The low-dose phenomenon: How bystander effects,genomic instability,and adaptive responses could transform cancer-risk models

Abstract

From the atomic bomb dropped over Japan to nuclear accidents at Three Mile Island, Chernobyl, and the Fukushima Daiichi Nuclear Power Station, there is strong public demand for information on the cancer risks from radiation exposure. In this article, the author explores some of the biological phenomenon that could alter or confirm current concepts of low-dose effects. Reviewing bystander effects, adaptive responses, and genomic instability, the author writes that these phenomena could revolutionize conventional understanding of how to approach cancer risk assessments in low-dose, possibly protracted, environments. Though current consensus supports a linear no-threshold model, evidence suggests that these biological responses just may overturn that thinking.

More than 60 years after the United States dropped atomic bombs on Hiroshima and Nagasaki, the specter of radiation-induced carcinogenesis continues to hang heavily over populations around the world. Nuclear reactor accidents, such as Three Island Mile and Chernobyl, managed to reliably stoke public fears over the years. And this was yet again the case in March 2011, when a 9.0 earthquake triggered a 45-foot tsunami, causing a triple meltdown at the Fukushima Daiichi Nuclear Power Station—contaminating both the land and ocean, and exposing both nuclear workers and the general public to doses beyond those associated with natural background radiation.

Ever since the atomic bombs were dropped on Japan, heralding the nuclear age, the public’s concern has been quite straightforward: When it comes to radiation exposure at low to moderate doses, what should or should not be considered when deciding the true extent of individual risk to radiation-induced disease? In responding to this question, there are aspects of emerging biology that deserve attention.

Thousands of studies have analyzed various aspects of radiation-induced carcinogenic events—and it is from these intensive efforts that a few consensus findings can be distilled. For one thing, in most tissue types, radiation exposure induces carcinogenic effects—like DNA damage in critical genes—that increase linearly with dose, from low to moderate doses; and, for another, in most tissue types, there is no apparent threshold in dose for these carcinogenic effects, both of which give rise to the linear no-threshold theory (LNT). ¹ Studies have not only found that there is a latent period between exposure and the appearance of the carcinogenic event, but that there is a lifetime risk associated with exposure to radiation, highlighting that the degree of the risk may change with time. Age at exposure is important: Studies have found that, in most cases, the younger a person is at exposure, the higher the risk of cancer. Another matter of consensus has been that, if exposure is concentrated in one particular organ—the brain, for example—an estimate of the total-body risk (the average dose over the entire body) is not a good measure of the overall risk, which might be higher to the particular organ.

The scientific community (BEIR, 2006) came to these generalized conclusions from biological experimentation and observation, mostly derived by establishing quantifiable phenomenological endpoints—for example, total tumor incidence per unit dose in exposed animals and other animal and cellular assays. But how do these results over the past decades relate to what can be discerned from molecular and genetic studies, which attempt to deduce mechanistic explanations for radiation-induced carcinogenesis?

At low dose, three phenomena are emerging as potentially important:

bystander effects, in which ionizing radiation hits one cell and transmits a signal, through cell-to-cell communication or through extracellular medium, to have an effect on cells that were not hit—that is, unirradiated cells show irradiated effects solely based on proximity;

adaptive response, in which a small priming dose or protracted dose reduces the effect of a subsequent small but larger dose;

genomic instability, in which initial damage to DNA alters it or its duplicating mechanism, so that additional genetic change is increased during subsequent divisions of a surviving cell.

Currently, there is a considerable push to understand bystander effects, genomic instability, and adaptive response after radiation exposure in low doses—less than 0.1 sievert (Sv), delivered acutely—or low-dose rate—less than 0.1 Sv per hour. All three phenomena have been observed at low doses of radiation and also at much later times following the initial radiation exposure.

Post-exposure: Frequency and latency

There are two aspects to cancer rates that must be considered: the likely frequency of a cancer-initiating event that leads to development of a cancer phenotype, and the latency period, or the amount of time that elapses before the cancer is biologically detectable. Radiation exposure and other environmental factors, like smoking or diet, may interact to reduce the latency period between initial exposure and the appearance of the disease.

