Columbia University Medical Center
Center for Radiological Research

NIH Program Project on
Radiation Bystander Effects: Mechanism
PO1-CA 49062-23

Project 2: Mechanisms of Bystander Mutagenesis

Project Leader: Tom K. Hei


Project 2 determines the incidence and mechanism of radiation-induced non-targeted/ bystander mutagenic response in vivo; to clarify the role of cyclooxygenase-2 (COX2) signaling pathways in the process; and to examine the incidence of genomic instability in bystander tissues in wild type and in genetically susceptible animals.


Research Aims

The central testable hypothesis is that COX2 mediates radiation induced bystander mutagenesis in vivo and that bystander cells are genomically unstable. A series of five inter-related specific aims are proposed to address these goals. The novel gpt delta transgenic mice and embryo fibroblasts from these animals will be used to correlate the findings under both in vitro and in vivo conditions. Using the gpt delta mice, we will determine the incidence and types of both Spi (deletions) and gpt mutations (point mutations), induction of oxidative DNA damages in the lung and breast tissues of animals exposed to graded doses of low LET X-rays targeted in lower abdominal cavity far away from the chest area at different time points. We will determine the incidence of mitochondrial DNA mutations and mitochondrial functions in these out of field tissues. We will generate COX2 knock out mice in the C57BL/6 gpt delta transgenic background and will use both the knock out animals and mouse embryo fibroblasts to ascertain the COX2 signaling pathways in mediating the non-targeted/ out of field response. We will further determine if bystander cells are genomically unstable in animals that are genetically susceptible to damaging effects of ionizing radiation.

Research Highlights

In vivo non-targeted mutagenic response

In the current funding period, we have used the gpt delta transgenic mouse system, established in the laboratory of Dr. Takehiko Nohmi, to examine in vivo non-targeted effects of ionizing radiation. An area 1cm x 1cm square in the lower extremity of eight weeks old animals was exposed to a single 5 Gy dose of X-rays while the rest of the body was shielded with lead. At specified time points post-irradiation animals were anesthetized and 0.1 ml of blood from each animal was drawn from the orbital sinus to determine cytokine (TGFß, TNFa, IL-6) levels by ELISA assay. Anesthetized animals were euthanized by decapitation and tissues (lung and liver) samples were obtained. The scattering dose was measured by insertion of mini-dosimeters into the lungs of the animals when irradiated and, for a 5 Gy dose, the scattered dose was estimated ~ 6 cGy. The following figure (Figure 1) illustrates the positioning of the animal during irradiation.

Figure 1.

Fig.1. Schematic diagram showing the irradiation set up to determine the non-targeted/ out of field effect of ionizing radiation.

Induction of Spi mutation and COX2 expression in non-targeted lung tissues after lower abdominal irradiation

The gpt (xanthine phosphoribosyltransferase) delta transgenic mice are established by integrating multiple copies of ? EG10 DNA with the redBA and gam genes into each chromosome 17 of C57BL/6J mice (Nohmi and Masumura 2004). Because wild-type ?-phage DNA replicate poorly in the presence of P2 lysogens in the host cells (called “sensitive to P2 interference” or “Spi”), only mutant ? phages that are deficient in the functions of both the redBA and gam genes are able to escape from P2 interference (called “Spi-”) and form visible clear plaques on a bacterial lawn. Simultaneous inactivations of both the redBA and gam genes, an indication of deletions in the gene loci region, provide an available method to quantify deletion mutations induced by various physical and chemical mutagens, such as X rays and alkylating agents. As shown in Figure 2A, out of field lung tissue demonstrated a three fold increase in Spi- mutation compared with controls. In contrast, lung tissues exposed to a whole body 6 cGy dose (simulating the scattered dose) showed little or no mutant induction.

The expression of COX2 protein in lung tissues at a series of time points after a single, 5 Gy X-ray irradiation of the lower abdominal area of animals was shown in Figure 2B. COX2 was induced at 1 hour after radiation exposure and the level steadily increased within the first 24 hours. Thereafter, COX2 expression decreased to basal level by 72 hours after irradiation. Quantification of Western blotting data showed that COX2 expression at the peak level (6 hr) increased by more than 20 fold relative to un-irradiated controls. mRNA level of COX-2, determined by RT-PCR, showed a similar tendency of COX2 induction as the protein level (data not shown).

