Modulation of Retinal Pigment Epithelial Cell Behavior by Agaricus Bisporus Lectin

Abstract

purpose. To determine whether Agaricus bisporus lectin (ABL) binds retinal pigment epithelial cells (RPEs), to conduct a preliminary viability study of RPEs exposed to ABL, and to evaluate the effects of ABL on RPE proliferation and RPE-mediated matrix contraction in vitro.

methods. Using cultured bovine RPEs, immunohistochemistry was used to study ABL binding. Morphologic and trypan blue exclusion techniques were used for toxicity studies. The effect of ABL on RPE proliferation was investigated by [methyl-³H]–thymidine incorporation. The effect of ABL on RPE-mediated matrix contraction was evaluated with RPE-populated three-dimensional collagen matrices.

results. ABL bound to RPE cells. This binding was inhibited by asialomucin. No change in RPE morphology or trypan blue exclusion compared with controls was observed in RPEs incubated with 5 to 60 μg/ml ABL for 3 days. Twenty-four-hour incubations of RPEs with ABL significantly inhibited RPE proliferation in a dose-dependent way, 40 μg/ml ABL inhibited proliferation by 83% (SE 14, P < 0.05). ABL showed a dose-dependent significant inhibition of RPE-mediated collagen matrix contraction over 3 days, with 93% inhibition compared with controls by 40 μg/ml lectin (P < 0.05). The inhibitory effect of ABL on proliferation and gel contraction was partly reversible after eliminating ABL from the culture medium.

conclusions. Bovine RPE cells bind ABL, and preliminary evaluations suggest that levels of ABL that are nontoxic to the cells potently inhibit RPE proliferation and RPE-mediated matrix contraction. ABL deserves further investigation as a potential inhibitor of RPE proliferation and cell-mediated matrix contraction in anomalous reparative processes such as proliferative vitreoretinopathy and as a laboratory tool for RPE behavioral studies.

Proliferative vitreoretinopathy (PVR) is an anomalous wound repair process typified by the formation of scar-like membranes on the retinal surfaces. ¹ The membranes may contain abundant retinal pigment epithelial cells (RPEs). ¹ RPE proliferation and RPE-mediated tissue contraction are thought to be fundamental to PVR membrane development. Drug prevention of PVR membrane formation has been based chiefly on antiproliferative and anti-inflammatory agents, but the results of such treatments have been disappointing, largely because these agents tend to be ineffective and toxic to the retina. ¹ There is a need for nontoxic agents that will specifically block cellular activities such as RPE proliferation and RPE-mediated membrane contraction in PVR. 

Lectins are ubiquitous carbohydrate-binding non-immunoglobulin proteins. They bind noncovalently to carbohydrates and are readily purified from a wide variety of sources. A range of carbohydrates occurs on all cell surfaces, and lectins have been used to explore cell membranes and to distinguish different cell types, because cells express distinct carbohydrates that can be detected by specific lectins. ² Furthermore, lectins binding to cell surface carbohydrates may affect the behavior of the cell. Thus, lectins such as peanut agglutinin and the lectin of the edible mushroom (Agaricus bisporus lectin; ABL), which both bind to the carbohydrate structure galactosyl β-1,3-N-acetyl-galactosamine, modulate the proliferation of malignant epithelial cells. ³ Peanut agglutinin increases colonic carcinoma cell division, whereas ABL inhibits proliferation of these cells. ³ Moreover, ABL inhibits proliferation of a range of other cells including Tenon’s capsule fibroblasts, and it inhibits contraction of collagen matrices by Tenon’s fibroblasts. ⁴ These effects occur without apparent cytotoxicity. ^{3 4} Because cell-mediated membrane contraction and proliferation are key RPE activities in PVR, we evaluated the effect of ABL on these RPE activities in vitro, having first determined whether ABL binds to (or affects the viability of) cultured RPEs. 

Materials and Methods

All reagents were of analytical grade, all lectin and peroxidase-conjugated lectin were obtained from EY Laboratories Ltd (San Diego, CA) and [methyl-³H]–thymidine was supplied by Amersham International (Amersham, UK). 

RPE Culture

Bovine RPEs were obtained and cultured as previously described. ⁵ Established cultures were maintained with minimal essential medium (MEM) containing glutamine and fungizone, penicillin and streptomycin, and 15% newborn calf serum (NCS) (GIBCO Europe Ltd., Paisley, UK). The cultures were kept at 37°C in the presence of 5% CO₂ and air. The RPEs reached confluence within 1 to 2 weeks, and subcultures between the fourth and seventh passages were used in the present study. The purity of the cultures was confirmed on the basis of cytokeratin staining as described previously. ⁵  

All experiments were conducted in MEM and 2% NCS. Some serum components in NCS bind ABL (e.g., IgA, fetuin), ⁶ neutralizing its activity, but 2% serum was required to maintain RPE viability. 

