Characterisation of Chromatin Landscapes of Pre-leukaemic and Leukaemic Stem Cells in Core Binding Factor AML and their Response to Epigenetic Therapy

Lead Research Organisation: University of Birmingham
Department Name: Institute of Cancer and Genomic Sciences


Acute Myeloid Leukaemia (AML) is a relatively common blood cancer that forms a significant proportion of my clinical work as a Haematologist. Whilst intensive chemotherapy regimens (which have changed very little since the 1980s(!) can achieve temporary clinical remission in about 80% of patients, 5 year survival is only 15% without going through risky procedures such as a stem cell transplant. Consequently, a key clinical question is to understand on a genomic level what causing all these patients to relapse, with a view to prevention.

RUNX1 is a gene that is of utmost importance in blood cell development and has been dubbed in its research community as the "master regulator of blood cell development". We have worked on understanding this gene and its interactions for many years in our group and we have considerable expertise. Mutations in RUNX1 also happen to be one of the commonest defects in AML. We aim to tie our expertise with this gene with the clinical problems of relapse in AML and aim to tackle 4 key objectives:

1) Characterise Pre-leukaemic Stem Cells
Chemotherapy is a rather blunt instrument that kills mainly the fastest-dividing cells. 'Pre-leukaemic stem cells' are blood cells with genetic changes that can eventually give rise to cancer but aren't yet fully cancerous. We propose that these cells are slow-dividing and hence resistant to chemotherapy. They can therefore survive in patients who are apparently in remission and lead to future relapses. We aim to purify pre-leukaemic stem cells from bone marrow of patients with RUNX1 mutations. We will then use state-of-the art technology such as single cell RNA sequencing to understand their biology.

2) Improve AML monitoring in Remission
Patients that achieve clinical remission are monitored for relapse by looking at the number of copies of a gene specific to their disease. However we still do not have good markers that would predict relapse. By characterising pre-leukaemic stem cells, this work aims to identify clinically relevant genes targets that can help patients be monitored better.

3) Investigate the Role of the Wilms Tumour 1 (WT1) Gene
We already know from the literature that patients that have high levels of the WT1 gene relapse quickly. WT1 has traditionally been thought of by cancer researchers as a gene protective against cancer so we aim to investigate why it harmful in AML. We will perform experiments on cell lines to find out where in the AML genome WT1 acts and other molecules it interacts with. We will also selectively alter levels of the gene in order to determine its role in relapsed leukaemia.

4) Investigate Response to Treatment by Epigenetic Drugs
At Birmingham, we have led a phase II clinical trial (RAvVA - currently the largest epigenetic therapies trial of its kind) which recruited 259 patients that were treated with combinations of 2 drugs that alter gene expression - Azacitidine and Vorinostat. However, it is not currently known what exactly these drugs are doing on a genomic level on the AML cells. We aim to elucidate this by comparing changes in gene expression in patients for whom these drugs have benefited compared to those for who they haven't. This can potentially highlight specific genes or signalling pathways to target that might give a more durable clinical response.

Technical Summary

Acute Myeloid Leukaemia (AML) is a clonal disorder of the myeloid lineage. Whilst up to 80% of patients treated with intensive chemotherapy can achieve a clinical remission, most will eventually relapse and die.

One type of AML, caused by mutations in the genes encoding the haemopoietic master regulator RUNX1 or its binding partner CBFB, is defined by the generation of RUNX1-ETO (in t(8;21) AML) or CBFB-MYH11 (in Inv(16) AML) fusion genes as a result of genomic translocations. There are also mutations in the RUNX1 gene itself. Monitoring translocation patients during remission shows that those with detectable fusion transcripts by RT-PCR are more likely to relapse. We hypothesise that these transcripts originate from pre-leukaemic or leukaemic stem cells (LSCs) that are insensitive to chemotherapy. We will purify LSCs from the bone marrow cells from AML patients at diagnosis and characterise their epigenome using ATAC-Seq and RNA-Seq. Single cell RNA-Seq will examine cellular heterogeneity. We will use this information as well as data from pre-leukaemic cell line models to examine enriched cells from patients in remission to identify additional markers for the presence of minimal residual disease.

