Effects of radiation felt by non-radiated neighbouring cells prompt a rethink of radiation risk, radiotherapy and radioprotection Dr. Mae-Wan Ho
Linear dose response relationships are routinely used in risk assessments of exposure to environmental hazards, and ionizing radiation is no exception. Typically, effects at high doses that kill cells, cause gene mutations and cancers, are back extrapolated to obtain an exposure limit at which the harm caused is considered miniscule or acceptable in view of the benefits gained. Ionizing radiation was widely believed to cause mutations by directly breaking the bonds of DNA molecules in the nucleus.
In the early 1990s, Hatsumi Nagasawa and John Little at Harvard School of Public Health, Boston, Massachusetts, discovered, to their surprise, that while a linear relationship applies to high doses of a-radiation (from 5cGy to 1.2 Gy, where cGy = 10-2Gy) (see Box), a much enhanced effect was obtained at very low doses of 0.03 cGy to 0. 25 cGy, when 30 to 45 % of the cells in a population of Chinese hamster cells exhibited sister chromatid exchange (SCE involving double-stranded DNA breaks). At that low dose of radiation, only 0.07 to 0.6 % of the nuclei should have been directly hit by an alpha-particle. Yet the frequency of SCE rose rapidly at very low doses reaching a plateau below 1 cGy, after which no further increase occurred with increasing dose, though a decline occurred at higher doses. That was the first indication that damaging signals may be transmitted from irradiated to neighbouring non-irradiated cells in a population, and they called it “the bystander effect” .
In another experiment they looked at mutation frequency of a specific enzyme, and found the same enhanced effect at very low dose. At the lowest dose of 0.83 cGy, the efficiency with which the alpha-particle can induce a mutation increases nearly five-fold; the mutation frequency was the same as that due a dose 100 times as great (0.83 Gy).
Using the then newly developed microbeam of very low dose alpha particles to target individual cells, researchers at Columbia University, New York, showed that hitting the cytoplasm was sufficient to induce mutation in the nucleus . They commented that low dose radiation is all the more dangerous because it does not kill the targeted cell, but allows its influence to spread widely to adjacent cells, thus multiplying the radiation effect (about 100 fold).
Absorbed dose, equivalent dose and effect dose
Radioactivity is measured physically as Curies (1 Ci = 3.7 x1010 disintegrations per second). But that does not take account of the energy of different kinds of radiation and their interaction with biological tissues.
The absorbed dose, Gray (Gy) is equal to and energy of 1 Joule/ kg absorbed.
The equivalent dose Sievert (Si), is weighted by biological potency of different kinds of radiation (1 for g-rays, b-particles, and X-rays, 20 for a-particles and 10 for neutrons). The effective dose also in Sievert takes into account the sensitivities of different tissues, applying weighting factors derived from previous epidemiological studies of radio-induced cancers. Thus, lots of judgements are used in arriving at the effective dose, based on a model of linear energy transfer (and linear dose response relationship) that has proven inapplicable for cells and organisms.
Since then, a wide range of bystander effects in cells not directly exposed to ionizing radiation have been found, which are the same as or similar to those in the cells that were exposed , including cell death and chromosomal instability.
Actually, radiation induced bystander effects have been described as far back as 1954, when factors that cause damage to chromosomes could be detected in the blood of irradiated patients. Carmel Mothersill and Colin Seymour at McMaster University published a key paper in 1997 showing that filtered medium from irradiated human epithelial cells can reduce the survival of unirradiated cells, suggesting that soluble factors produced by the irradiated cells were involved in the bystander effects .
Indeed, serum from cancer patients treated with radiotherapy also causes cell death and chromosomal instability in unexposed cells in culture, and this has been shown as far back as 1968 .
In 2001, researchers at Columbia University, New York used microbeams to target single cells with exactly defined numbers of a-particles. They found that hitting 10 % of the cells induced the same frequency of cancerous transformation as when every cell in the dish was targeted .
More recently, bystander DNA double-strand breaks were induced in a three-dimensional human tissue culture that is closer to in vivo conditions. The results obtained by the team led by Olga Sedelnikova at the National Cancer Institute, Bethesda, Maryland, were much more dramatic. In marked contrast to cultured cells in two-dimensions where maximal DSB occurred 30 minutes after irradiation, the incidence of DSBs in bystander cells reached a maximum between 12 to 48 hours after irradiation, gradually decreasing only over 7 days. At the maximum, 40 to 60 % of cells were affected . These increases in bystander DSBs were followed by increased apoptosis and micronucleus formation, loss of nuclear DNA methylation and increased fractions of senescent cells. The authors commented that treatment of primary tumours with radiation therapy frequently results in the growth of a secondary malignancy of the same or different origin. They raised the question on whether bystander effects could introduce negative complications in radiation therapy, such as genomic instability in normal tissues. They concluded that induced senescence might be a protective mechanism. On the other hand, failure of these protective pathways can lead to the appearance of proliferating, damaged cells and to an increased probability of oncogenic transformation.
