Cellular effects of long wavelength UV light (UVA) in mammalian cells.

UVA should receive significant consideration as a human health risk as it is a large proportion of the solar spectrum that reaches the earth's surface and because of its ability to penetrate human skin. It is only relatively recently that this has been recognized and this previously under-researched part of the UV spectrum is becoming increasingly well characterized at doses that are quite low in relation to those experienced by humans. Absorption of UVA in a cell leads to the production of reactive oxygen and nitrogen species that can damage major biomolecules including DNA and membrane lipids. Various types of damage induced in these molecules lead to significant biological effects including cytotoxicity, mutations and alterations in cell signalling pathways. Longer-term effects such as persistent genomic instability and bystander effects have also been observed following UVA treatment of mammalian cells and, as with ionizing radiation, this changes some of the fundamental thinking around tissue effects of irradiation. Antioxidants have been assessed extensively for their ability to protect against the biological effects of UVA and a number have been shown to be successful at least in-vitro, for example vitamin E and epigallocatechin-3-gallate. Other potential targets for protection are suggested through the increased understanding of some of the signalling mechanisms activated following treatment, for example the inhibition of NADPH oxidase is seen to reduce a bystander effect. The search for appropriate and successful photoprotective agents remains an important area of research.

Clinical and cellular ionizing radiation sensitivity in a patient with xeroderma pigmentosum.

XP14BR is a cell line derived from a xeroderma pigmentosum (XP) patient from complementation group C. The patient was unusual in presenting with an angiosarcoma of the scalp, treated by surgical excision and radiotherapy. Following 38 Gy in 19 fractions with 6 MEV electrons, a severe desquamation and necrosis of the underlying bone ensued, and death followed 4 years later. The cell line was correspondingly hypersensitive to the lethal effects of gamma irradiation. We had previously shown that this sensitivity could be discriminated from that seen in ataxia-telangiectasia (A-T). The cellular response to ultraviolet radiation below 280 nm (UVC) was characteristic of XP cells, indicating the second instance, in our experience, of dual cellular UVC and ionizing radiation hypersensitivity in XP. We then set out to evaluate any defects in repair of ionizing radiation damage and to verify any direct contribution of the XPC gene. The cells were defective in repair of a fraction of double strand breaks, with a pattern reminiscent of A-T. The cell line was immortalized with the vector pSV3neo and the XPC cDNA transfected in to correct the defect. The progeny derived from this transfection showed the presence of the XPC gene product, as measured by immunoblotting. A considerable restoration of normal UVC, but not ionizing radiation, sensitivity was observed amongst the clones. This differential correction of cellular sensitivity is strong evidence for the presence of a defective radiosensitivity gene, distinct from XPC, which is responsible for the clinical hypersensitivity to ionizing radiation. It is important to resolve how widespread ionizing radiation sensitivity is amongst XP patients.

Assessing radiation-associated mutational risk to the germline: repetitive DNA sequences as mutational targets and biomarkers.

This review assesses recent data on mutational risk to the germline after radiation exposure obtained by molecular analysis of tandemly repeated DNA loci (TRDLs): minisatellites in humans and expanded simple tandem repeats in mice. Some studies, particularly those including exposure to internal emitters, indicate that TRDL mutation can be used as a marker of human radiation exposure; most human studies, however, are negative. Although mouse studies have suggested that TRDL mutation analysis may be more widely applicable in biomonitoring, there are important differences between the structure of mouse and human TRDLs. Mutational mechanisms probably differ between the two species, and so care should be taken in predicting effects in humans from mouse data. In mice and humans, TRDL mutations are largely untargeted with only limited evidence of dose dependence. Transgenerational mutation has been observed in mice but not in humans, but the mechanisms driving such mutation transmission are unknown. Some minisatellite variants are associated with human diseases and may affect gene transcription, but causal relationships have not yet been established. It is concluded that at present the TRDL mutation data do not warrant a dramatic revision of germline or cancer risk estimates for radiation.