Figure 2.

Fig 2. A. Relative fold change in Spi ̵ mutant frequencies in non-targeted/ out of field lung tissues from animals exposed to a single dose of X-rays in the abdominal area when compared with non-irradiated controls. Whole body irradiation with doses of 5 Gy and 6 cGy dose served as controls. B. Western blot data showing COX2 induction in out of field lung tissue as a function of time post-irradiation. Note that in animals exposed to a 6 cGy whole body dose of X-ray, there was no induction of COX2

Induction of oxidative DNA damages in non-targeted lung tissue

There is evidence that prostaglandins induce production of reactive oxygen species (ROS) and oxidative DNA damage including formation of 8-hydroxydeoxyguanosine (8-OHdG). The level of oxidative DNA damage in out of field lung tissues was determined using immunohistochemical staining of 8-OHdG. Whole body irradiation dramatically induced 8-OHdG in lung tissues (Figure 3A). Likewise, non-targeted lung tissues from partially irradiated animals showed positive 8-OHdG staining relative to control animals at 24 hours after irradiation of lower abdomen. Consistent with the distribution of COX2 in the lung tissues, 8-OHdG was predominantly located in lung bronchial epithelial cells. Quantification of the staining showed that non-targeted irradiation induced an 8 fold increase in oxidative DNA damages compared to controls, while whole body irradiation with 5Gy X-rays resulted in a 12 fold increase (Figure 3B).

Figure 3.

Fig.3 Expression of 8-OHdG in out of field lung tissue (A) and quantification of the staining images (B).


Mouse embryo fibroblasts from COX2 knock out mice showed reduced bystander effects

Mouse embryo fibroblasts (MEF) from three cyclooxygenase-2 (COX2) genotypes (wild, heterozygote, and homozygote) were generated. Using these MEFs and the unique Columbia University microbeam, we found that when 10% of cells were irradiated with a lethal dose of 30 alpha particle through the nucleus, the induced micronucleus formation in COX2 wild type or heterozygote MEFs was significantly higher than that of background control (Figure 4). However, in COX2 null MEFs, there was no non-targeted/ bystander response. To further confirm this finding and to explore the mechanisms involved in the non-targeted signaling, we used strip Mylar dishes and a broad beam to ascertain the non-targeted response. Similar to the findings obtained using the microbeam, COX2 null MEFS showed a significantly decreased radiation induced-bystander response compared to the wild type and heterozygote MEFs. These results indicated COX2 play a critical role in alpha particle- induced non-targeted / bystander effects in vitro.

Figure 4.

Fig.4. Incidence of micronucleus formation in MEF from either wild type or COX2 knock out mice.

To determine the bystander signaling process, we used whole human genome microarrays and real time quantitative PCR to measure and validate gene expression profile in directly irradiated cells and compare the profiles with non-targeted/bystander cells as well as control irradiated human skin fibroblasts. Microarray analysis was performed using BRB-Array Tools; pathway and ontology analyses were conducted using Ingenuity Pathway Analysis and PANTHER, respectively. Furthermore, we studied signaling in irradiated and bystander cells using immunoblotting and semi-quantitative image analysis. Gene ontology suggested signal transduction and transcriptional regulation responding 30 minutes after treatment affected cell structure, motility and adhesion, and interleukin synthesis. We measured time-dependent expression of genes controlled by the NF-?B pathway; matrix metalloproteinases 1 and 3; chemokine ligands 2, 3 and 5 and interleukins 1ß, 6 and 33. There was an increased response of this set of genes 30 minutes after treatment and another wave of induction at 4 hours. We investigated AKT-GSK3ß signaling and found both AKT and GSK3ß are hyper-phosphorylated 30 minutes after irradiation and this effect is maintained through 4 hours. In bystander cells, a similar response was seen with a delay of 30 minutes. We proposed a network model where the observed decrease in phosphorylation of ß-catenin protein after GSK3ß dependent inactivation can trigger target gene expression at later times after radiation exposure. In Figure 5, we proposed a working model of signaling pathways in non-targeted response. Our current findings are the first to show that the radiation induced bystander signal induces a widespread gene expression response at 30 minutes after treatment and these changes are accompanied by modification through RAD9 and gap junctional proteins that mediates the PI3K-AKT-GSK3ß pathway.

Figure 5.

