Skin sensitization by misonidazole: a demonstration of uniform mild hypoxia.

Skin reactions on irradiated mouse feet were used to measure the radiosensitization of normal tissues by misonidazole (MISO). Fractionation schedules of 1, 2, 5 and 10 daily doses of X-rays were combined with either 100 mg/kg or 670 mg/kg MISO. When unanaesthetized mice were irradiated in air, significant sensitization was observed with both the high and low drug doses, in all fractionation schedules. There was no decrease in sensitization with fractionation, even using fractions as small as 5 Gy. This indicates that many of the cells in mouse skin may be marginally hypoxic, and that sensitization at low doses is possible. Irradiation in O2 without MISO rendered the skin more sensitive to X-rays than in air. MISO given 30 min before single doses of radiation further sensitized the skin, but for 10 fractions in O2 no MISO sensitization was detected. There was little evidence for cytotoxic killing in skin by MISO. Repair of radiation damage was slightly reduced when MISO was present, during or after irradiation.

Irradiation in 02 without MISO rendered the skin more sensitive to X-rays than in air. MISO given 30 min before single doses of radiation further sensitized the skin, but for 10 fractions in 02 no MISO sensitization was detected.
There was little evidence for cytotoxic killing in skin by MISO. Repair of radiation damage was slightly reduced when MISO was present, during or after irradiation. THE RAT10NALE for predicting a therapeutic benefit with the combined use of misonidazole (MISO) and X-rays is based on the assumption that tumours contain hypoxic cells which can be sensitized, whereas normal tissues do not. The assumption that all normal tissues are well oxygenated has been questioned (Fowler et al., 1965;Hendry & Sutton, 1978;Hendry, 1]979) and there are many singledose studies with X-rays + MISO which demonstrate significant radiosensitization of a variety of rodent normal tissues (Brown, 1975;Gonzalez & Breur, 1978, Hendry, 1978Hornsey & Field, 1979;Suzuki et al., 1]977; Yuhas et al., 1977;Yuhas, 1979) though other studies demonstrate no sensitization (Field & Morris, 1981;Travis et al., 1982;Van der Kogel, personal communication). If normal-tissue radiosensitization by MISO might be a problem clinically, it is important to determine which tissues are at risk, and to estimate the extent of normal-tissue sensitization, particularly at low drug an(1 X-ray doses. The available literature on normal-tissue sensitization is mainly restricted to large single doses of X-rays, usually combined with large doses of MISO (600-1000 mg/kg), which are obviously not very relevant to the clinic. If a very small proportion of acutely hypoxic cells existed in normal tissues, this would only be demonstrable at high X-ray doses, such as have been used in these studies, and would presumably be unimportant in a clinical regime using multiple X-ray doses of 2-3 Gy. However, Hendry (1979) has pointed out that many rodent normal tissues do not behave in a way consistent with a mixed population consisting mainly of radiosensitive oxic cells. Rather, the response of these tissues suggests either a rapidly cycling oxic/hypoxic state, probably due to fluctuations in blood flow, or an overall sub-optimal level of tissue oxygenation.
In tissues with a homogeneous hypoxia, MISO might be expected to sensitize the radiation response even at low X-ray doses, and in this situation the problem of a small amount ofnormal-tissue hypoxia becomes clinically relevant.
The present series of experiments was designed to measure the observed sensitization enhancement ratio (SER*) in mouse skin using either 100 mg/kg or 670 mg/kg MISO, with both single and multiple daily fractions of X-rays. Experiments were performed both in air and in normobaric 02 at room temperature (23-25°C) using unanaesthetized animals.
The possibility of there being some cytotoxic killing of skin cells by MISO was investigated by administering the drug immediately after each X-ray dose (1 or 10 fractions). It was also possible to use these data to examine the influence of MISO on repair of radiation damage, by comparing the repair increments for fractionated irradiations in the presence or absence of MISO.

