Two young girls with acute lymphocytic leukemia (ALL) receiving chemotherapy. They are sitting on a bed and are demonstrating some of the procedures and techniques used with intravenous chemotherapy. The girl on the left has an IV tube in the neck, the other girl’s IV is in her arm, which has an arm board which stabilizes the arm for IV needle insertion. Credit: Bill Branson (Photographer), Public domain, via Wikimedia Commons.
New research from Sweden has found DNA patterns can provide key insights into the diagnosis of childhood leukaemia. The research suggests treatment of leukaemia among children could be made more precise and potentially have fewer side effects with the help of more accurate methods of diagnosis.
The study, led by Umeå University, Sweden, reveals that analysing methylation patterns in the child’s leukemic cells DNA can enhance risk assessment, helping to ensure that only the children who need the most intensive treatment receive it.
ScienceVega spoke to Fernanda Schäfer Hackenhaar, who contributed to the study’s bioinformatic analysis, and study lead, Professor Sofie Degerman, both of Umeå University, to find out more.
ScienceVega: What is leukaemia?
SD: Acute Lymphoblastic Leukaemia (ALL) is the most common childhood cancer. This blood cancer originates from different bone marrow progenitor cells of B or T-cell type in the bone marrow which have gained genomic aberrations resulting in uncontrolled growth of these cells. The disease is treated with intensive chemotherapy and sometimes with stem cell transplantation. B-cell lymphoblastic leukaemia (B-ALL) accounts for 85% of cases, with T-cell lymphoblastic leukaemia (T-ALL) the much rarer variety in 15% of all paediatric ALLs. In Sweden, 60-90 out of 1,600,000 children are diagnosed with ALL each year, with around 10-15 among those diagnosed with T-ALL.
ScienceVega: What is the current approach when it comes to a child presenting with leukaemia today?
SD: In general when a child is diagnosed with T-ALL, a period of intensive chemotherapy starts that goes on for almost two years in different blocks of treatments that have different content; they are given a mixture of different drugs. The specified ways in which these drugs are given has improved survival, and almost 90% of children with the disease survive. But still, not all children survive and the short- and long-term side effects can be severe due to these intensive treatments. Given these are treatments for children, it is crucial we avoid the long-termd side effects because often patients will need to live with those for their whole life.
T-lymphoblastic cells of acute leukaemia in the bone marrow. In some cases, the cytoplasm is concentrated at one pole of the cell, forming ‘hand mirror cells’. Credit: Animalculist, CC BY-SA 4.0 <https://creativecommons.org/licenses/by-sa/4.0>, via Wikimedia Commons.
ScienceVega: What kinds of side effects can patients experience when they undergo these treatments?
FSH: It depends. There are long-term side effects and there are short-term side effects. Infections are a very common side effects because the chemotherapy supresses the immune system.. Of the long-term side effects, infertility is one of the most important. And because the chemotherapy affects the bone marrow of children there is a risk of damage to the bones which might affect growth.
SD: The treatment of ALL is associated with several side effects which most patients experience to some extent including, for example, susceptibility to infection, nausea, seizure, fatigue, and hair loss.
Unfortunately, a significant proportion also experience one or several long-term or late effects from their treatment after completing the treatment. These could be cardiac issues, cognitive (mental or learning) issues, tiredness (fatigue), fertility problems or secondary cancers.
Avoiding these side effects is crucial, and there is a need for novel molecular biomarkers that can support personalised treatment strategies by tailoring therapy intensity based on relapse risk in T-ALL.
DNA methylation profiling represents one such promising biomarker, with the potential to enhance current risk stratification approaches in the future.
3D Medical Animation still showing an increase in white blood cells of a person suffering from Leukaemia. Manu Sharma, http://www.scientificanimations.com, CC BY-SA 4.0 <https://creativecommons.org/licenses/by-sa/4.0>, via Wikimedia Commons.
ScienceVega: What is DNA methylation?
FSH: It is an epigenetic modification of the DNA that can be acquired in life through chronological ageing and environmental exposures, but it can also be genetically transmitted through generations. DNA methylation patterns can actually be used to predict an individual’s chronological age. So, DNA methylation is of biological importance because it is both inheritable and it can also be acquired.
SD: It is a bit like the ‘fingerprint’ of a cell that provides information about the cell identity. It can also give you information about what that kind of cell has experienced in terms of chronological age or cell divisions or even if something has gone wrong and it has turned into a cancer. Then for different cancers, you can identify unique kinds of fingerprints that tells us which cancer type we are looking at. You can also sometimes see different subgroups within the cancer based on these DNA methylation fingerprints.
DNA methylation is kind of a chemical modification of the nucleotides, which are the building blocks of the DNA. We analyse if this chemical modification is present or not on specific positions in the genomes that are called CpG sites. They are regions of DNA where a cytosine nucleotide is followed by a guanine nucleotide and when you have that specific combination of nucleotides, it’s possible to have methylation on the cytosine.
Those CpG sites are distributed throughout the genome. We know that there are around 30 million of these sites throughout the human genome and they form a pattern that is unique for each cell type. In some regions it’s methylated, in some it’s not, and this is contributing to regulation of gene expression; for example, it can regulate whether a gene is turned on or off. That means it is, in effect, a kind of memory for the cell to know which genes should be turned on, and which should be turned off. Since we have exactly the same nucleotide sequence in all cells, the cell needs to know which one of these genes should be used. For example a liver cell will express certain genes, which will be different to other cell types, and this is reflected in the DNA methylation patterns where we see these unique patterns.
But sometimes this regulation can go wrong and genes can be turned on when they shouldn’t, or can be silenced when they should be turned on. That is what can lead to cancer development or progression.
