Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2009 Jun 1.
Published in final edited form as: Toxicol In Vitro. 2008 Mar 4;22(4):1099–1106. doi: 10.1016/j.tiv.2008.02.013

A Bioluminescent Cytotoxicity Assay for Assessment of Membrane Integrity Using a Proteolytic Biomarker

Ming-Hsuang Cho 1, Andrew Niles 2, Ruili Huang 1, James Inglese 1, Christopher P Austin 1, Terry Riss 2, Menghang Xia 1,*
PMCID: PMC2386563  NIHMSID: NIHMS49482  PMID: 18400464

Abstract

Measurement of cell membrane integrity has been widely used to assess chemical cytotoxity. Several assays are available for determining cell membrane integrity including differential labeling techniques using neutral red and trypan blue dyes or fluorescent compounds such as propidium iodide. Other common methods for assessing cytotoxicity are enzymatic “release” assays which measure the extracellular activities of lactate dehydrogenase (LDH), adenylate kinase (AK), or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in culture medium. However, all these assays suffer from several practical limitations, including multiple reagent additions, scalability, low sensitivity, poor linearity, or requisite washes and medium exchanges. We have developed a new cytotoxicity assay which measures the activity of released intracellular proteases as a result of cell membrane impairment. It allows for a homogenous, one-step addition assay with a luminescent readout. We have optimized and miniaturized this assay into a 1536-well format, and validated it by screening a library of known toxins from the National Toxicology Program (NTP) using HEK 293 and human renal mesangial cells by quantitative high-throughput screening (qHTS). Several known and novel membrane disrupters were identified from the library, which indicates that the assay is robust and suitable for large scale library screening. This cytotoxicity assay, combined with the qHTS platform, allowed us to quickly and efficiently evaluate compound toxicities related to cell membrane integrity.

Keywords: 1536-well, NTP 1408 compound library, membrane integrity, cytotoxicity assay, protease release assay, qHTS, renal mesangial cells, HEK293 cells

Introduction

Cell-based assay methods are now routinely incorporated into high throughput screening activities for drug discovery and development (Digan, 2005; Boisclair, 2005). Automated dispensing and ancillary robotic systems make it possible to rapidly and efficiently interrogate a compound library for biological effects on a target cell population. As assay well densities increase and volumes plummet however, new physical constraints and practical hurdles commonly emerge that negatively impact reproducibility and sensitivity (Rose, 1997; Maffia, 1999). Therefore, new robust and scalable reagents are sought to produce higher quality data in a cost-effective manner.

Cytotoxicity is one of the most common biological parameters measured after experimental manipulation largely because it is easily measured and adheres to the dose-dependence paradigm of Paracelsus. It is well known that many drugs produce physiological and therapeutic actions at lower concentrations while the toxic effects including necrosis or apoptosis occur at higher concentrations. Regardless of the mechanism of cell death, mammalian cells after interacting with toxins undergo a series of dramatic structural and morphological changes that lead to loss of membrane integrity (Leist, 2001). This irreparable damage to cellular architecture allows free movement of previously excluded molecules into the cell, as well as their enzymatic contents to leak into the culture medium (Riss, 2004). Therefore, measurement of intracellular enzyme marker activity in the extra-cellular environment is the basis of several cytotoxicity assays for accessing cell membrane integrity.

Currently, several cell-based cytoxicity assays are available for determining cell membrane integrity. Dye uptake assays using neutral red, trypan blue and fluorescent compounds such as propidium iodide are traditional methods to assess cell viability and membrane damage. Trypan blue and fluorescent compounds such as propidium iodide are excluded by healthy, functional cell membranes but are able to traverse damaged cell membranes thereby preferentially staining dead cells. Conversely, neutral red accumulates only in the lysosomes of viable cells. The lactate dehydrogenase (LDH) release assay is widely used in in vitro toxicology studies. This assay is based on the measurement of LDH activity in the extracellular medium. Membrane integrity can be also evaluated by other enzyme release assays including adenylate kinase (Olsson et al., 1983) or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Corey et at., 1997), which are present in all cells. These enzymes are normally compartmentalized within the cell, but their activities are significantly increased in the extracellular environment as a result of cell death. However, these assays have their limitations including multiple reagent additions, low sensitivity, low-throughput and scalability, poor linearity, and the need for requisite washes or medium exchanges.

Although many assay methods have been developed for determining cytotoxicity in lower density formats, precious few have been validated for use in high throughput screening that requires simple, homogenous assay procedure with robust assay signal. Niles et al. recently described a novel proteolytic biomarker profile for cytotoxicity that embodies many biological attributes desirable for robust assay development (Niles et al., 2007). Subsequently, a homogenous luminescent assay was developed for measuring this biomarker that comprises a formulation containing an aminoluciferin-conjugated substrate, ATP, MgSO4 and a thermostable, recombinant luciferase (Figure 1). We describe here the development and validation of a bioluminescent cytotoxicity assay using a protease biomarker in a 1536-well plate format, by the screening of the1408 compound collection from National Toxicology Program (NTP) (Xia et al., 2008). We have identified several known and novel membrane disrupters by screening NTP library, which indicates that the assay is robust and suitable for large scale library screening.

