The radiation bystander effect and its potential implications for human health.

A long-held dogma in radiation biology has been that the biological effects of exposure to ionizing radiation occur as a result of damage in directly irradiated cells and that no effect would occur in neighboring unirradiated cells. This paradigm has been frequently challenged by reports of radiation effects in unirradiated or 'bystander' cells receiving signals from directly irradiated cells, an issue that may have substantial impact on radiation risk assessment and development of radiation-based therapies. Radiation-induced bystander effects have been shown in single-cell systems in vitro for an array of cancer relevant endpoints, and may trigger damage in more complex 3-D tissue systems. They may be mediated by soluble factors released by irradiated cells into the extracellular environment and/or by the passage of mediator molecules through gap-junction intercellular communication. To date, evidence that radiation-associated bystander or abscopal responses are effectual in vivo has been limited, but new data suggest that they may significantly affect tumor development in susceptible mouse models. Further understanding of how the signal/s is transmitted to unirradiated cells and tissues and how it provokes long-range and significant responses is crucial. By summarizing the existing evidence of radiation induced bystander-like effects in various systems with emphasis on in vivo findings, we will discuss the potential mechanisms involved in these observations and how effects in bystander cells contribute to uncertainties in assessing cancer risks associated with radiation exposure.

Role of connexin43 and ATP in long-range bystander radiation damage and oncogenesis in vivo.

Ionizing radiation is a genotoxic agent and human carcinogen. Recent work has questioned long-held dogmas by showing that cancer-associated genetic alterations occur in cells and tissues not directly exposed to radiation, questioning the robustness of the current system of radiation risk assessment. In vitro, diverse mechanisms involving secreted soluble factors, gap junction intercellular communication (GJIC) and oxidative metabolism are proposed to mediate these indirect effects. In vivo, the mechanisms behind long-range 'bystander' responses remain largely unknown. Here, we investigate the role of GJIC in propagating radiation stress signals in vivo, and in mediating radiation-associated bystander tumorigenesis in mouse central nervous system using a mouse model in which intercellular communication is downregulated by targeted deletion of the connexin43 (Cx43) gene. We show that GJIC is critical for transmission of oncogenic radiation damage to the non-targeted cerebellum, and that a mechanism involving adenosine triphosphate release and upregulation of Cx43, the major GJIC constituent, regulates transduction of oncogenic damage to unirradiated tissues in vivo. Our data provide a novel hypothesis for transduction of distant bystander effects and suggest that the highly branched nervous system, similar to the vascular network, has an important role.

Opposite modifying effects of HR and NHEJ deficiency on cancer risk in Ptc1 heterozygous mouse cerebellum.

Heterozygous Patched1 (Ptc1(+/-)) mice are prone to medulloblastoma (MB), and exposure of newborn mice to ionizing radiation dramatically increases the frequency and shortens the latency of MB. In Ptc1(+/-) mice, MB is characterized by loss of the normal remaining Ptc1 allele, suggesting that genome rearrangements may be key events in MB development. Recent evidence indicates that brain tumors may be linked to defects in DNA-damage repair processes, as various combinations of targeted deletions in genes controlling cell-cycle checkpoints, apoptosis and DNA repair result in MB in mice. Non-homologous end joining (NHEJ) and homologous recombination (HR) contribute to genome stability, and deficiencies in either pathway predispose to genome rearrangements. To test the role of defective HR or NHEJ in tumorigenesis, control and irradiated Ptc1(+/-) mice with two, one or no functional Rad54 or DNA-protein kinase catalytic subunit (DNA-PKcs) alleles were monitored for MB development. We also examined the effect of Rad54 or DNA-PKcs deletion on the processing of endogenous and radiation-induced double-strand breaks (DSBs) in neural precursors of the developing cerebellum, the cells of origin of MB. We found that, although HR and NHEJ collaborate in protecting cells from DNA damage and apoptosis, they have opposite roles in MB tumorigenesis. In fact, although Rad54 deficiency increased both spontaneous and radiation-induced MB development, DNA-PKcs disruption suppressed MB tumorigenesis. Together, our data provide the first evidence that Rad54-mediated HR in vivo is important for suppressing tumorigenesis by maintaining genomic stability.

