The management of desmoid tumors

Proceedings

of the 34th Annual ASTRO Meeting

93 HUMAN

TUMOR

Simon

N.

Institute

CELL

Powell, of

Cancer

REPAIR M.D.,

FIDELITY

CORRELATES

WITH

RADIORESISTANCE

Ph.D.

Research,

Cotswold

Road,

Sutton,

Surrey,

U.K.

Purpose: The efficiency of DNA repair is thought to be a significant determinant

of the cytotoxic effect of ionizing radiation. Ionizing radiation induces a wide variety of lesions in DNA, but the double-strand break is the lesion most closely associated with cell killing. Using assays based on DNA fragmentation, the number of double-strand breaks is observed to decrease with time following exposure to ionizing radiation. Repair, as thus measured, often shows little difference between tumor cell lines of widely different radiosensitivity. The question then arises whether it is the quality (or fidelity) of repair that differs between these cell lines. The mechanism by which DNA fragments are rejoined is poorly understood. The aim of this project was to use plasmid transfection to ask the question, what is the balance between correct-repair and misrepair? Materials and Methods: Using techniques described by Debenham et al. (Mutation Res. 199: 145-58,19881 a double-strand break was introduced into a coding sequence of circular plasmid DNA using a restriction endonuclease as a model for a radiation-induced double-strand break; it was then transfected as a linear molecule into human tumor cells and the subsequent cell-mediated restoration of the coding sequence, evidenced by intact gene function, was documented. The plasmid used in these experiments, pPMH16, is known to integrate into genomic DNA. Gene function was tested by the ability to grow colonies in selection media. The plasmid also contains a second selectable marker gene which was used to identify transfected cells, before the function of the damaged gene was tested. The proportion of transfected cells which had correctly restored the damaged gene gave a measure of repair fidelity. Results: A general trend for sensitive cells to show lower repair fidelity relative to resistant cells was observed. The type of double-strand cleavage of the plasmid (straight across or a four-base overhang) made little difference to the measured repair fidelity, in contrast to published studies in which restriction-enzyme breaks had been introduced into chromatin DNA. Specific comparison of parent lines and their radiosensitive clones showed significant differences in repair fidelity for a relatively small change in radiation response. Furthermore, when these same pairs were compared by measuring the reduction in DNA fragmentation with time, no differences were found. These observations have been supported by Southern analysis of how plasmids were processed by the cell lines under comparison. The results were consistent with misrepair of double-strand breaks involving error-prone recombination. Misrepair due to exonuclease-mediated degradation was thought less likely, as misrepair could be found after transfection by circular plasmid. Although multiple (l-101 copies of the plasmid were integrated, the copy number was shown to be similar in the cell lines compared. The number of intact copies of the damaged gene relative to the undamaged gene mirrored the observed repair fidelity. There was a suggestion that fewer copies of blunt-ended linear plasmid were integrated, and that blunt-ends were more frequently misrepaired by deletion relative to staggered ends, where misrepair by small lesions or insertions was found. In a comparison of two equally radioresistant bladder cancer cell lines, large differences in repair fidelity were observed. No differences in the integrated copy number were found, but the damaged gene was highly rearranged or deleted in the cell line with low repair fidelity. It was noted that this cell line sustained fewer double-strand breaks immediately following irradiation: in this case the difference in repair fidelity may partly be counteracted by the difference in initial damage. Conclusions: These studies have shown repair fidelity to correlate closely with radiosensitivity, especially when comparing genetically related lines. It is suggested that repair fidelity can be a significant determinant of residual damage resulting from the processing of doublestrand breaks, and hence the response to ionizing radiation.

94 A Quantitative Assay for Radiation-Induced DNA Damage Based on DNA Fragment Size Using Pulsed Field Gel Electrophoresis Theodore S. Lawrence, Mary A. Davis, Daniel P. Noxmolle, and Jonathan Maybaum Department of Radiation Oncology, Division of Cancer Biology, University of Michigan Medical Center, Ann Arbor MI 48109 Purpose: Although pulsed field gel electrophoresis (PFGE) appears to be a powerful tool for investigating radiation-induced DNA damage, previous studies have quantified damage only in terms of the fraction of activity released from the well. However, quantification of the distribution of fragment size could yield important information concerning the mechanism of damage. The development of biphasic linear pulse ramping gel electrophoresis has permitted resolution of DNA fragments in the range of 200 kilobase pairs (kbp) and 6 Mbp in a single gel. Therefore, we used this technique to develop a quantitative assay for radiationinduced DNA damage based on fragment size. Method: Log phase HT29 cells labeled with 14C thymidine were trypsinized and immediately embedded in low melting point agarose blocks (0.7% in a buffer (TBE) of Tris borate and EDTA). Plugs were lysed and treated with proteinase K for 24 h at 55’ C and were irradiated using ~CO. PFGE was performed according to the method of VanDevanter (Nut. Acids Res.,l992), using biphasic linear pulse ramping. A calibration curve of molecular weight versus distance migrated from the well was constructed using S. pombe and S. cerevisiae chromosomes as standards. The gel was sliced into 18 pieces (4-8 mm long) which were counted for t4C by liquid scintillation. The molecular weight range in each gel slice was determined from the calibration curve and the measured distance from the well to the start and end of each slice. The fraction of DNA as a function of length was fit using the Levenberg-Marquardt algorithm to an equation which describes the expected distribution of fragment sizes produced by random breakage of a polymer (Blocher, D. Int. J. Radiat. Biol. 57:13,1990). This equation has 2 parameters: the DNA double strand break (DSB) induction frequency (IF) (DSB bp-’ Gy-l) and the modal length of the chromosome (bp).