Figure 1.

Figure 1

Open in a new tab

Principle of the luminescent protease-release assay. AAF-aminoluciferin is not a substrate for recombinant luciferase, so viable cells generate only a modest luminescence background. Proteases released from compromised cells cleave the substrate liberating aminoluciferin which serves as a substrate for luciferase. The resulting “glow-type” signal is stable and proportional to the number of dead cells.

Materials and Methods

Reagents

Digitonin was obtained from Sigma-Aldrich (St. Louis, MO, U.S.A). The 1408 compound collection was provided by NTP (Xia et al., 2008). Among 1408 compounds, 55 of the compounds were represented twice in the collection, giving a total of 1353 unique compounds. Opaque white, clear bottom 96 well and opaque white, solid bottom 1536 well plates were obtained from Corning Costar (Acton, MA, U.S.A.). The ToxiLight® BioAssay Kit which measures released adenylate kinase (AK) was purchased from Cambrex Bio Science Rockland Inc. (Rockland, ME, U.S.A.). The aCella-Tox™ kit which measures the biomarker glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was obtained from Bachem (King of Prussia, PA, U.S.A). The CytoTox-Glo™ Cytotoxicity Assay which measures a distinct proteolytic biomarker profile was developed at Promega Corporation (Madison, WI, U.S.A.)

Cell Culture

Human embryonic kidney cells (HEK293) and Jurkat cells (clone E6-1) were purchased from the American Type Culture Collection (ATCC, Manassas, VA, U.S.A.). HEK 293 cells were cultured in DMEM medium (Invitrogen, CA) and Jurkat cells were maintained as a suspension culture in RPMI 1640 medium (Sigma-Aldrich, MO) supplemented with 10% fetal bovine serum (FBS). Human renal mesangial cells, obtained from adult kidney tissue, were kindly provided by Dr. Jeffrey Kopp (NIDDK/NIH, Bethesda, MD) and cultured in RPMI 1640 medium. The medium was supplemented with 10% FBS (Hyclone Laboratories, Logan, UT, U.S.A.), 50U/ml penicillin and 50μg/ml streptomycin (Invitrogen, CA, U.S.A.). Each cell line was maintained at 37°C under a humidified atmosphere and 5% CO2.

Cell viability and Enzyme Release Assay Comparison

To investigate the relative sensitivity, linearity, and signal stability of three different luminescent cytotoxicity assays, Jurkat cells were suspended in fresh RPMI 1640 with 10% FBS at a density of 100,000 cells/ml. The pool was divided into two equal fractions. One fraction was treated by mild sonication to cause cytotoxicity (impaired membrane integrity). Cell membrane damage was initiated using a Branson Microtip Sonifier (Model 450) using a 30% duty cycle for 20 pulses at a microtip limit of 4 (of 10). After treatment, less than 5% of the cells remained unstained and viable by trypan blue exclusion and microscopic examination. The other fraction was left untreated to represent the experimental contribution of normal viable cells. The two pools were subsequently blended in various proportions to represent viabilities from 100–0% (conversely, cytotoxicities from 0–100%). Each sample blend was added to plates in replicate 100μl volumes (10,000 cell equivalents/well). Toxi-Light® BioAssay Kit, aCella-Tox™, and CytoTox-Glo™ Cytotoxicity Assay reagents were prepared and added as directed by supporting technical literature from the manufacturers. After 30 seconds of mixing at 500 RPM on an orbital shaker (Ika-Schüttler MTS-4 S2, Wilmingon, NC), the plates were incubated at 23°C in a Me’Cour (Waltham, MA, U.S.A) water-jacketed plate incubator and luminescence measured using a BMG Labtech FluoStar™ (Offenburg, Germany) 15 and 60 minutes after the reagents were added.

Miniaturization of the Luminescent Protease Release Assay

The CytoTox-Glo™ Cytotoxicity Assay is a luminescent assay that measures the activity of protease(s) which are released as a result of cell membrane damage. The qHTS protocol and assay volumes are summarized in Figure 2. Briefly, HEK 293 cells were dispensed at 5μl/well (1,500 cells) and renal mesangial cells at 5μl/well (1,000 cells) in tissue culture treated 1536-well white assay plates (Greiner Bio-One North America, NC, U.S.A.) using a Flying Reagent Dispenser (FRD) (Aurora Discovery, CA, U.S.A.). Cells were incubated at 37°C overnight to allow for cell attachment, followed by addition of compounds via pin tool (Kalypsys). After compound addition, plates were incubated for 6hr at 37°C. At the end of the incubation period, 5μl/well of protease assay reagent was added, and the assay plates were incubated at room temperature (RT) for 10min. Luminescence from the assay plates was measured using a ViewLux plate reader (PerkinElmer; Shelton, CT, U.S.A.).

