Ultrastructural Changes after Artificial Retinal Detachment with Modified Retinal Adhesion

Abstract

purpose. Artificial retinal detachment is increasingly used in submacular surgery. However, overcoming physiological retinal adhesiveness by subretinal fluid injection is suspected to cause cellular damage and thus to limit visual rehabilitation. This experimental study was designed to examine the ultrastructural changes induced by retinal detachment under vitrectomy conditions and to evaluate factors that reduce adhesiveness and minimize cellular damage.

methods. Twenty-one pigmented rabbits underwent vitrectomy, and the vitreous cavity was perfused for 10 minutes with various solutions. These included variations in osmolarity (314 and 500 mOsM), Ca²⁺ ion concentration (Ca²⁺-supplemented, low Ca²⁺, active Ca²⁺ deprivation via 1 mM EDTA), temperature (19 °C and 34°C), and ischemia (5 minutes). Nonvitrectomized eyes served as the control. Consecutively, an artificial bleb detachment was created underneath the visual streak by injecting 1 mL of buffered saline solution subretinally. Eyes were enucleated within 3 minutes, fixed with 2% glutaraldehyde/0.1 M cacodylate buffer (pH 7.4) containing 100 mM sucrose and processed for transmission electron microscopy and scanning electron microscopy.

results. If a Ca²⁺-containing standard solution was used during vitrectomy, retinal adhesiveness was strong, and a forced bleb detachment caused substantial cellular damage characterized by swollen and fragmented photoreceptor outer segments and disruption of retinal pigment epithelial cells. Use of a Ca²⁺-free solution moderately reduced the adhesive strength with consequently less ultrastructural damage. Active Ca²⁺-deprivation further reduced the retinal adhesion, but may have induced damage as suggested by intracellular vacuolization. Hyperosmolarity and ischemic conditions had toxic effects on both the photoreceptors and RPE cells. In contrast, the use of a preheated Ca²⁺-free solution (34°C) substantially reduced retinal adhesiveness under vitrectomy conditions and hence ultrastructural damage.

conclusions. Artificial retinal detachment causes substantial ultrastructural damage in eyes with physiological retinal adhesiveness if performed under vitrectomy conditions similar to surgery in humans. The use of a preheated Ca²⁺-free physiologic saline solution seems to be suitable to reduce retinal adhesion sufficiently, without causing significant cellular damage.

Artificial retinal detachment is a critical step in submacular surgery, such as CNV membrane extraction, ^{1 2 3} 360° macular translocation, ^{4 5 6 7} limited macular rotation, ^{8 9} transplantation of autologous pigment epithelium, ^{10 11 12 13} or, more recently, translocation of full-thickness choroid–RPE patches. ^{14 15} However, overcoming the physiological photoreceptor–RPE adhesiveness by subretinal fluid injection has been shown to cause substantial cellular damage in the neuroperceptive interface. ^{16 17} Because anatomic and functional success depend on minimizing adverse effects, it is necessary to understand which conditions allow easy detachment of the retina. 

Marmor et al. have comprehensively shown in numerous studies that retinal adhesion can be modulated by changes in ion concentration, pH, temperature, osmolarity, oxygenation, and metabolic activity. ^{18 19 20 21 22 23} Particularly, it has been shown that reducing Ca²⁺ and Mg²⁺ ions leads to a substantial reduction of retinal adhesive strength ^{16 17 19 24} that is reversible on the restoration of normal ionic conditions. ²⁵  

Most of these studies on retinal adhesion have been performed ex vivo, ^{19 22 25 26} even though it is known that retinal adhesion decreases rapidly after enucleation. ²² However, it is important that conditions applicable to the surgical in vivo situation be examined. Preliminary studies in rabbits were promising ^{17 27 28} and may warrant additional investigations of the feasibility of modifying retinal adhesiveness under vitrectomy conditions similar to surgery in humans. In this experimental investigation, we have compared several conditions during vitrectomy—osmolarity, ionic changes, temperature, and metabolic activity (ischemia)—to define the conditions that mainly promote consistent detachment with ease and with minimal ultrastructural damage to the neural retina and the retinal pigment epithelium. 