For most of the first half of the twentieth century, many reports highlighted cancer rates among nuclear-industry workers, as well as people who had been exposed to radiation or radioactive materials through medical and diagnostic procedures. Most of these examples were anecdotal and did not lend themselves to strict scientific study due to a lack of clear dose estimates. ² Nonetheless, they did manage to draw scientists’ attention to the dangers of radiation exposure, eventually leading to both the systematic studies of the carcinogenic effects of radiation and the systematic efforts to limit human exposure. Recently, through more sensitive assays, advances in molecular staining for biological events, such as apoptosis (programmed cell death) or the detection of DNA-strand breaks, have increased the ability to spot deleterious events after low-dose radiation exposure. In addition, these advances have opened the door to examining molecular-response processes that occur in and between cells; this new understanding of molecular processes not only may affect the dose response or risk evaluation after low doses, it also could drive how the carcinogenic effects develop in a person’s body for years after the initial exposure to radiation.

Methods and evidence

There is a cautionary note that must be emphasized: All the tumor-incidence data used by epidemiologists look at the same final endpoint, i.e., an actual tumor; but by that endpoint, any molecular, cellular, and genomic pathway that contributed to the tumor development has already occurred. If bystander effects are present in human tissues and if adaptive responses are present in human tissues, they have already contributed to the dose response observed. Thus, from this point of view, the combined effects of these phenomena still demonstrate an apparent linear response. Mark Little, a radiation epidemiologist at the US National Cancer Institute, recently concluded in a study that there is no evidence so far that the LNT will likely show any change in the incidence for cancer risk as a result of non-targeted effects (Little, 2010).

The bystander effect

In 1992, radiation biologists John Little and Hatsumi Nagasawa, then both at the Harvard School of Public Health, published a groundbreaking report, which suggested cells that had not been hit by radiation, but that grew in the same dish, could later show damage to their DNA (Nagasawa and Little, 1992). This became known as the “bystander” effect. At the time, the significance was neither clear nor well-understood and became quite controversial. The difficulty in demonstrating the bystander effect was in the precision: Only one cell could be hit with radiation in order to demonstrate that the surrounding un-hit cells would be irradiated. Years later, in 1999, a group at Columbia University, led by radiation biologist Tom Hei, conducted elegant experiments using alpha particles that could be controlled and individually delivered, demonstrating that indeed, if a cell was hit with a single alpha particle, more than one cell was damaged. Hei and his colleagues proved that this bystander effect applies to DNA mutations, chromosome damage such as chromosome aberrations, and neoplastic transformation endpoints, i.e., cancer (Wu et al., 1999; Zhou et al., 2000, 2001). In their extensive writings over the years, the Columbia team has demonstrated that bystander effects can be seen in the DNA of nearby cells by irradiating only the cytoplasm of a cell, thus showing that radiation damage does not have to begin with DNA damage (Zhou et al., 2001). Others have shown that, in some cases, this bystander effect is mediated by direct communication between hit and un-hit cells via gap junctions (direct connections between adjacent cells); however, there is also strong evidence in some experimental cell-culture studies that, since the hit cell is untouched by any other cell, cells must be secreting signal molecules into the intracellular medium (Little et al., 2002; Lyng et al., 2002).

Researchers, such as Vitali Moiseenko at the London Regional Cancer Center in Ontario, Canada, have shown possible bystander effects in tissues. Using a lung model, Moiseenko and his colleagues demonstrated that irradiating the lower lobe of the lung not only causes free radical damage in the radiation field, but also causes free radical damage in the upper, unexposed portion of the lung (Moiseenko et al., 2000). In separate studies, other researchers have demonstrated that waves of oxidative damage continue to occur in the irradiated and unexposed lung tissue long after the radiation exposure was given (Khan et al., 2003). This work suggests that oxidative stress-response genes—some of which are involved in inflammatory responses—continue to respond long after an initial insult. To put it simply: Radiation damage produces stress in the tissue, and the tissue responds by pathways, which produce reactive oxygen species that cause further damage. This process can go on for a considerable time after the radiation exposure is over.

Until the early part of this century, most of the work on bystander effects was done using charged particles, a practical choice since no machinery was available to irradiate single cells with x-rays. Originally, some scientists suggested the bystander effects might be unique to cells irradiated with these charged particles. This changed in the early 2000s, when technical advances allowed the development of x-ray microbeams capable of exposing individual cells; today, several centers around the world have these beams. One is the Gray Cancer Institute in England, where a team recently explored low-linear energy-transfer bystander studies, proving that x-rays—and, thus, other similar and very common radiation sources like gamma rays—can also produce the bystander effect (Prise et al., 2003). Until their study, many scientists believed this effect could only occur after alpha particles hit single cells; alpha particles, however, are not often the cause of human exposure to radiation.

