www.elsevier.com/locate/asr

Advances in Space Research 35 (2005) 243–248

DNA DSB induced by iron ions in human fibroblasts:LET dependence and shielding efficiency

G. Esposito a,b, F. Antonelli a, M. Belli a,b,*, A. Campa a,b, V. Dini a,b, Y. Furusawa c,G. Simone a,b, E. Sorrentino a,b, M.A. Tabocchini a,b

a Laboratorio di Fisica, Technology and Health Department, Istituto Superiore di Sanita, Viale Regina Elena 299, 00161 Rome, Italyb Istituto Nazionale di Fisica Nucleare, Gruppo collegato Sanita-Sezione di Roma1, 00161 Rome, Italy

c National Institute for Radiological Sciences, 4-9-1 Anagawa, Inage, 263-8555 Chiba, Japan

Received 26 July 2004; received in revised form 22 October 2004; accepted 19 November 2004

Abstract

This paper reports on DNA DSB induction in human fibroblasts by iron ions of different energies, namely 5, 1 GeV/u, 414 and

115 MeV/u, in absence or presence of different shields (PMMA, Al and Pb). Measure of DNA DSB was performed by calibrated

Pulsed Field Gel Electrophoresis using the fragment counting method.

The RBE–LET relationships for unshielded and shielded beams were obtained both in terms of dose average LET and of track

average LET. Weak dependence on these parameters was observed for DSB induction.

The shielding efficiency, evaluated by the ratio between the cross sections for unshielded and shielded beams, depends not only on

the shield type and thickness, but also on the beam energy. Protection is only observed at high iron ions energy, especially at 5 GeV/

u, where PMMA shield gives higher protection compared to Al or Pb shields of the same thickness expressed in g/cm2.

� 2004 COSPAR. Published by Elsevier Ltd. All rights reserved.

Keywords: Space radiation; HZE particles; DNA DSB; PFGE; RBE–LET relationship; Radiation shielding

1. Introduction

Galactic Cosmic Rays represent one of the main

sources of radiation outside the magnetic field of the

Earth. They include high charge and energy (HZE) par-

ticles that are of special concern for their potential ef-fects on the astronaut�s health during long term space

flights. Risk reduction can be provided by effective radi-

ation shielding inside the spacecraft (Wilson et al., 1995,

1997). As a consequence of their interaction with a

shield, HZE particles fragment and deposit energy at

rates depending on their type and energy, and on the

nature and thickness of the shield. The biological effec-

0273-1177/$30 � 2004 COSPAR. Published by Elsevier Ltd. All rights reser

doi:10.1016/j.asr.2004.11.021

* Corresponding author. Tel.: +39 06 49902916; fax: +39 06

49387075.

E-mail address: m.belli@iss.it (M. Belli).

tiveness of the mixed field emerging from the shield

needs to be properly understood in order to evaluate

the shielding effectiveness. It is out of the question, for

HZE particles, to consider the possibility to have a

shield of reasonable weight which is sufficient to stop

the beam. Thus the primary particles, that have notchanged their identity in nuclear interactions, emerge

from the shield with lower energy. If this were the only

effect, then the shield would worse things, since deceler-

ated particles will generally have higher LET and greater

biological effectiveness. Therefore, the beneficial conse-

quences of a shield will have to derive from the nuclear

production of secondary particles, which will change not

only the energy, but also, in a substantial way, thecharge composition of the particles reaching the cellular

targets. The need then arises to build a shield such that

the emerging composite beam will be really less efficient

ved.

244 G. Esposito et al. / Advances in Space Research 35 (2005) 243–248

in damage induction. It is not obvious a priori that this

will be the case with any shield; then studies are neces-

sary to characterize the different situations.

To this purpose an international collaboration was

started some years ago aimed at studying the influence

of the shielding on biological effectiveness of HZE par-ticles for several cellular end points. In this framework

ground-based experiments have been performed using

the accelerators at the Brookhaven National Laboratory

(BNL), USA, and at the National Institute for Radio-

logical Sciences (NIRS), Japan.

