Human Platelet Suspension Stimulates Porcine Retinal Glial Proliferation and Migration In Vitro

Abstract

purpose. To characterize the cellular and molecular mechanisms underlying the efficacy of autologous platelet suspension adjuvant therapy in the treatment of macular hole.

methods. Platelet suspensions were prepared from whole blood samples obtained from informed volunteers. For proliferation assays, platelet suspensions or purified growth factors were added to semi-confluent cultures of porcine retinal glial cells for 24 hours, followed by[ ³H]thymidine for 15 hours, after which time cells were washed, solubilized, and counted for uptake of radioactive tracer. For cell migration assays, confluent glial cultures were scrape wounded and maintained in the presence or absence of platelet suspension or identified platelet constituents. Cell migration into the denuded area was scored as a function of time. In certain cases, specific pharmacologic inhibitors of growth factor action were added at the same time as platelet adjuvant or growth factors.

results. Platelet suspension adjuvant induced strong mitogenic and chemotactic responses in cultured glia, in a dose-dependent manner. Maximal incorporation of thymidine was two- to threefold that of control levels, with an ED₅₀ ∼5 × 10⁶ platelets/ml, and migration was enhanced up to 80-fold after 48 hours. Platelet suspension-induced proliferation was completely blocked by addition of 25 μM genistein, a tyrosine kinase receptor inhibitor. However, the same concentration only partially blocked the cell migration response. Addition of any single growth factor or protein identified from ELISA analysis, or a combination of all factors, did not significantly stimulate proliferation or cell migration.

conclusions. Human platelet suspensions exert both proliferative and chemotactic influences on retinal glial cells in vitro, suggesting that the same responses may occur in platelet-induced macular hole repair in humans. Growth factors or proteins that have been identified within the suspensions do not mimic these responses in vitro, implying that additional currently unidentified trophic activities are also present.

Macular holes (MH) are spontaneous breaks occurring at the central retina, with a prevalence of 3/1000 in people over 50 years of age and which are bilateral in 10% of patients. ¹ Retinal breaks occurring outside of the central retina can be treated through photocoagulation, permitting reattachment of torn retinal edges. ² However, because this approach produces tissue damage through inflammatory processes causing necrosis, it is not desirable in the central retinal field. Thus, MH remained largely untreatable until 1991, with the introduction of a surgical approach associating removal of the vitreous body, posterior hyaloid, and any epimacular substance present around the hole, followed by application of a gas tamponade for several days. ^{3 4} Today, the success rate of this surgery commonly attains or exceeds 70% to 80%. ^{5 6} This success rate was found to increase to >90% when purified native bovine tranforming growth factorβ ₂ (TGF-β₂) was injected into the operated area. ^{7 8} Such effects were not reproduced with recombinant TGF-β₂, and the approach was abandoned. ⁹ Renewed attempts at treating MH with combined vitrectomy and serum preparations, based on the reasoning that such fluids are a rich source of TGF-β, were highly successful, achieving wound closure in >90% of cases. ^{6 10 11 12 13 14 15 16}  

Glial cell involvement in MH repair was indicated from clinicopathologic studies in patients operated for MH ^{17 18} and by experiments in which surgically induced retinal tears in rabbits were found to be plugged by glia, retinal pigmented epithelia, and fibroblasts after injection of TGF-β₂ ¹⁹ or autologous serum. ²⁰ The use of serum in this animal model was found to encourage MH closure, with success rates > 90%. Serum components, especially fibronectin, have been shown to exert mitogenic and chemotactic influences on cultured rat retinal glia. ²¹ Thus, although the use of serum or derived treatments holds great promise for treatment of retinal breaks, their use is largely empiric, and the actual agents responsible for wound repair, as well as the target cells themselves, are mostly unknown. The goal of the present study was to use an in vitro model of one of the probable cellular sites of action, the retinal glia, to clarify the biologic effects engendered by human platelet suspensions (PS) and to examine possible candidate growth factors and serologic proteins for their precise roles in these effects. The results suggest that retinal glia are strongly influenced by PS both to proliferate and to migrate and that these activities are due to both tyrosine kinase receptor (TKR)-activated and non–TKR-activated pathways. Interestingly, these effects could not be attributed to any of the growth factors or proteins currently identified in PS, suggesting the presence of additional trophic agents. 

