2011 SABCS, San Antonio Breast Cancer Symposium, 6-10 December, San Antonio, USA

Energy supply in cancer

Dr Michael Pollak – McGill University, Montreal, Canada

The interesting thing that we’re studying is the relationship between energy balance and breast cancer. Although the estimates vary between studies, we know that obese women are more likely to die of cancer and one of the largest studies actually suggested that the factor is one of doubling. In other words, an obese woman with breast cancer may have a doubling of the risk of dying of breast cancer. So we all know that being obese is not good for your health in general – heart attacks, strokes, but these data suggest that actually cancer death is also affected by being too heavy. And the alarming aspect is that the prevalence of obesity around the world is greatly increasing, not only in the Western countries but also in the developing world. As soon as one exits poverty, severe poverty, one of the first things that happens to hungry populations is that they tend to overeat. So in countries like China, Southeast Asia, India, rapidly rising obesity.

So if we have, let’s say, a threefold increase in the proportion of women who are obese, and the obese women have a doubling of cancer death risk, we actually have a problem that could attenuate the progress that we’re seeing in breast cancer control. So that’s really the population science scene that interested us.

Another aspect is why? Why should a fat person be more likely to die of cancer? It’s an interesting question. We knew from prior studies that mice, if they’re starved, actually get much less cancer so caloric deficit is something that is protective but one didn’t realise that it operates throughout the spectrum of nutrition. So it’s not just that eating too little is protective, eating too much increases the risk. So what could be the players here? Why would that be? Our research suggests that it’s not really a question of how much you eat, per se, it’s about how much you eat relative to how active you are. So it’s energy balance, and what seems to be unhealthy is just eating more than you need to. So if you have a very physically demanding profession and you eat a lot, but you spend a lot of energy and don’t get fat, that’s fine. It’s if you have a sedentary life and you eat a lot, you don’t spend the energy, then you get fat and then you have this cancer problem.

Originally it was thought that maybe just eating too much gave the cancer cells too much energy and so having an excess of energy they just divided too much and that would explain it. But it’s not because the cancer gets more energy, because after you eat a lot you still have blood sugar before you eat, you still have blood sugar and lipids after you eat, so the actual change in the blood concentration of the nutrients is not enormous. Similarly, when you eat less you may feel hungry but your blood sugar does not fall. So we have a situation where varying how much you eat will really have an impact on the cancer situation but not apparently by affecting the amount of energy available to the cancers through the blood. So how could it work? The research that we have carried out suggests that it all works through hormones. Briefly, when you eat too much you alter the hormonal balance in your body in a way that raises the levels of the hormones that encourage growth. And that’s how the unhealthy diet, over-eating, impacts on the cancer behaviour, not actually by giving more energy to the cancer but rather by a hormonal signal to grow. Similarly, when you have a protective effect of eating too little, the cancer still has enough energy in the blood but there’s a drop in the growth stimulating hormones and so that’s why it’s protective. So it’s a kind of indirect hormonal mechanism by which the variation in how much you eat affects the cancer behaviour.

Why is it important? Once we understand, from these investigations, these clues, these investigations of how diet impacts cancer behaviour, we recognise that certain hormone levels are mediated in the effect. So, thereby, this investigation identifies new targets for drug therapies because we recognise, for example, if one of the hormonal mediators of the over-eating effect is IL-6, another might be insulin, there’s many examples of hormones whose level varies with how much you eat. So now that we understand that it’s those hormones that are driving the effect of diet on cancer prognosis, we can make drugs that affect those mediators. And that has been done. So now there are drugs that interfere with insulin signalling on cancer cells, that interfere with certain other hormones that are involved in the diet cancer relationship and some of these are showing great promise. So it was the identification of molecular targets as a result of trying to understand the mechanisms of diet effects.

Absolutely common sense, nothing that I’m saying is suggesting that we should try to use drugs where simple diet works. There are a couple of points, though. Sadly, it’s very hard to influence dietary behaviour and we would love to do that more effectively but we certainly try. But, more importantly, while the dietary clues identify the molecular targets, we can actually address those targets more effectively by drugs than by varying the diets. So if there’s a tumour that’s dependent on insulin, for example, we can much more effectively and completely block the insulin signalling with a drug than we could by telling the person just get very hungry. Or if it’s interleukin-6 that we want to lower, because that’s another diet related factor, we can block the interleukin-6 more effectively with a drug than even with a severe and probably impractical diet. So the idea is that the tumours that are influenced by the hormones can be addressed, to a limited extent, just by altering the diet but we can push that process and exploit that dependency further if we actually drug the hormone rather than just influence its level by dieting. To try to give you a more quantitative feel, if we were to notice that, let’s say, an insulin level twice normal in someone who is eating too much is stimulatory, we could use lifestyle interventions to return the insulin level, let’s say, from twice normal to 1.2 times normal. But we could abolish the signalling, or reduce the signalling to 0.25 by the use of a drug. So, again, the story here is that the investigation of the mechanisms of the dietary effect identifies the molecular targets but then once we know the molecular targets we can actually block them more effectively with a drug than even with the most severe diet.