Methods

Animals

Twenty-one pigmented rabbits weighing 3 to 4 kg divided into seven subgroups were included in the study. The rabbits were anesthetized using ketamine hydrochloride (25 mg/kg body weight) and xylazine hydrochloride (6.5 mg/kg body weight) by intramuscular injection. The pupil of the right eye was dilated with 3 drops of 1% topical tropicamide and 3 drops of 10% phenylephrine hydrochloride. All experiments in this study were performed in accordance with institutional animal guidelines and according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Surgical Procedure

Traction sutures were placed on the temporal and nasal rectus muscles of the right eye followed by a peritomy. Sclerotomies for a three-port pars plana vitrectomy were prepared 2.0 mm behind the limbus with the infusion cannula sutured at the 4-o’clock position, and two additional sclerotomies at the 2- and 10-o’clock positions. For vitrectomy, the fundus was visualized using a contact lens. A core vitrectomy was performed above the prospective site of the bleb detachment. The vitreous cavity of the rabbits was perfused for 10 minutes with one of the following solutions (Table 1) : (1) Physiologic saline solution (PSS) supplemented with Ca²⁺ (pH 7.2, HI211; Hanna Instruments, Kehl, Germany) at room temperature, 21°C (PSS+, Acri.Tec GmbH, Berlin, Germany); (2) Ca²⁺-free solution (pH 7.2) at room temperature, 21°C (PSS−, Acri.Tec); (3) active Ca²⁺-depriving PSS by supplementing 1 mM ethylene diamine tetra-acetate (pH 7.2) at room temperature, 21°C (PSS−/EDTA); (4) hyperosmolar Ca²⁺-free solution (pH 7.1) at room temperature, 21°C (PSS−/Osm) adjusted to an osmolarity of 500 mOsM by adding NaCl under the control of an osmometer (Type 15; Loeser Messtechnik, Berlin, Germany); (5) Ca²⁺-free solution (pH 7.2) preheated at 34°C (PSS−/Temp), which was obtained by immersion of the bottle in a water bath (37°C) and appropriate insulation of the bottle and tube; (6) Ca²⁺-free solution (pH 7.2) under ischemic conditions (PSS−/Ischemia), which are achieved by simultaneously rising the IOP to 76 mm Hg for 5 minutes; and (7) nonvitrectomized eyes, which served as the control. 

Subsequently, a 41-gauge Teflon micropipette was introduced and approximated to the retina in the midperiphery below the visual streak With the tip of the cannula held close to the retina, the jet stream created a small retinal hole and consequently a subretinal bleb. ¹⁷ Approximately 1 mL PSS was manually injected to assure a bleb formation sufficiently large for unequivocal histology. To avoid reflux the cannula was positioned at the site of the retinotomy as in macular rotation surgery. In the control group, the subretinal injection was directly performed transvitreally without any preceding vitrectomy or perfusion. The area of detachment was noted, but postfixation changes and the changes in the height of the bleb did not allow for valid interpretation of these data. 

All surgeries were performed by the same vitreoretinal surgeon (KUB-S). The surgeon was not informed which type of solution was used. The intraoperative ease of inducing the retinal detachment was graded by the surgeon immediately after the procedure using a scale from 0 to 10, with 0 denoting spontaneous detachment and 10 denoting maximum force to achieve retinal detachment. 

Transmission Electron Microscopy and Light Microscopy

The eyes were enucleated within 3 minutes after surgery and fixed at 4°C in 2% glutaraldehyde/0.1 M cacodylate buffer (pH 7.4) containing 100 mM sucrose. Twenty-four hours after fixation, the bleb areas were resected with a dissecting microscope, cut in half, buffer rinsed, and postfixed in 1% osmium tetroxide and 0.1 M cacodylate buffer (pH 7.4) for 3 hours at room temperature. The specimens were rinsed and embedded in Spurr’s resin after dehydration through a graded series of ethanol. For light microscopy, semithin sections (0.7 μm) were stained with toluidine blue and examined and photographed with a microscope (Orthoplan; Leica, Wetzlar, Germany). Ultrathin sections were stained with uranyl acetate and lead citrate, and examined under a transmission electron microscope (902 A; Carl Zeiss, AG, Oberkochen, Germany). 