It is now clear that bystander effects do occur and are a general phenomenon induced by all types of radiation. The development of studies in intact tissues, as opposed to cell culture, clarifies that bystander effects cause changes in cells in complete tissues, as well as in the surrounding tissue that was not hit by the radiation. ³ Scientists have noted that, for some types of tissue and types of radiation delivery, there is accumulating evidence that there is an increase in cancer induction—that is, above the linear response—in low doses.

If the bystander phenomenon is thought of as an effect in one cell that can be transmitted to surrounding cells in the tissue, it becomes clear that, since one cell touches or is near to many cells in three-dimensional space, the effect could easily be exponential until the signal fades away from the targeted cell. However, several reports have shown that bystander effects can alter probability of survival. For example, one study showed that when radiolabeling some cells in “nude” mice (genetically engineered) and reducing their “in vivo” cell survival, the cancer risk could be altered by removing potentially cancerous cells (Xue et al., 2002).

Genomic instability

Another significant discovery has been the observation of genomic instability after radiation. What this means is that cells with this phenomenon undergo progressive changes in genes that modify the cancer phenotype, gradually becoming a phenotype that can overcome tissue defenses, invade tissues, and spread to other organs and tissues in the body. ⁴ Several researchers are investigating this phenomenon, and some, like Hei et al. (2011) and William Morgan (2011), have shown that genomic instability can be induced in bystander cells by effects initiated in the hit cell. This type of effect can lead to an increase in initiated cells at low doses—and, therefore, might show an increased cancer incidence of an initial response.

Adaptive response

Possibly the most controversial action of radiation at low doses is the so-called “adaptive response.” Introduced in the early 1980s, this event was challenged for lack of experimental rigor until around 1987, when the late Sheldon Wolff published a seminal piece on the phenomenon. His paper highlighted that, if a priming dose—a small dose that has no discernible effect but that the cell can presumably detect and respond to—of about 5 millisieverts (mSv) was followed by a dose of 0.1 Sv, the effect on cells would be less than if the 0.1 Sv were given alone, without the priming dose (Shadley et al., 1987). ⁵ This study suggested that cells have a surveillance system that can detect very small doses of radiation, and, when activated, this surveillance system turns on the machinery that allows rapid response—such as DNA repair—when a larger dose of radiation comes along.

Wolff’s study inspired numerous reports on various attempts, both genomic and survival, that showed forms of this adaptive response (Feinendegen et al., 2007). The controversy surrounding the nature of this response still continues. Some studies have found that this is not the case for most human-exposure situations; thus, it is not clear if such responses are common or even occur in human tissue when it is exposed. If adaptive response mechanisms are present in human tissues, one would expect they would reduce the effect of radiation exposure; however, there is disagreement on whether this phenomenon only affects the survival of a cell—or both the survival and other endpoints, such as those that will produce a cancer.

This is an important distinction. Evolution probably dictates that a body or tissue will fight for the immediate survival of cells, as this gives the organism an advantage. It seems unlikely that an adaptive response would reduce or eliminate mutations in surviving cells, which could eventually cause a cancer to develop. So it remains to be seen whether adaptive responses will decrease or increase cancer incidence; more research is needed in this area.

Conclusion

Beyond the demonstration that these interesting biological processes occur and that they have cellular and molecular basis for action, there is now a significant effort to understand the mechanisms of these events and how they may contribute to cancer incidence—and, thereby, risk. And, as part of this effort, inflammatory response, apoptosis pathways, and cellular-repair pathways must be reviewed in more detail after low doses of radiation.

While the effects of bystander, genomic, and adaptive response pathways may already be included in standard models of radiation risk—because they have occurred in tissues during or after radiation exposure—there is still a strong likelihood that understanding specific risk for a given dose rate and low-dose exposure may have to rely on the fine-tuning effects of these phenomena.

More research is needed and more investigation into the relationship between bystander effects, genomic instability, and adaptive response is paramount before low-dose cancer risks are clearly delineated or even modeled based on these biological responses. A thorough mechanistic and quantitative understanding of these biological processes in the human context would offer the best chance of improving risk models for low-dose radiation exposures. Until then, quantifying the extent to which low-dose response might deviate from the LNT, if at all, will be difficult.