Although it is widely recognised that the spatial dis-

tribution of DNA damage depends on the radiation

quality and that correlated DSB are produced by thesame particle track (Goodhead, 1994; Holley and Chat-

terjee, 1996), to our knowledge there are no biological

data concerning the effects of shielding on DSB induced

by HZE particles.

In this paper we present data on DNA double strand

breaks (DSB) induction in human fibroblasts exposed to

accelerated iron ions in the absence and in the presence

of polymethylmethacrylate (PMMA), aluminium andlead shields.

2. Materials and methods

2.1. Cell culture and sample preparation

The normal human fibroblasts AG1522 cell line wasobtained from the National Institute of Aging (NIA)

cell repository (Coriell Institute for Medical Research,

USA). Cells were grown as monolayer in alpha-Mini-

mum Essential Medium, containing 1 mM glutamine,

supplemented with 20% foetal calf serum, 2% Hepes

buffer solution (1 M), 50 U/dm3 of penicillin and strep-

tomycin (all reagents from Gibco).

Cells were cultured for at least five generations toreach and maintain confluence in the presence of

1.85 · 103 Bq/cm3 14C-thymidine (NUNC). The labeled

medium was substituted with unlabeled one 24 h be-

fore irradiation. Cells were then detached by addition

of trypsin–EDTA, pooled, centrifuged and resus-

pended at the final concentration of about 1.3 · 106

cells cm�3 in 0.8% (w/v) low-gelling agarose (Sigma

Type VII) made up in PBSS/EDTA buffer. A volumeof 85 ll of this suspension was pipetted into 7 · 5 · 2

mm moulds (Bio-Rad) and allowed to form plugs at

4 �C.

2.2. Irradiation

AG1522 cells embedded in agarose plugs were irradi-

ated with Fe ions of various energies. Irradiation with 5and 1 GeV/u Fe ions was performed at the Alternate

Gradient Synchrotron (AGS) and at the NASA Space

Radiation Laboratory (NSRL) facilities of the Brookha-

ven National Laboratory (BNL), Upton, USA. Irradia-

tion with 414 and 115 MeV/u Fe ions was performed at

the Heavy Ions Medical Accelerator (HIMAC) of the

National Institute for Radiological Sciences (NIRS),

Chiba, Japan. c-rays from a 60Co source were used atthe Istituto Superiore di Sanita, Rome, Italy, as refer-

ence radiation.

Irradiations with 5 and 1 GeV/u and 414 MeV/u were

also performed in the presence of different shields,

namely PMMA, Al and Pb.

In order to avoid DSB repair during irradiation, plug

holders that allow to maintain the temperature at 0–4 �Cwere especially designed to fit the beam geometry at thetwo facilities.

Doses up to about 250 Gy were delivered to the sam-

ples. Dose rates of �10–15 Gy/min were used for un-

shielded beams. After shielding the dose rate changed

by factors ranging from 0.27 to 1.75, depending on the

beam energy and on the type and the thickness of the

shield.

Beam characteristics under the various irradiationconditions are reported in Table 1 (Durante et al., in

press; Miller and La Tessa, personal communications).

The track-average LET (LT) and the dose-average

LET (LD) listed in the table represent the first and the

second moment of the LET distribution, respectively,

so that their knowledge gives the expected value and

the width of this distribution, being its variance, in par-

ticular, given by LT Æ (LD � LT). It can be noted that forthe unshielded 5 GeV/u Fe beam the difference between

LD and LT reflects a significant contamination of lighter

ions, mainly protons, as shown by measurements with Si

detectors (Miller, personal communication).

Dosimetry was performed by ionization chambers at

BNL (Zeitlin et al., 1998) and by ionization chamber

and CR39 plastic nuclear track detectors at NIRS (Dur-

ante et al., 2002; Grossi et al., 2004).Shield insertion changes the dose rate measured at

the sample position. The ratio between the dose inci-

dent on the shield and that at the sample position (in

the same time interval) is defined here as the ‘‘dose rate

reduction factor’’ (DRRF) and listed in the last col-

umn of Table 1.