Scanning Electron Microscopy

After fixation, the excised bleb specimens were buffer rinsed, dehydrated in a graded ethanol series, and critical point dried through liquid CO₂ (CPD 030; Balzers Hochvakuum GmbH, Wiesbaden, Germany). Sheets with retina and RPE-choroid were mounted separately on aluminum stubs, sputter-coated with 30 nm gold (Sputter Coater S150B; Edwards High Vacuum, West Sussex, UK), and examined under a scanning electron microscope (DSM 950; Carl Zeiss, AG). The pathologist who processed the specimens for light and electron microscopy was not aware of the experimental protocol. 

Results

Photoreceptor–RPE Adhesion

Artificial retinal detachment was successfully achieved by injection of PSS into the subretinal space in all animals under all conditions. However, the size and height of the detachment bleb and the force necessary to induce the bleb varied among the different treatment modalities. It was most difficult to induce retinal detachment in the control group, where no prior vitrectomy had been performed (Table 1) . Vitrectomy followed by perfusion of a Ca²⁺-supplemented physiologic saline solution (PSS+) caused a subtle reduction of retinal adhesion (grade 8). A further reduction was achieved by perfusing the vitrectomized eyes with Ca²⁺-free PSS (PSS−), hence facilitating the retinal detachment moderately (grade 6). An accompanying ischemia induced by an elevated intraocular pressure did not further enhance the beneficial effect of the Ca²⁺-free solution (grade 6). In contrast, a solution that actively deprives the tissue of Ca²⁺ ions due to a supplemented chelating agent (EDTA) resulted in a substantially reduced retinal adhesion (grade 4). Preheating of the Ca²⁺-free PSS− was more effective than assumed (grade 3) and was even more successful in reducing the adhesive forces than was EDTA. 

Perfusion of the vitreous cavity with a hyperosmolar Ca²⁺-free solution accounted for the most pronounced effect on the retinal adhesion (grade 2). Under the present conditions, however, hyperosmolar physiologic saline (PSS; Acri.Tec) resulted in a rapid-progressive opacification of the posterior lens cortex. 

Histology and Ultrastructure

Retina from control eyes revealed scattered areas of adherent pigment that was clearly visible, even in the macroscopic overview (Fig. 1A) , and corresponded to fragmented RPE cells, as demonstrated in a light microscopy image (Fig. 1B) . In these eyes, Bruch’s membrane was partially denuded of RPE cells (Fig. 1C) . For a better depiction of this cellular damage, ultrastructural examination was performed and revealed areas of almost completely disrupted RPE plasma membranes and condensed actin filaments (Fig. 1D) . 

Similar damage was observed in vitrectomized eyes that had been perfused with Ca²⁺-containing PSS+. In the presence of physiological retinal adhesion, forced retinal detachment under vitrectomy conditions caused marked ultrastructural damage to both the photoreceptor outer segments (POS; Figs. 2A 2B ) and RPE cells (Fig. 2C) . Ultrastructural examination revealed swollen POS and mitochondria with partially opened and fragmented plasma membranes. Scanning electron microscopy demonstrated large sheets of pigment epithelium attached to the neural retina (Fig. 2C)and conversely, ragged POS adhering to the RPE cell layer in bundles (Fig. 2D) . 

In Ca²⁺-depleted conditions, the strong retinal adhesive forces were moderately diminished by PSS−. Consecutively, the retinal damage was reduced during the detachment procedure. No adhering pigment was detected macroscopically (Fig. 3A) . However, even though transmission electron microscopy showed the POS to be morphologically less damaged than in control eyes (Fig. 3B) , there were still widespread areas of POS adhering to the RPE layer (Fig. 3C) . Conversely, the damage to the RPE plasma membranes was still notable (Fig. 3D) , indicating that a Ca²⁺-free solution alone was insufficient to prevent substantial damage. 

To reduce further the amount of Ca²⁺ ions and hence the retinal adhesion, free Ca²⁺ ions in the vitreous and the subretinal interface can be actively withdrawn by using the chelating agent EDTA. Thus, the significantly reduced retinal adhesion during the bleb detachment resulted in a preserved retinal microarchitecture at the light microscopic level (Fig. 4A) . At the electron microscopic level, however, there was still evidence of swollen POS (Fig. 4B) . Of interest, the RPE layer showed an extensively vacuolated cytoplasm within a morphologically intact RPE cell, otherwise showing no alteration of the cellular organelles (Figs. 4C 4D) . Despite the active Ca²⁺ deprivation, the intercellular tight junctions (asterisk) were complete and attached. 

