Some recent data on chemical protection against ionizing radiation

Adv. Space Res. Vol. 12, No. 2—3, pp. (2)213—(2)221, 1992 Printed in Great Britain. All rights reserved.

0273—1177/92 $15.00 Copyright © 1991 COSPAR

SOME RECENT DATA ON CHEMICAL PROTECTION AGAINST IONIZING RADIATION M. Fatome, J. D. Laval and V. Roman Unite de Radioprotection, Centre de Recherches du Service de Sante des Armées, Grenoble, 24 Avenue des ~naquisdu Grésivaudan, 38702 La Tronche, France

ABSTRACT Once introduced in the organism, the radioprotectors are fastly degraded and that increases their toxicity, shortens their duration of action and renders them inactive after oral delivery. So, it was tried to protect them by their incorporation in vectors. When a cysteamine-liposomal suspension was orally delivered, it showed a radioprotective activity for about 4 hours. By using 35 S cysteamine, it was noted that its plasmatic concentration was increased. Freeze-drying of these preparations was a good mean of conservation if the samples were stored at 4°C. A good and sustained activity was also obtained after oral delivery of WR-2721 entrapped in microsj.heres. Otherwise, it was shown that after interacting with the polar heads of phospholipids, under determined conditions of pH and in fluid phase, aminothiols can penetrate inside the membrane and be entrapped in the internal medium of liposomes and as they penetrate, they can lessen the diffusion of oxygen in the lipidic bilayers. INTRODUCTION Most radioprotective compounds are aminothiols or derivatives among which phosphorotbioates are the most potent /1/. Generally, they lose their activity when the time elapsed between their delivery and the irradiation exceeds a few minutes, when their dose is lower than the maximum tolerable dose or when they are orally administered. These drawbacks cannot allow their delivery to intervention or rescue squads and particularly to spacemen. They can be explained by the fast oxidative degradation of these compounds once introduced in the organism /2/ or by their enzymatic alteration in the digestive tract, particularly in the stomach. So, we have tried to protect the best known radioprotectors from the degradation by incorporating them in suitable carriers. The first and the most studied were the liposomes which were used as radioprotectors carriers for testing the influence of the incorporation on the radioprotective effect and as membrane models in order to determine the nature of their interaction with compounds, the mechanism of penetration and the influence on oxygen diffusion of these coumpounds. Otherwise other types of carriers were studied, microcapsules, nucrospheres, nanoparticules, ions exchange resins particularly for the entrapment of phosphorothioates which could not be incorporated in liposomes, but the results were often disappointing. MATERIAL AND METHODS Liposomes were prepared according to BANGHAN /3/ from egg yolk lecithin and cholesterol in a molar ratio 4/1 dissolved in ether which was then eliminated in a rotatory evaporaror under reduced pressure. The dried lipidic film was dispersed in a 3 M cysteamine hydrochlohydric solution (25 g in 75 ml 5 mM phosphate buffer) and sonicated for 40 mm under a nitrogen atmosphere at 40°C. After a centrifugation for 40 mm at 105000 g, extravesicular cysteamine was removed by dialysis against a 5 mM phosphate buffer pH 7.5. Entrapped cysteamine was determined by the Eliman method /4/ after disruption of liposomes by 2-propanol. This method is a spectrophotometric estimate of the thiol groups by reaction with 5,5’-dithiobis (2-nitrobenzoate) and of absorbance at 412 were tim. The 1. measure These cysteamine suspensions given by concentration canula was about mg.mrmice (Charles River France) at a cysteamine dose of intragastric to CD9 COBS 500 mg.kg1. Mice were then exposed on different times after drug delivery (from 15 to 270 mm) to a gamma-irradiation from a 60Co source.

