[CANCER RESEARCH 49, 4435-4440, August 15. 1989]

Effects of the pH Dependence of 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl-

tetrazolium Bromide-Formazan Absorption on ChemosensitivityDetermined by a Novel Tetrazolium-based Assay1

Jane A. Plumb,2 R. Milroy, and S. B. Kaye

C.R.C. Department of Medical Oncology, University of Glasgow, 1 Horselethill Road, Glasgow GI2 9LX, United Kingdom

ABSTRACT

The tetrazolium dye, 3-(4,5-dimethylthia/ol-2-yl)-2,5-diphi-iiyltclra-zolium bromide (MTT), is reduced by live but not dead cells, and thisreaction is used as the end point in a rapid drug-screening assay. It canalso be used for accurate determinations of drug sensitivity but only if aquantitative relationship is established between cell number and M l'I -forma/an production. We have shown that reduction of MTT to M l'I -

forma/an by cells is dependent on the amount of MTT in the incubationmedium. The concentration required to give maximal MTT-formazanproductiondiffers widely between cell lines. The absorption spectrum ofMTT-formazan varies with cell number and with pH. At a low celldensity or a high pH, the absorption maximum is at a wavelength of 560to 570 nm. However, at a high cell density or a low pH, there are twoabsorption maxima; one at 510 nm and a second at about 570 nm.Measurements of absorbance at 570 nm underestimate MTT-formazanproductionand, hence, cell number at high cell densities. This error canresult in a 10-fold underestimation of chemosensitivity. Addition of abufferat pH 10.5 to the solubili/ed MTT-formazan productcan overcomethe effects of both cell density and culture medium on the absorptionspectrum. Provided that sufficient MIT is used and the pH of the MTT-formazan productis controlled, dye reductioncan be used to estimate cellnumbers in a simple chemosensitivity assay the results of which agreewell with a commonly used clonogenic assay.

INTRODUCTION

Although there are a number of chemosensitivity assayssuitable for use in vitro, few are applicable to all cell types, andthere is no universally accepted assay. Interest in a tetrazoliumdye-based assay has been stimulated following the adoption ofsuch an assay by the National Cancer Institute for use in itsdrug-screening program (1, 2). The assay relies on the abilityof live but not dead cells to reduce a water-soluble yellow dye,MTT,1 to a water-insoluble purple formazan product. Since the

substrate and product absorb at very different wavelengths, nowashing steps are required after removal of the aqueous mediumprior to solubilization of the MTT-formazan product. This is aclear advantage over those assays which use isotope incorporation as an end point, particularly when dealing with non-adherent cell lines. As designed for drug screening, the assayrequires minimal manipulation and is easily automated (1). Itis thus ideally suited to a program that involves a wide rangeof cell types and thousands of potential anticancer agents.However, when the assay is used to determine drug sensitivity,a number of shortcomings are apparent. MTT-formazan production is used as an estimate of surviving cell number followingexposure to a cytotoxic drug. Many attempts to use the assayhave failed to demonstrate a linear relationship between Mil -

Received 2/21/89; accepted 5/17/89.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

' This work was supported by a grant from the British Cancer Research

Campaign.2To whom requests for reprints should be addressed.3The abbreviations used are: MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphen-

yltetrazolium bromide; DMSO, dimethyl sulfoxide: DMEM, Dulbecco's modifiaiEagle's medium; PBS, phosphate-buffered saline; ID»,50% inhibitory dose.

formazan production and cell number at high cell densities (1,3, 4). This is particularly important, since the highest cellnumbers occur in the control, untreated wells, and these areused to determine the parameters of sensitivity to the cytotoxicdrug.

We have now established conditions under which MTT reduction can be used quantitatively to determine surviving cellnumbers in a chemosensitivity assay that can be applied to bothadherent and nonadherent cell lines. This assay has been compared with an alternative and widely used drug sensitivity assay.

MATERIALS AND METHODS

Cell Lines

Four non-small cell lung cancer cell lines were used: A549, CALU,and SK-MES obtained from the American Type Tissue Collection; andL-DAN, a squamous lung cancer cell line established in our ownlaboratory. The doubling times of the lines were about 24 h. The othertwo adherent cell lines were MCF-7, a breast cancer line obtained fromDr. K. Cowan, National Cancer Institute, and G-UVW, a glioma cellline established in our own department. The doubling times were about24 and 60 h, respectively. Two nonadherent cell lines were used. Thesewere two small cell lung cancer lines: GLC4, obtained from Dr. E.deVries, Department of Clinical Oncology, University of Groningen,The Netherlands; and NCI-HI87, obtained from Dr. D. Carney (5).The doubling times were about 24 and 48 h, respectively.

