A microfluidic model of bacterial keratitis to quantify the efficacy of anti-infective compounds under flow

AIMS: Bacterial infections of the cornea, also known as bacterial keratitis, are a major cause of blindness. The corneal surface is protected against infection by the flow of tear fluid, which displaces potential pathogens and debris. Whilst fluid flow is known to influence the behaviour of both bacterial and epithelial cells, most models of corneal infection use static conditions. Here, we develop a microfluidic model to directly visualise interactions between bacteria and human corneal epithelial cells (HCE2) in physiologically relevant flow conditions. METHODS AND RESULTS: We use high-resolution microscopy and automated image analysis to continuously quantify the rate at which Staphylococcus aureus (SH1000) and Pseudomonas aeruginosa (PAO1) adhere to HCE2 monolayers. We then use this system to investigate the effects of pre-treatment with unfractionated heparin (UFH), a putative therapeutic compound thought to inhibit bacterial adhesion. Using a traditional, well plate-based assay that lacks fluid flow, prior treatment of HCE2 monolayers with UFH significantly reduced adhesion of both species. In our microfluidic model we found that pre-treating HCE2 monolayers with UFH again sharply reduced S. aureus adhesion, but surprisingly, UFH pre-treatment had no effect on P. aeruginosa adhesion to corneal cells in the presence of fluid flow. CONCLUSIONS: Our findings demonstrate that incorporating fluid flow into in vitro infection models can lead to qualitatively different outcomes when screening putative anti-adhesion therapies. These findings, obtained using two widely studied model strains, illustrate how microfluidic models of infection can offer valuable insights into how flow influences bacterial keratitis and guide the development of new treatments.