Share this article.

Mapping the path to personalised cancer care

ArticleDetailDownload PDF
Professor Hannah Carter, from the University of California in San Diego, has helped to bring network mapping of protein-protein interactions to the search for precision medicine in the fight against cancer. Using specially developed bioinformatics tools, her team can integrate information from genetic analysis, protein structure and protein interactions to map the mutations which ‘rewire’ the biological system – opening up the possibility of personalised diagnostics and therapies.
Professor Carter believes mutations should be viewed as perturbations to a biological system rather than just a single protein molecule. At the University of California in San Diego (UCSD), she attacks this problem with approaches that have their basis in bioinformatics and machine learning; skills that she honed as a graduate student training in cancer genomics at Johns Hopkins University in Baltimore, US. There she developed tools to predict whether mutations could drive tumorigenesis or perturb protein function using machine learning. At UCSD she pursued additional training in systems biology and network analysis, enabling her to take that approach further and is mapping the broader impact of these perturbations on cellular activities.

Protein pathways
At the cellular level, information is transmitted by chains of interacting molecules, consisting predominantly of proteins. Through organisation into such pathways, proteins work together to create cellular behaviours. These pathways are often complex and include many interacting components; consequently, a system-wide approach is needed to probe the effects of functional mutations to any single pathway component. Prof Carter’s work focuses on the ‘interaction interfaces’ or surfaces at which these proteins come into physical contact with other proteins in dynamic interactions.
The correct configuration of these interacting interfaces is vital for protein function and correct pathway operation, some are highly conserved across species (performing the same function) and others are tightly specific to one particular interaction. This specificity and ubiquity is vital for the correct functioning of cellular machinery and close control of the organism’s homeostasis.

We found that surprisingly distant regions of the genome could affect whether or not a particular cancer gene was mutated in a tumour.

In cancer, key pathways that control cell growth and survival are perturbed, allowing the uncontrolled proliferation of the cell population and ultimately leading to tumour development and progression of the disease. Many of the gene mutations that result in aberrant proteins, and consequently improper pathway activities, are well documented. Genes that are ‘turned on’ or activated by mutations, resulting in increased growth and proliferation are termed oncogenes, those that are ‘turned off’ or inactivated by mutations that cause the generation of a non-functional protein, are called tumour suppressor genes.

A map of interactions between genetic risk factors and genes that later drive tumour development
Modelling the terrain
Prof Carter and colleagues have recently shown that, across cancers, acquired mutations selectively target the interfaces in protein-protein interactions (PPI). By mapping more than 1.2 million known cancer mutations onto 3D models of protein conformations, the team were able to show that mutations were unexpectedly concentrated to interaction interfaces on the surface of certain proteins.
Commenting on this protein structure-based analysis of cancer genes, Prof Carter said they had found, “an excess of core mutations in tumour suppressors, consistent with their frequent inactivation in cancer. However, when only surface residues were considered, we observed that mutations in both oncogenes and tumour suppressors tended to be located at PPIs.”
Prof Carter goes on to describe how many cancer genes have mutations at multiple interfaces, leading to the possibility of a unique ‘fingerprint’ of mutations in any one tumour. This view was supported by evidence of significant differences in patient survival depending on the distribution of so-called ‘hot-spot’ mutations on these interfaces.
The results of this study have the potential to improve patient care through optimisation of treatment to a specific individual. Understanding the mechanisms by which specific mutations influence tumour behaviour and the patient outcome could support more accurate prediction of which therapy will be effective and provide novel druggable targets for individual cancer types, thereby improving the chances of successful treatment.
Going down the RAS route
Taking this further, Prof Carter and colleagues have focussed on the Ras group of oncogenes. RAS is a protein which sits at the centre of a complex signalling network related to cell proliferation. All cells communicate with each other through a variety of signals, telling each other when to grow, divide or die. These messages need to be passed accurately across the cell membrane and on to the cellular processes which effect a response.
Like in the popular children’s game of passing a message along with a chain in a whisper, any disruption to the message can have serious repercussions to the final result. RAS has long been identified as a vital protein in the chain, the disruption of which can cause a cell to enter the unrestricted proliferation that underlies tumorigenesis.
Although apparently simple, the RAS system is deceptively complex and involves many diverse interactions with other proteins as well as localisation differences and conformational changes which affect the activity of RAS itself and the pathways it influences. This is partly responsible for the previously held belief that RAS was an ‘undruggable’ target for cancer therapeutics. This is now changing however as groups uncover previously unknown interaction sites and possibilities.
A new direction
Despite this complexity, Prof Carter and her colleagues were able to correctly predict the binding interfaces of well-known effectors using their modelling strategy. They achieved this by constructing a protein-protein interaction network which was specific for the various RAS isoforms (functionally similar proteins that have a similar but not identical amino acid sequence) and predicting 3D complexes involving RAS isoforms and interaction partners to identify the most probable interaction interfaces.
They further used the models they had created to investigate competitive binding scenarios and potential multi-protein complexes which would be compatible with the available RAS surface interaction interfaces. This method also allowed for the generally considered effects of somatic (not inherited) mutations on RAS protein interactions to be probed and validated.
TP53 mutations (triangles) are mapped to protein interactions (edges between circles). Different tumours carry mutations with TP53 and may respond differently to drugs.
The results of the study were presented in the context of the different locations of RAS activity, either at the cell surface or within the cytoplasm (interior matrix of the cell). This gave the distinct advantage of being able to provide a list of RAS protein interactions with possible cancer-related consequences, which could help guide future therapeutic strategies to target RAS proteins.
Letting the genes lead the way
Her latest research brings networks to the new problem of interpreting the interactions between hereditary or germline genes and those mutated in somatic cells. Many of the somatic mutation events that lead to tumorigenesis have been well characterised in an effort to better understand how cancer can be treated, but Prof Carter and her colleagues have shown that a person’s germline DNA is a strong predictor of which such events are likely to unfold.