Modulation of basal and squamous cell carcinoma by endogenous estrogen in mouse models of skin cancer.

Patched1 heterozygous mice (Ptch1(+/-)) are useful for basal cell carcinoma (BCC) studies, being remarkably susceptible to BCC induction by ultraviolet or ionizing radiation. Analogously, skin carcinogenesis-susceptible (Car-S) mice are elective for studies of papilloma and squamous cell carcinoma (SCC) induction. We previously reported a striking effect of gender on BCC induction in Ptch1(+/-) mice, with total resistance of females; likewise, Car-S females show increased skin tumor resistance relative to males. Here, we investigated the protective role of endogenous estrogen in skin keratinocyte tumorigenesis. Control (CN) and ovariectomized Ptch1(+/-) or Car-S females were irradiated for BCC induction or topically treated with chemical carcinogens for SCC induction. Susceptibility to BCC or SCC was dramatically increased in ovariectomized Ptch1(+/-) and Car-S females and restored to levels observed in males. Remarkably, progression of initially benign papillomas to malignant SCC occurred only in ovariectomized Car-S females. We explored the mechanisms underlying tumor progression and report overexpression of estrogen receptor (ER)-alpha, downregulation of ERbeta and upregulation of cyclin D1 in papillomas from ovariectomized Car-S relative to papillomas from CN females. Thus, an imbalanced ERalpha/ERbeta expression may be associated with estrogen-mediated modulation of non-melanoma skin carcinogenesis, with a key role played by cyclin D1. Our findings underscore a highly protective role of endogenous estrogen against skin tumorigenesis by diverse agents in two independent mouse models of skin cancer.

Effects of exposure of newborn patched1 heterozygous mice to GSM, 900 MHz.

Patched1 heterozygous knockout mice (Ptc1+/-), an animal model of multiorgan tumorigenesis in which ionizing radiation dramatically accelerates tumor development, were used to study the potential tumorigenic effects of electromagnetic fields (EMFs) on neonatal mice. Two hundred Ptc1+/- mice and their wild-type siblings were enrolled in this study. Newborn mice were exposed to 900 MHz radiofrequency radiation (average SAR: 0.4 W/kg for 5 days, 0.5 h twice a day) or were sham exposed. We found that RF EMFs simulating the Global System for Mobile Communications (GSM) did not affect the survival of the mice, because no statistically significant differences in survival were found between exposed and sham-exposed animals. Also, no effects attributable to radiofrequency radiation were observed on the incidence and histology of Ptc1-associated cerebellar tumors. Moreover, the skin phenotype was analyzed to look for proliferative effects of RF EMFs on the epidermal basal layer and for acceleration of preneoplastic lesions typical of the basal cell carcinoma phenotype of this model. We found no evidence of proliferative or promotional effects in the skin from neonatal exposure to radiofrequency radiation. Furthermore, no difference in Ptc1-associated rhabdomyosarcomas was detected between sham-exposed and exposed mice. Thus, under the experimental conditions tested, there was no evidence of life shortening or tumorigenic effects of neonatal exposure to GSM RF radiation in a highly tumor-susceptible mouse model.

Two-hit model for progression of medulloblastoma preneoplasia in Patched heterozygous mice.

Inactivation of one Ptc1 allele predisposes humans and mice to spontaneous medulloblastoma development, and irradiation of newborn Ptc1 heterozygous mice results in dramatic increase of medulloblastoma incidence. While a role for loss of wild-type (wt) Ptc1 (LOH) in radiation-induced medulloblastomas from Ptc1(neo67/+) mice is well established, the importance of this event in spontaneous medulloblastomas is still unclear. Here, we demonstrate that biallelic Ptc1 loss plays a crucial role in spontaneous medulloblastomas, as shown by high rate of wt Ptc1 loss in spontaneous tumors. In addition, remarkable differences in chromosomal events involving the Ptc1 locus in spontaneous and radiation-induced medulloblastomas suggest distinct mechanisms for Ptc1 loss. To assess when, during tumorigenesis, Ptc1 loss occurs, we characterized cerebellar abnormalities that precede tumor appearance in Ptc1(neo67/+) mice. We show that inactivation of only one copy of Ptc1 is sufficient to give rise to abnormal cerebellar proliferations with different degree of altered cell morphology, but lacking potential to progress to neoplasia. Furthermore, we identify biallelic Ptc1 loss as the event causally related to the transition from the preneoplastic stage to full blown medulloblastoma. These results underscore the utility of the Ptc1(neo67/+) mouse model for studies on the mechanisms of medulloblastoma and for development of new therapeutic strategies.

