Getting a grip on intractability

Pancreatic cancer cells - CRUK LRI Electron Microscopy Unit

​​

One of the cornerstones of our Research Strategy is to build capacity in research into cancers of unmet need: those we’ve identified due to their poor five-year survival and limited improvements in treatment in the past decade.

Two years post-launch, we’ve seen an increase in funding for research into lung, pancreatic and oesophageal cancers and brain tumours – together totalling £74.7 million in 2015/16. Not only are we spending more, but we’re also attracting more people into the area.

However, we’d like to get even more scientists thinking about cancers of unmet need, applying for grants, and helping to build effective research communities in this area. There are challenges of all flavours, from the most basic discovery research, to those directly related to patient benefit. So why might this field be of interest to you and your colleagues? We’ve been talking to a few of our researchers to find out what brought them to work on cancers of unmet need and what opportunities the research holds.

For CRUK, cancers of unmet need are those with poor five-year survival rates, and limited improvements in treatment in the past decade. We urgently need to transform the outlook for patients with tumours in this category, which includes lung, pancreatic, oesophageal cancers and brain tumours. However, despite highlighting lung, oesophageal and pancreatic cancers as areas of priority in our previous strategy, we did not see research effort increase by as much as is needed. This is why we’re now being far more proactive.

Research in each of the four tumour types holds different challenges, and communities are often smaller and less well developed than those studying more mainstream cancers. Therefore, we’ve been working with researchers to develop and support relevant activities in each area and cancer type by providing funding, supporting conferences and bringing people together. We know that tackling a task of this scale is going to require investment and innovation – and we’re committed to making it happen. And we’re confident that there are huge opportunities to carve out an outstanding research career in this area, whilst simultaneously having the chance to significantly improve patient outcome.

Cancers of unmet need share common features which make them particularly difficult to treat. Firstly, they are difficult to diagnose early, as in their early stages they can be asymptomatic or share symptoms with more innocuous problems. However, if they are diagnosed, they respond well to treatment at this early stage.

The problem of early diagnosis is one that has preoccupied Cambridge-based clinician scientist Professor Rebecca Fitzgerald for much of her career. Rebecca works on oesophageal cancer, where the odds of surviving the two main subtypes, adenocarcinoma and squamous cell carcinoma, improve remarkably if they are caught early. As Rebecca says, “Most adenocarcinoma patients will have come through a pre-malignant stage, called Barrett’s oesophagus, and if we really want to turn survival around, finding those people is very important”.

Cases of Barrett’s oesophagus aren’t picked up as often as they should be, and even if the condition is diagnosed and the patient enters a surveillance programme, progression of the disease may be missed. To solve this problem, Rebecca and her colleagues developed the Cytosponge, a gelatin-enclosed mesh sponge attached to a string. Swallowed by patients, the gelatin capsule dissolves and the sponge expands, such that when it’s pulled back out, it comprehensively samples cells from the whole oesophagus, entirely removing the risk of sampling bias. The cells are sent for analysis to determine whether Barrett’s is present or not using the newly established Barrett’s marker TTF3.

In theory there may be plenty of time to intervene before the disease changes from being relatively benign into something dangerously aggressive.

—Rebecca Fitzgerald

Following the success of trials showing that the Cytosponge was as good a diagnostic tool as endoscopy, and far more pleasant for patients, Rebecca’s team has received funding for a 4,000-patient clinical trial to pave the way for the Cytosponge to become the first-line assessment for Barrett’s oesophagus. If the trial is successful, the Cytosponge is cheap enough that it could eventually be used by GPs as a routine screening tool, potentially alerting many more people that they have Barrett’s.

However, the problems won’t end there. Only about 1-5% of people diagnosed with Barrett’s will go on to develop cancer, and the potential adoption of the Cytosponge, with a consequent increase in the numbers of cases of Barrett’s being detected, will only add to the numbers of ‘worried well’ whose chances of developing adenocarcinoma are slim. There is a huge need for robust biomarkers predictive of increased risk, but currently, they are few and far between. To help plug this diagnostic hole, Rebecca is leading the OCCAMS trial, a CRUK-funded multicentre study which is part of the International Cancer Genome Consortium (ICGC) global project to generate a comprehensive catalogue of genomic abnormalities in 50 tumour types. Phase 1 of the project, the sequencing of 500 tumours and corresponding healthy tissue, is nearly complete, and phase 2, taking longitudinal samples from patients with Barrett’s and oesophageal adenocarcinoma, has recently commenced. Like all ICGC data, the sequencing information from the OCCAMS project will be a freely available resource for all researchers.

