Myeloproliferative neoplasm driver mutations often acquired before birth, in childhood
Click Here to Manage Email Alerts
Patients acquired myeloproliferative neoplasm driver mutations early in life, often in utero, which then expanded at variable rates over their lifetime before diagnosis, according to study results.
These results, presented during the late-breaking abstract session of the virtual ASH Annual Meeting and Exposition, suggest early detection of these clones and calculation of their expansion rates may present an opportunity for preemptive intervention, according to the researchers.
“As physicians, some of the commonest questions we get asked by our patients with blood cancers are, ‘How long have I had it for?’ and ‘How fast did it grow?’ We set out to answer this in myeloproliferative neoplasms,” Jyoti Nangalia, MBBChir, PhD, clinician scientist at Wellcome Sanger Institute and University of Cambridge, said during her presentation.
Polycythemia vera, essential thrombocytopenia and myelofibrosis provide “a unique and tractable model for addressing this question because they give us a window into early cancer development, where normal blood production coexists alongside the cancer clone,” Nangalia added. “Importantly, these cancers are very well genetically characterized, with over 90% of patients harboring mutations in JAK/CALR/MPL, and 60% of patients have mutations in additional cancer-associated genes.”
Nangalia and colleagues evaluated data of 10 patients who presented with a myeloproliferative neoplasm when aged 20 to 76 years. Researchers took peripheral blood or bone marrow samples from each the patients to create single-cell derived hematopoietic colonies that underwent whole-genome sequencing, for a total of 952 sequences.
“Each whole genome of the colony reflected that of the single cell from which the colony was derived,” Nangalia said. “Right from the start of life, as our cells are dividing, mutations are being acquired and they are being passed down from generation to generation, such that at any one snapshot in time in our life, the mutations within individual cells represent natural barcodes that can be used to trace back the ancestry of those cells right to the start of life.”
Researchers found “an abundance” of driver mutations at the clonal level in genes associated with myeloproliferative neoplasms and cancer, but also more copy number changes than data had previously suggested, Nangalia added. They used these findings to create “phylogenetic trees” for each colony to understand the timing of when driver mutations appeared in patients.
In one patient diagnosed with essential thrombocytopenia at age 21 years, researchers found the acquisition of JAK2V617F occurred early in life, between 6.2 weeks after conception and 1.3 years of age.
“The pattern of branching downstream of JAK2 reflects how that single stem cell that acquired JAK2 is now clonally expanding into a clone of stem cells, and the rapidity of that branching tells us how quickly that clone was expanding,” Nangalia said.
In a patient diagnosed with polycythemia vera at age 31 years, researchers identified the myeloproliferative neoplasm clone in the middle of the phylogenetic tree, with JAK2 being acquired in childhood and another clonal expansion, harboring DNMT3A, also acquired in childhood and growing slowly until it reached a detectable clonal fraction.
Samples from the other patients continued to provide data suggesting JAK2 and DNMT3A mutations can be acquired in utero. In one patient, diagnosed with polycythemia vera at age 45 years, researchers found DNMT3A was acquired within days to weeks following conception.
“There was evidence of clonal evolution over 35 years and a JAK2 mutation followed by homozygosity for JAK2 followed by 1q amplification happening decades apart,” Nangalia said.
In some patients, JAK2 was acquired as a second mutation in an already expanded DNMT3A-mutated clone.
Also, in the eldest patient in the cohort, researchers observed a “driverless” clonal expansion.
“We know that clonal hematopoiesis can often lack driver mutations, which is thought to be the case in up to 50% of patients,” Nangalia said. “Here, we show that also has a single-cell origin, but what is driving clonal expansion in this individual in that particular clone, we do not know.”
Within individual patients across the cohort, researchers also recurrently found the same or similar genetic aberrations. For instance, one patient diagnosed with polycythemia vera in their 50s had three independent DNMT3A mutations and four independent acquisitions of homozygosity for their JAK2 exon 12 mutation.
“In other patients, we had independent acquisitions of 1q amplifications leading to myelofibrotic transformation, and in one patient we had multiple independent acquisitions of JAK2V617F,” Nangalia said. “This suggests that factors other than a driver mutation are predisposing patients to select for certain driver mutations, such as perhaps their germline background or their macroenvironment within the bone marrow.”
Next, researchers took longitudinal blood samples from these patients and resequenced the mutations identified in the phylogenetic trees in the whole blood. They evaluated the mutant clonal fractions in blood against the pattern of branching in the trees to infer the rates at which the clones were growing over the patient’s life.
They found widely variable clone growth rates. For example, in one patient with three mutant clones, their DNMT3A clone acquired in vitro grew 9% per year over their life, whereas their mutated JAK2 and TET2 clones grew at 233% per year, indicating a doubling-in-size time of every 7 months.
Even the same clone grew at different rates across patients. For instance, growth in JAK2V617F mutations ranged from 18% to 68% per year, suggesting “there are factors other than JAK2 that determine the consequences of acquiring it in individual patients,” Nangalia said.
Importantly, this rate of growth determined the latency to diagnosis. Patients with slow growth rates took 50 years to clinically present with disease, whereas those with fast growth rates presented with disease in as few as 10 years. Overall, the mean latency between JAK2V617F acquisition and clinical presentation was 34 years (range, 20-54).
Researchers then worked backward and evaluated JAK2 clonal fractions to calculate when the clones would be detectable ahead of disease presentation. They determined they were detectable 40 years prior to diagnosis in the case of slow-growing clones (18%/year) and 10 years prior to diagnosis in the case of fast-growing clones (233%/year).
“We think that if you can not only detect clones early, but then calculate their rates of growth with a repeat sample, you can then plot the growth trajectory of these clones and estimate the latency to a potential clinical manifestation, thus offering opportunities for early preventative strategies,” Nangalia said.