Cells were maintained in a mixture of Hani's FIO and DMEM

(50:50; Gibco, Paisley, Scotland) supplemented with glutamine (2 HIM)and fetal calf serum (10%, v/v) except for MCF-7, GLC4, and NCIH69.These lines were maintained in RPMI 1640 (Northumbria Biologicals,Cramlington, Northumberland, England) supplemented with glutamine(2 HIM)and fetal calf serum (10%, v/v).

Determination of Absorption Spectra

MTT and MTT-formazan (Sigma Chemical Company, Poole, Dorset, England) were dissolved in DMSO, and the absorbance was scannedbetween wavelengths of 350 and 700 nm. MTT-formazan produced bycells was obtained by incubation of cells in the presence of MTT for 4h. MTT was removed and the formazan crystals dissolved in DMSO.

Determination of the Optimal MTT Concentration

Cells were plated out in 200 n\ of medium at a concentration ofeither IO4 cells per flat-bottomed well or 5 x 10" cells per round-bottomed well (GLC4 and NCI-HI87) in 96-well microtiter plates(Linbro; Flow Laboratories, Irvine, Scotland). The first and last row of8 wells contained medium only. Plates were incubated for 24 h at 37*C

in an atmosphere of 2% CO2 in air. MTT (0 to 5 mg/ml) was dissolvedin PBS, and 50 ¡Awere added to each well (8 wells per concentration).Plates were wrapped in aluminum foil and incubated for a further 4 h.Medium and MTT were then removed from the wells, and the formazancrystals were dissolved in 200 n\ of DMSO. In some experiments, 25/il of Sorensen's glycine buffer (0.1 M glycine plus 0.1 M NaCl equili

brated to pH 10.5 with 0.1 N NaOH) were added at this stage. Platescontaining nonadherent cells were centrifugea at 200 x g for 5 minprior to removal of the medium. The absorbance was recorded in anenzyme-linked immunosorbent assay plate reader (Model 2550 enzymeimmunoassay plate reader; Bio-Rad Laboratories, Ltd., Watford, Eng-

4435

Research. on October 4, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

TETRAZOLIUM-BASED CHEMOSENSITIVITY ASSAY

Formazan Sigma

CDMSO)

Formazanproducedby cells

(DMSO)

300 500

Wavelength (nm)

Fig. 1. Absorption spectra of MTT and commercially prepared MTT-forma-zan dissolved in DMSO. Also shown is the spectrum of MTT-formazan producedby incubation of L-DAN cells (4000 per well) with MTT (5 mg/ml) for 4 h.

land) at a wavelength of 570 nm. The first and last rows of 8 wellswhich contained medium and PBS (SO/il) only were used to blank theplate reader.

Relationship between MTT-Formazan Production and Cell Number

Cells were diluted to a concentration of 2.5 x IO5 cells per ml.

Various volumes of this cell suspension (20 to 200 /il) were plated outin duplicate 96-well microtiter plates (8 wells per volume). Plates wereincubated for at least 24 h at 37°Cin an atmosphere of 2% CO2 in air.

To one plate, 50 n\ of MTT in PBS (concentration depending on cellline) were added, and the plate was processed as described above. Inthis case, the first row of 8 wells, which contained medium and MTT(50 fil) but no cells, was used to blank the plate reader. The mediumwas removed from the second plate, and 50 p\ of trypsin/EDTA (0.25%/1 HIM)were added to each well. After incubation for a few minutes, thecells from 4 replicate wells were pooled and counted with an electroniccounter (Coulter Electronics, Luton, England).

Chemosensitivity Measurements

Microtiter Assay. Adherent cells were plated out at a concentrationof 10' per well in 200 n\ of medium and allowed to attach and grow for

2 to 3 days. The medium was then removed from the wells and replacedwith either 200 n\ of fresh medium or 200 /¿Iof medium containingdrug. Cells were incubated with drug for 24 h after which the mediumwas replaced by drug-free medium. The medium was replaced at 24 h

intervals for a further 2 days. On the fourth day following drug addition,the cells were fed with 200 n\ of fresh medium containing 4-(2-hydrox-yethyl)-l-piperazineethanesulfonic acid buffer (10 IHM,pH 7.4). MTT(50 n\ of a predetermined concentration) was added to the wells, andthe plates were incubated and processed as described previously. Wellscontaining medium and MTT only were used to blank the plate reader.

