Comparing automated cell imaging with conventional methods of measuring cell proliferation and viability

Abstract

The ability to assess cell proliferation and viability is essential for assessing new drug treatments, particularly in cancer drug discovery. This study describes a new method that uses a plate reader digital microscopy cell imaging and analysis system to assess cell proliferation and viability. This imaging system utilizes high throughput fluorescence microscopy with two fluorescent probes: cell membrane-impermeable SYTOX green and nuclear binding Hoechst-33342. Here we compare this technology to other known viability assays, namely: propidium iodide (PI)-based flow cytometry, and sulforhodamine B (SRB) and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) based plate reader assays. These methods were assessed based on their effectiveness in detecting the cell numbers of two adherent cell lines and one suspension cell line. Automated cell imaging was most accurate at measuring cell number in both adherent and suspension cell lines. The PI-based flow cytometry method was more difficult to use with adherent cells, while the SRB and MTT assays had difficulties when monitoring cells in suspension. Despite these challenges, it was possible to obtain similar results when quantifying the effect of cytotoxic compounds. This study demonstrates that the digital microscopy automated cell imaging system is an effective method for assessing cell proliferation and the cytotoxic effect of compounds on both adherent and suspension cell lines.

Introduction

Quantifying drug cytotoxicity is a critical tool for broadening our understanding of cell biology and identifying new therapeutic compounds. It is important that the techniques used to analyze the effects of compounds on cell proliferation and viability are accurate and reproducible. In this study, a viability dye-based imaging assay using the digital microscopy automated cell imaging system is compared to three commonly used methods: a flow cytometry-based assay using propidium iodide (PI), the sulforhodamine B (SRB) assay, and the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay.

Each of the existing methods have limitations. The flow cytometry assay utilizes fluorescent stains, such as PI, to identify and quantify the proportion of viable and dead cells (Dengler et al. 1995). PI is a DNA-binding dye that is impermeable to viable cells as it is unable to cross intact cellular membranes, however, it labels dead cells. Improved technology has allowed for the determination of the flow rate of all cells, therefore allowing the direct number of viable cells in a population to be ascertained. Flow cytometry, however, does require the detachment of adherent cells. The cells also have to remain intact to register as a cell in the flow cytometer, and they have to retain their nucleus to be identified as a dead cell.

The SRB assay is a protein-based colorimetric assay that relies on the binding of SRB to basic amino acid residues in TCA-fixed cells (Skehan et al. 1990). This method assumes that dead cells are removed from the environment after cell fixation, and therefore SRB is only binding to protein of live cells. Electrostatically bound SRB is then solubilized by a weak base such as Tris and absorbance is measured. The amount of signal measured is proportionate to the amount of protein present, therefore an indication of cell number. Until recently, the SRB assay has been used by the National Cancer Institute (NCI), USA, to screen drugs on the NCI-60 panel of cancer cell lines (Holbeck et al. 2010). As of October 2023, the NCI changed their method to use a more sensitive luminescent probe, CellTiter-Glo®, enabling use of 384-well plates and therefore higher throughput of drug screens and cancer cell lines than the traditional 96-well plates (National Cancer Institute 2023). This assay is more difficult to perform with suspension or semi-adherent cells as viable cells can be lost during wash steps.

The MTT assay is a colorimetric assay that measures metabolically active cells through the conversion of colorless MTT to a colored formazan product (Mosmann 1983). The cleavage of the tetrazolium ring of MTT occurs in metabolically active mitochondria. Mosmann stated that the amount of formazan is directly proportionate to the cell number, when more accurately it is proportionate to the activity of their cellular oxidoreductases that reduce MTT.

The method described here, using the digital microscopy automated cell imaging system, employs the use of two fluorescent probes: live and dead cell nucleic acid-binding Hoechst-33342 (Bucevičius et al. 2018) and SYTOX green (Thakur et al. 2015) which is live cell impermeable and only stains nucleic acid of dead cells with compromised cell membranes. The use of two probes, the former which binds to all cells and the latter which only binds to dead cells, allows for the determination of the number of cells by analyzing microscopy images of whole wells from a 96-well plate. The assay is directly assessing viability, and there is no need to either detach or wash the cells, which should remove some of the limitations of other methods.

This study aims to compare the ability of the different technologies to measure alterations in cell growth and viability of both adherent and suspension cell lines.