Although highly effective in neutralizing retinal adhesion, perfusion with hyperosmolar PSS− caused a rapid, progressive toxic effect, with shriveling of the retina (Fig. 5A)and fragmentation of the POS (Fig. 5B) . The RPE monolayer was virtually covered with a layer of cellular debris that most probably derived from shrunken and fragmented POS and RPE cells (Fig. 5C) . More important, both the RPE plasma membrane and the cellular organelles were completely disrupted. 

In contrast, preheating of the Ca²⁺-free PSS− to 34°C seemed to have an additive effect in reducing the adhesive forces and did not show any adverse effects. The induced retinal detachment was performed with ease and resulted in an almost intact RPE cell layer (Figs. 6A 6B) . Areas with POS adhering to RPE cells (Fig. 6C) , detached single RPE cells (asterisk) and limited, denuded areas of Bruch’s membrane (asterisks) were only occasionally seen (Fig. 6D) . 

Ischemia, however, coupled with the perfusion of Ca²⁺-free PSS, resulted in a sheetlike disruption of the RPE layer with large areas of RPE cells adhering to the neuroretina (Fig. 7A) . Large areas of Bruch’s membrane were denuded of the covering RPE cell layer showing both areas with the hexagonal pattern of disrupted cell remnants (asterisks) and areas that lay completely bare (asterisk) as a sign of separation of intact cells (Fig. 7B) . 

Discussion

A normal retina adheres firmly to the RPE through a variety of complementary mechanisms and does not separate unless considerable force is exerted. It has been shown that overcoming this physiological adhesiveness results in significant pathologic stress and might damage both the neural retina and the RPE. ^{16 17} Our study confirms notable cellular damage to these delicate structures if artificial retinal detachment is performed under vitrectomy conditions similar to surgery in humans. Although detaching the macula may appear to be atraumatic to the surgeon, ultrastructural examination showed irreversible damage, with extensive areas of ragged POS and disrupted RPE cells. Hence, modifying the retinal adhesiveness before an artificial retinal detachment seems to be crucial in making submacular surgery as atraumatic as possible. Several studies have demonstrated that retinal adhesion can be modulated under experimental ex vivo conditions. Work on animal models demonstrated that photoreceptor-RPE adhesiveness is reduced by low pH, low calcium, low magnesium, and dark adaptation, ¹⁷ but no standard procedure has been established so far. 

In our experimental setting, we could show that both the retinal adhesiveness and the cellular damage can be substantially reduced if various irrigating solutions are used under vitrectomy conditions. However, retinal adhesiveness responded differently to different conditions and disclosed a varying degree of toxicity. 

Several studies indicated that a low-calcium environment seems to be promising in reducing retinal adhesion in vivo ²⁵ and in enucleated human eyes, ²² as it weakens retinal adhesion in rabbits to approximately 30% of normal. ²⁴ Faude et al. ^{16 17} were the first to report that the in vivo use of a Ca²⁺-free irrigation solution notably limits the structural damage caused by an artificial retinal detachment. 

Although the mechanisms that reduce adhesion are not completely understood, changes in the interphotoreceptor matrix, actin filament contraction, and subretinal transport processes are thought to be involved. ^{24 29 30} If used for macular translocation surgery, Ca²⁺-free PSS at room temperature facilitated artificial detachment moderately, albeit not convincingly, and showed no adverse effects clinically. ⁷ This was confirmed in our experimental series showing that Ca²⁺-free PSS reduced the adhesive strength only moderately. The damage to the POS and adherent RPE to the retina were still notable, indicating that a Ca²⁺-free solution alone might be insufficient to prevent ultrastructural damage through artificial retinal detachment. 