The dose per unit fluence for the unshielded beams

can be evaluated from the track-average LET (LT): D/F = 0.16LT/q, where D is the dose in Gy, F is the fluence

in lm�2, LT is in keV/lm, and the value q = 1 g/cm3 was

used. For the shielded beams, this value divided by the

dose reduction factor gives the dose at sample position

per unit fluence of particle impinging on the shield.

2.3. DSB measurements

After irradiation, the plugs were carefully removed

from the moulds and incubated in lysis solution (0.5

Table 1

Physical characteristics of the unshielded and shielded beams

E (MeV/u) Shield LD (keV/lm) LT (keV/lm) Dose per unit fluence (Gy lm2) Dose rate reduction factor

Material Thickness mm (g/cm2)

115 No (1) – 442 440 70.4 –

414 No (3) – 202 200 32.0 –

PMMA (1) 23 (2.7) 228 201 33.3 0.96

PMMA (1) 43 (5.1) 277 225 38.6 0.83

PMMA (2) 56 (6.6) 394 294 56.1 0.57

Al (1) 30 (8.1) 442 336 55.2 0.58

Pb (1) 10 (11.3) 354 318 53.3 0.60

1000 No (3) – 147 142 22.7 –

PMMA (2) 197 (23.2) 163 52 14.1 1.61

Al (1) 97 (26.2) 179 78 16.8 1.35

Pb (1) 26 (29.5) 170 90 19.9 1.14

5000 No (1) – 135 92 14.7 –

PMMA (1) 249 (29.4) 85 25 4.0 3.67

Al (1) 111 (30.0) 118 52 7.4 1.99

Pb (1) 26 (29.5) 134 98 13.2 1.12

In the second column the number in parentheses denotes the number of independent experiments performed in each case. The sixth column lists the

dose per unit fluence of particles impinging either on the sample for unshielded beams, or on the shields in the other cases. The dose rate reduction

factor reported in the last column is the ratio of the dose rate measured at the sample position without the shields to that with the shield.

G. Esposito et al. / Advances in Space Research 35 (2005) 243–248 245

mol/dm3 EDTA, pH 8.0, 1% sarkosyl, 0.5 mg/ml pro-

teinase-K) for 1 h at 4 �C followed by overnight incu-

bation at 50 �C. Then, plugs were washed three times

with Tris–EDTA buffer and stored at 4 �C in 4 ml of

0.5 mol/dm3 EDTA solutions, pH 8.0, until they were

subjected to calibrated Pulsed Field Gel Electrophore-

sis (PFGE) for DSB evaluation. DNA fragments were

separated using two different electrophoretic condi-tions. In the first one, electrophoresis was performed

in 0.75· TAE buffer (40 mM Tris–acetate, 2.0 mM

EDTA, pH 8.0) at 14 �C for 44 h at 2 V/cm (switch

time 1200–2400 s, angle 106�), followed by 4 h at 6

V/cm (switch time 7–114 s, angle 120�) (Newmann

et al., 1997). In the second condition, electrophoresis

was performed in 0.5· TBE buffer (44.5 mM Tris,

44.5 mM boric acid, 1.0 mM EDTA, pH 8.0) at14 �C for 18 h at 6 V/cm (switch time 50–90 s, angle

120�) (Lobrich et al., 1996).

After electrophoresis, each gel lane was cut in corre-

spondence of specific molecular weight markers chosen

to have conveniently spaced zone. The fraction of

DNA mass in each zone was derived by the normalized

activity in that zone measured by a liquid scintillation

counter, and the corresponding number of fragmentswas evaluated dividing the fraction of DNA mass by

the mean molecular weight of that zone. Then, the total

number of fragments (practically equal to the total num-

ber of DSB) induced in the range 5.7 Mbp to 23.1 kbp

was obtained by summing up the contributions of the

different zones considering both electrophoretic

conditions.