(2)213

M. Fatome et aL

(2)214

Pharmacokinetic assays were undertaken after oral delivery of these cysteamine liposomal suspensions previously with was 3~S measured cysteamine. finaldeproteinized specific 3 Bq.ml’ . The 35labeled S activity in The plasma, activity was 74x10 plasma fraction, liver and spleen. Moreover spectrophotometric thiol estimates were carried out on liver, spleen and deproteinized plasma fraction. The same determinations were made after oral delivery of free cysteamine. This study was practized on male CD COBS rats, 250 g in weight and fasted for 12 h before the experiment. Trials of freeze-drying of liposomal cysteamine samples were carried out. After several attempts, the best cryoprotective effect was obtained with a mixture of glucose and saccharose in weighing ratio 8:5 at the weighing concentration of 13 % with regard to the suspension. Samples of 1 ml of sugared cysteammne liposomal suspension were frozen in vials at -40°C, with a l°C.min~freezing rate and then freeze-dried. Vials were sealed and stored at 4°C or at 15°C. For the entrapment of the phosphorothioate WR-2721, the use of a microencapsulation technique incorporating the dispersed molecule in a polymerized matrix and producing microspheres was considered. Because of the very high hydrophilicity of WR-2721, a slightly modified solvent-evaporation technique /5/ was used. WR-272l (3 g) previously passed through 175 pm pore-size filters was dispersed in 100 ml of an acetonic ethylcellulose solution at 2 % (w/v), Metarin F (emulsifyng agent consisting of soya lecithin, mono and di-glycerides) (25 g) was dispersed in 475 g of vaseline oil in a reactor vessel in which was added the dispersed WR-272l phase while stirring at 800 g at 25°C over 1.5 h during which the solvent evaporated and the microspheres were obtained. They were then washed with hexane, passed through a boro-silicate filter (porosity 4), dried under vacuum at 20°C for 24 h and passed through 200 pm pore-size filters. The yield was about 95

%.

For the different physico-chemical studies, an aqueous liposome suspension (100 mg. cm .3) was prepared by vortexing in a 20 mmol ~j~•3 phosphate buffer dipalmitoyl-phosphatidylcholine (DPPC). The bulk pH was controlled with an Ingold microelectrode and eventually adjusted at pH 6 or pH 8 for experimenting outside the pKa region of the stearic spin probe which is around 6.8 and 7.4 in DPPC in gel phase and in fluid phase respectively. The ESR spectra were recorded on a Varian E-l09 spectrometer. The ESR labeling was done with stearic spin probes with a nitroxide free radical in position 5 (5NS) which explores the lipid-water interface or In position 16 (16NS) which explores the bilayer core. The molar ratio was about 1 spin probe for 250 DPPC. The incident microwave power varied from 2 to 400 mW. The oxygen transport was evaluated by the spin-exchange between the nitroxide free radical and the paramagnetic triplet state oxygen molecules dissolved in the lipidic matrix. The magnitude of this exchange can be followed by the ESR signal saturation with the increase in the incident microwave power. This saturation can be evaluated by the half-saturation parameter which is the power at which the signal is half as great as it would be without saturation. It is related to the oxygen diffusion concentration in the close vicinity of the nitroxide group /6/ /7/. The influence of different compounds on this parameter was tested : two aminothiols (cysteamine and WR-l065), two degradation products (hypotaurine and taurine) and a structural analogue, ethanolammne. RESULTS AND DISCUSSION The intragastric administration of the liposomal cysteamine suspension to mice at a cysteamine dose of 500 mg . kg1 led to a sustained radioprotective effect /8/. It is given in table 1 as a function of the time elapsed between the suspension delivery and the irradiation. No cysteamine-induced lethality was noted. A significant protection was observed over about 3.5 h with a DRF around 1.4. This result contrasts with the lack of protection when cysteamine is orally delivered as an aqueous solution and with the duration of action which does not exceed 30 mm after parenteral injection. It is also important to note that the association of free cysteamine and empty liposomes did not induce a significant protection and that entrapment is therefore necessary. After labeled cysteamine liposomal suspension oral delivery in rats, the radioactivity measured in plasma (figure 1), liver and spleen was higher than that assessed after free cysteammne oral delivery at the same dose ( 500 mg . kg’ ) /9/. Moreover, the percentages of 35 S activity in the deproteinized fraction of plasma were significantly lower when cysteamine was entrapped. From 10 to 120 mm after administration, these percentages progressively increased from 30 % to 70 %,

Chemical ProtecUon

(2)215

whereas with free cysteamine they were near 100 % as early as 10 mm (figure 2). These results could be explained by the binding of cysteammne to proteins by disulphide bridges, this fact implying an increased protection of the molecule from degradation and oxidation. This was confirmed by the higher values of thiol assays by Eliman method in plasma after entrapment. On the contrary, the thiol assays in liver and spleen showed that the higher levels of radioactivity are due to degradation products of cysteammne. TABLE 1 Radioprotective activity of cysteamine orally delivered to mice as liposomal suspension at 0.5 g.kg~ Delay (hours) between oral delivery and irradiation

D.R.F.

