Columbia University Medical Center
Center for Radiological Research

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

Project 3: Intercellular Communication and the Radiation-Induced Bystander Effect

Project Leader: Edouard I. Azzam


Traditionally, it was assumed that ionizing radiation (IR) produced genetic damage and other stressful effects only in cells where the radiation directly traversed the nucleus. However, it is now understood that the surrounding non-irradiated cells (i.e. bystander cells) are also affected through intercellular communication with the nearby irradiated cells [1]. Remarkably, elevated levels of DNA damage and oxidative stress also occur in progeny of the bystander cells (Fig. 1). The goal of Project 3 is to exploit fundamental knowledge on permeability properties of gap junction proteins (connexins) to investigate the role of different connexins in the propagation of radiation-induced non-targeted effects. The focus is on studies of biological changes in non-irradiated cells that are in the vicinity of irradiated cells (bystander effects), amplification of toxic effects among irradiated cells (cohort effects), and enhancement in the level of spontaneous genetic damage in progeny of bystander cells (genomic instability). The objective is to gain greater understanding of the mechanisms underlying the role of intercellular communication in the spread of IR-induced harmful effects. The results may provide insight into potentiating the benefits of cancer radiotherapy by enhancing its killing effects and protecting against the propagation of damaging effects to healthy tissue.

Fig.1. Ionizing radiation (IR) induces targeted and non-targeted (bystander) effects. Communication of stress-inducing molecules from cells exposed to IR propagates stressful effects, including oxidative stress, as well as genetic and epigenetic changes, to the bystander cells and their progeny. The induced effects may be similar in nature to those observed in progeny of irradiated cells.


Research Aims

The central hypothesis is that the cellular microenvironment modulates gap-junction gating, and thereby propagation of biological effects between irradiated and bystander cells. These events are modulated by redox modulated events and DNA repair, and result in transient and persistent changes in affected bystander cells and their progeny (Fig. 2). In four inter-related specific aims, we are investigating the effects of selective properties of connexin 26 (CX26), connexin 32 (CX32) and connexin 43 (CX43) channels on the proliferative capacity of irradiated and bystander human cells. Together with Project 2, we are examining the effects of cyclooxygenase-2 (COX-2) signaling on the gap junction permeabilities that mediate the expression of bystander effects. With Project 1, we are analyzing intercellular communication that results in regulation of the DNA repair and checkpoint proteins, Rad9 and Translationally Controlled Tumor Protein (TCTP), in bystander cells. With relevance to long term health risks, we are exploiting the resources in our Technical Core to assess, in normal cells, altered signaling pathways and genomic instability in progeny of bystander cells.

Fig.2. Mechanisms underlying ionizing radiation-induced bystander effects. Signaling molecules are propagated among irradiated and bystander cells through direct intercellular communication via gap junctions or through diffusible secretion in the surrounding environment. The expression of propagation of bystander effects is highly dependent upon the phenotype of both the irradiated and bystander cells.

Research Highlights

A layered cell culture strategy to examine gap junctional communication in the propagation of harmful effects of ionizing radiation

We use a layered cell culture strategy that allows isolation of pure bystander cell populations (Fig. 3) to investigate radiation-induced bystander effects. In this system, unirradiated cells are intimately co-cultured with other cells that were exposed to ionizing radiation. The cells destined to be bystanders (e.g., AG1522 normal human diploid fibroblasts) are grown to the confluent state on the bottom side of Transwell inserts with 1 µm pores. To this end, the cells are seeded on inverted inserts and allowed to attach, which occurs within ~ 1 h. The inserts are then re-inverted and placed in 6-well plates containing growth medium. The cells are grown with daily feeding with fresh medium. Typically, co-culture with irradiated cells occurs 2-4 days after a confluent monolayer of cells is formed across the bottom of the inserts. For the co-culture, irradiated cells are seeded on top of the insert and maintained with the bystander cells for different periods of time. Cells that are gap junction proficient establish junctional communication within 2 h of co-culture (Figure 4).

Fig.3. A layered tissue culture strategy for the study of IR-induced bystander effects.


Fig.4. Junctional communication between AG1522 cells grown on either side of a Transwell insert with 1 µm pores. AG1522 cells seeded on the top side of the insert were loaded with Calcein AM (green fluorescence) and CellTracker Orange (red fluorescence). Within 1.5 h of co-culture, significant transfer of Calcein AM occurred to AG1522 cells growing on the bottom side of the insert.