Lectin Histochemistry

Bovine RPEs were grown on eight chamber tissue culture glass slides (LabTeks; Nalge Nunc International, Life Technologies Limited, Glasgow, UK) as reported previously. ⁵ Peroxidase- or fluorescein isothiocyanate (FITC)–labeled ABL was added to the chambers at a concentration of 30 μg/ml for 1 or 8 hours, respectively. Controls consisted of preincubation of the labeled ABL with 10 mg/ml asialomucin for 5 minutes before the addition of the labeled lectin to the cells and preincubation of the chambers with 100μ g/ml unconjugated ABL for 1 hour before adding the conjugated lectin. The slides were washed with phosphate-buffered saline (PBS), and either peroxidase was developed with diamine benzidine or slides were mounted after fixation with methanol for 7 minutes at 20°C. Slides were dehydrated, mounted, and evaluated using DIC optics (Polyvar, Reichert–Jung, Austria) or epifluorescent photography, respectively. 

Morphologic Evaluation and Trypan Blue Staining

RPEs were seeded in 24-well plates (Corning Costar, High Wycombe, UK) at a concentration of 2 × 10⁴ cells/well. After 1 day, the cells were washed three times with MEM without serum (to remove the serum transferred with the maintenance medium). ABL was added in concentrations ranging from 5 to 60 μg/ml in MEM with 2% NCS. Controls were kept in MEM with 2% NCS. Three wells were used for each concentration, and the experiment was repeated twice. Cell morphology was evaluated daily for 3 days by phase contrast microscopy. Representative preparations were selected each day and stained with 2% trypan blue for 5 to 10 minutes. Stained and unstained cells were counted in each well. 

Cell Proliferation Assay

Cells were seeded in 24-well plates at a density of 1× 10⁴/well. After incubation for 48 hours, the wells were washed three times with PBS after which ABL was added (20, 30, 40 μg/ml in 0.5 ml MEM with 2% NCS). Cells were incubated for a further 24 hours. The cells then received a 1-hour pulse with 0.5μ Ci/well [methyl-³H]–thymidine. Each well was washed twice with PBS before cell precipitation with 0.5 ml/well of 5% trichloroacetic acid at 4°C. The precipitate was washed once with 5% trichloroacetic acid at 4°C and twice with 0.5 ml/well of 95% ethanol at 4°C and left to dry at room temperature. After solubilization in sodium hydroxide (NaOH), 0.3 ml of the precipitate was added to 1 ml Optima Gold MV scintillation cocktail (Packard, Pangbourne, UK), and the cell-associated radioactivity was determined using a Packard scintillation counter. 

Recoverability of inhibition of proliferation was evaluated by cell counting. Cells were seeded in 24-well plates as described above. After 24 hours, the cells were washed twice with PBS, and ABL at concentrations between 20 and 40 μg/ml was added in 2% NCS/MEM. After a further 24 hours, the lectin was removed by washing once with NCS and twice with PBS. Cells were then incubated for up to 7 days with 15% NCS/MEM and sample wells counted in triplicate daily. Cell counts were performed by washing the cells with PBS, trypsinizing with phosphate-buffered trypsin/EDTA at 37°C for 10 minutes, and counting in a hemocytometer. 

Collagen Matrix Contraction

For collagen matrix contraction experiments, the method published by Mazure and Grierson ⁷ was adapted to 24-well plates. Briefly, rat tail type I collagen (Sigma, Poole, UK) was dissolved in 0.1% (vol/vol) acetic acid in sterile distilled water. RPEs were counted after harvesting from maintenance cultures and then resuspended in MEM at a volume of 1.23 ml containing 5.76 × 10⁶ cells, sufficient for one 24-well plate. The cell suspension was mixed with 4.91 ml of 5 mg/ml collagen and with 2.86 ml of concentrated serum-free MEM containing glutamine, antibiotics, and NaOH. The collagen-cell mixture was then transferred in 350-μl aliquots to 24-well plates, ensuring that the matrix covered the bottom of the wells. The solution polymerized rapidly when incubated at 37°C in the presence of 5% CO₂, thus trapping the cells (at a density of 2.4 × 10⁵ RPEs per matrix) within the three-dimensional matrix. The matrices were detached from the edges and allowed to float in the wells by the addition of 1 ml of MEM with 2% NCS. Mushroom lectin was added at concentrations of 5, 10, 20, 30, and 40 μg/ml medium, and the preparations were incubated at 37°C in 5% CO₂ in air for at least 3 days. The surface area of each matrix was recorded photographically daily. Four wells were used for each concentration, and experiments were repeated three times. After 3 days, medium containing ABL was removed and gels were washed twice and then were incubated for an additional 4 days with medium containing 15% of serum. 

