Abstract
Metabolically active cells are able to convert the MTT [3-(4,5-dimethythiazol-2-yl)-2,5-diphenyltetrazolium bromide] dye to blue formazan. This is the basis of the MTT assay, which is among the most widely used screening methods to evaluate cell viability and proliferation. When testing the effects of cholesterol products on the viability of human pulmonary epithelial-like A549 cells using trypan blue staining (cell numbers) and the MTT assay, results were inconsistent. The MTT assay indicated greater than 50% loss of viability with exposure of cells to cholesterol, whereas there was no decrease in viability indicated by trypan blue exclusion and propidium iodide uptake. A similar decrease in MTT reduction was obtained upon cholesterol treatment in human lung microvascular endothelial cells (HLMVECs) and human coronary artery endothelial cells (HCAECs) without loss of viability. This suggested a direct interference of cholesterol with the assay. However, using a cell-free system, there was no decrease in the reduction of MTT by ascorbic acid during incubation with a similar concentration of cholesterol. Light microscopy revealed enhanced exocytosis of formazan granules in presence of cholesterol. Incubation with apolipoprotein A-1 decreased cholesterol-mediated inhibition of MTT assay. These studies indicate decreased MTT reduction as a result of enhanced exocytosis of formazan due to cholesterol. A careful validation of viability assay procedures is therefore suggested in experiments where cholesterol is a constituent, to avoid a potential bias in concluding results of cytotoxicity studies.
MTT [3-(4,5-dimethythiazol-2-yl)-2,5-diphenyltetrazolium bromide] reduction is a simple, sensitive and widely used colorimetric method for measuring in vitro cytotoxicity and cellular proliferation (Morgan 1998; Mosmann 1983; Loprevite et al. 2001; Zorilla, Balebona, and Morinigo 2001; Carvalho et al. 2002; Larsen, Gareis, and Frisvad 2002). The biochemical parameter measured is an enzymatic reduction of the tetrazolium ring of MTT to a water insoluble purple-to-dark blue formazan product (Liu et al. 1997). Although the cellular reactions involved in the reduction are not completely delineated, many mitochondrial as well as nonmitochondrial dehydrogenases and flavin oxidases are capable of reducing MTT (Mosmann 1983; Liu and Shubert 1997; Altman 1976; Burdon, Gill, and Rice-Evans 1993). Several reports exist wherein the MTT assay has been successfully validated by comparing other methods of viability estimation. The MTT assay has been found to correlate well with cell number within the limits of measurement of absorbance (Loveland et al. 1992; Smith et al. 1992). It has been a method of choice for drug screening in cancer research, various immunological investigations, and in the evaluation of in vitro cell compatibility with different biocompounds, metals and polymers (Loprevite et al. 2001; Carmichael et al. 1987; Styczynski et al. 2002; Wu et al. 2001; Iwata and Inoue 1993; Kreja and Seidel 2002; Ciapetti et al. 1993; Xu, Hoet, and Nemery 2002). The MTT assay also finds its application in the evaluation of metabolic activity of microorganisms (Muelas-Serrano, Nogal-Ruiz, and Gomez-Barrio 2000; Dias et al. 1999). The accurate assessment of cell viability is essential for identifying the cytotoxic potential of test drugs or compounds. For the MTT assay it is also important to exclude the possibility of MTT reduction by extracellular compounds as well. Antioxidants such as ascorbic acid, vitamin E,
Cholesterol also is an essential constituent (helper lipid) of lipid-based transfection reagents (Lenssen et al. 2002). The MTT-based colorimetric assay is frequently the method of choice for estimation of cell viability after genetic manipulations. The results of this study suggest use of extreme caution when interpreting results obtained using this method in the presence of cholesterol particularly in absence of appropriate controls.
