TGF-β2 in Aqueous Humor Suppresses S-Phase Entry in Cultured Corneal Endothelial Cells

Abstract

purpose. Corneal endothelium in vivo is arrested in G1, the phase of the cell cycle that prepares cells for DNA synthesis. In many cell types, transforming growth factor (TGF)-β inhibits proliferation by inducing G1-phase arrest. Evidence indicates that corneal endothelial cells synthesize mRNA for TGF-β1 and are also bathed in aqueous humor that contains TGF-β2 (mainly in a latent form). As such, this cytokine may maintain the corneal endothelium in a G1-phase–arrested state in vivo. The purpose of these studies was to determine the effect of exogenous TGF-β2 and TGF-β2 in aqueous humor on DNA synthesis in cultured corneal endothelial cells.

methods. Rat corneal endothelial cells were grown in explant culture and identified by morphology and reverse transcription–polymerase chain reaction using primers for specific corneal cell markers. Subconfluent cells were synchronized in the G0 phase (quiescence) by serum starvation for 24 hours. Serum was then added to the cells in the presence or absence of exogenous TGF-β2 or activated rat aqueous humor. [³H]Thymidine was added, and radioactivity was measured at various time points to detect DNA synthesis. Preincubation of exogenous TGF-β2 or activated rat aqueous humor with neutralizing antibody was used to test for cytokine specificity.

results. A linear increase in [³H]thymidine incorporation began approximately 16 hours after serum addition, and peak incorporation occurred at approximately 24 hours. Exposure of cells to serum plus TGF-β2 suppressed [³H]thymidine incorporation in a dose-dependent manner at concentrations ranging from 5 pg/ml to 5 ng/ml. [³H]Thymidine incorporation was also suppressed in cells exposed to serum plus rat aqueous humor diluted 1:10. Neutralizing antibody reversed the effects of both exogenous TGF-β2 and aqueous humor.

conclusions. Exogenous TGF-β2 and TGF-β2 in aqueous humor suppress S-phase entry of rat corneal endothelial cells. These results suggest that this cytokine in aqueous humor could help maintain the corneal endothelium in a G1-phase–arrested state in vivo.

We hypothesize that corneal endothelium is maintained in a growth-arrested state because of multiple antiproliferative factors, including those in the endothelial microenvironment. Evidence from studies of neonatal rats indicates a correlation between the formation of stable cell–cell and cell–substrate contacts and cessation of proliferation in corneal endothelium, suggesting an in vivo growth-arrest mechanism similar to that of contact inhibition in cultured cells. ¹ Within the anterior chamber, aqueous humor mediates immune privilege and helps regulate ocular cell proliferation, differentiation, and wound healing. ^{2 3 4} Because aqueous humor also bathes the corneal endothelium, it may contribute to regulation of its proliferation. Factors in aqueous humor that may contribute to this regulation include fibroblast growth factor (FGF), ^{5 6 7} epidermal growth factor (EGF), ^{5 6} transforming growth factor-β (TGF-β), ^{7 8} insulin-like growth factor-I (IGF-I), ^{9 10} platelet-derived growth factor, ^{11 12} endothelin 1, ^{13 14} and prostaglandin E₂. ¹⁵ Of the factors present in aqueous humor, TGF-β is of particular interest because it inhibits the G1-to-S-phase transition in several cell types ¹⁶ through the activation of cyclin-dependent kinase inhibitors. ¹⁷ TGF-β2 is the main TGF-β isoform in aqueous humor ^{8 18} and has been shown to inhibit lens epithelial cell proliferation. ^{19 20} Human, rabbit, and rat corneal endothelial cells in vivo express the three TGF-β receptor types (RI, RII, and RIII) needed to bind TGF-β2 and to transduce a TGF-β–induced signal. ^{21 22} TGF-β1 and -β2 mRNA and protein have been detected in corneal endothelium, lens epithelium, trabecular meshwork endothelium, and ciliary epithelium by immunocytochemistry and reverse transcription–polymerase chain reaction (RT-PCR), respectively, and are secreted by trabecular cells and ciliary epithelium in vitro. ^{23 24 25} Regulation of cell growth and metabolism by TGF-β2 through autocrine–paracrine mechanisms is well documented in some cell types, but the relationship between TGF-β2 and aqueous humor and corneal endothelium remains unclear. 

