Inf. J. Radiation Oncology Bid. Phys.. Vol. 15, pp. 699-702 0360~3016/88 $3.00 + .oO Printed in the U.S.A. All rights reserved. Copyright 0 1988 Pergamon Press plc

??Original Contribution

USE OF A COLORIMETRIC MICROTITER (MTT) ASSAY IN DETERMINING THE RADIOSENSITIVITY OF CELLS FROM MURINE SOLID TUMORS

TODD H. WASSERMAN, M.D.* AND PETER TWENTYMAN, PH.D.

MRC Unit and University Department of Clinical Oncology and Radiotherapeutics, Hills Road, Cambridge CB2 2QH, England

We assessed the use of a calorimetric assay for determination of radiosensitivity for cells taken directly from murine solid tumors. The assay uses microtier plates and measures the ability of viable cells to reduce a tetrazolium salt (MIT) to an insoluble form, a formazan salt. We established the dependency of the assay on the cell number and time of assay for two murine tumors (EMT-6 and IUF-1). We compared the MTT assay to the standard clonogenic assay and had good agreement of surviving fraction after radiation doses of 2 and 4 Gy. It is possible, therefore, to adapt the MTT assay for use with cell suspensions prepared directly from fresh murine tumors. This may provide a methodology for the determination of the clinical radiosensitivity of tumors including fresh clinical tumor specimens.

M’lT assay, Tetrazolium assay, Cell survival assay.

INTRODUCI’ION

There is a need for new assays of cellular response to cy- totoxic agents which can be completed in a rapid and economical manner. Such assays would have a variety of applications when used with established cell lines and with cells disaggregated from solid tumors. To this end, a number of microtiter assays are being developed. An assay which has been the subject of much recent study is the quantitative calorimetric assay using a tetrazolium salt as originally developed by Mossmann in 1983 and usually referred to as the “MTT assay.” This assay quan- titates living but not dead cells and has the advantage of being rapid, precise and semi-automatable. Some inves- tigators have recently demonstrated for established cell lines, good agreement between drug and radiation re- sponse curves obtained using the MTT assay and those obtained using other more traditional assays.2,3.4.1 ’ Sim- ilarly, drug response data for clinical tumor specimens have been obtained by one group’ using the MTT assay.

The assay measures the ability of viable cells to reduce MTT, (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetra- zolium bromide) a yellow tetrazolium salt, to a purple formazan precipitate, a process which requires active mi- tochondrial function. It is usually carried out on cells growing in 96-well microtitre plates. Following solubili-

zation of the formazan precipitate in an organic solvent, the optical density of the resulting solution is measured on a multiwell spectrophotometric plate reader (ELISA reader). The optical density of the formazan solution has been shown by us and others to be approximately pro- portional to the number of viable cells within certain lim- its 2,3,4.7, I 1

The National Cancer Institute’s anticancer drug screening program has recently been redirected towards in vitro studies using multiple lines of human tumor cells. The MTT assay is used as the standard determinant of drug responsiveness.

In this communication we describe studies designed to investigate the possible use of the MTT assay in deter- mining the radiation responsiveness of cells disaggre- gated directly from solid tumors.

METHODS AND MATERIALS

We used the MTT assay with a number of modifica- tions of the technique of Mossmann. We prepared sin- gle cell suspensions from fresh murine tumors (EMT-6 and RIF-1) grown in flanks of BALB-C and C3H mice respectively. Following excision and mincing of the tu- mors, single cell suspensions were prepared using bacte-

Presented in Part at the American College of Radiology Con- ference on the Prediction of Tumor Treatment Response held April 2 l-24, 1987, in Banff, Canada.

* Visiting Scientist, Professor of Radiation Oncology, Mal- linckrodt Institute of Radiology, Washington University, St. Louis, MO.

Reprint requests to: Todd H. Wasserman, M.D., Radiation Oncology Center, Mallinckrodt Institute of Radiology, 4939 Audubon Ave., Suite 5500 St. Louis, MO 63 110, U.S.A.

Accepted for publication 1 April 1988.