Figure 2.

Figure 2

Open in a new tab

qHTS protocol for luminescent protease release assay. HEK 293 cells were dispensed at 5μl/well (1,500 cells) and renal mesangial cells at 5μl/well (1,000 cells) in 1536-well white assay plates. Cells were incubated at 37°C overnight to allow for cell attachment, followed by addition of 23nL of compounds or controls via a pin tool. After compound addition, plates were incubated for 6hr at 37°C. At the end of the incubation period, 5μl/well of protease assay reagent was added, and the assay plates were incubated at room temperature for 10min. Luminescence from the assay plates was measured using a ViewLux plate reader.

Quantitative High-Throughput Screening (qHTS)

Compound formatting and qHTS were performed as described previously (Inglese et al., 2006; Xia et al., 2008). Two positive control compounds were dispensed on each plate: digitonin and tetraoctylammonium bromide in a concentration series from 0.5 nM to 100 μM in DMSO into columns 1 and 2, respectively. For additional controls, tetraoctylammonium bromide at 100 μM in DMSO was added to column 3 as a positive control and DMSO only was dispensed into column 4 as a negative control. The final concentration of DMSO in the assay was 0.45% (or 0.90% in wells that were double dosed).

As showed in Figure 2, 23nl of the test compounds in a DMSO solution were transferred via a pin tool, resulting in final concentrations of 0.59nM to 46μM of compound, and 0.45% DMSO. To achieve the highest final compound concentration of 92μM (DMSO concentration 0.9%), 23nl was transferred twice from the highest concentration of source plate into the assay plate; control plates with DMSO transferred twice were also included for comparison.

Data Analysis and Curve Fitting

Analysis of compound concentration-responses was performed as previously described (Inglese et al., 2006; Xia et al., 2008). Data for each titration point were first normalized relative to the tetraoctylammonium bromide control (100 μM, 100%) and DMSO-only wells (basal, 0%), and then corrected by applying a pattern correction algorithm using compound-free control plates (i.e., DMSO-only plates) at the beginning and end of the compound plate stack. Concentration-response titration points for each compound were fitted to the Hill equation (Hill, 1910) yielding concentrations of half-maximal activity (EC50) and maximal response (efficacy) values. Compounds were designated as Class 1–4 according to the type of concentration-response curve observed (Inglese et al., 2006; Xia et al., 2008). Briefly, Class 1.1 compounds have full efficacy (>80%) and complete curves. Class 1.2 compounds have partial efficacy (30–80%) and complete curve. Class 2.1 compounds have full efficacy (>80%) and partial curve. Class 2.2 compounds have partial efficacy (30–80%) and partial curve. Class 3 compounds display significant activity only at the highest concentration tested, and class 4 compounds show no concentration response. Compounds with Class 1.1, 1.2, 2.1, or 2.2 curves were selected for follow-up analyses as they represented high confidence data with good quality of fit (R2>0.9).

Results

Comparison of Bioluminescent Enzyme Release Assays

The linearity of the signal response achieved from the 0–100% cytotoxicity model was generally acceptable for the measurement of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and adenylate kinase (AK) during the first reading (15 minutes), but neither assay generated the signal window and linearity produced by the protease release assay (as indicated below). The most dramatic and striking differences among the three assays were evident after 60 minutes of reagent and sample contact (Figure 3A and 3B). At this time point, the GAPDH assay produced a proportional linear response only in the presence of modest changes in viability (10% or less cytotoxicity) and the AK assay demonstrated a substantial increase in background which effectively reduced its signal window. Conversely, the protease release assay produced a linear relationship (r2 = 1.0 and 0.9996) at both time points throughout the extreme degrees of cytotoxicity (0 – 100%) while maintaining a stable luminescent signal.

Figure 3.

Figure 3

Open in a new tab

The signal response of luminescent assay chemistries after 15 (A) and 60 min (B) of reagent and sample contact. Viable and cytotoxic fractions of Jurkat cells were blended in various proportions to represent 0–100% viability. Each well contained 10,000 cell equivalents of viable or cytotoxic cells. Reagents were prepared as directed by manufacturer instructions and added to assay wells. Luminescence was measured after 15 and 60min of reagent and sample contact at RT using a BMG Labtech FluoStar™. Linear best-fit correlations were calculated using GraphPad Prism.