MATERIALS AND METHODS
Materials
MTT [3-(4,5-dimethythiazol-2-yl)-2,5-diphenyltetrazolium bromide], ethanol, dimethylsulfoxide (DMSO), ascorbic acid, cholesterol acetate, 25-hydroxycholesterol, methyl-
Cell Culture
The human epithelial-like lung carcinoma cell line A549 was obtained from American Type Culture Collection (Rockville, MD). Cells were grown in 100-mm polystyrene tissue culture dishes (BD Falcon, Bedford, MA) in 10 ml of F-12K growth medium (Life Technologies, Rockville, MD) containing 10% fetal calf serum, penicillin (100 U/ml), and streptomycin (100 mg/ml) incubated at 37°C under a humidified atmosphere of air containing 5% CO2. Cells were routinely passaged by trypsinization and subcultured at an initial plating density of 0.5 million cells per plate. HLMVE cells and HCAE cells were purchased as frozen primary cultures from Clonetics (San Diego, CA). They were cultured in 10 ml endothelial cell basal medium (EBM-2) supplemented with vascular endothelial growth factor (VEGF), human fibroblast growth factor (hFGF), epithelial growth factor (EGF), hydrocortisone, ascorbic acid, insulin-like growth factor (IGF), GA1000 (gentamycin/amphotericin-B), and fetal bovine serum as per the manufacturer’s protocol in 100-mm tissue culture dishes. For treatment of A549 cells with cholesterol, stock solutions were prepared in ethanol. Equal volumes of ethanol were added to controls during such treatments. HLMVECs and HCAECs were treated with cholesterol/methyl-
MTT Assay
Reduction of water-soluble tetrazolium salt, 3-[4,5-dimethyl-thiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT), to the water-insoluble formazan was measured (Morgan 1998). Cells were plated in 96-well half-area tissue culture plates, and the medium was replaced by 100 μl of serum- and phenol red–free-DME/F12 1:1 mixture. Fifty microliters of MTT (4.0 mg/ml) in the serum- and phenol red–free DME/F12 (Dulbecco’s Modified Eagle’s medium and Ham’s F12 nutrient mixture, Sigma Chemicals, St. Louis, MO) mixture was added, and the plate was incubated for 4 h at 37°C. The purple formazan crystals thus formed were dissolved in 50 μl DMSO, and the optical densities of the wells on the plate were read at 540 nm using a plate reader.
Trypan Blue Exclusion
Trypan blue exclusion was performed by adding 25 μl of 0.1% trypan blue solution to 100 μl of cells suspended in phosphate-buffered saline (PBS). The cells that excluded the dye were counted on a hemocytometer as described earlier (Ahmad et al. 2001).
Propidium Iodide Staining
For the propidium iodide staining of nonviable cells, 1×106 cells were suspended in 1.0 ml PBS and propidium iodide (2 μg/ml, final concentration) was added. After incubating for 5 min on ice in the dark, flow cytometric analysis was performed as previously described (Ahmad et al. 2002).
Oxygen consumption was measured according to the method of Robinson and Cooper (1970), and as described before (Ahmad et al. 2001).
Statistical Analysis
All statistical calculations were performed with JMP software (SAS Institute, Cary, NC). Means were compared by one-way analysis of variance followed by two-tailed
RESULTS
Effect of Cholesterol on MTT Reduction
Using the MTT assay and conditions as described in Materials and Methods, the assay was linear in the range of 2000 to 15,000 A549 cells (Figure 1A ). Addition of cholesterol (0.1 μM, for 4 h at 37°C) to cell culture medium caused approximately 35% inhibition of color produced by formazan formation from MTT by A549 cells (Figure 1B ). More than a 50% decrease in absorbance was obtained when cholesterol concentration was increased from 1.0 μM to 15.0 μM. A further increase in cholesterol (30.0 to 60.0 μM) did not increase apparent interference as compared to controls where the solvent (ethanol) was used at similar dilutions and volumes. The experiment was also performed in absence of serum to overrule effect of serum components. Under these experimental conditions, the presence or absence of serum did not modify these findings.
