Understanding how an MLH1 promoter polymorphism predisposes to colorectal cancer

Lead Research Organisation: University of Oxford
Department Name: Wellcome Trust Centre for Human Genetics

Abstract

Colorectal cancer affects 1 in 20 people in the western world. Mutations that cause colorectal cancer can be inherited or, more often, acquired during a patient's lifetime. There are also some common genetic variants (small differences found in the DNA of many people) that can increase their risk of cancer by a small percentage. These variants usually do not have a big effect on the gene, but may influence how active the gene is in particular cells. One such colorectal cancer risk variant is found near the MLH1 gene. MLH1 is disrupted in about 15% of colorectal cancers, but instead of mutations in its DNA, it is repressed by acquired chemical marks on the DNA (called DNA methylation or epimutation). The common variant near MLH1 is strongly linked to increased DNA methylation and MLH1 gene repression, but we do not yet understand exactly how and why this leads to a higher risk of this specific type of colorectal cancer. The main aim of this project is to answer these questions.

We will work out how the common genetic variant near MLH1 causes different risks of MLH1 repression. We think that 1) the high risk variant binds to proteins that cause DNA methylation and that this represses MLH1 activity, or 2) the high risk variant disrupts the binding of proteins necessary to keep MLH1 active, reducing MLH1 expression and allowing methylation to spread across the DNA causing further reduction in expression, or 3) the high risk variant just affects the binding of activating proteins and that the less active gene is more easily silenced by DNA methylation than occurs by chance.

We will test these alternatives by
1) studying the genetic variant and MLH1 methylation and repression in a large number of patients with colorectal tumours
2) using cell lines to look for proteins which bind the DNA preferentially on one of the variants, and to study how MLH1 methylation and repression occur over time
3) finding out how the genetic variant influences cancer growth in mouse models of the human disease

The knowledge from these studies will help us to understand the effects of MLH1 disruption on tumours and ultimately provide information to help with patient diagnosis, prognosis and therapy. New treatments which remove methylation in tumours are being developed, and our findings, and the cell and mice models we create, should help to determine if and when these therapies could be used to treat colorectal cancer patients with MLH1 repression.

Technical Summary

Microsatellite-unstable (MSI+) tumours are commonly found in the colon where they comprise ~15% of all cancers. The underlying cause is mismatch repair (MMR) deficiency which in the majority of cases is caused by MutL homologue 1 (MLH1) promoter methylation. Why MLH1 is inactivated rather than other MMR genes and why methylation occurs rather than gene mutations are unknown, although there may be a causal link to the putative CpG Island Methylator Phenotype (CIMP). Intriguingly, a polymorphism, rs1800734, lies within the MLH1 promoter upstream of the translational start site of MLH1, and this SNP shows a strong association with the risk of MSI+ colorectal cancer (CRC), but not with MSI-negative cancers.

We shall
1)expand the analysis of the association between rs1800734 and CRC risk, and mine existing data for rare germline genetic variants near rs1800734
2)determine associations between rs1800734 genotype, MLH1 methylation, MLH1 expression and CIMP in human normal, benign and malignant colorectal samples and find the statistically most likely causal links
3)construct isogenic CRC cell lines with all three rs1800734 genotypes
4)investigate chromatin structure and transcription factor binding to identify molecules and processes differentially associated with rs1800734 alleles
5)modulate DNA methylation and follow subsequent methylation (re-)acquisition and consequences according to rs1800734 genotype.
6)repress MLH1 mRNA transcription and determine whether promoter methylation occurs as a consequence
7)set up a mouse model of rs1800734 and determine effects of each SNP genotype on MLH1 methylation and intestinal tumours in a Braf-mutant model of MSI+ carcinogenesis
8)if time permits, develop a mouse with a general tendency to MLH1 methylation and determine the influence of rs1800734 therein
These experiments will identify many of the mechanisms underlying the association of rs1800734 with MSI+ CRC and provide broader insights into MMR-deficient canc

