Biological effects of radiation in combination with other
physical, chemical or biological agents 

Annex L from Ionizing Radiation: Sources and Biological Effects 

United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) 1982 Report to the General Assembly

United Nations, New York, 1982

II. Physical Agents — III. Chemical Agents

II. PHYSICAL AGENTS

ANNEX L   CONTENTS Paragraphs INTRODUCTION 1-20 I. MODES OF INTERACTION 21-72 General approach 21-41 Surface of response and isobolic diagrams 42-49 Probabilistic assessment of the interaction 50-61 Theory and practice 62-72 II. PHYSICAL AGENTS 73-113 A. Combinations of various types of ionizing radiation 73-78 B. UV and ionizing radiation 79-85 C. Electromagnetic and ionizing radiation 86-93 Experimental data 86-91 Epidemiological evidence 92-93 D. Suboptimal temperature and ionizing radiation 94-103 High temperature 94-99 Low temperature 100-103 E. Magnetic fields and ultrasound 104-107 F. Dusts and fibres 108-113 III. CHEMICAL AGENTS 114-199 A. Inorganic compounds 114-120 B. Organic radiosensitizing compounds 121-136 C. Carcinogenic chemicals 137-157 D. The special case of tobacco smoke 158-183 General 158-159 Experimental data 160-168 Epidemiological evidence 169-183 E. Other drugs 184-199 IV. BIOLOGICAL AGENTS 200-217 General 200-201 Hormones 202-213 Infectious agents 214-217 Viral infections 214-215 Bacterial infections 216-217 V. CONCLUSIONS 218-237 VI. RESEARCH NEEDS 238-244 Page References 765

A. COMBINATIONS OF VARIOUS TYPES OF IONIZING RADIATION

73. The simplest type of interaction, where most of the reservations raised in the preceding sections do not apply due to the similarity of the underlying mechanisms, is that between two different types of ionizing radiation. Mixtures of high- and low-LET radiations have repeatedly been tested for the presence of synergistic or inhibiting effects due to the combination of two beams, since current understanding of radiation action is not sufficiently advanced to allow prediction of possible interactions. In other studies external irradiation was combined with internal or the effects of mixtures of radionuclides were tested.
74. Studies on the combined action of fast neutrons, heavy ions and x rays were stimulated by possible radiotherapeutical applications [N2, N4, N5, B16, D13, F5]. Interaction of sublethal reparable lesions produced by neutrons and x rays was shown in experiments where x-irradiation was delivered at different intervals after neutrons [N4]. The actual survival curves of cells in vitro lay between those to be expected on the basis of iso- and hetero-addition. Similar experiments with results in the same direction were performed with neon ions [N5]. Cells irradiated with ions, incubated for three hours and then exposed to x rays showed a partial restoration of the shoulder of the survival curve. However, the results of Durand and Olive [D13] are in disagreement with those reported above because they showed no recovery after combinations of neutron-neutron, x ray-neutron and neutron-x-rays. It should be pointed out that these experiments were not confirmed. A theoretical description of the interaction of high- and low-LET radiation based on the theory of dual radiation action was provided by Zaider and Rossi [Z3]. Within the frame of definitions accepted in this report their interaction would be confined to the envelope of additivity.
75. Some insight into the nature of the underlying processes may be provided by studies of repair. When tested at the tissue level, the rate of recovery from sublethal damage appeared to be independent of the radiation causing it [H12]. It was suggested that recovery from sublethal damage does apply to the low-LET component of the damage, whatever the radiation producing this damage [G3, H13]. Naturally, in the case of neutrons which cause relatively more lethal than sublethal damage the final effect will not be clearly determined until one of the two components, the sublethal, has been repaired at sufficiently long fractionation times [H15]. Further evidence shows [H14, F3] that tissues treated with neutrons or with x rays to similar levels of biological damage and then submitted to an x ray course appear to be more radiosensitive when neutrons had been delivered in the conditioning treatment. Thus, the presence of different components in the LET spectrum and the presence of different types of damage to be repaired (sublethal, potentially lethal) each with characteristic time parameters make the picture rather complex.
76. New studies on combined radiation treatments were reported on Chinese hamster cells in culture irradiated first by neon ions (LET = 180 keV/µm) and subsequently by 225 kVp x rays. Cell survival was the end-point analysed [N9, N10, NI1]. The results for the three levels of survival presented in Figure XIV show that the experimental points fall in the middle of an envelope of additivity formed by application of hetero-addition (upper curve) and isoaddition (lower curve). When the order of application is reversed (Figure XV) the envelope of additivity is reduced to a line, a situation illustrated earlier in Figure III c and e. Under these circumstances one would conclude for the presence of synergism, if not for the fact that the actual levels of survival do not change after the two sequences of treatment. This apparent paradox is explained by the fact that radiation may be considered to be synergistic with itself when survival is not exponential with dose.

Figure XIV. Isobolic diagrams at three levels of survival for Chinese hamster V79 cells after irradiation with neon Ions followed by x rays [N10]

Figure XV. Isobolic diagrams at three levels of survival for Chinese hamster V79 cells after irradiation with x rays followed by neon ions [N1O]

77. The above examples suggest that in cases where the sequence of treatments does not change the final outcome, the isobolic diagrams should be constructed with the order of treatment that gives the greatest possible area of the envelope of additivity. As to the interaction of ions and x rays, this does not appear in general to exceed the isoaddition limits (interaction of x rays with itself) and only in few experiments true synergism may be suspected. Further work on the mechanisms at the cell kinetics and molecular level could clarify the precise conditions of the interactions.
78. Moskalev et al. [M21], modelling the effects of nuclear fallout, gave ¹³¹I orally (0.3 kBq/g) to rats and irradiated them at the same time externally with gamma (~ 6 Gy) or beta (surface dose ~24 Gy) radiation. Other animals were only irradiated externally. Lethality at 90 days was about five times lower in the combined treatment group, which was attributed to changes in the hormonal state in the course of acute radiation sickness. Other studies were also reported on the yield of mammary tumours in rats following ^131I and external irradiation with x or gamma rays [M22]. For low iodine exposure (0.04—0.08 kBq/g body weight) an increased yield of tumours was seen in the combined treatment group; for high iodine exposure the reverse was true. Other experiments [V4, V5] tested tumorigenesis in rats with ¹³¹I (3.7 kBq/g) and external thyroid irradiation (up to 3 Gy). No significant effects of the combination were reported. The interaction in these cases is attributed to hormonal disturbances, about which more data will be provided in chapter IV. Luz et al. [L14] reported on an enhanced osteosarcoma induction in mice by the joint action of two radionuclides, a short-lived alpha-emitter (²²⁷Th at 190 kBq/kg) and a beta emitter (227Ac at 1.9 kBq/kg). A higher than additive osteosarcoma incidence was reported at 700 days post-exposure, amounting to interaction factors of about 1.7 in terms of final tumour incidence and of about 1.3 in terms of the time for 50% tumour appearance, as compared to the effects of the two doses given individually. The authors attributed the interaction to the stimulation of osteogenic cell division due to the protracted action of the low level ²²⁷Th formed from ²²⁷Ac or to the continuing activation of a virus by the same cause. It is clear however that without kinetic studies and accurate physical dosimetry at the level of the sensitive bone cells it would be difficult to validate the phenomenon as belonging to the class of synergistic effects.

 

B. UV AND IONIZING RADIATION

79. The combined action of UV and ionizing radiation has been examined repeatedly in various experimental systems of micro-organisms [H10, Y4]. The experiments of Haynes in E. coli B/r [H10] may be examined as a good quantitative example. Pre-irradiation of the bacteria with different UV exposures increases the final slope of the x ray survival curve. Changing the sequence of the agents leads to a disappearance of the synergistic interaction, but only if the cells are irradiated in rich medium. If cells are irradiated in buffer the order of irradiation is not so important. This suggests that post-irradiation events may affect the interaction mechanisms. Many experimental data point to these events in relation to repair mechanisms [M24]. It appears that the repair of single-strand DNA breaks induced by x-rays may be inhibited by prior UV irradiation. For some recent reviews of DNA repair mechanisms and their genetic control see references [D10, M23, S34].
80. Experiments are also available on mammalian cells in culture [H16]. Synchronized Chinese hamster cells were irradiated in mid-S phase with fixed doses of UV and then exposed to graded doses of x rays (case I); alternatively, fixed doses of x rays were followed by graded doses of UV (case II). In case I the resultant survival curve may be obtained by isoaddition showing that, despite the different nature of the molecular lesions, the damage by UV is fully additive with that of x rays. The UV survival curves in case II are higher than the theoretical curves obtained by isoaddition, but lower than those obtained by hetero-addition. The size of the shoulder on the combined action curves in case II is less than that of the pure UV survival curve. Thus, the damage produced by x-ray pre-irradiation is only partially additive with the subsequent UV damage. In mammalian cells, according to these data, the situation of survival additivity seems to prevail. These data have been analysed by others [L28] according to the molecular theory of cell survival.
81. Transformation of mammalian cells in vitro is also a relevant end-point. DiPaolo and Donovan [D17] tested the morphological transformation of Syrian hamster cells with UV and x rays. Irradiation by UV alone (254 nm) gave a yield of transformants linearly increasing with dose. X-irradiation produced no trans-formation at all. X-irradiation (2.5 Gy) followed by UV (1.5 J/m²) at 24, 48 or 72 hours resulted in a greatly increased yield of transformants. An interaction factor of 3, 11 and 2.2 may be calculated at the above time intervals, showing the interaction to be very time dependent. Increasing the UV dose to 3 J/m² led to a decrease of the 48-hour interaction factor, thus showing its dose dependence. The relevant biological mechanisms remain unclear owing to the lack of under-standing of the phenomenon of transformation.
82. A synergistic interaction of UV and x rays was found by Holmberg and Jonasson [H22] for chromosomal aberrations in human lymphocytes. Go cells were irradiated first by 254 nm UV (5 - 10 J/m²) and then by scalar doses of 260 kVp x rays (1.25 - 2.0 Gy). UV alone gave a very small yield of dicentrics; UV followed by x rays doubled the yield of x rays alone. The interval between the treatments was less than half of a minute. Reversing the order of administration did not change the interaction factor of about 2. When phytohaemoagglutinin-stimulated cells entering stage GI were used in the same experiments no synergism was observed [H23].
83. Experiments with chronic exposure to UV light of different spectral composition and parallel chronic or acute exposure to ionizing radiation were also reported. They involved complex biological end-points such as LD₅₀ or life span. Galanin et al. [G2] studied in mice and guinea-pigs the haemopoietic functions and the life span under conditions of combined chronic irradiation by UV and gamma rays (dose rate 0.5 Gy/day). The experiments showed that animals receiving the combined treatment lived longer and had haematological values closer to normal than controls receiving only the gamma treatment. Acute damage was also influenced in the same favourable way by the combination of chronic UV irradiation with lethal and sublethal doses of gamma radiation [L15]. In experiments by Yatzula [Y3] on rats the treatment by UV preceded or was made concurrently with x-irradiation or internal irradiation by ³²P. Again, a decrease in the LD50,30 was seen after the joint treatment. Animals under combined irradiation had a better recovery of the body weight and showed less severe skin reactions. Physiological adaptation mechanisms could be invoked to explain such effects.
84. A comparison of the carcinogenic action on rat skin of UV and ionizing radiation, was made by Burns and Albert [B11]. The predominant tumor type observed following UV irradiation was a keratoacanthoma; after electron irradiation epidermal tumours were mostly seen. The yield of keratoacanthoma in rats irradiated at four weeks of age by different doses of ionizing radiation up to 30 Gy and then exposed for different periods to high and low fluences of UV was not influenced by the ionizing radiation dose and depended primarily on UV exposure. Absence of interaction was also seen in the case of epithelial skin tumours. Only one UV treatment schedule (high-fluence, 25.2 10⁴ J/m², from 5 to 16 weeks of age) enhanced the yield of epithelial tumours for lower doses (5.5 and 11 Gy) but not for higher doses of electrons (17 Gy). However, some of this increase was also observed in the zero-dose group and neither of these increases was statistically significant at P = 0.05. The absence of oncogenic interaction between the two radiations is a particularly good illustration of the fact that there may be a difference in the targets specific to the two radiations.
85. The examples reviewed in this section illustrate several important points. They show that the type of interaction depends on the biological end-point studied, on the level of exposure of the agents applied, on the order of their administration, on the stage of cell cycle, state of growth of the cells, etc. Under these circumstances it is not surprising that no general conclusion about the character of the UV and ionizing radiation interaction may be drawn.