699

700 I. J. Radiation Oncology 0 Biology 0 Physics September 1988, Volume 15. Number 3

rial neutral protease as previously described.13 The me- dium used was Eagles MEM with Earles salts* supplemented with 20% newborn calf serum and antibi- otics. After preparation of the single cell suspension, cells were counted on a haemocytometer and appropriate di- lutions made. Aliquots of 200 ~1 of the suspension were pipetted into wells on 96-well microtitre plates.? The plates were then incubated at 37°C in an atmosphere of 92% sir/8% CO2 for a period of 4 or 5 days. The plates were then removed from the incubator and 20 ~1 of a 5 mg/ml solution of MTT in PBS was added to each well. The plates were then returned to the incubator for a fur- ther 4 hours. At the end of this time, the bulk of the me- dium was aspirated from each well and 200 ~1 of di- methyl sufoxide (DMSO) added. The plates were shaken for 10 minutes and read on a Titertek Multiskan plate reader (ELISA reader) using a blank of DMSO plus me- dium. Initial experiments were read at a wave length of 600 nanometers (nm) and later experiments at 550 nm. The optical density of a solution of cellular formazan in DMSO is approximately 1.7 times greater when read at 550 nm than at 600 nm.”

tional to cell number for both cell types, and we therefore decided to aim to work at an optical density (OD) in the range 0.5 to 1.0 for untreated cells. We found that the initial cell numbers per 200 ~1 well to produce such an OD were 1500 and 500 for EMT-6 cells with 4 or 5 days incubation respectively. For the RIF- 1 tumors, the corre- sponding cell numbers were 6000 and 3000. The graphs of these data are shown as Figures 1 and 2.

The cellular response to radiation doses of 0- 10 Gy (using 250 kv X rays at 0.65 Gy per minute) was tested. Cell suspensions were irradiated at a density of lo5 cells per ml in a volume of 50 ml of medium in a cell culture flask on ice. After the appropriate radiation dose incre- ments, the cell suspension was mixed and aliquots with- drawn.

Curves were also obtained relating optical density to the number of irradiated cells plated. Curves for EMT-6 cells are shown in Figure 3. The curves are parallel to those shown in Figure 1 but displaced to the right. Sim- ilar curves (not shown) were also obtained for cells from RIF- 1 tumors. It therefore appeared legitimate to use re- duction in optical density at days 4 or 5 as a measure of the radiation response of a given number of cells plated. For determination of radiation response using the MTT assay, the given number of cells per well were plated on day 0 and the optical density produced by Unix-radiated cells at day 4 or 5 provided a control value. The optical density produced by the same number of irradiated cells at day 4 or 5 was taken as a measure of radiation response (Table 1). For the EMT-6 tumor, 1500 cells/well were used on day 4 and 500 cells/well at day 5, and for the RIF- 1 tumor 6000 cells/well were used at day 4 and 3000 cells/well at day 5. The plating efficiencies in the clono- genie assay were 25% for the RIF-1 and 49% for the EMT-6 tumor systems respectively.

In some experiments, we used a selective adherence technique” to remove the non-tumor cells from the sus- pensions prepared from EMT-6 solid tumors. The radia- tion response of the original suspension could then be compared with that of the enriched suspension of tumor cells. The radiation survival of cells from both tumor types was also determined using a standard clonogenic assay with an incubation time of lo- 12 days and count- ing colonies of greater than 50 cells.” This was done to produce cell survival data for comparison with the results of the MTT assay.

We also prepared single cell suspensions from a num- ber of normal tissues, taken from non-tumor bearing mice. Suspensions of bone marrow cells were prepared by “flushing” cells from the femurs, while suspension of lung, liver, spleen and brain were prepared using bacte- rial neutral protease.

A comparison of the radiation dose response for each tumor cell type as determined by MTT assay and by clo- nogenic assay is shown in Table 1. The data are means of three separate experiments. Data for doses of 8 or 10 Gy showed poor correlation and are not presented. Each assay shows a dose-response relationship but the re- sponse is apparently greater as assessed by clonogenic as- say. This may reflect continued mitochondrial activity (and hence MTT reduction) in cells rendered clonogeni- tally sterile or it may indicate a discrepancy between the radiation responsiveness of the clonogenic population and that of the entire population which contributes to the MTT assay. The data at day 5 of the MTT assay gives better agreement, and in general, the longer the time to assay, the better the data agreement.

