Development and Application of Computational Methods for Profiling Cancers at Single Cell Resolution

Cancer is a complex disease with many factors which determine how rapidly cancer cells can grow and spread throughout the body. Significant differences exist within the cancer cell population of a patient. These differences shape the interaction of cancer cells with the surrounding healthy tissue, with dramatic variation between patients. This so called cancer heterogeneity has profound implication for patient prognosis, and is one of the primary challenges to developing effective cancer treatments. Recent technological advances now allow for the measurement of multiple aspects of individual cells within a cancer. This has created an opportunity to precisely characterize the set of mutations in each cancer cell, along with their functional consequences and how they impact interactions with surrounding cells. My group will develop statistical machine learning approaches to analyze the complex datasets generated by these technologies.

Working alongside clinicians and biologists at BC Cancer, part of the Provincial Health Services Authority (BC Cancer), we will apply these computational methods to study the evolution of metastatic breast cancer and the mechanisms of relapse in follicular lymphoma. Ultimately this research will provide important insights that can guide the development of better strategies for the diagnosis and treatment of these cancers.

Genomic Organization At Large (GOALS) predicts aggressive biological behaviour in prostate and breast cancers

Prostate cancer is the most commonly diagnosed form of cancer for men in North America. Prostate cancer deaths have been in decline since the mid-1990s after the discovery of Prostate-Specific Antigen (PSA), which, when used for screening, results in a steep increase in the number of early diagnoses. A large percent of these PSA-detected cases do not express clinically, are slow growing, and do not require treatment, and therefore do not contribute significantly to overall mortality. Conversely, some slow growing cancers are very aggressive and result in death. 

Treatment for prostate cancer can have significant negative impact on quality of life and healthcare costs, and should only be utilized when the cancer itself is likely to be fatal. Treatment recommendations are based on PSA levels , clinical staging, and Gleason scoring. Active surveillance is a preferred approach when the disease is low-risk and small. Significantly, 5-10% of individuals with low-risk disease treated up-front experience poor outcomes. Additionally, >40% of active surveillance patients may progress and require treatment – and half of those will ultimately fail treatment. The effectiveness of active surveillance is limited without a clinical tool to accurately assess risk of progression.

In small pilot studies, Dr. MacAulay’s lab has demonstrated the ability to predict aggressive behaviour in prostate cancers with >80% accuracy using a specific imaging technology that uses the measurement of GOALS in individual cells along with the cell’s position within the patient’s tissue.

DNA-PK inhibitors for use in combination with radiation therapy

Half of all cancer patients receive radiation therapy, impacting about seven million people worldwide each year. Enhancing tumour sensitivity to radiotherapy would have a far reaching and significant impact on patients with many kinds of cancer. 

Funded by a $5M grant from the Wellcome Trust, Dr. Minchinton’s lab has developed novel inhibitors of DNA-repair that can dramatically enhance the elimination of cancer cells with radiotherapy. He will improve his previously developed small molecule inhibitors of a DNA repair protein by developing therapeutic regimens to optimize their use for maximum anti-cancer benefit and minimize their effect on normal tissue. The overall aim of the project is to identify optimized inhibitors suitable for clinical candidate evaluation.

After the preclinical work, Dr. Minchinton will seek corporate partners to take the candidate into full clinical evaluation involving Phase I through III clinical trials. DNA damage repair mechanisms as a route to improved therapy could have a significant impact on the effectiveness of radiotherapy for cancer treatment.

Evolutionary determinants of treatment resistant high grade serous ovarian cancer investigated at single cell resolution

Women diagnosed with high grade serous ovarian cancer (HGS) continue to face poor prognosis, with ten-year survival at only 30-40%. Surgical cytoreduction followed by platinum and taxane-based chemotherapy result in clinical remission for a majority of patients. However, up to 80% of patients will suffer relapse because their disease is treatment resistant. Improved outcomes for HGS require both biomarkers of treatment resistance and development of additional treatments targeting tumour cells resistant to first line therapies.

Relapse in HGS is thought to result from the emergence of resistant tumour clones that evolve de-novo or are selected for during treatment. Dr. McPherson will leverage state-of-the-art single cell genome sequencing technologies to study the genomes of treatment naive primary and metastatic HGS samples, in addition to patient derived xenografts subjected to chemotherapeutic agents. His analysis will focus on understanding the genomic changes that confer treatment resistance, and the evolutionary dynamics that produce those changes.

An improved understanding of the mechanisms and dynamics of treatment resistance will result in improved ability to identify patients with relapse potential and provide targeted therapies to improve survival in HGS.