Statistical Analysis

ANOVA was used to explore between-group significance. Duncan’s multiple comparison test was used to detect homogeneous subsets. 

Results

Binding of RPEs by ABL In Vitro

Lectin histochemistry demonstrated specific binding of ABL to bovine RPEs, equally distributed by two different techniques (FITC and peroxidase). After 8 hours’ incubation, there was perinuclear accumulation of labeled ABL, best seen with FITC-labeled ABL (Fig. 1 A). This binding was abolished by preincubation with 100 μg/ml of unlabeled lectin and by 10 mg/ml asialomucin (Fig. 1B) . After only 1 hours’ incubation, labeled ABL was seen to be adherent to cell surfaces, best visualized by the peroxidase technique (Fig. 1C) . This binding and uptake also were abolished by preincubation with 100μ g/ml of unlabeled lectin (Fig. 1D) and by 10 mg/ml asialomucin. 

Cell Viability Study

No cell morphologic change was noticed during the 3 days of incubation with ABL (5–60 μg/ml) or in controls (Fig. 2) . Staining with trypan blue revealed more than 95% RPE viability after 3 days’ incubation at all concentrations of ABL from 5 to 60 μg/ml and in the controls (exact data not shown). 

Inhibition of RPE Proliferation by ABL In Vitro

ABL caused a significant dose-dependent inhibition of proliferation as assessed by thymidine incorporation (one-way ANOVA, P = 0.0002; F = 66; Fig. 3 A). Compared with wells incubated with media without ABL, 20 μg/ml ABL produced a 40% (SE 10) inhibition of thymidine incorporation in RPEs, whereas 40 μg/ml ABL caused an 83% (SE 14) inhibition. Duncan’s test for multiple comparisons indicated that each treatment group was homogeneous (P < 0.05). 

After removal of ABL from the medium, an increase in cell number was demonstrated for all ABL concentrations (Fig. 3B) . However, cells exposed to higher ABL doses exhibited a slower proliferation rate. Thus, by day 6, although the control cells (incubated without ABL) had reached a density of 17.04 ×10⁴ ± 0.92 cells/well, cells incubated with 40 μg/ml ABL had reached 10.40 × 10⁴ ± 1.12 cells/well (P < 0.05). 

Inhibition of Contraction of RPE-Populated Collagen Matrices by ABL

ABL caused a significant concentration-dependent inhibition of contraction (one-way ANOVA, P < 0.001, F = 14; Fig. 4 ). Duncan’s test for multiple comparisons indicated that 5 to 10μ g/ml ABL was moderately inhibitory (12% matrix contraction at day 3) and that 20 to 40 μg/ml ABL caused greater inhibition (6.2%—2.5% contraction at day 3; P < 0.05). Calculating inhibition as [1 − (test contraction/control contraction)] × 100, ABL in the range 5 to 40μg/ml produced inhibition of 43% to 93% (Fig. 4A) . 

After elimination of ABL from the medium, gel contraction was partly restored (Fig. 4B) . Percentage of gel contraction was calculated as[ gel size (day 3) − gel size (day 7)]/gel size (day 3) × 100. Compared with control gels that were not preincubated with lectin, the ability to contract collagen gels was restored to 90.9% for 5μ g/ml ABL, to 83.4% for 10 μg/ml ABL, to 64.7% for 20 μg/ml ABL, and to 23.2% for 40 μg/ml ABL. 

