High-Speed Full-Well Imaging for Assessing Cell Positivity

This application note highlights a high-content imaging and analysis platform that delivers rapid cellular positivity measurements, completing plate loading to dose-response curve analysis in under 6 minutes for ~1,000,000 cells. Using a live/dead assay, human epithelial cells were treated with an apoptotic compound and stained with Image-iT™ DEAD GREEN™ and Hoechst 33342. A 96-well plate was imaged with Araceli Endeavor® in 4min 40sec (0.27µm/pixel) or 3min 30sec (4×4 bin), followed by 2min 15sec analysis using Araceli Clairvoyance™, which segmented nuclei and scored live or dead cells based on green intensity. Binning reduced storage needs, improved signal-to-noise ratios, and increased speed. The assay’s exceptional green signal-to-noise ratio (200:1) enabled clear differentiation between live and dead cells, with dose-response curves and EC50 values aligning with kit standards. This streamlined workflow extends to other positivity-calling applications, making Endeavor and Clairvoyance invaluable for monitoring gene activation or cellular responses.

Measuring cell positivity is essential in high content imaging (HCI). With broad applications from measuring gene induction to viral infectivity and clonal analysis, assays scoring cells as positive or negative based on fluorescent intensity levels are used in virtually every high content screening lab. This Application Note looks at live/dead staining, a common cell positivity assay that quantifies dead cells based on a dye that selectively infiltrates permeable cells but not healthy cells. Membrane permeability provides straightforward assessment of cytotoxicity and apoptosis, with compromised cells exhibiting a characteristically permeant membrane (De Schuter, et al. 2021). Within HCI, assessing cell death and cytotoxicity can both ensure that potential screening hits with a toxicity liability are not carried forward or identify positive hits that cause cell death, in the case of antibiotic and chemotherapeutic screens.

Human epithelial cells were treated with a cell membrane disrupter followed by staining with a live/dead assay, imaging by Araceli Endeavor® high content imaging system and image analysis using Araceli Biosciences’ internal HCI analysis software. This live/dead assay uses Image-iT™ DEAD GREEN™ (Invitrogen, H10290), which stains the nucleus green when the cell membrane is permeable, generally indicative of apoptosis (De Schuter, et al. 2021), while live cells remain unstained (Figure 1). The total number of cells in the population were measured using nuclear dye Hoechst 33342, which stains DNA regardless of cell state. The results from this live/dead cell assay were near-identical to that reported in the assay kit documentation.

High content imaging with submicron resolution and full well coverage is usually prohibitively slow, but Endeavor’s ability to image 96-, 384-, or 1536-well microplates at 0.27 µm/pixel and imaging a majority of the well at speeds comparable to other high content systems’ low-resolution speed minimizes this issue. This 2-color assay takes a mere 4 minutes 40 seconds for a full scan. However, mammalian nuclei are almost always >10 µm2 (Lammerding et al 2011), so this submicron-level resolution is not necessary for a simple cell positivity assay. Some situations still warrant lower resolution images, such as assays and analysis pipelines optimized for 10x magnification or a desire to minimize file size. As Endeavor has fixed objectives, a bona fide 10x image isn’t possible, but down-sampling through pixel binning produces an equivalent effect. Here, the CMOS detector digitally bins the data, summing the specified number of pixels together to create a lower resolution image while also exponentially increasing fluorescent signal. The advantages of binning pixels for this live/dead cell assay are reported in this Application Note and compared with the non-binning results.

To perform this assay, human epithelial cells were treated with valinomycin, a potassium ionophore known for inducing apoptosis (Abdalah et al 2006). After 30 minutes of treatment, cells were stained with Image-iT DEAD GREEN and Hoechst 33342 in a 96 well high content microplate and imaged in 4 minutes 40 seconds/ microplate with full well coverage at full resolution on Araceli Endeavor (Figure 1). Current image times are expected to be significantly faster with current Endeavor 2.1 updates. Nuclei were then segmented in both green (permeable/dead nuclei) and blue (all nuclei), with the number of green cells ratioed to the total cell number in blue (Figure 2a). Additionally, intensity in green was measured within the total nuclear, defined by the Hoechst staining in blue, finding a >200-fold induction of total green nuclear intensity in dead cells (Figure 2b). Both measurements highlight the desired effect, with an EC50 of 26.7 µM for cell positivity (2a), and 26.3 µM based on intensity (2b), with R² values >0.96 for the nonlinear fit. This is nearly identical to the EC50 of 25.5 µM reported in Hela cells by the kit manufacturer.