MATERIALS AND METHODS
The left hind foot of unanaesthetized male WHT/GyfBSVS mice was irradiated in 02 or air. Irradiations were with 240 kV X-rays at 2-2 Gy/min, filtered with 0-24 mm Cu and 1 mm Al and with a HVL of 1-3 mm Cu. During the irradiations, mice were held in lead restraining boxes from which the left hind foot and thigh protruded. These lead boxes were loaded onto a perspex plate, using a series of anatomically positioned posts designed to hold the foot in position in the X-ray beam without constriction of the blood supply (Douglas & Fowler, 1976). A lead shield with a 6cm-diameter hole was used to collimate the X-rays and the feet of 5 mice were irradiated simultaneously in a vertical beam. For irradiations in 02 the whole apparatus was placed in a polythene bag and flushed with 02 flowing at 5-6 1/min at room temperature (23-25°C).
MISO was administered i.p. 30 min before irradiation. Drug doses of 100 and 670 mg/kg were tested with solutions made up in saline at concentrations of 3 and 20 mg/ml respectively; the injected volume was then varied according to mouse weight. Rectal temperatures were measured after the first and last of 10 fractions in mice treated with X-rays alone or X-rays+MISO.
After irradiation, skin reactions were scored 3 times weekly for erythema and desquamation, according to a previously published scale (Stewart & Denekamp, 1977). Average skin reactions for each treatment (5 mice/group) were calculated over the period 10-32 days after a single dose or an equivalent period for the fractionated schedules. After fractionated irradiation the skin reactions developed 1-2 days later than after a single X-ray dose, and equivalent scoring periods (e.g. 12-34 days) were chosen by matching the leading edges of curves for reaction versus time (Denekamp. 1975). for skin irradiated with single doses in air or 02, either without drug or 30 min after administering 670 mg/kg MISO. A significant enhancement of the skin reaction was obtained if mice were irradiated in 02 instead of air (dose-modifying factor (DMF) = 1.2). The skin response was further enhanced by MISO to give a common Data are shown for X-rays alone (xx), 100 mg/kg MISO (0) or 670 mg 2kg MISO (5 ) given 30 min before irradiation. Sensitization was seen with single doses but not with 10 fractions. sensitivity whether in air or 02. The SER' values are therefore higher for animals in air than for those in 02 SERs from these data are quoted in Table I. Fig. 2 summarizes single-dose and fractionation schedules for irradiation in air. For all the schedules tested, MISO significantly enhanced the radiation response; SERs are given in Table I. The degree of sensitization for a particular drug dose was similar whether the radiation was given as a large single dose or as many small fractions. Fig. 3 shows data from similar singledose and 10-fraction experiments after irradiation in 02. MISO sensitized the skin to single doses of X-rays (left panel)  but not to 10 small fractions (right panel). When MISO is given before irradiation, radiosensitization is measured together with any cytotoxicity due to its metabolic reduction in hypoxic cells to a toxic product (Hall & Roizin-Towle, 1975;Sutherland, 1974). In order to measure this cytotoxicity separately from radiosensitization, the MISO can be given after irradiation (Denekamp, 1978). Fig. 4 illustrates data from "post effect" experiments in skin. MISO was given 5 min after irradiation with single doses or 10 fractions, to mice breathing air or 02. For irradiation in air (top panel) there was no evidence of cytotoxic killing in single doses, but the 10-fraction dose-response curve with MISO lies slightly to the left of the curve for X-rays alone (DMF = 1P05 + 0.03). This shift is not significant, but suggests either a small amount of cell killing or an interference with X-ray repair processes in the presence of MISO (see Discussion). Single-dose irradiation in 02 (Fig. 4, bottom panel) with MISO after X-rays produced a similar DMF (1 0 to 1l08 + 0.07) again not significant. For 0 fractions in 02 there was slight radioprotection when MISO was given after each irradiation.
MISO has been shown to cause a dosedependent decrease in the body temperature of rodents (Johnson et al., 1980;Hirst, personal communication). In the present experiments a reduction in rectal temperature was seeii over the period when mice would normally have been irradiated (i.e. up to 60 min after injection). MISO at 100 mg/kg caused only a small drop from 38 8 to 37-8°C, but 670 mg/kg MISO caused a more extensive fall in body temperature, to W34C by 45 min after a single injection, and similarly to 35°C after the last of 0 fractions. Thus the core temperature of MISO-treated mice was 1-5°C below normal body temperature at the time of irradiation, an(d the foot-skin temperatures (which were not measured) may have been even lower.