We also know from the literature and from our preliminary data that the transcription factor WT1 is upregulated in CBF leukaemia and that such patients relapse earlier. To gain an understanding how RUNX1-ETO and WT cooperate, we will map global WT1 binding sites, knock it down or overexpress it to assess how WT1 contributes to the RUNX1-ETO specific transcriptional deregulation.

Finally, at Birmingham, the phase II RAvVA trial led by my supervisor Prof Craddock, looks at the efficacy of the epigenetic drugs Azacitidine and Vorinostat in patients with AML. By profiling the chromatin landscape of serial marrow from patients with RUNX1 mutations recruited to the trial, we will gain a unique insight into disease evolution and the emergence of resistance to therapy.

Planned Impact

I believe that this project has a significant impact and will have an impact on several groups:

First and foremost, the beneficiaries of this project will be patients with Acute Myeloid Leukaemia. The relapse rate is so high that even amongst patients that undergo procedures as drastic as an Allogeneic Bone Marrow Transplant, survival is only 40% at 4 years. The aim of this project is to understand the biology of pre-leukaemic stem cells with the hope that they can be targeted with therapies and prevent relapse.

The United Kingdom/Government Policy
There is growing emphasis upon personalised medicine in the NHS as evidenced by policy such as the 'Five Year Forward View' and Sir Bruce Keogh's 'Personalised Medicine Strategy'. It has been recognised in such policies that currently available drug treatment only benefits 30-60% of patients they are prescribed yet the current NHS drugs budget is huge at £12 billion a year. The cost of treating patients for the side effects of ineffective medication as well as the cost of treating patients with more advanced disease is also significant but hasn't been quantified. Consequently, there is a significant push for tests specific to a patient that can help guide treatment. There has also been investment in the 100,000 genomes projects to similarly better involved the role of genomics in disease. The fact that our data will be in the public domain means that other researchers will mine them in order to obtain more insight.
This project is absolutely compatible with such nationally identified priorities. It will identify the epigenetic landscape of pre-leukaemic cells in patients with AML. Together with studying the effects of epigenetic therapies from the RAvVA trial, this will greatly enhance our understanding how genomics contributes to disease by giving us clinical correlates of how modifying a particular gene's expression affects patient outcome. This should elucidate new and personalised drug targets.

Pharmaceutical Industry
In the long-run, this may well lead to collaborations with the Pharmaceutical Industry for creating new epigenetic therapies against newly identified targets. A recent example of this was the development of I-BET151, a specific epigenetic therapy that will revolutionalise the treatment of Mixed Lineage Leukaemia, by collaboration between the University of Cambridge and GlaxoSmithKline.

Improving Cancer Monitoring Services
Patients in remission from AML are monitored if they have an identified genetic marker of disease such as a RUNX1-ETO translocation. However, over 40% of patients do not have an identifiable genetic marker to monitor and so we often do not notice these patients until they have floridly relapsed. In addition, we do not know if we are monitoring the correct genetic markers and having discussed this with several of my senior clinical genetics colleagues, this is a cause of considerable anxiety. This study aims to find that exact genes that are aberrantly expressed and that correlate with clinical disease. Our strategy of identifying the gene expression profiles of pre-leukaemic and leukaemic stem cells will hopefully help to identify more markers.

Personal skills for future
I do not underestimate the personal benefit this project would have for me. Whilst I have 11 years of lab experience (part-time with my clinical medicine activities), this PhD fellowship would give me both dedicated time to focus upon my research as well as give me world-class training in 21st century medicine and "omics" technologies. I am determined to be a clinician-scientist in the future, despite the huge challenges of being excellent clinically and academically. By having a really good PhD training, I feel that this can be a really good springboard to achieving this aim.


10 25 50
Description Collaboration with Nick Hastie, University of Edinburgh 
Organisation University of Edinburgh
Country United Kingdom 
Sector Academic/University 
PI Contribution We are knowledgeable about molecular techniques
Collaborator Contribution Give advice about WT1
Impact None yet
Start Year 2017
Description Genetics Lab Collaboration 
Organisation West Midlands Regional Genetics Laboratory
Country United Kingdom 
Sector Public 
PI Contribution Providing ideas to them about minimal residual disease monitoring and biomarkers.
Collaborator Contribution Getting bone marrow samples from our collaboration here.
Impact Preliminary data for an MRC CRTF grant
Start Year 2014