The slow cell death response when screening chemotherapeutic agents

Abstract

Purpose: To examine the correlation between cell death and a common surrogate of death used in screening assays, we compared cell death responses to those obtained with the sulforhodamine B (SRB) cell protein-based "cytotoxicity" assay.

Method: With the SRB assay, the Hill equation was used to obtain an IC50 and final cell mass, or cell mass present at infinite agent concentrations, with eight adherent cell lines and four agents (32 agent/cell combinations). Cells were treated with high agent concentrations (well above the SRB IC50) and the death response determined as the time-dependent decrease in cells failing to bind both annexin V and vital fluorochromes by flow cytometry.

Results: Death kinetics were categorized as fast (5/32) (similar to the reference nonadherent Jurkat line), slow (17/32), or none (10/32), despite positive responses in the SRB assay in all cases. With slow cell death, a single exposure to a chemotherapeutic agent caused a slow, progressive increase in dead (necrotic) and dying (apoptotic) cells for at least 72 h.

Conclusions: Cell death (defined by annexin and/or fluorochrome binding) did not correlate with the standard SRB "cytotoxicity" assay. With the slow cell death response, a single exposure to an agent caused a slow conversion from vital to apoptotic and necrotic cells over at least 72 h (the longest time point examined). Here, increasing the time of exposure to agent concentrations modestly above the SRB IC50 provides a method of maximizing cell kill. If tumors respond similarly, sustained low doses of chemotherapeutic agents, rather than a log-kill, maximum tolerated dose strategy may be an optimal strategy of maximizing tumor cell death.

Fig. 1

The cell death response and the SRB cell mass assay. a To obtain the death response, cells were plated at 80,000 cells/well in a 24-well plate format and treated with agent concentrations above the IC50 as determined by the SRB assay. After 48 h, cells were detached, incubated with a fluorescent annexin and a vital dye, and analyzed by dual wavelength flow cytometry. The survival fraction at 48 h (SF48) was the fraction of cells negative for annexin and vital dye (represented by the gray quadrant). b For the SRB assay, cells were plated at 5,000 cells/well in a 96-well format and allowed to proliferate for 48 h. With no agent (No A), the growth control (GC) was the 100% value (solid gray circle). Increasing molar concentrations of agent [A] yielded data points (open circles) analyzed with the Hill equation to obtain a value of n (slope), an IC50, and a final cell mass (FCM). c ICM is not used in curve fitting or data reduction but can be greater than equal to or less than the FCM, yielding a categorization system cell death based on the SRB assay

Fig. 2

Comparison of cell death responses with responses shown in SRB assays. a Fast cell death responses of the CPT/786-0 and VBN/HeLa combinations. From left to right: linear/log plots of the SRB data fit to the Hill equation, flow cytometry dot plots of untreated cells, dot plots of treated cells, and a summary of SF48’s and SRB assay values. CPT and VBN caused a nearly complete conversion of viable (lower left) quadrant cells to annexin and/or vital dye positive populations. b Slow death responses with the CPT/A549 and PXL/MDA-MB231 combinations. Agents caused a partial conversion of viable cells to annexin and vital dye binding cells. c No death response with the CPT/PC-3 and 5-FU/U87 mg combinations. Viable cells did not convert to annexin or vital dye positive cells

Fig. 3

Cell line-dependent and time-dependent decreases in the SF cell death response parameter. a The SF decrease for the suspended Jurkat and adherent 786-0, A549, and PC-3 cell lines. Lines were obtained by fits to a linear equation or a simple exponential decay equation. b SF48’s as a function of cell line and agent for adherent cell lines and suspended Jurkat cells. Dotted lines are the cutoffs between fast and slow death responses, and between slow and no cell death responses

Fig. 4

Cell line doubling times. Squares are means and error bars are the standard errors of at least 6 values. Cell doubling times were from growth controls (Fig. 1b, GC’s) and initial cell masses (ICM’s) where k = [ln (GC/ICM)]/48 h]. Triangles are values from http://dtp.nci.nih.gov/docs/misc/common_files/cell_list.html except for HeLa and U87 cell lines which were from [20] and [21], respectively