Through the Smokescreen

For many years I was lucky enough to teach in a cancer biology course for third year natural science and medical students. Quite a few of those guys would already be eyeing up research careers and, within just a few months, some might be working on the very topics that came up in lectures. Nothing went down better, therefore, than talking about a nifty new method that had given easy-to-grasp results clearly of direct relevance to cancer.

Three cheers then for Mikhail Denissenko and friends who in 1996 published the first absolutely unequivocal evidence that a chemical in cigarette smoke could directly damage a bit of DNA that provides a major protection against cancer. The compound bound directly to several guanines in the DNA sequence that encodes P53 – the protein often called ‘the guardian of the genome’ – causing mutations. A pity poor old Fritz Lickint wasn’t around for a celebratory drink – it was he, back in the 1930s, that first spotted the link between smoking and lung cancer.

This was absolutely brilliant for showing how proteins switched on genes – and how that switch could be perturbed by mutations – because, just a couple of years earlier, Yunje Cho’s group at the Memorial Sloan-Kettering Cancer Center in New York had made crystals of P53 stuck to DNA and used X-rays to reveal the structure. This showed that six sites (amino acids) in the centre of the P53 protein poked like fingers into the groove of double-stranded DNA.

Central core of P53 (grey ribbon) binding to the groove in double-stranded DNA (blue). The six amino acids (residues) most commonly mutated in p53 are shown in yellow (from Cho et al., 1994).

So that was how P53 ‘talked’ to DNA to control the expression of specific genes. What could be better then, in a talk on how DNA damage can lead to cancer, than the story of a specific chemical doing nasty things to a gene that encodes perhaps the most revered of anti-cancer proteins?

The only thing baffling the students must have been the tobacco companies insisting, as they continued to do for years, that smoking was good for you.

And twenty-something years on …?

Well, it’s taken a couple of revolutions (scientific, of course!) but in that time we’ve advanced to being able to sequence genomes at a fantastic speed for next to nothing in terms of cost. In that period too more and more data have accumulated showing the pervasive influence of the weed. In particular that not only does it cause cancer in tissues directly exposed to cigarette smoke (lung, oesophagus, larynx, mouth and throat) but it also promotes cancers in places that never see inhaled smoke: kidney, bladder, liver, pancreas, stomach, cervix, colon, rectum and white blood cells (acute myeloid leukemia). However, up until now we’ve had very little idea of what, if anything, these effects have in common in terms of molecular damage.

Applying the power of modern sequencing, Ludmil Alexandrov of the Los Alamos National Lab, along with the Wellcome Trust Sanger Institute’s Michael Stratton and their colleagues have pieced together whole-genome sequences and exome sequences (those are just the DNA that encode proteins – about 1% of the total) of over 5,000 tumours. These covered 17 smoking-associated forms of cancer and permitted comparison of tobacco smokers with never-smokers.

Let’s hear it for consistent science!

The most obvious question then is do the latest results confirm the efforts of Denissenko & Co., now some 20 years old? The latest work found that smoking could increase the mutation load in the form of multiple, distinct ‘mutational signatures’, each contributing to different extents in different cancers. And indeed in lung and larynx tumours they found the guanine-to-thymine base-pair change that Denissenko et al had observed as the result of a specific chemical attaching to DNA.

For lung cancer they concluded that, all told, about 150 mutations accumulate in a given lung cell as a result of smoking a pack of cigarettes a day for a year.

Turning to tissues that are not directly exposed to smoke, things are a bit less clear. In liver and kidney cancers smokers have a bigger load of mutations than non-smokers (as in the lung). However, and somewhat surprisingly, in other smoking-associated cancer types there were no clear differences. And even odder, there was no difference in the methylation of DNA between smokers and non-smokers – that’s the chemical tags that can be added to DNA to tune the process of transforming the genetic code into proteins. Which was strange because we know that such ‘epigenetic’ changes can occur in response to external factors, e.g., diet.

What’s going on?

Not clear beyond the clear fact that tissues directly exposed to smoke accumulate cancer-driving mutations – and the longer the exposure the bigger the burden. For tissues that don’t see smoke its effect must be indirect. A possible way for this to happen would be for smoke to cause mild inflammation that in turn causes chemical signals to be released into the circulation that in turn affect how efficiently cells repair damage to their DNA.

Sir Walt showing off on his return                         to England

Whose fault it is anyway?

So tobacco-promoted cancers still retain some of their molecular mystery as well as presenting an appalling and globally growing problem. These days a popular pastime is to find someone else to blame for anything and everything – and in the case of smoking we all know who the front-runner is. But although Sir Walter Raleigh brought tobacco to Europe (in 1578), it had clearly been in use by American natives long before he turned up and, going in the opposite direction (à la Marco Polo), the Chinese had been at it since at least the early 1500s. To its credit, China had an anti-smoking movement by 1639, during the Ming Dynasty. One of their Emperors decreed that tobacco addicts be executed and the Qing Emperor Kangxi went a step further by beheading anyone who even possessed tobacco.

