TGF-β1, -β2, and -β3 In Vitro: Biphasic Effects on Tenon’s Fibroblast Contraction, Proliferation, and Migration

Abstract

purpose. To compare the effects of the three human transforming growth factor-β (TGF-β) isoforms and different concentrations of TGF-β on human Tenon’s capsule fibroblasts (HTF), with a view to delineating the role of this growth factor in the subconjunctival scarring response after glaucoma filtration surgery.

methods. Application of recombinant human TGF-β1, -β2, and -β3 (range 0–10⁻⁸ M) was assessed using several assays of HTF function: fibroblast-mediated collagen contraction, proliferation, and migration.

results. All three isoforms of TGF-β behaved in a similar manner in vitro. They each stimulated HTF-mediated collagen contraction, proliferation, and migration with a characteristic concentration-dependent response, with peak activities at 10⁻⁹, 10⁻¹², and 10⁻⁹ M, respectively, that were significantly different from control (P < 0.05). At concentrations above and below peak activities, HTF activity was reduced, demonstrating biphasic effects of TGF-β.

conclusions. TGF-β1, -β2, and -β3 have similar actions in vitro; this is demonstrated by their effects on several HTF-mediated functions. TGF-β induces a response in HTF that is concentration-dependent, with different functions being maximally stimulated at different concentrations. This biphasic response highlights the significance of the concentration profile of TGF-β at the wound site. These findings are important in filtration surgery, where constant changes in the local environment occur due to the passage of aqueous and the wound healing process. The varying levels of TGF-β in the aqueous and subconjunctival tissues may thus significantly modify the conjunctival scarring response.

Subconjunctival scarring at the wound site greatly affects the surgical treatment of glaucoma, and the wound healing response is the major determinant of filtration surgery outcome. ^{1 2} Although the introduction of the anti-scarring agents mitomycin-C and 5-fluorouracil has improved the results of filtration surgery, their use is associated with severe and potentially blinding complications. ^{3 4 5} This is probably because these agents work by nonspecific cell damage. Alternative and safer agents are therefore necessary, especially those with more physiological actions. However, identification of such agents requires further elucidation of the mechanisms involved in the cellular processes of wound healing in the eye. 

The growth factor transforming growth factor-β (TGF-β) is a multifunctional growth factor found throughout the body, intrinsically involved in the processes of scarring. ^{6 7 8 9} In the eye, of the three human isoforms, TGF-β2 appears to be predominant ^{10 11} and has been identified as important in the pathogenesis of several ocular scarring diseases (e.g., retinal fibrosis and cataract formation ^{12 13 14} ). 

TGF-β2 is also implicated in glaucomatous disease, with elevated levels of this factor being found in the aqueous of glaucoma eyes, compared with normals. ¹⁵ In addition, the aqueous appears important in the wound healing response after glaucoma filtration surgery, and compared with other growth factors found in aqueous, TGF-β is the most potent in stimulating in vitro conjunctival fibroblast functions. ^{16 17} However, these and other studies ^{12 18 19 20} have only investigated TGF-β1 and -β2; the role of TGF-β3 in aqueous is relatively unknown. 

Moreover, although TGF-β1 and -β2 are known potent stimulators of the scarring response in humans, ^{7 8 9 21 22 23 24} the actions of TGF-β3 in wound healing are unclear, with evidence in some studies that it may behave very differently from the other two human isoforms and inhibit the scarring response in vivo. ^{7 25} The role of TGF-β3, compared with TGF-β1 and -β2, in the processes of conjunctival scarring has still to be delineated. 

TGF-β production in the eye after surgery is locally derived from tissues and inflammatory cells. However, the profile of TGF-β activity in the aqueous is probably considerably altered by filtration surgery. Furthermore, the passage of aqueous through the filtration wound site will result in a constantly changing environment. The role of different concentrations of the different TGF-β isoforms on cells in various extracellular environments thus becomes an important consideration. 

We have used several in vitro assays to, first, compare the actions of TGF-β1, -β2, and -β3 on different components of the conjunctival scarring response (specifically fibroblast activity), and, second, assess the effect of varying TGF-β concentrations on cellular function. All the in vitro cell culture techniques in this study were based on human Tenon’s capsule fibroblasts (HTF). This is because HTF are believed to be the key cells involved in the subconjunctival wound healing response (with a number of essential functions, including: proliferation at the wound site, migration to and contraction of the wound, the synthesis of new extracellular matrix components, and, finally, remodeling of this new matrix to produce a scar). Our studies specifically focused on fibroblast proliferation, migration, and collagen contraction. 