New research from the University of Pittsburgh Pennsylvania throws further light on the implications of bystander effects for radiotherapy. It is customary for patients receiving bone-marrow transplant to undergo whole body irradiation to kill the bone marrow cells of the host so as to encourage repopulation by transplanted cells. The researcher found that irradiated mouse recipients significantly impaired the long-term repopulating ability of transplanted mouse haematopoietic stem cells (HSCs) 17 hours after exposure to irradiated hosts, and before the cells began to divide. There was an increase in acute cell death associated with accelerated proliferation of the bystander HSCs. The effect was marked by a dramatic down-regulation of c-Kit (a proto-oncogene), apparently because of elevated reactive oxygen species (ROS). Administration of an antioxidant chemical or ectopically over-expression of a ROS scavenging enzyme catalase improved the function of transplanted HSCs in the irradiated hosts . This obviously has implications for protecting patients during radiotherapy as well as those receiving bone-marrow transplant.
The bystander effect is largely a low-dose phenomenon, appearing at doses below 10 cGy . Higher doses often do not produce bystander effect possibly because the cells targeted are killed before they can influence non-targeted cells. As with the “war on cancer”, numerous attempts have been made to identify the genes or gene products involved in the bystander effects. And as in cancer, genes up-regulated or down-regulated are secondary to a state of electronic imbalance (see  Cancer a Redox Disease, SiS 54) created by the ionizing radiation, which breaks chemical bonds and generate free electrons (see Box 2).
How ionizing radiation can impact on health
Ionizing radiation comes from radioactive decay of unstable chemical elements, which are generated in the nuclear fission process in nuclear power reactors, or in linear accelerators that produce X-rays and electron beams (b-particles) for radiotherapy [12, 13]. In general photons or particles with energy above 10 eV (electron volts) are ionizing.
Nuclear fission is the splitting of the nucleus of a large atom into two, along with a few neutrons and release of energy in the form of heat and g-rays; about 0.2 to 0.4 % of fissions also produce a-particles (nuclei of helium-4 with two protons and two neutrons), or nuclei of tritium (one proton and two neutrons). The fission products are often unstable and hence radioactive; they undergo b-decay giving out b-particles, antineutrinos, and additional g-rays. Antinutrinos pass easily through ordinary matter; consequently, the major ionising radiations that can affect health are a- and b-particles, X-rays, g-rays and neutrons.
a- and b-particles are directly ionizing radiation; they interact directly with atoms, and if the energy is sufficient, knock outer electrons away to produce a free electron and a positively charged ion. A b-particle produces more than 100 ionizing events per cm in its track, whereas an a-particle produces more than 10 000 ionizing events per cm. But while a b-particle can travel for centimetres through tissues, a-particles travel for micrometres only. As the energy of each particle increases, so does the range. Consequently, external sources of a-particles are stopped by the skin, while external b-particles can penetrate into the body. However, inhaled or ingested sources of a-particles can do a lot more damage within the body.
X-rays and g-rays induce ionization indirectly through 3 principal mechanisms: Compton scattering where they are scattered from the outer electrons of atoms, transferring energy to the electrons, and if enough energy is transferred, give rise to a free electron and a positively charged ion. In the photoelectric effect, one of the inner electrons of the atom absorbs the energy of the X-ray or g-ray, and is ejected from the atom, again leaving a positively charged ion. Following this, one of the outer electrons ‘falls’ in to fill the vacancy, and X-ray is emitted from the atom. In pair formation, the x-ray or g-ray interacts with the electric field of the nucleus, and is converted into an electron and a positron, the positron in travelling through the tissue material will usually react with another electron and become converted back to two X-rays or g-rays.
Neutrons are scattered directly from the atomic nuclei of atoms, resulting either in losing energy that is released as g-rays or else it is absorbed by the nuclei resulting in a new nucleus (element) being formed. If the new nucleus is unstable, radioactive decay occurs creating a-, b- or g-rays. The second option can only occur if the neutron is sufficiently slow, and that is what happens in the nuclear fission process in nuclear power reactors.