Validation and Miniaturization of the Protease Release assay

The protease release assay was miniaturized into a 1536-well plate format with a final assay volume of 5μl per well. The mild detergents digitonin (Schulz, 1990) and tetraoctylammonium bromide (Kumar et al., 2003) were used as positive control compounds which facilitate damage to cell membranes. Both HEK 293 and renal mesangial cells were treated with these two compounds for 1hr. The leakage of proteases from cells was measured by addition of the protease substrate mixture. Figure 4 demonstrates that both digitonin and tetraoctylammonium bromide caused protease release from the damaged cells. The signal-to-background ratios for digitonin were 15.9 fold in HEK293 cells and 17.2 fold in renal mesangial cells. Similar signal-to-background ratios were found for the tetraoctylammonium bromide treatment groups, with 14.0 fold in HEK293 cells and 17.1 fold in renal mesangial cells. The EC50s of digitonin were 30μM in HEK293 cells and 18.8μM in renal mesangial cells, respectively. The EC50s of tetraoctylammonium bromide were 34.7μM in HEK293 cells and 14.4μvM in renal mesangial cells, respectively.

Figure 4.

Figure 4

Open in a new tab

Effect of digitonin and tetraoctylammonium bromide on protease activity. HEK293 and renal mesangial cells were incubated for 1 hr in the presence or absence of digitonin (100 μM) and tetraoctylammonium bromide (100 μM). After 1 hr incubation the protease substrate mixture was added to the assay plates and luminescence was measured using a ViewLux plate reader. Data represent mean ± SD from quadruplicates and are expressed as luminescence (RLU).

To determine the optimal time for capturing cell leakage of enzymatic contents caused by the treatment of compounds, the HEK 293 and renal mesangial cells were incubated with various concentrations of digitonin and tetraoctylammonium bromide for 1, 6 and 16 hr (Figure 5). With HEK293 cells (Figure 5A and B) the signal-to-background ratios for both digitonin and tetraoctylammonium bromide were the highest after 1hr of incubation (15.9 fold and 11.5 fold), followed by 6hr incubation (8.0 fold and 7.9 fold), and the lowest with 16hr incubation (3.3 fold and 5.5 fold). Similar effects were observed in mesangial cells (Figure 5C and D) with 1 and 6hr of treatment. However, both digitonin and tetraoctylammonium bromide no longer resulted in measurable protease activity in the medium after 16hr of treatment in mesangial cells (Figure 5C and D). These data suggest that both digitonin and tetraoctylammonium bromide have fast onset activity consistent with primary necrosis. Therefore, a 6hr incubation time was chosen for the larger compound screening effort in order to identify compounds that cause cell death by either necrosis or apoptosis. The activity of released proteases gradually decreased over time and disappeared 16 hr after the complete cell lysis/death.

Figure 5.

Figure 5

Open in a new tab

Time course of protease release in HEK 293 and renal mesangial cells. HEK293 cells were treated with digitonin (A) and tetraoctylammonium bromide (B) for 1hr, 6hr and 16hr. Renal mesangial cells were treated with digitonin (C) and tetraoctylammonium bromide (D) for 1hr, 6hr and 16hr. At the end of various time points, protease substrate mixture was added into the plates and luminescence was measured using a ViewLux plate reader. Data are from a single experiment performed in quadruplicate, representative of several experiments.

Quantitative High Throughput Screening (qHTS)

A set of 1408 known compounds provided by NTP was used to evaluate the luminescent protease release assay. To ensure data quality throughout the qHTS process, the concentration titration of tetraoctylammonium bromide, a positive control compound, was carried out in each assay plate with the HEK293 and renal mesangial cells. The concentration response curves of tetraoctylammonium bromide in 36 plates screened in both cell lines reproduced well (Figure 6), with an average EC50 value of 30.0 ± 7.3μM in HEK293 cells and 25.5 ± 4.3μM in renal mesangial cells, respectively. The average signal-to-background ratio for tetraoctylammonium bromide was 7 ± 2 and average Z′ factor was 0.6 ± 0.2 from 18 plates in HEK293 cells, while the signal-to-background ratio was 15 ± 2 and average Z′ factor was 0.87 ± 0.03 from 18 plates in renal mesangial cells (Table 1). Figure 7 shows data from a DMSO vehicle control plate (Figure 7A) and a high compound concentration (46 μM) plate (Figure 7B) from the screening in renal mesangial cells. All these data indicate that the protease release assay in the qHTS formatis robust and suitable for identifying compounds which cause membrane leakage and protease release from the cells.

Figure 6.

Figure 6

Open in a new tab

A 3D scatter plot of the intra-plate concentration response titration curves from 36 plates in the screen. In each plate, tetraoctylammonium bromide (TOAB) was used as a positive compound in HEK293 cells (blue) and renal mesengial cells (light blue). The TOAB concentration response curves were derived by fitting to the Hill equation, resulting in EC50s of 30.0 ± 7.3μM in HEK293 cells and 25.5 ± 4.3μM in renal mesangial cells, respectively.