Planned Impact

ACADEMIC IMPACTS
1) RESEARCH DELIVERABLES
(Please refer to section on Academic Beneficiaries for full list of research users). The deliverables will relate to the MLH1 gene and the SNP rs1800734 including its methylation, gene expression, role in tumorigenesis, association with colorectal cancer risk, mutation selection, cancer evolution, general epigenetic and gene regulation mechanisms and may provide information useful for medical genetics, screening, diagnostics and therapeutics. The proposal will directly relate to the field of colorectal cancer biology and more broadly the cancer, epigenetic, gene regulatory research fields and cancer medicine.
2) TRAINING THROUGH RESEARCH
The research assistant will be given full laboratory training in technical skills, project design, data analysis, manuscript preparation and critical literature review. Internal and external training courses and opportunites to work with collaborators will be encouraged whenever appropriate, equipping the post holder for future higher degrees or independent research.
3) BIOLOGICAL AND METHODOLOGICAL RESOURCES
The isogenic cell lines and transgenic mice created to study rs1800734, MLH1 and cancer predisposition, and any additional biological resources created during the project, will be an asset to those in the colorectal cancer field and will be made available to the research community after publication. New or improved methodology including developments of the CRISPR/Cas9 system may be useful to the wider community as described above and will be shared and published.
SOCIO-ECONOMIC IMPACTS
4) HEALTH AND MEDICINE
rs1800734 is known to increase colorectal cancer risk and an increase in understanding may help to decide its importance in future screening, diagnostic or treatment programmes. In addition, the importance of and ability to manipulate methylation at the MLH1 locus could have long-term prospects for therapeutics.
5) ENHANCING SKILLS OF MEDICAL AND SCIENCE STUDENTS
The University of Oxford expects medical and basic science students to undertake research projects as part of their course. This proposal would allow the PI to host and train undergraduate and masters students from Oxford and other leading universities.
6) PUBLIC ENGAGEMENT WITH RESEARCH
Cancer research is able to attract the attention of the general public. We will take available opportunities to explain the importance of basic research in the advancement of cancer screening, diagnosis and treatment.
 
Description IMPC- Understanding somatic MSI+ colorectal cancers using conditional coding and non-coding mutations in Mlh1
Amount £33,040 (GBP)
Organisation MRC Harwell 
Sector Academic/University
Country United Kingdom
Start 01/2018 
End 08/2018
 
Title The analysis of CRISPR-Cas9 off target effects using cytogenetic methods 
Description CRISPR-Cas9 is a powerful new technology for gene editing but adequate controls and methods for discovering off-target effects are not yet in regular use. We have developed quick and cost effective cytogenetic methods for identifying chromsomal abnormalities and large deletions. The methods include staining metaphase spreads followed by chromsomal counting and fluorescence in situ hybridisation using BAC probes at and flanking the CRISPR target locus. We find that these unwanted mutations are not readily detectable by Sanger sequencing but occur regularly in cancer cell lines, particularly those that already exhibit high levels of chromsomal instability. We have proposed in our publication in the CRISPR Journal that these methods should be routinely used in novel CRISPR mutated models to control for off target effects. 
Type Of Material Technology assay or reagent 
Year Produced 2019 
Provided To Others? Yes  
Impact Our paper describing this method was accepted for publication in The CRISPR Journal (December 2019). We have also disseminated our findings locally by personal communication with other researchers using CRISPR within the Welcome Centre for Human Genetics, University of Oxford and within the Department of Life Sciences, Brunel University. One hypothesis arising from this work is also that CRISPR-Cas9 might be used to cause an excess of chromosomal rearrangments in cancer cells with existing instability and we are planning preliminary experiments to test whether these effects could be severe enough to cause cell death. 
 
Title The creation of a mouse model of mismatch repair deficient colorectal cancer 
Description We have generated a mouse model of mismatch repair deficient colorectal cancer by breeding mice containing a conditional BRAFV600E oncogenic mutation with MLH1 conditional knockout mice. These mutations have been activated specifically in the intestine and result in increased numbers of polyps and faster growth than those found in mice carrying the BRAFV600E mutation alone. Preliminary analysis shows that these polyps are have microsatellite instability. We need to carry out further analysis but it is likely that this model will prove valuable for investigating the effects of exisiting and novel therapeutic technologies on intestinal cancers with microsatellite instablity. We are particularly interested in the mechanism by which immunotherapy successfully treats these cancers. Our collaborators, Dr David Church and his group are generating further MLH1 mouse crosses with APC mutated and POLE mutated strains. They are also developing methods to assess and compare immune activity in tumours from all three crosses. 
Type Of Material Model of mechanisms or symptoms - mammalian in vivo 
Year Produced 2019 
Provided To Others? No  
Impact We are still working on our analysis of this model and a comparison with other similar mouse strains. We expect the model to be useful for investigating the effects of exisiting and novel therapies on mismatch repair deficient colorectal cancer. 
 
Title Transgenic mouse models to study the MLH1 tumour surpressor gene 
Description The MLH1 gene is epigenetically silenced in a significant subset (approx 15%) of sporadic colorectal cancers. Previous work and literature have indicated a polymorphic variant in the promoter of MLH1 may influence the accumulation of DNA methylation and subsequent transcriptional silencing. We have created 4 transgenic mouse lines to model this: 2 replacing the reference low risk allele with the high risk, a single base pair deletion of the SNP and a 20 base pair deletion including the SNP. 
Type Of Material Model of mechanisms or symptoms - mammalian in vivo 
Provided To Others? No  
Impact We are still at an early stage of analysing the mice. 
 
Title Using next generation sequencing to detect allele specific biases 
Description We have used Illumina Miseq technology to accurately determine allele-specific differences in gene expression, DNA methylation and protein binding. This method is sensitive and reproducible, and can be used in cell line and animal models and patient samples. 
Type Of Material Technology assay or reagent 
Year Produced 2018 
Provided To Others? No  
Impact This has enabled us to measure the effect rs1800734 genotype on intermediate cancer phenotypes to uncover the mechanism by which it predisposes to colorectal cancer. 
 