C. ELECTROMAGNETIC AND IONIZING RADIATION

1. Experimental data

86. Many industrial,_, scientific, military and domestic appliances produce microwaves, electromagnetic radiation having frequencies of from approximately 10 to 10⁵ MHz. In some cases the same apparatus may produce very soft x radiation, as well as microwaves; in other instances ionizing radiation from other sources may be present in an occupational environment together with microwaves. The assessment of a possible combined action of these two agents is very difficult because exposure parameters for microwaves equivalent to the absorbed dose of ionizing radiation are absent [B23]. Even such a simple characteristic as the density of energy flux (DEF) is lacking in some experimental work. The quantitative expression and the underlying mechanisms of effects are far from clear. Differing views have been expressed on the nature of these mechanisms. Some authors consider the effects of microwaves to result from the dielectric heating of the tissues [M13]; others place the main importance on specific actions of the microwaves, particularly on the central nervous system [P9, G4]. A combination of these views has also been considered [B12].
87. In several early experiments [P18, M14, T3] the changes induced in the lethal action of radiation by microwaves were studied. The results consistently showed an antagonistic type of interaction on the lethality induced by ionizing radiation following pre-treatment of rats [P18], dogs [M14] and mice [T3] with microwaves. Later, interaction in the sense of additivity was reported for the same biological end-point [B13]. In more recent experiments Davydov et al. [D14] studied the lethality to mice after high exposure rates of micro-waves prior to acute gamma irradiation. Curves of the Rashevsky type describing the relationship between mean survival time and radiation dose were reported, and a shift to shorter survival times after combined treatment was observed. The animals were irradiated for 10 consecutive days with microwaves at a frequency of 2400 MHz with density of energy flux (DEF) 10, 20, 40 and 100 mW/cm², and exposure times of 40, 20, 10 and 4 minutes, respectively. It is interesting to point out that the degree of interaction was highest for DEF = 100 mW/cm² and depended rather on the intensive factor (DEF) than on the extensive one (energy administered). An approximately linear decrease in the LD50/30 with increasing DEF was observed. Extrapolation of this relationship to zero gamma dose would give a DEF value of about 325 mW/cm². The authors argue that at this level of DEF death might be brought about by microwave irradiation alone. If so, this would imply additivity of the two agents for a very complex end-point.
88. The state of the haemopoietic system of the animals in the course of the microwave pre-treatment described above was studied by Tichontchuk [T4] after 31 day irradiation at 100 mW/cm². Gamma radiation at 4 Gy was given after the last microwave treatment. The haematological parameters that were followed in the course of these experiments included the weight of spleen and thymus, and the number of cells in the circulating blood. An inhibitory action of the microwave irradiation alone on the haemopoietic system was noted, with a pronounced leukopaenia. The subsequent gamma treatment added further injury to the blood-forming organs, which could explain the additive type of interaction observed in [D14]. It is however fair to point out that other data on an inhibitory interaction of the same two agents have been reported in the literature [L19, F4, R8].
89. Rotkovska and Vacek [R8] exposed mice under conditions similar to those in [D14]. Lethality following low-LET radiation was again the experimental end-point tested. A definite decrease of lethality was observed if the mice were exposed after the x-ray treatment for five minutes to microwaves (2450 MHz, 1000 mW/cm²). Animals treated with microwaves showed also an increased number of haemopoietic stem-cells surviving and increased values of erythro- and myelo-poiesis. Differences of these results from those in [D14] could possibly be due to the reverse order of application of the interacting agents. A more recent paper by Rotkovska et al. [R14] provided further details of the therapeutic effect of microwaves on short-term mouse survival following whole-body x-irradiation and attributed the antagonistic effect of the two agents to an increased survival of the stem cells in the bone marrow.
90. For the purpose of the present document most interesting are the studies where clearly non-thermal, down to environmental, levels of microwaves were tested. Sakovskaya et al. [S25, S26] modelled in the animal a situation of chronic irradiation by microwaves and low-energy x rays. Female mice were irradiated in 31 or in 82 sessions, each delivered every second day and including 20 minutes microwave irradiation at DEF 2.5 and 5.0 mW/cm², and x irradiation (effective energy 10 keV) up to doses of 0.15 or 0.3 Gy. The end-points studied comprised body weight and weight of several organs (i); number of mice producing litters (ii); fertility (iii); weight of the litters at one month of age (iv); fraction of bone-marrow cells carrying chromosomal aberrations (v); lysozyme content of the blood serum (vi). Control groups and groups irradiated with only microwaves or x rays were also included in the experiments.
91. Statistically significant (P < 0.05) changes of the control values were obtained for the end-points (ii), (iii), (v) and (vi). A decrease of the lysozyme content of blood serum was observed, approximately to the same degree, for both the experimental groups irradiated with microwaves alone or with x rays alone. In the group receiving the combined irradiation the decrease of the enzyme was slightly greater, but the corresponding point on an isobolic diagram would fall into an envelope of additivity. For the chromosomal aberrations again an increased yield was seen in the groups receiving the separate treatment and again to approximately the same extent. The group under combined treatment showed a slightly higher incidence of aberrations, but the corresponding experimental points were not outside the envelope of additivity. Microwaves alone seemed to show a slight stimulating action on the fertility of the mice and on the fraction of mice producing litters; but in the combined treatment group a decrease of both end-points was observed. An additive type of interaction would appear more likely to apply in general to these experiments.

 

2. Epidemiological evidence

92. The combination of ionizing radiation and high-or low-frequency electromagnetic radiation is characteristic for a number of occupational environments in electronic and radiotechnical plants. Increased ambient temperature, constant electric or magnetic fields, sound pollution and vibration may also be part of these environments [02]. Wolfovskaya et al. [W3] studied the health of female workers employed in assembling, testing and vacuum pumping of high-voltage electronic equipment. They were exposed to electromagnetic fields of different frequencies (electric field strength 600–2500 V/m and magnetic field strength 50–320 aim) and to x irradiation at dose rates of up to 25 µGy/hour. The effects studied included frequency of functional disturbances of the nervous system, blood pressure, dysmenorrhoea, changes in the sedimentation rate of erythrocytes, thrombo- and leuko-cytopenia. It was claimed that the important determinant of the symptoms was x irradiation. However, a high percentage of disturbances of the nervous system found among these workers was attributed to microwave exposure. The nature of the end-points, their variability between groups and the lack of any precise dosimetry and statistical analysis make it difficult to validate such conclusions.
93. Similar comments may be made with respect to other epidemiological studies on the clinical effects of combined exposure to microwaves and ionizing radiation. A survey was reported of workers testing microwave generators [B9]. Four groups of people (200 subjects in total) were included in the survey: two groups worked under combined exposure to micro-waves and ionizing radiation; one was exposed to microwaves only and the last one was the non-exposed control. A group exposed to gamma radiation alone was not included in the study. Asthenia and migraine were characteristic complaints in all exposed groups. For the first two of them the symptoms were 20–50% more frequent than for the third group and between 2 and 2.5 times more frequent than in the fourth. Dysfunctions of the autonomic nervous system were 2 times more common in the groups with combined exposure than in the third group and 3–4 times higher than in controls. The combined action of ionizing radiation and microwaves was also investigated in workers by Lysina [L17] but loosely defined conditions of exposure render any judgement of the type of inter-action impossible.

D. SUBOPTIMAL TEMPERATURE AND IONIZING RADIATION

1. High temperature

94. Broad quantitative studies on the effect of heat on cells and the interaction between heat and ionizing radiation started in the 1960s, stimulated by the possible application in the treatment of cancer. Several conferences and symposia have by now taken place on this subject [C17, C18, P12, D20] and good reviews are available [D18, F2]. A full discussion of this subject is beyond the scope of this Annex, which will only briefly cover the most basic aspects. Heat alone may damage mammalian cells and tissues at temperatures of 42°C given for a sufficiently long time [F8]. Thermal inactivation curves as a function of the temperature or of the treatment time at a given temperature may be produced, having characteristics similar to those of the radiation inactivation curves. There are reasons to consider that the target for cell killing by heat may be plasma membranes [D18] but other targets such as lysosomal membranes or macromolecules cannot be excluded.
95. Treatment of cell cultures with heat increases their sensitivity to radiation in the sense that the final slope of the x ray survival curves becomes steeper after pre-heating [F8]. Thermal enhancement ratios as defined by equation (7) may be used to quantify the effect and for different cell lines the values of this ratio seem to correlate well with the sensitivity of the cells to heat [R9]. After pre-heating for 1 hour at 42.5°C the TER may reach values higher than 2 [R9]. These values seem to increase for irradiation at low dose rates [B24] because recovery from sublethal damage is sharply reduced by treatment with heat [L18]. A delay of rejoining of strand breaks in DNA [C19] and inhibition of DNA synthesis, including repair synthesis [S37], were observed after heat treatment. The targets for enhancement of radiation sensitivity by heat are different from the targets for simple heat inactivation and include all the repair systems and the chromatin [W10, D18, D26, S49].
96. The temporal pattern of treatment is very important for the synergistic interaction of radiation and heat [S35, D19, S36, 03]. Maximum interaction is usually observed with the simultaneous presence of the two agents and it declines as the interval between treatments increases. Figure XVI [F2] illustrates the time course of the decay of heat potentiation of x-ray damage in a variety of normal mammalian tissues for hyperthermia given either before or after irradiation.

 

Figure XVI. The time course of the decay of heat potentiation of x-ray damage in normal tissues for hyperthermia given at different times before or after irradiation. The tissue responses are normalized to the percentage of the maximum response which occurs, for each curve, at the shortest time intervals. Data from [F2]



 

97. Data on the life span of animals irradiated for the duration of their life under conditions of high environ-mental temperature are reviewed in Annex K. A study with pre-implantation mouse embryos exposed in vitro to 39°C immediately after x irradiation [V8] showed a great increase in the number of micronuclei in the cultured cells, indicating an enhanced chromosomal damage after combined treatment.
98. For animals, and for mammals in particular, it is difficult to foresee under what conditions a direct sensitizing action of high temperature in the environment might come about, since normally healthy mammals maintain a fine regulation of their body temperature. However, Dobrovolsky [D11] reported experiments on rats where the effects of chronic irradiation due to the daily intake of ³⁵S (0.55 MBq/kg), ⁴⁵Ca (1.3 and 2.8 MBq/kg) and ³²P (0.37 MBq/kg) in the course of a year were studied in combination with daily exposure for two hours to a temperature of 400C. Survival, body weight, fertility (mean number of litters per female), haematological parameters and histology of the ovaries were the end-points studied. During the first period of treatment the changes characteristic of a chronic radiation injury appeared to be aggravated by the combined treatments. During the second half of the treatment, however, the combination of irradiation and high temperature appeared to increase and accelerate repair processes. Fertility of the female animals depended on the mating time, but in general the combined action group had enhanced fertility in comparison with the control groups [D12].
99. Epidemiological investigations of combined actions were made on workers at metallurgical plants who were exposed to ionizing radiation and also received periodically high temperature exposure [M12]. Primary functional disturbances of the nervous system were seen with higher frequency in this group as compared to other groups of workers. People having a shorter occupational history were reported to show vascular dysfunctions attributable to dystonicity of the autonomic nervous system. For longer occupational times asthenia also accompanied the above symptoms in a higher percentage of workers. Owing to the obvious difficulties in the quantification of such subjective symptoms any judgement about this type of interaction should be reserved.