For the nonadherent cell lines, exponentially growing cells wereplated out at a concentration of 5 x IO3 cells per well in 100 n\ of

culture medium. Drug was added immediately in a volume of 100 /il attwice the final concentration. Plates were exposed to drug for 24 h andfed and processed as above.

Clonogenic Assay. Cells were plated out at a concentration of 7.8 xIO4cells per 25-cm2 flask (Falcon; Becton Dickinson, Oxford, England)

and allowed to attach and grow for 2 to 3 days. Flasks were then fedwith 5 ml of fresh medium or medium containing drug and incubatedfor 24 h. The medium was then removed, and the cells were washedwith 5 ml of trypsin/EDTA (0.25%/1 HIM)and incubated for a fewminutes at 37'C. Cells were resuspended in 10 ml of medium, and the

cells in the control, untreated, flasks were counted. They were dilutedto 100 cells per ml, and 5 ml were plated out into each of 4 Petri dishes(60-cm Nunclon; Gibco, Paisley, Scotland). Cells from the drug-treatedflasks were diluted according to the control flask and plated out asabove. The dishes were incubated at 37°Cin an atmosphere of 2% CO2

in air for 10 days. The medium was then removed, and the clones werewashed with PBS, fixed in methanol, and stained with crystal violet.Colonies of greater than 50 cells were counted.

RESULTS

Spectral Analysis. The absorption spectra of MTT and MTT-formazan are shown in Fig. 1. Also shown is the absorptionspectrum for MTT-formazan produced by incubation of cellsin the presence of MTT. For MTT, there is a single absorptionmaximum at a wavelength of 410 nm. MTT-formazan exhibitstwo absorption maxima at wavelengths of about 510 and 570nm (see also Fig. 5). However, the MTT-formazan producedby incubation of cells with MTT demonstrates a single absorption maximum at a wavelength of about 560 nm.

Effect of MTT Concentration on MTT-Formazan Production.Fig. 2 shows the amount of MTT-formazan produced when thecell line L-DAN is incubated with various concentrations ofMTT. MTT-formazan production increases with increasingMTT concentration up to about 4.5 mg/ml. This effect wasobserved regardless of whether cells were grown in F10/DMEMor in RPMI 1640 medium. A similar relationship is also apparent for the small cell lung cancer cell line, GLC4. However,for this cell line, the formazan production achieves a plateaulevel at an MTT concentration of only 1.2 mg/ml. Similarly,for the cell line, NCI-HI87, a plateau is achieved at an MTTconcentration of only 1.8 mg/ml, whereas for the cell line,

Fig. 2. Effect of MTT concentration onMTT-formazan production. Cells, either L-DAN or GLCt, were plated out at a density of5 x 10* per well and allowed to grow for 24 h.MTT (SO>ilof a solution of 0 to 5 mg/ml) wasadded (8 wells per concentration), and the platewas incubated for 4 h. The medium was removed, and the MTT-formazan crystals weredissolved in DMSO. Bars, SE.

0-4

0'3-

n 02-

L-DAN

12345

Concentration of MTT added [mg/ml)

4436

GLC4

04 08 12 16 20

Concentration of MTT added Cmg/ml)

Research. on October 4, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

TETRAZOLIUM-BASED CHEMOSENSITIVITY ASSAY

~ 05-

Cell number per weil Cx10~3)

Cell number per well (xlO*3)

Fig. 3. Relationship between MTT-formazan production and cell number forthe cell lines L-DAN and GLC,. Cells were plated out at a range of concentrationsand incubated for 24 h. Cell number was then estimated either by countingfollowing trypsin treatment if required or by incubation with MTT (L-DAN, 5mg/ml; GLC4, 2 mg/ml) for 4 h.

300 500

Wavelength Cnm)

Fig. 4. Absorption spectrum of MTT-formazan produced by incubation ofA549 cells with MTT (5 mg/ml) for 4 h at low (4 X IO3cells per well) and high(IO4 cells per well) cell densities. MTT-formazan crystals were dissolved in

DMSO.

MCF-7, the concentration is about 4.25 mg/ml (results notshown).

Relationship between MTT-Formazan Production and CellNumber. The relationship between MTT-formazan productionand cell number for the cell lines L-DAN and GLC4 is shownin Fig. 3. There is a linear relationship up to a cell number ofabout 2x10" cells per well. However, it should be noted that

neither regression line extrapolates back through the origin.Effect of Cell Number on the Absorption Spectrum of MTT-

Formazan. Fig. 4 shows the absorption spectrum of MTT-formazan produced following incubation of cells in the presenceof MTT. For low cell numbers (4 x IO3 cells per well), the

spectrum shows a single absorption maximum at a wavelengthof about 570 nm. In contrast, if the cell number is increased toIO4cells per well, the peak is much broader and the absorption

maximum is at a wavelength of about 510 nm. There is also ashoulder in this peak at a wavelength of about 580 nm.