RESULTS

We first established optimal cell numbers and incuba- tion times for non-treated tumor cells. Optical density in the range 0.1 to 1.2 (550 nm) was established as propor-

We were not able to demonstrate any difference be- tween the radiation response as determined by MTT as- say for an unseparated EMT-6 tumor cell suspension and a suspension enriched in tumor cells by the selective adherence technique (data not shown). The non-tumor cells therefore either make no significant contribution to MTT reduction or else have a similar radiosensitivity to the tumor cells.

For each of the normal tissues tested, plating of up to

* Gibco Biocult Ltd., Paisley, Scotland. t Falcon Plastics, Becton Dickinson, Cowley, Oxford, En-

gland.

MTT assay in determining the radiosensitivity of cells 0 T. H. WASSERMAN AND P. TWENTYMAN 701

1.4 -

I A***-J Day 5

6 1.2 - C-4 Day 4 8 -Day 3 $ l.O- (Average of 6 values) _________<_________J_------_---/---------_

c z .8-

? -I .6- a u F a .4- 0

.2 -

01 IO IO2 IO” IO4

NUMBER OF CELLS/WELL

Fig. 1. Optical density at 600 nanometers in the MTT assay for cells from in vivo EMT-6 (Cambridge) tumors. Data shown are for incubation time periods of 3,4, or 5 days and with the initial number of cells per well in the range of 60- 10,000.

1 O5 cells per well resulted in negligible reduction of MTT (optical density < 0.1) for either 4 or 5 days incubation.

DISCUSSION

It is clearly possible to adapt the MTT assay for use with cell suspensions prepared directly from fresh mu- rine tumors. It may also be adaptable for use with cells from fresh in vivo human tumors, and limited clinical data from Bernheim et al. ’ in Belgium and our labora- tory in Cambridge suggest that this is indeed possible. The assay is inexpensive, fast, semi-automatable, and re- quires low cell number and a short incubation time. We

1.6 r I

b**..A Day 5

6 1.2 +-+ Day 4

8 M Day 3

.2

(Average of

would hope that it will be possible to use the assay to meaningfully assess the responsiveness of tumor cell sus- pensions to a dose of 2 Gy, which is now felt to be an important determinant of radiation sensitivity.5.6*8.9 Be- cause cells from a variety of normal tissues do not give significant MTT reduction and because of the good agreement between data obtained using enriched and non-enriched suspensions of tumor cells, give hope that the normal tissue component of tumor cell suspensions will not seriously jeopardize the use of the MTT assay with cells from disaggregated tumors. The simple culture conditions used here for established murine tumors may not provide adequate conditions for the growth of cells

OL I 1 I1llll I I I ,,,,I 1 I I11111

IO2 IO” lo’ 10”

NUMBER OF CELLS/WELL

Fig. 2. Optical density at 550 nanometers in the MTT assay for cells from in vivo RIF-1 tumors with incubation times of 3,4, or 5 days and with the number of cells per well in the range of 1 ,OOO-20,000.

702 I. J. Radiation Oncology 0 Biology 0 Physics September 1988. Volume 15, Number 3

1.6 -

1.4 - t

z A......* Day 5

= 1.2 - 8

e-4 Day 4

; l.O-

_ Day 3

(Average of 6 values)

z z 8-

i J .6- J F g4

.2 -

0 I I I11111

IO IO2 IO3 IO’ NUMBER OF CELLS/WELL

Fig. 3. Same data as in Figure 1 except after a single dose of 4 Gy radiation, prior to incubation,

Table 1. Comparison of radiation response data for each tumor by clonogenic and MTT assays

RT dose in Gy Assay

Tumor

EMT-6 RIF- 1

2 Clonogenic* .90 (. 10) 0.80 (.08) MTT day 4t .90 (.14) 0.76 (09) MTT day 5 .88 (.09) 0.84 (.2 1)

4 Clonogenic .46 (.06) 0.45 (.08) MTT day 4 .72(.13) 0.63 (. 14) MTT day 5 58 (. 10) 0.64 (.14)

6 Clonogenic .15 (.03) 0.22 (.06) MTT day 4 .42 (.15) 0.43 (. 11) MTT day 5 .31 (.12) 0.36 (. 14)

( ) = + 1 standard deviation. * Surviving fraction as assessed by clonogenic assay. t Optical density as fraction of control in the MTT assay. Both data sets are means of three separate experiments.

from human clinical specimens and the addition of growth factors and other medium supplements is likely to be required.

In an in vitro study by Carmichael et a1.3 good agree- ment was seen between radiation dose response data for cell lines obtained by MTT and clonogenic assay. In that study, the MTT assay duration was varied with a longer assay time being used for lines with a longer doubling time. It remains to be established whether it will be neces- sary to adjust the assay time to optimize the radiation response data obtainable from tumor cell suspensions.

More data are needed to determine the ranking of clin- ically radiosensitive to resistant tumors in the MTT assay and how this ranking compares to such a ranking ob- tained using clonogenic and other assays. We believe fur- ther evaluation of this assay technique using a variety of experimental tumors and clinical tumor samples is indicated.

REFERENCES

1.

2.

3.

4.

5.

6.

7.

Bemheim, J., Van Belle, S., Roobol, C.: A novel in vitro chemoresistance test for human turnout-s by dye reduc- tion. Proc. ASCOS: 16, 1986. Carmichael, J., DeGraff, W.G., Gazdar, A.F., Minna, J.D., Mitchell, J.B.: Evaluation of a tetrazolium-based semiau- tomated calorimetric assay. Assessment of chemosensitiv- ity testing. Cancer Res. 47: 936-942, 1987. Carmichael, J., DeGraff, W.G., Gazdar, A.F., Minna, J.D., Mitchell, J.B.: Evaluation of a tetrazolium-based semiau- tomated calorimetric assay. Assessment of radiosensitiv- ity. Cancer Rex 47: 943-946, 1987. Cole, S.P.C.: Rapid chemosensitivity testing of human lung tumor cells using the MTT assay. Cancer Chemother. Pharmacol. 17: 259-263,1986. Deacon, J., Peclcham, M.J., Steel, G.G.: The radiorespon- siveness of human turnout-s and the initial slope of the cell survival curve. Radiother. Oncol. 2: 3 17-323, 1984, Fettil, B., Malaise, E.P.: Intrinsic radiosensitivity of hu- man cell lines is correlated with radioresponsiveness of hu- man tumors. Analysis of 101 published survival curves. Int. J. Radiat. Oncol. Biol. Phys. 11: 1699-1707, 1985. Mossmann, T.: Rapid calorimetric assay for cellular

8.

9.

10.

11.

12.

13.

growth and survival application to proliferation and cyto- toxicity assay. J. Immunol. Methods 63: 55-63, 1983. Rofstad, E.K.: Human tumour xenografts in radiothera- peutic research. Radiother. Oncol. 3: 35-46, 1985. Rofstad, E.K., Wahl, A., Brustad, T.: Radiation sensitivity in vitro of cells isolated from human tumor surgical speci- mens. Cancer Res. 47: 106- 110, 1987. Twentyman, P.R., Bleehen, N.M.: Changes in sensitivity to radiation and to bleomycin occurring during the life his- tory of monolayer cultures of a mouse tumour cell line. Brit. J. Cancer31: 68-74, 1975. Twentyman, P.R., Luscombe, M.: A study of some vari- ables in a tetrazolium dye (MTT) bases assay for cell growth and chemosensitivity. Brit. J. Cancer 56: 279-285, 1987. Twentyman, P.R., Watson, J.V.: Separation of clonogenic tumour cells from EMT-6 mouse mammary tumours. Brit. J. Cancer35(1): 120-122, 1977. Twentyman, P.R., Yuhas, J.M.: Use of a bacterial neutral protease for disaggregation of mouse tumours and multi- cellular tumour spheroids. Cancer Letters 9: 225-228. 1980.