The past decade has seen astonishing progress in our understanding of the genetic causes of cancers, and the rise of new ‘targeted’ drugs that directly exploit these genetic and molecular faults. Many of the new-style drugs are already showing considerable benefit to patients – yet sobering news is now emerging. Cancer, it turns out, is a lot more genetically complex than we originally believed. And this complexity is responsible for a major limitation in cancer therapy – drug resistance.
New technology is allowing scientists to scan the DNA code of tumours more rapidly and cheaply than ever before, looking for more common genetic faults as well as rarer gene changes. And instead of finding a small number of important cancer-causing genes in a given cancer, the picture that has now emerged is of many different possible combinations of genetic faults and abnormal biochemical switches that can combine together to drive cancer development. Worryingly, scientists are also discovering that many different cancer cell sub-types can be present within an individual patient’s tumour and may carry a range of cancer-causing faults, putting the concept of a “magic bullet” single cure-all cancer drug even further out of reach.
Landmark paper reveals complexity of cancers
In December 2010, The Institute of Cancer Research’s Professor of Cell Biology Mel Greaves published a landmark paper in the journal Nature showing that cancer stem cells – self-renewing cells that can drive the formation of new tumours – in the most common childhood leukaemia have “complex and diverse combinations of mutations, even within individual patients”. Professor Greaves found that cancer stem cells followed a process of Darwinian evolution, as ongoing genetic variation created new cell types, and natural selection allowed the strongest tumour cells to multiply.
In March this year, a study involving patients at the ICR’s partner hospital, The Royal Marsden, found substantial genetic variation between different parts of individual patient’s kidney tumours. Around two-thirds of genetic faults were not repeated between different biopsy samples of the same cancer. Then in May, an international collaboration led by the Wellcome Trust Sanger Institute and involving ICR scientists found 70 different combinations of mutations across 100 breast tumours. The team found 40 cancer-causing mutations in total – and revealed that while some tumours had just one, others had up to six.
This profound genetic variation helps explain why patients do not all respond in the same way to cancer treatments, and how resistance to treatment can develop. As genetic faults in tumours change with time and during treatment, targeted drugs that were previously effective may stop working.
More than 100 targeted cancer drugs have already been approved for patient use, and several hundred more are in development. The targeted drug imatinib, or Gleevec, has led to an 80 per cent drop in the number of people who die from chronic myeloid leukaemia, while vemurafenib, or Zelboraf, which the ICR and The Royal Marsden helped develop, has extended life for patients with malignant melanoma. Some targeted drugs have shown only moderate activity on their own, but significant benefit when combined with traditional cytotoxic chemotherapy.
Yet resistance, says Professor Paul Workman, deputy chief executive of the ICR, who was also this week awarded the 2012 Royal Society of Chemistry Entrepreneur of the Year Award, is “the biggest challenge for cancer research, as it affects virtually all treatments – including both cytotoxic drugs and the generation of new molecularly targeted agents”.
New technology brings hope of a solution
But now in an extensive article in Nature Biotechnology,1 Professor Workman and colleagues Dr Bissan Al-Lazikani and Dr Udai Banerji have set out a strategy for solving this problem. Through the use of new technology, including large-scale drug screening experiments and computer modelling – linked to the continuing efforts to discover and develop molecularly targeted drugs for all cancer genes and pathways – they provide a road map for finding “intelligent” treatment combinations that can overcome acquired resistance and ensure patients keep benefiting from targeted drugs.
“It has now become really clear that drug combinations currently provide the best route – and quite possibly the only way forward – to overcome the enormous problem of genetic complexity and heterogeneity of human cancer,” Professor Workman explains. “We face a major challenge, but we have powerful high-tech tools to tackle it.”
A rational approach to combining drugs
One of these tools is systematic high-throughput chemical screening, which is allowing scientists to assess thousands of drug combinations simultaneously. Clinical trials of new drug combinations are expensive – from $50 million to $100 million for a large randomised Phase III study – so it is important that they are only tested in patients if there is evidence that they are likely to be effective.
“Historically, drug combinations have come about when a pharmaceutical company makes both drug x and drug y and decides to try combining them – usually, there is no particular biological rationale for doing it. The more sophisticated approach now is to choose drugs so that they deliberately synergise with each other, and we can predict this using chemical screening – and also now using RNAi screening technology,” Professor Workman explains.
RNAi uses small sequences of genetic code that can bind to matching genes to stop them from carrying out their normal function. Each RNAi sequence matches a particular section of a given gene. Using this approach, scientists are identifying genes likely to be involved in resistance and can use this new technology to find out if drugs that block these genes should be used in combination with existing treatments.
First, they treat cancer cells with the primary drug. Then they take a library of RNAi sequences and screen all other genes in the genome looking for one that might change the cells’ response to the primary drug. Using a robot, each RNAi is added to a separate tiny well of cells, and cancer cell growth is measured. If the cancer cells become more sensitive to the original drug when a particular gene is “silenced” by RNAi then that gene can then be targeted, resulting in a two-drug combination that is potentially very effective and which would be investigated further. In some cases a drug might already be available to block the second gene. In other cases, a drug discovery project would need to be carried out.