To evaluate how pixel binning affects the results, the microplate was rescanned with a 4×4 bin, with 16 pixels summed at the level of the detector to yield 1 pixel of data. Rescanning the microplate took 3 minutes 30 seconds to yield full well data in 2 channels. Binning yielded similar results (Figure 2, red open circles) with an EC50 value of 30.2 for cell positivity. One advantage of binning is increased signal, with the 4x binned scan needing only a 1ms exposure time to detect the live/dead staining with comparable dynamic range as the original 15ms exposure in the full resolution scan. This means dim signals are detected more readily and exposure time is greatly reduced, important for dyes that easily bleach.

As the number of channels and microplates screened increases, file and hard drive size increasingly become an important consideration. Binning saves disk space, which is often a premium in large scale screening: for this microplate the 4x binned file size was 16 times smaller than the full resolution data (6.9 Gb instead of 110 Gb). These space considerations compound with more channels and the number of microplates being screened. Along with saving disk space comes greatly reduced data transfer times, as well as decreased analysis times (with less pixels to be analyzed). Binned live/dead data was analyzed 66% faster than the full resolution data with the same protocol. Binning within Endeavor offers greater flexibility and increased speed and signal, while decreasing the image file size, ideal in cases where submicron resolution is not needed.

This Application Note highlights the sensitivity of Endeavor and this live/dead assay, with 200:1 signal to noise ratio (Figure 2b). This leads to unambiguous identification of cell staining, resulting in a highly reliable assay, indicated by a Z’ score of 0.95. Endeavor can then be used to robustly quantify nuclear positivity with ease, here identifying dead cells with 85% well coverage and <4-minute scan time. The combination of coverage and speed is ideal for screens when it’s important to detect rare positivity events and when capturing subtle differences in cell death across an entire population matter. Beyond assays where viability is the primary metric, properly identifying dead cells can be essential to assay reliability and measurement accuracy, as dying cells stain differently than the rest of the study population, biasing analyses. The nuclear positivity workflow described here may be generalizable to other binary positivity assays such as transcription factor induction or cellular activation.

Human lung carcinoma epithelial cells (A549, ATCC) were seeded at 10,000 cells/well in a 96 well high content glass bottom microplate (Grenier 655891) and grown overnight at 37⁰C in 100 µL supplemented F-12 media. Cells were washed 1x in 100 µL phosphate-buffered saline (PBS), then valinomycin (Cayman 10009152) was applied in 1:2 dilution from 0.1 to 120 ng/mL in supplemented F-12 media, for a final volume of 100 µL/well, with 8 biological replicates. Cells were then incubated at 37°C and 5% CO2 for 24 hours. Image-iT DEAD Green stain was then applied at 1:75 dilution for 30 minutes before fixation for 15 minutes with 10% neutral buffered formalin. Then 1:1500 20mM Hoechst 33342 stain (Life Technologies 62249) was added followed by 2x washes with PBS. The microplate was then imaged in 2 channels on the Araceli Endeavor® in 1.3 software with maximized well coverage (4.72mm x 4.72mm/well), once using default resolution (0.27 µm/ pixel) with 13ms blue and 15ms green exposure times, and again a week later using a 4×4 bin (1.08 µm/pixel) with 3ms blue and 1ms green exposures.

Image analysis was performed using Araceli Clairvoyance analysis software. To calculate the ratio of dead cells to total cells by intensity, nuclei were defined and segmented using data from the blue channel (Hoechst stain) by machine vision tools and the green intensity quantified within the defined area. To calculate the ratio by cell positivity, machine vision tools defined the number of green positive cells colocalized with nuclear stain. Auto-fluorescent microplate edges were excluded from analysis. Graphing was done in GraphPad Prism™ with curve-fitting and EC50 calculated using the nonlinear regression function, using asymmetric (five parameter) fit.

Abdalah R, Wei L, Francis K, Yu SP. Valinomycin-induced apoptosis in Chinese hamster ovary cells. Neurosci Lett. 2006 Sep 11;405(1-2):68-73. doi: 10.1016/j.neulet.2006.06.055.

De Schutter E, Cappe B, Wiernicki B, Vandenabeele P, Riquet FB. Plasma membrane permeabilization following cell death: many ways to dye! Cell Death Discov. 2021 Jul 19;7(1):183. doi: 10.1038/s41420-021-00545-6. PMID: 34282124; PMCID: PMC8289853.

Lammerding J. Mechanics of the nucleus. Compr Physiol. 2011 Apr;1(2):783-807. doi: 10.1002/cphy.c100038. PMID: 23737203; PMCID: PMC4600468. Teledyne Photometrics. Binning. 2022 December 16. https://www.photometrics.com/learn/camera-basics/binning

Thermo Fisher Scientific. HCS Live/Dead Green Kit using HCS nuclearmask deepred. https://www.thermofisher.com/us/en/home/references/protocols/cell-and-tissue-analysis/protocols/hcs-live-dead-green-kit-using-hcs-nuclearmask-deep-red.html