We stumbled upon this because we were investigating the mechanisms of obesity on influencing cancer behaviour. But once we recognised that… let’s take the example of interleukin-6, interleukin-6 tends to be elevated in obese people and there’s a subset of cancers that behave more aggressively in obese people, probably because of their high interleukin-6 levels. But there are some people who have high interleukin-6 levels for genetic reasons and they’re not obese. So that in the end, some of these molecular targets that are identified through this way of investigating the obesity mechanisms may be relevant to many different kinds of cancers and even to thin people because, to stick with this example, while we recognise IL-6 may be part of the mediation or part of the mechanism by which obesity influences cancers, there are some people who have high interleukin-6 levels and are not obese, but these therapies would still be applicable for them.

One other aspect that really has captured some attention recently is the anti-diabetic drug, metformin. This is a very commonly used drug for the treatment of type 2 diabetes and some epidemiologic studies originally in Scotland suggested that, unexpectedly, the number of cancers in users of metformin was far lower than expected. So a diabetology group working in Scotland actually were just trying to understand why and how the population of diabetics were dying and so they undertook a survey, they took notes of why they died, what treatments they were on and, in a simple way, analysed their data. And they noticed that those diabetics who were on the drug metformin appeared to be dying of cancer at about one half the expected rate. It’s very unusual, usually when you do post-marketing surveillance of a drug you find a bad side effect that was unanticipated, so here we’re finding a drug that’s in very common use, very inexpensive, generic, and an apparent hint that there is an unexpected protective effect against cancer. Lots of scepticism for that result, it was a retrospective study; formerly it would be regarded as hypothesis generating, not the last word, but a number of groups around the world with great scepticism tried to debunk it and, in fact, they ended up supporting it. So that this increased the interest, although we have to be very careful because all of the studies were retrospective, no-one started a controlled clinical trial because it would be impractical really because remember the metformin is indicated for diabetes treatment.

So our laboratory and a few other laboratories thought that we could really perhaps look at this a little bit more rigorously by studying it under laboratory conditions, again with great scepticism, expecting that the metformin would do absolutely nothing to cancer cells. But again, to our surprise, it actually really did growth inhibit the cancer cells in a laboratory situation, so now the interest level really increased greatly around the research communities. A mechanism was identified by which the metformin acts as a growth inhibitor, it has to do with activation of a control system called AMP kinase. Really what the metformin does, fundamentally, is it affects the mitochondria of the cells so that they make less ATP, less usable energy, and when that happens the cell’s control system says, “Wait, I’m running out of energy. I’d better stop growing, I’d better stop using energy.” And as a consequence of that, you actually get a change in, for example, mTOR and other control systems of the cell where the cell is reacting to the metformin induced reduction in energy production by reducing energy consumption. That leads to a growth inhibitory action.

So with that rationale and that laboratory insight, together with the population studies that I mentioned, there has been increasing interest in the field and now a number of clinical trials are starting to investigate whether metformin really will or will not turn out to be a useful drug. It’s relevant in this context because it does affect the energy supply to the cells. The distinction to the points I made earlier where some dietary manipulations affect the hormones, the metformin actually can affect the usable energy that the cells really have through this mechanism on the mitochondria. It’s a complicated field, it’s an early field, it’s a very interesting field; one of the peculiar aspects is that since the drug is generic there’s no central co-ordination of this research, no-one owns the intellectual property, every drug store sells a lot of metformin, so many physicians and organisations of physicians can informally run a clinical trial on their own. There are no headquarters where all these trials are being tabulated in the way that it would be if it was a new compound with patent protection. So we have to have a lot of caution here because some of the early evidence, I would even say all of the early evidence, was retrospective. The laboratory evidence is interesting but, to play the devil’s advocate, you know it’s perhaps that the drug concentrations used in the laboratory are not achievable in the clinic. And now we have a clinical trials process that is a little bit less formal than would be the case if it was being handled by the private sector.

So there are provisos, there are lots of gaps in knowledge, but it’s another area of research related to this question of energy supply in cancer. And it’s dirt cheap, when we buy it for laboratory studies it’s $10-$20 a kilogram. If it would ever work it would really be a game-changer, it would be very disruptive to have an effective cancer drug because all the new cancer drugs that we hear about, is it $2,000 a month or is it $4,000 a month? This would be $20 or $30 a month. So it is a little bit of a peculiar story, in many ways, but one that should be on your radar screen because while it may fall apart with more rigorous investigation, it certainly has survived the initial sceptics who thought that within six months of the initial enthusiasm that the project or the concept would be completely disproven. It’s shown itself to be a more robust idea than that, but we really will need formal clinical trials and we certainly are not advocating people using it now.