Some of the free electrons generated by the ionizing radiation may be energetic enough to cause ionizations of their own; this is the secondary photoelectron effect of ionizing radiation.
When cells are irradiated, it is likely that ionization of one or more of the atoms on DNA molecules will occur in a direct hit, breaking the DNA chain or the links between chains. However, direct attack of radiation on the structure of DNA is not the only way radiation affect cells. The human body is about 70 % water; hence water is probably the most frequent target of ionizing radiation. Ionization of water leads to the formation of reactive oxygen species (ROS) (see Box 3) that damages DNA, lipids, proteins, carbohydrates, and other molecules. It is becoming increasingly clear that ROS is a major culprit in the bystander effect, as suggested by those who discovered the effect [1, 2]. This has been confirmed by more recent findings.
Reactive oxygen species generated from water 
Oxygen is the most important electron acceptor in the biosphere. It readily accepts unpaired electrons to give rise to a series of partially reduced species collectively known as reactive oxygen species (ROS). These include superoxide O·2-, hydrogen peroxide H2O2, hydroxyl radical HO· and peroxyl radical OO·, which may be initiate and propagate free radical chain reactions damaging to cells. Hydroxyl radicals are generated by ionizing radiation either directly from water, or indirectly by the formation of secondary partial ROS that are subsequently converted to hydroxyl radicals by metabolic processes. Gamma rays, beta and alpha particles are all able to ionize water to produce hydroxyl radicals, the most reactive, and therefore potentially the most hazardous. Hydroxyl radicals have a very short persistence time, while hydrogen peroxide is the most long-lasting. Hydrogen peroxide can diffuse freely and can generate hydroxyl radicals by reacting with free electrons:
H2O2 + e- → HO· + HO- (1)
Oxidative attack on proteins destroys their enzyme, receptor and other biological function; damage to DNA causes mutations and chromosomal rearrangements; and peroxidation of lipids destroys membrane structure and function.
More than 80 % of energy of ionizing radiation deposited in cells results in the ejection of electrons from water. Subsequent reactions with surrounding water results in the formation of several reactive species: eaq- (hydrated free electron) HO· (hydroxyl radical, the most important reactive oxygen species), H· (hydrogen radical), H2 (hydrogen gas) and H2O2 (hydrogen peroxide, a stable and diffusible reactive oxygen species). These products react rapidly with each other and with surrounding molecules. In the presence of O2, superoxide radicals (another reactive oxygen species) are formed:
eaq- + O2 → O·2- (1)
H· + O2 → O·2- + H+ (2)
Superoxide generates hydrogen peroxide on a longer time scale:
2 O·2- + H+ → O2 + H2O2 (3)
Because of their instability, most of the reactions generating the primary radical products will have taken place within 1 millisecond, but superoxide and H2O2 will persist and diffuse to more distant sites.
Cellular damage by hydroxyl radical attack depends partly on the antioxidant status of the cell and partly on the availability of reducing systems capable of reducing or activating superoxide or hydrogen peroxide. The cellular antioxidant status determines the intracellular concentration of ROS. It has been shown that the effects of H2O2 resemble those of ionizing radiation. Cells exhibiting high levels of SOD, catalase, and peroxidase activity are relatively less vulnerable to secondary effects of radiation. Glutathione peroxidase catalyses the reaction:
H2O2 + 2 GSH (reduced glutathione) → 2 H2O + GSSG (oxidized glutathione) (4)
The activity of this peroxidase depends on the availability of reduced GSH. Regeneration of GSH from GSSG by glutathione reductase requires reduced nicotinamde adenine dinucleotide phosphate (NADPH) as electron donor.
The hydroxyl radical can be produced from more stable ROS via the participation of an electron donor, and many transition metal ions can act as electron donors:
H2O2 + Fe (II) → Fe(III) + HO- + HO· (5)
Thus, hydroxyl radicals are generated from H2O2 at sites where reduced transition metals are present.