Methods

Cell culture

Two adherent human metastatic melanoma cell lines, MALME-3M and SK-MEL-2, were obtained from the NCI (National Institutes of Health, USA). Suspension Jurkat T-lymphoma cell line was obtained from ATCC (clone E6-1 TIB-152). Cells were cultured in T75 flasks and maintained in RPMI 1640 medium (Gibco™, Life Technologies, USA) supplemented with 10% (v/v) fetal bovine serum (Moregate Biotech, New Zealand), 100 units/mL penicillin and 100 µg/mL streptomycin (Gibco™, Life Technologies, USA) at 37 °C, 5% CO₂ and 21% O₂.

Drug preparation

Portimine was gifted from the Cawthron Institute, Nelson, New Zealand. A stock solution of 500 µM was prepared in dimethyl sulfoxide (DMSO, Fisher Bioreagents, USA) and all subsequently prepared concentrations contained 0.2% DMSO.

Hydrogen peroxide (H₂O₂) (Lab Serv, ThermoFisher Scientific, USA) was prepared by diluting stock solution 1:4 in milliQ water. H₂O₂ concentration was determined by measuring absorbance at 240 nm and calculated using the extinction coefficient (ɛ = 43.6). Stocks of differing H₂O₂ concentrations were then prepared by dilution in milliQ water.

Cell seeding and treatment

Adherent cells were dissociated from T75 flasks by incubation in TrypLE® Express trypsin (Gibco™, Life Technologies, USA). Dissociation was not required for suspension cells. All cells were counted using a hemocytometer. Cells were diluted in medium accordingly and 100 µL of cells were added to each well of 96-well flat bottom plate (Nunclon^TM Delta Surface, Thermo Fisher Scientific, USA), then left to settle for 24 h.

For assessing the accuracy of detection of cell number and its linearity by each method, a range of 625-20,000 cells were seeded per well in a total volume of 100 µL. Cell number was then assessed after 24 h.

For portimine and H₂O₂ treatment, 5,000 cells were seeded per well in a total volume of 100 µL. A total of eight plates were seeded per cell line, four of which were used for accuracy of cell number detection and time point zero (Tz, number of cells 24 h after seeding and at the time of treatment of cells with portimine and H₂O₂) analysis for each different assay, and the remaining four were used for portimine and H₂O₂ treatment for each different assay. After 24 h settling period, cell number accuracy and Tz plates were analyzed using each of the four assays described below. Drugs were added to cells and incubated for 48 h.

Analysis of cell growth inhibition and cell death

The number of cells determined at the time of treatment (Tz) for each method was used to determine cell growth inhibition and cell death. After 48 h of treatment with H₂O₂ or portimine, the final number of viable cells (Ti) was measured and normalized to the cell number at the timepoint of treatment (Tz) and untreated control cell densities (C). Net cell growth inhibition and cell death were determined using the following calculations:

$$\mathit{\text{If Ti}}\ge \mathit{\text{Tz}}:\text{\%\hspace{0.17em}}\mathit{\text{growth}}\hspace{0.17em}\hspace{0.17em}\mathit{\text{inhibition}}=\frac{\left(\mathit{\text{Ti}}-\mathit{\text{Tz}}\right)}{\left(C-\mathit{\text{Tz}}\right)}\times 100$$

$$\mathit{\text{If Ti}}<\mathit{\text{Tz}}:\text{\%\hspace{0.17em}}\mathit{\text{growth}}\hspace{0.17em}\hspace{0.17em}\mathit{\text{inhibtion}}=\frac{\left(\mathit{\text{Ti}}-\mathit{\text{Tz}}\right)}{\mathit{\text{Tz}}}\times 100$$

Values greater than 100% indicate increased cell proliferation compared to the untreated control, values between 0% and 100% indicate inhibition of cell proliferation, and negative values indicate cell death.

Digital microscopy automated cell imaging system assay

SYTOX green (ThermoFisher Scientific, USA) and Hoechst-33342 (ThermoFisher Scientific, USA) fluorescent dyes were added to cells at final concentrations of 1 µM and 500 µg/mL, respectively. Cells were incubated with fluorescent dyes for 1 h at 37 °C. For suspension cells, plates were incubated in an environmental chamber at 37 °C and 5% CO₂ in the microscope to allow cells to settle to the base of the well. Whole wells were then imaged on ImageXpress PICO automated cell imaging system (Molecular Devices) at 4x magnification using DAPI and FITC channels. “Apoptosis” analysis was employed using the CellReportXpress software to determine total cell numbers (Hoechst-positive), and within this population the number of dead (SYTOX green-positive) cells.