The purpose of the current studies was to determine whether TGF-β2 may contribute to regulation of proliferation in cultured corneal endothelial cells. For these studies, we cultured rat corneal endothelial cells and performed [³H]thymidine bioassays to determine the effect of exogenous TGF-β2 and TGF-β2 in rat aqueous humor on S-phase entry. 

Methods

Rat Corneal Endothelium and Fibroblast Explant Cultures

Adult male rats, treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, were used as a source of corneal cells. Corneas were removed and tissues were cultured using protocols essentially the same as those used for culture of rabbit corneal endothelium. ²⁶ To remove the epithelial cell layers, corneas were incubated for 1 hour at 37°C in 2.5 mM EDTA to disrupt hemidesmosomes located between basal cells of the epithelium and the underlying extracellular matrix. ²⁷ The epithelium was then removed as a sheet and stored at −80°C. For explant culture of the endothelium, corneas were cut in half and placed endothelium-side down in a 6-well tissue culture plate. Pieces were allowed to attach to the tissue culture plastic for approximately 5 minutes, after which 1 drop of culture medium was placed over the tissue. Culture medium, prepared according to an established protocol, ²⁶ consisted of Medium 199 (Gibco–Life Technologies, Gaithersburg, MD), 50 μg/ml gentamicin (Gibco), 25 ng/ml FGF (Biomedical Laboratories, Stoughton, MA), and 10% fetal bovine serum (FBS; Hyclone, Logan, UT). Corneal pieces were incubated overnight at 37°C in a 5% CO₂ humidified atmosphere. 

On the following day, 1 ml culture medium was gently added and cultures were incubated undisturbed at 37°C. After 5 days, 2 ml medium was added per well. Medium was changed every other day thereafter. After approximately 10 days, when a sufficient number of endothelial cells had migrated off the cornea, the corneal pieces were carefully removed, and the remaining endothelial cells were grown to confluence. For culturing of stromal fibroblasts, the epithelium was removed as described, and the endothelium was removed by mechanical scraping. The stroma was cut into small pieces that were then incubated in a 6-well plate in the same culture medium. 

Cell Identification

Cells were identified morphologically by phase-contrast microscopy and by RT-PCR detection of mRNA for specific markers, including collagen VIII as a cell marker for corneal endothelium ²⁸ ; keratin 14, a marker for corneal epithelium ²⁹ ; and decorin, a marker for both corneal endothelium and epithelium, but not for stromal fibroblasts. ³⁰ Any endothelial cell cultures found to be contaminated by fibroblasts were discarded. 

For RT-PCR studies, total RNA was prepared from confluent cultures of endothelial cells and stromal fibroblasts and from frozen corneal epithelium using reagent according to the manufacturer’s directions (TRIzol; Gibco). cDNA was prepared from 1 μg total RNA by RT in a volume of 20 μl using reagents from a commercially available kit (Promega, Pittsburgh, PA). Primers specific for each cell marker were designed using software (MacVector 5.0; Oxford Molecular Group, Oxford, UK). The primer pairs ³¹ used were: collagen VIII: upstream sequence, 5′-CCAAAGGAGAAGGTGGAGTTG-3′, downstream sequence, 5′-ACAATTCCTGGGATACCTGGGG-3′; keratin 14: upstream sequence, 5′-GATAAGGTGCGTGCCTGGAAG-3′, downstream sequence, 5′-CCATCTGCTCATACTGGTCCCTC-3′; and decorin: upstream sequence, 5′-ATATCTATGTGCCCCTACCGA-3′, downstream sequence, 5′-TTGCCGCCCAGTTCTATGACA-3′. 

PCR was performed in a reaction mixture containing 1 μg cDNA and 2.5μ M each of the upstream and downstream primers, plus reagents from a commercially available kit (Gibco). Specificity and yield of the PCR products were enhanced using the hot-start approach. ³² For all three experiments, PCR was performed for 40 cycles in a thermal cycler. Cycle conditions included denaturation at 95°C for 1 minute, annealing at the required temperature for 1 minute, and extension at 72°C for 2 minutes. A 5-minute extension was added at the end of the 40 cycles of PCR. Annealing temperatures were as follows: collagen VIII, 57.8°C; keratin 14, 58.4°C; and decorin, 53.2°C. PCR products and 100-bp DNA ladder molecular weight markers were electrophoresed in 1.5% agarose gels containing 0.5 μg/ml ethidium bromide and photographed. Glyceraldehyde-3-phosphate dehydrogenase (G3PDH; Clontech, Palo Alto, CA) acted as a positive control for the PCR. Negative control samples consisted of the PCR reaction mixture, including primers, but without cDNA. To ensure that the total RNA samples were not contaminated with genomic DNA, a negative control using 1 μg total RNA was substituted for cDNA in the PCR reaction mixture, along with 2.5 μM each of upstream and downstream G3PDH primers. No total RNA samples used in these studies yielded G3PDH PCR product under these conditions, indicating that the PCR products were obtained from mRNA and not from contaminating genomic DNA (data not shown). 

