Introduction

High-throughput screening of small molecule libraries, genetic reagents, and antibodies has been an indispensable tool for deciphering the mechanisms of disease and identifying potential therapeutics. Many diseases affect cell growth and division; therefore, assays of cell proliferation are especially important for high-throughput screening. High-throughput applications frequently employ end point assays that measure cell number by net metabolic activity (1,2), ATP content (3), or biomass (4). These whole-well assays can be accurate in their reflection of cell number but do not directly measure any aspect of the cell cycle and cannot distinguish reduced growth rates from increased death. Since DNA replication in the S phase of the cell cycle is a prerequisite for mitosis, proliferative capacity can be inferred from DNA synthesis through measurements of labeled nucleotide incorporation.

Cells replicating their genomes readily incorporate either radioactive or nonradioactive nucleoside analogs into their nascent DNA. Autoradiography of cells labeled with [³H]-thymidine provided early insights into the dynamics of DNA synthesis and cell proliferation in vivo (5,6,7). Labeling cultured cells with [³H]-thymidine allowed measurement of whole populations by scintillation counting (8), making this a rapid and accurate assessment of proliferative capacity of homogeneous cell populations. The thymidine analog 5-bromo-2′-deoxyuridine (BrdU) is a nonradioactive nucleoside that can be used to measure DNA synthesis on a per-cell basis when detected by microscopy or flow cytometry (9), allowing analysis of heterogeneous cell populations. Detection of incorporated BrdU requires both a monoclonal antibody and antigen retrieval by acid (9), heat (10), or nuclease treatment (11,12). Both immunodetection and enzymatic antigen retrieval are costly, while acid treatment can compromise other antigens, cell morphology, and reporter gene (e.g., GFP) function. A recently described chemical method of labeling DNA synthesis in cells and tissues uses the alkyne-substituted nucleoside 5-ethynyl-2′-deoxyuridine (EdU) as a reactive substrate for azide-substituted fluorescent dyes. During the copper-catalyzed alkyne azide cycloaddition (CuAAC or ‘click ’) reaction, a covalent bond is formed between the alkyne base and the azide dye. This reaction is rapid, specific, and requires minimal sample processing (13). We adapted this method without the use of commercial kits or custom reagents for use with 96-well plates on a high-throughput liquid handling platform. All experimental steps were automated, and the samples were processed in situ for analysis by automated quantitative fluorescence microscopy.

We chose MCF10A non-tumorigenic mammary epithelial cells as our model system because their proliferation can be manipulated by epidermal growth factor (EGF) and serum media supplements (14,15,16,17). We combined the chemical assay of DNA synthesis with measurement of DNA content to provide a more comprehensive cell cycle analysis. 4, 6′-diamidino-2-phenylindole (DAPI) is a quantitative DNA stain that can be used to resolve the G₀/G₁, S, and G₂/M phases of the cell cycle (18) and has been used previously to monitor cell cycle with high content imaging (19,20). Here we report high-content imaging analysis of DNA content and DNA synthesis by chemical detection in MCF10A cells grown in varied serum conditions, either in the presence or absence of epidermal growth factor (EGF). Similar analyses were conducted after treatments with inhibitors that target kinases downstream of EGF: EGF receptor (EGFR), mitogen extracellular kinase (MEK), and phosphoinositol-3 kinase (PI3K). We found that the CuAAC reaction with incorporated EdU produced intensely fluorescent nuclei with DNA content intermediate to the G₀/G₁ and G₂/M peaks. In these cells, DNA synthesis is dependent on EGF and, to a lesser extent, serum. Interrogation of intracellular signaling networks with kinase inhibitors in complete medium demonstrated that DNA synthesis requires MEK signaling and is partially dependent on PI3K signaling. Z′-factor analyses—a statistical calculation of assay uniformity and dynamic range—indicate this to be a robust high-throughput assay for quantifying the S-phase population. Together these data demonstrate the successful application of this new and efficient method of detecting DNA synthesis to high-throughput and high-content screening platforms.