A dose of hope

Medical staff use a Geiger counter to screen a woman for possible radiation exposure in Japan's Hitachi City, on March 16, 2011, after she was evacuated from an area within a 20-km radius of the Fukushima Daiichi nuclear plant. She tested negative for radiation exposure. The nuclear plant, run by Tokyo Electric Power Co., was damaged by the tsunami a week earlier.-HO/REUTERS

An MIT study provides a verifiable framework to estimate the biological response to radiation.

Evidence now seems to be mounting against the Linear No-Threshold (LNT) model of biological response to low-level nuclear radiation. LNT forms the basis of estimating long-term biological effects of ionising radiation (IR) such as cancers arising from environmental radiation exposure above the natural background radiation (NBR) levels, say from Fukushima- or Chernobyl-like nuclear accidents, as well as for setting international standards and public policy for limits of radiation exposure both to the general public and radiation workers.

Living cells are constantly bombarded by IR of various kinds and from diverse sources, including NBR from cosmic rays and natural radioactive isotopes in the environment. It is well known that IR has the potential to damage the DNA. But the body has mechanisms to correct any such damage. However, the background dose rate varies depending upon local geological conditions. This dose rate can be impacted by nuclear weapon detonations, nuclear power operations and nuclear accidents, which in certain circumstances can be much more than the NBR. Any cellular damage that is beyond the capacity of the bodys internal repair mechanisms can lead to cell malfunction or malignancy in the form of cancer.

The effects of IR on DNA are known to be significant for high radiation doses, such as those received by survivors of Hiroshima and Nagasaki. However, the effects of lower doses over longer periods are not so well understood and are more difficult to study.

The LNT model essentially allows one to deal with this uncertainty. The model assumes that biological damage is directly proportional to radiation dose at all dose levels. So observable effects of high doses are extrapolated to lower doses, and it is assumed that the same relationship applies to low doses with no observable effect. That is, there is no safety threshold and radiation is always harmful, never benign, and all levels of exposure come with some health risk, however small. Equivalently, if a particular dose of a given radiation results in one more cancer case in every thousand people exposed, one thousandth of that dose will result in one additional case in every million people and one-millionth will produce one additional case in a billion people. Further, the effect of several low-level exposures is the same as an acute exposure equal to the sum of these small exposures, which is what is meant by linearity of response to radiation.

This approach also forms the basis for public policy in the management of occupational and public exposure. In general, this approach has instilled in some people a deep sense of fear against all radiation and has significantly contributed to the anti-nuclear sentiment worldwide.

There are two opposing schools of thought in the study of biological effects of radiation the threshold model, according to which biological damage begins to occur at doses higher than a particular level, and the LNT model. Further, within the former there is the view that radiation at very small dose levels is, in fact, beneficial. This is what is known as radiation hormesis.

Recent advances in techniques of experimental biology, such as time-lapse photography of cellular processes, and an improved understanding of DNA repair mechanisms in the body have enabled the collection of more direct and conclusive evidence on the effects of low-dose radiation. The evidence for hormesis is perhaps much weaker than that for the threshold effect of IR. Of course, the new studies have all been in a few model animal systems. As such it will be considerably long before any new evidence is regarded as conclusive enough to have an impact on the international standards of radiation protection.

MIT study

The most recent of the new studies is the one by scientists at the Massachusetts Institute of Technology, which was published online in April and in print in May in the journal Environmental Health Perspectives. The MIT team, led by Werner Olipitz, addressed the question: How much additional radiation is too much? The study found no detectable DNA damage in mice exposed to prolonged low-dose radiation (400 times the NBR) for five weeks. It suggests that the existing guidelines for radiation protection may be too conservative. This is the first study to measure genetic damage seen at a level as low as 400 times the background.

This paper shows that you could go 400 times higher than average background levels and youre still not detecting any genetic damage. It could potentially have a big impact on tens if not hundreds of thousands of people living in the vicinity of a nuclear power plant accident or a nuclear bomb detonation, if we figure out just when we should evacuate and when its okay to stay where we are, said Jacquelyn Yanch, one of the authors of the paper.

It is interesting, says Bevin Engelward, another scientist from the team, that, despite the evacuation of nearly 100,000 residents, the Japanese government was criticised for not imposing evacuations on even more people. From our studies, we would predict that the population that was left behind would not show excess DNA damage.

According to Jacquelyn Yanch, most of the radiation studies on which the current conservative evacuation guidelines are based were done to establish safe levels for radiation in the workplace. These were based on acute radiation exposures, which were extrapolated to lower and lower dose rates. Given the enormous costs associated with making constraints on public policy too stringent (or too loose), these studies point to a significant need for additional knowledge regarding the impact of low dose-rate radiation, says the MIT paper.