Active Ca²⁺ (and Mg²⁺) deprivation by adding a chelating agent caused a further reduction of adhesiveness, to the extent that adherent pigment was completely avoided when artificial detachment was performed. Concurrently, the tight junctions remained intact, albeit highly Ca²⁺-dependent. ³¹ Nevertheless, active ion deprivation might be detrimental to the RPE cells, as multiple vacuolization within the RPE cytoplasm occurred. Although the significance remains unclear, a conclusive definition of a potential cytotoxicity should precede a further clinical evaluation. 

Similar limitations are valid for hyperosmolarity. Marmor ³² successfully produced a serous retinal detachment in rabbits 15 minutes after midvitreous injection of various hyperosmolar solutions up to 3000 mOsM. However, the author reported a toxic effect with permanent retinal degeneration and immediate deterioration of the electroretinogram. Similarly, our series showed that the intravitreous use of hyperosmolar solutions under vitrectomy conditions was extremely toxic. 

Marmor and Yao ²³ found that continuous oxygenation and active metabolic activity are vital to the maintenance of adhesive strength. However, the adhesive loss after transient ischemia described in the studies varies depending on the species used and the experimental conditions applied. ^{21 22 27 33} In our experiments, Ca²⁺-free PSS coupled with transient ischemia was not superior to Ca²⁺-free PSS alone in reducing adhesiveness. Under the present experimental conditions, ischemia preferentially resulted in disrupted RPE cells with sheetlike separation from Bruch’s membrane. This remarkable cellular damage may be explained by the longer ischemic period than in the series of Marmor and Yao ²³ ; but the efficacy of transient ischemia remains disappointing if used under vitrectomy conditions. 

Physical factors, such as temperature, are known to be modulators of the physicochemical properties of the interphotoreceptor matrix and have been shown to affect adhesiveness in donor eyes. ¹⁹ In vivo experiments, however, did not confirm the effect of temperature on retinal adhesion. ^{16 17} Contrary to these results, we found that a Ca²⁺-free PSS preheated to 34°C reduced retinal adhesion significantly, with minimal ultrastructural changes to the RPE cells or POS. Coupling the effect of low calcium with preheating of the solution seems to have an additive effect in regard to the reduction of adhesive forces, but not to the adverse effects. In view of a clinical applicability, it is also important that loss of adhesion under both elevated temperature and low calcium conditions is easily reversible within minutes by restoring normal environmental conditions. ^{25 33} However, the effect of intraocular temperature on the retinal function has not been conclusively defined yet: It has been shown that lowering the temperature of the eye below body temperature is accompanied by a decrease in peak and latency of the electroretinogram. ³⁴ On the other hand, it has been speculated that cooling the retina during vitrectomy may afford some protection against ischemic damage ³⁵ and phototoxicity. ³⁶  

In conclusion, our ultrastructural findings oppose the deliberate use of an artificial retinal detachment in eyes with unaltered retinal adhesiveness, as the cellular damage is considerable It can be shown that it is possible to reduce retinal adhesion artificially in vitrectomy conditions without causing significant damage to the photoreceptors or the RPE. But the long-term effects must be determined, if the method is to be used clinically with confidence. 

 Supported by the Köln Fortune Foundation and by a grant from the IZKF “BIOMAT” within the faculty of Medicine at the RWTH Aachen University.

 Submitted for publication April 28, 2006; revised June 7, 2006; accepted August 11, 2006.

 Disclosure: P. Szurman, None; S. Roters, None; S. Grisanti, None; S. Aisenbrey, None; U. Schraermeyer, None; M. Lüke, None; K.U. Bartz-Schmidt, None; G. Thumann, None.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Synopsis of the Irrigating Solutions and Conditions

Method	Solution	Osmolarity [mOsM]	Adhesive Strength	Adhesion RPE	Adhesion POS	Damage RPE	Damage POS
Control	None	—	10	++++	+++	+++	+++
Ca+	PSS+	312	8	+++	++++	++	+++
Ca−	PSS−	314	6	+	+++	++	++
Ca−−	PSS−/EDTA	300	4	−	+	+	++
Ca−/Osm	PSS−/Osm	500	2	−	−	++++	++++
Ca−/34°C	PSS−/Temp	314	3	+	+	+	+
Ca−/76mmHg	PSS−/Ischem	314	6	++++	++	+++	++

 Data show results for estimated retinal adhesive strength and the tissue damage. For grading of the intraoperative retinal adhesive forces, a scale of 0–10 was used, with 0 denoting spontaneous detachment and 10 denoting maximum force. Subgroups were: Ca+, Ca²⁺-containing solution; Ca−, Ca²⁺-free solution; Ca−−, Ca²⁺-depriving solution; Osm, hyperosmolar solution; Temp, preheated to 34°C; Ischem, ischemic conditions achieved by increasing IOP to 76 mmHg for 5 minutes.