Irradiated normal human fibroblasts propagate stressful effects to bystander cells, which result in significant DNA damage in progeny of the bystander cells

Bystander AG1522 cells were intimately co-cultured for 5 h with irradiated AG1522 cells. They were subsequently isolated (with 99.8% purity) and assayed for various stress markers. The results in Figure 5 show that the bystander cells experience stressful effects as illustrated by the increase in the level of P-p53Ser15.

Fig.5. Stress responses in bystander cells co-cultured with irradiated cells. AG1522 bystander cells were harvested after 5 h of co-culture with 0 (C), 40 cGy α-particles (α) or 400 cGy γ-rays (γ) irradiated cells. Equal amounts of protein were analyzed by Western blot for phosphorylation of p53 on serine 15, a marker of DNA damage. Expression ratios were normalized to loading control (α -tubulin).

Delayed effects in progeny of bystander cells were also examined. Subsets of the isolated bystander cells were propagated for 25 population doublings (PD), and micronuclei formation, a form of DNA damage that arises mainly from DNA double strand breaks, was assessed (Fig. 5). Relative to control, a significant increase in spontaneous micronuclei was observed in progeny of bystander cells that were co-cultured with cells exposed to 3.7 MeV a particles or 137Cs ? rays at mean absorbed doses that result in ~10% clonogenic survival (Fig. 6). This increase was not detected if 18-?-glycyrrhetinic acid (AGA), a reversible inhibitor of gap junction communication, was present during co-culture. In contrast, incubation of the co-culture with lanthanum chloride (La3+), an inhibitor of hemi-channels, did not result in attenuation of micronuclei in progeny cells.

Fig.6. Micronuclei formation in progeny of bystander AG1522 human fibroblasts at 25 population doublings (PD) following isolation from co-culture with irradiated AG1522 cells. (C) Control, (?) alpha-particle- or (?) ?-irradiated cells

Gap junctions consisting of CX26 and CX43, but not CX32, accelerate the expression of stressful radiation-induced bystander effects

We have shown by different approaches that GJIC contributes to the spread of IR-induced toxic and clastogenic effects from directly traversed normal or tumor human cells to bystander normal cells [2, 3]. We also found that GJIC amplifies toxic effects among irradiated tumor cells and that the process can be utilized to enhance the killing of GJIC-proficient tumor cells [4-6]. Another novel finding is that cell responses vary depending on the nature of CX. For example, CX26 and CX43 channels mediated intercellular effects leading to DNA damage in bystander normal cells. By contrast, the presence of CX32 channels led to intercellular effects that mitigated the deleterious effects, including DNA damage and cell killing [3, 7]. Thus, we propose that different CX channels may be exploited to enhance efficacy of tumor radiotherapy and to spare harm to bystander normal tissues [1].

Radiation quality, gap junctions and genomic instability

We have shown that the persistence of oxidative stress in progeny of irradiated and bystander cells greatly depends on radiation quality and dose. These findings are particularly important for understanding the long-term health risks of exposure to IR [8, 9]. For example, we found that the progeny of bystander cells are at a greater risk of neoplastic transformation [10], implying the risk of developing second cancers following therapeutic irradiation of the primary cancer.

We have shown that the persistence of oxidative stress in progeny of irradiated and bystander cells greatly depends on radiation quality and dose. These findings are particularly important for understanding the long-term health risks of exposure to IR [8, 9]. For example, we found that the progeny of bystander cells are at a greater risk of neoplastic transformation [10], implying the risk of developing second cancers following therapeutic irradiation of the primary cancer.

The cross-talk between junctional communication and COX-2 in expression of radiation-induced bystander effects

We have determined the crosstalk between junctional communication, through CX32 and CX26 channels, and COX2 in mediating the expression of radiation-induced bystander effects. COX-2 knockdown attenuates radiation bystander effects mediated by gap junctions and diffusible factor(s) [3].

Connexins, Rad9, TCTP and radiation bystander effects

We found that RAD9 and TCTP are up-regulated in bystander cells in a manner dependent on the expressed CX channel (Fig. 7). More importantly, we have discovered a novel role of TCTP in DNA damage sensing and repair (Fig. 8) [11].