The average dose due to NBR is about 3 milligray (3 mGy) per year per person. (1 gray = 1 joule of ionising radiation energy per one kilogram of tissue.) The MIT experiment studied the biological damage at a dose rate of 0.002 mGy/min (1,050 mGy/yr), which is about 400 times the average NBR and 200 times higher than the limit for radiation workers recommended by the International Commission on Radiological Protection (ICRP).

However, this is still in the realm of low dose rate as it is only about five times the level found in certain High Level Background Radiation Areas (HLBRAs) such as Ramsar in Iran, which is known to be the highest in the world, or about 30 times the dose rate in the highest radiation-emitting monazite sand regions of Kerala. It is also lower than the dose rate known to impact cancer and longevity studies in animals. The mice used in the study were exposed at this dose rate continuously over five weeks, which amounts to a cumulative dose of 105 mGy. This is roughly equivalent in biological effect to 105 millisieverts (mSv) in humans.

DNA damage

According to the authors, an exposure period of five weeks was chosen because previous studies have shown that a radiation level of about 100 mGy of IR delivered acutely or at one go has been shown to cause DNA damage. DNA damage by acute radiation exposure is known to be a precursory indicator of cancer over a decade later. Therefore, in the study the authors focussed on DNA damage and DNA damage responses. Types of DNA damage fall into two major classes: base lesions, in which the structure of the DNA base (nucleotide) is changed, and breaks in the DNA strand, importantly harmful breaks in both the strands of DNA called the double-strand breaks (DSBs). At the end of five weeks the team found no significant increase in either type.

In all, the experiment used 112 mice, which received a cumulative dose of 105 mGy. The source of radiation was a variable iodine-125 based low-dose irradiator. While the isotope iodine-125 is not found in nature, the authors argue that its gamma-ray emissions are a reasonable proxy for both NBR and environmental contamination because of a nuclear reactor accident or a nuclear explosion as the radionuclide of most concern is caesium-137, which is also a gamma emitter.

For studying DNA base lesions, the spleen was chosen because of radiosensitivity. Spleens of mice exposed to 400 times NBR for five weeks did not show any significant changes in the levels of base lesions. DNA damage occurs spontaneously even at background radiation levels, at the rate of about 10,000 lesions/cell/day. However, self-repair mechanisms within each cell are able to repair most of that damage. The researchers estimated that only about a dozen additional lesions would have been caused by the radiation delivered in the study and all of them seemed to have been repaired. The researchers also found that even if this radiation was delivered acutely, at a four orders of magnitude higher dose-rate of about 71 mGy/min (105 mGy delivered in 1.4 min), no significant change in the levels of base lesions was seen. This led them to conclude that there was no significant impact on the steady state levels of key DNA base lesions that were known to be formed in response to radiation and inflammation, regardless of dose-rate from 0.002 mGy/min to 71 mGy/min.

Radiation-induced DSBs of the DNA are known to be severely cytotoxic and mutagenic and can result in cancer. However, these occur less frequently than radiation-induced base lesions. Using the technique of micronucleus assay to detect these DSBs, the researchers found that while micronuclei formation an indication of chromosomal breaks was induced in the bone marrow of acutely exposed mice (105 mGy at 71 mGy/min), there was no significant difference in the micronuclei frequency in the exposed and the control groups when the same dose was delivered at lower dose-rate of 0.002 mGy/min.

Homologous recombination levels

The authors also studied radiation-induced DSBs in the pancreas using an alternative approach. They assayed for homologous recombination (HR), in which two similar or identical DNA molecules exchange nucleotide sequences, which cells use to repair DSBs accurately. The argument is that if a linear relationship between total dose and the number of DSBs is assumed, a radiation dose of 105 mGy is expected to induce about two DSBs per cell, which may be difficult to detect. However, using certain transgenic mice, called FYDR mice, that reveal modulations in HR levels, the researchers looked for radiation-induced changes in HR. They found that 105 mGy delivered either at a low dose rate or acutely did not induce any HR in the pancreatic cells.

Radiation exposure can also lead to changes in gene expression in certain genes as a response to the stress. These genes belong predominantly to the DNA damage response network. The researchers, therefore, looked for transcriptional changes in a set of genes known to be sensitive to low-dose radiation. They did not find any significant difference in gene expression between irradiated and control group mice. This result indicates that a 400-fold background radiation is not sufficient to affect radiation-sensitive genes in DNA damage response pathways, which implies absence of stress response. Using a much more sensitive technique that suppresses differences between individuals, the researchers, however, found that two genes had altered gene expression under acute conditions but there was no change at low dose rate. The finding suggests that there may be a dose-rate threshold for DNA damage response in terms of cellular gene expression.

