A number of investigators have observed that the use of 4-hydroperoxycyclop
hosphamide (4-HC) in multiwell plate cytotoxicity assays can be associated
with toxicity to cells in wells that contain no drug. Previous reports have
implicated diffusion of 4-HC decomposition products, and acrolein in parti
cular, as the active species. Purpose: The purpose of this study was to elu
cidate the species responsible for the airborne cytotoxicity of 4-HC, and t
o devise ways to minimize such effects in chemosensitivity assays. Methods:
To this end, analogues of 4-HC were synthesized to identify the contributi
ons of individual cyclophosphamide metabolites to cytotoxicity. The analogu
es were then tested for activity against three human breast tumor cell line
s (including a line resistant to 4-HC), and one non-small-cell lung carcino
ma line. Cytotoxicity was evaluated by assays that quantitate cellular meta
bolism and nucleic acid content. Results: Didechloro-4-hydroperoxycyclophos
phamide, a compound that generates acrolein and a nontoxic analogue of phos
phoramide mustard, gave no cross-well toxicity. In contrast, a significant
neighboring well effect was observed with phenylketophosphamide, a compound
that generates phosphoramide mustard but not acrolein. Addition of authent
ic chloroethylaziridine reproduced the airborne toxicity patterns generated
by 4-HC and phenylketophosphamide. Increasing the buffering capacity of th
e growth medium and sealing the microtiter plates prevented airborne cytoto
xicity. Conclusions: Since it is unlikely that phosphoramide mustard is vol
atile, these findings implicate chloroethylaziridine rather than acrolein a
s the volatile metabolite of 4-HC that is responsible for airborne cytotoxi
city. The fact that chloroethylaziridine is generated in amounts sufficient
to volatilize, diffuse across wells and cause cytotoxicity indicates that
it is an important component in the overall cytotoxicity of 4-HC in vitro.
Furthermore, these findings suggest that chloroethylaziridine may also cont
ribute to the toxicity of cyclophosphamide in vivo.