2. Low temperature

100. Cold-blooded animals are good experimental material for studies of the influence of low temperature on radiation sensitivity. Much relevant information was published in a specialized symposium [RIO]. In very general terms, regeneration of tissues [H24], lethality after fractionated irradiation [E7] and recovery from radiation injury in self-renewing tissues [E8, E9] in fish are considerably inhibited when the animals are kept at suboptimal temperatures. A detailed discussion of these data is beyond the scope of this report which is essentially centred on mammalian systems and on end-points of practical importance for man.
101. Trujillo et al. [T5] reported that RF/Un female mice showed a linear decrease of their ability to withstand a standard cold stress (6°C – 7°C for 14 days), as a function of increasing age. Mice exposed to protracted ⁶⁰Co gamma exposures at 0.5 Gy/day and then allowed to recover for 90 days showed a similar linear decrease with increasing radiation exposure in their ability to survive the same stress. This radiation induced effect was considered similar to life shortening through natural aging and was estimated to be equivalent to 9.3 days/Gy. Other data on the combined effect of duration of life irradiation in animals under conditions of low ambient temperature are discussed in Annex K.
102. In experiments by Gambino et al. [G5] rats were irradiated whole-body or on the adrenals only with a standard exposure of 5 Gy and then were exposed for 3 hours daily to 0°C. Reduced longevity, growth retardation, cataract, greying of the fur, and induction of tumours were the long-term effects seen in animals that had been whole-body irradiated, while animals irradiated only on the adrenals did not show such phenomena. The treatment at low temperature did not modify the incidence of these effects, with the exception of a slight reduction of the accelerated onset of tumours seen in whole-body irradiated animals. Since the treatment with low temperature as such gave rise to a reduction of the life span and to differential effects in the incidence of inflammatory and neoplastic conditions, the experiments are not easily interpreted [H17].
103. In many of the experiments described the temperature could not act as such, but as a condition producing physiological adaptive changes. Some information on the influence of miscellaneous physical treatments (permanent or transitory high altitude, high or low ambient temperature, mechanical damage, severe metabolic or physical stress) in respect to tumour induction in animals were already reviewed by the Committee in its 1977 report (Annex I) [U1]. The findings were on the whole negative. When interaction effects were reported they were not very large and explanations in terms of physiological adaptation mechanisms to the exposure conditions could readily be produced. In respect to life shortening, which at the low doses and dose rates of interest in radiation protection is mostly associated with tumour induction, some data are reported in Annex K. They concern low and high environmental temperature and specific and non-specific stress. Here again, the effects reported were marginal and often of antagonistic character, i.e., leading to an increased life span by the joint treatments. These effects could be explained on the ground that suboptimal living conditions frequently act by decreasing, rather than by enhancing, the susceptibility of the animals to the effects of radiation. However, even though the impression is in favour of the lack of positive synergistic evidence, the data are few, the effects unspecific and the underlying mechanisms obscure so that no definitive statement can be made.

 

E. MAGNETIC FIELDS AND ULTRASOUND

104. A fairly extensive body of literature exists on the effects of magnetic fields in biological systems [P14] but studies of their combined action with radiation are relatively few. A review is to be found in [N6]. This problem may conceivably be of practical significance for workers in thermonuclear fusion devices. It should also be recalled that the use of transversal magnetic fields to improve the dose distribution of high-energy electrons in radiation therapy has recently been envisaged.
105. Sikov [S38] tested various combinations of high-intensity magnetic fields with gamma-irradiation in mice. Of the various end-points considered (lethality, developmental changes, biochemical effects) only two appeared to be susceptible to the action of magnetic fields applied alone (2 to 4 10⁸ Tesla, T): audiogenic seizure and the level of tryptophan pyrrolase in liver. In both cases radiation alone had little effect and the results of the combined treatments could be attributed to the action of the magnetic field. A decrease in the slope of the probit line of mortality (gamma rays, 5.8, 7.5, 8.6 and 10 Gy) without change of the LD₅₀ value following the contemporaneous exposure to the 4 10⁸T field indicated a decrease of spectrum of radiosensitivity values induced by the joint treatment. Fields of 2 10⁸ T were inactive to this end.
106. Although some indications of synergism were reported for biochemical indices following localized liver irradiation [W11], experiments on the survival of cell cultures in vitro were negative in this respect [R11, N6]. Uniform magnetic fields of 1.4 10⁷ T in combination with radiation produced no changes in the form of the survival curves of cells in vitro or in the pattern of recovery from sublethal damage, as compared with radiation alone [R11]. Higher intensities of the magnetic field (2 10⁸ T) or non-uniform fields were also without effect for similar end-points in other experiments [N6]. There is therefore on the basis of presently available . evidence little ground to expect an enhancement of the effects of radiation by the joint application with magnetic fields.
107. Ultrasound is widely used for diagnostic and therapeutic purposes, as well as in many industrial appliances. Some experiments considered its possible interaction with ionizing radiation. Harkanyi et al. [H18] irradiated mice by ultrasound (800 kHz, exposures of 0.1, 0.5 and 1.0 W/cm²) followed two hours later by 0.5 Gy of x rays. Single-treatment groups were also set up at the same time. The yield of chromosomal aberrations in the bone marrow of the animals was taken as the end-point. None of the ultrasound exposures produced any significant increase over the spontaneous level, while the effect of the ionizing radiation dose was easily assessed. No change in this level of effect was found in the group of mice under-going the combined treatment.

 

F. DUSTS AND FIBRES

108. For many industrial environments the combination of radiation exposure and exposure to dusts is quite usual, as, for example, in mining, metallurgical industries, power plants and construction works. Many dusts and fibres have been shown to be carcinogenic or pathogenic by themselves. Direct experiments on mammals about the action of dusts are available [C3, C16, K9, P2, P10, P11]; concerning fibres, asbestos and other minerals have been given particular attention [W12, B25]. Since dusts may be soluble or insoluble, according to the different types of materials, studies of their combined action with radiation could be covered under the chemical or under the physical section, respectively. In the first instance the chemical compounds dissolved from dust particles would be the actual agents taking place in any combined action; in the latter the size and the distribution of the dust particles would be the parameters of relevance.
109. Panov et al. [P10] studied the respiratory and renal systems of rats after intra-tracheal instillation of a neutral ²¹⁰Po solution (37 kBq/rat) and quartz dust (50 mg in saline suspension). Lung fibrosis was found to be more pronounced in the combined treatment group. Malignant tumours of the respiratory tract were also said to be observed more frequently in this group, although no precise description of all the histological and statistical aspects of these tumours was presented in the work. Similarly, in kidneys glomerulo-tubular lesions were found more often in the group under the combined action of the two agents.
110. Ponomareva et al. [P11, P2] used different types of mineral dust with admixture of highly active thorium oxide. Rats were made to inhale or were instilled intra-tracheally for periods of time up to one year. The chronic action of these agents gave rise to inflammatory lung processes and to fibrosis. Tumours of the lung were also observed after 1.5 to 2 years. When an additional chronic whole-body irradiation course was given to the animals (gamma rays, 20 mGy/day, total dose 2.5 Gy) the lung tumour yield was increased by a factor of two, in comparison with the group under combined treatment and a group receiving external irradiation only. There were no experiments performed to define the specific role of dust in combination with internal or external radiation treatment. The data obtained from the group combining internal irradiation and dust were used to standardize conditions of occupational exposure including a combination of these agents [B15].
111. Experiments on the combined action of internal alpha irradiation (²³⁹PuO2) and chrysotile asbestos fibres (mean fibre length 1–10 µm) were performed by Sanders [S12, S24, S13]. Insoluble particles were administered to rats by intra-tracheal instillation. Animals receiving only the PuO₂ had a more homogeneous distribution of plutonium particles in the lung, while the combination of treatments led to a concentration of the radioactive particles within the asbestos-induced scars in the peribronchiolar regions of the lung. In groups receiving plutonium alone the pulmonary retention half-time of the nuclide was about 200 days; in the combined-treatment group it was 450 days. Correspondingly, the cumulative absorbed doses to the lung two years after instillation were 4 and 12 Gy. The incidence of pulmonary carcinoma was 4.5% in rats given the asbestos, 32% in rats receiving plutonium alone and 21% in the combined-treatment group. Thus, per Gy of absorbed dose, the incidence was about four times greater in the plutonium group than in the combined-treatment group. An explanation for the finding could be that by a reduction of the number of epithelial cells receiving alpha dose a reduction of the resulting yield of tumours could come about. In another series [S12] the two agents were injected intra-abdominally. The agents both tended to concentrate in the fibrous adhesions of the peritoneum and the omentum, inducing sarcomas and mesotheliomas to a final incidence which was not appreciably different from an expected sum of effects.
112. Lafuma et al. [L16] reported preliminary results of experiments with rats where internal or external irradiation were combined with intrapleural injection of crysotile asbestos. In a first series 8 rats were exposed to 3000 WLM (see definition of WLM in Annex D) of radon-222 over 1 month and they received about 70 days after the beginning of exposure, 2 mg of crysotile in suspension intrapleurally. As in the case of previous experiments with radon inhalation [L8] a very small proportion of animals developed lung tumours after radiation exposure and no mesotheliomas were observed at all. Exposure to crysotile only resulted in a very low incidence of mesotheliomas. However, the combined treatment led to the appearance of lung cancer in all rats, 7 of them being mesotheliomas. A clear synergism is here obtained. The same type of results was obtained in a second experimental series where whole-body mixed gamma-neutron reactor irradiation was given (2.3 Gy of 0.5 MeV neutrons with a gamma component of 0.75 Gy). The animals were injected with the same amount of crysotile 125 days after the radiation exposure. The results on lung tumor induction are given in Table 1 and show that, in addition to an increase in total tumours, mesotheliomas only appear in the irradiated group given crysotile intrapleurally. These preliminary data should be confirmed in larger experiments.
113. Sanders et al. [S22] studied the effects of beryllium oxide aerosol inhalation in combination with plutonium oxide aerosol on more than 600 rats. Aerosol particles were of micron and submicron sizes. Exposures up to initial alveolar depositions of 1 to 91 µg beryllium and 0.15 to 6.7 kBq of ²³⁹Pu were performed. The results obtained by the two agents given separately and by their combination (beryllium aerosol being introduced prior to plutonium aerosol) as total incidence of pulmonary tumours show that the changes in lung. tumour incidence due to the combination of the agents were insignificant. This in spite of the fact that the alveolar clearance of plutonium was decreased by exposure to beryllium and the translocation of plutonium to the thoracic nodes was increased.