Linking DNA damage to medulloblastoma tumorigenesis in patched heterozygous knockout mice.

Hemizygous Ptc1 mice have many features of Gorlin syndrome, including predisposition to medulloblastoma development. Ionizing radiation synergize with Ptc1 mutation to induce medulloblastoma only in neonatally exposed mice. To explore the mechanisms underlying age-dependent susceptibility, we irradiated Ptc(neo67/+) mice at postnatal day 1 (P1) or 10 (P10). We observed a dramatic difference in medulloblastoma incidence, which ranged from 81% in the cerebellum irradiated at P1 to 3% in the cerebellum irradiated at P10. A striking difference was also detected in the frequency of cerebellar preneoplastic lesions (100 versus 14%). Our data also show significantly lower induction of apoptosis in the cerebellum of medulloblastoma-susceptible (P1) compared to -resistant (P10) mice, strongly suggesting that medulloblastoma formation in Ptc1 mutants may be associated with resistance to radiation-induced cell killing. Furthermore, in marked contrast with P10 mice, cerebellum at P1 displays substantially increased activation of the cell survival-promoting Akt/Pkb protein, and markedly decreased p53 levels in response to radiation-induced genotoxic stress. Overall, these results show that developing cerebellar granule neuron precursors' (CGNPs) radiosensitivity to radiation-induced cell death increases with progressing development and inversely correlates with their ability to neoplastically transform.

The European Radiobiology Archives (ERA)--content, structure and use illustrated by an example.

The European Radiobiology Archives (ERA), supported by the European Commission and the European Late Effect Project Group (EULEP), together with the US National Radiobiology Archives (NRA) and the Japanese Radiobiology Archives (JRA) have collected all information still available on long-term animal experiments, including some selected human studies. The archives consist of a database in Microsoft Access, a website, databases of references and information on the use of the database. At present, the archives contain a description of the exposure conditions, animal strains, etc. from approximately 350,000 individuals; data on survival and pathology are available from approximately 200,000 individuals. Care has been taken to render pathological diagnoses compatible among different studies and to allow the lumping of pathological diagnoses into more general classes. 'Forms' in Access with an underlying computer code facilitate the use of the database. This paper describes the structure and content of the archives and illustrates an example for a possible analysis of such data.

The influence of sex on life shortening and tumor induction in CBA/Cne mice exposed to X rays or fission neutrons.

An experimental study of male and female CBA/Cne mice was set up at Casaccia primarily to investigate the influence of sex on long-term survival and tumor induction after exposure to high- and low-LET radiation. Mice were whole-body-irradiated at 3 months of age with fission-neutron doses of 0.1, 0.2, 0.4, 0.8, 1.2 and 1.8 Gy at the RSV-TAPIRO reactor (mean neutron energy 0.4 MeV, in terms of kerma, y(D) = 51.5 keV/micron), or with 250 kVp X-ray doses of 1, 3, 5 and 7 Gy. Control and irradiated animals were then followed for their entire life span. As a general finding, male CBA/Cne mice appear more susceptible to tumorigenesis than females. In particular, the incidences of induced acute myeloid leukemia and malignant lymphomas are significant only in male mice. Benign and malignant solid tumors of many types are observed in mice of both sexes, the most frequent being in the lung, liver and ovary. However, evidence for a radiation response is limited to the case of Harderian gland neoplasms. In addition, a comparison of the observed frequency of all irradiated compared to unirradiated animals bearing solid tumors shows that the total tumor occurrence is not altered markedly by radiation exposure. A decrease in survival time is observed for both sexes and radiation types and correlates well with increasing dose. Moreover, both sex and radiation quality appear to influence the life shortening. A similar dose dependence of survival time is found when tumor-free animals alone are considered, suggesting a non-specific component of life-shortening.

Sensitivity of C3H 10T1/2 cells to radiation-induced killing and neoplastic transformation as a function of cell cycle.