Fig. 5.  A model of the signaling pathways regulating radiation-induced non-targeted/ out of field effects through expression and secretion of soluble biologically active factors.

Generation of COX2 null phenotype in the gpt delta mice

Cyclooxygenase-2 (COX2) knock out phenotype in the gpt delta C57BL/6 background was successfully generated by serial back crossing. Male COX2 null mice and female gpt mice were mated and the resulting F1 generation of mice were all gpt +/– COX2+/–.  These F1 generation animals were then inter-crossed to generate F2 generation and roughly one quarter of the off-spring were found to have the gpt +/+ COX-2 +/+ genotype. Among the F2 mice, animals with the gpt +/+ COX-2 +/– genotype were then back-crossed to generate the F3 population and one quarter of the off-spring were expected to have the required genotype of gpt +/+ COX-2 –/– that we have used in subsequent experiments. A small 1 cm by 1 cm area of the lower abdominal area was irradiated with a 2 Gy dose of X-rays. After irradiation, out of field lung, liver, and mammary gland (female) were collected for mutation assay at 6 and 24 hours post irradiation. The spi- mutation frequency was significantly decreased in out of field lung tissue in COX2 null mice compared with COX2 wild type and COX2 heterozyous mice.

Radiation-induced non-targeted response in human neural stem cells and glioblastoma

Neural stem and progenitor cells (Neural Crest Cells or NCS) are sensitive to a variety of insults, including ionizing radiation, reactive oxygen species and various pro-inflammatory mediators such as interleukin-1, TNFa, COX2. In order to determine possible bystander effects of directly irradiated cancer cells on non-treated NSC, we used media transfer from irradiated NSC, embryonic astrocytes or glioblastoma cells to non-treated NSC. There were pronounced changes in apoptosis and differentiation of non-irradiated neural stem cells after such transfer as shown in Figure 6.

Figure 6.

Fig. 6. Using stem cells from glioblastoma (GB), we demonstrated the presence of a bystander response using a variety of endpoints.

(A-C) U87MG glioblastoma spheroid culture was established after 15 passages in serum-free media with epidermal growth factor (EGF) and fibroblast growth factor (FGF2). Immunostaining of spheroids attached to the fibronectin matrix revealed numerous CD133+ and SOX2+ glioblastoma stem cells. In addition, all cells express glial fibrillary acid protein (GFAP-positive), a glial cell marker. (D and E) Immunostaining and FACS analysis of CD133+ cells in U87MG spheroid culture indicating >70% stem cells. (F and G) Apoptosis of naïve NSC was induced by media transfer from non-irradiated and irradiated spheroid cultures of U87MG cells: GB-0 Gy and GB-10 Gy. Normalized (based on cell number) levels of apoptosis are shown in red. Anti-TRAIL Ab (5µg/ml) and normal IgG, which was used as a control, were added after media transfer, as indicated. These data highlighted the role of TRAIL in modulating the bystander response. (H) Naïve NSC was cultured in differentiation medium for neuronal cells (upper left) and glial cells (lower left). Irradiated medium from U87MG tumor cells strongly suppresses neuronal (upper right showing less green neuronal marker) and modestly on glial cell (lower right showing less green GFAP marker) differentiation.

4 cell


Genomic instability in bystander mammalian cells

Using human skin fibroblasts, the induction of genomic instability among the progeny of bystander cells were examined using the m-FISH assay. Approximately 20% of HSF cells were randomly selected by the image analysis system of the Columbia University microbeam and a lethal dose of 30 alpha particles were delivered to the nucleus of the cells. After irradiation, individual cells were cloned and the bystander, non-irradiated cells were expanded in cultures. The observed chromatid breaks per cell ranged from 0.8 to 1.4 in the control cells over the 3~4 week time point examined. In contrast, there was a significant increase in chromatid aberration per cells in bystander cells. There was a 2.2 fold increase in chromatid break by day 10. These data indicate that genomic instability manifest following ionizing radiation exposure is not dependent on direct damage to the cell nucleus but persists in the progeny of non-targeted/ bystander cells.


Investigators of Project 2




     Tom K. Hei, Ph.D.


     Project Leader








     Hongning Zhou, M.D.

     Research Scientist





     Vladimir Ivanov, Ph.D.

     Research Scientist




     Chai Yunfei, Ph.D.










     Roy Kong-Kwan Lam, B.Sc.

     Gradudate Student