Tumour oxygenation has been shown to play an important role in radiotherapeutic response both in experimental tumours and in the clinic. To eliminate microregions of hypoxia, i.e. low oxygenation, numerous strategies have been used, generally aiming either to increase oxygen delivery to the tumour (Horsman et al, 1994) or to directly target the hypoxic cells using radiosensitizers or hypoxic cell cytotoxic agents (Brown and Giaccia, 1994). Several recent reports have also suggested the use of angiogenesis inhibitors (Klauber et al, 1997;OÕReilly et al, 1997) or thrombocytic agents (Huang et al, 1997) to directly target the tumour vasculature. A major shortcoming in evaluating the use of any of these approaches, however, is the inability to adequately quantify and understand the accompanying changes in tumour physiology.The current study presents a method for combining several sophisticated techniques to obtain a comprehensive two-dimensional mapping of the relationships among tumour vascular configuration, oxygen transport and hypoxic development. First, tumour oxygen availability is spatially defined by measuring intravascular blood oxygen saturations (HbO 2 ) cryospectrophotometrically in a frozen tumour block. Second, hypoxic development, in relation to the intravascular oxygen availability, is quantified in an adjacent histological section, using immunohistochemical detection of a recently developed nitroheterocyclic hypoxia marker (EF5). Third, a combination of fluorescent and immunohistological stains is used to define the distribution of distances from the viable tumour cells to the nearest anatomical or perfused blood vessel.
MATERIALS AND METHODS
Mice and tumour modelsThe KHT tumour, a sarcoma maintained in vivo, was passaged approximately every 2 weeks by i.m. inoculation of single-cell suspensions prepared by a mechanical dissociation procedure (Thomson and Rauth, 1974). Using 6-to 8-week-old female C3H/HeJ mice (Jackson Laboratories, Bar Harbor, ME, USA), 2×10 5 KHT cells were inoculated i.m. into the hind limbs. Tumours were selected for analysis when they reached volumes of between 320 and 1100 mm 3 (as measured by callipers (volume = πqdiam-eter 3 /6). Guidelines for the humane treatment of animals were followed as approved by the University Committee on Animal Resources.
Injection of fluorescent stains and EF5 hypoxic markerLocalized areas of tumour hypoxia were assessed in frozen tissue sections by immunohistochemical identification of sites of 2-nitroimidazole metabolism. A pentafluorinated derivative (EF5) of etanidazole (synthesized by Dr R Vishnuvajjala of the National Cancer Institute) was injected i.v., 1 h before tumour freezing (0.2 ml of 10 mM EF5). Protein conjugates of EF5 have been used previously to immunize mice from which monoclonal antibodies were developed (Lord et al, 1993). These antibodies are extremely specific for the EF5 drug adducts that form when the drug is incorporated by Summary Despite the possibility that tumour hypoxia may limit radiotherapeutic response, the under...