Discussion

Our results demonstrate that ABL binds to RPE. RPEs adhere a number of lectins, including peanut agglutinin, ⁸ which binds to galactosyl β-1,3-N-acetyl-galactosamine (the oncofetal Thomsen Friedenreich antigen TFa). Because ABL also binds this antigen, we were not surprised to observe ABL binding to RPEs. The changes in the pattern of binding with time indicate that bound ABL is internalized, a finding consistent with observations concerning the fate of ABL in other cell types. ⁹  

The binding and uptake of ABL by RPE cells do not appear to alter RPE viability in vitro. RPEs incubated for 3 days with ABL concentrations up to 60 μg/ml (the highest studied level) show normal trypan blue exclusion and no discernible light microscopic change in morphology compared with control cultures. In addition, our data indicate that the effects of ABL on RPE proliferation and cell-mediated matrix contraction are recoverable after removal of the lectin from the media. Taken together, these results suggest that ABL has little or no toxicity for cultured RPEs. Indeed, evidence from studies of other cell types suggests that ABL generally is noncytotoxic at doses up to 200μ g/ml. ³  

Despite the apparent lack of effect of ABL on RPE viability in vitro, the lectin markedly inhibits RPE proliferation and RPE-mediated collagen matrix contraction in vitro even at concentrations of less than 30 μg/ml. The mechanism by which ABL modulates cell activities in general is not clear, although there is evidence that it interferes with nuclear protein import. ⁹ Whatever the mechanism of action, the lectin is known to influence the behavior of a variety of cell types. For example, the proliferative activity of colonic and breast carcinoma cells and mammary and Tenon’s fibroblasts ^{3 4} is, like that of RPEs, inhibited by ABL. However, in contrast to our observation that ABL inhibits RPE-mediated collagen matrix contraction and to a finding that ABL inhibits matrix contraction by Tenon’s fibroblasts, ⁴ a previous study involving dermal fibroblasts reports no effect of ABL on matrix contraction. ¹⁰ The latter investigation used serum at a concentration five times the level we used; because some serum components bind ABL (e.g., IgA), ⁶ the ABL may have been functionally neutralized by the high serum concentration. 

Inhibition of ABL activity by serum components could limit the potential antiproliferative and anticontractile use of the lectin (e.g., for PVR) in the presence of blood. Hemorrhage or leakage of plasma components may occur during PVR formation or surgery. ¹¹ On the other hand, it is possible that ABL binding to RPEs is augmented in PVR. PVR is a complication of retinal detachment, and evidence from PNA binding studies suggests that TFa expression by RPEs is augmented in retinal detachment. ⁸ Thus, ABL may represent a means of specifically controlling RPE activities in PVR without retinal toxicity. Indeed, given the lectin’s potential dual action on RPE proliferation and RPE-mediate membrane contraction and its dose-dependent titratability of effect on RPE behavior, the lectin deserves further investigation as a possible agent for the management of PVR and other anomalous wound repair disorders. Moreover, the lectin may represent a new laboratory agent with which to study further the relationship between RPE behavior and retinal pathology generally. 

 Submitted for publication January 21, 1999; revised April 27, 1999; accepted May 20, 1999.

 Commercial relationships policy: P.

 

Figure 1.

 

Distribution of ABL binding to cultured RPEs. (A) After 8 hours’ incubation (FITC-labeled); (B) ABL binding is abolished by preincubation of labeled lectin with 10 mg/ml asialomucin (magnification ×600, FITC-labeled); (C) after 1 hours’ incubation (peroxidase-labeled ABL); (D) ABL binding is abolished by preincubation of the cells with unlabeled lectin (magnification ×600, peroxidase labeled).

Figure 2.

 

Trypan blue staining of RPE cultures after a 3-day incubation with medium alone (A) or under the presence of 60 μg/ml ABL (B). Cells didn’t stain with trypan blue, and no morphologic differences were noticed between both culture conditions.

Figure 3.

 

Inhibition of RPE cell proliferation by Agaricus bisporus lectin. Thymidine incorporation by cultured RPE cells exposed to ABL for 24 hours. Cell proliferation (A) in medium without lectin was set to 100%. Bars represent the mean of two experiments each performed in triplicate. Error bars represent SEM. Recoverability of ABL effect on RPE cells measured by cell counting (B). One-hundred percent represents increase in cell numbers without previous lectin incubation. Bars represent the mean of two experiments each performed in triplicate. Error bars represent SEM.

Figure 4.

 

Inhibition of contraction of RPE-populated collagen matrices by Agaricus bisporus lectin. Contraction of RPE-populated collagen matrices after 3 days (A). Gel contraction in the absence of lectin was set to 100%. Bars represent mean of three experiments each performed in triplicate. Error bars represent SEM. Recoverability of ABL effect on RPE-populated collagen matrices measured by gel contraction after incubation for 4 days in medium without lectin (B). One-hundred percent represents gel contraction without previous lectin incubation. Bars represent the mean of two experiments each performed in triplicate. Error bars represent SEM.