Abstract
Purpose :
Fungal keratitis is a dangerous infection that can lead to loss of vision and the eye. To facilitate research into the pathogenesis of fungal infections, we developed a tissue-engineered 3D cornea model for infection with different fungal strains.
Methods :
The hemi-cornea model consists of epithelium and stroma, embedded in a snapwell insert (0.4 µm pore size). Immortalized human corneal epithelial cells (hTcEpi, Evercyte) were seeded on a stromal equivalent of 45,000 primary human keratocytes in collagen I matrix, compressed to a thickness of ~500 µm. Cultivation at the air-liquid interface for 8 days formed a multi-layered, nonkeratinized squamous epithelium.
Each model was infected with 5x104 conidia and incubated for up to 48 hours at 30°C. For infection, we used specimens isolated from human fungal keratitis swabs or scrapings, obtained from the German National Reference Center for Invasive Fungal Infections. Candida albicans served as a reference strain.
Proliferation and invasion of the fungus were visualized by H&E and immunofluorescent staining. Cytotoxicity was quantified by LDH assay. IL-8 in the supernatant was measured by ELISA.
Results :
Three Fusarium strains as emerging causes of fungal keratitis were chosen for infection of the model.
Regarding cell adhesion and tissue invasion, F. keratoplasticum showed the fastest tissue invasion and onset of proliferation, 8 hours earlier than F. petroliphilum and F. falciforme. 24 hours after infection, F. keratoplasticum exhibited significantly higher cytotoxicity (50.2% ±1.8) compared to F. petroliphilum (37.3% ±1.6) or F. falciforme (25.2% ±3.7). Compared to C. albicans, Fusarium strains was 0.6-fold lower in cytotoxicity. Secretion of pro-inflammatory cytokine IL-8 by the host cells was reduced in Fusarium-infected models compared to C. albicans (0.3-fold on day 2, 0.5-fold on day 3 post-infection).
A Candida-infected corneal button excised during keratoplastic surgery after 2 weeks of unsuccessful anti-mycotic therapy was compared to the model in a time course. Histological similarity was highest at 48 hours post-infection of the model.
Conclusions :
Fungal keratitis can be recreated in vitro, revealing detailed information about the time course, extent of tissue invasion, and host response of different fungal strains. Scalability and reproducibility of the tissue-engineered model may allow testing of anti-fungal therapies.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.