We used HeLa cells as a model system. These cells, which lack endogenous connexins, were stably transfected with either empty vector (henceforth coined HeLa parental) or inducible vectors containing CX26 or CX32 cDNA (henceforth coined HeLa Cx26 or HeLa Cx32). In the transfected cells, the expressed connexins localize in the plasma membrane, and form functional gap junctions. We have previously shown that HeLa CX26 and HeLa CX32 cells also form functional gap junctions with normal human diploid AG1522 fibroblasts [7]. When confluent cultures of HeLa cells expressing either CX32 were exposed to very low mean absorbed doses of 3.7 MeV α-particles, TCTP level was upregulated to a greater extent in cells from HeLa CX32+ than HeLa CX32- cultures. Six-fold increase in HeLa Cx32+ exposed to a mean dose of 0.6 cGy by which less than 4% of cells is traversed by a particle track through the nucleus versus 3-fold increase in HeLa CX32- cells (Fig. &). Note that in the case of 3.7 MeV α-particles, the generated δ-rays would have a range of only ~1 μm. Whereas Rad9 was up-regulated to a similar extent whether HeLa cells expressed CX32 or not, its basal level was strikingly 14-fold higher in HeLa Cx32+ than in HeLa CX32- cells. Those changes in TCTP observed in HeLa cells expressing CX32 were not observed in HeLa cells expressing CX26. Together the data strongly support the role of gap junction selectivity in regulating Rad9 and TCTP expression in bystander and irradiated cells.

Fig.7. Western blot analyses of TCTP and Rad9 expression in HeLa cells. HeLa cells CX 32 was expressed [(+) doxycycline] and cells were irradiated in the confluent state and were harvested for analyses 3 h later.

Our accumulating results indicate that expression of CX32 channels promote effects that prevent the propagation of harmful effects [3, 4, 7]; the data in Fig. 7 indicate that intercellular communication through these channels induce the upregulation of Rad9, a radiation protective protein, and TCTP. In follow up experiments, we examined whether TCTP has role in radioprotective functions.

Role of the Translationally Controlled Tumor Protein in DNA Damage Sensing and Repair

We tested the hypothesis that TCTP plays a critical role in response to DNA damage and that this function is essential particularly for the survival and genomic integrity of irradiated cells. We showed that upon exposure of normal human cells to IR, the TCTP protein level was greatly increased, with a significant enrichment in nuclei. The protein was also upregulated in tissues of low-dose-irradiated mice. Interestingly, TCTP up-regulation was dependent on early sensors of DNA damage, specifically ATM and the DNA-dependent protein kinase (DNA-PK). Importantly, it was associated with protective effects against DNA damage. Like in the case of cells treated with DNA repair inhibitors, TCTP-deficient cells also failed to repair γ-ray-induced chromosomal damage. In chromatin of irradiated cells, TCTP was found to physically interact with ATM and to exist in a complex with γH2A.X, in agreement with its distinct localization with the foci of the DNA damage marker proteins γH2A.X, 53BP1 and P-ATM (Fig. 8) [11].


Fig.8. TCTP interacts with components of DNA damage sensing and repair. (A) Benzonase-treated chromatin-enriched fractions that were isolated 30 min after irradiation (IR) with 0 (-) or acute 50 cGy (+) from confluent U2OS cells were immunoprecipitated with anti-ATM or anti-TCTP antibodies, or normal mouse or rabbit serum (PI). Immunoblots were then reacted with antibodies against ATM, TCTP, ?H2A.X or H2AX. (B) Benzonase-treated nuclear extracts isolated 30 min after exposure of U2OS confluent cells to 0 or acute 50 cGy were immunoprecipitated with anti-TCTP, anti-p53 or control anti-TBP antibodies. Mouse or rabbit preimmune serum (PI) was used as a control. Immunoblotting was performed using antibodies against p53, TCTP or TBP. (C) Immunoblotting of TCTP, Ku70 and Ku80 in benzonase-treated nuclear extracts of control unirradiated U2OS confluent cells after immunoprecipitation with either normal serum (PI) or antibodies against TCTP, Ku70 or Ku80. (D) Untreated or ?-irradiated (acute 100 cGy) AG1522 asynchronous cells were pre-extracted, fixed 1 h later, and immunostained in situ with anti-TCTP, anti-P-ATM (S1981), anti-?H2A.X, or anti 53BP1 antibodies. Bars, 10 µm. (E) Quantitative assessment of co-localization of TCTP foci with those of P-ATM (S1981) (left panel), ?H2A (middle panel) and 53BP1 (right panel) in AG1522 asynchronous cells at 1 h after exposure to 50, 100 or 200 cGy.

