinal quadrants, lens status, or duration of symptoms. 

Studies in monkey eyes showed that the height of RD correlated to the amount of outer retinal damage as documented by photoreceptor cell degeneration.²¹ This observation may be due to reduced diffusion of oxygen and other nutrients to the outer retina in higher detached retina.^2,22 Ross and coworkers²³ showed that the preoperative height of macular detachment in humans is an important factor regarding final visual recovery.²³ We identified TIMP-1 as a factor increasing with greater RD height. Further TIMP-2 levels were elevated in eyes with higher compared with more shallow detachment heights. A physiological balance between TIMPs and MMPs is an important feature in retinal morphology and function and is impaired during detachment of the retina. TIMPs are key enzymes for the modulation of extracellular matrix (ECM) by inhibiting MMPs, which degrade ECM components. Both enzymes are constitutively expressed in the retinal ganglion cell layer and in the outer retina.^24,25 It is still unclear which exact role TIMPs play in inhibiting or promoting retinal fibrosis in the time course of a detached retina. However, data from animal models and human subjects support a pathogenic role of imbalanced TIMPs in eyes with RD.^26,27 Zacks²⁸ identified TIMP-1 as one of the genes with the highest increase in transcription levels up to 28 days after experimentally-induced RD in rats. To the best of our knowledge, our data link for the first time increasing expression levels of TIMPs with higher detachment heights in humans. In addition, levels of TIMP-4 were slightly, yet not statistically significantly, increased in our pseudophakic RD subgroup with more than two quadrants affected. Further, we were able to demonstrate a weak, yet statistically significant association of declining concentrations of TIMP-1 in cases with long-lasting RD, a finding, which has only been reported in a murine RD model before.²⁶ We suggest that the downregulation of TIMP levels in eyes with long persisting RD lead to increased MMP activity and may predispose the retina to pathologic remodeling potentially leading to PVR and redetachments after surgical repair. 

Besides detachment height, extent of RD plays a pivotal role in the final visual acuity outcome. In 1981, Tani and coworkers²⁹ observed that only 7% of eyes with total RD, but 47% of eyes with two quadrants detached reached a final visual acuity of 20/50. In the current study, we defined the extent of RD by number of quadrants involved and correlated this number to intravitreal proteins. Our data revealed that IL-8 and TGFβ-3 significantly increased in the vitreous with the number of retinal quadrants detached. Interluekin-8 is secreted by both RPE cells and macrophages and serves as a chemotactic protein for inflammatory cells.^30,31 Elevated levels of IL-8 in the vitreous of RD eyes with a greater detached retinal area indicates an inflammatory reaction eventually due to underlying ischemic conditions, as RD leads to hypoxia in the corresponding outer retinal areas. The simultaneous upregulation of TGF-β 3 suggests that apoptotic processes as well as biological activities involved in ECM synthesis occur at the same time to a greater extent the more retinal quadrants are involved.^32,33 Studies have shown that macrophages and microglia are present in detached retina and clear debris, mainly apoptotic photoreceptors, into the subretinal space.^34,35 Active macrophages secrete the pro-oxidant enzyme MPO, which is expressed in the retina,^36,37 and was, together with the pro-inflammatory cytokine IP-10, higher expressed within the RD group with more than two quadrants affected. Myeloperoxidase has been shown to increase reactive oxygen species (ROS) potentially resulting in retinal tissue damage.³⁸ 

Comparing lens status to intravitreal cytokines and chemokines revealed clinically highly interesting findings adding fundamental understanding to previously reported clinical observations. The “SPR study,” which compared scleral buckling with primary vitrectomy in RRD, showed that in pseudophakic patients initial scleral buckling procedures required more surgeries than initial primary vitrectomy.³⁹ Results from another large multicenter trial investigating cases with RRD revealed that pseudophakic eyes were more likely to have a recurrent detachment or complication after the first procedure.⁴⁰ Our data showed that eyes of pseudophakic patients with great RD heights contained excessive levels of the proinflammatory protein VCAM-1. Upregulation of this adhesion molecule in the retinal microvasculature promotes leukocyte migration, hence contributing to an increased inflammatory state.⁴¹ Similar to our findings Zacks²⁸ reported significant upregulation of VCAM-1 at 1 and 7 days after experimental induction of RD in rats. Another study analyzing VCAM-1 in subretinal fluid of human eyes with RRD found increasing levels with progression of PVR severity.⁴² Limb and coworkers⁴³ proposed that molecular inflammatory pathways mediated by VCAM-1 and leading to fibrotic processes in eyes with proliferative diabetic retinopathy resulted from activation of RPE, endothelial or Müller cells. It is thus conceivable that in eyes with RRD similar mechanisms occur placing, in particular, pseudophakic eyes with high RD height at increased risk for developing PVR. Interestingly, our analysis revealed that pseudophakic eyes with shallow RD height contained significantly lower concentrations of NCAM in comparison with pseudophakic eyes with higher RD height or phakic RD/control eyes. NCAM is a glycoprotein known to be responsible for adhesion between RPE and neural retina and for maintaining regular retinal ganglion cell number and survival.^44,45 It is intracellularly modified by degradation processes mediated by calpains, which belong to the papain superfamily.⁴⁶ A recent study suggested that calpain-mediated cellular mechanisms lead to lysosomal downregulation of autophagy specifically associated with photoreceptor cell death.⁴⁷ 

