The background of our study is that patients with never-smoking lung cancer often ask, ‘How did I get this? I lead a healthier lifestyle, I eat normally, I exercise, how did I get lung cancer?’ And the truth of the matter is that we don’t know. There are several mechanisms that have been postulated, including air pollution as one of them, associated with lung cancer in never-smokers, as well as radon exposure and germline familial genetics that may predispose you to lung cancer, all of which have plausible associations.

In the context of never-smoking lung cancer and air pollution, however, the underlying mechanistic basis was not known and we were interested in studying whether air pollution actually directly causes lung cancer and trying to unravel that conundrum.

What was the methodology behind the study?

We took three different approaches: the first approach was to look to see whether there was an association globally and within England, Taiwan and Korea between rising levels of PM2.5 and EGFR mutant lung cancer, a subtype of lung cancer that commonly occurs, probably five times more common, in never-smokers. We found a correlation of rising PM2.5 levels and EGFR mutant lung cancer in Public Health England data, in Taiwan and in South Korea.

So the next step was to say, well, is it causative? Does air pollution cause lung cancer? We took three different mouse models, each one you can induce a mutation in EGFR or KRAS in normal lung tissue. When we do that and then expose mice to pollution, to PM2.5s we see an increase in the number, size or grade of tumours in those mouse models, giving us some compelling evidence that air pollution is actually directly influencing the growth of tumours and causing tumour progression and probably initiation. 

Then the last step was to ask, well, how does this happen if air pollution is not causing mutations? Because we know from studying lung cancer in never-smokers it has got a very low mutational burden so there is no environmental carcinogenic signature. So any causative mechanism would have to be independent of causing DNA mutations. What we find is that exposure to air pollution in mice and humans results in release of interleukin-1 beta, probably from macrophages in epithelial cells that causes, we think, transdifferentiation of a rare population of epithelial cells into a progenitor stem-like state. Now, on its own those cells are probably just part of a natural wound healing process, inflammatory process, but if that cell happens to harbour an oncogenic mutation such as EGFR or KRAS then it can form a tumour. 

The last step, consistent with the Berenblum model of tumour promotion that I mentioned in my talk, there would have to be mutations present in normal tissue for that to happen. So we biopsied normal tissue in a study called PEACE, which is an autopsy programme running from UCL and Dr Mariam Jamal-Hanjani. What we find is that we can identify oncogenic mutations in over 50% of biopsies in EGFR or KRAS. So these mutations are existing probably in a healthy adult like me, I hope, and the vast majority won’t ever go on to cause cancer at a population level. The difficulty is if that mutation happens to be in the wrong cell at the wrong time when exposed to pollution then that cell can transdifferentiate into a stem cell or a cell with stem-like properties and form a tumour. That’s the model we think is happening here.

Were there any other key results you were interested in?

What we’re interested in now is really understanding how widely applicable this is to other environmental carcinogens that aren’t causing DNA mutations and also answering a key question – could tobacco be causing cancer, or contributing to cancer, through promotion as well as mutations? Because both the classical model of tumour initiation through mutations and this old model proposed by Isaac Berenblum in 1947 based on tumour promotion may be equally applicable in different circumstances or, indeed, one environmental carcinogen might initiate and promote cancer. There’s no reason to believe otherwise, that tobacco would do any different. Tobacco smoke is clearly a pollutant of sorts and one might expect it to have the same sort of promoter effect. If that’s the case that raises questions about the safety of vaping and other such things in the context of tobacco cessation because I would be worried that perhaps vaping might promote cancers too if we could prove that tobacco did as well.

How do you think these results will impact future treatments?

I don’t think right now these results will directly impact today the treatment or prevention but I think it opens up an opportunity to start thinking about molecular cancer prevention. If these environmental carcinogens are triggering this inflammatory response that’s enabling that first nascent cell that has a mutation to expand, then perhaps we could intervene and block that process. Because I don’t think we can stop the mutational process because that’s a natural hallmark of aging and we can’t do anything about aging right now, but maybe we could intervene in that inflammatory axis, much as they did in the CANTOS trial with canakinumab, to block one or two of those key inflammatory mediators to prevent that first step in tumour initiation.