Collection of Rat Aqueous Humor

Rat eyes were proptosed with hemostatic forceps, and a small incision was made in the limbus with a no. 10 scalpel blade. With a small, flame-polished glass pipette, the cornea was penetrated through the incision paralleling the iris–lens plane. Aqueous humor was passively obtained by capillary action without touching the iris, lens or corneal endothelium. Aspiration of aqueous humor was halted before the anterior chamber collapsed. Aqueous humor specimens were collected in siliconized microfuge tubes and immediately placed on dry ice for transport to a −80°C freezer for storage. Less than 10 μl aqueous humor was obtained from one rat eye. Thirty rats were needed to collect enough volume of aqueous humor for use in[ ³H]thymidine bioassays. 

[³H]Thymidine Bioassay

Confluent primary cultures of rat corneal endothelial cells were trypsinized and seeded at a density of 10⁴ cells per well in a 96-well tissue culture plate. This seeding density was optimal for maintaining a subconfluent population of cells throughout the experiment. Cells were incubated overnight at 37°C in culture medium containing 10% serum to ensure attachment to the tissue culture plastic. Attached cells were synchronized in the G0 (quiescent) phase by incubation for 24 hours in serum-free medium. After synchronization, cells were maintained for specific periods of time in culture medium containing 10% serum, supplemented or not with exogenous porcine TGF-β2 (R& D Systems, Minneapolis, MI) or rat aqueous humor. Media used for the synchronization step and for the[ ³H]thymidine bioassay did not contain FGF, as did the primary culture medium. [³H]Thymidine (NEN Life Science Products, Boston, MA) was added at a concentration of 25 μCi/well, and cells were incubated for specific periods of time, depending on the experiment. Cells were then trypsinized for 30 minutes at 37°C and [³H]thymidine incorporation was measured in counts per minute (cpm) using a liquid scintillation counter (model 1205 Betaplaten; Wallac, Gaithersburg, MD). Duplicate wells were used for each time point and treatment condition. Experiments were repeated two to three times. In some experiments, culture medium was incubated overnight at 37°C with an excess of TGF-β2 neutralizing antibody (R&D Systems) before addition to the cells. 

Analysis of Data

Data are presented as mean ± SEM. Student’s paired t-test (two-tailed) was used to analyze scintillation counts. P < 0.05 was considered to be statistically significant. 

Results

Identification of Rat Corneal Endothelium

Phase-contrast microscopy and RT-PCR verified that cells obtained from the explant cultures of rat corneal tissue were corneal endothelial cells. Confluent cultures consisted of a monolayer of generally polygonal cells that exhibited morphologic characteristics remarkably similar to those consistently obtained in confluent cultures of rabbit corneal endothelial cells. ²⁶ In contrast, confluent cultures of stromal fibroblasts exhibited a typical dense, spindle-shaped appearance (Fig. 1) . Therefore, by morphology alone, it was easy to distinguish endothelial cells from fibroblasts. RT-PCR for collagen type VIII, decorin, and keratin 14 distinguished endothelial from epithelial cells. As anticipated, PCR products of the size expected for collagen VIII and decorin were obtained from cultured endothelial cells. No PCR product was observed for keratin 14, a marker for epithelial cells (Fig. 2 and Table 1 ). 

Kinetics of S-Phase Entry

[³H]Thymidine incorporation studies first determined the kinetics of S-phase entry in rat corneal endothelial cells grown in serum alone. Subconfluent cultures were synchronized in the G0 phase by incubation for 24 hours in serum-free medium. Serum was then added to a final concentration of either 1% or 10% and, at the same time, [³H]thymidine was added at a concentration of 25 μCi/well. Samples were taken for scintillation counting every 8 hours for 48 hours. Figure 3 is a representative example of [³H]thymidine incorporation in the presence of 1% serum. Under these conditions, radioactive counts were not detectable until approximately 16 hours after serum addition. Incorporation then increased in a linear fashion up to approximately 24 hours after which counts plateaued. These results suggest that, when stimulated with 1% serum, cells took a minimum of 16 hours to traverse the G1 phase. S-Phase entry was initiated approximately 16 hours after serum addition and was maximal at 24 hours. 