Similar protein expression kinetics as NCAM was found for cathepsin D when evaluating the effect of lens status in RD. Like calpains, cathepsins are part of the papain superfamily and are lysosomal proteases associated with degradation of phagocytosed photoreceptors and extracellular matrix but also preservation of retinal photoreceptors with deficiency inducing apoptosis.^48–50 Besirli and coworkers⁵¹ reported upregulation of cathepsin D and B upon induction of RD in rats as part of protective autophagy activation within the retina. Comparable results with upregulation of cathepsin genes up to 28 days after RD in rats were reported by Zacks.²⁸ Our data obtained from human vitreous do not confirm reported upregulation of cathepsin D in eyes with RD. Downregulated levels of cathepsin D and NCAM in our specific pseudophakic subgroup may reflect a direct or, in case of NCAM, indirect dysfunction of lysosomal activities in RD. 

Selected other factors, which are upregulated in RD eyes irrespective of specific RD morphologies, are MCP-1, MIP-1α, PDGF AB/BB, IL-6, and BDNF. Excessive levels of proinflammatory MCP-1 in the vitreous of eyes with RD play a crucial role in mediating photoreceptor apoptosis as shown in experimental mouse studies.⁵² Further, RD leads to increased MCP-1 expression in the Müller glia and to increased macrophage infiltration in the detached area.⁵² Similar to MCP-1, MIP-1α mediates the recruitment of monocytes, particularly in areas of ischemic retinopathy and is also upregulated in our RRD group.⁵³ Another major growth factor upregulated in the vitreous of RD eyes is PDGF, which is produced by a variety of cells including activated macrophages, platelets, or endothelial cells.⁵⁴ Platelet-derived growth factor has been shown to act as a chemoattractant for inflammatory cells and fibroblasts, thereby enhancing inflammation and ECM formation.⁵⁵ Furthermore, the development of contractile retinal membranes and tractional RD is linked to the presence of PDGF.²⁰ It is also involved in other retinal pathologies such as neovascular AMD and may reflect the impact of photoreceptor stress in a detached condition from the nurturing RPE.⁵⁶ Increased levels of IL-6, as observed in our patient cohort, have been linked with the protection against photoreceptor apoptosis as described by Chong and coworkers⁵⁷ in a murine RD model. Upregulation of IL-6 levels has been demonstrated in a rat RD model before and occurs within 24 hours of detachment as a potential survival factor for photoreceptor cells in the early phase of RD.⁵⁸ In ischemic retinal tissue IL-6 upregulation enhances survival of retinal ganglion cells.⁵⁹ On the contrary, a clinical study reported an association between upregulated concentrations of IL-6 in human subretinal fluid and the development of fibrotic membranes after primary RRD repair.⁶⁰ These data point to an early protective role of IL-6 for retinal tissue with a still uncertain function at later stages during the pathogenesis of RD. Brain-derived neurotrophic factor is another factor upregulated in the RD group in our clinical study, which has been described as neuroprotective in a feline RD model by reducing Müller cell activity.⁶¹ Similar to IL-6 it may help to reduce initial damage to photoreceptors caused by detached retinal layers. 

In summary, our study demonstrated that lens status and extent of RD had significant influence on the expression of intravitreal cytokines and chemokines. The results obtained from our study not only give explanations to observed clinical findings in patients with RRD but also have direct clinical implications. We confirm and extend previously reported observations of increased intravitreal proinflammatory, profibrotic, and proapoptotic factors in eyes with RRD by correlating them to clinically relevant characteristics and identifying conditions, which put patients at risk of permanent retinal damage. Particularly, conditions such as greater RD height, more extensive RD area and/or the presence of an intraocular lens appear relevant in promoting neurodegenerative pathways. Our findings add fundamental understanding to previously reported observations of poorer outcome in severe presentation of RRD. In addition we identified proteins in the vitreous activated upon retinal separation and potentially protecting neuroretinal tissue. Finally, a complete understanding of the intraocular fluid milieu is crucial for the development of potential therapeutics administered in conjunction with retinal surgery to facilitate the restoration of best functional and morphologic outcomes in patients after RRD. 

Acknowledgments

The authors thank Petra Waidhofer-Soellner for excellent technical work. 

Disclosure: A. Pollreisz, None; S. Sacu, None; K. Eibenberger, None; M. Funk, None; D. Kivaranovic, None; G.J. Zlabinger, None; M. Georgopoulos, None; U. Schmidt-Erfurth, None