Towards individualized treatment for pancreatic ductal adenocarcinoma (PDAC)

Pancreatic cancer kills almost 5,000 Canadians each year and if progress is not made to improve outcomes, the annual number of deaths will double by 2030. In 80% of patients, the cancer has spread at the time of diagnosis, and is not operable. Most of these patients die within one year due to the lack of effective therapies and the fact that clinicians have no clear guidance on which existing treatment option would work best for individual patients.

Precision medicine in cancer has gained a lot of attention in the last decade, as it may provide the best approach to treating tumours on an individual basis. Cancer treatment does not benefit from the one-size-fits-all approach because individual tumours, even if affecting the same organ, are biologically different, which can impact their response to treatment. Tumour subtyping, a method by which scientists identify the unique characteristics of individual tumours, is critical for precision medicine enabling personalized treatment based on the tumour's specific biological traits. Advances in the understanding of cancer subtypes have revolutionized treatment in multiple cancers, but we have yet to uncover pancreatic cancer subtypes that can help with treatment decisions.

Our goal is to define clinically meaningful pancreatic tumour subtypes, and study their impact on tumour aggressiveness and response to treatment. These findings will be rapidly translated to the clinic to have immediate impact on treatment selection for patients. We will perform detailed genetic and molecular analysis of patient tumour samples to investigate the distinct molecular characteristics. The patients will be enrolled in a clinical trial at the BC Cancer Agency and will be provided with detailed and cutting edge analyses of their tumours to help the clinical team guide further therapy decisions. 

Currently, over 90% of diagnosed pancreatic cancer patients are not expected to survive five years. Our program has the potential to dramatically change the trajectory of pancreatic cancer and improve outcomes for thousands of Canadians diagnosed with the disease.

Determining the virulence determinants of Fusobacterium nucleatum to define diagnostic and therapeutic targets for colorectal cancer

A substantial portion of the cancer burden worldwide is attributable to infectious agents (viruses or bacteria). Some of these can directly cause cancers, others can facilitate cancer development, and the rest may have no causative role but their existence can indicate the presence of a cancer or risk of developing one.

Recently, Fusobacterium nucleatum, a bacterium present on mucosal surfaces, has been found to be highly elevated in a subset of colorectal cancers. F. nucleatum is an invasive bacterium that can cause acute oral and gastrointestinal infections and can act as a pro-inflammatory agent, thus it is a reasonable candidate for having a facilitating role in tumorigenesis. However, F. nucleatum is also well recognized as a benign resident of mucosal surfaces in the absence of pathology. The reason why F. nucleatum may in some cases be pathogenic and at other times an apparently benign, commensal organism is not yet completely understood.

The overall goal of this study is to identify gene(s) associated with F. nucleatum virulence, and to determine how expression levels of these genes are modulated during infection using RNA-Seq. The Canadian Cancer Society estimates that currently 12 percent of all cancer deaths in Canada are attributed to colorectal cancers; a tendency toward late diagnosis indicates a dire need for simple strategies to help detect colorectal cancers early. The finding that F. nucleatum is strongly associated with a significant number of colorectal cancers cases raises the possibility of developing a simple diagnostic pre-screen for the disease, enhancing early detection rates. The proposed work will identify the F. nucleatum genes that are associated with the disease, creating a signature that will markedly increase specificity of new screening tests. Moreover, this study will indicate how pathogenic F. nucleatum strains cause disease, dramatically increasing our knowledge of this enigmatic bacterium and its interactions with host cells that lead to oncogenesis.

Armed with this new knowledge, it will be possible to develop novel diagnostics, and create new tools such as vaccines to combat, and even prevent, infection. Knowledge translation activities for this study will include presenting results at conferences, writing papers and building on the network between the BC Cancer Agency and our anaerobic bacteriology collaborators at the University of Guelph.

Toward personalized immunotherapy: defining mechanisms of immune suppression across the molecular subtypes of ovarian cancer

Ovarian cancer affects approximately 1,700 women per year in Canada. Current treatment involves surgery and chemotherapy, which is initially effective in most cases. However, most patients relapse with chemotherapy-resistant tumors within a few years of treatment; this highlights the urgency for new, effective treatment strategies. Encouragingly, the immune system has a strong influence on survival in ovarian cancer. Tumors that are densely infiltrated by T cells (a type of immune cell) are linked to improved prognosis. However, a large proportion of patients lack dense T cell infiltrates. Instead, T cells are trapped in the surrounding stromal regions of the tumor and fail to make direct contact with tumor cells.