Methods

Cell Culture

Human Tenon’s fibroblasts were propagated from 0.5-cm³ tissue explants as previously described. ²⁶ The tenets of the Declaration of Helsinki were followed, and institutional human experimentation committee approval was granted. All cells used were from the same initial population, between the 3rd and 6th passages. Fibroblasts used in the study were cultured at 37°C in 5% humidified CO₂ in air and fed every 3 days with renewed culture medium (Dulbecco’s modified Eagle’s medium [DMEM] containing 10% [vol/vol] newborn calf serum, 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptokinase, 0.25 μg/ml fungizone, and 50 μg/ml gentamicin; all from Gibco Life Technologies, Paisley, UK). For all assays, cultured HTF from monolayers were trypsinized, pelleted, and counted using a hemocytometer. 

Materials

The exogenous TGF-βs used in these studies were all recombinant human proteins. TGF-β2 and -β3 were generous gifts of Ciba–Geigy (Basel, Switzerland) and TGF-β1 was a gift from Gregory Schultz (University of Florida, Gainesville). Working solutions of TGF-β were prepared in DMEM/1% bovine serum albumin (BSA; Sigma, Dorset, UK). The concentrations of TGF-β used in the study were as follows: 0, 10⁻¹⁵ M, 10⁻¹⁴, 10⁻¹³, 10⁻¹², 10⁻¹¹, 10⁻¹⁰, 10⁻⁹, and 10⁻⁸ M. 

Fibroblast-Mediated Collagen Contraction Assays

One milliliter free-floating collagen type-1 (Sigma) lattices were prepared using the method previously described. ²⁷ Both three-dimensional (3D) and two-dimensional (2D) models were investigated. For 3D lattices, 250 μl aliquots of a collagen/cell suspension mixture (250,000 cells/ml) were pipetted into single 24-well plates and allowed to polymerize. For 2D models, collagen lattices were prepared without any cells, after which HTF (in serum-free DMEM) were placed on the surfaces of the lattices (20,000 cells/48-well plate). Each lattice underwent treatment with 500 μl/well of different concentrations of TGF-β. They were incubated up to 14 days at 37°C in 5% humidified CO₂ in air and medium, and TGF-β was replaced every 3 days. 

Measurements of the collagen lattice area were performed from digitized images (model VQV-100; Casio, Tokyo, Japan), which were analyzed using ImageTool software (UTHSCSA software, University of Health Sciences San Antonio, Texas). The results from the sextuplicate wells were expressed as mean percentage reduction in surface area (compared with original surface area, in mm²). 

Fibroblast Proliferation Assays

The effects of different concentrations of all three TGF-β isoforms on fibroblast proliferation were investigated using several methods: a manual count technique; by demonstrating cell DNA activity using bromodeoxyuridine (BrdU) and Ki67 immunocytochemistry; and, finally, using the colorimetric WST-1 assay. 

Fibroblast monolayers (from 96-well plate; initial density, 2000 cells/well) were treated with different concentrations of TGF-β. On days 0, 3, 7, 14, and 30, monolayers were fixed with 100% methanol, stained with hematoxylin for 2 minutes, washed gently, and air-dried. Cells were manually counted in 10 randomly selected high-power fields (magnification, ×400). The results for each experiment were expressed as a mean percentage compared with control (with 95% confidence intervals [CIs]). 

Fibroblast monolayer and 3D lattices were assessed for BrdU uptake and Ki67 immunostaining. On days 1, 3, and 7 after treatment, HTF were pulsed with 10 mM BrdU (Amersham, Aylesbury, UK) in DMEM/1% BSA for either 4 or 12 hours, after which monolayers were gently washed in phosphate-buffered saline (PBS), fixed in 95% ethanol/5% acetic acid for 30 minutes at room temperature, air-dried, and frozen at− 20°C. Three-dimensional collagen lattices were washed in PBS and fixed in 10% formal saline. Bromodeoxyuridine incorporation was demonstrated by immunocytochemistry using a monoclonal mouse anti-BrdU antibody (Amersham) and immunoprecipitation with diaminobenzamine (DAB). For Ki67 assessment, monolayers were fixed in 70% methanol (Fisons) at −20°C for 30 minutes, and 3D lattices were fixed in 10% formal saline, embedded in paraffin wax, and sectioned before being rehydrated in PBS. Specimens were then microwaved to unmask antigenic sites. Immunocytochemistry was performed using a biotin–streptavidin peroxidase technique with a monoclonal mouse anti-Ki67 antibody (Dako, Sussex, UK) and immunoprecipitation with DAB. 

