Vitreous Microparticle Shedding in Retinal Detachment: A Prospective Comparative Study

Abstract

Purpose: Microparticles (MPs) are membrane-derived vesicles measuring less than 1 μm in diameter. They are shed from nearly every activated or preapoptotic cell and may exhibit biologic activities in inflammation or apoptosis settings. The main purpose of this study was to determine whether MP shedding was higher in the vitreous of patients with retinal detachment (RD).

Methods: This was a prospective, comparative study. Levels of vitreous MPs (including phosphatidylserine [PS]-expressing MPs, photoreceptor cell–derived MPs, and photoreceptor cell–derived MPs expressing PS) and soluble proinflammatory factors (i.e., monocyte chemoattractant protein-1, intercellular adhesion molecule-1, and IL-6) were analyzed by flow cytometry. Samples were obtained from 49 eyes undergoing RD surgery and 41 control eyes.

Results: Vitreous levels of all the MPs studied were significantly increased in the RD group. Vitreous MP levels were correlated with levels of at least one proinflammatory factor depending on MP subsets. Concerning clinical parameters, vitreous PS-expressing MP and PS-expressing photoreceptor cell–derived MP levels were higher depending on the duration of RD at surgery, the detached retina surface, and the macula status and were found more sensitive than proinflammatory factors only for the duration of RD at surgery.

Conclusions: Vitreous concentrations of MPs (mainly derived from photoreceptor cells) are higher after rhegmatogenous RD and found to be correlated with soluble proinflammatory factors.

Rhegmatogenous retinal detachment (RD) is the leading sight-threatening surgical ophthalmologic emergency.¹ Detachment of the neurosensory retina from the RPE can lead to photoreceptor cell death.^2,3 Understanding the pathophysiologic mechanisms taking place after RD is mandatory so that innovative strategies can be developed, thus limiting vision loss. 

Cytokines are extracellular proteins regulating cell activation or death and play a pivotal role in triggering inflammation or apoptosis. An increase in vitreous concentrations of proinflammatory cytokines, including TNF-α,^4,5 IL-1β,^4,6 and BFGF⁷ has been documented in RD. The role of cytokines in the occurrence of proliferative vitreoretinopathy (PVR), which is the most common cause of failure in RD surgery, is also well documented.^8,9 However, the relationship between mechanical stress represented by RD and expression of inflammatory cytokines remains poorly understood. 

Microparticles (MPs) correspond to small vesicles—between 0.1 and 1 μm in diameter—produced by budding and blebbing from the plasma membrane of nearly all cells in response to activation or apoptosis.¹⁰ Microparticles express cell surface markers from the cells from which they derive, which can be useful for identification of their origin. For instance, glycoconjugates specific from photoreceptor cells can be detected on photoreceptor cell–derived MPs by lectins from arachis hypogaea peanut agglutinin (PNA).¹¹ In addition, MPs may express phosphatidylserines (PSs)—which can be revealed using its ligand annexin V—on their surface.¹² Phosphatidylserine exposure at the MP surface is involved in several MP functions, including their procoagulant activity,¹³ inhibition of both inflammation and specific immune responses,¹⁴ and their uptake by professional phagocytes (i.e., macrophages and immature dendritic cells),¹⁵ but also by nonprofessional neighboring cells (e.g., RPE cells).¹⁶ This may play a central role in cell-to-cell communication and during inflammation¹⁷ and apoptosis.¹³ In addition to acting as key messengers between cells, MPs are connected to cytokines, in that MP shedding from healthy or damaged cells can be enhanced by cytokines.¹⁸ Microparticles were detected in human plasma and various tissues and fluids such as atherosclerotic plaques¹⁹ and synovial fluid.²⁰ In the eye, MPs were overexpressed in the vitreous in proliferant diabetic retinopathy.⁹ Overall, the biologic activity of MPs is associated with inflammation and/or apoptosis. 

This study aimed to investigate the levels of MPs in the vitreous of patients with RD—a situation associated with photoreceptor cell death—and their relationship with concentrations of proinflammatory factors, including monocyte chemoattractant protein-1 (MCP-1), intercellular adhesion molecule-1 (ICAM-1), and IL-6. We speculated that photoreceptor-derived MPs may reflect photoreceptor cell death and could be an additional factor associated with inflammation in RD. 

Methods

Patients who underwent vitrectomy in the ophthalmology department of the University Hospital of Besançon (Besançon, France) between September 2013 and April 2014 were included. The RD group included consecutive patients presenting with RD requiring vitrectomy surgery. The control group included patients who had undergone vitrectomy surgery for idiopathic epiretinal membrane (ERM) or macular hole (MH). 