Effect of pH on MTT-Formazan Absorbance. Fig. 5 showsthe absorption spectrum of a solution of MTT-formazan inDMSO (200 pi) in the presence of glycine buffer (25 /¿I)equilibrated at a range of pH values. For pH 7.5, a phosphate bufferwas used. When solubilized in DMSO alone, there are twoapparent absorption maxima: one at a wavelength of about 510nm and a smaller peak at about 570 nm. A similar spectrum isobserved following addition of a small amount of buffer equilibrated to a low pH (3.5). However, if the pH of the buffer is

raised, the spectrum shifts to the right and a single absorptionmaximum, at about 560 nm, is observed above pH 8.5. Theheight of the peak increases with increasing pH and reaches amaximum at pH 10.5.

Effect of Buffer Addition on the Relationship between MTT-

Formazan Production and Cell Number. Fig. 6 shows the relationship between MTT-formazan production and cell numberfor the cell line A549 both before (O) and after (•)addition ofbuffer to the wells (25 n\, 0.1 M, pH 10.5). The absorbance ofthe MTT-formazan is increased for all cell numbers, but theincrease is most marked at high cell numbers. Furthermore,following addition of buffer, the regression line extrapolatesback through the origin. Addition of 4-(2-hydroxyethyl)-l-pi-perazineethanesulfonic acid buffer (10 HIM,pH 7.4) during theincubation with MTT did not alter the slope of the line butreduced the scatter of points about the line (results not shown).Although results are shown for A549, this effect was observedwith all cell lines studied.

Effect of Culture Medium on MTT-Formazan Absorption.Table 1 shows the effect of different volumes of culture medium(F10/DMEM containing 10% fetal calf serum) on the absorbance of a standard solution of MTT-formazan. Addition ofculture medium results in an increase in the absorbance of theMTT-formazan, but the increase is greater for small additions

400 450 500 550 BOO 650

WAVELENGTH (nm)

Fig. 5. Absorption spectrum of a solution of commercially prepared MTT-formazan in DMSO or in DMSO (200 /<!)plus glycine buffer (25 /.I) equilibratedat a range of pH values. For pH 7.5, a phosphate buffer was used.

i

10 20 30 40

Cell number per well (x10~3)

Fig. 6. Effect of buffer addition on the relationship between MTT-formazanproduction and cell number. Cells (A549) were plated out at a range of concentrations and allowed to adhere for 24 h. Cells were then either trypsinized andcounted or incubated with MTT (5 mg/ml) for 4 h. Medium was removed, andMTT-formazan crystals were dissolved in DMSO, and the absorbance was recorded at 570 nm (O). Buffer (25 /J, 0.1 M glycine, pH 10.5) was added to eachwell, and the absorbance was again recorded at 570 nm (•).

4437

Research. on October 4, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

TETRAZOLIUM-BASED CHEMOSENSITIVITY ASSAY

Table I Effect of culture medium and buffer on MTT-formazan absorbanceA standard solution of MTT-formazan in DMSO was plated out in a microtiter

plate (200 nl/well), and various volumes of F10/DMEM plus 10% fetal calfserum were added to each row of 8 wells. The plate was read before and afteraddition of buffer (25 u\, 0.1 M, pH 10.5).

Volume of mediumadded(¡X)0

51020304050DMSO0.203

±0.001"

0.484 ±0.0190.563 ±0.0050.456 ±0.0030.389 ±0.0050.354 ±0.0040.353 ±0.003A570DMSO

+buffer0.429

±0.0030.454 ±0.0060.438 ±0.0100.405 ±0.0070.381 ±0.0050.388 ±0.0040.409 ±0.005

' Mean ±SE of eight wells.

•1.8x10 M01.1x10 "7M

«r10 io"8 io'8 ter7 io"6 id"5

Doxorubicin concentration (Ml

so 1.2x10 ~8M

KT» 10-8 10"'

Vincristine concentration CM)

Fig. 7. MTT-formazan production per well determined 3 days after exposureof cells to various concentrations of doxorubicin or vincristine for 24 h. MTT-formazan crystals were dissolved in DMSO, and the absorbance was noted at 570nm both before (•)and after (O) addition of buffer (25 n\, 0.1 M glycine, pH10.5). , 50% of the control cell absorbance values as used to determine theID50.