I)ISCU SSION
These data clearly demonstrate that the skin of unanaesthetized mice is sensitized by MISO, even at a drug dose as low as 100 mg/kg and radiation fractions as low as 5 Gy (Fig. 2). The data have been analysed to determine whether cells are at an intermediate 02 tension, or a small proportion of cells are at a very low 02 tension, e.g. those over a critical distance  sensitized by MISO. Fig. 5A shows the calculated effect of sensitizer on a mixed population of oxic and hypoxic cells (1% and 5% hypoxic fractions). Virtually no effect of the sensitizer is seen at X-ray doses below 10 Gy, because the response is dominated by the oxic cells. SER' values calculated from these hypothetical curves are plotted as a function of X-ray dose in the lower panels of Fig. 5 This approach is similar to that adopted by Hendry (1979) for comparing measured OER values for normal tissues treated in 02, anoxia or air. Fig. 6 summarises the SER' values derived from the data in Fig. 2 for mice irradiated in air. A wide range of X-ray dose per fraction can be covered, both for different levels of reaction within one fractionation scheme, and because single dose and fractionated data are available. It is clear that with 670 mg/kg MISO the SER' is constant over a wide range of X-ray dose. The SER' values do not vary significantly from 5 Gy to 32 Gy per fraction. The data from the low drug dose are more equivocal. Some sensitization is observed at 5-6 Gy per fraction, but there is a tendency to higher values at 25-32 Gy, though the error bars overlap.
This analysis demonstrates that the response of unanaesthetized mouse skin is more consistent with a uniform population of slightly hypoxic cells than with simply a very small fraction of severely hypoxic cells. Other workers (e.g. Dixon, 1967;Withers, 1967) have also concluded that murine normal tissues have a uniform level of mild hypoxia rather than a small proportion of severely hypoxic cells. This view is supported by a variety of studies which were reviewed by Hendry (1979), comparing anoxic, aerobic and fully oxygenated irradiations. The present data (Fig. 3) demonstrate some MISO sensitization even in 02-breathing mice; this indicates that pure 02 at atmospheric pressure does not fully sensitize all cells in mouse skin. Unexpectedly, 670 mg/kg MISO was more effective in sensitizing skin of air-breathing mice than changing the inspired gas from air to pure normobaric 02, despite MISO being less efficient than 02 on a molar basis (Adams, 1977;Suit et al., 1981). Sensitization was obtained with low and high doses of MISO for single X-ray treatments, but not with 10 fractions in 02. This would suggest that pure 02 increases the 02 tension in most of the skin cells, leaving a somewhat resistant subpopulation that can only be detected at high dose levels, 1*0-1-1 1 1-1 2 1*0-1-1 1-2-1-3 1-0-1-3 1-0-1-3 1-2 1-1 1-5 1-3 1X1 1-5 1 3 1-3 1-5 1 6-1 8 0.8-1.* 1.0 1.0 1*0 1.0 as in Fig. 5A. This could also explain the absence of MISO sensitization in skin clone experiments for irradiation in 02  where the dose range was < 20 Gy.
The experimental normal tissues in which MISO sensitization of X-ray damage has been tested in rodents are summarized in Table II. Three quarters of the studies show a small SER'. Published information for fractionated treatments in normal tissues is very sparse (Suit et al., 1981) but this information is clearly needed to interpret the relevance of mouse data to current clinical trials with repeated small X-ray doses. No MISO sensitization of human skin under ambient conditions was observed in the early studies of Dische et at. (1976). Similarly, no enhanced normal-tissue damage has been found in the first 200 patients treated with MISO at Mount Vernon Hospital (Dische et al., 1979). The only clinical trial in which enhanced normaltissue reactions have so far been reported is the Italian study (Arcangeli & Nervi, 1980)  sites. However, increased clinical skin and bowel reactions were found after irradiation in hyperbaric 02 (Dische, 1979;Henk et al., 1977) which would again be consistent with uniform mild hypoxia rather than a small fraction of acutely hypoxic cells.