A team led by Aleksei Ermakov at the Research Centre for Medical Genetics, Russian Academy of Medical Sciences in Moscow Researchers showed that an extracellular DNA (ecDNA) derived from the cell genome participates in the bystander effect induced by X-ray exposure in human lymphocytes and human umbilical-vein epithelial cells . Their previous work suggested that radiation-sensitive cells undergoing apoptosis serve as a source of ecDNA fragments that diffuse in the medium and bind to DNA receptors on the surface of bystander cells. Bystander effects could be stimulated by ecDNA of irradiated cells but not by ecDNA produced by normal cells. In a new study, the team tested the idea that the difference between the two types of ecDNA is due to DNA oxidation events occurring during and after irradiation. They compared the production of NO (nitric oxide, a free radical and reactive oxygen species) and ROS in human endothelial cells that were irradiated at a low dose radiation, or exposed to the ecDNAR extracted from the media conditioned by irradiated cells, or exposed to the genomic DNA oxidized in vitro by treatment with H2O2, (DNAo1), or H2O2 plus uv light (DNAO2more strongly oxidizing). They found that all three treatments gave similar responses. The production of NO at 2h was suppressed at low doses of 0.03 Gy and 0.1 Gy but increased at 0.5 Gy or higher. Similarly, the ecDNAR extracted from media conditioned by irradiated cells decreased NO but not the extracellular DNA from non-irradiated cells; the oxidized DNA o1 and more so DNAO2 also reduced NO. ROS levels in general were increased in all three treatments by 1.2 to 1.8-times the controls with ecDNAR and oxidized DNA o1 and DNAO2 to larger extents than the direct radiation, or the bystander effect due from the conditioned medium.
Other researchers have shown that the major source of ROS in endothelial cells is the activity of NAD(P)H-oxidases, predominantly one encoded by the NOX4 gene. Irradiation with 0.1 Gy and treatment with ecDNAR led respectively to a 3-fold and 1.7 fold increase in NOX4 mRNA, while oxidized DNA stimulated transcription 5-15 fold compared with unoxidized DNA.
Also in previous work by the Russian team, the bystander effect involves DNA-binding to Toll-like receptor TLR9. This was confirmed by blocking the TLR9 response with chloroquine and oligonucleotide 2088, which suppressed the increase in ROS production and eliminated the effects of ecDNAR.
The team suggested that the bystander effect-like properties of ecDNAR and oxidized DNA may be used for the development of novel anti-tumour therapy that may stimulate cell death without actual irradiation, or synergistically with reduced irradiation doses.
Another way low dose ionization radiation can be amplified and appear as bystander effects is through scattering of photons through the tissues. Photons or particles can bounce off one target atom and strike another, generating a further free electron (see Box 2).
A research team at the Maria Sklodowska-Curie Memorial Cancer Centre and Institute of Oncology Gliwice Branch, Poland investigated direct and bystander effects induced by scattered radiation in two human cell lines – normal bronchial epithelial cells BEAS-2B and lung cancer epithelial cells A549 – placed in a bath of water at different depths and subjected to irradiation by 6 MeV photon beam or 22 MeV electron beam (5Gy maximum dose), and examined for apoptosis and micronucleated cells .
They found that for electron radiation both the numbers of apoptotic and micronucleated cells were greater than expected from the corresponding received dose, and the discrepancy between observed and expected becomes larger with increased medium depth. At a depth of 15-17 cm, the observed was ten times the expected, while micronucleated cells was about 2-3 fold. For photon radiation the biological effect did not differ significantly from expected value because photon radiation penetrates the medium better. When cells were placed outside the radiation field or under a shield, differences from expected dose were also found for both photon and electron, but no depth dependence was observed. For cells exposed outside the field of the photon beam, apoptosis was again about 7-10 fold the expected while micronuclei formation was 4-5 fold. For shielded cells under photon irradiation, apoptosis was about 3-fold while micronuclei was about 1.2-fold. For cells exposed outside the radiation field of the electron beam, again, a 10-fold difference from the expected, and for micronucleated cells, 1.5 to 4-fold in BEAS cells, and 4-7 fold in A549 cells. All the irradiated cell medium, when added to non-irradiated A549 cells gave a 2-fold increase in micronucleated cells and a 2-fold increase in apoptotic cells, regardless of the dose of irradiation or whether it was inside the beam, outside the beam or shielded.
Apart from the bystander effects mediated through the exposed cell medium, these experiments indicate that secondary photoelectron scattering may be involved in the biological effects of low-dose radiation. This has been suggested by research published in the early 1990s . Monte Carlo track structure methods were used to illustrate the importance of low-energy electrons produced by low linear-energy-transfer radiations. These low-energy secondary electrons contribute substantially to the dose in all low-LET irradiations, and account for up to nearly 50 % of the total dose imparted to a medium when irradiated with electrons or photons. Up to 50 % of secondary electrons themselves can also undergo further scattering and to generate more free electrons. For most ionizing radiations, nearly 50 % of all ionizations are due to secondary electrons with starting energies less than 1 keV.
