We developed an image processing protocol to perform time lapse fluorescent analysis in microfluidic biochips. Using three probes to follow three parameters: the number of cell nucleus, the mitochondrial activity and the cell death, we have investigated the behaviour of HepG2/C3a under microfluidic culture conditions over a period of 10 h. The results have shown that cell response was independent of their location in the biochip, and cells aggregated after 4 h of perfusion whereas they spread in the Petri dish control. The mitochondrial activity increased with time in biochips when compared to Petri dishes. Nevertheless, the cell population presented mono-modal mitochondrial activity illustrating no specific cell subpopulation. The variance analysis of the parameters (via the interquartile calculation) revealed an increase in cell heterogeneity in the first 6 h to 7 h of cultures. Then the variance reached a plateau illustrating a stabilisation towards a homogeneous cell population in the biochips. Despite the fact that we also found a mono-modal cell population in Petri dishes, the variance depicted a continuous increase of the population heterogeneity. To highlight the potential of the image processing method for drug screening, we then performed a staurosporine toxicity assay. 5 μM treatments after 10 h of exposure led to 65% of cell death in the biochips whereas only 11% was detected in control. Altogether, these results illustrate the potential of our image processing technique combined with a microfluidic approach for wider real time cytotoxicity assays using on chip methods.