10.5), the absorbance at 570 nm was greatly increased in thewells exposed to low concentrations of drug and in the controluntreated wells; the ID50 in this instance was 1.1 x 10~8. A

similar effect was observed for the other cell lines (results notshown) and when vincristine was used (Fig. 7).

Effect of Time after Drug Removal on Chemosensitivity. Table2 shows the sensitivity to doxorubicin determined by the micro-titration assay for the cell line, A549. Cells were exposed todrug for 24 h, and MTT was added at various times after drugexposure. The apparent sensitivity of the cell line increases withincreasing time over the first 3 days after drug removal. After3 days, there is no further increase in sensitivity.

Determination of Chemosensitivity by the Microtiter Assayand Correlation with a Clonogenic Assay. Table 3 shows thechemosensitivities of the two nonadherent small cell lines, NCI-HI 87 and GLC'j, and of the adherent cell lines to doxorubicin

as determined by the microtitration assay. The overall error ofthe estimates is small (coefficient of variation, 18%) and issimilar for both adherent and nonadherent cell lines. Alsoshown in Table 3 are the sensitivities of the adherent cell linesto both doxorubicin and vincristine determined both by themicrotitration assay and by a standard clonogenic assay. Thestandard errors for the two assays are similar, and there is closeagreement between the two assays for all cell lines and for bothdrugs.

DISCUSSION

The end point of a Chemosensitivity assay is usually anestimate, either direct or indirect, of surviving cell numbers.Use of a tetrazolium dye as the end point relies on the abilityof cells to reduce the dye in a quantitative manner. We haveshown that, when cells are incubated with MTT, the amount ofMTT-formazan produced depends upon the concentration ofMTT in the incubation medium. MTT-formazan productionincreases with increasing MTT until a concentration is reachedat which MTT-formazan production is maximal. This optimumconcentration differs widely among cell lines. For example, forthe two small cell lines, NCI-HI87 and GLC4, a concentration

Table 2 Sensitivity ofA549 to doxorubicin determined by the microtitrationassay and estimated at various times after drug removal

Time after drugremoval (days) ID»(nM)

140 ±28°

120 ±2238 ±738 ±6

than for large additions. Table 1 also shows the absorbancevalues obtained after addition of buffer (25 n\ per well, 0.1 M,pH 10.5) to this same plate. Although there is still a slightincrease in absorbance in the presence of 5 ¡Aof culture medium,addition of up to 50 n\ of medium results in only a very smallreduction in MTT-formazan absorption. It should be noted thataddition of buffer to the MTT-formazan alone resulted in amarked increase in the absorbance.

Effect of Buffer Addition on the Apparent Chemosensitivity toDoxorubicin and Vincristine. Fig. 7 shows MTT-formazan production per well 3 days after exposure of cells to variousconcentrations of either doxorubicin or vincristine for 24 h.When the MTT-formazan crystals were dissolved in DMSOand the absorbance noted at a wavelength of 570 nm, the IDSOfor doxorubicin for the cell line L-DAN was 1.8 x IO"7.

However, after addition of buffer (25 /xl per well, 0.1 M, pH4438

" Mean ±SE of six estimations.

Table 3 Chemosensitivity to doxorubicin and vincristine estimated by themicrotitration assay and by a clonogenic assay

ID50 (nM)

DrugDoxorubicinVincristineCelllineNCI-HI

87GLC4L-DANA549WILSK-MESCALUG-UVWMCF7L-DANA549WILMicrotitration36±2"(6)*12

±3(6)23±4(7)23±4(7)57±6(7)33±2(10)1

1 ±1(6)19±6(4)28±1(3)12

±1(6)44±10(5)41±4 (7)Clonogenic24

±7(6)38±7(6)34±8(4)21±7(4)5±3(3)16±9(4)30±4(3)12

±2(6)22±4(5)20±8 (4)

Mean ±SE.Numbers in parentheses, number of estimations.