The exclusion criteria encompassed any history of vitreoretinal surgery on the eye studied, diabetic retinopathy, or vitreous hemorrhage. Patients less than 18 years of age were also excluded. 

This study was approved by the regional Institutional Ethics Committee (Comité de Protection des Personnes CPP 13/385) and adhered to the Declaration of Helsinki. All patients signed written informed consent. 

Baseline Evaluation

Prior to surgery, all patients underwent a comprehensive ophthalmologic examination including measurement of best corrected visual acuity (BCVA) (measured on the Early Treatment Diabetic Retinopathy Study scale and then converted to logMAR) and indirect ophthalmoscopy (QuadrAspheric; Volk Optical, Mentor, OH, USA). 

In RD patients, the number, size, and location of tears were collected, as well as the detachment surface (expressed as a number of detached quadrants [Q]: <1, 1, 2, ≥3) and PVR stage (according to the revised Retina Society Terminology Committee classification²¹: graded A, B, or C) in a well-documented drawing obtained at baseline and complemented by the surgery video recording. The macula status was defined clinically as follows: macula-on RDs represented RDs that did not involve the retinal area between the vascular arcades, whereas macula on-off and macula-off RDs corresponded, respectively, to partial or complete detachment of the area located between the vascular arcades. 

All patients underwent a spectral-domain optical coherence tomography (SD-OCT) (Spectralis; Heidelberg Engineering, Heidelberg, Germany) examination to measure the foveal height of the RD using the calipers of the manufacturer's software on a horizontal 30° (length) line B-scan passing through the fovea. 

Surgical Procedure

Undiluted vitreous fluid samples (400 μL) were collected at the beginning of a standard three-port 23-G pars plana vitrectomy with the infusion line switched off (Accurus; Alcon, Johns Creek, GA, USA). Neither cataract surgery nor scleral buckle was combined to the vitrectomy surgery in RD group. In the control group, some patients may have undergone combined cataract surgery and vitrectomy, but the vitreous samples were always collected at the beginning of the surgical procedure so that it could stay pure and undiluted. All samples were sent to the French Blood Agency biomonitoring platform (EFS Bourgogne-Franche-Comté, Besançon, France) within 4 hours, in a dedicated box maintaining the samples at a steady +4°C, in order to avoid any sample deterioration and artefactual MP production.²² 

Postoperative Evaluation

All RD patients underwent a renewed clinical and imaging examination at the 1-, 3-, and 6-month postoperative visits (M1, M3, and M6, respectively). The SD-OCT imaging (Spectralis; Heidelberg Engineering) was performed to obtain the central macular thickness (CMT). The axial length was measured using an IOL Master500 biometer (Carl Zeiss Meditec, Jena, Germany). 

Microparticle Analysis

Microparticle Isolation.

Microparticles were isolated from vitreous samples according to the International Society of Thrombosis and Haemostasis recommendations.²⁰ Briefly, samples were centrifuged with the following protocol: 2500g for 15 minutes, decantation, and then 2500g for 15 minutes again.²³ All the samples were then frozen and stored at −80°C and thawed quickly at 37°C for 5 minutes before analysis by flow cytometry. 

Microparticle Quantification by Flow Cytometry.

Flow cytometry was performed using a NAVIOS Cytometer (Beckman Coulter Immunotech, Villepinte, France) by two independent examiners unaware of the patient status. Vitreous MPs were identified with the following reagents: allophycocyanin-conjugated annexin V (BD Pharmingen, Franklin Lakes, NJ, USA) and lectins from arachis hypogaea PNA conjugated with FITC (Sigma-Aldrich Corp., St. Louis, MO, USA) to detect photoreceptor cell–derived MPs, as described.⁹ Three MP subsets are thus analyzed: total annexin V⁺ MPs, photoreceptor cell–derived PNA⁺ MPs, and photoreceptor cell–derived PNA⁺ annexin⁺ MPs. Two types of beads were added to vitreous samples: the first one (Megamix-Plus) to determine a standard-sized area corresponding to MPs and the second one (Cytocount) to obtain MP counts. Fluorescent beads with three defined diameters (0.3, 0.5, and 0.9 μm, Megamix-Plus FSC; BioCytex, Marseille, France), were selected to cover the theoretical MP size range (0.1–1 μm) and used to define a standard-sized area (0.3–1 μm-eq). The number of MPs was calculated on the basis of the known number of Cytocount beads (Dako, Trappes, France) added to the sample. Data were analyzed using the Kaluza software (Beckman Coulter, Inc., Brea, CA, USA). 