Materials and Methods

Retinal Glial Cell Culture

Whole eyes were obtained from freshly killed pigs at a local abattoir and transported rapidly to the laboratory on crushed ice. Neural retina, free of retinal pigmented epithelium, were dissected from the globes and placed in serum-free Dulbecco’s modified Eagle’s medium (DMEM; Gibco BRL Life Technologies, Cergy-Pontoise, France). Retinal glia were isolated and cultured from retinal cell suspensions using previously published techniques. ²² Cells were harvested from 50% Percoll gradients (Pharmacia Biotech, Uppsala, Sweden), examined for viability by trypan blue exclusion, and seeded into 10-cm culture dishes in DMEM supplemented with 10% fetal bovine serum (FBS), with one retina per dish. The medium was replaced with fresh DMEM/10% FBS after 24 hours to remove Percoll beads and again after 4 days. After 7 to 8 days in vitro, cells were trypsinized and replated in fresh medium at 5 to 10 × 10⁴/cm² in 24 × 18- and 6 × 35-mm tissue culture plates. Glass coverslips were included in some wells for use in immunocytochemistry. After 3 days in vitro, medium was replaced by a chemically defined serum-free medium for a further 2 days [DMEM supplemented with insulin/transferrin/selenium mix (Sigma-Aldrich, Saint Quentin Fallavier, France) and sodium pyruvate]. 

Human Platelet Suspension Preparation

PS was prepared from fresh blood samples drawn from informed fasted volunteers (members of the laboratory). PS was prepared using the same protocol for MH surgery ¹² : 16 ml venous blood was collected on 4 ml anticoagulant (ACD-A: citric acid–dextrose–formula A) and centrifuged at 150g for 10 minutes. Then 5 ml of platelet-rich plasma supernatant was mixed with 0.6 ml ACD-A and centrifuged at 1500g for 10 minutes to obtain pelleted platelets. The platelet-poor plasma supernatant was removed, and the pellet was resuspended in 0.75 ml 0.9% NaCl. Such preparations typically contained ∼10⁹ platelets/ml and in most assays were used within 12 hours. In some studies, it was necessary to store platelet suspensions at 4°C for 24 to 48 hours. 

Growth Factors and Inhibitors

Based on enzyme-linked immunosorption assay detection of different growth factors and serologic constituents present in PS, ^{12 23} the following purified growth factors and proteins were purchased: native porcine platelet-derived growth factor (PDGF), native porcine TGF-β1, recombinant human RANTES (all from R & D Systems, Abingdon, UK); tissue culture grade human epidermal growth factor (EGF) and recombinant human basic fibroblast growth factor (FGF-2) (both from Euromedex, Souffelweyersheim, France); native purified platelet factor 4 and β-thromboglobulin (both from STAGO, Paris, France); and native purified human plasma fibronectin (Sigma-Aldrich). Anti-TGF-β (all isoforms) and PDGF-neutralizing antibodies also were purchased from R & D Systems. Anti-thrombospondin–neutralizing antibody was the generous gift of C. Legrand, INSERM U.353, Paris, France. The following growth factor inhibitors also were purchased: genistein (pan-TKR blocker ²⁴ ), 5′-methylthioadenosine (MTA, specific blocker of FGFR ²⁵ ) (both from Sigma-Aldrich); staurosporine (specific blocker of PDGFR ²⁶ ), and tyrphostin 23 (specific blocker of EGFR ²⁷ ) (both from Euromedex). 

Cell Proliferation Assays

PS, defined growth factors, or pharmacologic agents were added directly to the culture medium. Each reagent was tested at a range of dilutions to construct dose–response curves, and control wells were treated with buffer or vehicle alone. After 36 hours,[ ³H]thymidine (1 μCi/well, in 10 μl, specific activity, 50 Ci/nmol; ICN Pharmaceuticals Inc., Costa Mesa, CA) was added for 12 hours, and the cultures were rinsed three times in phosphate-buffered saline (PBS) and solubilized in 0.5 ml 1 M NaOH. Aliquots were mixed with scintillation cocktail and counted in a liquid scintillation counter (1211 minibeta; Pharmacia LKB, Piscataway, NJ). Each experiment was repeated independently a minimum of three times, with each treatment being performed in triplicate or quadruplicate wells. 