Propidium iodide-based flow cytometry assay

The propidium iodide (PI) flow cytometric assay has been widely used for the evaluation of cell death in different experimental models (Riccardi and Nicoletti 2006). Adherent cells were trypsinized using 100 µL TrypLE® Express trypsin to dissociate from the well bottom. Trypsinisation was stopped by the addition of 150 µL medium and pipetted up and down to ensure complete dissociation from the well bottom. Propidium iodide (PI, Sigma-Aldrich, Merck, Germany) was added to cells to get a final concentration of 2 µg/mL. For suspension cells, PI was added to cells to get a final concentration of 2 µg/mL. Cells were analyzed using CytoFLEX flow cytometer (Beckman Coulter, USA) and CytExpert 2.4 software, which recorded percentage and concentration (cells/µL) of PI-positive and PI-negative cells based on flow rate (excitation 562 nm and emission 690 ± 25 nm). From this, the total number of PI-positive and PI-negative cells would be calculated by multiplying the concentration by total volume (250 µL).

Sulforhodamine B assay

SRB assay was performed as previously described (Holbeck et al. 2010). In brief, adherent cells were fixed with 50% (w/v) ice-cold trichloroacetic acid (TCA, Sigma-Aldrich, Merck, Germany) for 1 h at 4 °C. Suspension cells were centrifuged for 5 min at 1000 g to settle at the bottom of the plate before fixing with 80% (w/v) ice-cold TCA. Fixed cells were then washed with milliQ H₂O before the plate was air dried. A total of 50 µL of SRB solution (0.4% (w/v) SRB sodium salt (Santa Cruz Biotechnology, USA) in 1% acetic acid) was added to cells and incubated for 10 min at room temperature. Excess SRB was removed with five milliQ washes. Bound SRB was solubilized in 100 µL of 10 mM Tris and absorbance was read at 515 nm on the Multiskan SkyHigh Microplate spectrophotometer (Thermo Scientific™, ThermoFisher Scientific USA). Cell-free control wells were included in analysis and subtracted from absorbance readings of treatment wells to account for background signal.

MTT assay

The MTT assay was adapted from previously described methods (Mosmann 1983; Alley et al. 1988; Morgan 1998). For adherent cells, medium was removed from cells and 100 µL of 0.5 mg/mL MTT (Sigma-Aldrich, Merck, Germany) made up in RPMI medium was added to wells and incubated for 3 h at 37 °C. For suspension cells, the plate was first centrifuged for 5 min at 1000 g to settle cells at the well bottom before media was carefully removed without disturbing cells at the bottom of the well and MTT was added. After incubation, an equal volume of solubilization solution (89% isopropanol, 10% Triton-X-100, 0.1 M HCl) was added to cells. Formazan crystals formed were dissolved by pipetting. Absorbance was read at 570 nm on the Multiskan SkyHigh Microplate spectrophotometer. Cell-free control wells were included in analysis and subtracted from absorbance readings of treatment wells to account for background signal.

Statistical analysis

For each individual experiment, six individual wells for each treatment condition (n = 6 technical replicates) were used to calculate the mean value. Biological replicates were assessed, defined as cells from different passages.

All graphs were prepared and statistical analysis was performed using GraphPad Prism version 9. GI₅₀ and LC₅₀ were determined using a non-linear fit of variable slope. The GI₅₀ value is the concentration at which 50% of cell proliferation is inhibited. The LC₅₀ value is the concentration at which 50% of cell death is achieved. Analysis of the sigmoidal curve fitted was used to assess the difference in GI₅₀ and LC₅₀. One-way ANOVA analysis with Tukey’s multiple comparisons was used to compare GI₅₀ and LC₅₀ values calculated for each assay, where p < 0.05 was considered statistically significant.

Results

Accuracy of detection of cell number

The first objective was to assess the ability of each method to accurately quantify the number of cells seeded at different densities. Figure 1 depicts full well imaging and viability-dye based analysis of wells using the integrated software of the digital automated microscopy cell imaging system.