More details of the experimental procedure and DSBcalculation can be found in Belli et al. (2002).

3. Results

3.1. Effectiveness of unshielded and shielded Fe beams

When the number of DSB per unit DNA length in-

duced by Fe ions, in absence or in presence of the differ-

ent shields, is plotted versus dose and versus particle

fluence incident on the sample (unshielded beams) oron the shield, linear relationships are obtained in all

cases (data not shown). The slope values, yD and yF,

respectively, obtained from best fits of the data are listed

in Table 2. It is noted that yF represents the effect cross-

section per unit DNA length. The same table also re-

ports the RBE for unshielded and shielded beams, ob-

tained from the yield values yD and from the c-raysvalue yD = 5.27 ± 0.41 · 10�3 Mbp�1 Gy�1 (Dini et al.,in press). Last column in Table 2 lists the shielding pro-

tection factor (SPF), defined as the ratio between the yFvalues of unshielded and shielded beam. The conversion

from dose at the sample position to fluence on the

shield, obtained as explained at the end of Section 2.2,

allows one to find the following relation: yF(un-

shielded)/yF(shielded) = SPF = DRRF Æ yD(unshielded)/yD(shielded).

3.2. RBE–LET relationships

Fig. 1 shows the RBE for unshielded and shielded Fe

ions as a function of LD and of LT. In both cases there is

a scarce dependence of the RBE on these parameters.

This circumstance, together with the extent of the error

bars, does not allow to precisely identify an RBE maxi-mum, although the data suggest it could be in the LD

Table 2

DSB yields (yD), RBE, cross-sections values per unit length for DSB induction (yF) of the unshielded and shielded beams, and shielding protection

factors (SPF)

E (MeV/u) Shield yD (109 bp Gy)�1 RBE yF (10�7 bp�1 lm2) SPF

Material mm (g/cm2)

115 No – 6.45 + 1.05 1.23 + 0.22 4.55 ± 0.74 –

414 No – 6.79 ± 0.69 1.29 ± 0.16 2.17 ± 0.22

PMMA 23 (2.7) 7.07 ± 0.54 1.34 ± 0.15 2.37 + 0.18 0.92 ± 0.12

PMMA 43 (5.1) 6.86 ± 0.51 1.30 + 0.14 2.66 ± 0.20 0.82 ± 0.10

PMMA 56 (6.6) 5.33 ± 0.50 1.01 ± 0.12 2.99 ± 0.28 0.73 ± 0.10

Al 30 (8.1) 4.77 ± 0.67 0.91 ± 0.14 2.61 ± 0.36 0.83 ± 0.14

Pb 10 (11.3) 7.25 ± 0.76 1.38 ± 0.18 3.88 ± 0.41 0.56 ± 0.08

1000 No – 6.95 ± 0.68 1.32 ± 0.17 1.58 ± 0.16 –

PMMA 197 (23.2) 6.16 ± 0.62 1.17 + 0.15 0.85 + 0.08 1.87 ± 0.26

Al 97 (26.2) 6.81 ± 0.89 1.29 ± 0.20 1.14 + 0.15 1.38 ± 0.23

Pb 26 (29.5) 8.72 ± 0.58 1.66 + 0.17 1.74 ± 0.12 0.91 ± 0.11

5000 No – 9.00 ± 0.72 1.71 ± 0.19 1.32 ± 0.11 –

PMMA 249 (29.4) 7.81 ± 0.56 1.48 ± 0.16 0.31 + 0.02 4.23 ± 0.45

Al 111 (30.0) 8.20 ± 0.59 1.56 + 0.16 0.61 ± 0.04 2.19 ± 0.24

Pb 26 (29.5) 7.13 ± 0.51 1.35 ± 0.14 0.94 ± 0.07 1.41 ± 0.15

Fig. 1. RBE values for DSB induction plotted versus (a) the dose-

average LET (LD) or (b) the track-average LET (LT). Unshielded

beams: open circles; PMMA shielded beams: solid triangles; Al

shielded beams: solid circles; Pb shielded beams: solid squares. The

points and the errors have been obtained as explained in the text.