And paying the price

If you’re thinking maybe we should get a touch more Draconian in our anti-smoking measures, it’s worth pointing out that the Chinese model hasn’t worked out too well so far. China’s currently heading for three million cancer deaths annually. About 400,000 of these are from lung cancer and the smoking trends mean this figure will be 700,000 annual deaths by 2020. The global cancer map is a great way to keep up with the stats of both lung cancer and the rest – though it’s not for those of a nervous disposition!

References

Denissenko, M.F. et al. ( (1996). Preferential Formation of Benzo[a]pyrene Adducts at Lung Cancer Mutational Hotspots in P53.Science 274, 430–432.

Cho, Y. et al. (1994). Crystal Structure of a p53 Tumor Suppressor-DNA Complex: Understanding Tumorigenic Mutations. Science, 265, 346-355.

Alexandrov, L.D. et al. (2016). Mutational signatures associated with tobacco smoking in human cancer. Science 354, 618-622.

Obesity and Cancer

Science, you could say, comes in two sorts. There’s the stuff we more or less understand – and there’s the rest. We’re pretty secure with the earth being round and orbiting the sun, the heart being a pump connected to a network of tubes that keeps us alive, DNA carrying the genetic code – and a few other things. But human beings are curious souls and we tend to be fascinated by what we don’t know and can’t see – why the Dance of the Seven Veils caught on, I guess.

Scientists are, of course, the extreme example – they spend their lives pursuing the unknown (and, as Fred Hoyle gloomily remarked, they’re always wrong and yet they always go on). But in this media era they pay a public price for their doggedness because they get asked the pressing questions of the moment. Is global warning going to finish us off soon, why is British sport generally so poor and – today’s teaser – does being fat make you more likely to get cancer?

A few facts go a long way

The major cancers have become familiar because the numbers afflicted are so staggering – but the one good thing is that the epidemiology can tell us something about the disease. Thus for cancers of the bowel, endometrium, kidney, oesophagus and pancreas and also for postmenopausal breast cancer there is clear evidence that being overweight or obese makes you more susceptible. In other words, if you compare large groups with those cancers to equally large numbers without, the disease groups contain significantly more people who are fat. We should add that the above list is conservative. A number of other cancers are almost certainly more common in those who are overweight (brain, thyroid, liver, ovary, prostate and stomach tumours as well as multiple myeloma, leukaemia, non-Hodgkin lymphoma and malignant melanoma in men).

Sizing up the problem

The usual measure is Body Mass Index (BMI) – your weight (in kilograms) divided by the square of your height (in metres). A BMI of 25 to 29.9 and you’re overweight; over 30 is obese. In England in 2009 just over 61% of adults and 28% of children (aged 2-10) were overweight or obese and of these, 23% of adults and 14% of children were obese. And every year these figures get bigger.

How big is the risk?

Impossible to say exactly – for one thing we don’t know how long you need to be exposed to the risk (i.e. being overweight) for cancer to develop but in 2010 just over 5% of the total of new cancer cases in the UK was due to excess weight. That’s another conservative estimate, but it means at least 17,000 out of 309,000 cases, with bowel and breast cancers being the major sites.

What’s going on?

Showing an association is a good start but the important thing is to find out which molecules make that link. For obesity and cancer detail remains obscure but broad outlines are emerging, summarised in the sketch. In obesity fat (adipose) cells increase in both number and size (so it’s a double problem: more cells – and the fat cells themselves are fatter). As this happens other cells are recruited to adipose tissue and, from this cellular cooperative, signalling proteins are released that have the potential to drive tumours. This picture is similar to that of the microenvironment of tumours themselves, where many types of cell infiltrate the new growth. Initially this inflammatory and immune response aims to kill the tumour but if it fails the balance of signalling shifts so that it actually helps the tumour grow. In addition to signals from fat cells themselves, obesity is usually associated with increased levels of circulating growth hormones (e.g., insulin) and of lipids, both of which may also promote tumour development.

Thus many signals with cancerous potential arise in obese individuals. In principle these could initiate tumour growth or they could accelerate it in cancers that have started to develop independently of obesity. So it is complicated – but at least as new signalling strands emerge they offer new targets for drug therapy.

In obesity abnormal signals from fatty tissue can combine with others arising from perturbed metabolism to help cancers develop

Reference

World Cancer Research Fund (WCRF) Panel on Food, Nutrition, Physical Activity, and the Prevention of Cancer (WCRF, 2007).