Understanding the genetic 'patchwork' of ovarian cancer will lead to more effective treatments. Patchwork quilt
Cancer is not one disease, but a multitude.
We often say that there are more than 200 types of cancer – lung, prostate, bowel, breast and all the rest – but this is a fairly broad definition, based on the rough number of different principal cell types in the body.
Thanks to advances in gene-reading (sequencing) technology over the past few years, we now know that each person’s cancer is as unique as they are. And as cancers grow and spread, the cells within them evolve and change in different ways. This means that within a patient’s body, individual tumours or groups of cells within a single tumour can be genetically distinct, even though they all started from the same place.
Scientists use the term tumour heterogeneity to describe this patchwork made up of clusters of different cancer cells. In fact, it’s been known about for many years, but it’s only recently that we’ve been able to look in detail at the genetic threads that create it.
Knowing about heterogeneity helps to explain why cancers are so difficult to treat successfully once they’ve spread through the body. Cancer drugs or radiotherapy may kill some of the cancer cells, but others will have gained genetic changes that make them resistant. And once the vulnerable cells have been picked off, any remaining resistant ones can just keep on growing.
So understanding this patchwork is vital if we’re to make meaningful progress in treating advanced cancers and save more lives. And a new study from scientists at our Cambridge Institute, published in the journal PLoS Medicine, reveals what’s going on in a type of ovarian cancer, paving the way for new approaches for treatment.
To find out more, we spoke to Dr James Brenton, who led the research. As a doctor working in Addenbrooke’s Hospital in Cambridge, he treats women with ovarian cancer every day. And although fewer women are getting ovarian cancer nowadays, survival from certain types has changed little over the years. One such form of the disease is high-grade serous ovarian cancer.
“The problem here is that most women with this type of cancer already have advanced disease when they’re diagnosed,” says James. “We think that this type of tumour probably starts in the fallopian tube, but by the time it’s detected it’s just a big mess, and the cancer cells are all over the place in their abdomen. The other thing is that these cancers often have a fault in a gene called TP53, which makes the p53 protein that normally protects us against cancer – so they grow and spread very quickly.”
Back in 2010, James and his team published a paper looking at the genetic makeup of high-grade serous ovarian cancer in cells growing in the lab that had been taken from tumours before and after treatment with carboplatin – the main drug used for the disease.
“What we found is that when the cancer comes back after treatment, it’s genetically very different from when it’s originally diagnosed. So we wanted to see if the same thing was happening in patients, rather than cells growing in the lab.”
The other thing James and his team wanted to do was question a common assumption: that the more pieces making up the genetic patchwork of a patient’s disease, the worse their outcome.
In many ways this is a sensible idea – the more different clusters of cells there are, the greater the chance that some of them will evolve resistance to treatment. But it’s not that simple.
“It could be different between cancers,” James explains. “Also we know that some ovarian cancers have faults in their BRCA1 or BRCA2 genes, which reduce their ability to repair DNA damage caused by the treatment. So, counter-intuitively, that might actually mean they respond better to therapy.”
Recently we’ve seen several studies looking at tumour heterogeneity in a range of different types of cancer, some of which we’ve helped to fund (for example, lung, bowel and kidney cancers). But what James and his team wanted to know was not just whether there were genetically different cells in a patient’s cancer, but to put some numbers on exactly how different they were.
Mapping the patches
To do this, James enlisted the help of Florian Markowetz and Roland Schwarz at the Cambridge Institute, both experts in using computer software to analyse genetic relationships. Together, they trawled through DNA sequences of 135 separate tumour samples from 14 women who had been diagnosed with high-grade serous ovarian cancer, taken from various places the disease had spread to inside their body, both before and after carboplatin treatment.
This allowed them to get an idea of the degree of heterogeneity in each woman’s cancer – whether it was a patchwork of a few large blocks of closely-related cancer cells, or an intricate collage of many genetically different groups.
Then they looked to see how this variation was related to how the women had fared after being treated with the disease.
“The bottom line is the more detailed the patchwork is – in other words, the more heterogeneity there is – the higher the chances that the cancer will come back quickly, and that the patient will die from the disease,” James explains.