Description Collaboration with Eric O'Neill's group in the Oxford Institute for Radiation Oncology, including cosupervision of PhD student, Michael Eyres 
Organisation University of Oxford
Department Gray Institute for Radiation Oncology and Biology
Country United Kingdom 
Sector Academic/University 
PI Contribution I have provided assistance in designing and creating mouse transgenic lines and training and ongoing supervision in mouse model experimental design and procedures.
Collaborator Contribution The O'Neill group has provided information and resources allowing me to initiate experiments to target de novo DNA methylation to specific target sites using Cas9/CRISPR technology.
Impact The O'Neill lab has generated 1 new transgenic mouse line with another currently at the chimera stage.
Start Year 2016
 
Description Daniela Furlan 
Organisation University of Insubria
Country Italy 
Sector Academic/University 
PI Contribution Analysis of DNA and RNA from colorectal cancer samples to investigate allele specific MLH1 expression and promoter methylation
Collaborator Contribution Sharing of patient samples and scientific expertise
Impact Data still under analysis
Start Year 2018
 
Description David Church and laboratory 
Organisation University of Oxford
Department Nuffield Department of Medicine
Country United Kingdom 
Sector Academic/University 
PI Contribution We have generated a mouse model of microsatellite unstable colorectal cancer by breeding mice containing a conditional BRAFV600E oncogenic mutation with MLH1 conditional knockout mice. These mutations have been activated specifically in the intestine and result in increased numbers of polyps and faster growth than those found in mice carrying the BRAFV600E mutation alone. Preliminary analysis shows that these polyps are have microsatellite instability.
Collaborator Contribution Dr Church and his group are generating further MLH1 mouse crosses with APC mutated and POLE mutated strains. They are also developing methods to assess and compare immune activity in tumours from all three crosses.
Impact A new mouse model of MSI+ colorectal cancer with the potential for two further models of hypermutated colorectal cancer. This project is very much a work in progress at the moment with encouraging preliminary results.
Start Year 2019
 
Description Ian Tomlinson Birmingham 
Organisation University of Birmingham
Country United Kingdom 
Sector Academic/University 
PI Contribution Analysis of RNA and DNA from patient samples collected by Ian Tomlinson to investigate allele specific MLH1 mRNA expression and promoter methylation
Collaborator Contribution Sharing of patient sample sets
Impact Data still under analysis
Start Year 2017
 
Description Timothy Maughan and the S:CORT consortium 
Organisation University of Oxford
Department Department of Oncology
Country United Kingdom 
Sector Academic/University 
PI Contribution We have used samples and data collected by the S:CORT consortium to analyse the effect of the SNP rs1800734 on MLH1 expression and methylation in early colorectal lesions
Collaborator Contribution Providing data and DNA and RNA from early colorectal cancer lesions as well as regular meetings offering support and advice on this sub-project and the project as a whole.
Impact Work from this collaboration forms part of a manuscript currently under review at Scientific Reports.
Start Year 2017
 
Description Tracy O'Mara and other members of the ECAC consortium 
Organisation QIMR Berghofer Medical Research Institute
Country Australia 
Sector Academic/University 
PI Contribution Using patient genotyping data provided by Tracy O'Mara and colleagues, we have carried out a candidate SNP analysis on the effect of rs1800734 on the risk of mismatch repair deficient endometrial cancer. We have shown that, in contrast to our published work on colorectal cancer (Thomas et al, 2019), this SNP has no effect on endometrial cancer risk which ties in with our own in vitro studies. We now have a manuscript including this data ready for submission to Clinical Epigenetics
Collaborator Contribution T racy A. O'Mara, Daniel D. Buchanan, Emma Tham, Miriam Mints, Anne Keränen and Amanda B. Spurdle have previously collected and analysed several large cohorts of endometrial cancer patient genotyping data for use in genome wide association studies. For this collaboration they provided additional analyses on the mismatch repair status of each patient allowing the cohorts to be stratified. This allowed us to carry out the analysis of the effect of rs1800734 risk on MMR deficient endometrial cancer which now forms part of a manuscript in preparation.
Impact A manuscript ready for submission to Clinical Epigenetics
Start Year 2019
 
Description Research presentation for Disease Mechanisms Master's students, Brunel University 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Postgraduate students
Results and Impact 9 Masters students and 3 invited undergraduate students attended a 2 hour presentation/workshop based on my long term research and including work done under this grant.
Year(s) Of Engagement Activity 2020
 
Description Speaker at Alumni symposium at St Catharine's College Cambridge 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Postgraduate students
Results and Impact The symposium was intended to bring together researchers from a wide range of disciplines to discuss the nature and impact of their work with a non-specialist student audience. The main aim was to inform and encourage discussion with the audience.
Year(s) Of Engagement Activity 2018
URL https://mcr.caths.cam.ac.uk/symposium