Table 1.

Assay performance as measured by plate QC statistics

Cell line S/B Z′ factora No. of plates
HEK293 6.47±1.73 0.60±0.19 18
Mesangial cell 15.2±1.83 0.87±0.03 18
Open in a new tab
a

Z′ factor = 1−[3*(SD of positive control + SD of basal)/(median of positive control − median of basal)]

Figure 7.

Figure 7

Open in a new tab

Screen performance in 1536-well plate format. Scatter plot of a DMSO plate (A) and a compound (46 μM) plate (B) in protease release assay in renal mesengial cells. The signal-to-background ratio was 18.2 and the Z′ factor was 0.8 in the DMSO plate, while the signal to background ratio was 16.9 and the Z′ factor was 0.9 in the compound plate.

In this qHTS campaign, each compound from the NTP library was screened at fourteen concentrations (from 0.59nM to 92μM). The concentration response curves of all the compounds have been classified into four major classes (1–4) as described in the Methods section. With HEK 293 cells, 11 compounds were classified as class 1.1, 1.2, 2.1 or 2.2 (Table 2), while 16 compounds with class 1.1, 1.2, 2.1 or 2.2 curves were found in renal mesangial cells (Table 3). Of these 11 compounds positive in HEK 293 cells, colchicine, which has two copies in this library (Xia et al., 2008) displayed the same EC50 of 0.1μM and the same curve class of 2.2. The other 54 compounds with two copies in this library were negative with curve class of 4 in HEK 293 cells. These results indicate that compound potencies and efficacies generated immediately after the qHTS screening are intra-experimentally reproducible, indicating the reliability of our qHTS assay in 1536-well plate format.

Table 2.

Potencies (EC50) for compounds from the NTP library that were active in HEK 293 cells following 6hr of exposure.

Sample ID Compound Name EC50 (μM) Curve class
AB07944366-01 tetra-N-Octylammonium bromide 25.1 1.1
AB07944329-01 Kepone 31.6 1.1
AB08001301-01 a-Solanine 10.0 1.2
AB08548268-01 Rhothane (TDE) 50.1 2.1
AB08582918-01 Digitonin 39.8 2.1
AB07985386-01 Captan 50.1 2.1
AB07944367-01 Cetylpyridinium bromide 63.1 2.1
AB07990332-01 Colchicine* 0.1 2.2
AB08001092-01 Pentaerythritol triacrylate 39.8 2.2
AB07944343-01 Colchicine* 0.1 2.2
Open in a new tab
*

Colchicine has been tested in the duplicates.

Table 3.

Potencies (EC50) for compounds from the NTP library that were active in mesangial cells following 6hr of exposure.

Sample ID Compound Name EC50 (μM) Curve class
AB07944716-01 Zinc pyrithione 25.1 1.1
AB07944366-01 tetra-N-Octylammonium bromide 25.1 1.1
AB07944331-01 Domiphen bromide 10.0 1.1
AB08080834-01 Tetramethylthiuram disulfide 12.6 1.2
AB08582918-01 Digitonin 31.6 2.1
AB07944404-01 Captan 90-concentrate (solid) 50.1 2.1
AB08001301-01 a-Solanine 39.8 2.1
AB07944329-01 Kepone 39.8 2.1
AB08002816-01 Malachite green oxalate 12.6 2.1
AB08007268-01 Benzethonium chloride 39.8 2.1
AB08548268-01 Rhothane (TDE) 39.8 2.1
AB07944367-01 Cetylpyridinium bromide 50.1 2.1
AB07985386-01 Captan 31.6 2.2
AB07990339-01 Ziram 50.1 2.2
AB07944034-01 p -n–Nonylphenol 50.1 2.2
AB08001092-01 Pentaerythritol triacrylate 31.6 2.2
Open in a new tab

Compounds identified from Protease Release Assay

Cytotoxicity caused by a 6 hr-exposure to the NTP library was evaluated in HEK293 and mesangial cells using the luminescent protease release assay. Among the 16 compounds found positive in HEK 293 (Table 2) and/or renal mesangial (Table 3) cells, there are two known cell membrane disrupters including a-solanine (Erik et al., 1996) and zinc pyrithione (Gibson et al., 1985). The results obtained from both cell lines show that six of the 16 positive compounds are detergents including tetra-N-Octylammonium bromide (Kumar et al., 2003), digitonin (Schulz, 1990), cetylpyridinium bromide (Blonder et al., 1976), p-n-nonylphenol (Li, 2007), benzethonium chloride (Li, 2007), and domiphen bromide (Fukushima et al., 2006). The destruction of cell membranes by these detergents leads to protease release and cell death. There were 15 compounds that induced protease release in renal mesangial cells, while 9 compounds were found positive in HEK 293 cells, suggesting that renal mesengial cells were more sensitive to compound-induced membrane disruption. However, colchicine, a known microtubule inhibitor (Modriansky et al, 2005), induced protease release in HEK 293 cells, but not in the renal mesangial cells, suggesting that cell type may play a role in compound sensitivity.