Role for non-homologous end-joining in the repair of UVA-induced DNA damage.

PURPOSE: The biological significance of long-wavelength ultraviolet (UV) light, UVA, is increasingly realized, but the precise nature of the cellular damage responsible for the effects of this radiation is still not clear. It has been reported that UVA can induce double-strand breaks in DNA, but the biological significance of these is not known. We have therefore examined the UVA sensitivity of a cell line deficient in non-homologous end-joining, the major pathway for the repair of DNA double-strand breaks in mammalian cells in order to determine the biological importance of UVA-induced DSB. MATERIALS AND METHODS: Xrs-6, a Chinese hamster ovary cell line mutant for XRCC5 (Ku80) was compared with its parental CHO-K1 cell line for its sensitivity to UVA radiation (365 nm) using both a clonogenic assay and the micronucleus assay. RESULTS: Xrs-6 cells were sensitive to the cytotoxic effects of UVA. This resulted in the formation of chromosome damage, as measured by the micronucleus assay, which this cell line was unable to repair. CONCLUSIONS: Owing to the nature of the repair defect in these cells, these results imply that DNA double-strand breaks are produced in cells following UVA irradiation, that the non-homologous end-joining repair pathway is involved in their repair and that they are produced with sufficient frequency to have biological significance.

The use of DNA double-strand break quantification in radiotherapy.

DNA double-strand breaks (DSB) are an important direct consequence of treating cells with ionizingradiation. A variety of evidence points toward DSBs being the key damage type linked to radiation-induced lethality. In particular, the link between DSB and chromosome breakage, which in turn closely correlates with cell death in some cell types, is strongly supportive of this concept. There has been much interest in the possibility of using measures of strand breaks as a pretreatment test of radiation response. This has largely been in the context of assessing inherent cellular sensitivity through damage induction or repair parameters. A number of studies have produced hopeful results, but overall there has been no parameter that can reliably predict radiosensitivity. This may be due to the inadequacies of the assays, but it is more likely to reflect the fact that the radiosensitivity of cells is dictated by a whole series of events; alterations in many of these can alter the overall response. In addition, it is now recognized that cell-signalling pathways form an essential part of the cellular response to damage, and these can be triggered by damage other than DSB. It is therefore possible that while DSBs are clearly important--and they may be the single most important lesion in some types--other damage types may be significant triggers of cell death pathways after ionizing radiation treatment.

Variation in sensitizing effect of caffeine in human tumour cell lines after gamma-irradiation.

BACKGROUND AND PURPOSE: We have investigated whether the protective role of the G2 checkpoint has increasing importance when the p53-dependent G1 checkpoint is inactivated. MATERIALS AND METHODS: We have studied the differential effect of caffeine by clonogenic assays and flow cytometry in three human tumour cell lines with different functionality of p53 protein. RESULTS: The radiosensitizing effect of caffeine (2 mM) expressed itself as a significant decrease in surviving fraction at 2 Gy and a significant increase in alpha-values in RT112 and TE671, both with non-functional p53. However, no radiosensitizing effect was seen in cells with a normal p53 function (MCF-7 BUS). Two millimoles of caffeine also caused important changes in the cell cycle progression after irradiation. MCF-7 BUS showed a G1 arrest after irradiation and an early G2 arrest but those cells that reached the second G2 did not arrest significantly. In contrast, TE671 exhibited radiosensitization by caffeine, no G1 arrest, a G2 arrest in those cells irradiated in G2, no significant accumulation in the second G2 but an overall delay in release from the first cell cycle, which could be abrogated by caffeine. RT112 was similar to TE671 except that the emphasis in a G2 arrest was shifted from the block in cells irradiated in G2 to those irradiated at other cell cycle phases. CONCLUSION: The data presented confirm that p53 status can be a significant determinant of the efficacy of caffeine as radiosensitizer in these tumour cell lines, and document the importance of the G2 checkpoint in this effect.