Cytokine Quantification

The MCP-1, ICAM-1, and IL-6 capture beads (CBA Human Soluble Protein Flex Sets; BD Biosciences, Franklin Lakes, NJ, USA) were used to determine proinflammatory factor vitreous concentrations, together with the BD CBA Human Soluble Protein Master Buffer Kit, according to the manufacturer's recommendations. Cytometric bead array was performed on a CANTO II flow cytometer (BD Biosciences). After acquisition, data were analyzed using FCAP Array software (BD Biosciences). 

Statistical Analysis

The number of subjects was calculated with nQuery Adisor, based on the Noether GE method²⁴ (α risk = 0.05, power = 0.9, unilateral test, probability that a control subject has fewer MPs than a RD patient = 0.7). Based on this hypothesis, the number of subjects to include was 36 per group. 

Microparticle and proinflammatory factor levels are expressed as the mean ± SD. 

Descriptive statistics and the Mann-Whitney test were used for statistical comparisons between the two groups, with the 2-tailed P < 0.05 considered significant. The Kruskal-Wallis nonparametric analysis was performed to compare levels between more than two groups, followed by Dunn's multiple comparison test. The nonparametric correlation with calculation of the Spearman r was performed. Statistics were calculated using the commercially available software program GraphPad Prism 6 (GraphPad, Inc., San Diego, CA, USA). A multivariate stepwise linear regression was performed using IBM SPSS statistics (Chicago, IL, USA) 19.0 to determine which independent factors (age, axial length, proinflammatory factor levels) had a statistically significant effect (defined as P ≤ 0.05) on MP concentration. The significant level at entry was 0.15 and for remaining in the model 0.05. 

Results

Demographics

Ninety eyes of 89 patients were included in this study. The RD group consisted of 49 eyes (48 patients), whereas the control group was composed of 41 eyes (41 patients) divided into two subgroups: idiopathic ERM (n = 25) and MH (including impending macular hole) (n = 16). The RD group encompassed 28 male and 20 female patients from 46 to 85 years of age (mean ± SD: 65 ± 10.6 years), with an average axial length of 24.9 ± 2.1 mm. The RD group included 32 phakic eyes. There was no significant difference between the RD and control groups except for the age, which was more advanced in the control group (P < 0.05). 

Characterization of Vitreous MPs

The major finding was represented by the higher MP levels in the RD group (versus the control group). All the MPs studied (i.e., annexin V⁺ PS-expressing MPs, PNA⁺ photoreceptor cell–derived MPs, and photoreceptor cell–derived PNA⁺/annexin V⁺ PS-expressing MPs) were higher in the vitreous of RD patients (by 3-fold, 1.4-fold, and 3.6-fold, respectively) in comparison with the control group (Mann-Whitney test, P < 0.01; Fig. 1). 

Figure 1

 

Comparison of MP levels between the RD and control groups. Mann-Whitney test. *Significant difference (P < 0.05). Horizontal bars represent medians.

Subgroup analysis showed that all the vitreous MPs analyzed were higher in the RD group than the ERM group, and all except the PNA⁺ MPs, compared with the MH group (Fig. 2). 

Figure 2

 

Comparison between the RD and control subgroup MP levels. ERM, epiretinal membrane (n = 25); MH, macular hole (n = 16). Data are given as median (horizontal bar), 25th–75th percentile (boxes), and minimum and maximum (whiskers). Kruskal-Wallis analysis. *Significantly different from the RD group (P < 0.05).

Interestingly, vitreous PNA⁺ MP levels were also higher in the MH group than the ERM group (P < 0.05). 

Characterization of Vitreous Proinflammatory Factors

Vitreous MCP-1 and IL-6 concentrations were higher in the RD group, by 3.9-fold and 10.8-fold, respectively, than the control group (P < 0.01). Higher vitreous ICAM-1 levels, however, did not reach statistical significance in the RD group (Fig. 3). 

Figure 3

 

Comparison of proinflammatory factors (cytokines) between the RD and control groups. Horizontal bars represent medians. Mann-Whitney test. *Significant difference (P < 0.05).

There was no difference in vitreous proinflammatory factor concentrations, except for IL-6 (higher in the MH group) between the two control subgroups. 