These ovarian cancers had extraordinarily chaotic DNA. “You see big chunks of DNA being copied or lost. Each of these changes could potentially encompass hundreds of genes, so now we need to work out which are the important ones that are driving the cancer to grow and become resistant to treatment,” says James.
The team also compared the genetic makeup of cancer samples taken early on, before women were given carboplatin, to later cancer cells that had developed resistance to the treatment and were continuing to grow. In doing so, they shed light on a long-standing puzzle: the origins of drug resistance.
James and his team noticed that certain gene faults, previously linked to treatment resistance, were present in tumours before treatment started, but only in a very small proportion of the cancer cells. One example is a gene called Neurofibromin-1 (NF1). Mistakes in NF1 are often found in advanced high-grade serous ovarian cancer, and it seems to play a role in making cancer cells grow again after treatment.
“We found that the later drug-resistant tumours have a high proportion of cells carrying a faulty version of NF1. But when we looked carefully, around five per cent of the cancer cells prior to treatment had it too. And it was even there in the original tumour sample from the fallopian tube, but only in a few cells.”
This suggests that the genetic seeds of resistance are already present at the very earliest stages of cancer, and that the population of cells carrying them expands and grows over time.
Unfortunately, it’s very difficult to test patients for these rare gene faults when they’re first diagnosed, although there is a technique that might help: so-called ‘liquid’ biopsies, which can detect DNA shed from tumours into the bloodstream (and which we’ve written more about here). Can the early signs that groups of cells carrying particular genes faults are starting to grow, be spotted in the blood?
“We want to be able to use this technique to monitor patients in real-time as they have their initial treatments, and see what is changing in their cancer at a genetic level,” James tells us. “Clearly, if we could see changes occurring earlier, we might be able to do something to make treatment more effective – such as switching to a different drug if it looks like the cancer is becoming resistant. But we need to be able to pick up on these changes quickly, while there’s still a chance to make a difference.”
At the moment, it’s not possible to measure the amount of heterogeneity and put a precise prediction on survival – but James is hopeful that he’ll get there.
“While we’ve shown that there definitely is a relationship between the amount of heterogeneity and survival, we can’t put precise numbers on it yet as we don’t have enough data,” he says. “But I think this kind of approach will eventually allow us to make those correlations. High grade serous ovarian cancer is a disease where there are extraordinary genetic changes, and it’s very complex to figure it all out. Now we need to work out ways we can do this on a larger scale, over hundreds of patients.
“We did this study looking at a lot of samples from a few patients, but we want to know if we can develop methods where we just need a couple of samples from many women.”
From patchwork to plan
From a scientific point of view, this is fascinating. But what we really need to do is turn it into more effective treatments. This is still very much a work in progress, and something that the Cambridge team is now turning its attention to, although there are a few exciting ideas in the pipeline.
James says: “We need to be thinking about how we schedule chemotherapy and other treatments to make it very unfavourable for these resistant cells to grow – we need to make life very uncomfortable for them.”
He thinks it’s important to look at the ways doctors give drugs to patients, and whether tweaking the dosing or timing could be beneficial for patients with lots of heterogeneity in their cancer.
“Is it such a good idea to give big doses of drugs every few weeks, or might it be better to give smaller doses more regularly?” James asks. “We need to think about how to schedule chemotherapy in different ways, whether we can use the immune system to tackle cancer as well as drugs, and even if we can alter the evolution of the disease in the body.”
Research like this is enabling scientists to reveal the genetic landscape of cancer within the body for the first time, and it should lead to urgently-needed improvements in the way that women are treated for ovarian cancer. And by agreeing to take part in studies like James’s, it’s these women who will ultimately lead to advances for future patients.
“I’m not sure how much they know about tumour heterogeneity,” he says, “They just want something to make the cancer go away and make them feel better. But they all really want to help our research.”
Cooke S.L. et al.(2010). Genomic analysis of genetic heterogeneity and evolution in high-grade serous ovarian carcinoma., Oncogene, PMID: http://www.ncbi.nlm.nih.gov/pubmed/20581869
Schwarz R.F. et al. (2015). Spatial and Temporal Heterogeneity in High-Grade Serous Ovarian Cancer: A Phylogenetic Analysis., PLoS medicine, PMID: http://www.ncbi.nlm.nih.gov/pubmed/25710373