Glutathione modulates the level of free radicals produced in UVA-irradiated cells.

We have developed an assay to detect reactive oxygen species (ROS) generated by UVA radiation utilising chemical probes which become fluorescent upon oxidation. Using a human bladder carcinoma cell line (MGH-U1) and spontaneously immortalised keratinocytes (HaCaT), we have shown a UVA (narrow band 365+/-5 nm) dose-dependent increase in fluorescence by flow cytometry following loading of the cells with either dihydrorhodamine 123 (DHR) or 2',7'-dichlorodihydrofluorescein diacetate (DCFH-DA). The UVA response of both DHR and DCFH was enhanced by elevation of intracellular levels of the photosensitiser protoporphyrin IX by incubation for 2.5 h with 5-aminolaevulinic acid. Depletion of the antioxidant glutathione (GSH) using the inhibitor D,L-buthionine-sulphoximine (BSO), resulted in an increase in the UVA-induced fluorescence of DCF but not of rhodamine 123. Conversely, raising intracellular GSH levels with N-acetyl cysteine (NAC) had relatively little protective effect in terms of degree of induced fluorescence.

Cells from individuals with SOD-1 associated familial amyotrophic lateral sclerosis do not have an increased susceptibility to radiation-induced free radical production or DNA damage.

Oxidative stress may play a role in the pathogenesis of familial amyotrophic lateral sclerosis (FALS). Superoxide dismutases (SODs) are enzymes that can influence free radical processes in irradiated cells and there is some evidence that manipulation of SODs can affect survival of cells after radiation treatments. SOD-1 associated FALS mutants may have an altered radiation response due to an enhanced generation of hydroxyl radicals or a compromised ability to neutralize free radicals. We have investigated the ability of the lymphoblastoid cell lines from FALS patients with SOD-1 gene mutations, patients with sporadic ALS and controls to handle oxidative stress induced by ionising radiation by measuring levels of intracellular reactive oxygen species and production of DNA double-strand breaks. Levels of reactive oxygen species, expressed as the slope of the relative fluorescence of a radical-reactive fluorochrome, in the cells from familial ALS patients with SOD-1 gene mutations (2.14+/-1.06 Gy(-1)) and patients with sporadic ALS (1.38+/-0.21 Gy(-1)) were not significantly different from the controls (1.54+/-0.39 Gy(-1)). No significant difference was observed in the production of DNA double-strand breaks between three groups. The ability of lymphoblastoid cells from FALS patients with SOD-1 gene mutations to scavenge radiation-induced free radicals is not compromised nor is their ability to protect DNA damage induced by ionising radiation.

Apoptosis after gamma irradiation. Is it an important cell death modality?

Apoptosis and necrosis are two different forms of cell death that can be induced by cytotoxic stress, such as ionizing radiation. We have studied the importance of apoptotic death induced after treatment with 6 Gy of gamma-irradiation in a panel of eight human tumour cell lines of different radiosensitivities. Three different techniques based on the detection of DNA fragmentation have been used, a qualitative one--DNA ladder formation --and two quantitative approaches--in situ tailing and comet assay. No statistically significant relationship between the two quantitative assays was found (r= 0.327, P = 0.159) so these methods seem to show different aspects of the process of cell death. The presence of the DNA ladder related well to the end-labelling method in that the least amount of end labelling was seen in samples in which necrotic degradation rather than apoptotic ladders were seen. However, as the results obtained by the comet assay are not in agreement with the DNA ladder experiments, we suggest that the distinction between the degraded DNA produced by apoptosis and necrosis may be difficult by this technique. Finally, although apoptosis has been proposed to be dependent on p53 functionality, and this may explain differences in cellular radiosensitivity, no statistically significant relationship was found between these parameters and apoptosis in the eight cell lines studied.