[This work] creates a validated resource of inherited variants that govern where and how cancer develops.

In studying interactions, across thousands of patients, between a patient’s inherited genome and the characteristics of the tumour that they developed, Prof Carter says, “we found that surprisingly distant regions of the genome could affect whether or not a particular cancer gene was mutated in a tumour”.
Some of the gene variants resulted in an 8-fold increase in the incidence of a particular somatic mutation and the information available includes identification of which tissues are most at risk and which genes are likely to be involved. Prof Carter commented that the study “creates a validated resource of inherited variants that govern where and how cancer develops, opening avenues for prevention research.”

The ultimate destination
All these studies fit with Prof Carter’s wider goal of making progress on three major areas of need for cancer medicine. The first is identifying inherited risk factors for cancer to aid early intervention; the second is to develop bioinformatic tools which can help researchers and physicians to understand how genetic variation influences cancer risk and progression; the third strand is the vital step of translating molecular measurement data from tumours into therapeutic opportunities which will benefit patients. As a recipient of a 2013 NIH Director’s Early Independence Award, a Siebel Scholar and a 2017 CIFAR Azrieli Global Scholar, Prof Carter is opening possibilities of personalised healthcare which have the potential to revolutionise patient therapy and outcomes. This is precision cancer medicine in the round: aiding prevention by identifying susceptibilities and improving diagnostic tools; whilst providing the opportunity for uniquely tailored therapies to those already gripped by this deadly disease.

How does your research contribute to the ultimate aim of personalised medicine in cancer care?
We seek to enable interpretation of a profile of genetic alterations in the context of the cellular system in order to advance precision medicine beyond the one gene-one drug model. Central to this idea is recognising that different mutations in the same gene can create differences in patient outcomes that must be understood at the pathway rather than the protein level. In the future, we expect to project cancer risk or plan therapy on the basis of the state of the system, rather than from a list of perturbed genes.

References

  • H. Billur Engin, Daniel Carlin, Dexter Pratt and Hannah Carter, Modelling of RAS complexes supports roles in cancer for less studied partners, BMC Biophysics 2017,10 (Suppl 1):5. doi:10.1186/s13628-017-0037-6.
  • Engin HB, Kreisberg JF, Carter H (2016), Structure-Based Analysis Reveals Cancer Missense Mutations Target Protein Interaction Interfaces. PLoS ONE 11(4): e0152929.
    doi:10.1371/journal.pone.0152929.
  • Hannah Carter, Rachel Marty, Matan Hofree, et al. Interaction Landscape of Inherited Polymorphisms with Somatic Events in Cancer. Cancer Discov 2017;7:410-423. Published OnlineFirst February 10, 2017. doi:10.1158/2159-8290.CD-16-1045.
Research Objectives
Prof Hannah Carter’s research focuses on computationally modelling how DNA mutations in tumour genomes impact intracellular biological processes and cellular behaviours, and how these cellular level changes cause cancer.
Funding

  • NIH
  • CIFAR

Collaborators

  • Trey Ideker 
  • Maurizio Zanetti
  • Joan Font-Burgada

Bio
Dr Carter is an Assistant Professor at the University of California, San Diego Department of Medicine. Her research addresses challenges in computational cancer genomics and precision cancer medicine. She is a recipient of a 2013 NIH Director’s Early Independence Award, a Siebel Scholar and a 2017 CIFAR Azrieli Global Scholar.
Contact
Assistant Professor Hannah Carter
University of California, San Diego
Medicine
9500 Gilman Drive #0688
CA La Jolla 92093-0688
USA

E: [email protected]
W: https://scholar.google.com/citations?user=nlWQnXMAAAAJ&hl=en
W: https://medschool.ucsd.edu/som/medicine/divisions/med-genetics/research/pages/default.aspx

Creative Commons Licence

(CC BY-NC-ND 4.0) This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. Creative Commons License

What does this mean?
Share: You can copy and redistribute the material in any medium or format
Related posts.