0.25

*

1.30

0.5

1.45

1

1.35

1.5

1.40

2

1.35

3.5

1.30

4.5

1

Free cysteamine

*

1

Empty liposomes

*

1

Free cysteamine Empty liposomes Delay : 0.5 h

+ *

1.10

30000

DPM

TA’T~ I

20000

/

10000

0

I-Il ‘60

L6O

10

DELAY (MINUTES)

Fig. 1 : 35 S radioactivity in plasma (d.p.m.m11) after oral administration of (—) free cysteammne and (- -) cysteamine liposomal suspension to rats (mean ±s.d). JASR

12:2/3—0

(2)216

M. Fatome

aoL

PERCENT

120

i00~~ BO

eo

/\

40-

/

I 20

DELAY (MINUTES) Fig. 2 Percentage of plasmatic ~5 S radioactivity in deproteinized plasma fraction after oral delivery of (—) free cysteamine and (- -) cysteamine liposomal suspension to rats. The higher digestive absorption of cysteammne when entrapped and its protection in the digestive tract during and after digestive absorption can account for the observed radioprotection, but its mechanism remains unclear. Although the possibility of small absorption of unilamellar vesicles through the intestinal wall has not been excluded /10/ /11/, phospholipids are expected to be hydrolysed into lysolipids by phospholipases in the digestive tract, and liposomes are known to be easily disrupted and turned into mixed micelles by bile salts /12/. An explanation could be an interaction between the phosphate ions of the polar heads of phospholipids and the cysteamine as described by Kranck /13/ by means of spectrophotometric, dielectric and conductivity measurements.

CYSTEA~CNE CONCENTRATION

~ (mg.nir’)

“~¼

a.,

‘‘‘~~~‘‘‘

‘~•

2~)O ‘ ‘ DELAY (DAYS)

360

‘

‘

1) assessed at different delays Fig. 3 : Cysteamine concentration (mg.ml after preparation in freeze-dried cysteamine liposomal suspension samples stored at (—) 4°C or (- -) 15°C.

Chemical Protection

(2)217

STJWIIVAL RATE 120 (percent)

I-

..

_I\

60 40 20’ S.

I_I..’

‘‘‘I’’’

100

,

I

200 300 DELAY (DAYs)

400

Fig. 4 : Survival rate of mice after oral delivery of freeze dried cysteamine liposomal suspension after rehydratation of samples at different times of storage at (—) 4°Cor (- -) 15°C.The delivered dose of cysteammne was equal to 500 mg.kg~.

The stability of freeze-dried liposomal cysteamine samples gave the following results. For samples stored at 4°C, no significant decrease in cysteamine concentration and in radioprotective activity in mice after oral delivery of 1 of rehydrated samples were observed for 1 year (all prepared samples were 40 pl.g used by this time) (figures 3 and 4). No aspect modification was also noted. On the contrary, for samples stored at 15°C, the cysteammne concentration and the radioprotective activity decreased as early as the third month. This activity was abolished on the sixth month (figures 3 and 4). These results were in accordance with the browning of the preparations. This modification is characteristic of autooxidative processes. So, freeze-drying seems a good mean for conservation of cysteamine incorporated in liposomes, particularly if they are stored at 4°C. This result is noticeable, owing to the very great sensitivity of aminothiols to oxidation. The incorporation of WR-2721 in microspheres led to a decrease in toxicity after oral delivery to mice, as compared with that obtained with its administration as suspension in oil. A drawback was represented by the high viscosity which precluded the mntra-gastric delivery of a dose higher than 2400 mg.kg1 For radioprotective tests, the WR-272l dose was equal to 1800 mg.kg~ . The DRF value was around 1.7 when mice were irradiated 1 or 2 hours before the oral administration. It fell to 1.3 when the delay was equal to 3 h /14/. So the microencapsulation of WR-272l led to a significant and sustained radioprotection by oral way. However, the duration was lower than that obtained when cysteamine was entrapped in liposomes. This fact could be explained by the very high hydrophihtcity of WR-272l /15/. As it has been described for water-soluble drugs incorporated in microspheres, a fraction is to be found on or near the surface. A rapid loss of this fraction could lead to a shortening of its action duration /16/. Moreover, it induces a fragility of the microspheres. Anyway, the incorporation gives a protection of WR-2721 against hydrolysis in the gastro-intestinal tract, particularly in the stomach, and a progressive release of this molecule. These results show that it is possible to protect radioprotectors against degradation in the organism and to improve their activity by incorporating them in suitable carriers, such as liposomes and microspheres. However all formulations are not efficient. For example, it is the case of polyisobutylcyanoacrylate nanoparticles in which the amount of WR-2721 or WR-1065 was too low for giving a radioprotection, the delivered quantity being limited by the toxicity of the polymer /17/. Otherwise, it was impossible to incorporate WR-2721 in liposomes and for aminothiols, the concentration was not very high. So, simultaneously, physico-