Cell-age sensitivity to both cell killing and neoplastic transformation induced by radiation was investigated using synchronized populations of C3H10T1/2 cells. Mitotic-cell suspensions, collected using a mitotic shake-off procedure, were irradiated with 4Gy 250 kVp X-rays or 0.5 Gy fission neutrons from the RSV-TAPIRO reactor at CR-Casaccia. For study of cell killing the mitotic-cell suspensions were either irradiated immediately after collection, or plated for subsequent irradiation, which was performed every hour, covering an interval of 17 h. The response pattern observed was similar after X-rays and neutron irradiation, but the magnitude of the variation through the cell cycle was smaller in the case of neutrons (1.3- compared with 5-fold). For study of neoplastic transformation induction the irradiation was performed immediately after collection, i.e. in M phase, or at later times corresponding to mid-G1, G1/S and G2 phases. The sensitivity of the G2/M phase was examined by irradiating the cells with 4Gy X-rays while still attached to the flask bottom, and dislodging them after 25 min. SimilarLy to cell survival, the transformation frequency showed a small variation after neutron irradiation (1.4- compared with 3.1-fold) for the phases examined.

Neoplastic transformation of C3H 10T1/2 cells: a study with fractionated doses of monoenergetic neutrons.

As most occupational and environmental exposures to ionizing radiation are at low dose rates or in small dose fractions, risk estimation requires that the effects of the temporal distribution of dose are taken into account. Previous in vitro studies of oncogenic transformation, as well as in vivo studies of carcinogenesis induced by high-LET radiation, yielded controversial results concerning the presence of an inverse dose-rate effect. The present study tested the influence of one scheme of dose fractionation of monoenergetic neutrons on neoplastic transformation of C3H 10T1/2 cells. Neutrons of 0.5, 1.0 and 6.0 MeV were used. Cells were exposed to doses of 0.25 and 0.5 Gy, given acutely or in five fractions at 2-h intervals. The acute and fractionated irradiations with each energy were done on the same day. No significant difference between the two irradiation modes was found for both cell inactivation and neoplastic transformation at all energies. These results are in agreement with our data for fractionated fission-spectrum neutrons from the RSV-TAPIRO reactor.

Neutron-induced tumors in BC3F1 mice: effects of dose fractionation.

An experimental study of the biological effectiveness of multifractionated low doses of high-LET radiation was carried out using BC3F1 male mice. They were treated with whole-body irradiation with five equal daily fractions of fission neutrons to yield cumulative doses of 0.025, 0.05, 0.10, 0.17, 0.25, 0.36, 0.535 and 0.71 Gy at the RSV-TAPIRO reactor (mean neutron energy 0.4 MeV, in terms of kerma, y D = 51.5 keV/microns, dose rate 0.004 Gy/min) and were followed for their entire life span. The statistical method described by Peto et al. (IARC Monograph, Suppl. 2, 1980) to establish the existence of a carcinogenic effect in long-term animal experiments was applied to the data sets. This analysis was done for myeloid leukemia and for the presence of selected solid tumors. Myeloid leukemia was absent in the control group and was rarely found in irradiated animals. However, a positive significant trend was found in the dose ranges 0-0.17 Gy and higher. Epithelial tumors were induced at doses from 0.17 Gy on. Tumor occurrence was evaluated further as final incidences with age adjustment for the differences in mortality rates. Survival and incidence data for selected classes of tumors after 0.17, 0.36 and 0.71 Gy were compared with those from a previous experiment at corresponding doses given acutely (dose rate between 0.05 and 0.25 Gy/min). This indicated no marked overall influence of the time regimen of neutron irradiation on survival and tumor induction.

The dose-response relationships for tumor induction after high-LET radiation.