Figure 1.

Representative microscope images and analysis from the digital microscopy automated cell imaging system. SK-MEL-2 melanoma cells after 48 h treatment with 1 nM portimine, probed with Hoechst-33342 and SYTOX green. Fluorescence images were taken in DAPI and FITC channels (C, G) and superimposed with transmissible light images (A, E). Analysis of images was performed, displayed in the images on the right with red highlighting viable cells (Hoechst-positive, SYTOX green-negative) and green highlighting dead cells (Hoechst-positive, SYTOX green-positive) (B, D, F, H). (A–D) 4x microscopic images of whole wells of a 96-well plate are displayed. (E-H) Zoomed in images of cells, demonstrating nuclear Hoechst-33342 staining of all cells (blue) and SYTOX green staining of dead cells (green).

Using the two adherent melanoma cell lines and one suspension T-lymphoma cell line, cells were counted using a hemocytometer, and different numbers of cells were seeded in 96-well plates. The cells were left to settle for 24 h before analysis. Images taken 24 h after seeding with the automated cell imaging system demonstrate seeding densities of cell lines (Figure 2). Figure 3 depicts the detection of cells using the digital microscopy automated cell imaging assay (Figure 3(A)), flow cytometry assay (Figure 3(B)), SRB assay (Figure 3(C)), and MTT assay (Figure 3(D)).

Figure 2.

Representative images of cells at different seeding numbers of 1250, 5000 and 20000 cells. Imaged on the automated cell imaging system with DAPI and FITC fluorescent channels, with (top panels) and without (bottom panels) superimposed transmissible light channel. Whole wells were imaged and analyzed.

Figure 3.

Detection of cells seeded at varying densities using the automated cell imaging system (A), flow cytometry (B), SRB (C) and MTT (D) assays. Two adherent melanoma cell lines (MALME-3M, black; SK-MEL-2, green) and one suspension T-lymphoma cell line (Jurkat, orange) were seeded in 96-well plates at different densities: 625, 1250, 2500, 5000, 10000 and 20000 cells. After a 24-h settling period, each assay was performed and cell density was measured as total number of cells counted (A–B) or absorbance (C–D). The grey dashed line (A-B) indicates y = x linearity. Error bars are the standard error of the mean from n = 4 biological replicates.

The automated cell imaging system reliably measured viable numbers of cells, regardless of whether the cell line was adherent or suspension, as demonstrated by the number of cells counted being consistent across all three cell lines. The number of cells counted was higher than what was seeded, particularly in the higher cell numbers, which can be explained by the expected proliferation that will have occurred within 24 h. Linearity was consistent across all cell lines up to the highest cell density (Figure 3(A)).

For the suspension cell line, the flow cytometry assay performed exceptionally well, linearly detecting viable cells representative of the cell numbers seeded (Figure 3(B)). However, for the two adherent cell lines this assay did not accurately measure the number of viable adherent cells. This is likely due to the trypsinization step that is required to dissociate adherent cells from the base of the well before detection by flow cytometry. Trypsinization may not have been effective at completely dissociating all cells from the base of the plate wells, therefore a proportion of the cells may not have been analyzed. Additionally, the trypsinization process sometimes results in multiple cells clumping together and being analyzed as one, resulting in an underestimation.

In comparison, the SRB and MTT assays were effective for the adherent cell lines (Figure 3(C–D)), although there was large variability within experiments. The MTT assay lost linearity at high cell density for both adherent cell lines. The SRB and MTT assays performed poorly for the suspension cell line, resulting in low absorbance changes that did not reflect the seeded cell numbers. This is most likely due to loss of cells during the cell fixation step for the SRB assay, and the centrifugation and media removal step for the MTT assay.

Viability assays

Each of the assays were assessed for their ability to measure effects on cell proliferation and viability after drug treatment. Two types of treatment were selected: oxidant treatment with hydrogen peroxide (H₂O₂) or with the cytotoxin, portimine (Cuddihy et al. 2016). A total of 5,000 cells were seeded in two 96-well plates and left to settle for 24 h, at which point one of the plates was used to quantify cell numbers at the time of treatment (Tz). After 48 h of treatment with H₂O₂ or portimine, the final number of viable cells (Ti) was measured and normalized to starting cell numbers (Tz) and untreated control cell densities (C). This allowed for net cell growth inhibition and cell death to be determined. Values greater than 100% indicated increased cell proliferation compared to the untreated control, values between 0% and 100% indicated inhibition of cell proliferation, and negative values indicated cell death.