OCCAMS seeks to map the molecular genetic landscape characterising the journey of oesophageal lesions from metaplasia to adenocarcinoma. The hope is that there will be a step change in our molecular understanding of oesophageal adenocarcinoma and that this knowledge can be applied to Cytosponge samples to identify which patients are at high risk for progression to cancer who need preventative treatment. The task is large but not insurmountable, says Rebecca: “One of the big problems with this cancer is that it doesn’t have a recurrent driver gene – it’s a disease dominated much more by multiple tumour suppressor genes being lost. However, there’s some evidence for very gradual evolution interspersed with crises, so in theory there may be plenty of time to intervene before the disease changes from being relatively benign into something dangerously aggressive.”

But where does Barrett’s oesophagus come from? And can we learn any lessons on cancer risk by studying the evolution of the disease based on concurrent analysis of the genetic and architectural features of Barrett’s? This is the research focus of Dr Stuart McDonald, of London’s Bart’s Cancer Institute, and he’s recently proposed a theory about the origin of the lesion that may overturn current dogma. Stuart has recently been awarded a CRUK Programme Foundation Award, a funding scheme we introduced to provide support for mid career researchers to establish or develop their own group. Stuart, originally an immunologist, came into Barrett’s research through his interest in the inflammatory response. But in the course of his work, detailed phenotypic characterisation of biopsy samples led him to an unexpected conclusion: “The traditional view is that Barrett’s glands develop from the normal lining of the oesophagus, but actually, they look really similar to pre-cancerous changes seen in the stomach lining,” Stuart says. “What we think is that when patients get heartburn, it’s corrosive, so it strips the normal oesophageal epithelium, leaving behind an empty ulcerated landscape. And that’s then preferentially colonised by stomach epithelium, which is better adapted to acid conditions and over time exhibits features of intestinal metaplasia – much like that observed in H. pylori infection in the stomach.” The idea, though still controversial, is gaining ground, although as Stuart says, “there isn’t a smoking gun link yet – we’re still working on that.”

A further observation made by Stuart’s lab and other groups is that there are a broad range of identifiable gland types in Barrett’s lesions. These range from those that are very similar to the acid-secreting glands of the stomach to those that look like intestinal crypts – indeed intestinalisation (defined by the presence of mucin secreting goblet cells), is diagnostic for Barrett’s in some countries. However, we do not know the clonal relationship between these gland types nor do we know the cancer risk each possess. “We have preliminary data suggesting that gland phenotypes themselves evolve, and that patients at risk of developing cancer show an increased diversity of gland types.” Understanding how this changes over time in conjunction with genetic analysis may permit the generation of predictive tools to identify patients at risk of developing cancer and those that are not.

Professor Gerard Evan seeks to understand cancer through the medium of an oncogene, in his case MYC, one of the key drivers of many cancer types. Drawn by the excellent mouse models that are available, and by the intractability of the tumours, his lab at the University of Cambridge has recently started working on lung and pancreatic cancer, and the switch of focus has already thrown up some interesting new puzzles. Gerard has been especially struck by the dissonance between the phenotype and genotype of the two diseases. “Lung and pancreatic cancers are about as different as you can get in terms of what they look like,” says Gerard. “Why is this, when, at least to a first approximation, they’re driven by the same oncogenic mutations?”

Gerard’s solution to this conundrum harks back to an observation made by the 19th century physician Rudolf Virchow that cancers are wounds that never heal: “Our data clearly indicate that MYC and RAS hack into the regenerative programme already present in each type of tissue. So the natural evolutionary niche in which tissues find themselves will dictate a lot about their response to oncogenic mutations, in the same way that it dictates their response to normal injury. And it will also influence how they respond to therapy.” In Gerard’s view, this idea, that cancers take their identity from their tissue of origin, means that they can only be understood within that context. Studying isolated cancer cells in tissue culture is now only a limited weapon in the researcher’s armoury.

MYC’s role in tumour formation has stretched to include hiding cancers from the immune system. In March 2016, the Felsher lab at Stanford University reported that in a mouse model of T-cell acute lymphoblastic leukaemia (T-ALL), MYC up-regulates the ‘invisibility’ ligand PD-L1, thwarting the immune response to the cancer. Gerard’s lab also sees PD-L1 upregulation by MYC, but intriguingly, how this happens varies depending on cancer type: when MYC is switched on in pancreatic cancer models, PD-L1 is upregulated on the tumour cells themselves, whereas in a non-small cell lung cancer model, it is switched on solely in the macrophages that flood the tumour, despite MYC being activated only in the tumour cells. “It’s fascinating that it’s so complicated”, says Gerard, “there is a code we don’t yet understand that differs from tissue to tissue – the number of chemokines and cytokines is distinct in each tissue type and the types of cells that enter and exit the tumour also differ”.