Fig. 2. Shielding protection factor (SPF) plotted versus the shield

thickness expressed in g/cm2. PMMA shielded beams: solid triangles;

Al shielded beams: solid circles; Pb shielded beams: solid squares.

246 G. Esposito et al. / Advances in Space Research 35 (2005) 243–248

range 120–170 keV/lm. The second column of Table 1denotes the number of independent experiments per-

formed in each case. When more than one experiment

was available the corresponding point in Figs. 1(a) and

(b) is the average over the experiments and the error is

the related standard error. When only one experiment

was available the average and the standard error were

obtained from the values measured in 2 or 3 indepen-

dent electrophoretic runs.

3.3. Effect of shielding

In order to compare the efficiency of the different

shields used, the SPF has been plotted as a function of

the shield thickness expressed in g/cm2 (Fig. 2). Different

relationships were found for beams with different en-

ergy. In particular, for 414 MeV/u Fe ions, the presence

of the shields gives SPF lower than, or close to, unity,

independently of thickness and material. The results also

G. Esposito et al. / Advances in Space Research 35 (2005) 243–248 247

suggest that SPF values decreases on increasing the

PMMA thickness, although they are affected by large

uncertainties. The Student�s t-test, performed to evalu-

ate the statistical significance of the observed difference

between the SPF values for 6.6 and for 2.7 g/cm2

PMMA, showed it is significant at a confidence levelof 90%, but not at a confidence level of 95%, indicating

that the SPF decrease with the thickness is likely,

although the error bars do not allow to exclude the con-

trary with a very high level of confidence. In contrast to

the 414 MeV/u Fe beam, the SPF values measured with

1 and 5 GeV/u Fe beams are higher than or, in one case

only, close to unity. For these energies, PMMA gives a

higher SPF than Al or Pb shields having similar thick-ness. The higher protection exerted by PMMA is espe-

cially observed for the 5 GeV/u Fe beam.

4. Discussion

Although the PFGE technique used in this study to

analyse DNA fragmentation requires high doses in or-der to produce DNA fragments in amounts large en-

ough to be detected in a wide range of molecular

weights, it has the advantage of allowing direct calcula-

tion of the number of DSB (Lobrich et al., 1996), with-

out any assumption on the breakage mechanism. Most

of the alternative methods, on the contrary, usually per-

form such evaluation assuming a random breakage

mechanism for DSB induction. However, this is not ex-pected to hold, especially for high-LET radiation.

Moreover, the validity for low doses of the DSB

yields as measured by PFGE is further supported by a

recent paper (Rothkamm and Lobrich, 2003). The

authors evaluated the DSB induction and repair in hu-

man cells exposed to very low doses of X-rays, and

showed that the formation of c-H2AX foci (that are ex-

pected to represent DSB) is linearly related to the dosein the range 1.2 mGy to 2 Gy, with a regression line hav-

ing the same slope as that obtained by measuring DSB

with PFGE in the dose range 10–80 Gy.

The experimental data obtained with Fe ion irradia-

tion show no clear RBE dependence on LET, both in

terms of LD and LT. Only a flat maximum between

120 and 170 keV/lm can be envisaged. It can be noted

that also parallel experiments concerning chromosomeaberrations in lymphocytes (Durante et al., in press)

and cell death in AG1522 cells (Bettega, personal com-

munication) did not show good correlation of RBE with

LET, although RBE values are larger than those here re-

ported for DSB induction. In effect, the RBE for DSB

induced by Fe beams may be underestimated because

of the limitation in the measured fragment size range.

In fact, recent measurements of small fragments (1–23kbp) induced by 1 GeV/u Fe ions (either without shield

or after a PMMA shield) have shown that they contrib-

ute for about 30% of the total measured fragments,

while a much smaller contribution was observed in this

range after irradiation with c-rays (Dini et al., in press).

