Noble-metal nano-hole arrays that generate EOT phenomenon are a promising label-free surface plasmon resonance (SPR) biochemical sensor [1]. Cellular response detection by other SPR sensors (Kretschmann prism and photonic crystal) has been successful [2]. However, limited by the high cost of earlier nano-hole fabrication techniques such as FIB/EBL, the EOT-based biochemical sensors had only ~10x10 μm scale of nano-holes and thus were applied mostly in detecting molecules [1], which need only a small area of sensing elements. Recently, a template-stripping technique to fabricate large area (~1x1 cm) of nano-holes was reported [3], which enables EOT-based measurement of cellular events. Prior studies [4] showed that the cell adhesion on the nano-holes can result in a change in effective refractive index (RI), which, in turn, leads to a spectral shift in the transmission peak. Therefore, we hypothesize that the EOT-based sensors are capable of label-free monitoring the cellular effects of toxic chemicals that change cell adhesion conditions.

The integrated microfluidic device and measuring system shown in Fig. 1 was used to monitor the cellular response quantified by the EOT spectral shift in association with chemical stimuli. Gold film perforated with nano-hole arrays on a glass substrate was fabricated by depositing 100-nm-thick gold on a nano-hole-patterned silicon template and template-stripped onto the glass slide with UV curable epoxy [3]. PDMS microfluidic channels with an inlet and an outlet were fabricated by soft lithography and align-bonded with a gold-film-coated glass slide. After washing and sterilization of the device, cells were cultured on the cystimine-modified gold surface till confluent. Then the device was placed under a microscope connected to a spectrometer to record the spectral response (Fig. 2) to chemical stimuli applied via the microfluidic channels. Using the spectrum at t=0 as baseline, the spectral shift at different times was acquired as a function of the chemical concentration. In addition to chemical stimuli reported here, the same system can be used to characterize the cell response to electrical stimuli in the future.

We applied three types of chemical stimuli to HEK293 cells in experiments. The first was trypsin that breaks the collagen links between cells and the substrate. Fig. 3 shows the spectral shift for 0.05% trypsin in PBS. The spectral shift has a logarithmic time dependence, which fits well with the first order reaction mode of enzyme. The spectrum was blue-shifted because the effective RI near gold surface decreased as the cell were gradually detached from the substrate. The second stimulus was lipopolysaccharide (LPS), which is a cell toxic chemical known to flatten and expand cells. Fig. 4 plots the spectral shift for different concentrations of LPS, showing a red-shifted spectrum with increased time and concentration, which was attributed to a greater RI near the sensor surface due to increased cell adhesion on the gold nano-holes. The third stimulus was sodium azide, which, in contrast to LPS, makes cells to transform and shrink. As expected, the spectrum in Fig. 5 was blue-shifted, since the effective RI dropped in response to weaker cell adhesion. These results demonstrate the capability of our EOT-based microfluidic biosensor for label-free monitoring of cellular responses to external stimuli.