Fibroblast proliferation in the monolayer and 2D models was also assessed using the WST-1 (Boehringer–Mannheim GmBH, Mannheim, Germany) assay. On days 0, 3, 6, 9, 14, and 30, 10 μl of WST-1 was added to the 200 μl of medium in each 96-well plate. A standard calibration curve was constructed at each of these time points, by assessing the absorbance of known cell densities set up in triplicate, and determining the logarithmic equation (R ² calculated for each experimental run). The results were expressed as a percentage of mean absorbance compared with control (with 95% CI). 

Fibroblast Migration Assays

Cell migration consists of several components, including chemotaxis, ²⁸ chemokinesis, ²⁹ and haptotaxis. ³⁰ For the purposes of the present study, both chemotaxis and chemokinesis were assessed with HTF, using a Transwell assay system with tissue culture inserts (Costar; High Wycombe, UK) for 24-well plates. 

The chemotaxis assay was performed as previously described. ³¹ Cultured HTF in serum-free DMEM were placed in the inner chamber (10,000 cell/well), and serum-free DMEM/1% BSA medium containing TGF-β1, -β2, and -β3 at different concentrations was added to the outer chamber. Cells were left to migrate over a period of 16 hours at 37°C in 5% CO₂ in air. Similarly, a checkerboard analysis of migratory activity was set up as previously described, ^{31 32} with different concentrations of TGF-β in the inner and outer chambers. 

For each well, the total number of cells that had migrated to the undersurface of the membrane was calculated. Sets of triplicate membranes were counted. Results were expressed as the mean number of migrated cells (±95% CI) compared with the negative control with only serum-free medium in both chambers (±95% CI). 

Statistical Analysis

For each assay, quantitative data were statistically analyzed at individual time points using a one-way ANOVA (SPSS for Windows; SPSS Inc., Cary, NC). The observed significance levels from multiple comparisons were adjusted using the Bonferroni test, with P < 0.05 indicating significance. 

Results

Effects of TGF-β on HTF-Mediated Collagen Contraction

Treatment with TGF-β1, -β2, and -β3 was found to similarly stimulate 3D and 2D collagen contractions. Contraction was induced in a biphasic, concentration-dependent manner, with greatest activity at 10⁻⁹ M in both models (Fig. 1) . 

On days 10 and 14, all TGF-β isoforms significantly stimulated contraction compared with control at concentrations of 10⁻¹¹ to 10⁻⁸ M in 3D lattices and 10⁻¹² to 10⁻⁸ M in 2D lattices (P < 0.05). Although no differences were found between the three isoforms in the 3D matrix, TGF-β1 appeared more stimulatory than TGF-β2 or -β3 in the 2D matrix model. 

Comparison of TGF-β concentrations showed significant differences in contraction on days 7, 10, and 14 in 3D models, and on days 3, 7, 10, and 14 in 2D models (P < 0.05). Significant differences were also found between concentrations of 10⁻¹² to 10⁻⁸ M in 3D lattices and 10⁻¹³ to 10⁻⁸ M in 2D lattices, with respect to their stimulation of contraction throughout the course of the study (P < 0.05). 

Effect of TGF-β on HTF Proliferation

TGF-β stimulated HTF proliferation in monolayer and 2D lattice models. All isoforms behaved similarly, inducing proliferation in a biphasic, concentration-dependent manner, with greatest activity at 10⁻¹² M. 