For cells grown in 10% serum, the kinetics of S-phase entry were similar to those of cells grown in 1% serum (compare Figs. 3 and 5 ). As expected, the amplitude of [³H]thymidine incorporation was greater with 10% serum, indicating that the extent of DNA synthesis was serum concentration dependent. Cultures grown in 10% serum showed an additional small peak of[ ³H]thymidine incorporation approximately 8 hours after serum addition. As will be discussed, this peak may have been caused by a wave of DNA synthesis in a small population of unsynchronized cells. 

Effect of Exogenous TGF-β2 on S-Phase Entry

One set of experiments was conducted to determine the effect of exogenous TGF-β2 on [³H]thymidine incorporation. For these studies, subconfluent cultures of rat corneal endothelial cells were synchronized in serum-free medium as described. Culture medium plus 10% serum was added alone or with TGF-β2 at concentrations ranging from 0.005 to 95.2 ng/ml.[ ³H ]Thymidine (25 μCi/well) was added 16 hours after serum addition, and cells were harvested for scintillation counting 8 hours later. As seen in Figure 4 , TGF-β2 suppressed [³H]thymidine incorporation over a broad concentration range. A shallow linear decrease in incorporation was observed between 0.005 ng/ml (the lowest concentration tested) and 3.5 ng/ml. At concentrations higher than 3.5 ng/ml, [³H]thymidine incorporation plateaued. Even at a concentration of 0.005 ng/ml, TGF-β2 exerted an almost 50% reduction in [³H]thymidine incorporation, compared with that in untreated control samples, reflective of the apparent sensitivity of these cells to this cytokine. In subsequent experiments, TGF-β2 was used at a concentration of 5 ng/ml, because this concentration consistently produced a maximal suppressive response. 

A second study compared the kinetics and extent of S-phase entry in cells incubated in the presence or absence of TGF-β2. In this study, cells were synchronized in serum-free medium for 24 hours as described. Medium containing [³H]thymidine (25 μCi/well) was added to each well together with 10% serum alone or serum supplemented with 5 ng/ml TGF-β2. Cells were then harvested every 8 hours for a period of 32 hours. As seen in Figure 5 , radioactive counts increased linearly between 16 and 24 hours after serum addition. In cultures supplemented with TGF-β2,[ ³H]thymidine incorporation was suppressed by approximately 65% at the 24-hour time point and by approximately 35% by 32 hours after serum addition. Pretreatment with a neutralizing antibody against TGF-β2 reversed this suppressive effect. 

Effect of Rat Aqueous Humor on S-Phase Entry

Because aqueous humor normally contains TGF-β2, studies were conducted to determine the effect of aqueous humor on[ ³H]thymidine incorporation. Rat aqueous humor was collected and then subjected to three rounds of freeze–thaw to activate latent TGF-β. Rat corneal endothelial cells were synchronized in the G0 phase as has been described. Cells were then placed in culture medium containing 10% serum alone or in the same medium supplemented with aqueous humor diluted 1:2, 1:5, or 1:10.[ ³H]Thymidine was added, and cells were harvested for scintillation counting every 8 hours for 40 hours. Results showed a dose-dependent suppression of[ ³H]thymidine incorporation (data not shown). Figure 6 shows the suppressive effect of a 1:10 dilution of aqueous humor. This dilution inhibited [³H]thymidine incorporation by approximately 34% compared with untreated control samples (P = 0.002). This effect was very similar to that obtained in the previous experiment when exogenous TGF-β2 was added to the culture medium (P > 0.1). The suppressive effect of aqueous humor was totally reversed after pretreatment of aqueous humor with neutralizing antibody that specifically binds TGF-β2, indicating that the suppressive effect of aqueous humor was caused by this cytokine. 