Research. on October 4, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

TETRAZOLIUM-BASED CHEMOSENSITIVITY ASSAY

of about 2 mg/ml (0.4 mg/ml final) gives maximum MTT-formazan production, whereas for the non-small cell lines andfor MCF-7, a concentration of about 5 mg/ml (1 mg/ml final)is required (Fig. 2). This difference cannot be explained in termsof the medium used, since it is seen between the small cell linesand MCF-7, all of which were grown in RPMI 1640 medium.Furthermore, the concentration of MTT required to give maximum MTT-formazan production by L-DAN was independentof the medium used. The screening assay (1) uses a finalconcentration of 0.4 mg/ml for all cell lines, and this is theamount used in the original assay (3), although it has beenobserved that, for a T-lymphoma cell line, an MTT concentration of 1 mg/ml gives optimal MTT-formazan production (6).We have found that, for most cell lines, a final concentrationof 1 mg/ml is more than sufficient, but we cannot recommendthis optimal for the assay, since we have observed that for somecell lines an excess of MTT can result in a reduction in MTT-formazan production.

Previous attempts to use tetrazolium dye reduction as anestimate of cell number have failed to demonstrate a linearrelationship between the amount of dye reduced and cell number, particularly at high cell numbers and absorbance values ofgreater than 0.8 (1,3, 4). Our observations demonstrate that,if the optimum MTT concentration is determined, a linearrelationship exists between cell number and MTT-formazanproduction up to a cell number of about 2 x IO4 cells per well

(Fig. 3). However, when the regression line is extrapolatedback, it does not pass through the origin such that MTT-formazan is apparently produced in the absence of any cells.Since both medium and MTT are used to establish the background absorbance when reading the plate, this observationmight indicate that the true relationship is not linear. It shouldbe noted that points can often be joined by a straight line whena curve is clearly a better fit (for example, see Fig. 1; Ref. 4).

It was noted that commercially prepared MTT-formazan isnot the same color when dissolved in DMSO as that producedby incubation of cells in the presence of MTT. This observationwas confirmed by spectral analysis of the two MTT-formazans.Commercially prepared MTT-formazan exhibits a major absorption maximum at a wavelength of 510 nm and a shoulderin the peak which may represent a second absorption maximumat around 560 to 580 nm. In contrast, that produced by cells(4000 per well) exhibits a single absorption maximum at 560to 570 nm (Fig. 1). This difference has also been reportedelsewhere (2) but was attributed to the presence or absence ofserum when MTT-formazan is dissolved in DMSO. This cannot be the only explanation, since we have shown that, if thecell number is increased to IO4 cells per well, contaminatingserum is still present, yet the absorption spectrum of the MTT-formazan product resembles that of the commercially preparedMTT-formazan (Fig. 4). A similar effect is seen when mediumis added to a solution of MTT-formazan in DMSO. Additionof a small volume of medium (5 to 10 fil) results in a muchgreater increase in absorbance at 570 nm than does addition ofa larger volume (20 to 50 n\). Furthermore, our observationsindicate that the absorption spectrum of MTT-formazan is pHdependent. At an acid pH (3.5), the spectrum resembles that ofthe commercially prepared MTT-formazan in DMSO alone,and the absorbance measured at 570 nm is low. However, if thepH of the solubili/al MTT-formazan is increased, the spectrumapproaches that of the MTT-formazan produced by cells (4000cells per well; Fig. 4 cf. Fig. 5), and the absorbance at 570 nmis markedly increased. At a pH of 10.5, the absorption spectrumexhibits a single peak at 570 nm, and the absorbance at 570

nm reaches a plateau value which is not affected by any furtherincrease in pH (Fig. 5).

We cannot explain the changes in the absorption spectrumof MTT-formazan. When MTT is reduced, the tetrazolium ringis opened and the quarternary amine is converted to a tertiaryamine. A second tertiary amine which formed one of the bondsto the quarternary amine becomes bonded to a hydrogen atom.It could be hypothesized that displacement of this hydrogen athigh pH results in changes in the absorption spectrum of themolecule. Whatever the explanation, the spectral shifts producea real problem when dye reduction is used to estimate cellnumbers. At low cell numbers, the absorption maximum is atabout 560 to 570 nm, while at high cell numbers there are twopeaks: the larger at 510 nm and a smaller peak at 570 nm.Measurement of absorbance at a single wavelength is thusinappropriate. In an attempt to overcome this problem, weraised the pH of the MTT-formazan product to 10.5. At thispH, a single peak is observed regardless of cell number or thepresence of culture medium. Addition of a small quantity ofbuffer at pH 10.5 to the solubili/ed MTT-formazan product

resulted in a significant increase in the absorbance reading forwells with a high cell density (50% increase for 3 X IO4 cells

per well; Fig. 6). Furthermore, it resulted in a clear linearrelationship between MTT-formazan production and cell number up to cell numbers of 5 x IO4 per well and absorbance

readings of 1.5 (Fig. 6). This range was well within the limitsrequired for a chemosensitivity assay even for cell lines with arapid doubling time. It is of interest that a strong base has beenused to enhance the sensitivity of the nitroblue tetrazolium test,but in this instance, it was claimed to increase the solubility ofthe formazan product in DMSO (7).