The relevance of the present mouse results to the response of human tissues in patients undergoing radiotherapy clearly depends on whether tissue oxygenation in a rodent resembles that in humans. Mouse skin is thinner than human skin, and the hair follicles at least are known to derive some of their 02 from the ambient gas phase, rather than through the vasculature (Potten & Howard, 1969). Skin is also a major thermoregulatory tissue, and may be grossly influenced by ambient experimental conditions. However, the present results (Table I) and others in the literature (Hendry, 1979;Denekamp et al., 1981) indicate that many normal-tissue cells may be closer to radiobiological hypoxia than is often supposed. This might make them easy to protect against radiation injury, but it also means that the.y may be sensitized by radio-sensitizers like MISO. Such a uniform low 02 tension would imply that most of the cells in certain tissues are below the venous 02 tension, or that the 02 K value in vivo differs from that in vitro (Withers, 1967;Hendry, 1979;Denekamp et al., 1981). Sensitization of normal tissue by MISO should be taken into account in assessing its therapeutic value for experimental tumours. A therapeutic benefit will only exist if there is more sensitization in tumours than in normal tissues. Early experiments with skin irradiation in 02 indicated no significant MISO sensitization at doses below 25 Gy . We have therefore published therapeutic comparisons of MISO-treated tumours with those receiving X-rays alone on the basis of no skin-damage enhancement . These comparisons show a slight decrease in the benefit of MISO compared with the present sets of skin data. The single-dose sensitization in tumours is however much larger than for skin (e.g. SER' 1.7-2.4). Therefore there is still a big therapeutic gain for murine tumours treated with single doses of X-rays +MISO which is, however, less marked for fractionated treatments.
The present single-dose and fractionated data can be compared to investigate the effects of MISO on the repair capacity of mouse skin. A reduced ability to repair potentially lethal radiation damage has been demonstrated in vitro and in vivo with MISO given before or after irradiation of both oxic and hypoxic cells (Guichard et al., 1979;Nakatsugawa & Sugahara, 1980;Sakamoto & Aritake, 1981). Repair of radiation damage has been estimated from dose-response curves in the present series of experiments (Figs. 2 & 3) by comparing the doses required in single and fractionated treatments to give the same skin reaction. These experiments do not allow us to distinguish between repair of sublethal and potentially lethal damage (PLD). Thus we are measuring total repaired damage, including any PLD there might be. Repair increments (DN-Dl/N-1)* have been calculated for X-rays alone, MISO given before X-rays and MISO after X-rays. In Fig. 7 (Fowler et al., 1972(Fowler et al., , 1974Denekamp, 1973;Douglas & Fowler, 1976). The present data for X-rays alone, whether in air or 02, agree well with published values. MISO given before or after * D = X-ray dose; N = number of fractions. (see text). Mice sensitized by MIISO given be-fore X-rays had a reduced repair capacity, particularly at high X-ray doses. (In air: x, X-ray only. *, MISO pre X. *, MISO post X. In O2: +, X-ray only. 0, MISO pre X. C], MISO post X.) irradiation caused an apparent reduction in repair, particularly at high doses/ fraction (i.e. in the 2-fraction experiment with MISO given before irradiation in air).
The reduction in body temperature associated with high doses of MISO could influence the skin response in several ways. A reduced core temperature could lead to considerable peripheral vasoconstriction, which would limit the available 02; but this would also reduce 02 consumption and hence increase the 02 diffusion distances. These two opposing effects would respectively protect or sensitize the skin, but their relative magnitude is not known. If lower temperature in the 1MISO-treated mice is contributing to the apparent sensitization of mouse skin, it would imply that further 02 diffusion in hypothermic tissues predominates over the effects of vasoconstriction. This temperature effect is likely to be more important for skin than for other normal tissues.
Significant radiosensitization by MISO has been seen in mouse skin with both high and low doses of MISO, and with doses of radiation as low as 5 Gy. Very little cytotoxicity has been found, but there is a small decrease in repair of radiation injury in the presence of MISO. These results appear to support the thesis of Hendry (1979) and others that rodent normal tissues contain a large proportion of cells at a critical intermediate 02 tension which makes them marginally radioresistant.
WVe should like to thank Roche Products Ltd, Welwyn Garden City, for providing the misoniclazole. WN'e should also like to to thank Professor J. F. Fowler and Dr J. Hendry for their helpful criticism of this manuscript, Dr M. Joiner and Dr A. C. Begg for invaluable help with computer modelling for the hypoxic fraction estimates and the Cancer Research Campaign for financial support of this project.