6th International Kidney Cancer Association Symposium,6—7 May, 2011, Warsaw

Advances in understanding the genetic basis of kidney cancer

Professor Tim Eisen (Cambridge Cancer Centre, UK)

For many years kidney cancer was a bit of a Cinderella subject in that there were treatments which occasionally worked but pretty toxic but most patients just had the toxicities and no benefit. But in about 2002 a series of studies started with antiangiogenic drugs, the first two out were sunitinib and sorafenib which in about three-quarters to four-fifths of patients caused the disease to stabilise or shrink. This is an order of magnitude better than what we had been used to before and for that reason the whole field has been revolutionised so that now there are six treatments which are in various stages of use for patients with advanced kidney cancer.

At the same time as this has been going on, and indeed for the decade or two decades before that, the surgeons had made huge progress in terms of their surgical techniques so that many patients now don’t have nearly the same loss of renal function that they would have done twenty or thirty years ago where a whole kidney was removed, whereas nowadays only part of a kidney, the diseased part of a kidney with a normal rim around it, would be removed, thus preserving renal function and making later complications like strokes and cardiovascular disease much less common.

What progress has been made in the advanced disease setting?

So the surgeons have made a lot of progress, their surgical techniques now allow them sometimes to operate laparoscopically i.e. through a keyhole, a rather large keyhole but still a keyhole or, and sometimes and, to remove only part of the kidney, the part which is diseased with a rim of normal tissue around it. That has very long term benefits in that it reduces the complication rate with particularly cardiovascular disease being a problem in people who have reduced renal function.

Although we’ve made a lot of advances in surgery and in advanced disease, the traditional oncological paradigm, if you like, is to bring treatments which we know work in the advanced disease setting forward to the adjuvant disease setting or the adjuvant setting where we’re actually trying to improve the cure rate. We don’t have any proven treatment at the moment in the adjuvant setting although there are large studies which are investigating this. So there has been a study in the United States with around 1,900 patients in it called the ASSURE study which has compared one year of placebo with one year of sunitinib and one year of placebo with one year of sorafenib. We’ll all be very interested in whether sunitinib or sorafenib does better in that situation, although in fact the study is not powered to look specifically at that question. Indeed, one thing which will hamper the analysis of that study is that the adjuvant treatments seem to be much more difficult for the patient to tolerate than the same treatment in the advanced disease setting. We’re used to that in breast cancer and colorectal cancer but not to the same extent that we have seen this in the use of antiangiogenic treatment in the adjuvant setting and we don’t know why that is. So studies which are continuing now, such as the SORCE study of which I’m the Chief Investigator, have introduced a low dose period before going up to the full target dose in this patient group. And also we’ve changed the emphasis away partly from achieving the absolutely highest target dose to keeping the patient on a tolerable treatment, because these are treatments which we may want to continue for up to three years. So that means people have a life to get on with and they must be able to get on with that life. It means that the treatment has to be just part of their life, not the governing feature of it.

And you’re working on the molecular differences in kidney cancer?

Ironically, for a disease where until the last nine years or so we’ve really had not much to offer, we have known quite a lot about the molecular biology of the disease and we’ve known most about the commonest kind of renal cell carcinoma which is clear cell carcinoma of the kidney. And we’ve known for many years now that the Von Hippel-Lindau gene is abnormal, either because the gene is silenced by methylation or mutated and therefore the Von Hippel-Lindau gene product, which is a very large protein, doesn’t work. More recently we’ve understood that the consequence of this gene not working is that you have deregulated expression of a number of targets, such as vascular endothelial growth factor and platelet derived growth factor, which leads to  a very high angiogenic impulse and that’s what we see in kidney cancer. But that alone would not be enough to cause a kidney cancer, we think. We think there are other targets, downstream targets of Von Hippel-Lindau which do alter the cellular behaviour and cause a tumour. Interestingly, and this is really hot news because it has only been published in the last few months, we’ve known for some years that if you knock out the Von Hippel-Lindau gene in mice you get a sick mouse but you don’t get kidney cancer. And we don’t know why that is but the exciting discovery that people have made, particularly Bin Teh and Andy Futreal at the Sanger Centre, have made in the last few months is that there is a whole series of other genes all of which seem to affect the activity of the chromatin structure, so these are chromatin remodelling genes which are also abnormal. It may be that the Von Hippel-Lindau abnormality causes unopposed up-regulation of a number of signals and it’s a tried and tested pathway that often that will lead to senescence, so this is the case in melanoma and in other disease types too. And one way of getting round that problem from the tumour’s point of view is to deregulate the control of the chromatin structure and that may be why these chromatin remodelling genes and the Von Hippel-Lindau gene are working together to cause kidney cancer.