The importance of these observations becomes apparentwhen MTT reduction is used to estimate cell numbers in achemosensitivity assay. If the MTT-formazan crystals are dissolved in DMSO alone, the control cell number is markedlyunderestimated and the apparent ITX,,is about 10-fold higherthan is obtained by a standard clonogenic assay (Fig. 7 cf. Table3). Addition of buffer (pH 10.5) to these same wells resulted ina significant increase in the estimated cell number in the controlwells and in wells exposed to low concentrations of doxorubicinsuch that the apparent II).,„agrees well with that obtained bythe clonogenic assay (Fig. 7 cf. Table 3).

Previously, it has been observed that MTT-formazan absorption is greatly increased in the presence of 10 n\ or less ofculture medium but only slightly increased in the presence of10 to 40 n\ of medium (8). We also observed this effect (Table1). However, addition of the buffer at pH 10.5 causes a significant increase in the absorbance of MTT-formazan and overcomes most of the variability in absorbance reading caused bythe presence of culture medium (Table 1). This is an importantobservation, since for the nonadherent cell lines, it is notpossible to remove all of the medium from the wells prior tothe addition of DMSO. In the screening assay, it was notedthat the variability between wells was greater for nonadherentcell lines, and it was suggested that the residual medium left inthe wells may be responsible (1). We do not observe any increasein variability with nonadherent cell lines, and this could beexplained by the inclusion of the buffer at the final stage of theassay. Although it was suggested that this problem would beovercome by the introduction of an alternative tetrazolium dyewhich is reduced to a water-soluble formazan product (XTT),there are major problems encountered in the use of this dye (9).Control of the pH during incubation of cells with MTT, throughincorporation of 4-(2-hydroxyethyl)-l-piperazineethanesul-

4439

Research. on October 4, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

TETRAZOLIUM-BASED CHEMOSENSITIVITY ASSAY

fonie acid buffer into culture medium used for this step, alsodecreased the variability in MTT-formazan production betweenwells.

The protocol described in this paper also differs from othertetrazolium-based chemosensitivity assays in that cells are exposed to drug for a defined time and then grown in drug-freemedium. Although we describe a 24-h drug exposure time, itcan be varied to suit alternative experimental protocols. Previous reports describe assays designed to be used in a drug-screening program where speed and ease of automation areimportant determinants of the assay protocol (1, 2). The assaythus included continuous drug exposure for a period of up to11 days. In one a limit of 7 days was chosen in order to avoidthe need to feed the cells during the assay (1). Thus, the drugexposure time and growth period are combined. For studies ofdrug sensitivity, particularly where small changes in sensitivityare critical, a growth period is essential. It allows for eliminationof cells killed by the drug and distinguishes clearly betweenthose cells which survive drug treatment but can no longerproliferate and those which continue to proliferate (1). Theimportance of this is apparent from a decrease in the ID50 overthe first 3 days after drug exposure (Table 2). Another advantage is that differences in cell survival at different drug concentrations are amplified. This effect is maximized if the cells arefed daily, such that growth is not limited by nutrient supply.

The disadvantage of continuous drug exposure is that it takesno account of drug stability or of drug metabolism, the productsof which may be either active or inactive. The half-life of somemitomycins in tissue culture medium is about 2 h and that ofdoxorubicin only 15 h (10). Furthermore, there are additionalproblems which are specific to the drug used. For example, thecardiotoxicity of doxorubicin is associated with alterations inmitochondria! enzyme activities (11). Although the exact siteor sites of MTT reduction in intact cells are unclear, it is knownthat mitochondria! dehydrogenase activities can reduce the dye(12, 13). Thus, with continuous exposure, doxorubicin couldinterfere with MTT reduction as a result of nonlethal damageto the cells. Furthermore, the absorption maximum for doxorubicin is between 500 and 600 nm, and thus it may wellcontribute to absorbance measurements at 570 nm as notedpreviously (1). Therefore, we use a defined drug exposure timeand include a growth period of about 3 cell doubling times.

Thus, the assay described differs markedly from previouslydescribed tetra/oliimi dye-based assays. Cells are incubated withMTT, but the MTT concentration is predetermined for the cellline in order to ensure maximum MTT-formazan production.