 

III. CHEMICAL AGENTS

A. INORGANIC COMPOUNDS

114. Changes in the physical and chemical characteristics of the water matrix of biological systems may bring about changes in radiosensitivity. Chinese hamster cells were exposed to media containing deuterium oxide (D₂O) following ⁶⁰Co gamma irradiation and cell survival was scored as the end-point [B17]. Under these conditions the cell response to radiation was greatly enhanced. Depending on the concentration and the treatment time of D₂O, dose modification factors of up to 4.5 could be found. Pre-irradiation incubation had, on the contrary, a very slight effect on the radiation response. The sensitizing effect of D₂O depended clearly on the conditions of cell metabolism, since it was influenced by the type of media and by the temperature. It was found that the radiation damage capable of interacting with the deuterium oxide was repaired by the cells when they were kept for three hours at 37°C in the growth medium and split-dose experiments suggested that the sublethal damage repair capacity was reduced in the presence of D₂O. The heat sensitivity of the cells was unaffected by D₂O and the enhancement of radiation response induced by heat was also independent of the presence of D₂O.
115. Some natural mineral components of the diet may change the radiation response of the animals [K10]. Rats were kept on diets with low (50 mg/d Ca and 0.2 mg/d F) or high (150 mg/d Ca and 3 mg/d F) content of calcium and fluorine and after 5 weeks of such diet were given radioactive ⁹⁰Sr. As a result of the combined treatments the haemopoietic system of the first group of animals was more severely damaged and their mean life span shortened by 50–70 days, as compared with the group with high Ca and F in the diet. The protective action of the high Ca and F diet is achieved at intakes of the two minerals not higher than the upper limits of physiological intake for humans. Similar results were obtained if external gamma irradiation was added to the internal ⁹⁰Sr irradiation. In other experiments rats were subjected only to gamma irradiation and to diet changes. In all cases the survival at short term and the life span proved to be higher in groups with high calcium and fluorine intakes.
116. The different trace metals found in the air, food and water of some parts of the industrialized world [T7] may alone induce adverse health effects, including malignancies and teratological effects, at sufficiently high concentrations. They may also conceivably combine with the action of ionizing radiation at the background level or under special conditions of exposure. The universal spread of these metallic contaminants make studies of their possible combined action particularly important. Data on the combined action of silver ions and radiation in bacterial systems (spore or vegetative stage) have been provided by Richmond and Powers [R12] and Held and Powers [H25].
117. Lead chloride (PbC1₂) in concentrations of 0.1 and 1 µg/cm³ was studied in combination with radiation (doses of approximately 1 Gy) for its ability to induce various effects in vitro on embryonic systems [S15]. The number of nucleated cells per mouse embryo, the labelling and mitotic indices and the number of micronuclei per cell were among the effects scored. At both concentrations a synergistic increase of the micronuclei was found, accompanied by an inhibition of embryonic development. For cadmium, the combined effects with radiation were found to be additive in the same system [M26]. Lead was studied by Kudrizkaya [K3] for its capacity to damage spermatogenesis in the mouse. Exposure was given chronically over a period of about six months up to cumulated concentrations of 0.3 mg/g of lead chloride and 81 kBq/g of ⁹⁰Sr, administered in drinking water. Testis weight or the number of spermatocytes were unaffected by lead alone, while ⁹⁰Sr significantly decreased the control values of both end-points. Combination of the treatments produced a final effect which was lower than that caused by radiation alone, an antagonistic type of interaction. Lappenbush [L25] injected adult male rats with cadmium chloride (125 to 250 mg) intraperitoneally for 30 days twice per week and subsequently irradiated them with x rays. The 60-day survival was unaffected by doses of the contaminant lower than 125 µg. The radiation LD_50/30 was found to decrease linearly with increasing exposure to cadmium. The numbers of red and white cells in the peripheral blood were affected by the combined treatment in a complex way.
118. Platinum (cis-dichloro-bis platinum, DBCP) and radiation affected the survival of ovarian-derived Chinese hamster cells in culture according to a synergistic type of interaction [C8]. Chromatide aberrations were also induced in higher percentages. In order to observe synergism the chemical had to be administered between four hours before and two hours after irradiation. Two or three days elapsing between the chemical and the radiation treatment abolished the interaction. The synergistic effect was considered to result from radiation-induced single-strand breaks in the DNA which occurred in linear proportion to dose, opposite to a single platinum complex intra-strand cross-link which occurred linearly with respect to platinum concentration. The combination of the two lesions led to lethality. A simple mathematical model to describe the experimental data was developed.
119. The nitrocompounds, especially the oxides, are rather common pollutants of the air. Sensitization of anoxic bacterial spores was reported when they were irradiated in NO₂-saturated water [P15]. A study is available in mammals [K4] where inhaled plutonium-239 under the form of plutonium pentacarbonate ammonium (69 kBq/kg of lung tissue) was administered to rats, after which the animals were also made to inhale nitrogen oxide (0.09 mg/1) or chlorine (0.05 mg/1) for 15 minutes. After the combined treatments the incidence of lung cancer was almost doubled as compared with the irradiation treatment alone. Tumours were multifocal and different types of tumours were seen in the combined than in the single treatment. Pneumosclerosis was also enhanced in the combined treatment group.
120. Occupational situations where exposure to ionizing radiation may be accompanied by exposure to other detrimental chemicals should not be uncommon in industrial practice, but epidemiological data in this field are very rare. In one case observations were carried out on workers exposed to gamma rays for industrial radiography and also to vapours of hydrofluoric acid (HF) [S21]. The changes investigated (levels of T-lymphocytes, C-reactive protein and auto-antibodies) were mainly immunological. The group under the combined influence of radiation and the toxic chemical was reported to have lower levels of T-lymphocytes and higher levels of C-reactive protein and auto-antibodies than the groups exposed to only one of the agents.

B. ORGANIC RADIOSENSITIZING COMPOUNDS

121. The present section includes what is essentially a review of substances which may enhance the radiation response of biological systems, and are called radiosensitizers. Compounds inhibiting the radiation response are called radioprotectors. In many cases these substances were specifically developed for their protective or sensitizing properties. The study of radioprotective chemicals has been strongly pursued [R2, M7, M29, B29]. More recently, the application of such compounds in clinical tumour therapy has been discussed [Y7]. Also, a new field has grown and is still rapidly expanding, that of the radiosensitizing compounds [M10, R6], whose potential in clinical radiotherapy is being tested.
122. The relevant data will be reviewed briefly because it seems unlikely that situations will arise in which these substances may pose significant problems of public or occupational health. For a review of biological effects, mostly lethal, of the combined action of acute irradiation with other common industrial poisons (at high toxic levels) the reader is referred to Tiunov et al. [T2]. Annexes J and I of the 1977 report [UI] reviewed the action of radioprotective and radio-sensitizing chemicals in respect to the production of embryonic and foetal damage by radiation and of tumour induction, respectively. The available information on the action of chemical radioprotective drugs for life-shortening effects in animals is reviewed in Annex K of this report.
123. Several classifications of radiosensitizing substances have been proposed [MIO, P16, S39], based on their mechanisms of action. Keeping in mind that in some cases the molecular mechanisms are still unknown and that some agents may act through more than one mechanism, one classification may be as follows: 1. Agents modifying the primary radiation chemical processes, including (a) electroaffinic agents and (b) iodine compounds; 2. Agents interacting with DNA metabolism (DNA-base analogues); 3. Antibiotics and other agents interfering with repair processes (see section III C); 4. Agents reacting with nucleophilic groups (SH groups); 5. Other radiosensitizing agents.
124. The best known example of the first class of agents is oxygen whose level in biological systems at the time of irradiation greatly influences the yield of radiation effects. A massive body of literature exists on the action of oxygen and the interested reader is referred to [A11, P1]. A large number of electroaffinic compounds or hypoxic cell sensitizers is also known but their detailed discussion is beyond the scope of this Annex. Such compounds may contain one of the following chemical groups: the carbonyl (CO), the aldehyde (CHO), the nitro (NO₂), the cyano (CN) groups, homo- and hetero-cyclic rings. Stable free radicals are also electroaffinic agents. The radiosensitive properties of such compounds are manifest when they are present in biological systems at the time of irradiation or if they are irradiated separately and then immediately added to the biological system.
125. The same is true of the iodine compounds which may also change the concentration of radiation-induced free radicals. If cells are exposed to irradiated iodoacetamide within milliseconds after irradiation cell killing takes place, which is not observed if irradiated cells are exposed to non-irradiated iodoacetamide, thus showing the role in sensitization of short-lived transient compounds [D9]. Radiosensitization takes place also with other iodine compounds: iodide, iodoacetic acid, iodopropionic acid, methyliodide, p-iodophenol, iodobenzoic acid and others. Reactions with -SH groups may account for part of the sensitizing effect of some of these compounds [M7].
126. Attention has recently been given to the radiosensitizing properties of iodine contrast media used in radiodiagnostics [S43, N7, A2, M25]. Sensitizing effects on bacterial killing were first reported [S43] and then an increased yield of chromosomal aberrations in peripheral lymphoctyes of children undergoing x-ray angiocardiography with contrast media [A2, N7]. It has also been reported that sensitization of mammalian cell killing by iodine compounds would occur for x but not for gamma rays [M25]. These data are explained by the difference in doses due to photoelectric effect in the case of x rays. An accurate physical dosimetry should clarify this issue.
127. Quinones are unsaturated carbonyl compounds with conjugated structures and electron affinic properties. Several quinones and their derivatives have been found to sensitize bacterial and yeast cells under oxygenated and anoxic conditions [A4, MII, S16, S17]. It has been postulated that the sensitization of E.coli B/r by vitamin K5 is mediated by radiolytically produced hydroxyl radicals [S16]. Diphenylquinone was found to enhance the action of radiation in mice [A4]. In some bacterial systems under anoxia the value of DMF could be about 3 (10-³ M indanetrion monohydrate [B7]) or even 4 (100 ppm vitamin K5, [S16]). Newly synthesized isoindole quinones showed promising characteristics when tested in vivo on soft tissue sarcomas transplanted into mice [C13].
128. Electroaffinic compounds containing nitro groups can specifically increase the radiosensitivity of anoxic cells, leaving that of oxygenated cells unchanged or even decreased. These properties would be advantageous for tumour radiotherapy [A5, D8, H4, H8, P17, Y5]. The radiosensitiziation by mizonidazole was proved to occur for hypoxic mammalian cells in vitro and in vivo [A12]. Under aerobic conditions no sensitizing effect of the compound at any stage of the cell cycle was observed [P17] and under anoxia the strongest effect occurred in middle-S. Toxicity of the agent under anoxia requires low exposures to the agent.
129. Yuhas and Li [Y5] studied the effects of the compound at a concentration of 6 mM in combination with the radioprotective compound cysteine (8 mM) on mammalian cells in culture, showing protection under conditions of oxygenated irradiation and sensitization under anoxia. Hall et al. [H8] tested eight different nitro compounds: for all of them the DMF was an increasing function of the concentration and for some it reached a value of about 3.5, equalling the average value of the OER in the cells tested. In general, 2-nitroimidazoles were more effective sensitizers than 5-nitroimidazoles. Other nitrocompounds, the nitrofurans, may be even more effective, specifically under anoxia [R3, R6]. Sensitization by nitrocompounds was greater when they were administered prior to irradiation [D8]. Radiosensitizing properties were also described for nitrogen-containing stable free radicals such as triacetoneamide-N-oxyl (TAN) [ES, B21] and 2,2,6,6,-tetramethyl-4-piperidinol-N-oxyl (TMPN) [P6].
130. DNA base analogues belong to the second class of radiosensitizers. Extensive studies were made especially on halogenated DNA base analogues such as 5-fluorouracyl (5-FU), 5-bromouracyl (5-BU) or 5-bromo-2-deoxyuridine (5-BUdR) [K13]. Significant enhancement of killing was shown for viruses, bacterial and mammalian cells [S18]. Some attempts for a clinical application of these substances have also been reported. For a review of the relevant studies see [M7, M10].
131. The third class of radiosensitizers will be considered in section III.C. Here various substances capable of modifying the biochemical cellular processes should be mentioned, belonging to classes 4 and 5. Several organic chemicals capable of enhancing radiation damage share the property of being -SH reactive. Since -SH compounds are known to be radio-protectors, the correlation has been investigated between the ability to bind -SH groups and the capacity to sensitize the cells to the action of radiation. Bruce et al. [B8] found that the capacity to sensitize was well correlated to the amount of p-hydroxymercuribenzoate bound to cells.
132. Sensitization of anoxic cells is an important goal for tumour radiotherapy [A13, R6]. Radiosensitization of bacterial cells under anoxia by N-ethyl-maleimide (NEM) was shown as early as 1960 by Bridges [B6]. Other data on bacterial and mammalian cells are also available [L12, M18, K11]. A DMF of 1.5 with human cells in vitro irradiated with x rays was reported by Klimek [K11]. The known property of NEM to bind -SH groups led to the hypothesis [L12] that NEM could bind the free non-protein -SH groups, thus preventing DNA repair through donation of hydrogen from these groups. Other experiments by Klimek and Zemanova [K12] showed that under concentrations of NEM too low to inhibit DNA synthesis a high proportion of the original free thiol groups was still present, thus implying other mechanisms for NEM sensitization. However, the role of intracellular thiol groups would also be supported by experiments of Sinclair on oxygenated [S40] and anoxic [K17] Chinese hamster cells exposed to NEM and radiation. Repair of lethal damage is inhibited by the presence of NEM but the mechanisms of such an inhibition are still unknown.
133. Another organic compound that may produce cytological changes is carbon tetrachloride (CC1₄). Its administration to animals induces, for example, liver cell proliferation [A6] similar to that induced by partial hepatectomy. Cole and Nowell [C20] examined the effect of CC1₄ on the induction of hepatomas in fast neutron irradiated mice with doses of 1.7 to 3.1 Gy. At various times after irradiation some animals received the compound subcutaneously. Sixty-one percent of animals receiving the combined treatment developed hepatomas, as compared to 19% of the mice irradiated only. Since CC1₄ alone produced no hepatomas, the interaction factor is approximately 3. Histologically the tumours were similar in both groups but tumours of larger size were more frequent in the combined modalities group. The authors concluded for a promoting effect of CC1₄ in liver cancerogenesis. Procaine hydrochloride, a local anaesthetic acting on cell membranes, has been shown to sensitize bacterial and mammalian cells to the action of radiation [S19, S20].
134. Alkylating agents may react with DNA bases and thus directly influence the radiosensitivity of cells. The alkylating agent spirohydantoin mustard (SHM) was tested in combination with x-irradiation on brain tumour cells in vitro [D15]. The enhancement of cell killing was greatest when the cells were irradiated four hours before the drug treatment. The doses ranged from zero to 20 Gy and the subsequent chemical treatment with SHM lasted one hour at concentrations of 0, 2, 3, 4 and 5 µg/ml. The results were normalized and the corresponding isobolic diagrams were built. At levels of cell killing down to 10% a synergistic interaction was apparent, although for lower levels of survival down to 0.1% the interaction turned into an additive one. This and another paper [D2] by the same authors are some of the rare examples where the analysis of the interaction type was carried out according to the approach outlined in chapter I of this Annex, involving the use of isobolic diagrams.
135. The same brain tumor cells cultured in vitro were exposed for one hour to 1, 3, 5, 7.5 µg/ml of 1,3-bis(chloroethyl)-1-nitrosourea (BCNU), followed 15 hours later by a series of x-ray doses of up to 20 Gy [D2]. Survival curves for the x rays alone, the BCNU alone and for the combination of both agents were obtained and on their basis isobolic diagrams for survival levels of 1, 2 and 3 log cell kill were constructed as in Figure XVII. The figure shows the experimental points for the combined treatment connected with a dashed line; all the points except one fall into the envelope of additivity applying at each survival level. The point at the lowest level of survival (7.5 µg/ml of BCNU, 4 Gy of x rays) falls outside the respective envelope, although the displacement is not so great that it might not be explained by experimental uncertainties.