Differences between a human bladder carcinoma cell line and its radiosensitive clone in the formation of radiation-induced DNA double-strand breaks in different chromatin substrates.

It is well established that DNA-associated proteins, as well as soluble free-radical scavengers, can significantly influence the amount of damage inflicted in DNA by ionising radiation. It is not known, however, to what degree there is variation between cell lines in the effectiveness of these cellular components to protect DNA. In this study we have examined the level of strand break induction in a human bladder carcinoma cell line, MGH-U1, and its radiosensitive mutant, U1-S40b, when soluble scavengers and DNA-associated proteins were progressively removed. DNA double-strand breaks were measured using pulsed-field gel electrophoresis when cells were irradiated after lysis in solutions containing various salt concentrations. The two cell lines showed only a small, non-significant difference in damage induced in intact cells but isolated nuclei and chromatin devoid of non-histone proteins showed significantly more damage in the U1-S40b cells. Once the histone H1 was removed again there was no difference between the cell lines in the damage induced. We conclude that the different components of the cellular defences against free radical attack can have different influences in different cells. It is not clear whether this has an influence on the cellular sensitivity to the killing effects of radiation but it does suggest that artificial manipulation of the different components of the system may not affect overall damage induction to the same degree in all cells.

Radiosensitive human tumour cell lines show misrepair of DNA termini.

Physical measures of the rejoining of radiation-induced breaks in DNA strands are limited in terms of sensitivity and the fact that they do not assess the fidelity with which the rejoining occurs. In this report, transfection of cleaved plasmid has been used as a probe for repair in three radiosensitive tumour cell lines and shown them to have low repair fidelity compared with resistant cells. Errors in the repair of linear plasmid were found by Southern analysis, in keeping with the measured repair fidelity. Radiosensitive tumour cells showed few errors in the uptake and integration of circular plasmid, in contrast to ataxia-telangiectasia (A-T) cells. In the neuroblastoma HX142, the repair of blunt-ended linear plasmid was associated with deletions of > 1 kb; staggered-ended linear plasmid was repaired with small insertions and circular plasmid integration was intact in > 60% of the copies. The neuroblastoma SKN.SH, processed staggered-ended plasmid by insertions of a variety of sizes, but processed circular plasmid largely error-free. In contrast, A-T cells (AT5BIVA) had the same spectrum of errors irrespective of the form of plasmid transfected. Cell fusion between HX142 and AT5BIVA showed complementation to a resistant phenotype, suggesting that misrepair in the tumour cell did not result from somatic mutation in the ATM gene. In conclusion, radiosensitive tumours show evidence of misrepair of DNA termini, with a mechanism which is functionally and genetically distinct from that in A-T cells.

Radiosensitivity and double-strand break rejoining in tumorigenic and non-tumorigenic human epithelial cell lines.

Radiosensitivity and repair of DNA damage induced by ionizing radiation and restriction enzymes were investigated in three human epithelial cell lines: two tumorigenic squamous carcinoma cell lines (SCC-4 and SCC-25), and a non-tumorigenic epidermal keratinocyte cell line (RHEK-1). Sensitivity to ionizing radiation was determined using a clonogenic cell survival assay, which showed SCC-4 to be more radiosensitive than SCC-25 and RHEK-1, which in turn displayed about equal sensitivity. Using DNA precipitation under alkaline conditions for the analysis of induction and repair of DNA single-strand breaks (ssb), an increased level of ssb induction was found for SCC-4 while the efficiency of ssb repair was about equal in all three cell lines. Using pulsed-field gel electrophoresis (PFGE) for the measurement of induction and repair of DNA double-strand breaks (dsb), no consistent differences were detected between the three cell lines. A plasmid reconstitution assay was used to determine the capacity to rejoin restriction enzyme-induced dsb in whole-cell extracts prepared from the three cell lines. In these experiments, dsb rejoining was shown to be significantly reduced in the most radiosensitive SCC-4 cell line while it was about equal in RHEK-1 and SCC-25. The results indicate that plasmid reconstitution in cell-free extracts is a sufficiently sensitive assay to detect differences in repair capacity among tumour cell lines of different radiosensitivity which remain undetectable by DNA precipitation and PFGE.