Cell Migration Assays

One milliliter of fresh, chemically defined medium containing the test reagent (PS, growth factor, protein, or inhibitor) was added to 35-mm wells containing confluent glial cultures. In some experiments, anti-thrombospondin or TGF-β–neutralizing antibodies were added together with PS, in excess of the estimated antigen concentrations present in PS (40 μg/ml anti-thrombospondin and 20 μg/ml anti–TGF-β for 30 × 10⁶ platelets). Immediately after, a small area of the culture was denuded by scraping with a 0.5-mm-wide sterile blade. Microscopic observation ensured the complete removal of glial cells within this area. The mark produced in the plastic substrate was used as the zero point, and cell movement across this border was recorded as the number of cells per microscope field at 24, 48, and 72 hours. Each experiment was performed independently at least twice, in duplicate wells for each treatment. 

Immunocytochemistry

To confirm the purity of passaged glia used for these studies, coverslips containing confluent cultures were fixed for 15 minutes in 4% paraformaldehyde in PBS, rinsed, and permeabilized with 0.1% Triton X-100 in PBS. After blocking for 10 minutes in PBS containing 0.1% bovine serum albumin and 0.1% Tween 20 (buffer A), cells were incubated for 2 hours with anti-glial fibrillar acidic protein polyclonal antibody (GFAP) (DAKO S.A., Trappes, France) diluted 1:400 in buffer A. After washing, antibody binding was visualized with goat anti-rabbit IgG/BODIPY FL (Molecular Probes Europe BV, Leiden, The Netherlands), 10 μg/ml in buffer A for 1 hour. Coverslips were washed thoroughly, mounted, and observed using a Nikon Optiphot 2 fluorescence microscope (Nikon, Melville, NY). 

Statistics

Measures of statistical significance were performed using the parametric Peritz f test for two populations. ²⁸ In the present study the accepted levels of probability were* P < 0.01 and **P < 0.001. 

Results

Culture Purity

Labeling of fixed cells with anti-GFAP antibody revealed that> 99% of cells present were immunopositive, showing strong staining of the cytoskeleton (Fig. 1) . Contamination of cultures with other cell types including neurons, fibroblasts, and microglia was <0.5% (data not shown). 

Mitogenic Effects of PS

Addition of PS to quiescent glial cultures led to average two- to threefold increases in thymidine uptake (Fig. 2) . The maximal response was elicited by 26 × 10⁶ platelets/ml, whereas higher concentrations (>50 × 10⁶/ml) led to a marked drop off from the plateau. The ED₅₀ was ∼5 × 10⁶ platelets/ml. In preliminary efforts to characterize the molecular nature of these mitogenic effects, pelleting of platelets from freshly prepared PS revealed that the remaining platelet-poor supernatant exerted only small mitogenic effects (data not shown). PS was heated to 100°C for 10 minutes before addition to glia, and such treatment completely abolished mitogenic activity. Dialysis (molecular weight cutoff, 8 kDa) of freshly purified PS for 24 hours only slightly reduced activity (Fig. 2) . Repeated cycles of freeze-thawing followed by dialysis as above gave the same results. PS stored for 24 hours at 4°C retained full activity compared to freshly prepared samples, and ACD-A alone had no effect (data not shown). 

A series of trials was conducted to examine the contribution of growth factors contained within the PS to the observed proliferative effect. Simultaneous addition of PS (26 × 10⁶ platelets/ml) and increasing concentrations of genistein showed a dose-dependent inhibition of glial proliferation. Genistein (25 μM) led to >95% inhibition of glial thymidine uptake, whereas serial dilutions of the inhibitor led to gradual recuperation of the PS-induced response (ED₅₀ ∼3 μM) (Fig. 3) . Direct toxic effects of genistein were excluded by initial trials examining trypan blue exclusion after 24 hours in the presence of the drug: concentrations < 50 μM were not toxic, whereas higher doses led to decreased viability (data not shown). We also tested potential inhibitory effects of molecules reported as specific growth factor blockers. All the following inhibitors were tested for direct toxicity by trypan blue exclusion as described above and at the concentrations used showed no deleterious effects on glial viability after 24 hours (data not shown). Staurosporine (1 μM), tyrphostin-23 (100 μM), and MTA (500 μM) all completely inhibited the PS-induced increase in thymidine uptake (Fig. 4) . Staurosporine (1 μM) alone did not affect basal uptake, whereas tyrphostin-23 (100 μM) alone and MTA (500 μM) alone both significantly decreased basal uptake levels (by 70% and 30%, respectively). Simultaneous addition of PS and neutralizing anti-PDGF antibody did not significantly reduce the mitogenic effects compared to PS alone (95% of the stimulatory effect remained) (Fig. 3) . Addition of the vehicle buffer alone (0.1% ethanol) had no effect. 