Discussion

The corneal endothelium is an excellent model system for studying cell cycle regulation, both in vivo and in tissue culture. This is the first known report of the isolation and culture of rat corneal endothelial cells. Rat corneal endothelial cells grow readily in culture, permitting an examination of cell-cycle regulatory mechanisms.[ ³H]Thymidine incorporation studies indicate that, in cells grown in 1% serum, the minimum length of the G1 phase in these cells was 16 hours. [³H]Thymidine incorporation peaked at 24 hours, indicating that most responsive cells entered the S-phase within the 16- to 24-hour period. The finding of peak [³H]thymidine incorporation at 24 hours is within the range for cultured corneal endothelial cells observed by others. ^{33 34} Unpublished observations from this laboratory indicate that approximately 6% of the population of rat corneal endothelial cells are refractive to serum starvation, even for 48 hours. In contrast to cells grown in 1% serum, cultures grown in 10% serum consistently yielded a small peak of[ ³H]thymidine incorporation approximately 8 hours after serum addition. These counts may result from the initiation of DNA synthesis in this apparently refractive, nonsynchronized cell population. In cultures grown in 1% serum, this peak may not have been detected because of the lower amplitude of[ ³H]thymidine counts obtained with this serum concentration. 

The finding that TGF-β2 suppresses S-phase entry in rat corneal endothelial cells confirms and extends results of previous studies from this laboratory ³⁵ in which flow cytometry and bromodeoxyuridine staining showed that exogenous TGF-β1 and TGF-β2 induce suppression of S-phase entry in rabbit corneal endothelial cells. The finding of TGF-β–induced suppression of the S phase in corneal endothelial cells is similar to results of studies in other cell types in which TGF-β has been shown to arrest cells in the mid to late G1 phase of the cell cycle. ³⁶ The results of the current studies appear to be physiologically relevant in that aqueous humor suppressed S-phase entry to an extent similar to exogenous TGF-β2, the concentrations at which TGF-β2 was suppressive were within the range of concentrations reported for this cytokine in aqueous humor, and the suppressive effect of aqueous humor was reversed by preincubation with TGF-β2–neutralizing antibody. These findings suggest that, of the multiple growth factors present in this ocular fluid, TGF-β2 is most likely to contribute to the antiproliferative microenvironment of the endothelium. 

Investigators have had different results when studying the effect of TGF-β on proliferation of corneal endothelial cells. Using a methylene blue absorbance assay, Hongo et al. ³⁷ reported that TGF-β1 at concentrations of 0.1 to 10 ng/ml did not influence growth of rabbit corneal endothelial cells in culture. In contrast, Plouet and Gospodarowicz ³⁸ and Rieck et al. ⁷ found that TGF-β1 stimulated proliferation in cultured bovine and human corneal endothelial cells, respectively. A number of factors most likely contribute to the different results obtained. One is that different species of corneal endothelial cells were studied. It is possible that the same cell type from different species may have different receptor affinities for TGF-β1 or TGF-β2 or may process TGF-β–induced signals differently. The relative effects of TGF-β1 and TGF-β2 may be different or may differ with the concentrations tested. Another important difference between the studies reported here and those of Plouet and Gospodarowicz and Rieck et al. is in the specific treatment protocol and method of assessing the TGF-β–induced effects. In the protocol used for the studies reported here, cells were synchronized in serum-free medium to enrich for a mitotically quiescent population before exposure of the cells to TGF-β2. In addition, the effect of TGF-β2 was studied by assessing the kinetics of [³H]thymidine incorporation. In the other studies, cells were incubated in the presence of TGF-β for several days, after which the effect of this cytokine on proliferation was determined by cell counting. 

Our data indicate that exogenous TGF-β2 suppresses S-phase entry of rat corneal endothelial cells in a dose-dependent fashion at concentrations ranging from 0.005 to 5 ng/ml. This finding appears to be physiologically relevant, because Ishida et al. ³⁹ reported that the total TGF-β2 concentration in rat aqueous humor is 3.15 ± 1.25 ng/ml. Similar results were reported by Hu et al., ⁴⁰ who found that TGF-β2 inhibited DNA synthesis of cultured uveal melanocytes at concentrations of 0.03 to 10 ng/ml. In general, TGF-βs in vivo are secreted in a latent form and subsequently become activated; therefore, the relative amount of active TGF-β2 available to the corneal endothelium in vivo may be significantly less than the total amount of TGF-β2 measured in aqueous humor. Investigators have detected a range of active TGF-β concentrations in this ocular fluid. Ishida et al. ³⁹ reported the active form of TGF-β2 in rat aqueous humor to be 580 ± 140 pg/ml. Cousins et al. ⁴¹ measured the amount of mature TGF-β in human, rabbit, bovine, porcine, and mouse aqueous humor and found a range from undetectable levels in mouse to 495 pg/ml in rabbit. The finding in the current studies that TGF-β2 suppressed S-phase entry by approximately 50% at the lowest concentration tested (5 pg/ml) suggests that even low concentrations of active TGF-β2 could affect proliferative activity of rat corneal endothelium. A number of mechanisms are known to activate TGF-β, including denaturing treatments, plasmin-mediated proteolytic cleavage of TGF-β latency-associated peptide, ionizing radiation, or interaction with thrombospondin ^{42 43 44} ; however, the mechanism that activates latent TGF-β in aqueous humor is still poorly understood. Corneal endothelial cells themselves may be capable of activating latent TGF-β, because they express, on their surfaces, proteases such as plasmin, ⁴² calpain, ⁴³ and thrombospondin, ⁴⁴ that can activate this cytokine. 