In monolayer assays (Fig. 2) , TGF-β1, -β2, and -β3 displayed similar concentration-dependent activities. All three TGF-β isoforms at concentrations of 10⁻¹³ to 10⁻¹⁰ M stimulated proliferation more than control after day 7; and on day 30, all isoforms at all concentrations significantly stimulated HTF proliferation compared with control. For each concentration of TGF-β, significant differences (P < 0.05) were found between proliferation on days 3, 7, 14, and 30 after treatment. Significant differences were also found between concentrations of 10⁻¹⁴ to 10⁻⁹ M with respect to their stimulation of monolayer HTF proliferation throughout the course of the study (P < 0.05). 

TGF-β induced similar effects in the 2D model. All isoforms showed concentration-dependent activities similar to those seen in the monolayer models (Fig. 3) . Maximal stimulation of HTF proliferation occurred at a concentration of 10⁻¹² M, as for monolayer peak activity. On day 30, all isoforms significantly stimulated HTF proliferation compared with control at concentrations of 10⁻¹³ to 10⁻¹⁰ M. For each concentration of TGF-β, significant differences (P < 0.05) were found between proliferation on days 6, 9, 14, and 30 after treatment. 

HTF proliferation was not detected in 3D lattices, with either BrdU uptake or Ki67 immunohistochemistry techniques. 

Effect of TGF-β on HTF Migration

TGF-β stimulated HTF chemotaxis. All isoforms showed similar concentration-dependent activities in stimulating HTF chemotactic migratory activity. Maximal stimulation of HTF migration occurred at a concentration of 10⁻⁹ M (Fig. 4) . All isoforms and concentrations, apart from TGF-β2 10⁻¹² M, were found to significantly stimulate migratory activity compared with control. 

Checkerboard analysis of TGF-β2 activity suggested that migration of HTF was stimulated mainly by chemotaxis (Table 1) . This effect was apparent from the results because it can be seen when equimolar concentrations are present in the outer and inner chambers; there was no real trend toward an increase in the number of migrated cells with an increase in TGF-β2 concentration. The concentration with maximal activity was found to be 10⁻⁹ M. Chemotactic activity (94.5 migrated cells) at this concentration was found to be greater than chemokinetic activity (35.75 migrated cells). Between groups, analysis with ANOVA, using the concentration of TGF-β2 in the outer chamber as the determinant, showed significant differences between the migratory activity stimulated by all the different TGF-β concentrations (P < 0.05). Using Bonferroni’s modification to compare concentration gradients, it was possible to identify at which concentrations chemotactic activity was significantly greater than chemokinesis (P < 0.05, Table 1 ). 

Discussion

HTF perform an essential role in conjunctival scarring and provide a useful in vitro model of cellular activity during the scarring response. We have shown that human TGF-β1, -β2, and -β3 behave similarly to each other with respect to their actions on HTF. All three isoforms of TGF-β stimulate in vitro fibroblast activity, as demonstrated by their effects on assays of fibroblast-mediated collagen contraction, fibroblast proliferation, and fibroblast migration, and suggesting that TGF-β has a stimulatory effect on conjunctival scarring. This stimulation occurs in a biphasic, concentration-dependent manner, with different peak activities associated with different fibroblast functions. 

TGF-β stimulation of fibroblast-mediated collagen contraction occurred with a biphasic response with peak activity at 10⁻⁹ M. These results were similar both for 2D and 3D lattices. However, previous studies have mainly assessed TGF-β1, and there is no work comparing the effects of the three TGF-β isoforms on wound contraction. In addition, assays have used 3D collagen lattices only, and there is but one study, recently reported, investigating ocular cell-mediated collagen contraction ³³ and showing a dose-related stimulatory response to 4 × 10⁻¹³ to 4 × 10⁻¹⁰ M TGF-β1. Similar results have been obtained with BHK-21 (infant hamster kidney), 3T3-L1 (mouse embryo) fibroblasts, and human foreskin fibroblasts. ^{34 35} In an experiment comparing TGF-β2 to the TGF-β1 isoform in relation to their ability to contract rabbit dermal fibroblast–populated collagen gels, Pena et al. ³⁶ showed equal effects with 2 × 10⁻¹⁰ M of each isoform. The mechanism by which TGF-β may stimulate collagen contraction may via its effect in altering the expression of the integrin family of cell adhesion receptors, specifically theα ₂β₁ receptor ^{37 38} and α_1,2,3,5 andβ ₁ subunits. ³⁹  