Discussion

Useful enzymatic surrogates of cytotoxicity have common attributes. First, the marker must be ubiquitous and conserved amongst the diverse array of cell lineages and phenotypes employed in common experimental culture. Second, the marker activity must not be positively or negatively modulated by any stimulus other than cytotoxicity. Third, the marker should retain sufficient activity (half-life) in culture medium to be measured in a reasonable time frame. Lastly, the marker should be detectible by robust and sensitive assay, free from undo chemical interferences. Three markers currently meet these criteria and are detected with luminescence chemistry.

Luminescent ATP-cycling chemistries used in cytotoxicity detection offer improved sensitivity and dynamic range while avoiding artifactual data from fluorescence interference. AK and GAPDH activities are harnessed to generate ATP in a manner that is proportional to the number of dead cells. The ATP produced by these non-lytic assays is detected in a coupled, simultaneous reaction with luciferase reagents (Squirrel and Murphy, 1997). Although these assay systems are initially sensitive and linear and represent a considerable improvement over colorimetric and fluorescent techniques, they suffer from high background interferences (serum) and unsteady light output kinetics. These limitations lead to loss of linearity across the cytotoxic spectrum, limiting the quantitative value of the assay and hindering interpretation of the data.

We developed an assay system to measure a recently discovered novel protease surrogate of cytotoxicity. This proteolytic activity, released into culture medium upon cell death, can be measured with a luminogenic, cell-impermeant peptide substrate. This technology utilizes an amino-derivatized luciferin and a luciferase-based detection reagent that generates luminescence proportional to proteolytic activity (O’Brien et al., 2005). This cell-based protease-release assay measures cell membrane integrity homogenously, which has only one step of reagent addition, with luminescent readout. Following exposure of toxic compounds the intracellular proteases release from cell into the culture medium. These released proteases cleavage the luminogenic substrate containing the AAF sequence, which releases a luciferase substrate (aminoluciferin), resulting in the luciferase reaction and the production of “glow-type” signal. Thus, the compound cytotoxic effect can be measured as the increased luciferase activity (Figure 1). This coupled chemistry generates large dynamic ranges with excellent linearity which allows for unprecedented sensitivity in high density formats. Although the dead-cell protease also has a finite half-life of enzymatic activity, it is vastly better than AK and GAPDH and lasts an average of 100–200% longer (Niles et al., 2007). Ultimately, this signal stability contributes to greater flexibility in measurement times and mitigates underestimation of profound cytotoxicity.

When compared to a traditional LDH release assay, the protease release assay can identify membrane disrupters that also inhibit or otherwise interfere with LDH enzyme activity. For example, kepone was found to cause protease release in this protease release assay in a concentration-dependent manner, with EC50s of 31.6μM and 39.8μM in HEK293 and renal mesangial cells, respectively. However, kepone showed no or poor concentration responses in the LDH release assay (data not shown). This negative result is likely the result of kepone inhibiting the LDH enzyme activity (Anderson and Noble, 1977). In both HEK 293 and renal mesangial assay models, the signal-to-background ratio was higher with the luminescent protease release assay (6–15 fold) than with the LDH release assay (2–3 fold). In addition, the extra stop addition step in the LDH release assay was eliminated in the protease release assay, making the latter assay more robust and appealing compared to the LDH release assay. On the other hand, the number of active compounds identified in the luminescent protease release assay (0.7–1.2%) was smaller when compared to the LDH assay (1.5–2.5%). Although not rigorously examined, the relative disparity in these results may be partially explained by false positives in the fluorescence assay or luminescence quenching by intensely colored compounds from the collection. Nevertheless, the luminescent protease release assay offers a novel readout for compound toxicity profiling, and should prove a useful tool to assess compound toxicity in miniaturized formats.