I hypothesize that the infiltration of T cells is inhibited by suppressive mechanisms in these stromal regions and with better understanding, these mechanisms can be reversed by immunotherapy. One objective of this project is to determine whether T cells that are trapped in stromal regions are capable of recognizing tumor cells. If so, then these T cells have the potential to recognize and eradicate tumors. Another objective is to identify and then block the signals by which stromal cells carry out suppressive functions. I will assess the effects on T cell infiltration and tumor regression following this blockade. This project will facilitate the development of new treatments that release T cells from the suppressive effects of stroma to launch more powerful attacks against ovarian cancer and related malignancies. The possibilities of using off-patent fibrosis drugs for cancer treatment will be investigated; this might result in an inexpensive, effective new form of immunotherapy, thus reducing costs and increasing the number of patients benefitting from these approaches. Since the BC Cancer Agency’s Deeley Research Centre (BCCA-DRC) is able to perform clinical trials, the work can be directly implicated into clinical research.

This research will be presented at both national and international conferences and published in international peer-reviewed journals. The BCCA-DRC’s clinical trials program will also provide me with ongoing opportunities to speak to patient support groups, clinicians, and lay audiences at forums focused on education, awareness and philanthropy.

A comprehensive screen for oncogenic microRNA mutations in an acute myeloid leukemia cohort and across the Cancer Genome Atlas

Acute myeloid leukemia (AML) is a cancer in which blood cells grow out of control. Blood cells have to suffer at least two mutations to become cancerous: one to make them grow faster, and another to stop them developing normally. However, even with whole genome sequencing, in some patients we have been unable to find both mutations using existing methods, and we need to look deeper.   

MicroRNAs are one place we can look. These are small pieces of RNA which reduce the production of proteins by targeting specific messenger RNAs. We know that cancers tend to have more or less of some microRNAs, and that many of these play important roles in cancer biology. However, whole-genome studies have mainly looked at the amounts of well-known microRNAs, without looking deeply at mutations of the microRNAs themselves, which can completely change their targets. Smaller studies have shown that microRNA mutations (as well as normal variations between people) can be important drivers of cancers, but nobody has investigated these at the genome-wide scale.    

I will examine mutations of microRNAs in the genomes of around 200 AML and myelodysplastic syndrome patients. I will measure the effects of each mutation on messenger RNA levels. I will then look especially in patients in which two driver mutations could not be found to see whether any microRNA mutations could be oncogenic. The results will increase our understanding of the biology of AML, thereby leading to new research into improved therapies. They will also improve our ability to diagnose AML, which will give more information to doctors and patients making difficult decisions on treatments.    After analysing our local dataset, I intend to similarly analyse all cancers in the Cancer Genome Atlas (TCGA) data set. Since the microRNA sequencing for the TCGA was performed at the Michael Smith Genome Science Centre in Vancouver, this is an excellent opportunity to extract further value from a locally-produced resource.   

For knowledge translation activities, I intend to present this work at the annual meetings of the American Society for Hematology and the International Society for Computational Biology. Further, I will write up the AML analysis for submission to Genome Research or Leukaemia, and the later work applying the method to the TCGA data to a similar (or higher-impact) venue. Lastly, I will release the source code to perform the analysis as an open source software package.

Upstream regulators of the Hippo signaling pathway in liver development and cancer

Liver cancer is the fifth most common cancer and the leading cause of cancer deaths worldwide, primarily because of late diagnosis and scanty therapeutic options. Animal studies have demonstrated that the Hippo intracellular signalling pathway is critical in regulating liver size and liver cell fate and is a potential tumour suppressor in the liver.

Loss of cell-cell adhesion is associated with the progression and poor prognosis of liver cancer. In this project, we will explore whether loss of cell-cell adhesion regulates Hippo signalling in liver cells. We are particularly interested in how loss of cadherin molecules can regulate the Hippo signalling pathway and subsequently contribute to liver development and cancer. The findings generated from this proposal will increase our understanding of the underlying molecular mechanisms involved in liver development and tumourigenesis.

Physical activity, sedentary behaviour and gene-environment interactions in cancer

Dr. Boyle’s research will investigate the role that physical activity and sedentary behaviour play in the development of non-Hodgkin lymphoma, multiple myeloma, and breast cancer.

This project aims to: 1) examine the associations between physical activity and sedentary behaviour and the risks of non-Hodgkin lymphoma and multiple myeloma, and 2) investigate whether the effects of physical activity and sedentary behaviour on the risks of non-Hodgkin lymphoma and breast cancer are modified by particular genes.

These research questions will be investigated using data from four case-control studies (three of which were conducted in British Columbia), as well as pooled data from an international consortium of case-control studies.

This research will provide new and important information about the associations between physical activity, sedentary behaviour, and cancer. Identification of modifiable risk factors is particularly important for the prevention and control of these cancers, as little is known about their etiology.