A major advantage of the approach is that, so as to leave no stone unturned, the RNAi screening is “unbiased”. Rather than just testing genes that have experimental evidence of involvement in resistance, they can use a library of RNAis that block all “druggable genes” – that is, genes for which targeted drugs have already been or could be developed.
Using this technology, ICR scientists have already identified promising new drug combinations. Professor Alan Ashworth, the ICR’s chief executive, has been a pioneer of this approach.
Network maps of cancer resistance
In addition to using unbiased screening, another approach to tackling resistance is by first understanding how it is occurs, and then trying to prevent it from the outset. There are two main ways in which cancer cells can acquire resistance to new treatments – and for each, new technology has become available that can help scientists understand and ultimately overcome it.
The first way resistance can happen very quickly occurs at the biochemical level in cells. Normally, information that cells receive from their environment – including other cells – prompts a cascade of changes that ultimately gives the cell an instruction, such as dividing or changing shape. Faults in these biochemical signalling pathways can lead to uncontrolled cell growth and cancer, so one of the major focuses for drug discoverers is to create new drugs that target and block molecules involved in these signalling pathways. However, cancer cells can often find a new way to get their growth signal through, either by changing the original pathway or by finding another way around the block.
“When we hit one part of the biochemical network with a drug, the network responds. It’s like if one particular underground or metro line has a fault, the passengers can take a different line to get to their destination,” Professor Workman says. “Cancer cells can operate in this same way, allowing them to escape the effects of targeted drugs.”
One solution is to block all the pathways at once with a combination of drugs – but first scientists must make sure they have uncovered all the possible pathways. A promising approach to drawing this full ‘transport map’ is network biology. Scientists combine all of the world’s knowledge about cancer genes, resistance pathways and drugs in powerful computers, and then use mathematical modelling to try to work out how it all fits together. In this way they can predict drug combinations that might have the most powerful effects to overcome biochemical resistance mechanisms. The ICR has invested heavily in network biology research and in the high-performance computing that is essential to support it.
One class of drugs already discovered at the ICR – HSP90 molecular chaperone inhibitors– already have the ability to simultaneously block multiple lines in cancer cells’ transport maps. But even more sophisticated methods are needed and can be discovered using the modelling approach.
“Already, scientists have used mathematical modelling to predict all the genes that interact in the simple organism yeast” says article co-author and computational biologist Dr Bissan Al-Lazikani. “While mapping the pathways involved in cancer development is hugely more complex, the progress made so far indicates that it will be possible.”
Staying one step ahead of the cancer
On top of biochemical mechanisms, another way for cancer cells to acquire resistance is to exploit the process of Darwinian evolution and natural selection, where a random mutation creates a single cancer cell that can survive the selective pressure of a drug, which then multiplies to become the dominant cancer cell type – or clone – within the tumour.
Again, computer modelling can help. Dr Al-Lazikani explains: “By using an approach called evolutionary modelling, researchers can work backwards to identify the genetic make-up of the very first cancer cell that developed in a patient. By looking at all the patterns of mutations since that point, they can then identify different sub-populations of cells within the same cancer, predict drug resistance paths likely to develop within these individual populations in future, and head them off before they start to cause problems.”
“In the future, we want to be able to do experiments in the lab – and using powerful computers – that will predict what the resistance mechanism will be in a patient, and then be ready with the next treatment when this starts to happen,” Professor Workman says. “The aim is to always keep one step ahead of the cancer.”
Adapting treatments in real time
At the moment, gene screening techniques used in the clinic generally detect the dominant type of cancer cell. But with increasingly powerful “deep” sequencing, scientists are now able to scan the genome of multiple cancer clones at the same time, looking for all the different sub-populations with different faults.
“The near future will see the advent of cocktails of drugs that target all the different faults present within patients’ tumours at diagnosis,” says co-author and medical oncologist Dr Udai Banerji. “Patients will be genetically screened periodically throughout their treatment, so their drug cocktails can be adapted to suit the changing genetics of the tumour in real time.”
“Ideally, we want to know enough to really clobber the cancer upfront with, say, up to five or six perfectly suited targeted drugs that knock out all the gene faults in the different sub-populations and cure it there and then,” Professor Workman says. “But the reality is that you still only need one rogue cancer clone to grow out to allow the tumour to become resistant. However, it may be that cancer patients will have 10 years or more of remission before this happens, rather than the 5 to 10 months that they often have now in the difficult cancers. Tuberculosis and HIV were only controlled when we started combining drugs, and we believe it will be the same for cancer.
“I am very optimistic that by using this combinatorial targeted approach and trying to stay ahead of the game, we will greatly improve survival and indeed achieve cures in the currently difficult-to-treat cancers. It is a huge challenge – but, thanks to the scientific and technological advances that the whole field has made, we have a fantastic opportunity to work together and move this new approach forward as fast as possible.”
- Bissan Al-Lazikani, Udai Banerji & Paul Workman. Combinatorial drug therapy for cancer in the post-genomic era. Nature Biotechnology 30:679–692 (2012)