This paper presents a review of several studies conducted in our laboratory to examine the carcinogenic effects in mice of high-LET radiation and, for comparison, of low-LET reference radiation. For some specific end-points the following conclusions can be formulated: i) the dose-response curves for myeloid leukemia and malignant lymphoma can be interpreted in terms of induction and inactivation; in particular, the data confirmed that a linear dependence of the induction on dose is adequate to describe the response to fission neutrons, while a pure quadratic dependence is consistent with the experimental data for low-LET radiation; ii) in the liver, a marked age-dependence was demonstrated for radiation-induced tumors with a much higher susceptibility in young than in old mice; also for these tumors the dose-effect curves can be described by a linear and a quadratic relationships for high- and low-LET radiation, respectively; iii) data on ovarian tumor induction suggested threshold-like dose responses: these peculiar shapes as well as the absence of a clear radiation quality dependence of the curves are difficult findings to explain using a simple model of radiation action, and they might better be related to a non-stochastic effect of hormonal imbalance following irradiation.

Neutron carcinogenesis in mice: a study of the dose-response curves.

Several experimental studies have been conducted with the objective to improve our knowledge of the types of dose-response relationships for radiation carcinogenesis in mice exposed to single acute doses. The experimental results on tumor induction have already been published and are here summarized with emphasis on the dependence on radiation quality, age at irradiation, and sex. These data indicate that the bone marrow, liver, and ovaries of the mice tested have an appreciable susceptibility to radiation carcinogenesis. However, the shape of the dose-response relationship depends on the tissue exposed. The data also confirm that a linear relationship is adequate for a conservative description of the dose-effect curves after exposure to low dose of neutrons, while a purely quadratic dependence is not inconsistent with the experimental data obtained using low-LET radiation. Other information which stems from the present analysis is that the susceptibility to radiation induction of liver tumor by fission neutrons decreases in old age. Finally, the experimental data on induction of ovarian tumors suggest a threshold-like dose response.

Absence of a dose-fractionation effect on neoplastic transformation induced by fission-spectrum neutrons in C3H 10T1/2 cells.

We have investigated the effect of fission-spectrum neutron dose fractionation on neoplastic transformation of exponentially growing C3H 10T1/2 cells. Total doses of 10.8, 27, 54, and 108 cGy were given in single doses or in five equal fractions delivered at 24-h intervals in the biological channel of the RSV-TAPIRO reactor at CRE-Casaccia. Both cell inactivation and neoplastic transformation were more effectively induced by fission neutrons than by 250-kVp X rays. No significant effect on cell survival or neoplastic transformation was observed with split doses compared to single doses of fission-spectrum neutrons. Neutron RBE values relative to X rays determined from data for survival and neoplastic transformation were comparable.

Age-related susceptibility of mouse liver to induction of tumors by neutrons.

The induction of liver tumors has been studied in BC3F1 male mice after acute whole-body irradiation with fission neutrons and X rays, given at different ages. In prenatally irradiated mice, a small effect is seen after doses of 0.3 to 2.1 Gy of X rays, and a more pronounced effect is found after neutron doses of 0.09 to 0.62 Gy. At 3 months of age the animals show a higher incidence after X-ray doses from 2 Gy, and for neutrons from 0.17 Gy. At 19 months of age, liver tumors are rarely induced by either type of radiation. These findings are confirmed by the statistical analysis of trend. The possibility of deriving dose-response relationships for liver tumors was also investigated. In the dose ranges where the risk appears to increase as a function of the increase in dose, the data points for neutrons were well fitted by a linear expression. A pure quadratic relationship best fitted the X-ray data. Using these expressions, the RBE for neutrons was 28 at 0.09 Gy for prenatal irradiation and 13 at 0.17 Gy for irradiation at 3 months. This suggests the existence of a risk of developing liver tumors after exposure to radiation, and fetal liver seems to be particularly sensitive to neoplastic transformation. This risk may be negligible at low doses (less than 1 Gy) of low-LET radiation, or with exposure at older ages.

The dose-response relationships for myeloid leukemia and malignant lymphoma in BC3F1 mice.

The present analysis of data on the induction of lymphoma and myeloid leukemia in BC3F1 mice has indicated some new and interesting aspects regarding the shapes of the dose-effect curves. The incidence data can be interpreted by radiobiological models of the induction process coupled with cell inactivation. In particular, for malignant lymphoma the dose-response curve after X rays can be described assuming a quadratic model corrected for cell inactivation, while the incidence data after fission neutrons are best fitted by a linear model which also allows for cell inactivation. Myeloid leukemia has also been induced in BC3F1 mice. The bell-shaped dose-response curves observed after irradiation with either X rays or neutrons are explained by assuming simultaneous initial transforming events and cell inactivation with the data for cell inactivation at higher doses being in agreement with data reported for other strains of mice. A value for relative biological effectiveness of 4 is obtained at the lowest neutron dose used. The value of the inactivation parameters can be compared with those of the cell inactivation probability per unit dose for the bone marrow hematopoietic stem cells, which are believed to be the target cells for these tumors.

