An ideal cancer chemotherapeutic prodrug is completely inactive until metab
olized by a tumour-specific enzyme, or by an enzyme that is only metabolica
lly competent towards the prodrug under physiological conditions unique to
the tumour. Human cancers, including colon, breast, lung, liver, kidney and
prostate, are known to express cytochrome P450 (CYP) isoforms including 3A
and 1A subfamily members. This raises the possibility that tumour CYP isof
orms could be a focus for tumour-specific prodrug activation. Several appro
aches are reviewed, including identification of prodrugs activated by tumou
r-specific polymorphic CYPs, use of CYP-gene directed enzyme prodrug therap
y and CYPs acting as reductases in hypoxic tumour regions. The last approac
h is best exemplified by AQ4N, a chemotherapeutic prodrug that is bioreduct
ively activated by CYP3A. This study shows that freshly isolated murine T50
/80 mammary carcinoma and RIF-1 fibrosarcoma 4-electron reduces AQ4N to its
cytotoxic metabolite, AQ4 (T50/80 K-m = 26.7 mu M, V-max = 0.43 mu M/mg pr
otein/min; RIF-1 K-m = 33.5 mu M, V-max = 0.42 mu M/mg protein/min) via AQM
, a mono-N-oxide intermediate (T50/80 K-m = 37.5 mu M; V-max = 1.4 mu M/mg
protein/min; RIF-1 K-m = 37.5 mu M; V-max = 1.2 mu M/mg protein/min). The p
rodrug conversion was dependent on NADPH and inhibited by air or carbon mon
oxide. Cyp3A mRNA and protein were both present in T50/80 carcinoma grown i
n vivo (RIF-1 not measured). Exposure of isolated tumour cells to anoxia (2
h) immediately after tumour excision increased cyp3A protein 2-3-fold over
a 12 h period, after which time the cyp protein levels returned to the lev
el found under aerobic conditions. Conversely, cyp3A mRNA expression showed
an initial 3-fold decrease under both oxic and anoxic conditions; this ret
urned to near basal levels after 8-24 h. These results suggest that cyp3A p
rotein is stabilized in the absence of air, despite a decrease in cyp3A mRN
A. Such a 'stabilization factor' may decrease cyp3A protein turnover withou
t affecting the translation efficiency of cyp3A mRNA. Confirmation of the C
YP activation of AQ4N bioreduction was shown with human lymphoblastoid cell
microsomes transfected with CYP3A4, but not those transfected with CYP2B6
or cytochrome P450 reductase. AQ4N is also reduced to AQ4 in NADPH-fortifie
d human renal cell carcinoma (K-m = 4 mu M, V-max = 3.5 pmol/mg protein/min
) and normal kidney (K-m = 4 mu M, V-max = 4.0 pmol/mg protein/min), both p
reviously shown to express CYP3A. Germane to the clinical potential of AQ4N
is that although both normal and tumour cells are capable of reducing AQ4N
to its cytotoxic species, the process requires low oxygen conditions. Henc
e, AQ4N metabolism should be restricted to hypoxic tumour cells. The isofor
m selectivity of AQ4N reduction, in addition to its air sensitivity, indica
tes that AQ4N haem coordination and subsequent oxygen atom transfer from th
e active-site-bound AQ4N is the likely mechanism of N-oxide reduction. The
apparent increase in CYP3A expression under hypoxia makes this a particular
ly interesting application of CYPs for tumour-specific prodrug activation.