Limitations of the study

In conclusion, by using highly sensitive techniques, the study found that low dose-rate radiation (400-fold NBR) delivered over five weeks in vivo in an animal model did not affect DNA base lesion levels, micronuclei frequency, homologous recombination levels or gene expression associated with DNA damage response. Also, an acute exposure of equal dose did induce DNA damage and DNA damage response. The authors, however, also point out the limitations of the experiment. Despite the use of highly sensitive assays for DNA damage response, it remains possible, says the paper, that genetic changes are induced by low dose-rate radiation, but such changes are below the limits of detection for the assays used.

The authors also pointed out that an earlier experiment had found that low dose-rate radiation did induce chromosome aberrations in vitro but the dose rate in that experiment was 10-fold higher than the rate used here. Most DSBs in the present experiment were found to have been rapidly repaired. But the researchers observed that a minor proportion of DNA breaks were complex and had been found to be resistant to repair in other experiments, and such breaks could be there at undetectable levels.

Another important limitation is the type of radiation used. The biological impact of DNA damage varies with the type of radiation. The gamma rays and X-rays used in this study are known as low Linear Energy Transfer (LET) radiation. (LET refers to the average amount of energy lost by radiation over a defined distance, say the energy deposited in 10 cells, as it interacts with matter. Some types of radiation, such as alpha particles, deposit a large amount of energy over a small distance (a few cells) and are called high-LET radiation. Low-LET radiation, on the other hand, penetrates tissues very easily with very little deposition of energy.) Therefore, high-LET radiation is more effective in producing DNA damage than low-LET exposure for a given dose and can cause more complex breaks. However, the researchers pointed out that elevated radiation levels from a contaminated environment were only due to low-LET radiation (particularly from I-125 and Cs-137). While the limitations arising from the types of assays used in the experiment and the use only of low-LET radiation are important aspects, the experiment demonstrated that the potentially harmful impact of radiation was greatly reduced in low dose-rate situations, which challenged the currently followed LNT model for establishing international policy guidelines and regulations for radiation protection.

In the context of the findings of this MIT study, it is important to point out the findings of a study from Lawrence Berkeley National Laboratory (LBNL) last December, which threw important light on DNA repair mechanisms at work at low doses. The study, led by Sylvain Costes, was published in PNAS ( Proceedings of the National Academy of Sciences) in the United States. The researchers observed DNA repair mechanisms in vitro and showed that they worked more efficiently at low IR doses, thus challenging the LNT theory from a different perspective. The finding complements the MIT finding that there is no significant DNA damage at low dose rates even if the exposure is continued over a long period.

The study used time-lapse images of cells to study their response to IR exposure at different dose levels. These images showed that repair proteins tended to concentrate around parts of DNA that suffered DSBs, which are termed as radiation-induced foci (RIF). Over time, the broken strands of DNA moved within the cell to cluster in larger RIFs known as DNA repair centres. What they found was that DSBs, even if they were 1-2 micrometres apart, rapidly clustered to form repair centres. According to the researchers, with increasing radiation dose, multiple repairs could be taking place simultaneously in the repair centres, leading to more errors in the repaired DNA. However, at low levels, it was unlikely that any cell would have to repair more than one DSB at a time and hence the repair mechanism worked better at low doses. The study found 15 RIF/Gy after 2 Gy exposure compared with about 64 RIF/Gy after 0.1 Gy exposure.

It is, therefore, likely that at even lower dose rates of the MIT experiment, the RIF yield would be much higher, resulting in an efficient repair mechanism to restore damaged DNA strands rapidly. Our discovery, says the PNAS paper, of DSB clustering over such large distances casts considerable doubts of the general assumption that risk to IR is proportional to dose, and instead provides a mechanism that could more accurately address risk-dose dependency of IR. But, given that there has been considerable controversy over the existence of RIF and DNA repair centres, it must be borne in mind that it was an in vitro model and may not be completely representative of cellular processes in vivo.

Though these studies are not conclusive enough to revise the present guidelines based on the LNT model, they point to the need for more studies. The MIT study was on a mice model system, and according to Engelward, many studies have shown that mice and humans share similar responses to radiation. But more importantly, the findings of the MIT and Berkeley experiments have provided a framework for future research so that estimation of biological response to IR at low doses, and consequently public policy, is no longer rooted in guesswork but grounded in real verifiable biological evidence.