The kinetics of cell death during compound exposure can vary greatly. Therefore, choosing the time point for assaying cytotoxicity is challenging for both the protease and LDH release assays. Assaying for the biomarkers too early or too late can result in false negative findings. For instance, cytotoxicity as a result of apoptosis may require several hours or days, so premature time point testing would not be predictive of cytotoxicity (Riss and Moravec, 2004). Conversely, time point testing long after biomarker activity has degraded may greatly underestimate or miss cytotoxicity. Most detergents produce catastrophic cell lysis that causes protease and LDH release within seconds to 1hr. In the present study, we found that detergents produced measurable cytotoxicity after 1hr of treatment and reached the peak at 6 hr (data not shown) but time-dependent degradation of biomarker activities caused them to score as negative after 16hr of exposure. It is thus critical to measure the time course of the protease or LDH release assays for each compound if detailed study of the compound is required. It is precisely for this reason that we have now developed a method to measure the proteolytic contents of previously viable cells (Niles et al., 2007). The protease release assay contains an optional lysis reagent that can be used to also produce a same-well viability measure. This is possible because the chemistry produces a stable luminescent signal that is proportional to proteolytic activity in the medium. Therefore, an indirect viability measure can be obtained by first measuring experimental cytotoxicity, then measuring total protease activity after the addition of a lytic agent. Subtraction of experimental cytotoxicity values from the total values yields the proportion of cells that were viable prior to optional lysis.

The qHTS platform plays an important role in toxicity screening because response must be measured over a broad compound concentration range (0.59nM – 92μM) to fully characterize the biological activity. It is well known that the cytotoxicity of all compounds is concentration-dependent and that concentration-response curves are frequently biphasic (bell shape). For example, malachite green oxalate started to induce protease release at 9.2μM, plateau at 20.6μM, and the signal started to decrease at 46 μM and disappeared at 92μM in renal mesangial cells. These types of compounds could be missed if only screened at a single concentration. Therefore, using qHTS platform in toxicity screening provides more accurate assessment of compound toxicological activity and is suitable for the profiling of compound toxicity.

In summary, we have described the development and validation of a cell-based protease release assay for screening compound toxicity in 1536-well plate format. This assay has been miniaturized and is suitable for larger scale library screening. The data from qHTS allowed us to quickly and efficiently evaluate compound toxicities including potency and efficacy. From NTP library screening, we found 16 toxic compounds including a couple of known membrane disrupters and 6 detergents. The luminescent protease release assay combined with the qHTS platform approach and other cytotoxicity assays can effectively identify toxicological compounds.