Further analysis by fitting sigmoidal curves allowed the interpolation of GI₅₀ and LC₅₀ values, as summarized in Table 1 for each cell type, treatment, and assay. The GI₅₀ value is the concentration where 50% of cell proliferation is inhibited. The LC₅₀ value is the concentration where 50% of cell death is achieved.

Table 1.

GI₅₀ and LC₅₀ values for cells treated with hydrogen peroxide (H₂O₂) and portimine.

Cell type	Treatment	Assay	GI50 (95% CI)	LC50 (95% CI)
MALME-3M	H2O2 (mM)	Automated cell imaging system	0.5 (0.3–1.0)	0.5 (0.3–1.0)
		SRB	0.4 (0.3–0.5)	0.7 (0.6–0.9)
		MTT	0.4 (0.3–0.4)	0.6 (0.6–0.8)
		Flow cytometry	0.4 (0.3–0.5)	1.0* (0.8–1.7)
	Portimine (nM)	Automated cell imaging system	0.6 (0.4–0.8)	1.8 (1.2–3.3)
		SRB	0.5 (0.4–0.7)	1.4 (1.0–2.1)
		MTT	0.6 (0.4–0.8)	2.1 (1.2–100)
		Flow cytometry	0.8 (0.5–1.2)	–
SK-MEL-2	H2O2 (mM)	Automated cell imaging system	0.1 (0.1–0.2)	0.5 (0.4–0.6)
		SRB	0.2 (0.1–0.2)	0.6 (0.5–0.8)
		MTT	0.1 (0.0–0.2)	0.6 (0.4–0.9)
		Flow cytometry	0.2 (0.1–0.2)	1.0 (0.7–1.4)
	Portimine (nM)	Automated cell imaging system	2.8 (0.9–3.8)	14.3 (9.9–32.2)
		SRB	1.7 (1.0–2.7)	11.7 (7.2–28.6)
		MTT	2.0 (1.3–2.8)	34.0 (20.2–80.6)
		Flow cytometry	2.7 (1.5–4.2)	–
Jurkat	H2O2 (mM)	Automated cell imaging system	< 0.1 (0.0–0.02)	< 0.1 (0.0–0.04)
		SRB	< 0.1 (0.0–0.5)	–
		MTT	< 0.1 (0.0–0.5)	–
		Flow cytometry	< 0.1 (0.0–0.06)	< 0.1 (0.0–0.8)
	Portimine (nM)	Automated cell imaging system	0.4 (0.3–0.5)	1.2 (0.9–1.8)
		SRB	0.1 (0.0–0.3)	–
		MTT	0.5 (0.2–0.8)	2.5 (1.0–100)
		Flow cytometry	0.4 (0.3–0.5)	1.7 (1.2–2.4)

Values are determined from three biological experiments, with lower and upper 95% confidence intervals (CI) included in brackets. * indicates statistically significantly different value compared to other methods, determined by one-way ANOVA with Tukey’s multiple comparisons.

Compared to automated cell imaging, flow cytometry appeared to underestimate cell death in adherent melanoma cells when treated with both H₂O₂ and portimine. The MTT and SRB assay overestimated cell death in MALME-3M adherent melanoma cells treated with H₂O₂, however this was not seen in SK-MEL-2 adherent melanoma cells and there was no difference when assessing the effect of portimine on both cell lines. With the suspension Jurkat T-lymphoma cell lines, both the SRB and MTT assay underestimated H₂O₂-induced cell death compared to the digital microscopy automated cell imaging assay, with both SRB and MTT recording less cell death at higher concentrations of H₂O₂ compared to lower concentrations. Similarly, the SRB assay underestimated cell death when suspension Jurkat T-lymphoma cell lines were treated with portimine. Additionally, the SRB assay appeared to overestimate inhibition of cell proliferation, although there was large variability.