Gerard’s view that cancer phenotype is of the utmost importance, is one that resonates with Dr Simona Parrinello, a cell biologist based at Imperial College’s Hammersmith Campus who studies the mechanisms of invasion of glioma, a type of brain cancer. “Mutations in the tumours will affect signalling in different ways depending on which gene is affected, but different mechanisms seem to converge to produce a common phenotype of invasiveness”, says Simona.​

We need to know how invasion is driven. It will have profound implications for therapy.

—Simona Parrinello

The key to more effective treatment for glioma is to get to grips with the glioma stem cells, which share many characteristics with the normal neural stem cells in the brain. Brain cells are highly plastic, and this plasticity translates in glioma into extreme genetic heterogeneity and a pro-invasive predisposition; this combination of continual evolution and rampant spread are what makes glioma so difficult to treat.

In addition to the normal tools of the cell biologist’s trade, Simona’s lab uses intravital imaging to peer into the brains of mice. Cells are implanted under a transparent cranial window, allowing longitudinal imaging as tumours develop. “It tells you about the ecosystem of the tumour”, Simona says, “using fluorescent markers, we can label many specific cell types and structures within the brain in the vicinity of the tumour, and look not just at the tumour cells but how they interact with their surroundings. And we can also use what we know about the genetics of the human cancer to make mutations in normal cells, and see how different mutations drive invasion.” Simona thinks that this cell biological approach could help make sense of the heterogeneity of gliomas and improve our understanding of invasion overall: “These are key questions – we need to know how invasion is driven as it could have profound implications for therapy”.

Dr Steve Pollard, of Edinburgh’s MRC Centre for Regenerative Medicine, brings another perspective to the glioma problem: “Now is a good time to think about cancer biology with a developmental biologist’s hat on” he says, “it’s not new, but it’s come back again as the field moves on from analysing the molecular biology of cancer into cancer in a tissue specific context. Stem cell methodologies like culture conditions, sorting, and how to grow and transfect cells are all tools we can apply now to human brain cancers. And CRISPR has completely and utterly transformed the field, as we can do rapid gene targeting now in primary human neural stem cells.”

Steve’s lab works on neural stem cell specific transcription factors, and his focus, like Simona’s, is to strip out all the heterogeneity and focus on universal features common to all gliomas. His approach is to target the ‘stemness’ of the cells driving a glioma, and change them into something less lethal: “We want to target lineage identity and self-renewal rather than just cell proliferation” he says, “the big advantage is that the proteins involved might be tissue-specific so there may not be many side effects.” A lot is known about the transcription factors important for self-renewal and maintenance of neural stem cells, but drugging these is notoriously difficult. Instead, Steve’s lab is working on their mechanism of action to see whether there are co-factors or partners that may be vulnerable to drug targeting. Neural stem cells have only very limited function in the adult brain, so a drug able to destroy all the neural stem cells may not cause catastrophic damage.

In a disease where human tissue is an essential resource, standardisation of cell lines and making them available to the community is incredibly important. To that end, Steve has successfully coordinated a bid for a CRUK Centres’ Network Accelerator Award, shared between the University of Edinburgh, the Francis Crick Institute, UCL and The Institute of Cancer Research. The Glioma Cellular Genetics Resource will provide high quality well-characterised cellular models of glioma with matched normal stem cells, together with all the validated genome editing tools that the community needs. Steve’s hope is that in addition to its more practical functions, it will act as a community hub, where anyone working on glioma can get the information and datasets they needed to get a project underway: “The zebrafish, mouse and fly communities have these kind of databases and repositories and we should be thinking about them for human cancers as well – making it more like a model system.”

The extra financial support CRUK has given to cancers of unmet need is paying dividends – more and more researchers are clearly turning their attention to this underexplored area. However, the journey towards a cure may be a long one. CRUK will be supporting these fledgling communities every step of the way as they take on the challenges that lie ahead. Identifying and solving tough clinical challenges; linking basic to translational research; provocative thinking, and strategic provision of communal resources: these factors, coupled to excellent science, will go a long way towards cracking the difficult problem posed by cancers of unmet need. We have every confidence in our researchers’ abilities to make the scientific advances so desperately needed by thousands of patients, and in return, we pledge that we’ll be there to help you.

Discover more

Research opportunities for cancers with substantial unmet need

CRUK Science Blog: Making a special case – how to boost brain tumour research

This story is part of Pioneering Research: our annual research publication for 2015/16.