Acknowledgments

We gratefully acknowledge Adam Yasgar and Paul Shinn for the compound management. We also thank Dr. Raymond Tice and Kristine Witt for their critical comments for the project. This research was supported by the Intramural Research Programs of the National Toxicology Program, National Institute of Environmental Health Sciences and the National Human Genome Research Institute, National Institutes of Health, and the NIH Roadmap for Medical Research Molecular Libraries Program.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Anderson BM, Noble C., Jr In vitro inhibition of lactate dehydrogenases by kepone. Journal of Agricultural and Food Chemistry. 1977;25:28–31. doi: 10.1021/jf60209a004. [DOI] [PubMed] [Google Scholar]
  2. Blonder E, Klibansky C, DE Vries A. Effects of detergents and choline-containing phospholipids on human spleen glucocerebrosidase. Biochimica et Biophysica Acta. 1976;431:45–53. doi: 10.1016/0005-2760(76)90258-7. [DOI] [PubMed] [Google Scholar]
  3. Boisclar M, Iwata K. HTS using cell-based assays for drug discovery. Am Assoc Cancer Res Educ Book. 2005 April 2005;:363–368. [Google Scholar]
  4. Corey MJ, Kinders R, Brown L, Vessella R. A very sensitive coupled luminescent assay for cytotoxicity and complement-mediated lysis. Journal of Immunological Methods. 1997;207:43–51. doi: 10.1016/s0022-1759(97)00098-7. [DOI] [PubMed] [Google Scholar]
  5. Digan ME, Pou C, Niu H, Zhang J. Evaluation of division-arrested cells for cell-based high throughput screening profiling. Journal of Biomolecular Screening. 2005;10:615–623. doi: 10.1177/1087057105276474. [DOI] [PubMed] [Google Scholar]
  6. Erik A, Keukens J, Vrije TD, Jansen LAM, Boer HD, Janssen M, Kroon AIPMD, Jongen WMF, Kruijff ED. Glycoalkaloids selectively permeabilize cholesterol containing biomembranes. Biochimica et Biophysica Acta. 1996;1279:243–250. doi: 10.1016/0005-2736(95)00253-7. [DOI] [PubMed] [Google Scholar]
  7. Fukushima M, Kikuchi A, Tatsumi K, Tanaka F. Separation of fulvic acid from soil extracts based on ion-pair formation with a cationic surfactant. Analytical Sciences. 2006;22:229–33. doi: 10.2116/analsci.22.229. [DOI] [PubMed] [Google Scholar]
  8. Hill AV. The possible effects of the aggregation of the molecules of haemoglobin on its dissociation curves. Journal of Physiol (London) 1910;40:4–7. [Google Scholar]
  9. Inglese J, Auld DS, Jadhav A, Johnson RL, Simeonov A, Yasgar A, Zheng W, Austin CP. Quantitative high-throughput screening: a titration-based approach that efficiently identifies biological activities in large chemical libraries. Proceedings of the National Academy of Sciences USA. 2006;103:11473–11478. doi: 10.1073/pnas.0604348103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Kumar S, Sharma D, Kabir-ud-Din Role of quaternary bromides in changing the solubilization site of n-Heptylamine in cationic micellar solutions. Journal of Surfactants and Detergents. 2003;6:339–344. [Google Scholar]
  11. Leist M, Jaattela M. Four deaths and a funeral: From caspases to alternative mechanisms. Nature Review. 2001;2:1–10. doi: 10.1038/35085008. [DOI] [PubMed] [Google Scholar]
  12. Li MH. Effects of nonionic and ionic surfactants on survival, oxidative stress, and cholinesterase activity of planarian. Chemosphere. 2007 doi: 10.1016/j.chemosphere.2007.08.032. in press. [DOI] [PubMed] [Google Scholar]
  13. Maffia AM, Kariv I, Oldenburg K. Miniaturization of a mammalian cell-based assay: luciferase reporter gene readout in a 3 microliter 1536-well plate. Journal of Biomolecular Screening. 1999;4:137–142. doi: 10.1177/108705719900400307. [DOI] [PubMed] [Google Scholar]
  14. Modriansky M, Dvorak Z. Microtubule disruptors and their interaction with biotransformation enzymes. Biomedical Papers of the Medical Faculty of the University Palacky Olomouc Czechosovakia. 2005;149:213–5. [PubMed] [Google Scholar]
  15. Niles A, Moravec R, Hesselbeth P, Scurria M, Daily W, Riss T. A homogeneous assay to measure live and dead cells in the same sample by detecting different protease markers. Analytical Biochemistry. 2007;366:197–206. doi: 10.1016/j.ab.2007.04.007. [DOI] [PubMed] [Google Scholar]
  16. Niles A, Moravec R, Scurria M, Daily W, Bernad L, McNamara B, Moraes A, Rashka K, Lange D, Riss T. MultiTox-fluor multiplex cytotoxicity assay technology. Cell Notes. 2006;15:11–15. [Google Scholar]
  17. Niles A, Worzela T, Scurria M, Daily W, Bernad L, Guthmiller P, McNamara B, Rashka K, Lange D, Riss T. Multiplexed viability, cytotoxicity and apoptosis assays for cell-based screening. Cell Notes. 2006;16:12–15. [Google Scholar]
  18. Niles A, Scurria M, Bernad L, McNamara B, Rashka K, Lange D, Guthmiller P, Worzella T, Riss T. Measure relative numbers of live and dead cells and normalize assay data to cell number. Cell Notes. 2007;18:15–20. [Google Scholar]
  19. O’Brien M, Daily W, Hesselberth E, Moravec R, Scurria M, Klaubert D, Robert Bulleit R, Wood K. Homogeneous, bioluminescent protease assays: caspase-3 as a model. Journal of Biomolecular Screening. 2005;10:137–148. doi: 10.1177/1087057104271865. [DOI] [PubMed] [Google Scholar]
  20. Olsson T, Gulliksson H, Palmeborn M, Bergstrom K, Thore A. Leakage of adenylate kinase from stored blood cells. Journal of Applied Biochemistry. 1983;5:437–445. [PubMed] [Google Scholar]
  21. Riss T, Moravec R. Use of multiple assay endpoints to investigate the effects of incubation time, dose of toxin, and plating density in cell-based cytotoxicity assays. Assay and Drug Development Technology. 2004;2:51–62. doi: 10.1089/154065804322966315. [DOI] [PubMed] [Google Scholar]
  22. Riss T, Moravec R. Introducing the CytoTox-ONE homogeneous membrane integrity assay. Cell Notes. 2002;4:6–9. [Google Scholar]
  23. Rose D, Lemmo T. Challenges in implementing high density formats for high throughput screening. Laboratory Automation News. 1997;2:12–19. [Google Scholar]
  24. Squirrel D, Murphy J. Rapid detection of very low number of micro-organisms using adenylate kinase as a cell marker. A practical guide to industrial uses of ATP luminescence in rapid microbiology. 1997:107–113. [Google Scholar]
  25. Schulz I. Permeabilizing cells: some methods and applications for the study of intracellular processes. Methods in Enzymology. 1990;192:280–300. doi: 10.1016/0076-6879(90)92077-q. [DOI] [PubMed] [Google Scholar]
  26. Xia M, Huang R, Witt KL, Southall N, Fostel J, Cho M, Jadhav A, Smith CS, Inglese J, Portier CJ, Tice RR, Austin CP. Compound cytotoxicity profiling using quantitative high-throughput screening. Environmental Health Perspectives. 2008;116:284–291. doi: 10.1289/ehp.10727. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES