Multicell spheroids were exposed to DNA-damaging agents with the aim o
f determining whether prompt DNA damage could be predictive for cell k
illing and drug resistance. Chinese hamster V79 cells, SiHa human cerv
ical carcinoma cells, and WiDr human colon carcinoma cells were grown
as spheroids and exposed to N-methyl-N'-nitro-N-nitrosoguanidine (MNNG
), 4-nitroquinoline-1-oxide (4NQO), doxorubicin, etoposide, actinomyci
n D, 1-(2-nitro-1-imidazolyl)-3-aziridino-2-(RSU 1069), 3-amino-1,2,4-
benzotriazine-1,4-dioxide (tirapazamine) and nitrogen mustard. Average
DNA damage measured using the alkali comet assay generally correlated
with cell killing irrespective of exposure times or drug concentratio
n, However, better predictive power was achieved by using DNA damage l
evels in individual cells to identify the fraction of cells containing
sufficient numbers of DNA strand breaks to cause death. Using this co
ncept of a ''threshold'' for DNA damage, cell survival could be predic
ted for exposure to 4NQO, tirapazamine, nitrogen mustard, RSU 1069, an
d actinomycin D and was largely independent of cell type. The threshol
d value varied for each drug. For 4NQO, tirapazamine, and RSU 1069, DN
A damage equivalent to about 10,000 strand breaks/cell was not toxic t
o cells of any spheroid type. Conversely, for actinomycin D, any DNA d
amage above background levels (approximate to 100 breaks) was toxic fo
r all three cell types. For some DNA-damaging drugs, the lack of corre
lation between DNA damage and cell killing was also informative. For e
toposide and doxorubicin, no common threshold for cell killing could b
e determined, consistent with the hypothesis that DNA damage is only o
ne of the actions of these drugs leading to cell death. For MNNG, the
tail moment threshold varied significantly for the different spheroid
types, probably indicating differences in repair. Overall, for five of
the eight drugs, DNA damage measured using the comet assay was an eff
ective and quantitative method of predicting drug cytotoxicity in comp
lex multicelled systems.