Risk assessment and radiation protection have been based on extrapolation from known epidemiological data that mainly relate to high dose effects that assume a linear dose-response relationship even at very low doses . This is clearly untenable in view of the bystander effects at low doses, which amplify the effective dose and harm caused.
The best available evidence suggests that bystander effects are mediated by ROS. ROS is well-known to be involved in general oxidative stress, with many downstream effects that mirror bystander effects: DNA breaks, genome instability, cell death, cancer, including cell senescence and aging , and cataracts . It is notable that these effects are appearing as significant health impacts linked to the Chernobyl fallout  (Chernobyl Deaths Top a Million Based on Real Evidence, SiS 55). The pro-nuclear lobby and regulators should stop denying these impacts and governments should devote much more resources to studying them instead, to prevent repeating the humanitarian disaster in the wake of the Fukushima meltdown (see  Truth about Fukushima, SiS 55).
The involvement of ROS also suggests that antioxidant interventions should be considered as a mitigation of bystander effects in those exposed or still being exposed to the Fukushima and Chernobyl fallouts. This is a matter of some urgency. Among the most promising findings are the well-known benefits of green tea in cancer prevention (see  Green Tea Against Cancers, SiS 33), and its many antioxidants polyphenols that probably account for reducing risks of heart disease, cancers, Alzheimer’s obesity, arthritis, diabetes, and a host of other conditions associated with oxidative stress (see  Green Tea, The Elixir of Life? SiS 33). New research from the Radiation and Cancer Therapeutics Lab at Jawaharlal Nehru University, New Delhi, and the Central University of Gujarat in India indeed shows that one of the main green tea polyphenol, EGCG (epigallocatechin-3-gallate) is most efficient at protecting DNA against g-radiation induced breaks both inside and outside the cell, and also protects cells against radiation-induced cell death, lipid peroxidation and membrane damage (see  Green Tea Compound for Radioprotection, SiS 55).
As far as cancer radiotherapy is concerned, the bystander effects mean that the radiation beam will cover a wider area than the physical beam, and the potential harm may outweigh the presumed benefit. The same goes for diagnostic radiology, as it occurs at doses that might induce more harmful bystander effects than the potential benefit the procedure might deliver. It is also possible that antioxidants could offer radioprotection against these procedures.
Article first published 28/05/12
Got something to say about this page? Comment
There are 3 comments on this article so far. Add your comment above.
Rory Short Comment left 28th May 2012 21:09:49
It would seem that we mess with radiation at our own peril. Do microwaves also have bystander effects I wonder?
Susan Rigali Comment left 29th May 2012 06:06:01
To Rory & Mae-Wan Near ultraviolet, visible light, infrared, microwave, radio waves, and low-frequency RF (longwave) are all examples of non-ionizing radiation. Visible and near ultraviolet may induce photochemical reactions, or accelerate radical reactions, such as photochemical aging of varnishes or the breakdown of flavoring compounds in beer to produce the "lightstruck flavor". Near ultraviolet radiation, although technically non-ionizing, may still excite and cause photochemical reactions in some molecules. This happens because at ultraviolet photon energies, molecules may become electronically-excited or promoted to free-radical form, even without ionization taking place. The light from the Sun that reaches the earth is largely composed of non-ionizing radiation, since the ionizing far-ultraviolet rays have been filtered out by the gases in the atmosphere, particularly oxygen. The remaining ultraviolet radiation from the Sun is in the non-ionizing band, and causes molecular damage by photochemical and free-radical-producing means that do not ionize. wikipedia on non-ionizing radiation. Mae-Wan, I have noticed when cooking with solar thermal energy that foods have a longer shelf life. After reading this I am wondering if such damage could occur to food at high solar thermal temperatures. What I have read indicates the thermal heat is from infrared radiation, I have to wonder to what extent ultra-violet is absorbed? Since my main interest is in processing and preserving foods I suppose a slower process with lower temps would be ideal for the individual, but high temperatures for larger scale production as an alternative cooking source probably needs some review. Do you know of any studies or sources for this type of information? Thank you
Michael Godfrey Comment left 29th May 2012 22:10:16
A very pertinent paper not only relating to the generally gung-ho attitude to radiation in the profession but also to current radiation from Japan. The latter is now also in the marine food chain as confirmed by Caesium 134 in blue-fin tuna off San Diego 5 months after the Fukushima disaster. Even more evidence that we need to maximise antioxidant levels on a daily basis and that those behind the FDA and Codex must be prevented at the highest level from their protracted attacks on the health industry.