Does radiotherapy affect the outcome of the comet ... - BIR Publications

The British Journal of Radiology, 75 (2002), 608–614

E

2002 The British Institute of Radiology

Does radiotherapy affect the outcome of the comet assay? 1

¨ LLER, PhD, 1T BAUCH, PhD, 1C STREFFER, PhD and 2D VON MALLEK, MD W-U MU

1

Institut fu¨r Medizinische Strahlenbiologie und 2Klinik und Poliklinik fu¨r Strahlentherapie, Universita¨tsklinikum Essen, D45122 Essen, Germany

Abstract. This study was designed to assess possible effects of fractionated radiotherapy (5 or 10 fractions at 2 Gy per fraction) on the DNA repair capacity of lymphocytes, as measured by the comet assay. 50 patients with various tumour types were chosen. They had received no chemotherapy during the 6 months prior to radiotherapy and did not receive cortisone. 10 ml of heparinized blood was collected before radiotherapy, after 5 fractions and after 10 fractions. Lymphocytes were isolated and analysed using the comet assay. On average, no effect on DNA repair capacity was observed that could be attributed to radiotherapy. On an individual basis, there were a few patients who showed a comparatively pronounced variability in their response to radiotherapy (three patients with a relative coefficient of variability of more than 30%). There was some indication of a weak correlation between poor repair capacity and severe side effects in normal tissue. We also found that alcohol in particular, and smoking to some extent, may impair repair capacity during radiotherapy. Age, gender, field size, medication and tumour entity showed no effect on repair capacity. In recent years the so-called ‘‘comet assay’’, or single cell gel electrophoresis, has gained considerable interest in various fields of application [1–3]. Among these, the assay is used to give an explanation for unexpectedly pronounced side effects during radiotherapy [4, 5] or, as it is hoped, to predict individual radiosensitivity before radiotherapy starts [6, 7]. The reason for using the assay in this context is its ability to provide information regarding the repair capacity of single cells. Detailed descriptions of the comet assay can be found in the literature [8]. The assay determines the amount of DNA damage (single and double strand breaks and conformational changes) in a radiation-exposed cell after removing most of the non-DNA material of this cell and applying a weak electric field to the remaining DNA embedded in an agarose gel. Dependent on the amount of DNA damage, DNA partly moves into the ‘‘comet tail’’ (Figure 1). When DNA damage is higher, more DNA is found in the tail. The comet assay can be used to measure initial and residual levels of radiation-induced damage, and the rate of DNA repair following irradiation of cells. In lymphocytes, 3 h at 37 ˚C is usually sufficient to reduce the amount of DNA in the tail

induced by 2 Gy X-rays almost to control levels [9]. In the past, the lymphocytes of several patients who reacted with severe side effects in the normal tissue during or after radiotherapy have been analysed [5]. In several of these cases, a poor repair characteristic of the lymphocytes was detected, i.e. repair was slow and high residual damage after 3 h at 37 ˚C was observed. As these assays were run following some fractions of radiation, the pertinent question emerged whether radiotherapy itself was responsible for the outcome of the comet assay. Therefore we decided to check this possibility systematically. The lymphocytes of 50 radiotherapy patients were analysed before their first treatment, after 5 fractions and after 10 fractions (at 2 Gy per fraction). Exclusion criteria were chemotherapy during the past

Received 2 November 2001 and in revised form 21 March 2002, accepted 2 April 2002. Present address for D von Mallek: Klinik fu¨r Nuklearmedizin, Universita¨t Bonn, Sigmund-FreudStraße 25, D53127 Bonn, Germany. 608

Figure 1. Comet immediately after a 2 Gy exposure in vitro. The British Journal of Radiology, July 2002

Radiotherapy and the comet assay

6 months and cortisone treatment accompanying radiotherapy. Analyses were not restricted to a specific tumour entity. In addition to the central question of whether radiotherapy itself affects, on average, repair kinetics of lymphocytes as measured by the comet assay, the following questions were also addressed:

N N N

Are there individual differences in repair response in the course of radiotherapy? Is there a relation between repair capacity and side effects? Do field size, age, gender, alcohol, smoking, medication and tumour entity affect repair kinetics?