To further evaluate whether this effect of cholesterol was affected by cell density, cells were plated at varying densities and 30 μM cholesterol was added (Figure 1C ). A greater than 50% decrease in measured absorbance values was obtained throughout the experiment irrespective of cell density or numbers.
Cholesterol/methyl-
Effect of Cholesterol on Cell Viability
We also assessed the viability of cells using the trypan blue exclusion method (Figure 2A ). Addition of cholesterol at concentrations from 15 to 240 μM for 4 h at 37°C did not produce a significant change in the number of trypan blue excluding cells as compared to the solvent control. Observation of cholesterol-exposed cells with phase-contrast microscopy also did not reveal any morphological change relative to nonexposed controls (Figure 2B and C ). Quantitation of nonviable cells using propidium iodide staining with detection by flow cytometry indicated no loss of viability (Figure 2D ).
There was no apparent cytotoxicity in the endothelial cells at these concentrations of cholesterol and time of incubation (data not shown).
Effect of 25-Hydroxycholesterol and Cholesteryl Acetate on MTT Reduction
25-Hydroxycholesterol (present as 0.2% impurity in cholesterol) did not cause a significant change in the absorbance as compared to the control (Figure 3). Treatment with cholesteryl acetate (Figure 3) also did not alter MTT reduction.
Effect of Cholesterol on Ascorbic Acid-Mediated MTT Reduction
Ascorbic acid is known to reduce the tetrazolium compound to formazan (Chakrabarti et al. 2000). We utilized this property of sodium ascorbate to investigate whether cholesterol interferes directly with formazan formation in solution. As seen in Table 2, no significant change in the color formation occurred upon addition of cholesterol to the MTT (4.0 mg/ml)/ascorbate (0.5 mM) system.
Effect of Cholesterol on Exocytosis of Formazan
Whether this effect was due to modulated exocytosis was followed by light microscopy (Figure 4) using A549 and human coronary artery endothelial cells (HCAECs). Light microscopic examination of cholesterol- (1.0 μM, 4 h) and MTT-treated cells revealed enhanced exocytosis of formazan from A549 and HCAECs at 30 min of incubation time.
Apolipoprotein A-1 mediates cholesterol efflux from cells (Fielding and Moser 1982). Pretreatment of A549 cells with apolipoprotein A-1 abolished the cholesterol-mediated decrease in MTT reduction at lower cholesterol concentrations (Figure 5). However, higher cholesterol concentrations were unaffected by the presence of apolipoprotein A-1.
Effect of Cholesterol on Oxygen Consumption
We then investigated whether cholesterol treatment affects metabolic processes like oxygen consumption. Measurement of oxygen consumption in cholesterol (1 μM)-treated A549 cells (4 h at 37°C) revealed a significant increase in cellular oxygen consumption (150 ± 0.14 nmol oxygen/mg protein/min in control versus 250 ± 0.31 nmol oxygen/mg protein/min in cholesterol-treated cells).
Oxygen consumption values with and without cholesterol treatment in endothelial cells were not significantly different (Table 3) under our conditions of treatment.
DISCUSSION
Cell proliferation and viability assays are indispensable for the establishment of optimum culture conditions, assessment of cytotoxicity, quantification of apoptosis, and for measurement of effects of growth factors, drug/therapeutic agents, and various environmental pollutants on cell growth. The MTT-based colorimetric assay is the simplest, most rapid, and one of the most sensitive methods used for estimation of cell viability. However, when examining the effects of biologic compounds that have not been described previously, it is essential to ensure that the effector does not interfere with the test itself.
We have used the MTT assay for assessment of viability of A549 cells (Uhlson et al. 2000), as reported previously (Kreja and Seidel 2002). In our hands the assay was linear in the range of 2000 to 15,000 A549 cells. Importantly, the MTT assay is sensitive to the metabolic state of cells. There could be considerable variations in the detection of range of linearity depending on the type of cell used (Morgan 1998).
