EACR 2016

Understanding mutational processes in human cancer

Dr Ludmil Alexandrov - Los Alamos National Laboratory, Los Alamos, New Mexico

I’ll talk today about mutational processes, so these are the processes that cause human cancer, and I’m going to focus on something called mutational signature, so these are the patterns of mutations induced by mutational processes on cancer genomes.

Can you tell us about the “clocklike” mutational processes?

We have identified a number of mutational processes in human cancers, some of those associate with age so they accumulate as the person gets older. We refer to these processes as clocklike mutational processes and we think they are responsible for the majority of mutations that accumulate in normal somatic tissue.

What methodology did you use?

The way we identify mutational process and the way we identify the clocklike mutational process is by looking at cancer genomics data and using machine learning algorithms to understand which are the different patterns of mutations imprinted on cancer genomes. In regards to clocklike mutational processes we looked at processes that accumulate with age so we took patients for whom we have information about their age at diagnosis and we looked whether any mutational process increases its mutational burden with the increase of the patient’s age.

How does this relate to treatment and diagnosis?

A lot of my work on mutational signatures and mutational processes is mostly focussed actually on cancer prevention - understanding what are the processes that cause cancer and then based on that trying to develop cancer prevention strategies.

How do you see bioinformatics influencing cancer research in the future?

I think bioinformatics will have a very, very big impact on cancer research both for treatment and for prevention we can think about predicting response from different drugs, whether these are targeted inhibitors, whether these are immunotherapies, we can think about developing computationally, as I said, understanding what the processes that cause cancer and then trying to think of how to prevent those processes.

What will be your focus in the coming years?

The main interest is to try to understand the incredible difference in cancer incidence across the world. So there are places across the world that have 5-10 times higher incidence of cancer compared to others. So an example would be oesophageal cancer in China is about twenty times higher than oesophageal cancer in Italy. So trying to understand how this has happened will be the focus of my research for the next few years.

Are there differences between mutational signatures of different cancers?

There are mutational signatures that are common across all cancer types, clocklike mutational signatures is one example for that, another example is mutational signatures such as by the family of deaminases. But there are a lot of mutational signatures that are confined to single cancer types, for example we have a mutational signature associated with tobacco chewing and, as you can imagine, this has been found only in the mouth cancers of tobacco chewers. Now, if we compare two people which have the same cancer type, if you look at the mutational signatures in two people who have the same cancer type usually they are completely different. So even though we call it breast cancer or lung cancer we’re going to see very different mutational signatures operative in these cancer types, indicating the cancers most likely arose by different means.

Can you predict cancer from the signatures?

This is very much what we do right now, we look at single patients and for a single patient we can say in these patients the cancer in this patient was caused by these mutational processes because we see this set of mutational signatures.

Does this mean that there will be higher treatment cost?

I think some of the mutational signatures could be valuable for treatment and these are signatures reflecting failure of DNA repair mechanisms and we have different inhibitors or different drugs targeting repair mechanisms but the main value there will be actually screening large scale populations and saying in this population the reason these people get cancer is because they do a, b and c and what can we do to prevent a, b and c? 90% of lung cancers are caused by tobacco smoking so we know how to prevent the majority of lung cancers. But if you have the question how can we prevent the majority of breast cancers most of the time the answer is harder.

What methods are used to find mutational signatures?

To find mutational signatures we need to sequence the genome of a cancer as of right now. So we need to take a biopsy of the cancer, we need to sequence it and then we need to do bioinformatics analysis of the sequencing data and then we can say which are the mutational signatures. So it’s more complex than a clinical test at the moment.

Is there anyone else working on similar studies as you?

I think there are several large groups, obviously there is the Sanger Institute with Max Stratton working on that; the Broad Institute with Gad Getz there working on that; focus in Singapore, Korea, Japan, so there are quite a lot of things. It’s not as popular as immunotherapy but it’s getting quite a lot of people excited about, again, trying to understand what causes the cancer.

What does winning the Young Investigator Award mean for your research?

It means more visibility, it means that I’m going to be able to tell more people about what I do and hopefully, again, excite more people about my research and about cancer prevention.