Conclusions were based on the shape of the repair curves over 3 h of repair, and on the area under the repair curve (the smaller the area under the curve, the better the repair).

Materials and methods Patients Intentionally, blood was collected from patients affected by various tumour diseases (Table 1) and analyses were not restricted to one tumour entity. All 50 patients gave their informed consent to this study. Heparinized blood was collected immediately prior to the first radiotherapy fraction, then again Table 1. Tumour entities studied Region

Entity or site

No. of patients

Head and neck

Oropharynx Salivary gland Hypopharynx Larynx Oral cavity Auditory tube Neck metastasis Maxillary sinus Floor of the mouth Oro-hypopharynx Cheek Thyroid glanda Orbita Meningiom Bronchial Breast Cervix Prostate Anal NHL abdomen NHL stomach Ovary

6 5 3 3 2 1 1 1 1 1 1 1 2 1 3 6 2 2 1 1 1 1 4

Brain/orbita Thorax Breast Abdomen

Bone/soft tissue a

This tumour was not included in analysis because therapy had been recently carried out.

The British Journal of Radiology, July 2002

after the fifth and tenth fractions (2 Gy per fraction). Patients who had received chemotherapy within 6 months prior to radiotherapy were excluded, as were patients receiving cortisone. Severity of side effects was scored according to Sauer [10]; 05without effect, 15marginal effects, 25medium effects, 35severe effects, and 45life threatening effects.

Comet assay The procedure of the comet assay is described in detail by Bauch et al [9]. Only those aspects necessary for the understanding of this publication will be addressed here. Unstimulated lymphocytes were isolated from heparinized blood samples by using the Ficoll– Hypaque technique [11] and cultured in vitro for 1 day. Such a strategy is advisable because the isolation stress itself is sometimes sufficient to induce DNA damage that can be detected in the comet assay. In order to assess DNA damage and DNA repair, isolated lymphocytes were exposed to 2 Gy of X-rays generated by a Stabilipan X-ray machine (Siemens, Erlangen, Germany) at 240 kVp, a dose rate of 1 Gy min21 and a copper filtration of 0.5 mm. Exposure was carried out on ice. Those cells that were checked for initial radiation damage were suspended immediately after radiation exposure in warm agarose (0.75%) for several seconds and distributed on a slide that had been precoated with agarose (0.1%). After gelling at 4 ˚C, all non-DNA material was removed in a sodium dodecylsulfate solution (pH 9.5) for 15 min at 18 ˚C. After washing in distilled water at 18 ˚C, electrophoresis was carried out at pH 8.4, 2.5 V cm21 and 16 mA, for 4 min at 10 ˚C. Electrophoresis conditions were chosen in such a way that just a very short tail appeared in unexposed DNA, which ensures small additional effects by, for example, low radiation doses, are detected. Slides were washed again, dried and stored at 220 ˚C. Before analysis the slides were rehydrated in distilled water and the comets stained with propidium iodide. Those lymphocytes whose repair capacity was monitored were allowed to repair DNA damage for up to 3 h at 37 ˚C in medium. At various time intervals after radiation exposure the cells were transferred into agarose and processed as described above.

Analysis of comets and statistics

131

Iodine

At least 40 comets per data point were analysed using a self-designed, interactive image analysis system [12]. The ratio of the amount of DNA in the comet tail to that in the head region was used 609

W-U Mu¨ller, T Bauch, C Streffer and D von Mallek

Results

AuC[a.u.]

Does radiotherapy itself affect the results of the comet assay?

as a measure for the amount of DNA damage, i.e. the higher the amount of DNA in the tail, the more DNA damage. Repair capacity was assessed by plotting residual damage after specific times of repair (up to 3 h) and calculating the area under the respective repair curves. Area under curve (AuC) shows a very strong correlation to residual damage after 3 h of repair (Figure 2). Statistics were performed using GraphPad software Prism (version 3.01) (GraphPad Software Inc., San Diego, CA). Most comparisons were run using Student’s two-sided t-test or one-way analysis of variance; in a few clearly nonnormal distributions the Kruskal-Wallis test was applied. The relative inter-patient coefficient of variance (CV) was 14.3%, the relative average intra-patient CV 15.3%.