Another source of RBE underestimation may come

from the heat-labile site production (Rydberg, 2000)

during the lysis treatment, that gives probably a moresignificant contribution in c-irradiated than in Fe-irradi-

ated cells. However, these corrections would not appear

large enough to bring the values of the RBE for DSB

close to those for chromosome aberrations. This is con-

sistent with the idea that cellular effects such as chromo-

some aberrations and cell inactivation are related to

unrepaired or misrepaired DNA damage, rather than

to the initial DNA damage, and that the proportion of(unrepaired + misrepaired)/initial DNA damage de-

pends on radiation quality. This observation points

out a possible role of features of DSB induction other

than their number, such as their spatial correlation, that

can affect damage processing and reparability. Parallel

experiments are underway aimed at determining and

analyzing the size distribution of DNA fragments in-

duced by Fe beams of various energies, either with orwithout shielding.

Although the RBE values are affected by large exper-

imental uncertainties, differences are observed in RBE

among different beams having almost the same LD or

LT, which can reasonably be ascribed to their different

composition. This implies that the data obtained with

different beams cannot be adequately described by a

common RBE–LET relationship, neither in terms ofLD nor of LT.

Insertion of shields modifies the DNA breakage ob-

served after a given number of particles incident on

the shield, as expressed by the yF values listed in Table

1. This effect is ascribed to two factors, namely the

change in the dose to the sample per particle incident

on the shield (i.e., change in dose rate on sample) and

that in biological effectiveness. Both of them reflect vari-ations in the beam quality (i.e., types and energies of the

beam components) due to the shield traversal.

For 1 and 5 GeV/u Fe ions, nuclear interactions caus-

ing projectile and target fragmentation are dominant.

For example, it can be evaluated that only 9% of the pri-

mary 1 GeV/u Fe ions survive after passing the 19.7 mm

thick PMMA shield (Miller, personal communication).

The presence of lighter fast ions (of relatively lowLET) in the shielded beams then explains the decrease

in dose rate. On the contrary, for the less energetic

beams, namely 414 MeV/u Fe ions, insertion of the

shield increases the dose rate on the sample, suggesting

that the LET increase due to the slowing down of the

primary particles dominates over the LET decrease

due to fragmentation.

As shown in Section 3.1, SPF reflects variations inboth yD and dose rate. For 5 and 1 GeV/u Fe beams,

there is an overall protection after the shields, due to

248 G. Esposito et al. / Advances in Space Research 35 (2005) 243–248

the significant dose rate reduction after shield insertion.

A similar consideration can explain the SPF reduction

(to less than unity) observed with 414 MeV/u beam,

since in this case the dose rate increase after shield inser-

tion gives the main contribution. Moreover, it can be

noted that with this beam the protection offered by thePMMA shield does not increase (rather, it probably de-

creases) on increasing its thickness.

Results obtained with the same beam and shields of

comparable thickness (in g/cm2) but made of different

materials show that protection depends on the material,

PMMA resulting more effective than Al and Pb.

In conclusion, the present study showed that: (i) the

yield of DSB induced by HZE particles depends weaklyon both LD or LT; (ii) the insertion of a shield can de-

crease or increase the DSB induced per particle incident

on the shield, depending on the beam energy, mainly be-

cause of changes in the dose rate; (iii) PMMA in general

offers better protection than Al and Pb against DSB

induction.

Finally, it can be concluded that it is not obvious that

shield insertion necessarily offers a protection againstbiological effects of space radiation, and that radiobio-

logical experiments such as those here described are

important to develop a general strategy to reduce the

radiation risk to the astronauts by means of shielding.

Acknowledgements

This work was supported by the Italian Space Agency

(Project ‘‘Influence of the shielding on the space radia-

tion biological effectiveness’’) and by NASA (Project

‘‘Space radiation shielding: biological effects of heavy

ions after traversal of different shielding materials’’.)