Relationship Between Vitreous Proinflammatory Factors and MPs

Levels of annexin V⁺ MPs in the vitreous of RD patients were correlated with MCP-1, IL-6, and ICAM-1 concentrations (r = 0.61, P < 0.01; r = 0.60, P < 0.01; and r = 0.18, P < 0.05, respectively). A similar correlation was observed between vitreous PNA⁺/annexin V⁺ MPs (i.e., photoreceptor cell–derived MPs expressing PSs) and IL-6 and MCP-1 levels (r = 0.49 and r = 0.55; P < 0.01). When considering PNA⁺ MPs, their levels were only correlated with vitreous MCP-1 concentrations (r = 0.20; P < 0.05). Multivariate analysis showed that vitreous MCP-1 and ICAM-1 levels and axial length (obtained in 59 eyes) were independently related to vitreous annexin V⁺ MP levels. 

Relationship Between Biologic and Clinical Data

There was no relationship between preoperative BCVA and vitreous MP concentrations. On the other hand, vitreous ICAM-1, IL-6, and MCP-1 levels were moderately correlated with BCVA (r = 0.37, r = 0.36, and r = 0.37, respectively; P < 0.01). 

Subgroup analysis, based on the duration of RD at time of surgery (i.e., <7, 7–13, and ≥14 days), showed that vitreous annexin V⁺ MP and PNA⁺/annexin V⁺ MP levels were higher in all RD subgroups than the control group (Kruskal-Wallis analysis with Dunn's multiple comparison test; P < 0.01). However, only recent RDs (<7 days) displayed higher PNA⁺ MP levels than the control group (P < 0.05). Overall, vitreous MP levels were significantly higher within the first week after RD. Then they started to decrease until eventually reaching the control group's MP levels. Concerning vitreous proinflammatory factors, IL-6 and MCP-1 were higher in all RD subgroups than the control group (P < 0.01). A trend toward higher vitreous proinflammatory factor levels related to RD duration at time of surgery was observed but did not reach statistical significance (Fig. 4). 

Figure 4

 

Comparison of MP and cytokine levels between duration of RD subgroups. Data are given as median (horizontal bar), 25th–75th percentile (boxes), and minimum and maximum (whiskers). Kruskal-Wallis analysis. *Significantly different from control group (P < 0.05).

Subgroup analysis based on the surface of the detached retina (<1Q, 1Q, 2Q, and ≥3Q of RD) showed a significant increase in vitreous annexin V⁺ MP in all RD subgroups compared with the control group (Kruskal-Wallis analysis followed by Dunn's multiple comparison test; P < 0.01). Vitreous PNA⁺ annexin V⁺ MPs were only higher in the 1Q and 2Q subgroups. At the same time, vitreous IL-6 and MCP-1 concentrations were higher in all the subgroups from 1Q of detached retina (Supplementary Fig. S1). 

According to the macula status (macula on, off, and on/off), vitreous annexin V⁺, and PNA⁺/annexin V⁺ MP levels, together with IL-6 and MCP-1 levels, were higher in all RD subgroups compared with controls (Supplementary Fig. S2). 

Comparison of the PVR status to vitreous MP concentrations showed great variation depending on the MP subpopulations considered and on which PVR subgroups were considered. No clear conclusion could be drawn. In contrast, vitreous MCP-1 and IL-6 levels were higher in all PVR subgroups than in the control group (P < 0.01; Supplementary Fig. S3). 

At baseline, there was no relationship between RD height (when measurable: n = 19) and vitreous MP levels. Vitreous IL-6 levels, on the other hand, were positively correlated with RD height (r = 0.62; P < 0.01). 

Postoperative BCVA increased progressively from 3 months of follow-up (Wilcoxon matched-pairs signed rank test; P < 0.05). However, there was no correlation between vitreous MP levels at the time of surgery and postoperative BCVA, whereas vitreous ICAM-1 concentrations were positively correlated with the postoperative BCVA at all visits (r = 0.29, r = 0.32, and r = 0.35, respectively; P < 0.05). 

Thirty percent of RD patients (n = 15) developed a secondary ERM during the follow-up and 16.3% (n = 8) developed a CME. ICAM-1 levels were correlated with a higher risk of developing ERM and macular edema (Mann-Whitney test, P < 0.01) increasing CMT at the 6-month visit. 

Discussion

Vitreous concentrations of MPs and the three proinflammatory factors analyzed here were significantly higher in patients with RD than in the control group. 

Microparticles are known to play an important role in cell-to-cell communication²⁵ and are implicated in homeostasis regulation,²⁶ endothelial dysfunction,²⁷ cancer progression,^28,29 and virus propagation.³⁰ Every living cell can produce MPs during its life cycle. Presence in a biologic fluid of MPs from different cell origins is thought to be a detectable signature of cell activation or apoptosis.³⁰ For instance, Chahed et al.⁹ studied the vitreous concentrations of MP in proliferative diabetic retinopathy (PDR). They reported increased retinal, vascular, and circulating cell-derived MP in the vitreous of PDR patients. Furthermore, they demonstrated the in vitro ability of these MPs to stimulate endothelial cell proliferation and angiogenesis. 

