Human corneal fibrosis: an in vitro model

Abstract

Purpose: Corneal injury may ultimately lead to a scar by way of corneal fibrosis, which is characterized by the presence of myofibroblasts and improper deposition of extracellular matrix (ECM) components. TGF-beta1 is known to stimulate overproduction and deposition of ECM components. Previously, an in vitro three-dimensional (3-D) model of a corneal stroma was developed by using primary human corneal fibroblasts (HCFs) stimulated with stable vitamin C (VitC). This model mimics corneal development. The authors postulate that with the addition of TGF-beta1, a 3-D corneal scar model can be generated.

Methods: HCFs were grown in four media conditions for 4 or 8 weeks: VitC only; VitC+TGF-beta1 for the entire time; VitC+TGF-beta1 for 1 week, then VitC only for 3 or 7 weeks; and VitC for 4 weeks, then VitC+TGF-beta1 for 4 weeks. Cultures were analyzed with TEM and indirect immunofluorescence.

Results: Compared with the control, addition of TGF-beta1 increased construct thickness significantly, with maximum increase in constructs with TGF-beta1 present for the entire time-2.1- to 3.2-fold at 4 and 8 weeks, respectively. In all TGF-beta-treated cultures, cells became long and flat, numerous filamentous cells were seen, collagen levels increased, and long collagen fibrils were visible. Smooth muscle actin, cellular fibronectin, and type III collagen expression all appeared to increase. Cultures between weeks 4 and 8 showed minimal differences.

Conclusions: Human corneal fibroblasts stimulated by VitC and TGF-beta1 appear to generate a model that resembles processes observed in human corneal fibrosis. This model should be useful in examining matrix deposition and assembly in a wound-healing situation.

Figure 1.

(A) Construct thickness as measured with the confocal microscope software. T1, T1–1w, and T1–4w are statistically significant compared with C at both 4 and 8 weeks (P < 0.05). Week 8 is statistically significant compared with week 4 in each condition (P < 0.05). Note that the longer the exposure to TGF-β1, the thicker the construct. (B) Optical thick sections of the constructs at 4 and 8 weeks confirm the trend in thickness that was observed with the confocal wholemounts (A). Please note that in all constructs, the cells stratified, produced their own matrix, were flat and elongated, and aligned and parallel with the porous membrane. Also, the cell density at the top and bottom of the construct appeared to be higher than in the middle. In T1 week 8, there appeared to be a change in matrix integrity in the area below the arrow. Bar, 50 μm.

Figure 2.

High-magnification TEM (31,000×) showing cell–matrix interaction and matrix condition. The fibril orientation changed direction more than once in all the conditions at both 4 and 8 weeks; however, by 8 weeks, T1 fibril integrity appeared to be decreasing. Of interest, at both the 4- and 8-week time points, the fibrils in T1–1w and T1–4w appeared to be longer than those in C. Bar, 0.5 μm.

Figure 3.

Fibril diameter as measured in approximately 10 high-magnification (31,000x) TEM images. Ten fibrils were measured per photo per group and time point. At least 100 fibrils were measured per group. At 4 weeks, the fibril diameter of C, T1, and T1–1w all appeared to be comparable to that in vivo (30–33 μm). Compared with C, T1 is not statistically significant at 4 weeks, whereas, T1–1w versus C is significant (P < 0.05). At 8 weeks, the differences observed in C versus T1, T1–1w, and T1–4w are statistically significant (P < 0.05). Week 8 is statistically significant compared with week 4 in each condition (P < 0.05).

Figure 4.

Confocal images of full-thickness constructs at 4 weeks. Images show the maximum projection of all the planes of focus for each sample. C or control (No TGF-β1) was compared with T1 (TGF-β1 present for the entire culture time). Fibrosis markers—(A, B) SMA, (C, D) type III collagen or Col III, and (E, F) EDA-Fn—were observed by indirect immunofluorescence. The expression of SMA, type III collagen, and EDA-Fn all increased when TGF-β1 was introduced into the system (T1). Bar, 50 μm.

Figure 5.

Indirect immunofluorescence of fibrotic markers—SMA (A, B), type III collagen (C, D), and EDA-Fn (E, F)—on full-thickness constructs at 8 weeks that were treated with TGF-β1 for various times: T1–1w: TGF-β1 treatment for the first week, then construct medium for the remaining 7 weeks; and T1–4w: construct medium for the first 4 weeks, then TGF-β1 treatment for the remaining 4 weeks. Single plane-of-focus confocal images were selected from the brightest area of immunofluorescence. Four- and 8-week T1–1w constructs appeared to be similar; therefore, since T1–4w was collected only at 8 weeks, we compared T1–1w at 8 weeks. With the addition of TGF-β1 for only 1 week (T1–1w), there appeared to be a slight increase in the fibrotic markers (A, C, E). As the time of TGF-β1 exposure increased (T1–4w), so did the amount of myofibroblasts (B), type III collagen (D), and EDA-Fn (F). Of interest, the point at which TGF-β1 was added to the system (first week or last four) did not seem to be important; however, the length of time TGF-β1 was added to the system appeared to affect the expression of the fibrotic markers. A 1-week treatment appeared to increase the markers, but not to the same extent as the 4-week treatment. Bar, 50 μm.

Figure 6.

Indirect immunofluorescence of proliferation marker, Ki67 (green) in full-thickness constructs without (C) and with (T1) TGF-β1 for 4 weeks. Images show the maximum projection of all the planes-of-focus for each sample. Blue: iodide, a marker of all cell nuclei. Bar, 50 μm.