The question of the effect of radiotherapy on the comet assay triggered these analyses. If one looks at the mean values of those patients who took part in this study (Figure 3), the answer is straightforward; there is no indication that, on average, the repair kinetics of lymphocytes are affected by conventionally fractionated radiotherapy after 5 and 10 fractions of 2 Gy. This is obvious from the shape of the repair curves (Figure 3a) and from the AuCs (Figure 3b). One should keep in mind that the same patients were studied before therapy and after 5 and 10 fractions. There is a certain possibility that Figure 3a is somewhat misleading, because initial damage was normalized, i.e. the damage seen immediately after a 2 Gy in vitro exposure, to 100%, as is frequently done in comet assay studies. Thus, if initial DNA damage in the lymphocytes after 5 or 10 fractions were higher than that of the lymphocytes before radiotherapy, the number of damaged sites to be repaired would be higher. However, Figure 4 demonstrates the amount of initial damage after 2 Gy in vitro exposure is the same for all three data sets. On an individual basis, there was some variability in response to radiotherapy. Part of this variability is due to experimental errors. In three cases we found a relative coefficient of variability of more than 30%. However, we could not identify any reason why, in these three cases, variability was so high.

(a)

(b)

residual damage after 3 h of repair [%] Figure 2. Relation between residual damage after 3 h of repair and area under curve (AuC). (---) 95% confidence intervals of the linear regression line; r250.62; degrees of freedom544; p,0.01). a.u., arbitrary units.

Figure 3. Repair characteristics of lymphocytes of tumour patients before therapy (bT), after 5 fractions of 2 Gy (dT/10) and after 10 fractions of 2 Gy (dT/20). (a) Repair kinetics up to 180 min. Figure shows arithmetic means and 95% confidence intervals. Due to methodological problems not all of the 49 patients under study could be analysed. n, number of successfully analysed patients. (b) Area under curve. Arithmetic means and 95% confidence intervals are shown. Differences among the three groups are not significant (p.0.05) using analysis of variance. a.u., arbitrary units. 610


area under curve [a.u.]


Figure means before dT/20,

4. Initial DNA damage after 2 Gy. Arithmetic and 95% confidence intervals are shown. bT, therapy; dT/10, after 5 fractions of 2 Gy; after 10 fractions of 2 Gy.

Is there a relation between repair capacity and side effects?

Field size (cm2) Figure 6. Effect of field size on repair capacity. Comparison was performed between the field sizes lower than the median and higher than or equal to the median. The two right-hand columns show a comparison between field sizes exceeding 1000 cm2 and all smaller field sizes. a.u., arbitrary units.

The background for analysing patient lymphocytes by comet assay is the idea to identify individuals who may react with severe side effects to radiotherapy. Thus, it is of some interest to look for a possible relation between repair capacity and side effects. Severe side effects are, fortunately, the exception and not the rule and therefore few people are affected. This means that a group of just 50 patients is far too small to obtain a definite answer. But a certain trend might be expected even in such a small group. And, indeed, there are two indications that support this basic idea. First, there is a small, although non-significant, increase in the median of the groups of patients with no or only marginal, medium, and severe side effects (Figure 5a). Second, the patient with the worst repair kinetics (Figure 5b, the uppermost curve) had severe side effects. One must keep in mind

that this study was not designed to give evidence for a relationship between repair capacity and side effects; to do this many more individuals would need to be analysed. This study was carried out to look for possible effects of radiotherapy itself on the outcome of the comet assay.