The authors are indebted with M. Durante for coordi-

nating these projects and to W. Schimmerling and

F.A. Cucinotta for their encouragement and support.The authors are also indebted to J. Miller, C. Zeitlin,

C. La Tessa and T. Kanai for the beam parameter eval-

uation at BNL and at NIRS. Thanks are due to M. E.

Vazquez for the excellent laboratory support and to

the AGS and NSRL staff at BNL, as well as to the HI-

MAC staff at NIRS.

References

Belli, M., Cherubini, R., Dalla Vecchia, M., Dini, V., Esposito, G.,

Moschini, G., Sapora, O., Simone, G., Tabocchini, M.A. DNA

fragmentation in V79 cells irradiated with light ions as measured by

PFGE. I. Experimental results. Int. J. Radiat. Biol. 78, 475–482,

2002.

Dini, V., Antonelli, F., Belli, M., Campa, A., Esposito, G., Simone, G.,

Sorrentino, E., Tabocchini, M.A. Influence of PMMA shielding on

DNA fragmentation induced in human fibroblasts by iron and

titanium ions. Radiat. Res. (in press).

Durante, M., Gialanella, G., Grossi, G., Pugliese, M., Scampoli, P.,

Kawata, T., Furusawa, Y. Influence of the shielding on the space

radiation biological effectiveness. II. Chromosomal aberrations.

Radiat. Res. (in press).

Durante, M., Gialanella, G., Grossi, G., Pugliese, M., Scampoli, P.,

Kawata, T., Yasuda, N., Furusawa, Y. Influence of the shielding

on the induction of chromosomal aberrations in human lynpho-

cytes exposed to high-energy iron ions. J. Radiat. Res. 43, S107–

S111, 2002.

Goodhead, D.T. Initial events in the cellular effects of ionizing

radiations: clustered damage in DNA. Int. J. Radiat. Biol. 65, 7–17,

1994.

Grossi, G., Durante, M., Gialanella, G., Pugliese, M., Scampoli, P.,

Furusawa, Y., Kanai, T., Matsufuji, N. Chromosomal aberrations

induced by high-energy iron ions with shielding. Adv. Space Res.

34, 1358–1361, 2004.

Holley, W.R., Chatterjee, A. Clusters of DNA damage induced by

ionizing radiation: formation of short DNA fragments. I. Theo-

retical modeling. Radiat. Res. 145, 188–199, 1996.

Lobrich, M., Cooper, P.K., Rydberg, B. Non-random distribution of

DNA double-strand breaks induced by particle irradiation. Int. J.

Radiat. Biol. 70, 493–503, 1996.

Newmann, H.C., Prise, K.M., Folkard, M., Michael, B.D. DNA

double-strand break distributions in x-ray and a-particle irradiatedV79 cells: evidence for non-random breakage. Int. J. Radiat. Biol

71, 347–363, 1997.

Rothkamm, K., Lobrich, M. Evidence for a lack of DNA double-

strand break repair in human cells exposed to very low X-ray

doses. PNAS 100, 5057–5062, 2003.

Rydberg, B. Radiation-induced heat-labile sites that convert into DNA

double-strand breaks. Radiat. Res. 153, 805–812, 2000.

Wilson, J.W., Kim, M., Schimmerling, W., Badavi, F.F., Thibeault,

S.A., Cucinotta, F.A., Shinn, J.L., Kiefer, R. Issues in space

radiation protection: galactic cosmic rays. Health Phys. 68, 50–58,

1995.

Wilson, J.W., Konradi, A., Miller J., Cucinotta, F.A. Shielding

Strategies for Human Space Exploration. NASA CP-3360, Spring-

field, VA, 1997.

Zeitlin, C., Heilbronn, L., Miller, J. Detailed characterization of the

1087 MeV/nucleon Iron-56 beam used for radiobiology at the

Alternating Gradient Synchroton. Radiat. Res. 149, 560–569, 1998.