When assessing the effect of drugs on cells, these assays are typically used without consideration of cell numbers at the time of treatment (Tz), resulting in data being displayed as a percentage of untreated control at endpoint of treatment. While this form of presenting viability data is used by many, it should be understood that results may not provide clarity on the effect of the drug, whether it be inducing cell death or simply inhibiting cell proliferation. Figure 5 depicts the same data as Figure 4, however, only the endpoint data is considered. Using endpoint data, it is unclear whether the concentration where 50% of the control is achieved is due to 50% cell death, 50% of cell proliferation being inhibited, or a combination of both. The addition of a Tz plate provides much more insight into the effect of the drug and therefore should be added to any experiments assessing the effect of drugs on cells.

Figure 5.

Treatment of two adherent melanoma cell lines (MALME-3M, SK-MEL-2) and one suspension T-lymphoma cell line (Jurkat) with hydrogen peroxide (H₂O₂) and portimine, displayed as percentage of control using four different assays. These results use the same raw data as Figure 4 but without consideration of Tz, therefore data can only be plotted as a percentage of untreated control.

Figure 4.

Effect of hydrogen peroxide (H₂O₂) and portimine on cell proliferation and viability, measured using four different assays. Two adherent melanoma cell lines (MALME-3M, SK-MEL-2) and one suspension T-lymphoma cell line (Jurkat) were assessed for their sensitivity to H₂O₂ and portimine after 48-h treatment. Error bars represent the standard error of the mean from n = 3 biological replicates.

For example, for cells treated with H₂O₂, using only endpoint data it appears that there is almost complete loss of cell viability from 1 mM for melanoma cell lines (Figure 5(A,C)) and from 0.1 mM for Jurkats (Figure 5E). However, when the Tz plate is considered, for some of the methods, particularly flow cytometry, only 50% of cell death is achieved at 1 mM for melanoma cells (Figure 4(A,C)) and 0.1 mM for Jurkats (Figure 4(E)). Similarly, for MALME-3M cells treated with portimine, it may be interpreted that 1 nM reduces cell growth by 50% or causes 50% cell death when looking at end point data alone (Figure 5(B)). However, when the Tz plate is considered, it clearly demonstrates that cell growth is completely inhibited at 1 nM portimine but limited cell death has occurred (Figure 4(B)).

Discussion

We have assessed the ability of a high-throughput automated cell imaging system to accurately quantify the effect of drugs on cell proliferation and viability by comparing to more traditional assays. The advantages and disadvantages for each method are summarized in Table 2. The automated cell imaging system relies on fluorescence microscopy and direct read outs of cell numbers based on the detection of Hoechst-positive and SYTOX green-positive cells. This imaging approach allows for the generation of accurate data that can be confirmed by visual analysis of microscopy images. The propidium iodide (PI)-based flow cytometry method also relies on a similar approach of a fluorescence-based assay and single-cell analysis, analyzing the positive or negative staining of a proportion of the cell population within the well. However, it has been demonstrated that this method is effective for suspension cells but not adherent cells, likely due to the preparation steps required prior to analysis, including trypsinization of adherent cells, interfering with the accurate measurements of cell number. It is also important to note that the whole population of cells is not able to be assessed with the PI-based flow cytometry method. The automated cell imaging system has very few preparation or cell handling steps involved in the method, saving time and making it highly reproducible resulting in reduced variation between users and across laboratories. Additionally, the whole well is analyzed, ensuring that all cells are accounted for in assessing viability status. This is also true for both the SRB and MTT assays, which assess the entire well of cells. These two assays were effective at analyzing the number of cells seeded for adherent cells, however, not for suspension cell lines. Ineffective fixation for the SRB method and ineffective centrifugation for the MTT assay cause difficulties, resulting in loss of viable cells. Furthermore, both these assays make assumptions that either dead cells dissociate from the base of the well (SRB assay) and that formazan formation is directly proportionate to cell number (MTT assay). While these hold true under most circumstances, it is possible that viable cells dissociate and are lost in wash steps for the SRB assay, which could result in overestimation of cell death. For the MTT assay, differences in cellular metabolic activity mean that formazan is not directly proportionate to cell number, but rather to the activity of cells which can be influenced by altered mitochondrial activity. A compound that decreased this metabolic activity could be misinterpreted as being cytotoxic. It is important to still consider the limitation of cell confluency that affects all methods, by impacting on both growth characteristics and subsequent analysis. There is always an upper limit of how many cells can be assessed in a fixed area of a well. Cells should be seeded at a density that ensures that 100% confluence is not achieved during experimentation.

Table 2.

