Human in vitro Model Reveals the Effects of Collagen Cross-linking on Keratoconus Pathogenesis

Abstract

Keratoconus (KC) is a corneal thinning disorder that leads to severe vision impairment As opposed to corneal transplantation; corneal collagen crosslinking (CXL) is a relatively non-invasive procedure that leads to an increase in corneal stiffness. In order to evaluate the effect of CXL on human corneal stromal cells in vitro, we developed a 3-D in vitro CXL model, using primary Human corneal fibroblasts (HCFs) from healthy patients and Human Keratoconus fibroblasts (HKCs) from KC patients. Cells were plated on transwell polycarbonate membranes and stimulated by a stable vitamin C. CXL was performed using a mixed riboflavin 0.1% PBS solution followed by UVA irradiation. Our data revealed no significant apoptosis in either HCFs or HKCs following CXL. However, corneal fibrosis markers, Collagen III and α-smooth muscle actin, were significantly downregulated in CXL HKCs. Furthermore, a significant downregulation was seen in SMAD3, SMAD7, and phosphorylated SMADs -2 and -3 expression in CXL HKCs, contrary to a significant upregulation in both SMAD2 and Lysyl oxidase expression, compared to HCFs. Our novel 3-D in vitro model can be utilized to determine the cellular and molecular effects on the human corneal stroma post CXL, and promises to establish optimized treatment modalities in patients with KC.

Figure 1

UVX-1000 illumination system/3-D in vitro model.

Figure 2

Live/Dead assay: shows the effect of CXL on corneal cell viability, HCF, and HKC were exposed to UVA irradiation + riboflavin for three different time periods (3 min, 5 min, and 10 minutes).

Figure 3

MTT assay quantification: HCF, HKC controls, and HCF and HKC CXL. Data was normalized to HCF controls and a fold regulation is plotted. One way ANOVA for a total n = 4 data sets. *(P = 0.0015), **(P = 0.0086), ***(P = 0.0008).

Figure 4

Cell hydration profile post CXL. Figure (A) represents that water uptake percentage quantified over a period of 40 minutes. Figure (B) shows the swelling ratio in both cell types after CXL treatment.

Figure 5

Cell migration: HCFs, and HKCs were scratched and the relative cell migration distance was quantified at 0 hr, 4 hr, 24 hr and 48 hr time points.

Figure 6

Protein expression for α-SMA in HCF, HKC controls, and HCF, HKCs treated with CXL Quantification of protein bands that are normalized to the loading control. n = 4, and error bars represent standard error of the mean. One way ANOVA was performed. *(P = 0.0466), **(P = 0.0030), ****(P = 0.0003).

Figure 7

Collagen protein quantification post CXL: (A) Collagen I, **(P = 0.0078), (B) Collagen III, *(P = 0.0356), **(P = 0.0076), (C) Collagen V, *(P = 0.0142), **(P = 0.0097).

Figure 8

Quantification of SMAD6 and SMAD7 expression in HCFs and HKCs post CXL. Western blot analysis shows protein expression for (A) SMAD6, significant downregulation in HCF and HKC controls ***(P = 0.0003), similar pattern in HCF and HKC treated by CXL***(P = 0.0007). (B) SMAD7, significant downregulation shown in HKC control cells compared to HCF controls**(P = 0.0031).

Figure 9

Quantification of protein expression for (A) SMAD2, (B) SMAD3, (C) pSMAD2, and (D) pSMAD3 following CXL treatment, n = 4. All samples were repeated at least three times. *p < 0.05 was considered to be statistically significant **p < 0.01.

Figure 10

Protein expression for LOX in HCF, HKC controls, and HCF, HKCs treated with CXL. Quantification of protein bands are normalized to the loading control. N = 4, and error bars represent standard error of the mean. One way ANOVA was performed. *(P = 0.0120), **(P = 0.0015).