Mitogenic Effects of Identified Growth Factors

To examine the role of growth factors known to be present in the PS on glial proliferation, these factors were tested at a wide range of concentrations covering those estimated to be present in PS, either alone or in combination, and in the presence or absence of inhibitory drugs. PDGF (0.8–50 ng/ml) led to slight increases in glial uptake of thymidine (maximal 10% to 20% increase relative to basal levels), representing <20% of PS assays run on parallel wells (Fig. 5A ). TGF-β₁ used at 0.04 to 10 ng/ml did not significantly alter (either increase or decrease) glial proliferation in these cultures (Fig. 5B) . PDGF and TGF-β₁ also were tested at higher concentrations, up to 80 ng/ml either alone or in combination, but in no case was glial proliferation significantly stimulated compared to controls (data not shown). RANTES applied at 0.8 to 50 ng/ml also had no effect (Fig. 5C) . FGF-2 or EGF (0.3 and 1 ng/ml, respectively) did not stimulate glial proliferation, whereas at 50 ng/ml FGF-2 was strongly mitogenic but EGF had only small effects (Fig. 5D) . Combination of several factors at the concentrations estimated to be present in PS by ELISA (PDGF, 80 ng/ml; TGF-β₁, 80 ng/ml, RANTES, 80 ng/ml; FGF-2, 0.8 ng/ml; EGF, 0.8 ng/ml) also did not stimulate glial cell proliferation (Fig. 5D) . 

Cell Migration Effects of PS, Growth Factors, and Proteins

A second aspect of PS effects on glial behavior concerned glial migration. Confluent glial cultures were scrape wounded and maintained in the presence of different test substances (Fig. 6) . Compared to control cultures (defined medium or ACD-A only), addition of PS induced a dose-dependent increase in glial migration, with an ED₅₀ ∼5 × 10⁶ platelets/ml (Fig. 7) . PS was very effective at stimulating cell migration, with an average 16-fold increase in cell numbers and a maximal effect of 80-fold, traversing the scrape border by 48 hours (Figs. 6 8) . Addition of the different inhibitors at the same time as PS gave results different from those observed for thymidine incorporation. Genistein led to only partial inhibition of cell migration, 20% less than adjuvant alone but still ∼55-fold higher than nontreated controls (Figs. 6 8) . Tyrphostin-23 and staurosporine (100 and 1 μM, respectively) reduced migration by 17% and 30%, respectively, compared to adjuvant alone. MTA (500 μM) led to complete block of cell migration, but this was probably due to long-term toxic influences as glia died and detached from the culture dishes at 48 hours (Fig. 8) . Addition of purified growth factors and cytokines, either alone or in combination (see above) (PDGF, TGF-β₁, EGF, FGF-2, and RANTES) had no significant effect on glial migration (Figs. 6 8) . Simultaneous inclusion of PS and blocking anti-TGF-β antibody did not differ from PS alone (Fig. 8) . As with mitogenic assays, preheating of PS at 100°C for 10 minutes nullified completely effects on migration (Fig. 9) . Addition of the purified platelet-associated protein platelet factor 4 or β-thromboglobulin did not stimulate migration, and inclusion of anti-thrombospondin antibodies with PS did not reduce cell migration compared to PS alone (Fig. 9) . Anti-thrombospondin antibody alone had no effect, whereas addition of 5% FBS enhanced migration. Addition of increasing concentrations of fibronectin (0.1–100 μg/ml) produced small increases (maximal twofold) in glial migration compared to parallel control cultures (Fig. 10) . 

Discussion

The data presented in this study indicate that human platelet suspensions (PS) are capable of stimulating both proliferation and migration of cultured porcine retinal glial cells. These effects are mediated through both tyrosine kinase receptor (TKR)- and non–TKR-linked pathways. Surprisingly, those growth factors and proteins that have been identified in PS do not mimic these effects when applied as purified substances, suggesting that additional as yet uncharacterized molecules are responsible. 