(2)2 18

M. Fatome a aL

chemical studies were undertaken for a better knowledge of the interaction of aminothiols with liposomes, and of the mechanism of their entrapment and therefore for the obtention of a better yield of incorporation. The different physico-chemical studies showed that cysteammne presents by adsorption a high affinity for the lipid-water interface up to a cysteamine dipalmitoyl-phosphatidylcholine molar ratio equal to 1/1. There was an electrostatic binding between the positive charges of cysteamine (with external pH around 7.4, below the pKb of 11 of the amino group) and the negative charges of the polar heads of the phospholipids /18/. It is important to note that no such interaction was observed with WR-2721. Results suggest that as structural defects appear near the main transition of temperature of DPPC (around 41°C) cysteammne penetrates the bilayers, with the thiol extremity presumably directed towards the hydrophobic core. This interaction is dynamic and presents fast and permanent fluctuations in the bulk water and the bilayer. It is ruled by a dynamic parameter and is pH-dependent. At the lipid interf ace, cysteamine reactivity is different from reactivity in the bulk-water. It depends on the membrane potential and on the partition coefficient of protonated and non-protonated forms. These coefficients depend on the membrane organization, on its chemical composition and on the nature of the interacting molecule. For example the pKb of cysteamine which is equal to U in water is

apparently replaced by a pHeq of 5.4 at the interf ace. In these conditions, if a gradient of pH is created between the internal medium of DPPC liposomes and the external medium, (for example with pH values of 4 and 7.4 respectively), cysteamine is protonated in the external bulk water and can interact at the lipid-water interface. Its PHeg in lipids being equal to 5.4, it loses its proton and this neutral form being more lipophilic than the NH 3~forts, it can penetrate inside the lipidic bilayer and finally reach the inner interf ace where it is reprotonated and entrapped in the internal medium. Thus, this medium loses a proton and therefore, its pH increases and the penetration stops when this pH reaches the PHeq of cysteamine. So, the mechanism of entrapment is dynamic and acts as a cysteamine pump. It has been observed that this penetration starts at pH 4.9 and reaches its maximum at pH 5.9. We are now trying to experimentally check this model. A preliminary study seems to indicate that with the creation of the precedent pH gradient, the incorporation is 10 to 20 fold higher than that obtained without gradient (figure 5). This influence of pH gradient on incorporation in liposomes has been described for some lipophilic drugs such as dubicaine /19/.

~7.43~MEA

E.O3~~MEA

Fig. 5 : Influence of pH gradient on penetration and entrapment of cysteamine in DPPC liposomes - A ten fold increasing in entrapment has been noted. The spectral half saturation parameter which is related to the oxygen diffusionconcentration product was studied according to the temperature. For 5-nitroxide stearic acid which explores the interface, in oxygenated system, it appeared two increases, one at 33°Cwhich is the pretransition temperature of DPPC the other at 41°Cwhich is the main transition temperature. This last increase was very important and even appeared as a discontinuity, this critical point involving a total disorganization of the structures and a maximal transport of oxygen. In deoxygenated medium, no increase of P 1/2 was observed. The addition of cysteammne to oxygenated medium with an increasing in the cysteammne/DPPC ratio from 1/20 to 1/1 strongly decreased the P 1/2 values, and at the highest ratio value, they were even lower than that obtained in deoxygenated medium (figure 6) /20/. These decreased values remain stable over at least one day in oxygenated medium. This observation shows that this effect is done without cysteamine consumption. For 16-nitroxide stearic acid, the decreasing in P 1/2 was a function

Chemical Protection

(2)219

of the extravesicular pH, it started at pH 4.9, reached its maximum at pH 5.9 and thus depended on the cysteamine penetration in lipidic bilayers (Table 2). This effect on oxygen diffusion was observed with WR-l065 but not with ethanolamine, hypotaurine and taurine, this involving the role of the thiol group.