In summary, we have established rat corneal endothelial cells in tissue culture and have demonstrated that both exogenous TGF-β2 and active TGF-β2 in rat aqueous humor can inhibit S-phase entry in these cells. Based on these findings, we suggest that, at least in rat, TGF-β2 in aqueous humor contributes to the G1-phase arrest observed in corneal endothelium in vivo. As stated earlier, we believe that the corneal endothelium in vivo is maintained in a growth-arrested state by multiple antiproliferative factors. We have now identified at least two factors that contribute to this state: stabilization of cell–cell and cell–substrate contacts, and TGF-β2; however, this does not exclude the very likely possibility that other, intrinsic mechanisms also help maintain G1-phase inhibition in these cells. The relevancy of our finding that TGF-β2 induces suppression of S-phase entry in rat corneal endothelial cells remains to be directly demonstrated in human corneal endothelium. 

 Supported by Grant RO1 EY05767 from the National Eye Institute (NCJ) and the Pearle Vision Foundation (NCJ).

 Submitted for publication September 18, 1998; revised May 25, 1999; accepted June 28, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Phase-contrast micrographs of confluent rat corneal endothelial cells (A) and stromal fibroblasts (B) demonstrate differences in morphology of these two cell types. Confluent cultures of endothelial cells exhibit a generally polygonal morphology, whereas confluent stromal fibroblasts exhibit a dense, spindle-shaped appearance. Magnification, ×110.

Figure 2.

 

RT-PCR detection of mRNA for G3PDH (lanes 1 through 3), collagen type VIII (lanes 5 through 7), decorin (lanes 9 through 11), and keratin 14 (lanes 13 through 15) in cultured rat corneal epithelial cells (lanes 1, 5, 9, and 13), stromal fibroblasts (lanes 2, 6, 10, and 14), and corneal endothelial cells (lanes 3, 7, 11, and 15). A 100-bp cDNA ladder is shown in lanes 4, 8, and 12. The brightest band in each ladder represents 800 bp.

Table 1.

 

RT–PCR Characterization of Cultured Corneal Cells

PCR Product	Epithelium	Fibroblasts	Endothelium
G3PDH (452 bp)	+	+	+
Collagen VIII (274 bp)	−	−	+
Decorin (410 bp)	+	−	+
Keratin 14 (488 bp)	+	−	−

Figure 3.

 

[³H]Thymidine bioassay reveals the kinetics of S-phase entry in culture medium containing 1% FBS. Values are the mean ± SEM of three separate experiments conducted in duplicate.

Figure 4.

 

TGF-β2 dose-dependent suppression of [³H]thymidine incorporation in rat corneal endothelial cells. Values for[ ³H]thymidine incorporation are the mean ± SEM of three separate experiments conducted in duplicate.

Figure 5.

 

Kinetics of TGF-β2–induced suppression of[ ³H]thymidine incorporation. Rat corneal endothelial cells were cultured in medium containing 10% FBS alone (⋄), in the same medium supplemented with 5 ng/ml TGF-β2 (□), or in medium containing TGF-β2 preincubated with 60 ng/ml TGF-β2 neutralizing antibody (▴). Values are the mean ± SEM of three separate experiments conducted in duplicate.

Figure 6.

 

Suppression of [³H]thymidine incorporation by TGF-β2 in aqueous humor. Rat corneal endothelial cells were cultured in medium containing 10% FBS alone (♦), in the same medium containing rat aqueous humor diluted 1:10 (□), or in medium containing FBS and 1:10 aqueous humor preincubated with 300 ng/ml TGF-β2–neutralizing antibody (▵). Values are the mean ± SEM of two separate experiments conducted in duplicate.