Our results demonstrate that all three isoforms of TGF-β produce a similar effect in stimulating HTF proliferation in a concentration-dependent manner, with maximal activity at 10⁻¹² M. This occurs in both monolayer and 2D matrix models, with no proliferation detected in the 3D model. In one of the few studies assessing effects of TGF-β on ocular cell proliferation, Kay et al. ⁴⁰ recently compared the three TGF-β isoforms on HTF. They demonstrated that all isoforms behaved similarly, with stimulation of HTF proliferation in a dose-dependent manner, although they studied a much narrower range (4 × 10⁻¹² to 4 × 10⁻¹⁰ M), and hence did not establish a biphasic response. TGF-β1 and -β2 have been shown to have a similar dose-dependent effect on human foreskin and dermal fibroblasts, ⁴¹ although some authors suggest that TGF-β on its own in serum-free medium has no effect on fibroblast proliferation. ⁴² Compared with a variety of“ stressed,” attached collagen gels, cell proliferation in a“ relaxed,” free-floating 3D collagen matrix has been shown not to significantly occur, ^{43 44 45} although this phenomenon is controversial. ^{42 46} TGF-β–induced proliferation has also been demonstrated in C3H 10T (mouse embryonic fibroblasts), ⁴⁷ NRK-49F (rat kidney fibroblasts), and AKR-2B (mouse embryo-derived) fibroblasts. ⁴⁸  

Fibroblast proliferation induced by TGF-β may occur via its modulation of c-fos, c-myc, and c-sis expression ^{49 50} or its induction of cyclin D and strong downregulation of p27 expression, leading to passage from G1 to S phases of the cell cycle. ⁵¹ Kay et al. ⁴⁰ have suggested that TGF-β may have an indirect mitogenic effect on HTF, via its induction of fibroblast growth factor-2 (FGF-2). Recent studies found that TGF-β effects in NRK fibroblast proliferation are mediated by connective tissue growth factor (CTGF). ⁵² CTGF is known to be a cysteine-rich mitogenic peptide, the secretion and synthesis of which are selectively induced by TGF-β at the level of gene expression. It is believed to mimic many of the actions of TGF-β on mesenchymal cells only, it has no action on epithelial cell lines. ⁵³ However, CTGF alone is not capable of stimulating anchorage-independent growth of fibroblasts but appears to act as a downstream mediator of the growth promoting effects of TGF-β. 

Our results indicate that all TGF-β isoforms show similar concentration-dependent activities in stimulating HTF chemotactic migratory activity and that TGF-β chemotactic activity is greater than chemokinetic activity. Maximal stimulation of HTF migration occurred at a concentration of 10⁻⁹ M, with all isoforms showing a biphasic response, and another peak (albeit less significant than 10⁻⁹ M) of migratory activity being demonstrated at 10⁻¹³ M. A biphasic response to TGF-β has also been demonstrated in neutrophil chemotaxis, where the authors compared isoform activity and suggested potency in the order of TGF-β2 > TGF-β3 > TGF-β1. ⁵⁴ They postulated that TGF-β–induced migration occurred via its stimulation of fibronectin production and explained the biphasic response by the fact that at high concentrations TGF-β causes excessive fibronectin secretion, retarding neutrophil migration. ⁵⁴ TGF-β also strongly stimulates peripheral monocyte chemotaxis, ⁵⁵ attributed to high affinity receptors for TGF-β found on their cell surface, making them able to respond to fentomolar concentrations. The effects of TGF-β on ocular cell migration have been assessed in a number of studies on the cornea, where it appears to be primarily inhibitory, ^{56 57} and is more potent than either epidermal growth factor or FGF. ⁵⁸ Other ocular cell migratory activities that have been investigated include trabecular meshwork cells, where both of TGF-β1 and -β2 were assessed, with maximal activity found to be at concentrations of 4 × 10⁻¹⁵ to 4 × 10⁻¹³ M. ⁵⁹ In this cell type, however, platelet-derived growth factor was shown to be a more powerful chemoattractant. Like our findings, these authors demonstrated that TGF-β migratory activity was predominantly chemotactic rather than chemokinetic. Unfortunately, higher concentrations of TGF-β were not fully investigated in these ocular cell migration studies, with most experiments using a range of TGF-β concentrations with a maximum of less than 4 × 10⁻¹³ M. 