A comparison of p53 and p16 expression in human tumor cells treated with hyperthermia or ionizing radiation.

To assess the potential relationship between p53 and p16 proteins in the cellular response to stress, we have examined the levels of these proteins in a series of human tumor cell lines after treatment with either ionizing radiation or hyperthermia. We found that cells with abnormal radiation-induced G1 arrest (non-functional p53) had significantly higher constitutive levels of p16 than cells showing a normal G1 arrest (functional p53). Time-course experiments were done to test the effect of gamma-irradiation on intracellular levels of p16. The pattern of changes in p16 response was similar in all cell lines studied, and p16 expression was not related to cellular sensitivity to radiation or to the level of p53 induction after treatment. We also provide evidence that short-term exposure to high temperature causes p53 accumulation. Hyperthermia-induced p53 accumulation was greatest in those cells exhibiting the highest radiation-induced p53 accumulation, suggesting a possible relationship between p53 induction after these 2 different stresses. p16 synthesis was also induced in different cell lines after heat treatment, and this response was independent of p53 functionality. When we compared the level of p16 expression with the extent of G0/G1 arrest induced by heat, a linear correlation was found, raising the possibility that p16 may be involved in the control of cell cycle progression in response to heat treatment.

Hprt- mutation spectrum in a closely related pair of human bladder tumour cell lines after gamma-irradiation at different dose-rates.

The spectrum of deletion sizes in mutants of two human bladder carcinoma cell lines has been examined. The cell lines were MGH-U1 and a radiation-sensitive subline (U1-S40b) that has been developed in this laboratory. Three groups, each of 20-30 mutants at the hprt locus were investigated: arising spontaneously, or induced after exposure to 10 Gy gamma-radiation either at high dose-rate (2 Gy/min) or low dose-rate (0.01 Gy/min). Data on the mutation frequency of the two cell lines at low dose-rate were obtained to supplement previously published data at high dose-rate. The mutation frequency was lower in U1-S40b than in MGH-U1 both for high and low dose-rate irradiation. The presence of intact copies of each of the nine hprt exons was examined using multiplex PCR, supplemented by single-exon PCR. The incidence of small hprt mutations (i.e. leading to no change in the size of the PCR products) was the same for spontaneous mutations in the two cell lines; for radiation-induced mutants it was higher in U1-S40b. The incidence of total deletions (i.e. no positive exon amplification) was lower in U1-S40b both for high and low dose-rate irradiation. The results are consistent with the hypothesis that large deletions tend to lead to the loss of adjacent essential genes and thereby to the death of potential mutants.

A role for molecular radiobiology in radiotherapy?

As we learn more about the cellular response to radiation and its genetic control, new avenues are opened up that have the potential to have a significant impact on radiotherapy practice. The recognition of the importance of the control of DNA damage induction and repair, cell cycle arrest and apoptosis gives us the primary areas to investigate, and the improvements in molecular technology make the application of our new knowledge more feasible. It can only be hoped that specific means can be found to assist in the prediction of normal tissue and tumour radiosensitivity and to manipulate sensitivity when that is desirable.

Cell-cycle variation in DNA migration in pulsed-field gel electrophoresis.

Pulsed-field electrophoresis is being used extensively in the gene mapping studies and in the analysis of DNA strand breakage by ionizing radiation. We have evaluated the relationship between the fraction of S phase DNA in a cell population and its ability to modify the migration of DNA in pulsed-field gel electrophoresis. We have shown that increasing the proportion of S phase DNA reduced the effective rate of migration of MGH-U1 cellular DNA. This effect was observed after treatment with ionizing radiation or the restriction enzyme Not I. However, when radiation-induced damage was studied using intact cells, only the DNA with 70 percent S phase showed apparent differences in damage induction. These studies therefore provide data to indicate the percentage of S phase cells at which overall DNA migration might be affected significantly.

