I’ve been asked to comment on the recent Costes/Bissel paper (see http://www.pnas.org/content/early/2011/12/16/1117849108.full.pdf+html)
which concluded that “extrapolating risk linearly from high dose as done with LNT could lead to overestimation of cancer risk at low doses.”
This in vitro study examines patterns of radiation-induced foci (RIF) in a human immortalised cell-line. Within a few seconds/minutes of being exposed to ionising radiation, DNA damage-sensing proteins (eg the p53 binding protein) apparently localize at DNA double-strand breaks (DSBs) resulting in the formation of stainable areas in cells: these are called radiation-induced foci (RIF). RIF are used to assess DNA damage and DNA repair kinetics. However, uncertainty exists on the precise relationship between DSBs and RIF – in particular on whether there’s a linear relationship between them.
The study’s main finding is of a lower RIF production rate per Gy at high doses compared to low doses: 15 RIF?Gy after 2 Gy exposure compared to ~ 64 RIF?Gy after 0.1 Gy. The authors state
“We observe that as …DSBs increase in a cell, the number of RIF does not increase proportionally and the kinetics of RIF formation/disappearance is altered; RIF appear faster but remain longer in the cells as dose levels increase. These nonlinear processes cast considerable doubts on the general assumption that risk to IR is proportional to dose and could be interpreted as the consequence of DNA repair centers in human cells.”
Well, yes, it could be so interpreted …..but it might not necessarily be the case.
The devil is in the detail I’m afraid, but I shall try to keep it simple.
The crux is whether their conclusions come from their experimental data or whether they come from the authors’ interpretation of the data. See the example in the previous para.
In another example, they state “As a result of this [DSB] clustering, a single RIF may reflect a center where multiple double strand breaks are rejoined. Such multiple repair activity increases the risks of broken DNA strands being incorrectly rejoined and that can lead to cancer.”
My reading is that this conclusion is dependent on their interpretation, rather than on any data actually showing this. Yes, a single RIF may reflect a centre where many DSBs are repaired, but equally it may not. And more DSB breaks could be misrepaired at clustered repair centers, but that is just their hypothesis and it’s not shown by any data here. DSB breaks might also be linearly repaired at cluster centres – how do we know?
They say this study points to non-linear responses at low doses, but there is a problem here. The study examines doses between 0.1 and 2 Gy and as Keith Baverstock says the authors have to contend with the fact that, over this range, the dose response for cancer is observed to be linear or as near as we can see. It’s true that below 0.1 Gy it’s hard to be sure that the dose response relationship is linear, but above 0.1 Gy – a lot of evidence shows it’s linear.
Another problem is that, as they state “although their data bring additional evidence of the existence of DNA repair centers in human cells, the mechanisms by which such clustering take place remain entirely unknown at this point…” This means that their jumping to conclusions about non-linear responses is inappropriate.
My main caveat about this study is that it repeatedly makes strong comments about the dose-response relationship which are unjustified. For example, they twice say “These nonlinear processes cast considerable doubts on the general assumption that risk to IR is proportional to dose…”. This is a single in vitro study, on one cell-line: for the authors to make such a strong conclusion is ill-advised.