(a)

(b)

Do other factors affect repair kinetics? There are, of course, other factors that might affect the comet assay and it is interesting to look for such factors in a group of 50 patients. First of all, the effect of the size of the area exposed to radiation during therapy was checked. One might expect that increasing field size has some impact, because the dose on the circulating lymphocytes will increase. Figure 6 reveals that no such effect can be observed. Irrespective of

Figure 5. Individual repair characteristics before radiotherapy. (a) Box and whisker plots of the area under curve of Figure 3b in dependence on the severity of side effects (0+1, n515; 2, n514; 3+4, n510). (b) Repair kinetics up to 180 min of all patients analysed. a.u., arbitrary units. The British Journal of Radiology, July 2002

611

W-U Mu¨ller, T Bauch, C Streffer and D von Mallek Table 2. Effects of various factors on the repair capacity of lymphocytes before and during radiotherapy Factor

AuC¡SD (n)

Age

,median

¢median

bT dT/10 dT/20

42.4¡12.3 (23) 43.4¡13.7 (23) 51.0¡14.7 (22)

45.8¡18.2 (23) 40.2¡10.6 (24) 43.4¡15.9 (21)

Gender

Female

Male

bT dT/10 dT/20

45.1¡14.2 (18) 43.2¡13.4 (20) 41.7¡14.3 (16)

43.4¡16.4 (28) 40.7¡11.3 (27) 50.5¡15.7 (27)

Alcohol

No

Yes

bT dT/10 dT/20

39.7¡16.9 (7) 39.8¡12.6 (8) 47.2¡9.0 (7)

41.1¡13.7 (6) 41.7¡8.8 (6) 65.8¡19.7 (6)

Smoking

No

Yes

bT dT/10 dT/20

43.6¡14.8 (13) 38.9¡11.2 (14) 45.4¡8.5 (12)

50.0¡22.8 (11) 43.0¡11.7 (12) 55.6¡18.7 (11)

Drugs

No

Yes

bT dT/10 dT/20

46.6¡13.3 (6) 40.9¡12.0 (5) 46.8¡10.3 (6)

42.7¡13.1 (28) 40.7¡12.2 (30) 44.3¡17.0 (31)

Tumour entity

Mamma carcinoma

All others

bT dT/10 dT/20

44.9¡8.3 (4) 40.1¡16.5 (5) 44.2¡8.2 (6)

44.0¡16.0 (42) 42.0¡11.8 (42) 44.3¡9.6 (43)

AuC, area under the repair curve; SD, standard deviation.

field size, repair capacity is the same. This also holds true for large areas greater than 1000 cm2. In addition, the following factors were checked: age, gender, alcohol, smoking, medication, and tumour entity (Table 1). No effect whatsoever could be detected for age, drugs or tumour entity. Tumour entity means that we checked the repair characteristics of patients with the most frequent tumour type in our study (mamma carcinoma, n56) against all other tumours types. In addition, 26 head/neck tumours were tested against all other tumour types; again, no significant differences were observed. One statistically significant difference (p,0.05) was found; alcohol consumption resulted in markedly worse repair after 10 fractions of radiotherapy. There was a similar response in the same direction for smoking, but this response was not significant (p50.10). Two aspects, however, have to be kept in mind when interpreting these results. First, this study was not specifically designed to test for alcohol and smoking effects, therefore there is the possibility that we observed significance just by chance. Second, it was very difficult to obtain reliable information, in particular on alcohol consumption, so that the groups studied 612

are very small (n56 for alcohol consumers, n58 for abstainers). At first glance there seems to be a small effect through gender (p50.07 for the difference between females and males after 10 fractions). This difference, however, is probably owing to the fact that all 6 alcohol consumers were men, and that 9 of the 11 smokers were also men. Thus, there seems to be no effect due to gender, but an effect due to alcohol and smoking.