We have also shown that TCTP interacts with major elements of non-homologuous end joining and homologuous recombination modes of repair of DNA double strand break, a particularly harmful form of DNA damage. These interacting proteins include the DNA-binding subunits, Ku70 and Ku80, of DNA-PK, and filamin A, respectively. In addition, TCTP physically interacts with p53, a critical protein involved in maintenance of genomic integrity, and TCTP knockdown shortened the radiation-induced delay in G1 phase of the cell cycle. The latter effect was associated with attenuated induction of p21Waf1, an inhibitor of master regulators of the cell cycle. The loss of normal G1 checkpoint control disrupts DNA repair and is an early step in carcinogenesis, which highlights the role of TCTP in regulating a process that maintains healthy survival.  Together, our results identify TCTP as a new member of a group of proteins involved in DNA damage response (Fig. 9) [11].

Fig.9. The role of TCTP in DNA damage sensing and repair. TCTP is upregulated by ionizing radiation; it interacts with elements of DNA damage sensing and repair, and modulates radiation-induced cell cycle checkpoints. (IR, ionizing radiation; P, phosphorylation; MRN, MRE11–RAD50–NBS1; NHEJ, non-homologous end joining; HR, homologous recombination; HSPs, heat shock proteins).


Investigators of Project 3

Edouard I Azzam

Edouard Azzam, PhD
Project Leader

Sonia de Toledo, PhD
Adjunct Assistant Professor

Ye Zhao

Ye Zhao, Ph.D.
Post-Doctoral Fellow

Nicholas Colangelo
MD/PhD student

Neeha Sharma

Neha Sharma
PhD student

Jason Domogauer
MD/PhD student


Jie Zhang

Jie Zhang, MD, PhD

Manuela Buonanno, PhD

Narongchai Autsavapromporn, PhD

Géraldine Gonon, PhD



1. Azzam, E.I., et al., The Ionizing Radiation-Induced Bystander Effect: Evidence, Mechanism and Significance, in Pathobiology of Cancer Regimen-Related Toxicities, S.T. Sonis and D.M. Keefe, Editors. 2013, Springer: New York, NY. p. 42-68.
2. Autsavapromporn, N., et al., Gap junction communication and the propagation of bystander effects induced by microbeam irradiation in human fibroblast cultures: the impact of radiation quality. Radiat Res, 2013. 180(4): p. 367-75.
3. Zhao, Y., et al., Connexins and Cyclooxygenase-2 Crosstalk in Expression of Radiation-Induced Bystander. British Journal of Cancer, 2014. In press.
4. Autsavapromporn, N., et al., Intercellular communication amplifies stressful effects in high-charge, high-energy (HZE) particle-irradiated human cells. J Radiat Res (Tokyo), 2011. 52: p. 408-414.
5. Autsavapromporn, N., et al., The role of gap junction communication and oxidative stress in the propagation of toxic effects among high-dose a-particle-irradiated human cells. Radiat Res, 2011. 175(3): p. 347-57.
6. Autsavapromporn, N., et al., Participation of gap junction communication in potentially lethal damage repair and DNA damage in human fibroblasts exposed to low- or high-LET radiation. Mutat Res, 2013. 756(1-2): p. 78-85.
7. Autsavapromporn, N., et al., Human Cell Responses to Ionizing Radiation are Differentially Affected by the Expressed Connexins. J Radiat Res (Tokyo), 2013. 54(2): p. 251-259.
8. Autsavapromporn, N., et al., Genetic changes in progeny of bystander human fibroblasts after microbeam irradiation with X rays, protons or carbon ions: The relevance to cancer risk. Int J Radiat Biol, 2014. Submitted.
9. Buonanno, M., et al., Long-term consequences of radiation-induced bystander effects depend on radiation quality and dose and correlate with oxidative stress. Radiat Res, 2011. 175(4): p. 405-415.
10. Buonanno, M., S.M. de Toledo, and E.I. Azzam, Increased frequency of spontaneous neoplastic transformation in progeny of bystander cells from cultures exposed to densely-ionizing radiation. PLoS One, 2011. 6(6): p. art. no. e21540.
11. Zhang, J., et al., Role of the translationally controlled tumor protein in DNA damage sensing and repair. Proc Natl Acad Sci U S A, 2012. 109(16): p. E926-33.