In 2018, leukaemia accounted for 32.5% of all cancer cases in children aged 0-14 years globally. Credit: AMISOM Public Information, CC0, via Wikimedia Commons.
ScienceVega: How does methylation provide detailed insights about the properties of cancer cells?
SD: In general it provides information about the cell type which it originates from, but it can also provide information about the specific subtype of cancer as our classifier does. It tells us something about the biology which can provide information about the risk of getting a relapse, but we still have a lot of things to learn about the role of DNA methylation patterns and the treatment outcome of leukaemias.
We have several ongoing studies trying to understand the role of DNA methylation and how different DNA methylation patterns can be associated with how you should treat these patients.
ScienceVega: How did you approach your research and what analysis methods did you use?
FSH: Currently we are using DNA methylation arrays. Methylation arrays cannot analyse the whole 30 million CpG sites, but we can analyse almost a million sites, which has been giving good insights into the genome-wide DNA methylation information. For leukaemia, that has been working quite well, but we are interested in using whole genome sequencing strategies to see the whole genome profile of DNA methylation.
A DNA methylation array. Credit: Sofie Degerman/ Umeå University.
ScienceVega: When did you realise this could be helpful in the case of T-cell lymphoblastic leukaemia?
SD: I started looking into this more than ten years ago! We did our first analysis of T-ALL patients in 2013 and at that time we only had access to Swedish paediatric patients, but it was really interesting to see such big differences within the disease group; some of the patients had another methylation profile to the other ones. What we then did was to define a panel of those specific CpG sites, which we called the CpG island methylator phenotype (CIMP) panel, and that panel consists of around 1,000 CpG sites. By using it we can classify each leukemic sample to see if they are CIMP-low or CIMP-high depending on how much methylation they have on those 1,000 CpG sites.
In this study we realised that there are subgroups within T-ALL that have different DNA methylation patterns. We had already looked at prognosis at that time and there were big differences in the outcome for the ones that had a CIMP-low profile and those with a CIMP-high profile. But we only had 43 Swedish patients, which was too few to really make a conclusion.
Luckily this is a very rare disease, but that makes it more difficult as a researcher to be able to study it. Therefore we have to collaborate and we now work with clinicians and scientists in many European countries. This latest study is really a result of team science, which has allowed us to analyse samples from all five Nordic countries as well as the Netherlands. That collaboration has made it possible for us to analyse 348 paediatric T-ALL patients. The reality is that if we had done that study locally in Umeå, it would have taken our entire careers, or longer due to the rare T-ALL diagnosis.
The patients in our study were treated with modern protocols and we could combine our CIMP classifier with the follow up data after the first so-called ‘induction treatment’, which is where a patient gets a treatment for one month and then it’s evaluated to see if there are leukemic cells left. This is what is really driving the future treatment after the 30-day evaluation in terms of how the treatment might continue. If we combine that information with the CIMP classifier, that could further improve the prediction of the outcome for these patients.
ScienceVega: What is specific about T-ALL that means the DNA methylation analysis can have a positive impact on diagnosis and treatment?
SD: First of all we offer a molecular biomarker that can be analysed at the diagnostic stage. You can learn from the biology of the leukaemia and predict if the patient is at risk of getting a relapse. If you have a higher risk or lower risk of getting a relapse, then potentially that could be used as a support to adjust the treatment intensity for the patients.
Of course, it is now up to the clinicians to evaluate if and how this kind of information could be introduced into the clinics because it’s always a big responsibility to reduce the intensity of a treatment if you know it’s effective, but you also have a risk of side effects. It’s a difficult decision for the clinicians to make, but as scientists, we think that this could be a strong biomarker for predicting the risk of relapse.
FSH: We believe in studying epigenetics in T-cell acute leukaemia because if you’re looking into other cancer types, or even B-cell leukaemia, you have genetic alterations such as a mutation in specific genes that give you information about the severity of the disease and how to treat the patient better with more or less chemotherapy.
But T-ALL, until this moment, doesn’t include specific genetic mutations in the treatment stratification. There are mutations, of course, but most of them are not associated with the prognosis of the patients, and it might be that the epigenetics can provide this information.
Professor Sofie Degerman in the lab. Credit: Mattias Pettersson.
ScienceVega: Can the analysis be used in other leukaemias or other conditions?
SD: Epigenetics is already in use in the research and diagnosis of brain tumours. Methylation analysis is already an important support for subgrouping and therapy prediction. But the DNA methylation classifier we have developed, has been developed specifically for T-ALL, so I don’t think it can be used directly on other subtypes.
Instead you can use the same technology but look at other specific CpG sites and have unique classifiers for each of the different diseases.
ScienceVega: What further research would you like to do in this area?
SD: We are looking at a broader set of CpG sites, more genome-wide, to learn if we can use the information of the DNA methylation to further subgroup the leukaemias. We are also looking into other techniques of estimating DNA methylation, such as sequencing techniques. We are also doing some preclinical analysis on cell cultures and mice models to evaluate how you can use different treatment strategies to be as effective as possible when you have different methylation profiles.
In the last year new publications from different scientists have provided new knowledge about the complex biology of T-ALL regarding genetics and transcriptomics, highlighting that there might be many different subgroups of T-ALL. An important step is to integrate information about the DNA methylation profiles with genomic aberrations and gene expression alterations. By combining all these molecular markers, we hope to provide information of how these patients should be treated most effectively.
This is a complicated subject, and it is not yet in the clinic. We hope to get DNA methylation analysis into the clinics, but it’s not yet there.
Sofie Degerman
Professor of Biomedical Laboratory Science
and