Discussion Radiation therapists would like to have a tool that allows the quantitative assessment of radiosensitivity of normal tissues. Meanwhile, research has identified quite a number of factors that affect individual radiation sensitivity [13]. One of these factors is definitely the repair capacity of cells. The role of repair is most evident in a number of genetic syndromes with hypersensitivity against radiation [14]. The comet assay gives valuable information on repair characteristics, a fact that initiated several studies to check the role of this assay in the assessment of individual radiosensitivity [4, 5, 7, 15–18]. In several cases the comet assay is applied during radiotherapy and, therefore, it is reasonable to systematically look for possible interference of radiotherapy itself on the assay. On average, radiotherapy did not affect the outcome of the assay, although lymphocytes, the cell type used in this study, were exposed to radiation during therapy. Thus, one could expect some radiation damage in the lymphocytes analysed making the results of the comet assay unreliable. There are at least two reasons why this did not happen. First, those lymphocytes that were exposed during therapy had plenty of time to repair radiation damage because they were used for the assay, at the earliest, 24 h after the last fraction. Second, those lymphocytes that were severely damaged during radiotherapy had a high risk of dying, so that we did not see them in the assay at all. One is, of course, also inclined to use the data obtained in a study of 50 patients to answer questions for which the analyses were not originally designed. However, one must be careful when proceeding in such a way as applying many statistical tests to the same data set will result in significant results simply by chance. The number of those chance outcomes depends on the chosen significance level. Thus, the best that can be hoped for is to obtain the opportunity to generate new hypotheses that can be tested in a second approach. In our study, the effect of alcohol and smoking were conspicuous in so far as both showed either a trend (smoking) or a significant The British Journal of Radiology, July 2002


effect (alcohol) on repair after 10 fractions (20 Gy) of radiotherapy. One could assume some interaction between radiation and alcohol/ smoking. As outlined above, however, this hypothesis should be specifically tested in another study. The other factors checked (age, gender, medication, tumour entity and field size) did not show an effect on repair either before or during radiotherapy. This, of course, can also be a chance finding, because in a series of analyses there will always be some results that are insignificant although actually a difference exists. For age, gender and medication, however, this result agrees with the result of a previous series of analyses [5]. Fortunately, severe side effects through radiotherapy are comparatively rare. Thus, in a cohort of just 50 patients it is highly unlikely to find several ‘‘over reactors’’. But one can look, at least, for a relationship between poor repair capacity and side effects. A very weak relation was actually observed and that individual with the worst repair characteristic indeed had severe side effects. This supports our finding that patients with poor repair characteristics show a high risk of reacting with severe side effects, whereas not all patients with severe side effects also show a poor repair capacity [5]. This latter observation is caused by the multiple factors that are responsible for the overall radiation sensitivity, e.g. genetic predisposition, genomic instability and oxygen tension, and repair is just one of these factors. A very similar conclusion was obtained by Alapetite et al [4], who specifically addressed the question of a possible relation between over reacting to radiation therapy and poor repair capacity. They analysed the lymphocytes of 17 breast cancer and 9 Hodgkin’s disease cases using the alkaline version of the comet assay, and compared results with those of 22 patients with average reactions and 24 healthy donors. 7 of the 17 breast cancer over reactors reproducibly showed very slow repair during the first hour, and the most severe cases were found among these 7 patients. Only one of the nine Hodgkin’s disease patients showed poor repair kinetics, but again this was the case with the most severe reaction. These results also point to our aforementioned conclusion that not all severe side effects can be explained by poor repair characteristics. However, one should be very cautious if an individual reveals such poor repair characteristics. A second study of the same group on thyroid tumour risk after a history of head and neck irradiation in childhood also indicates that a sub-group of these individuals (11 of 13) shows poor repair kinetics, The British Journal of Radiology, July 2002

linking risk of a secondary tumour after radiotherapy and impaired repair capacity [19]. Obviously, several assay systems are necessary to obtain a complete, quantitative picture of individual radiosensitivity. Studies such as that presented here suggest that the comet assay may be one component of such a battery of test systems.

