An in vitro cell irradiation protocol for testing photopharmaceuticals and the effect of blue, green, and red light on human cancer cell lines

A LED-based cell irradiation system was built that can irradiate a 96-well plate with monochromatic light at controlled temperature and with a built-in dark control. This system was used to study the response of six human cancer cell lines to blue, green, and red light.


II. LED irradiation setup
: Diagram of the printed circuit board (PCB) used for each LED array. Figure S2. Schematic of the electronics of the 455, 520 and 630 nm LED arrays.  S5 III. LED power density setup and analysis, Integrating sphere

S3
The light power density of the LED arrays was measured either using a lab-built setup using physical sensing, or chemical actinometry. The physical sensing setup consisted of an integrating sphere, which was positioned underneath the 96-wells plate to simulate the irradiation conditions during cell experiments ( Figure S4A). The integrating sphere was mounted in a custom-made holder that aligned the 6 mm aperture of the integrating sphere with the 6 mm diameter of a single well in a 96-wells plate, while the LED array was placed on top of the well plate (with its lid), like during cell irradiation. The integrating sphere position was adjustable over ten wells thus providing a diagonal representative set of wells that could be measured individually. The integrating sphere was connected via an optical fiber to an Avantes CCD spectrometer. The spectrometer, in combination with the sphere and fiber, were spectrally calibrated directly before measurement, using an Avantes calibration lamp, to report the spectrum in absolute irradiance units (µW·cm −2 ·nm −1 ), where the surface here refers to that of the aperture of the integrating sphere. The power supply was set to the appropriate voltage, 29.9, 27.9, or 20.7 V for the 455, 520, or 630 nm LED arrays, respectively. The integrating sphere was then placed in the setup and centered under a specific well, for example D4 ( Figure S4B). The LED system was then placed on top of the setup and an average of several scans was measured. These steps were repeated for all ten wells (B2, C3, C7, D4, D8, E5, E9, F6, F10, and G11). An average from the ten data points was taken to determine the values for the entire array. The emission maximum and full width at half maximum (FWHM) bandwidth for each LED array wavelength was determined from the emission spectra (Table 1). Using Excel and OriginPro, the total power density at the bottom of each well (mW·cm −2 ) was calculated by integrating the area under the spectral irradiance vs.
wavelength curve, multiplying by the surface area of the integrating sphere aperture (diameter = 6 mm, surface 0.28 cm 2 ), and dividing the resulting power (mW) by the surface area of a well (0.32 cm 2 ). Average power densities and standard deviations were 10.5 ± 0.7, 20.9 ± 1.6, and 34.4 ± 1.7 mW·cm −2 for the 455, 520, and 630 nm LED arrays, respectively (Table 1). Using an integrating sphere provides a straightforward physical method to test the LEDs, but assumes that the single LED measured per 8-LED column is not significantly different from the remaining seven. Therefore, chemical actinometry (in all 96-wells) was used as a comparison for the blue and green LED arrays. Note: it was impossible to find a chemical actinometer for red light compatible with aqueous solutions, so this comparison was not done for 630 nm. = 9538 + 0.05 Equation S1 The plot of the number of mol of Fe 2+ produced vs. irradiation time (s) for the 455, 520, and dark setups are shown in Figure S5B.  approximately 75-80% of the absorbance measured for the central wells with increased standard deviation (> 20%, see Figure S6). These border wells excepted, the central 60 wells show uniform plate irradiation within 10% standard deviation. Dark controls were S8 used to verify that cells maintained in the dark were receiving negligible photon flux Figure   S5B). The power densities using chemical actinometry for the LED arrays were 10.2 ± 0.9 mW·cm −2 (455 nm) and 16.6 ± 1.3 mW·cm −2 (520 nm). The 455 nm power density values obtained by actinometry and integrating sphere correlate very well, which validates the physical sensing method using the integrating sphere. By contrast the 520 nm measurement slightly deviates (Table 1), which is attributed to the very low absorbance of the ferrioxalate actinometer at 520 nm. 520 nm actually represents the ultimate wavelength at which the ferrioxalate actinometer can be used. As stated above, an additional limitation of chemical actinometry, is the incompatibility of chemical actinometers for red light with plastic plates, therefore no comparison could be done between the two methods at 630 nm. In the following, it is the power density values measured using the integrating sphere that was used for all dose calculations during blue, green, or red light irradiation experiments. Figure S6. Example of actinometry absorbance results for 15 s irradiation at 520 nm. The average absorbance at 510 nm of the central 60 wells was 0.32 ± 0.03 AU (gradient colors with black writing yellow = lowest value to orange = highest value). The average of the border values was 0.25 ± 0.04 AU, where values below average are shown in red text and above average in green text.

V. LED array thermal control
The temperature in well D6 was measured to determine the thermal effects of the Ditabis thermoblock, either in combination with the fans cooling for the LED array, or without the fans for the dark control. In each well of two 96-wells plates, 200 µL phosphate buffered saline (PBS) was added. A thermocouple was used to measure the temperature over time in both "dark" and "irradiated" systems. The temperature was measured every 5 minutes up to 45 minutes and each measurement series was repeated three times.
To maintain a temperature range between 35-37 °C, the target temperature of the Ditabis thermostat needed to be set to 39 °C and PBS was pre-warmed in a water bath S9 (37 °C). Under such conditions, the temperature in well D6 was stably maintained between 35.5 and 37.5 °C for the blue, green, red, and dark systems.

VI. Effective concentrations (EC 50 ) values for rose bengal and methylene blue
Two standard PDT dyes were chosen for evaluating the irradiating setup on a known light-  Figure S7). 5 Methylene blue is a blue dye that absorbs in the red and has been used in a variety of PDT and aPDT type I and II applications ( Figure   S7). 6 Effective concentrations (EC 50 values in µM) and the 95% confidence intervals for rose bengal and methylene blue using the standardized conditions of this manuscript are reported in Figure S8.    Figure S8A shows the dose-response curves for dark (black circles) as well as blue (blue circles, 455 nm, 6 J.cm −2 ), green (green circles, 520 nm, 6 J.cm −2 ), and red (red circles, 630 nm, 6 J.cm −2 ) irradiated responses. Methylene blue samples were dark treated (black triangles) and only irradiated using red light (red triangles, 630 nm, 6 J.cm −2 ). The irradiated/light control cytotoxicity values for A and B were 100%. The SRB absorbance of ten technical replications was averaged for one experiment (n = 3). The absorbance values were exponentially fitted to determine the exponential growth curves using GraphPad Prism, non-linear fit of exponential growth. The doubling times were determined with 95% confidence intervals. S15