1/2 mY

400

a

‘

‘

20

60 TEMPERATURE

Fig. 6 : Half saturation parameter (P 1/2) values as a function of the temperature observed with 5 - nitroxide stearate at pH8: (1) Pure DPPC in oxygenated medium; (2) Pure DPPC in deoxygenated medium; (3) cysteamine/DPPC molar ratio 1:1. in oxygenated medium; (4) cysteainine/DPPC molar ratio 1 : 5 in oxygenated medium. The discontinuity in pure DPPC oxygenated system is observed at 41°C which is the main transition temperature of DPPC. TABLE 2 P 1/2 parameter of 16-nitroxide stearic acid as function of the extravesicular pH in cysteammne/DPCC molar ratio = 1:1 - Decreasing starts as cysteammne penetrates inside the bilayer (pH 4.9) and reaches its maximum at

pH 5.9.

I

pH

I

1/2(mW)

p

4

170

4.5

165

4.9

165

5

130

5.4

110

5.9

45

6

60

6.5

60

7

50

The mechanism of this action of aminothiols is not entirely clarified. An interaction of these compounds with oxygen is unprobable because of the long duration of the effect in open atmosphere, NMR spectra in D 20 or in DPPC-D20 suspension did not detect, in the same conditions, an oxidative degradation of

(2)220

M. Fatome a aL

cystamine for at least 5 hours. An interaction o ~minothiols with nitroxide stearic acid spin probes can also be ruled out because they are in large excess. Finally, an interaction of cysteamine with DPPC can be hypothetized. Oxygen diffusion in phospholipidic membranes can be described by an axially synunetric tensor with the long axis perpendicular to the bilayer plane /21/. Penetration of aminothiols inside the bilayers could induce a specific hindrance which would lessen or abolish this oxygen diffusion. Anyway, this action is important to consider because it could be extrapolated to some extent to biological membranes, their lipidic matrices being fluid at physiological temperatures. It could be considered as an additional argument for considering a cellular hypoxia as one mechanism of radioprotection by aminothiols /22/. CONCLUSION In conclusion, carriers such as liposomes or microspheres are able to improve the radioprotective effect of radioprotectors, particularly by enhancing the duration of their action and by leading to an effect after their oral delivery. Their plasmatic concentration is increased and their release is progressive. Freeze-drying of aminothiol liposomal suspensions seems a good mean of conservation of aminothiols if samples are kept at 4°C. Liposomes are also a suitable tool for studying the mechanism of interaction of radioprotectors, their mechanism of penetration inside the membranes and some aspects of their mechanism of action. So, under determined conditions of pH and in fluid phase, aminothiols after interacting with the interface can penetrate inside the membrane and be entrapped in the internal medium of liposomes and as they penetrate, they can lessen the diffusion of oxygen in the lipidic bilayers. AKNOWLEDGMENTS The authors are grateful to F. Leterrier, F. Berleur, C. Lecomte and A. Vachon for helpful assistance and valuable comments. Parts of this work were supported by grants from Institut Henri Beaufour and Direction des Recherches Etudes et Techniques. REFERENCES 1

D.E. Davidson, M.M. Grenan and T.R. Sweeney, Biological characteristics of some improved radio-protectors, in : Radiation Sensitizers. Their use in the clinical management of cancer, ed. L.W. Brady, New york, 1980, p. 309.

2

D.M. Ziegler, L.L. Poulien and R.B. Richerson, Oxidation metabolism of sulfur containing radioprotective agents, in Radioprotection and Anticarcinogens, ed. O.F. Nygaard and M.G. Simic, New York, 1983, p. 191.

3

A. Bangham, M. Standish and .3. Warkins, Diffusion of ions across the lamellae of swollen phospholipids, 3. Mol. Biol, 13, 328 (1965).

4

G.L. Ellman, K.D. Courtney, V. Andres and R.M. Featherstone, A new and rapid colorimetric determination of acetylcholinesterase activity, Biochem. Pharmacol. 7, 88 (1961).