What does this mean for future drug development?

I think this could mean a number of things for future drug development. I think it could have important prognostic impacts so that if we understood the molecular biology of an individual’s tumour we might have a much clearer idea of what the future holds for that individual, which would be very useful for the individual but would also be useful therapeutically to know whether we needed to jump in with treatment or not.

It’s difficult to predict, but I think that the most important possible outcome of this sort of work is that it might be able to allow us to predict two things. First of all, whether we should think that a tyrosine kinase inhibitor such as sunitinib or pazopanib would be effective against a particular tumour genotype, or whether we would be better trying a mammalian target of rapamycin mTOR inhibitor instead, like everolimus. That is the basic thing that this might help us predict.

More specific predictions, such as how well would the patient respond to treatment, and particularly which agent should we use, I think look far less likely. And the reason I think they’re far less likely is that there are no convincing data at the moment that implicate any of the kinase targets, with the exception of the main angiogenic targets which we know all of these drugs hit. So it seems unlikely to me, with the knowledge that we have today, that we’ll be able to pick and choose between these agents if we understood the molecular biology of the tumour with these VHL and chromatin remodelling genes. But that’s a pure guess and I may be being pessimistic.

What’s the state of play for biomarkers in kidney cancer?

Although the VHL and chromatin remodelling genes are not biomarkers in themselves, we’ve had considerable difficulty with biomarkers such as VEGF expression because VEGF is in all our platelets and platelets degranulate and there is a lot of VEGF; VEGF goes up and down through the menstrual cycle, VEGF changes in response to all sorts of insults and injuries and other stimuli. So that has not proved very productive and I’m not sure how many more legs there are in that sort of endeavour.

I think more interesting might potentially be the concept of circulating tumour DNA, which is an approach that we in Cambridge and others are taking, whereby you use very novel technologies to detect with astounding sensitivity and specificity DNA which has arisen from a tumour. Now clear cell kidney cancer is actually a very good candidate for that because the large majority of patients’ tumours will have an abnormality in VHL or one of these other genes, the large majority will. In a way it doesn’t matter too much what the abnormality is, as long as you can identify it specifically and then you look specifically for that abnormality in tiny, tiny quantities in the peripheral blood. Now that could be a biomarker, I don’t think you could put it any stronger than that and there are lots of ifs and buts in what I’m saying. But this does give a tumour specific abnormality to look for and when you combine that with the sensitivity and specificity of the latest technological advances, I think we are in a better position to investigate this than we’ve ever been in before.

This is DNA, and that’s an important point. So this is just circulating tumour DNA, it does not depend on an intact tumour cell which is an important distinction because there have been lots of investigations into circulating tumour cells, and they are an order of magnitude more difficult to conduct than, for example, free DNA originating from a tumour, which we know we can detect in a large proportion of patients.

Is the circulating DNA from the primary tumour?

We think so. I think that we can’t definitively say whether the circulating tumour DNA has arisen from a primary or a secondary or whatever. One thing you could do, though, and I’m very interested in adjuvant therapy because all of these exciting new drugs that we have do not improve the cure rate, they improve the control rate for an undefined period of time, but a period of time. Then the disease inevitably progresses and we lose control and the patient ultimately dies. The prospect of improving the cure rate would be, for me, is the holy grail and I think that one of the problems of doing adjuvant studies, particularly in a relatively uncommon tumour type like renal cell carcinoma, is that in a working lifetime you can probably only do one or two adjuvant studies because we have no surrogate for failure of treatment. So I think of circulating tumour DNA as a possible biomarker that we might use to look for the advent of relapsing tumour, a molecular signal for a relapsing tumour, which would indicate failure of adjuvant therapy. So that might accelerate the development of adjuvant treatments, and I think that possibly is one of the most exciting ways of investigating this.

We simply don’t know, of course it would be great to have a biomarker for the activity of advanced disease, but I think there are complexities there. So, for example, I’m not sure we know ideally how to use our existing agents. Many of the agents that we have will never cause any significant tumour shrinkage, they will stabilise the tumour. But if I had a tumour, I would prefer to have a slow growing tumour than one which responds very quickly and then grows back very quickly. So the field is pretty well virgin territory but I do think that because of the technological advances we are now in a position to start investigating some of these questions and there is considerable patient benefit that could arise from this.