 

Figure XVII. Isobolic diagram for the combined action of BCNU and x rays on 9L rat brain tumour cells [D2]



 

136. The scope of this brief overview of chemicals capable of modifying radiation sensitivity is simply that of illustrating the very wide range of processes whose alteration may in turn lead to a synergistic or antagonistic interaction in irradiated biological systems. The data reviewed are as such of little relevance for the main scope of this Annex, because the doses of radiation used are usually very high (up to several Gy, depending on the susceptibility of the systems tested) and the concentrations of the chemicals often toxic. It is however appropriate that they should be mentioned because the processes governing different aspects of cell radiosensitivity might also be relevant at lower levels of exposure. No practical situation where the above mechanisms would be significant can currently be envisaged at the low doses of interest for the purpose of the present Annex.

 

C. CARCINOGENIC CHEMICALS

137. Organic substances which are known to have carcinogenic properties should be discussed separately. Some, such as the alkylating agents, have already been mentioned in section III. B. Carcinogenic agents are usually divided roughly into initiators and promoters, following the two-stage theory of carcinogenesis [B18]. It is known however that such a subdivision is not rigid because many agents share the properties of both classes. In combination experiments it may be expected that the final tumour yield may depend on the properties of the interacting agents, as well as on the order and time pattern of their administration. A potent initiator followed by an active promoter might be expected to give the highest carcinogenic response and reversal of their order of administration a drastic reduction of this response. Another important trait is the spectrum of the tumours induced, as some agents may be extremely specific in this respect and their inter-action with radiation might change this selectivity.
138. Precise quantitative data were provided by DiPaolo et al. [D21, D17, D22] and Kennedy et al. [K14] on the morphological transformation of mammalian cells in culture in regard to the interaction between ionizing radiation and the carcinogen benzo(a)pyrene or the promoting agent phorbol ester. The experiments elucidated the dose-time relationships for an effect of special significance for practical purposes, showing enhancement factors of up to 9 fold, depending on the conditions of exposure and on the doses of the agents interacting. This series of experiments also attempted to elucidate the mechanisms of interaction. For the same biological end-point, the promoter 12-0-tetradecanoylphorbol-13-acetate (TPA) administered after x-ray or neutron irradiation to C3H/10 T 1/2 cells in culture was shown to act synergistically, with complex relationships as a function of the radiation type and dose [H27].
139. A study showing an increased yield of leukaemia in mice pre-irradiated with x rays and subsequently treated with methylcholanthrene was published by Furth and Boom [F9]. An increased yield of leukaemia in mice induced by x rays, methylcholanthrene or oestrogens was also shown by Kawamoto et al. [K6] when the animals were simultaneously treated by urethan. After Berenblum investigated the interaction of x rays and urethan in mouse leukaemogenesis in great detail and showed that the order of administration of the combined agents was of decisive importance [B2, B26] many other authors reported enhancement of leukaemia under the same agents [D4, L22, V6, G1] particularly in young animals [B1, L23] where enhancement is especially pronounced. This may be on account of differences in the drug distribution or catabolism as a function of age [C2]. Data have also been reported for croton oil [I2], myleran [U5] and novoembicyn [A8] in conjunction with radiation.
140. Combined treatments of pre-implantation mouse embryos in vitro with x rays or phenols (which are promoting and mutagenic chemicals in some test systems) showed that the effects were, at most, additive [M30]. Schmahl and Kriegel irradiated mouse embryos in utero at 11–13 days p.c. (1 Gy at each time) and injected the pregnant mothers at 17 days p.c. with 0.5 mM/kg of ethylnitrosourea [S42]. Tumour development was followed postnatally up to 18 months. Results from this series are shown in Table 2, with the inter-action factor calculated according to equation (5). If one considers total tumours as the end-effect, inter-action appears to be antagonistic ( = 0.44). If one takes each category of tumours separately, one may conclude for at least one clear case of synergism and one of antagonism for leukaemia and hepatomas, respectively. This appears to be a good example of a change in the tumour spectrum brought about by the combined treatment. However, for more definitive statements exposure-response curves within a broader range of values for both the single and the combined actions would be required.
141. Much work has been carried out on the skin, the tissue where the two-stage mechanism of carcinogenesis was originally identified [B18] and can be more easily tested. Electrons or UV light in association with other carcinogens usually result in a higher yield of tumours than any of the agents administered alone. This applies to methylcholanthrene [C14] and to 7,12-dimethylbenz(a)anthracene (DMBA) [E6, S28]. However, a recent report [B19] on this latter substance in association with 0.8 MeV electrons (5–25 Gy) in respect to carcinogenesis of rat skin showed that the tumour yields were approximately equal to the sum of the yields induced by the separate treatments, so that prior irradiation did not appear to alter the susceptibility of rat skin to DMBA carcinogenesis.
142. The case of 4-nitroquinoline-l-oxide (4NQO) has been particularly well analysed. When applied in combination after beta rays from ⁹⁰Sr-⁹⁰Y (both agents at doses that did not separately induce tumours) it appeared to have a synergistic effect for skin tumour induction in mice. Reversing the order of administration of the treatments led to a much smaller yield of tumours by about an order of magnitude [H5]. When the interval between beta irradiation and subsequent chemical treatment was made to vary between 11 and 408 d, the tumour induction rate was found to be almost at the same level for all treatment times, indicating that the latent carcinogenic change induced by skin irradiation may persist for a very long time and remain available for subsequent interaction with the 4NQO [H6]. Finally, caffeine was found to further increase the incidence of malignant tumours in mouse skin when painted after beta rays and 4NQO treatment [H7].
143. Croton oil, a typical promoter of skin neoplasia from which TPA is extracted, gives uncertain results when combined with radiation: enhanced effects with UV [E2] and electrons [S29] or absence of any enhancement [G6, B20] have in fact been reported. It may be said in very general terms that the concepts of initiation and promotion may be verified on the skin also in the case of drug-radiation interactions. However, the results of combined treatments on the skin could also be interpreted on different grounds and some of the previously mentioned experiments [H5, E2, E6] would in fact be regarded by others [N3] as clear examples of co-carcinogenesis by chemical and physical agents.
144. The situation with respect to other tumours or to systemic leukaemogenesis is definitely more difficult to interpret. In the case of the lung, urethane (which specifically induces adenomas in mice) has been used in association with x rays at various doses and dosages. A reduction in the incidence of tumours (both as percentage incidence and as tumours/animals) has been obtained in one experimental series [F6]: cell killing by the high radiation dose in the urethane-induced tumours was held responsible for the effect. Recalculation of these data by others [L21] led however to a different interpretation. Additive effects of radiation and urethane were reported in another series, and the final outcome of the treatments was deemed to depend on two competing phenomena, cell killing and cell transformation, whereby, depending on the dose of the two agents, any effect may become possible. Immunological phenomena might also interfere in this case to make the picture very complex [C15].
145. Procarbazine (PCB), a drug used frequently in the treatment of the Hodgkin's disease, is a known carcinogen in experimental animals since it gives rise to pulmonary adenoma and leukaemia in mice, mammary tumours in rats and acute myelogenous leukaemia in primates. Hybrid (BALB/c x DBA/2) F₁ mice were given this drug and ionizing radiation at different times to test for possible synergistic effects [A7]. Single-treatment groups received 300 mg/kg PCB weekly for four weeks, a dose effective for induction of pulmonary adenoma and leukaemia; or 0.6 Gy/d of 300 kVp x rays for five d, a dose which did not result in tumours of the lung. Combined-treatment groups received radiation three days or three weeks before PCB or PCB three days before irradiation at the above dosages. The experiments were terminated within 12 weeks with killing of the surviving animals. Pulmonary adenomas in mice receiving both agents were significantly increased over the level of induction by PCB alone. Thymomas were also increased significantly in the animals given the drug three days before or after irradiation. The authors concluded for a synergistic effect of the combination and hypothesized that an increase of the normal tendency of mice to develop pulmonary adenoma would be at the origin of the interaction. Immunosuppression combined with direct cellular damage might also be responsible for the effect.
146. Among studies where combinations of chemical carcinogens and radiation were tested, the experiments of Metivier [M6] regarded the action of PuO₂ given by inhalation, in combination with benzo(a)pyrene (BP) or dimethylnitrosamine (DMNA), compounds which are widespread environmental pollutants. Both carcinogens were given after the exposure to the PuO₂: BP (2 x 5 mg) was administered intra-tracheally in association with haematite 2–3 weeks after the nuclide; DMNA (2 or 20 ppm) was given orally, added to the drinking water. Tumours of the lung and of other sites, histological types of tumours, invasiveness and survival time were the principal end-points investigated.
147. BP alone led to a small increase of tumour incidence above the control level. PuO2 (0.63 kBq) produced similarly a slightly increased incidence. Both agents combined produced an appreciable increase in the number of tumours with an increased invasiveness. Survival time reflected closely the results on tumour incidence, being essentially unchanged for the two agents given alone and practically halved by their combination. At least on qualitative grounds, a synergistic interaction was operating in these experiments, the latency period of the tumours in the combined treatment group being evidently shorter. For higher levels of PuO2 (6.3 kBq) a synergistic action might also be present, but its expression (particularly with regard to survival time) is much less clear. In the case of DMNA no synergistic action with respect to alpha radiation alone was found. At high concentrations (20 ppm) the latter drug produced a subacute intoxication and no synergistic effect. It was reported however that inhalation of PuO2 in association with DMNA did result in an increased tendency of liver tumours to metastasize into the lungs.
148. In the experiments of Little et al. [L10] the inter-action between benzo(a)pyrene (BP) and alpha radiation of ²¹⁰Po, was examined. The experiments were performed on hamsters and the two agents were administered by intratracheal instillation, absorbed on haematite particles or dissolved into physiological saline. In a first series of experiments the two agents were administered simultaneously in 15 weekly instillations (0.3 mg BP + 0.2 kBq ²¹⁰Po/treatment). Under these conditions the results were compatible with an additive interaction of the two agents.
149. In a second series of experiments BP was given 15–18 weeks after the administration of a single dose of 1.5 kBq ²¹⁰Po. While BP alone (2.4 mg in eight weekly instillations of 0.3 mg) or ²¹⁰Po alone produced practically no lung tumours, the combinations of both agents resulted in a clear synergistic effect, with 17% of the animals developing frank tumours of the lung. Physiological saline and gelatine were mostly used as carriers. When the administration of BP preceded the ²¹⁰Po treatment no increase of tumour induction was seen. It is remarkable that when the second treatment consisted of saline alone, without BP, a sharp increase of the tumour yield was seen, compared to the ²¹⁰Po treatment alone. The instillation of isotonic saline could act as a non-specific stimulus to cell proliferation [Lll] and subsequent experimental work [L24] appeared to lend support to this hypothesis. Autoradiographic experiments showed that after treatment by BP or by saline the epithelial cells of the hamster lung undergo a wave of mitoses. This enhanced proliferation would be essential for the expression of the radiation-transformed cells.
150. A biochemical approach to the study of mechanisms of interaction between radiation and chemicals in the case of lung tumour induction was followed by Queval and Beaumatin [Q1]. These authors studied the correlation between the capacity by various substances of inducing pulmonary enzymes and their ability to shorten the latency period of the lung tumours in rats, following inhalation of radon daughters. The research established that compounds such as benzoflavone, methylcholanthrene and benzopyrene are highly effective in enzyme induction and capable, at the same time, to shorten the latent period of tumour appearance.
151. Large experimental series were carried out on the combined effects of radiation and inhalation of uranium ore dust and diesel oil exhaust fumes at the Pacific Northwest Laboratory [C16]. The experiments on hamsters involved about 600 animals non-exposed or exposed to radon and radon daughters, uranium ore dust and diesel engine exhaust, alone or in various combinations. Squamous cell carcinomas developed in only a few of the animals exposed to radiation and they were always preceded by a squamous metaplasia of the alveolar epithelium. In general, however, the hamster lung was found to be rather refractory to the malignant transformation and did not even develop lesions that could be classified as pre-cancerous when exposed to levels of the above agents which were regarded as realistic for life exposure regimes. Thus, the hamster lung under these conditions may not be a useful model for pulmonary cancerogenesis in man.
152. Knizhnikov et al. [K9] modelled another case of industrial exposure by a combination of shistose ash, benzo(a)pyrene and ²¹⁰Po. The mice were exposed to ash alone, ash with BP or with ²¹⁰Po, and to the triple combination of the agents together. The yield of lung tumours and their latency period were studied and at the levels used the yield was reported to increase from 35% (ash only) to 61% (triple combination). The latency period decreased in the same two groups from 300 to 200 days. An interaction factor may be calculated from these data of about 1.3, indicating some synergistic interaction. Other control groups were included in this series.
153. The intragastric administration of 3-methylcholanthrene followed by x rays [S30] or fission neutrons [S4] produced no more than additive effects for induction of mammary adenocarcinoma. However, the same chemical applied locally on the brain, in association with beta irradiation resulted in an antagonism which was proportional to radiation dose [MI5].
154. There are many different experiments concerning a variety of other tumours. X rays alone or in combination with benzo(a)pyrene produced the same incidence of neoplasia [K7]. Dibutylnitrosamine (DBNA) or 4-ethylsulphonyl-naphtalene-l-sulphonamide (ENS) combined with x rays showed no effect on tumours of the urinary bladder but a reduction of the mammary tumour incidence [F7]. A synergistic action on the production of liver and gastric carcinoma by fission neutrons in combination with N,N'-2,7-fluorenylenebisacetamide (2,7-FAA) was reported, but no interaction for intestinal tumours was found [V7]. Localized x-irradiation in association with the same drug administered in the diet accelerated the induction of hepatomas [N1] and similar effects were reported with the association of x rays and o-aminoazotoluene [K8] and of ¹⁴⁴Ce and dimetylaminoazobenzene (DBA) [M16]. Experiments on additive carcinogenic effects of 9,10-dimethyl-1,2-benzanthracene or 1,2,5,6-dibenzanthracene in association with chronic internal irradiation from ⁹⁰Sr were also reported [Z1, Z2].
155. The mutagenic substance N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) was tested in combination with whole-body fission neutron irradiation for its carcinogenic properties on the gastrointestinal tract of rats. A high incidence of gastric and duodenal carcinomas was found after the MNNG treatment but the neutrons did not produce any tumour. Combining the two treatments did not change the effect of MNNG [V3]. Survival and tumour induction were tested in three strains of rat following x-irradiation in association with urethane. In spite of some interesting differences observed between strains, the overall effect of the joint treatment was not greater than the sum of the separate effects at the dosage level studied [M17].
156. In work by Sanders [S12], besides the combined action of ²³⁹PuO₂ and asbestos, the combined action of ²³⁹PuO₂ with benzo(a)pyrene was also studied, both agents being administered intraperitoneally. The action of BP alone produced mostly abdominal sarcomas. The combination of BP and 13.3 kBq of ²³⁹PuO2 resulted in an approximately additive yield of sarcomas. Other tumours which are characteristic of the plutonium action were also produced. The administration of BP increased the translocation of plutonium to liver and lung, which points to the need that possible metabolic effects leading to different dose patterns in various organs upon the joint administration of two substances should be taken into account when discussing the results of combined actions.
157. In conclusion, it appears that the evidence reviewed is very conflicting. The number of substances tested is large and the amount of information relating to each of them very little. The pathogenesis of the tumour systems tested is complex and the conceptual distinctions between induction and promotion cannot be held in many instances. In some cases the application of chemicals after irradiation may enhance the tumour yield by comparison with the opposite order of application. In other cases, the association of treatments may actually decrease rather than enhance the induction of neoplasia when the toxicity of the combined agents outweighs their additive carcinogenic properties [U2]. No definite conclusions with respect to any class of tumours may therefore be drawn before the dose, the dosage schedule, the order of administration and modalities of the combined treatments are properly and thoroughly explored, which is very rarely the case in the contributions that have been reviewed.