Inheritance of immune responsiveness, life span, and disease incidence in interline crosses of mice selected for high or low multispecific antibody production.

High (H) and low (L) antibody responder lines of mice separated by selective breeding present a maximal interline difference in antibody (Ab) response to Ag of different specificities (general genetic regulation). The analysis of SRBC agglutinin response in H line, L line, F1 hybrids, F2, and backcross segregants demonstrates that Ab responsiveness is a polygenic trait regulated by the additive interaction of 5 to 7 independent loci, with an incomplete dominance (44% +/- 7%) of the high response character, and a 30% +/- 10% impact of the environmental factors. The life span of H, L, F1, F2, and backcross populations is correlated positively with 2-ME-resistant agglutinin response (r = 0.97, p less than 0.001) and negatively with 2-ME-sensitive agglutinin response (r = 0.95, p = 0.01) (interpopulation correlation). Similar correlations are also observed in individuals of the various populations, especially in F1 x L backcross, in which the largest phenotypic variance is found. The positive correlation between Ab responsiveness and life span was confirmed by ELISA titration for distinct IgG isotypes (intrapopulation correlation). Malignant lymphomas and chronic nephritis were the two most common diseases observed. The age-adjusted incidence of such diseases, which is largely affected by environmental factors, accounts for the longer life span of H, as compared with L, mouse populations. The longevity of the 30% or less survivors, chiefly determined by the rate of physiologic aging, is a polygenic character regulated by the cumulative interaction of 3 to 7 independent loci, with a complete dominance of the long life trait and an impact of the environmental factors of about 60%. Thus we have grounds for regarding general Ab responsiveness and life span as polygenic traits regulated by a small number of identical or closely linked gene loci, and immune responsiveness as a defense mechanism against neoplastic and inflammatory diseases.

Late somatic effects in mice after total lymphoid irradiation.

Late somatic effects of total lymphoid irradiation have been investigated in BC3F1 mice. A total X-ray dose of 34 Gy was distributed in 17 daily fractions. The cumulative mortality curve is shifted in time because all the irradiated mice died earlier than the unirradiated controls. There was a 24% shortening of life span. A marked increase of solid tumor incidence, mostly due to skin cancers, was observed (66% vs 30%). In contrast, the incidence of malignant lymphomas was greatly reduced in irradiated mice (6% vs 49%). Furthermore, nephrosclerosis was a common finding in the irradiated group (38% vs 8%). Death-rate analysis revealed an association between life shortening and the presence of solid tumors and nephrosclerosis at death.

Tumor induction and life shortening in BC3F1 female mice at low doses of fast neutrons and X rays.

Extension of previous investigations at this laboratory regarding life shortening and tumor induction in the mouse has provided more complete dose-response information in the low dose region of X rays and neutrons. A complete observation of survival and late pathology has been carried out on over 2000 BC3F1 female mice irradiated with single doses of 1.5 MeV neutrons (0.5, 1, 2, 4, 8, 16 cGy) and, for comparison, of X rays (4, 8, 16, 32, 64, 128, 256 cGy). Data analysis has shown that a significant life shortening is observable only for individual neutron doses not lower than 8 cGy. Nevertheless, assuming a linear nonthreshold form for the overall dose-effect relationships of both radiation qualities, an RBE value of 12.3 is obtained for the 1.5 MeV neutrons. The induction of solid tumors by neutrons becomes statistically significant at individual doses from 8 cGy and by X rays for doses larger than 1 Gy. Linear dependence on neutron dose appears adequate to interpret the data at low doses. A separate analysis of ovarian tumor induction substantiates the hypothesis of a threshold dose for the X rays, while this is not strictly needed to interpret the neutron data. A trend analysis conducted on the neoplasm incidence confirms the above findings. Death rates have been analyzed, and a general agreement between the shift to earlier times of these curves and tumor induction was found.