When evaluating the viability of A549 cells in presence of cholesterol using the MTT assay, we encountered conflicting results when compared with results obtained using trypan blue exclusion. Studies using trypan blue exclusion did not indicate any loss of viability. Therefore, the effect of cholesterol on the assessment of viability of A549 cells using the MTT assay was explored further. Quantitation of nonviable cells using propidium iodide staining indicated no loss of viability, further suggesting an inhibition of MTT assay by cholesterol rather than a true loss of viability. 25-Hydroxycholesterol is usually present as 0.2% impurity in cholesterol. Our results indicate that the cholesterol-mediated decrease in MTT reduction is not due to the presence of this impurity. Treatment with cholesteryl acetate also did not alter MTT reduction, suggesting that the free hydroxyl group in cholesterol is required for this effect.
Cholesterol did not cause a direct interference of formazan formation in solution by ascorbic acid. This proved that, within the cells, cholesterol treatment leads to certain metabolic changes that result in decreased formazan formation. An initial decrease in formazan production from MTT in neuronal PC12 cells has been observed upon treatment with toxic
Regulation of the transport of MTT formazan-containing vesicles by free cholesterol previously has been reported in neuronal cell cultures (Liu, Peterson, and Schubert 1998). In absence of any exogenous lipids, the presence of free cholesterol within these cells affected the type of formazan crystal produced and enhanced their exocytosis. Importantly, in the present study using A549 cells, we obtained a decrease in color formation with MTT with or without serum or exogenous lipids (or lipoproteins also present in serum). Apolipoprotein A-1 mediated faster efflux of cholesterol, decreased the efflux of formazan along with cholesterol, hence allowing it to be reduced by the cells. This further confirmed enhanced exocytosis as one possible mechanism of inhibition of MTT reduction by cholesterol.
Cholesterol-mediated modulation of cellular oxygen consumption is relatively unexplored. Bjorndheden and Bondjers (1987) measured oxygen consumption in aortic segments of cholesterol-fed rabbits. Their results indicated enhanced oxygen consumption per DNA with an increased degree of atherosclerotic involvement in these segments. In this study in A549 cells, an increase in oxygen consumption was observed upon cholesterol treatment. The decreased MTT reduction with enhanced oxygen consumption indicated possible mitochondrial uncoupling, caused by cholesterol treatment in these cells. We were then interested in determining whether vascular endothelial cells like HLMVECs and HCAECs were similarly affected by cholesterol treatment. A similar mitochondrial uncoupling phenomenon, if present in these cells, could be pertinent to vascular dysfunction in atherosclerosis. Our results indicate that oxygen consumption of HLMVECs and HCAECs are not affected by cholesterol treatment. Khan et al. (2003) also have recently addressed the importance of plasma membrane cholesterol in the maintenance of the oxygen gradient in cells.
The MTT assay is a nonradioactive, rapid, and economical method. Moreover, it is superior to other formazan based assays (Parlo and Coleman 1984; Cory et al. 1991) in that (i) it does not include the use of an intermediate electron acceptor like phenazine methosulphate (PMS); (ii) it is not rapidly reduced by cofactors such as NADH and NADPH (Goodwin et al. 1995); and (iii) oxidant-generating conditions do not interfere with the test (Berridge et al. 1996; Ukeda et al. 2002). For these reasons it has been used extensively in determining the cytotoxicity and therapeutic potential of cancer therapy drugs. Our studies indicate that the method used to assay cellular viability needs to be carefully chosen. In particular, the MTT assay does not appear useful in studies of cholesterol metabolism or toxicity, nor in experiments where cholesterol is a constituent.
Footnotes
Figures and Tables
These studies were supported in part by grant R825702 from the Environmental Protection Agency and grant no. ES01448 from the National Institute of Environmental Health Sciences (NIEHS), NIH. The authors are grateful to Dr. Robert C. Murphy and Dr. Melissa Pulfer for critically reading this manuscript.