One of the most important sites of TGF-β activity in the eye is the aqueous humor, identified as a factor influencing the wound healing response in filtration surgery. ¹⁶ This has been supported by the demonstration that compared with other growth factors found in the aqueous, TGF-β has been shown to be the most potent in stimulating HTF activity. ⁶⁰ TGF-β2 is the predominantly expressed isoform in the aqueous. ^{12 16 18} However, it is important to note that in all these studies only TGF-β1 and -β2 isoforms (and not TGF-β3) were analyzed. The average concentration of active TGF-β2 in normal aqueous is between 0.73 and 10.98 × 10⁻¹¹ M, compared with serum where levels of active TGF-β in normals lie between 12 × 10⁻¹¹ and 16 × 10⁻¹¹ M. 

An important finding by Tripathi et al. ¹⁵ is the demonstration that aqueous levels of TGF-β2 are significantly raised in glaucomatous (primary open-angle glaucoma) eyes compared with age-matched controls (1.8 versus 0.71 × 10⁻¹¹ M active TGF-β2), in a sample of patients undergoing routine cataract surgery. The authors hypothesized that this difference may account for the excessive extracellular matrix deposition seen in the trabecular meshwork of glaucomatous eyes, leading to decreased aqueous outflow and raised intraocular pressure. Production of TGF-β in the aqueous is believed to be derived from local tissues. Again, studies have identified TGF-β2 as the predominant isoform produced by the iris and ciliary body ⁶¹ and trabecular cells. ⁶² In addition, it has been shown that when secreted, it is mainly in its latent form. 

Although the aqueous is a major factor in glaucoma filtration surgery, production of TGF-β will also be locally derived from the wound site. TGF-β isoforms can be produced by a variety of cells: TGF-β1, released predominantly from degranulating platelets; TGF-β2, locally produced in the aqueous, as discussed above; and TGF-β3 from inflammatory cells. At least two factors would be expected to alter the amount of TGF-β secreted and its degree of activation at the filtration wound site. First, surgery will invariably initiate the blood clotting cascade and, second, cause a breakdown in the blood–aqueous barrier. Plasmin and thrombospondin ^{63 64} produced by these events can then activate all three TGF-β isoforms. Hence, the profile of TGF-β activity at the wound site may be considerably altered by filtration surgery, complicated by the fact that the passage of aqueous through the filtration wound will result in a constantly changing environment. 

The implications of different peak activities of TGF-β–induced fibroblast functions may be explained physiologically. In a wound environment, the two early functions of fibroblasts are migration and proliferation. TGF-β is initially released by inflammatory cells and platelets at the wound site. At relatively low concentrations, it can act as a stimulant for fibroblast proliferation and as a weak chemoattractant (range, 10⁻¹³ to 10⁻¹² M). At this stage, a provisional matrix is deposited, which attenuates fibroblast proliferation. The concentration of TGF-β in the wound would then probably be much higher due to the stimulated increase in fibroblast number. Previous authors have suggested that at concentrations of >10⁻⁹ M, TGF-β is a potent stimulant of collagen production. ¹⁷ Thus, at around 10⁻⁹ M, TGF-β activity is adapted to collagen matrix deposition, with stimulated functions of fibroblast-mediated contraction, secondary to stimulated fibroblast migration and matrix remodeling. Hence, during normal evolution of the scarring process, the key functions of the fibroblast depend on its environment. In particular, the effects of growth factors determine fibroblast activity at any one time. Our in vitro results thus suggest that the biphasic effects of TGF-β determine the response of fibroblasts and that the role of TGF-β during HTF-mediated conjunctival wound healing is very much dependent on its concentration at the wound site. 

In summary, we have demonstrated that TGF-β is a potent stimulant of HTF fibroblast activity, suggesting its stimulatory role in the conjunctival scarring response. Because all three isoforms are probably present during the wound healing response after glaucoma filtration surgery, an important finding has been the fact that they behave in a similar manner. Finally, TGF-β actions are characterized by different concentration-dependent effects and different peak activities for stimulating various fibroblast functions, and its biphasic characteristics have implications for the timing and development of the conjunctival scarring response. Hence, TGF-β appears to be an important component of conjunctival scarring. Its potent effects make it a possible target agent for modulating the scarring response after glaucoma filtration surgery. 

 Supported in part by Wellcome Trust Vision Research Fellowship, London, UK and Guide Dogs for the Blind Association, London, UK.