In the current study, levels of MPs originating from the photoreceptors (PNA⁺), and/or expressing PS (annexin V⁺) were higher in the vitreous of patients presenting RD. Apoptosis observed after RD in animal models³¹ and humans^32,33 may contribute to the shedding of MPs by preapoptotic retinal cells.^34,35 In fact, the mechanical separation of the neurosensory retina from the RPE produces early-onset changes at a cellular level such as microvillosity retraction of RPE cells and shortening of photoreceptor outer segments.^36,37 These changes also occur at a molecular level with the trigger of biologic cascades controlling cell activation,^38,39 proliferation, and inflammation. Such cell activation and mechanically induced cell damage⁴⁰ observed after RD can also contribute to the release of MPs in the vitreous of RD patients.⁴¹ Preliminary data in 15 samples show that no significant platelet-derived MP levels (3.1 annexin V⁺ CD41⁺/μL + 1.4, range, 0–8; vitreous annexin⁺ MPs found in these matched samples: 231/μL + 221, range, 12–978) can be found in the vitreous of RD patients. This suggests that MPs are rather produced locally and mainly from photoreceptor cells. 

The inflammatory cascade occurring after RD was studied extensively, and it has been shown that cytokine concentrations in the vitreous reach their peak at different times after the onset of RD.^42,43 Cytokine concentrations measured in the current study are in accordance with the data reported in the literature, with higher concentrations of IL-6^5,11,44 and MCP-1,^45,46 both of which are proinflammatory cytokines. Interestingly, we found a significant correlation between annexin V⁺ MPs and IL-6, ICAM-1, and MCP-1, the latter known to mediate photoreceptor death.⁴⁷ The assumption that can be made is that annexin V⁺ MP shedding is closely related to MCP-1 release by retinal and blood-derived cells, a relationship already demonstrated in the synovial fluid by Berckmans et al.⁴⁸ and Distler et al.,⁴⁹ who showed that MPs derived from immune cells can stimulate synovial fibroblasts by inducing cytokine synthesis (including MCP-1 and IL-6). Furthermore, it should be underlined that MPs by themselves may contain IL-1 bioactivity, an important inflammatory factor^50–52 known to trigger the production of IL-6^53,54 and MCP-1^55,56 by RPE cells. These cytokines could in turn activate inflammatory cells to generate more MPs, resulting in a positive feedback loop.^57,58 This hypothesis is supported by the RD duration-based analysis, which shows that the MP peak in the vitreous is reached rapidly after the occurrence of the detachment during the first week, which is concomitant with photoreceptor apoptosis^12,59 a maximum of 2 to 3 days after the occurrence of the RD and then rapidly decreasing after 7 days. 

One of the drawbacks of this study was the difficulty in accurately estimating the duration of RD based on subjective symptoms reported by the patient. Ideally, we should have included only patients with the same RD duration, implying a multicentric design (to include more patients), which would also have raised major issues in terms of sample preservation and analysis, because inappropriate handling and transport of the samples can influence MP levels. In addition, the method for analyzing MPs requires a specialized technical platform that is not available in every center. 

We should also mention the difficult task of comparing a “diseased” group to a “nonhealthy” control group. Our control group encompassed patients presenting with vitreomacular interface diseases (ERM and MH), and as we could see, there are some significant differences between the control subgroups in terms of MP and cytokine vitreous levels. For obvious ethical reasons, no healthy patients could have been included in this study, even if the possible results would have reflected a more physiologic state. 

In summary, MP shedding increases in the vitreous of eyes presenting with RD. The increase in PNA⁺ (i.e., photoreceptor cell derived) MPs is also correlated to MCP-1, a major proinflammatory cytokine involved in photoreceptor apoptosis. Microparticles could be one of the missing links between the mechanical stress induced by RD and the occurrence of inflammatory and cell activation cascades, leading to photoreceptor cell death. In a translational approach, further studies conducted in animal models would certainly enhance our understanding of the specific biological potency of these MPs, which may eventually serve as prognosis factors or as therapeutic targets in photoreceptor rescue strategies. 

Acknowledgments

Supported by the University Hospital of Besançon, France (Clinical Research and Innovation Delegation, DRCI): API-RFC 2012 Project. 

Disclosure: P. Tumahai, None; P. Saas, None; F. Ricouard, None; S. Biichlé, None; M. Puyraveau, None; C. Laheurte, None; B. Delbosc, None; M. Saleh, None