Treatment of macular hole (MH) through a combination of vitrectomy and autologous blood preparations has been very successful (success rates > 90%) and is becoming generalized throughout Europe. There are currently no data available on success rates of adjuvant therapy in different forms of MH, but given the high success rate in treatment of simple MH by surgery alone, ^{3 4 5 6} such approaches would be particularly interesting in cases of recurrent MH or failed surgery. Except for histopathologic observations in humans ^{18 29} and in experimental rabbit models, ^{19 20} the possible cellular and molecular pathways through which PS exerts its effects are largely unknown. The original reason for testing PS was based on its known rich content in TGF-β, since TGF-β₂ had been shown to be effective in earlier in vivo studies, ^{5 6} although recombinant TGF-β₂ later proved to be ineffective. ⁷ In turn, the justification for testing TGF-β came from its published effects on stimulating extracellular matrix synthesis and trophic effects for many cells. ³⁰ In the in vitro trials used here, TGF-β is unlikely to contribute significantly to glial responses for several reasons. TGF-β₂ concentrations in PS would probably have been too low (∼10 pg: ^{Ref. 12)} to induce any cellular division. Neutralizing pan-TGF-β antibody was unable to inhibit PS-stimulated migration, which is important in light of the presence in PS of high levels of thrombospondin, a known activator of latent TGF-β. ³¹ Furthermore, heating is known to convert the latent precursors of TGF-β into active forms, ³² whereas in our studies heating destroyed PS effects. Finally, TGF-β₁ even when used at high concentrations was ineffective in stimulating glial proliferation or migration. TGF-β₁ is known to stimulate glycosaminoglycan synthesis in cells ³³ and in cultured Müller glia (unpublished results) and so may still be involved in wound repair by stimulating synthesis of matrix components. In the present study we did not examine the possible effects of PS or identified proteins on other cell types thought to be present in MH plugs, such as retinal pigment epithelium and fibroblasts, which may produce additional trophic factors. 

The general characteristics of PS-induced glial proliferation evoke growth factor-mediated responses. PS activity was dose-dependent, showing a sigmoid curve reaching a saturable plateau and with higher doses actually leading to a reduction in the biologic effect, as has often been observed for known growth factors (e.g., FGF-2 ³⁴ ). The activity also could be destroyed by heating but was not removed by dialysis, suggesting a polypeptidic nature of the active components. The mitogenic effects of PS were completely blocked by treatment with the pan-TKR blocker genistein, as well as with more specific growth factor inhibitors. Genistein is thought to block TKR activation through perturbing association of phospholipase C-γ1 with membrane receptors. ²⁴ Genistein has been previously reported to inhibit FGF-2 stimulation of corneal endothelial cell proliferation. ³⁵ The doses required to inhibit cell growth in the two studies were very different: whereas we observed∼ 75% inhibition of PS effects with 10 μM genistein, this same dose had no significant effect on FGF-2–stimulated corneal endothelial proliferation. Staurosporine is reported to be 100 times more potent at inhibiting PDGF than EGF in fibroblasts. ²⁶ Tyrphostin-23 has been published as a selective inhibitor of EGF receptors in keratinocytes. ²⁷ 5′-Methylthioadenosine (MTA) has been shown to specifically inhibit FGF-2–induced fibroblast proliferation through binding to the FGF receptor. ²⁵ These three inhibitors all fully suppressed the mitogenic effect of PS, yet the specificity of these pharmacologic probes in the present model is doubtful, a conclusion mainly based on the data that purified PDGF, EGF, and FGF-2 neither stimulate porcine retinal glial proliferation or migration when used at doses similar to their concentration in PS, nor do they further stimulate such activities when added to PS. Although the effects of MTA may be due to direct toxicity, those of staurosporine and tyrphostin may function through inhibiting other TKR or additional signaling molecules such as protein kinase C. ³⁶ Further studies will be necessary to determine the specific actions of these drugs in retinal glial cells. 