 Submitted for publication April 20, 1999; revised August 6, 1999; accepted September 13, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Graphs show the effects of different TGF-β isoform concentrations on HTF-mediated collagen contraction in the 3D matrix model (A) and on the 2D matrix model (B) at 10 days after treatment. Results are expressed as mean % reduction in original collagen matrix surface area with 95% CI. TGF-β stimulated HTF-mediated collagen contraction in 2D and 3D models, with biphasic effects. All isoforms showed similar concentration-dependent activities, with significant differences in activity compared with control at concentrations of 10⁻¹¹ to 10⁻⁸ M. Peak activity was seen at a concentration of 10⁻⁹ M, when maximal stimulation of HTF-mediated contraction in both models occurred. Multiple comparisons with respect to different concentrations and at each time point were performed using ANOVA and Bonferroni’s modification. Although no differences were found between the three isoforms in the 3D matrix, TGF-β1 appeared more stimulatory than TGF-β2 or -β3 in the 2D matrix model. *All isoforms significantly different (P < 0.05) compared with control. Error bars = 95% CI.

Figure 2.

 

HTF proliferation in monolayers was investigated using a manual count technique. Fibroblast proliferation in monolayers are shown as mean cell number as a percentage of control (with 95% CI) with respect to stimulation by TGF-β at different concentrations at 30 days after treatment. All isoforms showed similar concentration-dependent, biphasic activities, with significant differences in activity compared with control at all concentrations on day 30. All three TGF-β isoforms at concentrations of 10⁻¹² to 10⁻¹⁰ M consistently stimulate proliferation more than control at the time points studied. Peak activity was seen at a concentration of 10⁻¹² M, when maximal stimulation of HTF proliferation occurred. Multiple comparisons with respect to different concentrations and at each time point were performed using ANOVA and Bonferroni’s modification. *All isoforms significantly different (P < 0.05) compared with control. Error bars = 95% CI.

Figure 3.

 

The WST-1 technique was used to assess HTF proliferation in 2D matrix model, as illustrated at 30 days after treatment with TGF-β1, TGF-β2, and TGF-β3. Comparisons using ANOVA with Bonferroni’s modification highlighted significant differences between TGF-β and control, as illustrated. All isoforms showed biphasic, concentration-dependent activities similar to those seen in the monolayer models. Maximal stimulation of HTF proliferation occurred at a concentration of 10⁻¹² M, as for monolayer peak activity. On day 30, all isoforms significantly stimulated HTF proliferation compared with control at concentrations of 10⁻¹³ to 10⁻¹⁰ M. For each concentration of TGF-β, significant differences (P < 0.05) were found between proliferation on days 6, 9, 14, and 30 after treatment.* Activity significantly different compared with control (P < 0.05). Error bars = 95% CI.

Figure 4.

 

The mean numbers of migrated cells compared with control, after 16-hour exposure to chemoattractant (TGF-β in outer chamber of Costar Transwells), are shown with 95% CI. All isoforms and concentrations apart from TGF-β2 10⁻¹² M were found to significantly stimulate migratory activity compared with control. *Activity significantly different compared with control (P < 0.05). Error bars = 95% CI.

Table 1.

 

Checkerboard Analysis of TGF-β2 Effect on HTF Migration Showing Chemotactic and Chemokinetic Activity

	TGF-β2 Concentration (log M)	Outer Chamber (No Cells)
				0	−14	−13	−12	−11	−10	−9
Inner chamber (with cells)	0, †	1.67	7.25	13.25	7	8.5	11.5	94.5*
		−14	2.67	23.25	46.25*	15.25*	19.5	63*	88.5*
		−13	2.16	6	14.5	25*	5.75	7.75	14.5
		−12	5	17.75	21.75	5.75	6	11	25
		−11	2	13.75	19.25	10	15	20.25	28
		−10	8.3	25.75	23.75	19	13.5	15	44
		−9	5.3	6.25	19.5	12.25	7.5	19.25	35.75

 Values shown are the mean number of cells migrated from the inner chamber to the undersurface of the porous membrane.

*  Concentration gradient between chambers where TGF-β2–stimulated chemotaxis is significantly greater (P < 0.05) than chemokinesis (diagonal column, bold italic).

†  Chemotactic activity only (first row).