179

180

Radiation

Oncology,

Biology, Physics

Volume 24, Supplement

1

Results: The S. oombe and S. cerevisiae chromosomes were seuarable mto dtscrete bands, defining a resolved region of 200 kbp to 6 Mbp. ‘Within this zone, the log of the molecular weight was linearly related to the distance migrated (correlation coefficient 0.99 f 0.01 (N=4)). A random breakage model produced excellent fits of the fraction of radiation-induced DNA fragments in each gel slice taken from the resol;ed region. It was found that the modal chromosome size (estimated at approximately 100 Mbp) had a minimal influence on the predicted fragment distribution. In contrast, the estimates for IF critically affected the fit. The estimates of IF in 3 experiments using 20 Gy were 9.4 + 0.4, 7.5 f 0.3, and 8.2 f 0.2 x 10e9 DSB bp-t my-1, which are in the range of those reported by others using far higher doses of radiation. Preliminary data suggest that it will be possible to obtain estimates of the IF using 5- 10 Gy doses. Conclusion: These data demonstrate that biphasic linear pulse ramping gel electrophoresis permits the experimental determination of the DSB induction frequencies using doses of radiation approaching those employed in the clinic. We anticipate that this technique will be applicable to a variety of investigations aimed at understanding radiation-induced DNA damage. Supported

by NIH grants CA 53440 and CA 44173 and by an American

Cancer Society Career Development

Award (91-210)

95 HEAT

SHOCK

PROTEIN

Gloria C. Li and Ching-Hsien Departments

HSWO

PROTECTS

CELLS

FROM

OXIDATIVE

STRESS

Lee

of Medical Physics

and Radiation

Oncology,

Memorial

Sloan Kettering

Cancer Center, New York, NY 10021

Purmse: Retroviral-mediated gene transfer experiments show that rodent cells become heat resistant when stably and constitutively expressing a cloned human gene encoding an intact human 70 kDa heat shock protein (hsp70). The purpose of this study is to investigate whether overexpression of hsp70 protects cells from oxidative stress. Specifically, we examine the hydrogen peroxide (H202)-induced cytotoxicity in Rat-l cells and transfected Rat-l cells which express high level of human hsp70 protein. Materials and Method%:Rat fibroblasts

(Rat-l)

were grown in Dulbecco’s

Modified

Medium

supplemented

with 10% fetal calf

serum. The transfected cells (M21, Bg2, Sm2) were routinely maintained in medium containing antibiotic G418 (200 pg/ml) [1,2]. M21 cells express intact human hsp7O; Bg2 cells express a mutant human hsp70 missing codons 128-418 (ATP-binding domain); Sm2 cells express a mutant human hsp70 missing codons 438-618 (nucleolar localization). For H202 cytotoxicity studies, monolayers of exponentially growing cells were exposed to graded doses of H202 at 370C for 1 hr. The cells were rinsed three times with phosphate buffered saline, uypsinized, counted, and plated for colony formation and determination of ce!l survival [3]. For heat shock experiments, cells were exposed to 45OC for O-90 min and survival was assayed as above. w: Cells expressing higher levels of the hsp70 protein generally tolerate thermal stress better, whereas cells expressing either of two mutated hsp70-encoding genes, one with a 4-base pair out-of-frame deletion and one with an in-frame deletion of codons 436 -618 are heat sensitive. These results provide strong evidence that expression of hsp70 leads directly to thermai tolerance. Interestingly, cells expressing a mutant hsp70 missing codons 120-428 (the ATP-binding domain) are nevertheless heat resistant. This result suggests that ATP-binding appears to be dispensible in the hsp70-mediated protection of cells from thermal stress [2]. In parallel, cells expressing high level of intact human hsp70 are more resistant to H202, whereas cells expressing the mutant human hsp70 missing codons 436-618, are H202 sensitive. Suptisingly, cells expressing the mutant hsp70 missing its ATP-binding domain and being heat resistant are also resistant to H202, Conclusion: Our results show that the constitutive and stable expression of human hsp70 gene confers heat resistance and resistance to H202, which implies a direct link between expression of a functional mammalian hsp70 and cell survival during heat shock or oxidative stress. Furthermore, our data suggest that the ATP-binding may not be necessary for the protective function of hsp70. Perhaps hsp70 lacking the ATP-binding domain can still bind to cellular proteins and prevent their aggregation upon thermal or oxidative stress. On the other hand, ATP binding and/or hydrolysis may be used to enable hsp70 to dissociate from its substrates or to facilitate the dissociation of aggregated proteins. Li GC, Li L, Liu YK, Mak JY, Chen L and Lee WMF, PNAS 88, 1681-1685, LI GC. Li L. Liu YK. Rehman M and Lee WMF. PNAS 89. (In Press) 1992 Spitz D, Dewey WC and Li GC, J. Cellular Physiology 131, !+54-373,1987

1991