None of the growth factors known to be present in PS-evoked significant responses in porcine retinal glial cultures. TGF-β was dealt with earlier and is known to either stimulate or inhibit satellite cell proliferation in the presence of PDGF, FGF-2, and EGF. ³⁷ We were surprised by the general lack of effect of PDGF, which is known to be a powerful mitogen for many cell types including glia. ³⁸ PDGF has been shown to stimulate rat ²¹ and human ^{39 40} retinal glial proliferation, while having no effect on rabbit ⁴¹ or guinea pig ⁴² retinal glia. Because human retinal glia are sensitive to PDGF, this factor may contribute to MH repair in humans but was relatively ineffective in either of the in vitro assays used in the present study, despite the use of platelet-purified protein from the homologous species (pig). In these studies, maximal effects of PDGF were only ∼20% of those observed for PS and were obtained with 5- to 10-fold higher doses of PDGF than would have been present in PS. In addition, anti-PDGF neutralizing antibody did not reduce PS-induced glial proliferation, and PDGF is heat stable, ⁴³ whereas PS activity was heat labile. It should be noted that although PDGF has no mitogenic effect on rabbit retinal glia, ⁴¹ PS did produce healing of retinal breaks in this species in vivo, ²⁰ suggesting that growth factors other than PDGF are actively stimulating glial growth. EGF and FGF-2 were only present in trace amounts and at these concentrations were without effect. Interestingly FGF-2 was a strong mitogen for pig retinal glia when used at 3 nM, whereas EGF still only induced weak effects. This was not due to loss of activity of EGF, inasmuch as in parallel cultures it strongly stimulated proliferation of rat retinal glia (unpublished results), which are known to contain high numbers of EGF receptors. ⁴⁴ None of the different combinations of some or all growth factors identified in PS produced a notable effect on cultured porcine retinal glial proliferation or migration. As the TKR inhibitor genistein was such an effective blocker of PS-induced glial proliferation, these data suggest that other growth factors acting through TKR-activated pathways are present and necessary for the biologic effects. 

In contrast to proliferative effects of PS, genistein was only a mild blocker of PS-induced glial migration. The majority of the effect remained after genistein treatment, suggesting the predominance of non-TKR pathways in mediating this behavior. As with mitogenic influences, these activities were dose-dependent, heat-labile, and resistant to dialysis. Growth factors known to be present in PS did not significantly stimulate glial migration. Purified cytokines and proteins known to be abundant in PS [RANTES ⁴⁵ ;β -thromboglobulin, ⁴⁶ platelet factor 4, ⁴⁷ and thrombospondin ⁴⁸ ] were all without effect on glial migration. Serum fibronectin adsorbs readily to the surface of platelets and is well known to stimulate cellular adhesion and migration, ²¹ but had only a small effect in the present study. The concentrations tested were in the range of the fibronectin content of PS, calculated from western blot analysis of PS and serial dilutions of purified fibronectin (∼2 μg/30 × 10⁶ platelets, data not shown). However, it should be noted that platelet-associated fibronectin differs from that in serum. ⁴⁹ There are several additional possibilities that were not tested in the present study. PS also contains small amounts of vascular endothelial (CD, unpublished results) and insulin-like growth factor 1. ¹⁰ CD9 is present in platelets ⁵⁰ and has also been shown to be expressed in the central nervous system, where it is thought to play multiple roles. ⁵¹ Hence, it is a possible candidate for stimulating migration of adult retinal glial cells. 

In conclusion, human PS stimulate both proliferation and migration of porcine retinal glia in vitro, and it is reasonable to suppose this reflects the mode of action in repairing MH in vivo. Whereas the proliferative effects could be entirely accounted for by TKR-dependent growth factors, migratory effects were mostly due to non-TKR activity. Further studies are in progress to isolate and characterize the active component(s) of PS. 

 Supported by Fédération Nationale des Aveugles et Handicapés Visuels, La Fondation de l’Avenir, Université Louis Pasteur, DRET/DGA, ADRET, L’Etablissement Français des Greffes, and ESSILOR International.

 Submitted for publication April 26, 1999; revised July 30, 1999; accepted September 15, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Immunocytochemical labeling of primary cultured porcine retinal glia, 6 days in vitro. (A) Nomarski image of elongated cells; (B) anti-glial fibrillar acidic protein antibody labeling of same field. Arrow depicts representative cell. Bar, 10μ m.

Figure 2.

 

Dose–response curve of PS-induced thymidine incorporation by porcine retinal glial cells. PS were added at increasing concentrations (numbers on x-axis refer to platelet numbers × 10⁶/ml) or after dialysis (PS dial) or boiling (PS 100°C). Control wells received vehicle buffer (ACD-A) alone.* P < 0.01, **P < 0.001, relative to control wells.

