Human head and neck squamous cell carcinoma cells transfected with mutant TP53 (SAS/mp53) or with neo vector as a control (SAS/neo) were inoculated subcutaneously into both the hind legs of Balb/cA nude mice. Mice bearing the tumors received 5-bromo-2'-deoxyuridine (BrdU) continuously to label all proliferating (P) cells in the tumors. The mice then received a hypoxic cytotoxin, tirapazamine (TPZ) or TX-402, with or without a vascular targeting agent (VTA), ZD6126. Another group of mice given ZD6126 received a series of test doses of gamma-rays while alive or after tumor clamping to obtain hypoxic fractions (HFs) in the tumors. After each treatment, the tumor cells were isolated and incubated with a cytokinesis blocker (cytochalasin-B), and the micronucleus (MN) frequency in cells without BrdU labeling [quiescent (Q) cells] was determined using immunofluorescence staining for BrdU. The MN frequency in total (P+Q) tumor cells was determined from the tumors that were not pretreated with BrdU. Both hypoxic cytotoxins showed significantly greater toxicity toward SAS/mp53 and Q than SAS/neo and total tumor cells, respectively. The sensitivity to TX-402 was significantly higher than that to TPZ in both total and Q tumor cells of both tumors. The significant enhancive effect by ZD6126 combined with each hypoxic cytotoxin was similar irrespective of p53 status, and slightly greater for total than Q cells probably because of a more marked increase in the size of the HFs in total than Q cells on the use of ZD6126 in both tumors, resulting in a reduction of the difference in the sensitivity to the hypoxic cytotoxin between total and Q cells. In the treatment of conventional cancer therapy-resistant Q tumor cells or p53-mutated tumor cells, the use of hypoxic cytotoxin was very promising either alone or when VTA was co-administered. TX-402 might be more promising than TPZ, although further study of the toxicity to normal tissue is needed.