5

Y. Pongpaibul, .1.C. Price and C.W. Whirworth, Preparation and evaluation of controlled indomethacine microspheres, Drug Dev and Industr. Pharm. 10, 1597 (1984).

6

W.R. Subczinski and .1.S. Hyde, The diffusion concentration product of oxygen, in lipidic bilayers, using the spin label T. method, Bioch. Bioph. Acta, 643, 283, (1981).

7

A. Kusumi, W.D. Subczinski and .1.S. Hyde, Oxygen transport parameter in membranes as deduced by saturation recovery measurements of spin lattice relaxation times of spin labels, Proc. Nat. Acad. SL. USA, 79, 1854, (1982).

8

V. Roman, F. Bocquier, F. Leterrier et M. Fatome, Action radioprotectrice de la cysteamine incorporée dans des liposomes administrés par voie orale a la souris, C.R. Acad. Sc Paris, 295, 191 (1982).

Chemical Protection

9

(2)221

D. 3askierowicz, F. Genissel, V. Roman, F. Berleur and M. Fatome, Oral administration of liposome-entrapped cysteamine and the distribution pattern in blood, liver and spleen, mt 3. Radiat. Biol 47, 615 (1985).

10 D.S. Deskhmukh, W.D. Beak and H. Brockerhoff, Can intact liposomes be absorbed in the gut ?, Life Sc. 28, 239 (1981). 11 R.E. Pagano and 3.N. Weinstein, Interactions of liposomes with mammalian cells, Ann. Rev. Bioph. Bioengin 7, 435 (1978). 12 F. Paltauf, The intestinal absorption of 1,2- and 1,3-dialkyl glycerol ethers and of diether phospholipids, Biol. Bioph. Acta, 176, 818 (1969). 13 M. Kranck, M.A. Rix-Monteil and D. Vasileslu, Demons tration of an interaction of the radioprotector cysteamine with lecithin, Physiol. Chem. Phys. 13, 429 (1981). 14 M. Fatome, F. Courteille, J.D. Laval and V. Roman, Radioprotective activity of ethylcellulose microspheres containing WR-2721, after oral administration, mt. 3. Radiat. Biol, 52, 24 (1987). 15 3.M. Yuhas, M.E. Davies, D. Glover, D.A. Brown and M. Ritter, Circumvention of the tumour membrane barrier to WR-272l, adsorption by reduction of drug hydrophilicity, mt. 3. Radiat. Oncol. Biol. Phys. 8, 519 (1982). 16 E. Tomlinson and 3.3. Burger, Incorporation of water-soluble drugs in albumin microspheres, in Methods in enzymology. Vol. 112 Drug and enzyme targeting, ed. K.S. Widder and R. Green, New york, 1985, p. 75. 17 M. Fatome, .3.D. Laval and M. Massot, Incorporation of WR-272l and WR-1065 in nanoparticles: Physico-chemical and radioprotective studies, Acta Pharm. Techno. 35, 238 (1989). 18 F. Berleur, V. Roman, D. Jaskierowicz, M. Fatome, F. Leterrier, L. TerMinassiaw-Saraga and G. Madelmont, The binding of the radioprotective agent cysteamine with the phospholipidic membrane head group-interface region, Biochem. Pharmacol 34, 3071 (1985). 19 L.D. Mayer, K.F. Wong, K. Menon, C. Chong, P.R. Harrigan and P.R. Cullis, Influence of ion gradients in the transbilayer distribution of dibucaine in large unilamellar vesicles, Biochem 27, 2053 (1988). 20 A. Vachon, C. Lecomte, P. Braquet, V. Roman, M. Fatome and F. Berleur, Oxygen diffusion concentration in phospholipidic model membranes. An e.s.r. saturation study, 3. Chem. Soc. Faraday Trans. I, 83, 177 (1987). 21 Y.N. Molin, M.M. Salikhov19. and K.L. Zamarev, : Spin exchange, ed Springer Valay, New York, 1980, p. 22 3.W. Purdie, E.R. Inhaher, H. Schneider and 3.L. Labelle, Interaction of cultured mammalian cells in the WR-272l and its thiol WR-1065 implications for mechanisms of radioprotection. mt ~. Radiat. Biol, 43, 517

(1983).