New research from the Cancer Grand Challenges Rosetta team shows that metabolic profiling using a variety of mass spectrometry techniques can be used to stratify tissues according to their underlying mutations in colorectal cancer, and that such techniques can be used to identify new potential targets for cancer treatment.
Rosetta, led by Professor Josephine Bunch of the National Physical Laboratory, UK, is one of the teams tackling our 3D Tumour Mapping challenge. The team has been working to develop a ‘Google Earth’ for cancer that can zoom in from the whole tumour right down into the metabolome (fats, proteins and sugars produced by cellular processes) of tumour cells.
We spoke to Dr Johan Vande Voorde, associate scientist in the group of Rosetta co-investigator Professor Owen Sansom at the Cancer Research UK Beatson Institute, to find out more about the team’s newest study, published this week in Nature Metabolism.
What are the key findings from your new study?
When the Rosetta team started, one of the first things we wanted to understand is how does the metabolic landscape of the intestine change when we start to introduce cancer-driving mutations that are very common in colorectal cancer. Our lab has developed genetically engineered mouse models that allow us to study the effects of these mutations.
In this study, we first utilised mouse models of hyperproliferation – these are not cancer models, but the intestine cells within the mice are genetically engineered to carry common mutations known to drive colorectal cancer, which cause the cells to grow quickly.
To understand what was happening on the metabolic level as a result of these mutations, we worked with the National Physical Laboratory and Imperial College London who have access to a technology called REIMS (rapid evaporative ionising mass spectrometry) to analyse tissues from our mouse models.
REIMS, developed by Rosetta team member Professor Zoltan Takats (Imperial College London, UK), is the technology that underlies the iKnife – a surgical instrument that is coupled with a mass spectrometer. As the surgeon makes tissue incisions during surgery, the vapour is forced into a mass spectrometer and informs on the metabolic constitution of the tissue that is being resected. It’s been shown before that this technology can discriminate between normal tissue and cancerous tissue.
In our hyperproliferation models, REIMS was able to separate tissues according to their underlying genotype. This is useful information to have because cancers are targeted in very different ways depending on the mutations they carry.
To test whether using metabolic profiling by REIMS could have clinical potential for classifying colorectal cancer tissue, we next used REIMS to profile tissue samples taken from 24 people undergoing surgery for colorectal cancer or polyps. We used whole exome sequencing to identify the genetic drivers causing the cancer or polyps in the tissues.
We used this discovery cohort to build a predictive model for KRAS mutations, which are present in 40-50% of colorectal cancer cases. In this limited set of clinical samples, we found that the REIMS technology could be applied to stratify samples according to whether or not they carried a KRAS mutation.
Our findings provide preclinical and clinical evidence that we can use the metabolic profile of intestine samples to understand what the genetic drivers are within a tissue.
In addition to REIMS, what other mass spectrometry techniques did you use?
REIMS is really good for getting a snapshot of the metabolites within a tissue. But we also looked at the spatial aspect of the metabolome within the intestines of our models using mass spectrometry imaging and found that the metabolic profiles within a tumour are different between different compartments and cell populations.
We wanted to go into more detail to find what the different metabolites are that are driving the separation between genotypes. We used liquid chromatography mass spectrometry to analyse the intestinal tissue from our hyperproliferation models and stumbled across an interesting finding. Upon loss of the tumour suppressor gene APC in our models, which occurs in about 80% of colorectal cancer cases, there was dysregulation of a biochemical pathway called the methionine cycle.
We zoomed into this phenotype of interest using patient samples and mouse models of hyperproliferation and colorectal cancer and found that a particular enzyme within the methionine cycle, called adenosylhomocysteinase (AHCY), was upregulated in colorectal cancer compared to normal intestine, suggesting that it could be a potential target for colorectal cancer.
We took these findings into the lab to explore the effects of inhibiting AHCY in organoids, our hyperproliferation model and in an APCmin mouse model – a very well characterised model for intestinal tumorigenesis. When we inhibited AHCY, we found that we could reduce cellular proliferation in the organoids and hyperproliferation model, and significantly impair tumour growth in the mouse model – indicating the potential of AHCY as a cancer target.
Why is this study and its findings important? How do they feed into addressing Rosetta’s challenge?
I think the strength of the study is that, as a Cancer Grand Challenges team, Rosetta has come together and harnessed a lot of different technologies. By combining mass spectrometry imaging technologies and traditional liquid chromatography mass spectrometry, we've presented a pipeline where we can show that the various technologies can be used for different aspects of interrogating the metabolome in the context of tumour biology.
These technologies have been around for a while but, in this study, we have brought them together, and we’ve shown that it's powerful to use them in combination. Our findings suggest that we can exploit these technologies and the information that we get from them for stratification of patients or for looking at the specific alterations in compartments of the tumour for different cell populations.
Rosetta’s challenge is all about mapping the tumour metabolome in various ways, including in a spatial way, and that's exactly what we've done. We’ve harnessed sophisticated preclinical models of colorectal cancer and patient samples to comprehensively understand the effect of common oncogenic mutations in colorectal cancer on the metabolome. This is helping us to understand how we can use metabolic profiling for classifying cancer tissues and identifying new potential treatment targets.
What are the next steps for this research?
We're very keen to validate the model that we developed for stratifying tissues according to KRAS status. To do this, we want to look at a much more extensive patient data set to find out whether we can use this as a tool for patient stratification.
We’ve not yet fully exploited our data. So far, we’ve only explored the ability of these data to differentiate tissues according to their KRAS status, but of course this is not the only player in cancer biology – there are many different mutations that are known to impact the metabolome quite significantly. We want to understand if we can find other mutations that can be used for stratification within the metabolic data we have generated.
We also want to investigate whether our spatial metabolomics data could reveal other vulnerabilities in colorectal cancer that could be exploited therapeutically to target cancer.
For our AHCY work, we want to explore its potential as a cancer target in more complex models and to find out whether we can interfere with other cancer processes in colorectal cancer, such as metastasis, to determine how broadly this could be applied as a new target.
As told to Bethan Warman. Find out more about this research in Nature Metabolism.
The Rosetta team is funded by Cancer Research UK. You can find out more about the team’s work here.