 

D. THE SPECIAL CASE OF TOBACCO SMOKE

1. General

158. Tobacco smoking is a widespread habit of many human populations in spite of a well documented association between smoking and lung tumour incidence. The relationship between annual death rate from this cause and number of cigarettes smoked per day is reported to be linear with slope of about 10-⁴ [D5, D23] and with incidences of lung cancer rising from about 0.07 10-³ per year in non-smoking males to 3 10-³ per year for male persons smoking 35 or more cigarettes per day. The chemical composition of tobacco smoke is very complex and includes more than 1000 identified compounds [S9], a number of which are aromatic hydrocarbons that have been shown to act as carcinogens. Smoke and tobacco tar also contain a number of tumour promoting and co-carcinogenic agents.
159. The two-stage nature of the carcinogenic action of tobacco smoke was shown by a classical experiment on mouse skin by van Duuren et al. [VI]. The initiator, 1,2-dimethylbenz(a)anthracene (DMBA) acted in this case as an initiator and cigarette smoke condensate (CSC) as the promoter. Five weekly applications of CSC after a single application of DMBA greatly increased the rate of tumour appearance, by shifting the latency period from 450 d (for DMBA alone) to approximately 100 d. The initiating action of the tar components is relatively low compared to that of DMBA. In this particular case the initiator was a chemical substance, but any other carcinogenic agent, ionizing radiation in particular, could be effective in combination with the promoters contained in the smoke concentrate. This point was proven experimentally by McGregor [M27, M28] who treated rat skin with beta radiation and subsequently applied CSC. Rats treated with CSC only produced no tumours. A two- to three-fold increase in the numbers of skin tumours was observed in the groups under combined action, as compared with the animals exposed to beta radiation alone. It should, however, be realized that only few agents can be considered as pure initiators or promoters, the rule being that many carcinogenic agents have the properties of both classes of substances and sometimes to various degrees, according to the different animal models tested.

2. Experimental data

160. Various examples of interaction between radiation and tobacco smoke have been reported in animals [C5, C6, C16]. In experiments by Chameaud et al. [C5, C6] rats were exposed to radon inhalation in special chambers. They developed respiratory cancers as a function of exposure and exposure rate, starting from a control background incidence of practically zero. Similarities could be shown histologically between these tumours and human lung tumours. Inhalation of cigarette smoke in these animals did not result in malignant transformation of the respiratory cells but only in benign lesions of the bronchial epithelium and lung parenchyma [C7]. In interaction experiments the exposures to radon daughters was chosen to be 100, 500 and 4000 WLM, because it was shown in previous tests that the incidence of lung cancers of respectively 1-2, 5-10 and 30-40% would result from them [L8]. Cigarette smoke inhalation was carried out under standardized conditions for periods of 15 min ten times per day, four days per week, for one year. No change in the animals' life span was seen after this treatment. An elaborate classification of the pathology was set up to follow the spread of tumours at death.
161. For the highest radiation exposure (4000 WLM) the incidence of lung cancer was 34% and it increased to 68% in animals also exposed to smoke. At 500 WLM, 7% and 28% were the corresponding figures and at 100 WLM, 0% and 3.3%. Since smoking was without effect, equation (22) in a simplified form may be used to analyse the data. Accordingly, the values of the inter-action factors for the above groups are 2, 4 and 00. Pathologically, tumours appeared to be more advanced in animals receiving the combined treatments, indicating that neoplastic lesions developed earlier in these animals. Microscopically, the same tumour histotypes were found in the irradiated group and in the group with combined exposure. The authors pointed out the similarity between these findings and those in uranium miners and proposed their rat tumour system as a good model system for the human situation [C7, L8].
162. Another interesting aspect of the laboratory experiments with rats which is in accordance with some results from epidemiological studies on uranium miners is the complex dependence of the lung tumour yield in the animals on the exposure rate and on the level of exposure. Human data strongly suggest that lung cancer may be produced more efficiently at low than at high exposure rates [L6, K1], in the sense that per unit dose higher incidences of tumours are produced at low than at high dose levels. It should be pointed out that low doses are usually obtained at low dose rates. In the experiments of a French laboratory the incidence of lung cancer in the rat per 10⁶ WLM changed from over 200 at cumulative exposures of around 175 WLM to 46 at 8000 WLM [L8]. 
163. The above mentioned French group investigated in further experiments the temporal aspects of a combined treatment in rats of radon daughters and tobacco smoke, by reversing the order of administration of radiation and tobacco smoke with respect to the previously cited experiments [C7, C8]. In this case radon exposure followed exposure to smoke [L9], without any enhancement of carcinogenesis. This observation is in keeping with the notion that tobacco smoke has a promoting action. It was not possible to examine the relationship between the level of exposure to smoke and tumour incidence, since higher levels of exposure led to a toxic action of some tobacco constituents and, on the other hand, lower exposures required an excessive number of animals for statistical validation of the data. 
164. The effect of grading the exposure to tobacco these experiments [C22]. 
165. Grading the exposure to BF (25, 9, 3 mg/kg/ week) and to radon daughters (6000, 3000, 500 and 100 WLM) gave 12 possible combination groups [L9]. Preliminary data showed that the reduction of the latent period was dependent on the product of the parameters characterizing exposure to each agent, as though a lower dose of one could be compensated by a higher exposure to the other in a multiplicative manner. Such a dependence resembles to some extent the "relative risk model" proposed by Lundin et al. [L6] to account for epidemiological data in uranium miners. 