Summary of four assays assessed, including the method of analysis, single cell or population-based analysis, and the advantages and disadvantages of each method.

Assay	Method of analysis	Single cell vs population-based	Advantages	Disadvantages
Automated cell imaging system	Fluorescence	Single cell (although does assess whole population of cells in well)	Good for adherent and suspension cells Whole well analyzed Minimal steps Cells able to be visualized Environmental control High throughput Low running costs	Imaging can take some time, dependent on the number of fluorescent channels being used Acquisition and maintenance costs if equipment not established
Propidium-iodide based flow cytometry	Fluorescence	Single cell	Good for suspension cells High throughput Low running costs	Adherent cells dependent on sufficient trypsinisation Acquisition and maintenance costs if equipment not established
SRB	Colorimetric	Population	Good for adherent cells High throughput Low running costs	Suspension cells lost in wash steps Assumption that all dead cells dissociate from well Multiple steps Long incubation periods Acquisition and maintenance costs if equipment not established
MTT	Colorimetric	Population	Good for adherent cells High throughput Low running costs	Suspension cells lost in wash steps Assumption that formazan production is directly proportionate to cell number Multiple steps Long incubation periods Acquisition and maintenance costs if equipment not established

Despite the differences identified when assessing the ability to detect cells at different seeding densities, the overall impact on assessing the effects of H₂O₂ and portimine on cell proliferation and viability did not differ greatly between the four methods. This is likely due to the fact that the Tz data, which the endpoint data is normalized to, is measured in the same method, therefore any downfalls of methods are negated. This further emphasizes the benefit of including a Tz plate in analysis, which additionally provides insight into the effect of drugs on cell proliferation and cell death. Added benefits of the automated cell imaging include the ability to effectively compare adherent and suspension cell lines, as well as the reduced time it takes to perform the assay, the ability to include other dyes of interest for staining, microscopic imaging of cells which may provide insight into morphological changes occurring to the cells and mechanisms of cell death, and environmental control allowing for long term imaging studies.

The National Cancer Institute (NCI) recently announced a change from using the SRB assay in 96-well plates to a new method utilizing fluorescent probe CellTiter-Glo on a 384-well plate (National Cancer Institute 2023). Reasons stated for the change included the highly automated increased throughput with fewer cells required due to the smaller well size and increased sensitivity of the fluorescent probe, and additionally, fewer handling steps involved in the protocol resulting in higher reproducibility. The automated cell imaging system is similar to this new method, utilizing fluorescence imaging for analysis. It is also able to analyze 384 well plates, however, this has not been tested in this study. One limiting factor of the automated cell imaging system is the time required to image a plate. Acquiring images of the whole well in two fluorescent channels (DAPI and FITC) takes longer than an individual fluorescent read on a plate reader, which would also be the case for the CellTiter-Glo method now used by the NCI. However, the benefit is that microscopy images provide more insights into the cell morphology. Therefore, it may be beneficial to use the automated cell imaging system in conjunction with the new method employed by the NCI, whereby high throughput screening of drugs is used first to identify compounds of interest, followed up by the automated cell imaging system that may uncover details of cellular responses and mechanisms of cell death.

Quantitative optical imaging has seen dramatic methodological advancements over the past few years and shown to be a transformative approach to analysis of single-cell biology. Quantitative microscopy is likely to change significantly in the next decade toward measuring specific biomolecules and subcellular structures (Bagheri et al. 2022). As discussed by Bagheri et al., substantial progress is needed throughout the workflow, particularly in data generation, management, analysis and sharing data in standardized ways for quantitative optical imaging to reach its full potential. If the automated imaging system described in this paper is to be utilized more frequently in research, validation against other methods will be valuable. We recommend that standard operating procedures are established to ensure consistent quantification between different research laboratories.

Funding Statement

This research was supported (in part) by a contract from the Health Research Council of New Zealand, the Richdale Charitable Trust, the Maurice Wilkins Center and a University of Otago PhD scholarship.

Author’s contributions

TF, conceptualization, lab work, analysis, manuscript preparation; SH, lab work, analysis, manuscript preparation; LCDS, digital microscopy automated cell imaging system funding acquisition, established methodology; MBH, conceptualization, editing of manuscript, funding acquisition; MPP, conceptualization, editing and manuscript submission.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The authors confirm that the data supporting the findings of this study are available within the article.