The formazan product is dissolved in DMSO but, in contrastto other assays, the pH is then adjusted to 10.5 so as toovercome the effects of culture medium and cell density on theMTT-formazan absorption spectrum. In addition, cells areexposed to drug for a defined time, and surviving cells areallowed to grow for two to three cell-doubling times in theabsence of drug. During this time, the culture medium isreplaced daily such that the growth rate is not limited bynutrient depletion. These modifications clearly detract from theadvantages of the assay in terms of automation. However, wedescribe a rapid and simple chemosensitivity assay that givesresults closely similar to those obtained by a clonogenic assaywhich can be used for both adherent and nonadherent cell lines.

REFERENCES

1. Carmichael, J., DeGraff, W. G., Gazdar. A. F., Minna, J. D., and Mitchell,J. B. Evaluation of a tetrazolium-based semiautomated colorimetrie assay:assessment of chemosensitivity testing. Cancer Res., 47: 936-942, 1987.

2. Alley, M. C, Scudiere, D. A., Monks, A., Hursey, M. L., Czerwinski, M. J.,Fine, D. L., Abbot, B. J., Mayo, J. G., Shoemaker, R. H., and Boyd, M. R.Feasibility of drug screening with panels of human tumor cell lines using amicroculture tetrazolium assay. Cancer Res., 48: 589-601, 1988.

3. Mosmann, T. Rapid colorimetrie assay for cellular growth and survival:application to proliferation and cytotoxicity assays. J. Immunol. Methods,65: 55-63, 1983.

4. Cole, S. P. C. Rapid chemosensitivity testing of human lung tumor cellsusing the MTT assay. Cancer Chemother. Pharmacol., 17: 259-263, 1986.

5. Carney, D. N., Gazdar, A. F., Bepler, G., Guccion, J. G., Marangos, P. J.,Moody, T. W., Zweig, M. H., and Minna, J. D. Establishment and identification of small-cell lines having classic and variant features. Cancer Res., 45:2913-2923, 1985.

6. Denizot, F., and Lang, R. Rapid colorimetrie assay for cell growth andsurvival. Modifications to the tetrazolium dye procedure giving improvedsensitivity and reliability. J. Immunol. Methods, 89: 271-277, 1986.

7. Rook, G. A. W., Steel, J., Umar, S., and Dockrell, H. M. A. A simple methodfor the solubilization of reduced NBT and its use as a colorimetrie assay foractivation of human macrophages by -, -interteron. J. Immunol. Methods, 82:161-167, 1985.

8. Twentyman, P. R., and Luscombe, M. A study of some variables in atetrazolium dye (Mil) based assay for cell growth and chemosensitivity. Br.J. Cancer, 56: 279-285, 1987.

9. Scudiero, D. A., Shoemaker, R. H., Paull, K. D., Monks, A., Tierney, S.,Nofziger, T. H., Currens, M. J., Seniff, D., and Boyd, M. R. Evaluation of asoluble tetrazolium/formazan assay for cell growth and drug sensitivity inculture using human and other tumor cell lines. Cancer Res., 48:4827-4833,1988.

10. Beijnen, J. H., Van der Nat, J. M., Labadie, R. P., and Underberg, W. J. M.Decomposition of mitomycin and anthracycline cytostatics in cell culturemedium. Anticancer Res., 6: 39-44, 1986.

11. Praet, M., Laghmiche, M., Pollakis, G., Goormaghtigh, E., and Ruysschaert,J. M. In vivo and in vitro modifications of the mitochondria! membraneinduced by 4'Epi-adriamycin. Biochem. Pharmacol., 35:2923-2928, 1986.

12. Hess, R., and Pearse, A. G. E. Pathways of reduced pyridine nucleotideoxidation in rat-brain homogenate demonstrated by a tetrazolium method.Biochim. Biophys. Acta, 71: 285-294, 1963.

13. Slater, T. F., Sawyer, B., and Stranii, U. Studies on a succinate-tetrazoliumreducíasesystem. III. Points of coupling of four different tetrazolium salts.Biochim. Biophys. Acta, 77: 383-393, 1963.

4440

Research. on October 4, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

1989;49:4435-4440. Cancer Res   Jane A. Plumb, R. Milroy and S.B. Kaye  by a Novel Tetrazolium-based AssayBromide-Formazan Absorption on Chemosensitivity Determined3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium Effects of the pH Dependence of

  Updated version

  http://cancerres.aacrjournals.org/content/49/16/4435

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/49/16/4435To request permission to re-use all or part of this article, use this link

Research. on October 4, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from