Dr Chunxiao Song, a group leader and chemist at the Ludwig Institute for Cancer Research, University of Oxford, is developing novel tools to analyse liquid biopsies for pancreatic and oesophageal cancer. Here Chunxiao tells us about his latest project, jointly funded by CRUK and Oregon Health and Science University (OHSU), developed in collaboration with Dr Thuy Ngo. Together they aim to use epigenetic and transcriptome technology to detect cancer earlier and provide information on where the cancer originated from.
“Thuy and I were postdocs together in the lab of Stephen Quake at Stanford University, California, where we worked in parallel on new approaches for biological measurements. I developed a tool for measuring epigenetic information in cell-free DNA, while Thuy developed a tool for transcriptome measurements in cell-free RNA. Based on these works we were both recruited to start our own labs – I went to Oxford and Thuy to OHSU.
At Oxford my first big grant was a CRUK Early Detection Project Award, which started in May 2018. I’m using my technology to develop epigenetic maps for a range of cancers including lung, pancreatic, gastric, colorectal and breast cancer, hepatocellular carcinoma and glioblastoma.
The only limitation for applications is good ideas
When I first heard about the CRUK-OHSU Project Award it felt as though the grant had been tailor-made for us, since the application needed to be a joint one from investigators at a UK institution and OHSU. Thuy and I already knew we worked well as a team, but initially, I was concerned that because I already had a CRUK award, it would preclude us from applying. When I got in contact with CRUK they reassured me that the only limitation on applications was the number of good ideas that you had!
We decided to concentrate on early detection in pancreatic and oesophageal cancer for three reasons. Firstly, these two cancers are often diagnosed late and have particularly low survival rates. Secondly, I showed my technology works well in pancreatic and oesophageal cancer and Thuy showed her technology works well in pancreatic cancer. And thirdly, at Oxford I have access to oesophageal cancer blood samples and, at OHSU, Thuy has access to pancreatic cancer blood samples.
New technology helps identify cancer tissue of origin
Our project uses liquid biopsies – a test that looks for DNA and RNA shed by tumours (and in fact all cells) in a patient’s blood sample. Liquid biopsies offer multiple advantages over tissue biopsies in that they are pain free, non-invasive, cheaper and altogether less risky for patients. They can also be used to screen for genetic alterations, such as mutations and copy number variations, which distinguish patients with cancer from healthy people.
However, while you can use the genetic information to detect cancer and track treatment, there are few clues so far about where the cancer comes from.
We’ve found that adding in information about the epigenome and transcriptome provides important clues to where the cancer cells originated from that is vital for early detection.
My work focuses on detecting and measuring 5-methylcytosine (5mC) and 5-hydroxymethylcytosine (5hmC), two of the most common epigenetic marks in the mammalian genome. The rationale is that 5mC marks repressed genes and 5hmC marks expressed genes. Identifying genes that are expressed or repressed makes it possible to identify the tissue that they originate from, since genes are switched on and off in early foetal development to form individual organs.
I have developed new measuring technologies that use less harsh chemicals than the standard approach. This causes less DNA degradation, which makes it easier to measure small quantities of cell -free DNA.
Thuy has focused her research on measuring cell-free RNA. This is even more difficult than measuring cell-free DNA – but she has developed a special protocol that’s really gentle and doesn’t cause degradation of the RNA.
Considering two technologies together provides a fuller picture
We plan to use both technologies on blood samples taken from patients with oesophageal and pancreatic cancers. For the oesophageal part we’ll be testing samples from 100 patients with oesophageal cancer, 100 patients with Barrett’s oesophagus (a precancerous state for oesophageal cancer) and 100 healthy controls; while for the pancreatic part we’ll be testing samples from 100 patients with pancreatic cancer and 100 healthy controls.
To combine the epigenome and transcriptome information we’ll use machine learning to create classification models that distinguish cancer patients from healthy controls. In the case of oesophageal cancer, we also plan to identify signatures defining the precancerous condition.
We believe that the two technologies together will provide a much fuller picture than either alone, although we are keeping an open mind that they might each be better suited to different types of cancer. If our project proves successful we’d eventually like to run a clinical trial comparing the results of our tests to the gold standard of medical imaging.
Advice for applicants: talk to CRUK early on
When I first arrived in the UK, I was cautious about applying to CRUK for funding because I thought the fact that my first degree was in chemistry and that my PhD had nothing to do with cancer, would disqualify me from applying. However, when Dr Alexis Webb, a CRUK Research Funding Manager, visited Oxford, she explained that CRUK wants to attract applicants from other areas of research, like chemistry, physics, and engineering sciences, particularly for the early detection awards where there is an emphasis on developing technology.
My CRUK applications proved surprisingly easy. You just need to put together a proposal of around 2,500 words, fill in an application form, and then address the reviewer’s comments on a point-by-point basis. There was no requirement for an interview. My advice to anyone thinking of putting in an application is to contact the CRUK funding managers early on because they can help guide you through the process from the beginning.”