References 1. McKelvey-Martin VJ, Green MHL, Schmezer P, Pool-Zobel BL, De Me´o MP, Collins A. The single cell gel electrophoresis assay (comet assay): a European review. Mutat Res 1993;288:47–63. 2. Fairbairn DW, Olive PL, O’Neill KL. The comet assay: a comprehensive review. Mutat Res 1995; 339:37–59. 3. Olive PL. DNA damage and repair in individual cells: applications of the comet assay in radiobiology. Int J Radiat Biol 1999;75:395–405. 4. Alapetite C, Thirion P, de la Rochefordie`re A, Cosset J-M, Moustacchi E. Analysis by alkaline comet assay of cancer patients with severe reactions to radiotherapy: defective rejoining of radioinduced DNA strand breaks in lymphocytes of breast cancer patients. Int J Cancer 1999;83:83–90. 5. Mu¨ller W-U, Bauch T, Stu¨ben G, Sack H, Streffer C. Radiation sensitivity of lymphocytes of healthy individuals and cancer patients as measured by the comet assay. Radiat Environ Biophys 2001;40: 83–9. 6. Streffer C, Bauch T, Stu¨ben G, Zo¨lzer F, Mu¨ller W-U. Predictive assays for individualization of cancer therapy. In: Kogelnik HD, Sedlmayer F, editors. Progress in Radio-Oncology VI. Bologna, Italy: Monduzzi, 1998:501–7. 7. Mu¨ller W-U, Bauch T, Streffer C. The application of comet and micronucleus assay for the assessment of individual radiation risk. In: Baumstark-Khan C, Kozubek S, Horneck G, editors. Fundamentals for the Assessment of Risks from Environmental Radiation. Dordrecht, The Netherlands: Kluwer Academic Press, 1999:127–34. 8. Bo¨cker W, Rolf W, Bauch T, Mu¨ller W-U, Streffer C. Automated comet assay analysis. Cytometry 1999;35:134–44. 9. Bauch T, Bo¨cker W, Mallek U, Mu¨ller W-U, Streffer C. Optimization and standardization of the ‘‘comet assay’’ for analyzing the repair of DNA damage in cells. Strahlenther Onkol 1999;175: 333–40. 10. Sauer R. Tumorklassifikation und -dokumentation. In: Scherer E, Sack H, editors. Strahlentherapie (4th edn). Berlin, Germany: Springer, 1996:269–91. 11. Boyum A. Isolation of leucocytes from human blood. Further observations. Methylcellulose, dextran, and ficoll as erythrocyteaggregating agents. Scand J Clin Lab Invest 1968;97 (Suppl.):31–50. 12. Bo¨cker W, Bauch T, Mu¨ller W-U, Streffer C. Image analysis of comet assay measurements. Int J Radiat Biol 1997;72:449–60. 13. National Radiological Protection Board. Genetic heterogeneity in the population and its implications for radiation risk. Chilton, UK: NRPB, 1999:1– 47. 613

W-U Mu¨ller, T Bauch, C Streffer and D von Mallek 14. Arlett CF. DNA repair defects and human radiosensitivity. In: Molekulare und Zellula¨re Prozesse bei der Entstehung Stochastischer Strahlenwirkungen, Bd. 33. Stuttgart, Germany: Gustav Fischer, 1995:229–37. 15. Betti C, Davini T, Giannessi L, Loprieno N, Barale R. Microgel electrophoresis assay (comet test) and SCE analysis in human lymphocytes from 100 normal subjects. Mutat Res 1994;307:232–3. 16. Marples B, Longhurst D, Eastham AM, West CM. The ratio of initial/residual DNA damage predicts intrinsic radiosensitivity in seven cervix carcinoma cell lines. Br J Cancer 1998;77:1108–14. 17. Eastham AM, Marples B, Kiltie AE, Orton CJ, West CM. Fibroblast radiosensitivity measured using the comet DNA-damage assay correlates with clonogenic survival parameters. Br J Cancer 1999;79:1366–71.

614

18. Myllyperkio MH, Koski TR, Vilpo LM, Vilpo JA. Gamma-irradiation-induced DNA single- and double-strand breaks and their repair in chronic lymphocytic leukemia cells of variable radiosensitivity. Hematol Cell Ther 1999;41:95–103. 19. Leprat F, Alapetite C, Rosselli F, Ridet A, Schlumberger M, Sarasin A, et al. Impaired DNA repair as assessed by the ‘‘comet’’ assay in patients with thyroid tumors after a history of radiation therapy: a preliminary study. Int J Radiat Oncol Biol Phys 1998;40:1019–26.