Figure 3.

 

Genistein inhibits PS-induced glial proliferation in a dose-dependent manner. Co-incubation of a fixed amount of PS (26 × 10⁶/ml) and serial dilutions of genistein (gen, 0.3–25μ M, values on x-axis) showed a dose-dependent inhibitory effect of genistein. Genistein concentration of 25 μM reduced thymidine incorporation (expressed as percentage maximal incorporation of PS alone) to those of control wells. Inclusion of anti-PDGF antibody with PS (PS/aPD) did not reduce incorporation relative to PS alone.

Figure 4.

 

Effects of specific growth factor receptor inhibitors on PS-induced glial proliferation. Co-incubation of PS with different inhibitors (Staurosporine, SS; Tyrphostin 23, TP; 5′-methylthioadenosine, MTA) all led to reduction of thymidine uptake to control levels (P values above PS/SS, PS/TP, and PS/MTA relative to PS 26, <0.001). SS alone did not affect proliferation, whereas both TP and MTA inhibited proliferation relative to untreated wells (P values above TP and MTA relative to controls,<0.001). PS alone significantly stimulated proliferation relative to controls (**P < 0.001).

Figure 5.

 

Dose–response effects of PDGF (A), TGF-β₁ (B), RANTES (C), and FGF-2/EGF/mixed growth factors (D) on glial proliferation. Addition of each growth factor/cytokine over a wide range of concentrations did not lead to marked glial mitogenic responses. Only FGF-2 used at 50 ng/ml evoked a strong response (**P < 0.001).

Figure 6.

 

Photomicrographs of representative PS treatments on porcine retinal glial migration. Confluent glial cultures were scrape-wounded (dotted line between large arrows), and cell numbers colonizing the denuded area scored after 24, 48, or 72 hours. (A) Control well, 48 hours; (B) PS, 26 × 10⁶ platelets/ml; (C) PS, 26 × 10⁶ platelets/ml and 25 μM genistein; (D) TGF-β₁, 20 ng/ml. Notice the large numbers of cells in (B) and (C) (platelets are visible as small dots), but not in (A) or (D). Scale bar, 30 μm.

Figure 7.

 

Dose- and time-dependent PS stimulation of glial migration. Increasing amounts of PS (3, 10, or 30 × 10⁶ platelets/ml) were added to scrape-wounded wells, and migrating cells were counted at different times. Compared to control wells (scrape-wounding only), both higher concentrations were significantly different at all times examined, and the lower concentration was significantly different at 48 and 72 hours (**P < 0.001). The migratory effect of PS also follows a saturable dose–response curve.

Figure 8.

 

Effects of defined growth factors and growth factor receptor blockers on retinal glial migration. Whereas PS (30 × 10⁶ platelets/ml) led to an ∼16-fold rise in migrating cell numbers by 48 hours, defined growth factors did not influence this parameter, either alone or in combination. Inclusion of neutralizing TGF antibody with PS did not reduce migration relative to PS alone. Genistein led to only partial, although statistically significant, block, as did staurosporine and tyrphostin (*P < 0.01,** P < 0.001 relative to PS alone). MTA completely blocked cell migration at this time, but glia appeared unhealthy and MTA was probably directly toxic.

Figure 9.

 

Effects of heat treatment, purified platelet proteins and blocking antibodies on glial migration. Boiling of PS completely removed stimulation of migration (**P < 0.001 relative to PS 10 × 10⁶ platelets/ml alone). Purified PF-4 andβ -TG were without effect, and co-incubation of PS with neutralizing TSP antibody was not different from PS alone. Fetal bovine serum (FBS) did stimulate glial migration (*P < 0.01 relative to control).

Figure 10.

 

Effects of purified fibronectin on glial migration. Scrape-wounded cells were maintained in the presence of fibronectin (FN) (0.1–100 μg/ml) or defined medium and scored for migrating cells per microscope field at 24, 48 and 72 hours. Data are expressed as a percentage relative to controls at 24 hours. FN led to small but significant increases in glial migration, especially at higher concentrations.

The authors acknowledge expert technical assistance by Valérie Forster for tissue cultures and René Marshall for photographic work.