Relationship between DNA damage, rejoining and cell killing by radiation in mammalian cells.

The prevailing hypothesis on the mechanism of radiation-induced cell killing identifies the genetic material deoxyribonucleic acid (DNA) as the most important subcellular target at biologically relevant doses. In this review we present new data and summarize the role of the DNA double-strand breaks (dsb) induced by ionizing radiation and DNA dsb rejoining as determinants of cellular radiosensitivity. When cells were irradiated at high dose-rate, two molecular end-points were identified which often correlated with radiosensitivity: (1) the apparent number of DNA dsb induced per Gy per DNA unit and (2) the half-time of the fast component of the DNA dsb rejoining kinetics. These two molecular determinants, not mutually exclusive, may be linked through a common factor such as the conformation of DNA.

Relationship between p53 status and radiosensitivity in human tumour cell lines.

We examined the relationship between p53 levels before and after irradiation, radiation-induced cell cycle delays, apoptotic cell death and radiosensitivity in a panel of eight human tumour cell lines. The cell lines differed widely in their clonogenic survival after radiation, (surviving fraction at 2 Gy: SF2=0.18-0.82). Constitutive p53 protein levels varied from 2.2 +/- 0.4 to 6.3 +/- 0.3 optical density units (OD) per 10(6) cells. p53 after irradiation (6 Gy) also varied between the cell lines, ranging from no induction to a 1.6-fold increase in p53 levels 4 h after treatment. p53 function was also assessed by G1 cell cycle arrest after irradiation. The cellular response to radiation, measured as G0/G1 arrest, and the induction of apoptosis were in good agreement. However, a trace amount of DNA ladder formation was found in two cell lines lacking G1 arrest. Overall cellular radiosensitivity correlated well with the level of radiation-induced G1 arrest (correlation coefficient r=0.856; P=0.0067), with p53 constitutive levels (r=0.874, P=0.0046), and with p53 protein fold induction (r=-0.882, P=0.0038). Our data suggest that (1) the constitutive p53 level, (2) G1 arrest after irradiation, or (3) the p53 protein response to radiation may be good predictive tests for radiosensitivity in some cell types.

Application of a bromodeoxyuridine-Hoechst/ethidium bromide technique for the analysis of radiation-induced cell cycle delays in asynchronous cell populations.

A flow cytometric technique utilizing the continuous incorporation of bromodeoxyuridine (BrdU) into asynchronous cells to measure radiation-induced cell cycle delay is described. Following the incorporation of the BrdU label the cells are stained with ethidium bromide and the bis-benzimidazole Hoechst 33258. These fluorochromes have differential staining patterns. Hoechst 33258 fluoresces blue and is quenched by BrdU incorporated into cellular DNA during S phase. Ethidium bromide fluoresces red and is not quenched by BrdU. Therefore in cells that are cycling and synthesizing DNA new G1 and G2 compartments are created and this can be used to measure cell cycle delays following ionizing radiation to asynchronous cells. We have used this technique to evaluate two cell lines: a normal diploid human embryo fibroblast cell line MRC 5, which has inducible p53 and shows delays at both G1 and G2 checkpoints, and the human cervix carcinoma cell line HX 156. This cell line has been infected with human papilloma virus (HPV) 16, and therefore has inactivated p53 function and is blocked only at the G2 checkpoint. Using this method, cell cycle-dependent effects relating to the G2 block can be observed. The radiation-induced G2 block differs from that induced by drugs or heating in that cells are blocked in G2 irrespective of the phase of the cell cycle they are treated in. This method allows these different types of G2 block to be quantified.