Figure 1.

 

In control retinal sections, a macroscopic overview (A) of isolated neurosensory retina–facing photoreceptors shows considerable pigment adherence to the neural retina in the bleb area, indicating strong adherence of the retina to the RPE. Light microscopy (B) reveals that the adhering pigment corresponds to fragmented RPE cells (*). The basal fragments (*) of these disrupted RPE cells remain attached to Bruch’s membrane (C) with fragmented POS. Ultrastructural examination (D) of this area reveals details of damaged RPE cells with disrupted plasma membrane. The long, thin, dense structure in the basal residual part of this RPE cell represents a bundle of condensed actin filaments.

Figure 2.

 

With PSS+, light microscopy reveals an intact retinal architecture (not shown), but ultrastructural examination discloses swollen POS (A). Scanning electron microscopy of the neurosensory retina shows disruption of the photoreceptor layer with partial abrasion of outer segments (B). Correspondingly, areas of wide-spread abrasion of the RPE monolayer from Bruch’s membrane also show the harmful effect of mechanical separation of RPE and retina, if strong retinal adhesion is present (C). In other areas, scanning electron microscopy shows adherence of fragmented POS to the RPE layer (D).

Figure 3.

 

With the use of PSS−, a Ca²⁺-free solution, retinal adhesion was moderately reduced. Hence, the macroscopic overview shows the retina to be free of adhering pigment (A). (*) Myelinated nerve fibers close to the optic nerve head. POS appear less fragmented, but still swollen when observed at the electron microscopic level (B). Areas of POS adhering to the RPE layer appear to be less frequent, but are still present (C). Although the RPE appears as an intact monolayer at the light microscopic level (not shown), several areas with disrupted RPE plasma membrane (*) can be identified at the electron microscopic level (D).

Figure 4.

 

Active Ca²⁺ deprivation by EDTA (PSS−/EDTA) further reduced retinal adhesion, but led to an alteration of the RPE cell plasma, revealing multiple vacuoles of unknown origin. Retina with intact microarchitecture (A) reveals ultrastructurally bloated POS (B). POS adherence to the RPE is only minimal, but a highly vacuolated cell cytoplasm is evident (*), even at the light microscopic level (C). This is supported by the electron microscopic examination demonstrating an exhaustive vacuolization within morphologically intact RPE cells (D). Simultaneously, neither the cellular organelles nor the intercellular tight junctions (*) were altered.

Figure 5.

 

Hyperosmolarity (PSS−) resulted in shriveling of the retina (A) and notable fragmentation of the POS (B), with shrunken and deformed blood cells, fragmented POS, and cytolytic RPE remnants covering the RPE monolayer (C). Ultrastructurally, complete fragmentation of both the RPE plasma membrane (* *) and the cellular organelles (*) indicates the substantial toxic effect of hyperosmolarity (D).

Figure 6.

 

Ca²⁺-free solution at an elevated temperature (PSS−, preheated to 34°C) resulted in a moderate reduction of retinal adhesion that left the RPE monolayer mostly intact, as shown in light (A) and scanning electron microscope images (B). Areas of POS adherent to the RPE layer (C) and conversely, disrupted single RPE cells (*) or entire RPE sheets (* *) were detected occasionally (D), but were infrequent.

Figure 7.

 

Additional PSS−, IOP-induced ischemia caused widespread disruption of the RPE layer with exhaustive adhesion of RPE sheets (*) to the retina (A), resulting in large areas of denuded Bruch’s membrane (B). Bruch’s membrane is either completely bare (*) or shows the residual sole plate of disrupted cells, as indicated by the hexagonal pattern (* *).

The authors thank Judith Birch and Barbara Wallenfels-Thilo for excellent editorial assistance.