166. Modelling of chronic inhalation of radon daughters and tobacco smoke simultaneously was carried out on experimental animals at the Battelle Northwest Laboratories [C16]. The temporal aspect of the administration of the combining agents differed from the experiments of the French group already reviewed [C5, C6], where exposure to smoke followed the radon treatment. The experiments comprised seventy beagle dogs: twenty of them were exposed to radon, uranium ore dust and cigarette smoke; twenty to smoke only; and twenty to radon plus uranium ore dust. The other animals served as the controls. Exposure to tobacco smoke was performed through special masks during several daily sessions. 

167. Animals that developed lung tumours had in general cumulative exposures to radiation in excess of 13 000 WLM. This dose level is about two orders of magnitude higher than that reported to cause lung cancer in man. The possibility was therefore considered that the longer life span of the human species might allow more tumours to appear while, for the same tumour incidence, much of the exposure in dogs would be "wasted", i.e., ineffective in producing additional tumours. Differences in histotype between human and dog respiratory neoplasms were also noted. Cigarette smoke had a reducing effect on the radiation lung cancers (2 cases out of 20 animals) as compared to animals non-exposed to smoke (8/20). It was suggested —but in the absence of direct experimental evidence—that smoke through an increased production of mucus might result in a lower dose of radiation to the target cells; alternatively, smoke might stimulate mucociliary clearance. Changes in the lung that were associated to tobacco smoke were emphysema, chronic bronchitis and bronchiolitis, lung fibrosis. The antagonistic effect of tobacco smoke on lung tumors induced by radon daughters was confirmed in a very recent report of smoke may to some extent be studied by the use of chemicals which are constitutents of tobacco smoke or tar, although it should be kept in mind that in this case the mechanism of action could be rather different. Morin et al. [M5] examined the effect of inhaled radon daughters in combination with the I.P. administration (25 mg/kg/week for 13 weeks) of benzo-5, 6-flavone (BF), a substance which is not in itself a carcinogen. Treatment with BF was started at three months after the end of radon exposure at 6000 WLM during about two months. One hundred percent of the animals developed lung tumours (multifocal, invasive epidermoid type with a latency period of 3 months), as compared with an expected 50% within 15 months after radiation exposure given alone. When BF administration was started 16 months after radon exposure, no difference was seen with respect to the group receiving only radon. This was taken as evidence that the promoting action of BF was exerted during the period of latency of the radon induced malignancies.

168. It may thus be concluded that reasonable dose-response relationships for lung tumour induction in experimental animals may be obtained for exposure to ionizing radiation. The separate effects of tobacco smoke may also be studied, but testing their combined action poses serious problems. The temporal sequence of administration is very important; there are probably differences in target cells with respect to the two agents; there may be other unknown factors complicating the picture; the mechanisms of induction have not been sufficiently clarified. It may be tentatively proposed that a common feature of many experiments in animals (and of some epidemiological series in man) is a promoting action of the smoke (or some of its constituents), leading to a shortening of the latency in tumour appearance. Whether this might be due to a non-specific stimulating action on the proliferation of the respiratory epithelia or to a specific effect of some smoke constituents is impossible to say at present.

169. Uranium miners are exposed to radon and radon daughters. They represent the first occupational group on which extensive epidemiological surveys were made of the effects of radiation in combination with tobacco smoke. The exposure levels for this group of workers are usually expressed in WLM: for the equivalence of this operational unit with other radiation units, see Annex D. In an epidemiological survey [L2] 3414 miners exposed to up to 10⁴ WLM from the year 1950 were followed up to September 1967. Against 251 deaths expected during this interval of time, 398 deaths were actually observed, the main causes for the excess being violent deaths (120 observed versus 51 expected) and malignant tumours of the respiratory tract occurring ten or more years after beginning of work in the uranium mines (62 observed versus 10 expected). The time relationship and the increase in cancer mortality as a function of radiation exposure indicate a causal relationship between the two variables.
170. Information about the smoking habits of the miners were collected during the survey and also in an annual census of uranium miners which was started in 1963. Standardized mortality ratios of lung cancer by smoking categories [H3] were used for calculation of the expected death rate for respiratory cancer. The reference population was a random sample of adult males from the United States and the ten expected cases mentioned above were calculated according to these data. It was found instead that the cases expected would be 16 for the same total population of 3414 miners if the lung tumour incidence among males of four Colorado plateau states in the United States would be taken as the reference control. Table 3 shows the distribution of observed and expected respiratory cancer deaths between smokers and non-smokers. The increase in the number of cancer cases is attributable to irradiation by inhaled radon daughters. The relative excess of risk between smokers and non-smokers is the same (3.9 against 4.0 for the two categories, respectively). If one calculates the increase in cancer incidence due to irradiation per person year at risk, one finds 1.7 10^-3 for smokers and 1.7 10^-4 for non-smokers, the difference being attributed by some to a 10-fold synergistic increase of the risk for the smoking miners.
171. A more accurate analysis shows however that this could be a misleading argument. It should be realized that the statistical significance to be attached to the number of tumours observed in the non-smoking group is very low, owing to the small number of cases observed. The estimate of the probability of tumour induction obtained from this number is therefore affected by a large error. A statement such as the preceding one of a ten-fold increase in risk in the smoking population, would be equivalent to using for the assessment of the interaction factor the formula 
     
                                                                   = (P_ot — P_t1) / P_o2           (36)
     
     where the signs 1 and 2 refer to smoking and radiation, respectively. In fact, this formula cannot be used under the circumstances, because of the mentioned low statistical significance of the term Po2 (the probability of respiratory cancer death in the non-smokers) and of the absence of the term Pot from the numerator and denominator.
172. According to the reasoning presented in chapter I of this Annex, when Po_l and Po2 are small, one calculates the interaction factor (t) by the formula
     
                                                            
     
     Table 4 shows the results of separately analyzing the data for the period 1950-1967 [L2] (A: top line) and for the last four years of the same period, 1964-1967 [Al] (B: bottom line). As may be expected, risk estimates based on the most recent period of observation are higher, excluding re-evaluated estimates of spontaneous risk and risk of smoking. The interaction factors are however close enough to each other and indicate a synergistic interaction. In view of the low statistical significance of the results, other indirect evidence may be of great value.
173. Archer and collaborators [Al, A3, L6] point out some of this evidence. In a larger group of uranium miners 207 lung cancers were identified; all of these individuals except three were cigarette smokers [Al]. Since it is known that 71% of miners are smoking, it is clear that in the above group smokers are over-represented. Another observation relates to the age at diagnosis: in the 207 people mentioned, 17 stopped smoking eight or more years before diagnosis; 16 stopped between four and eight years; 19 smoked less than 15 cigarettes/day (light smokers). Controls were chosen to match as closely as possible the exposed individuals in relation to age at the start of mining, cumulative radiation exposure, years of hard rock non-uranium mining. All of them smoked 20 or more cigarettes per day and none stopped smoking more than one year before diagnosis. The comparison showed that non-smokers or those who stopped smoking eight or more years before developing lung tumours had a mean age at diagnosis three years greater than smoking controls. Light smokers differed from controls by a year and a half, and those who stopped smoking between four and eight years before diagnosis differed from controls by less than one year. The results support the hypothesis that cigarette smoking acts in these miners as a promoting agent [Vl] by decreasing the length of the latent period. These conclusions were strongly supported by an update of the earlier uranium miners mortality studies in the United States [A3]. The incidences of lung cancer between different categories of smokers are shown in Figure XVIII [A3].

174. High rates of lung cancer in smoking persons are also observed in workers of industries that use known carcinogenic substances such as chromates [01, L7] and asbestos [S10]. In addition, among persons developing lung cancers in the same groups of workers, smokers were over-represented. These data may be taken to show a non-specific promotive influence of tobacco smoking. At the present time there is by no means a full understanding of these mechanisms in smoking individuals. It could be that tobacco smoke contains enough initiators and promoters to give the observed yield of respiratory cancer. Alternatively, in case of smoking acting apparently alone, some environmental factors may provide the initiating stimulus and the role of smoking might be essentially promotive. The chemical composition of smoke itself might even not be the decisive factor in the promoting action of this agent. As shown by experiments of Little et al. [L10] reviewed previously, the irritation of the respiratory epithelia by non-specific physical or chemical agents (instillation of saline solutions, for example) could have a promoting effect.
175. It could of course be debated if promotion as such is an effect to be included under the general heading of synergistic. Two extreme situations may be visualized in this respect. There could be, on the one hand, a forward displacement in time of the tumours appearing, but with a final yield of tumours not different from the situation in which promotion is not operating. Alternatively, a continuously increasing rate of tumour appearance might take place, leading finally to an incidence higher than that in the absence of promotion. A variety of intermediate situations could also operate between these two extremes. Clearly, if the final tumour incidence would be taken as the reference end-point, the first of the two situations depicted would not come under the definition of synergism, while the latter would. But if instead, more correctly, the length of tumour-free life lost is taken to be the reference parameter and it is assumed that smoking alone could cause cancer, both situations, as well as all the inter-mediate ones, would be rightly described as synergistic interactions.
176. In their 1979 paper Lundin et al. [L6] gave a more elaborate quantitative treatment of respiratory cancer death in uranium miners. A log-normal distribution of the time elapsed from exposure to diagnosis based on other experimental and theoretical evidence [M4] was assumed to apply. It was also assumed that this distribution would have a standard deviation of 0.17609 in log t units, where t is the number of years elapsed after the beginning of exposure. The choice of the standard deviation was rather arbitrary and, according to criteria developed in reference [M4], somewhat below the range expected usually. The parameter describing exposure (when the risk from earlier was added to that of later exposures) was the Eff WLM (k) for the year k, defined as
     
                                                
     
     where O < j < k; w(k–j) is the proportion of the area under the log-normal distribution density curve which is bounded by the interval from (k – j – 1/2) years to (k – j + 1/2) years; and WLM(j) is the exposure in WLM during the year j.
177. Two alternative hypotheses were examined: that the increase in absolute risk might be proportional to radiation exposure, in which case the risk increase would be independent of the rate associated with cigarette smoking, aging or other environmental factors (a); that the increase in relative risk may be proportional to radiation exposure (b). In this case the increase in risk should be proportional to that risk which would have affected the miners in the absence of radiation. The analysis of both models was carried out based on the form of the temporal distribution of the respiratory cancer deaths. 
178. Three temporal parameters were used, namely, the age of the miners, the calendar year and the years after the beginning of exposure. Computations were as follows: (a) for the absolute risk hypothesis
     
                                                
     
     where X_rad = av (Eff WLMx). In the above formula n_x is the predicted excess of lung cancer deaths among miners in stratum x and the av (Eff WLM_x) is averaged over all the person-years at risk in the stratum; (b) for the relative risk hypothesis
     
                                                
     
     in which n_x is proportional to the product of the expected number of lung cancers E_x in the stratum, multiplied by the exposure X_rad. The symbols a_A and a_R in the above equations are coefficients applying to the two situations postulated. The sum of n_x + E_x gives the total predicted number of lung tumours for each particular set of parameters.
179. Calculations were made of the number of deaths according to the age category and with assumptions of mean latency times of 5, 10 and 15 years. The relative risk model gives results which are closer to observations and latency times of 10 to 15 years fit the data best. One of the parameters which is most strongly influencing the expected number Ex is the smoking category (Table 5). It may be assumed that Ex is proportional to smoking exposure X_sm. Then the predicted excess of respiratory tumour deaths will be proportional to both smoke and radiation exposure as
     
                                                
     
     The probability of developing lung cancer per person per year is shown in Table 5 for four smoking categories. The last column of the table gives the corresponding interaction factors and shows that the highest value is for former smokers and the lowest for heavy smokers, being intermediate for light smokers. This result comes about through an insufficient increase in P_ot by comparison with P_tl for heavy smokers, an observation which contradicts the previous conclusion about the applicability of the relative risk model. The authors interpret this observation as evidence against a possible "synergism" defined as an increase in the total radiation risk of lung tumour development. Such a risk they consider to be approximately the same for all categories of smokers and somewhat higher than for non-smokers. They classify the observed increase in the lung tumour death as promotion. However, as already discussed before, there is good ground to describe it as synergism (see paragraph 175).
180. Lundin et al. [L6] do not exclude the possibility that for longer time intervals the yield of lung tumours in non-smokers might be the same as that of the smokers exposed to radon daughters. One preliminary report on lung cancer in Swedish iron miners seems to support this possibility [R15]. In an other epidemiological study of metal and iron-ore miners in Sweden carried out by Axelson and Sundell [A9] the risk for the non-smokers was claimed to be higher. However, the size of the groups analysed was rather small and the statistical significance of the observed effects correspondingly low. Also, the methodology of the case-control study was not fully described in the publication [A9] and it raises some questions in the form presented.
181. Long latency time for lung cancer and the incidence dependence on the dose rate could also obscure the final picture. Uncertainties in the distribution of miners between exposure categories could lead to distortions in the estimates of risk. The lower risk of the American uranium miners could be justified to some degree by their possible misclassification into higher exposure categories [Sil, K1]. It should also be mentioned that for these miners the observation time elapsed from the beginning of work in the uranium mines is not much longer than 20–30 years, which might be insufficient for the development of lung cancer among non-smoking individuals. Another important factor could be the exposure rate which was lower for the Swedish than for the miners in the United States. It has already been mentioned that lower exposure rates may bring about a higher total yield of potential tumours. Different exposure rates can be met in epidemiological studies with Czechoslovak [S11, K1] and Canadian [H20] miners.
182. Enhanced mortality for chronic respiratory diseases other than cancer resulting in pulmonary insufficiency (pneumoconiosis, pulmonary fibrosis, emphysema) and for acute conditions (pneumonia, asthma) could also be the result of combined radiation and tobacco smoke exposure [A3]. An increase in the rate of mortality from these diseases for uranium miners in the United States was clearly observed which might be related to radiation exposure. It is interesting to note that the rate is highest for light smokers, so that at high exposure levels mortality is twice as high as for heavy smokers. Some possible interaction between radiation and smoke is also evident for these diseases but, at this point, the contribution of other ambient conditions like siliceous dust or diesel fumes should also be considered, for which data are very scarce.
183. The epidemiological data discussed point to a synergistic interaction between tobacco smoke and radiation exposure in the sense discussed under paragraph 175. Non-specific effects induced by some component of tobacco smoke could be responsible for the results described. Thus, changes in the production of mucus, a slower rate of clearance of radioactive particles by the ciliary action and metaplasia of the epithelia might result in a higher dose delivered to the target cells in smokers than in non-smokers. Against this general proposition is however the observation that promotion by tobacco smoke is still found when smoke is applied long after radiation exposure. Clearly, these questions cannot be settled now with the limited information available. Quantitation of the degree of synergism is also impossible with the necessary degree of precision and significance, owing to the low number of tumours observed, particularly among the non-smoking individuals, and to the complex temporal pattern of lung cancer development.

E. OTHER DRUGS

184. It may appear somewhat artificial to separate in this section substances which are utilized for their pharmacological properties in clinical medicine from other organic substances mostly developed for their radiosensitizing actions. The separation may be made on the ground that interaction with radiation may be incidental in the former case but is pursued as a specific goal in the latter. In spite of a widespread and increasing use of many drugs in modern societies, it is difficult to visualize situations where the combined effects of any of them with radiation may pose significant problems in public health. The cases of inter-action in the treatment of specific diseases where the combined use of radiation and drugs might increase the risk of undesirable effects on the patient may be more important. However, in most of the work reviewed radiation doses were very high and, irrespective of the nature of the interaction (synergistic or antagonistic), extrapolating the findings to lower levels may be very difficult or impossible in view of the modification of the form of the dose-effect relationships that might occur at low doses.
185. Antibiotics are widely used in clinical medicine and some of them are also used in combination with radiation for cancer chemotherapy [P3, P4, P5, P16]. Among them, actinomycin D was shown to have a synergistic interaction with radiation on Chinese hamster cells in culture. Elkind et al. [E3] related this effect to the ability of the drug to impair recovery of sublethal damage, as shown by a reduction of the shoulder of the survival curves at low doses (2–5 10-³ µg/ml). Ten-fold higher doses given before irradiation increased the exponential slope of the survival curves. Time was also an important parameter in these experiments because no synergism was observed with treatments by actinomycin later than 10–12 h post-irradiation or when the drug was applied more than six hours before irradiation [E4].
186. The molecular basis for the action of actinomycin D on cells is due to its proven ability to bind to DNA and thus to create a steric hindrance to the synthesis of RNA. Interesting studies on the combined action of this drug and of another drug, cordycepin, were reported on two cell lines in culture by Robertson et al. [R13]. For actinomycin D interaction factors of 1.2 and 1.3 in the two cell lines were reported with x rays. The survival parameters affected were both the exponential slope and the shoulder of the survival curve, but mostly the former. With cordycepin the inter-action factors were 1.1 and 2.2, respectively, and the main parameter affected was the extrapolation number. The nature of the differences described led the authors to some hypothesis on their molecular basis.
187. Actinomycin D was also tested in preimplantation mouse embryos in tissue culture for its combined effect with radiation [S15]. The concentration of the drug was here several orders of magnitude lower (10-⁴ µg/ml) than in previously reported experiments by Robertson [R13] and the drug alone was ineffective at these concentrations in retarding the development of the embryos to the blastocyst stage. Combining the drug with tritiated water led to a higher effect than that of tritiated water alone, with interaction factors between 2 and 4, depending on the tritium concentration in the culture medium. The lower values were observed at the highest concentrations. This observation could be explained by the more effective inhibition of the repair processes at low radiation doses. A re-analysis of these data in terms of isobolic diagrams [S48] showed that the results of combined treatment fell clearly outside of the envelope of additivity in the direction of synergism. The possibility that the shape of the dose-response relationships may be changed by the combined treatments has been discussed in this context [S39]. Other experiments showing a radiosensitizing action of actinomycin D have also been reported [M7].
188. In humans, Wara et al. [W2] studied the effect of actinomycin D on the induction of radiation pneumonitis occurring 1–3 months after the irradiation of lung metastases in 41 patients. Doses of the order of 7.7 Gy were necessary to elicit pneumonitis in 5% of the patients and these doses were reduced to 5.5 Gy (DMF = 1.4) when an actinomycin treatment was given along with radiation. Similar DMF were obtained for radiation-induced intestinal injury, oesophageal lethality and pulmonary lethality in LAF₁ mice treated with this drug at the same time [P5]. The values of the DMF obtained for a variety of chemotherapeutic drugs in the above experimental series are given in Table 6 and a detailed review of experimental and therapeutic findings related to this topic has been written by Phillips and Fu [P3]. In patients treated for erythema of the skin d'Angio et al. [D6] reported a DMF of 3.4 by combining x irradiation and actinomycin D.
189. Dritschilo et al. [D7] investigated the mechanisms of the combined action with radiation of actinomycin D and adriamycin. Non-toxic levels of actinomycin and minimally toxic levels of adriamycin produced suppression of potentially lethal damage repair in plateau-phase Chinese hamster cells in culture. For actinomycin this suppression persisted as long as the drug was present in the culture medium, but as soon as it was removed prompt repair took place. This suggested that suppression did not act through fixation of injury to a non-repairable state. Adriamycin was different because cells exposed to it could eventually proceed to repair potentially lethal injury even in the presence of the drug, after an initial delay of the repair processes.
190. Redpath et al. [R4] studied the effect of combining adriamycin (2 or 1 mg/kg, 5 daily fractions) and x irradiation (10 Gy/fraction, 5 daily fractions) on mouse tissues. Enhancement of damage was seen for lung and foot skin damage, when the interval between the beginning of the radiation and of the drug course was within two to seven days. In another series the radiation sensitivity was studied after the single dose of 1 mg/kg intraperitoneally. No effect was found in this case. Experiments were performed in the same laboratory on the radiosensitivity of ICR male mice irradiated whole-body with fast neutrons (mean energy 25 MeV) or photons (6 MeV), in combination with a single dose of adriamycin (10 mg/kg) [C9]. The L^D50/6 for photons was reduced from 13 to 10 Gy; that of neutrons from 5.6 to 4.3 Gy. The RBE for gut damage was unaltered by the addition of adriamycin. The data indicated that for drug administration 16 hours before or after the radiation exposure the interaction will be the same.
191. A cell cycle dependence of the synergistic interaction of a drug with radiation was shown for dihydroxyanthraquinone (DHAQ), a potential cancer chemotherapeutic agent similar to adriamycin and actinomycin D [K5]. The survival of x-irradiated Chinese hamster cells in combination with different exposures to DHAQ was the end-point of this study. DHAQ had a toxicity which was more pronounced during the early phases of the cell cycle. After combined treatment a synergistic effect was noted for cells in the S phase, but in all other phases of the cycle additivity prevailed. In asynchronous populations DHAQ enhanced the radiation-induced cell lethality primarily by increasing the slope of the radiation dose-survival curve.
192. Lucanthone (Miracyl D) has long been used in the treatment of schistosomiasis. The drug has a hetero-cyclic ring structure resembling that of actinomycin D. A synergistic type of interaction of this drug with ionizing radiation has been shown for HeLa cells. This effect decreased with the time lag allowed between radiation and the treatment with lucanthone [B4]. The same publication refers also to increased 30-day lethality in mice given 4 Gy of total body radiation and a simultaneous injection of 180 mg/kg of the drug.
193. The influence of lucanthone in combination with x-irradiation was also studied on V-79 cells and on spheroids [D3]. The treatment of asynchronous cells with 5 µg/ml of the drug led to a progressive decrease in the proportion of cells in G₁ and to an accumulation of S-phase cells. The toxicity of the drug was noted only during this latter phase of the cell cycle. In general, the survival of the single cells after the combined treatment was lower, owing to a reduced capacity of the cells to accumulate and repair sublethal damage. For equal levels of drug toxicity, the radiation-modifying effect of the drug was greater in the spheroids, pointing to a larger interaction in the system which has greater capacity for accumulation and repair of the sublethal radiation damage.
194. Lucanthone has also been shown to be active in respect to induction of developmental defects in mice. Pregnant animals (8 days p.c.) were given 70 mg/kg of the drug and treated one hour later with 0.5 Gy of x-radiation. The treatment resulted in a distinct synergistic increase of the eye abnormalities of the embryos [M8]. The above studies were further developed with a decrease of the x-ray dose down to 0.01 Gy [M9]. Pregnant mice of the strains F/A and NMRI were irradiated at eight days p.c. with 140 kVp x rays, with or without treatment with lucanthone. The foetuses were observed 4–5 d after irradiation for the presence of macro- and microscopic developmental defects (post-implantation loss, growth retardation, eye abnormalities, exencephaly, cleft palate and limb defects). There was a strain specificity with respect to the sensitivity to lucanthone given alone, the NMRI mice being more susceptible to lower doses of the drug. A dose of 0.01 Gy was reported to produce a statistically significant increase of the abnormalities and combination of the two treatments gave rise to a synergistic interaction. Some strain specificity was also found for the combined effects, because the F/A mice were more susceptible to the joint action. Higher doses of radiation (0.5 Gy) producing an approximately 4-fold increase of the control abnormalities were also reported to produce synergism. Other data on the enhancement of radiation effects by antibiotics (ledermycin, reverine) were also reported [M9].
195. Bleomycin was reported to potentiate the radiation damage in rat brain tumour cells of the line 9L [H9]. The drug enhanced cell lethality mostly through an increase of the slope of the radiation dose-survival curve, its Do decreasing from 3.7 to 2.1 Gy in the presence of the drug. There was also a more modest decrease in the capacity for accumulation of sublethal damage, shown by a decrease of the Dq from 3.2 Gy in the absence to 2.9 Gy in the presence of the drug. This was evidence for an inhibition of repair of the sublethal radiation damage. Other authors observed an additive effect of bleomycin and x rays with sm