Perspective
Back to Top
Robert A. Brodsky, MD
This study suggests that genetic mutations linked to myeloproliferative neoplasms can occur during childhood, or even before birth, and proliferate in the body for many years before causing symptoms. Nangalia and her team were able to trace the ancestry of different blood cells and estimate the time at which the patient acquired JAK2 mutations and other important mutations linked to myeloproliferative neoplasms. The results suggest that there may be untapped opportunities to detect these conditions much earlier and to potentially intervene and prevent disease development. You can just imagine, with such a long latency period — and diseases that take another 10, 15 or 20 years to cause problems — what an impact of slowing this down even 15% would have.
This is fascinating research, and [raises additional questions] about what it means. What about patients who harbor these mutations for years and then need chemotherapy? Will this be accelerated? Would you consider using these patients as bone marrow donors, where you’re putting replicative stress on this? Again, could this process be accelerated?
Perspective
Back to Top
This fascinating study convincingly showed that the canonical driver mutations of this disease are acquired decades before diagnosis and, in some cases, the earliest mutations may have occurred during development in utero, with serial acquisition of additional mutations happening over decades.
Although this was a basic science project to better understand the disease biology, we can imagine next directions for research and implications. The results have many potential implications for patients, not least of which is being able to explain to our patients how these diseases arise.
In addition, identifying a mutated clone population of cells with the JAK2 mutation, and then estimating the trajectory of its growth, can provide a sense of the risk for development of a myeloproliferative neoplasm. Specifically, patients in whom that clone expanded most rapidly had a shorter latency to diagnosis, so you could imagine a future where we identify JAK2-mutant clones in the peripheral blood of healthy people and survey them over time to get a sense of the growth trajectory, perhaps working that into a model or risk score that predicts which individuals are at higher risk for evolution to a myeloproliferative neoplasm.
That information by itself would be useful to patients, so they can plan and prepare based on their risk. But, what would be most impactful is intervention. This research sets the stage for that, although it wasn’t the subject of the study. We currently don’t have the tools at our disposal to slow the growth of the clone over time, but that’s something a lot of people will be thinking about in the future.
The researchers conducted whole-genome sequencing on more than 950 colonies, which is a massive amount of data and analysis for just 10 patients, so there’s no way that scales population-wide in the foreseeable future. But, their finding of growth in the clone size being related to the latency of the disease is specific to the JAK2 clone, so in principle that’s all you’d have to look at, which can be done with cost-effective and high-throughput techniques.
The field of clonal hematopoiesis has developed over the last few years and is grappling with these same questions. How do we identify and survey at-risk individuals? How can we think about interventions to reduce that risk?
In fact, that type of work is going on in many different centers around the world. Researchers are establishing clinics to characterize individuals with clonal hematopoiesis to better identify those who are and aren’t at risk for a variety of illnesses, including cardiovascular disease. The vast majority of people with clonal hematopoiesis never develop a blood cancer, so mapping out their natural history and thinking about interventions is an active area of investigation.
This work also might have implications for high-risk groups, such as patients getting cytotoxic chemotherapy. In some cases, these mutated clones have a fitness advantage and will expand under the selected pressure of chemotherapy and radiation. Another high-risk group are those with an inherited predisposition to blood cancer. Both of these groups are targets of current investigations.
The findings of Nangalia and colleagues are novel, although they tie into some prior work. From a larger body of research, we know the general paradigm is that serial acquisition of mutations and epigenetic changes leads to disease playing out over time. The time between acquisition of these mutations is often measured in years, although the timescale observed by Nangalia showed a much longer latency than has been mapped before.
The idea that some of these initiating mutations might happen in utero also is not new. The seminal work of Sir Melvyn Francis Greaves, FMedSci, FRS, and his colleagues in the U.K. mapped this out for childhood leukemias in a series of elegant studies of monozygotic twins, showing very convincingly that lesion occurs during development. These two individuals track back to a single cell, and one twin developed leukemia and the other one didn’t, or sometimes they both develop leukemia but at very different latencies.
Also at ASH, Sahand Hormoz, PhD, and colleagues presented results of a study in which they used a single-cell approach to construct these pedigrees and trajectories of the evolution of patients with myeloproliferative neoplasms. They used whole-genome sequencing to also show a latency of decades from the acquisition of that JAK2 mutation until diagnosis. But, they also used single-cell RNA sequencing, which allowed them to layer the mutational hierarchy on a developmental hierarchy to compare the lineage trajectories of wild-type and JAK2-mutant cells. In many ways, these two studies are very complementary.
This work raises several future directions of research. First, like any good scientific study, this needs to be replicated. The researchers used sophisticated mathematical models to infer the timing of acquisition of mutations. They didn’t collect samples throughout those 70 years, including in utero; they collected one sample and inferred the time course going back. Like any mathematical model, there are assumptions built in, so the work should be replicated by other groups using different methods to confirm the findings.
Second, this work should be expanded to a larger population, as there were only 10 patients in this study. Almost every patient told a new chapter in the story, so by studying more patients we’ll likely learn additional details, even if the general take-home points don’t change.
Third, it would be interesting to apply these methods to other blood cancers and study this in other contexts. This was a great place to start because these are chronic leukemias that develop and remain relatively stable over decades, allowing enough time to elapse to sample and figure out the timing and order of acquisition. Acute leukemias have a more explosive onset, but still there is an indolent acquisition of mutations before a final mutation that leads to presentation. However, in this study, researchers took a single cell and grew colonies of thousands of cell copies, to have enough material to make a library with enough cells to do whole-genome sequencing. It is often not feasible to broadly sample wildtype and mutant colonies in other leukemias.
Admittedly, as fascinating and informative as this research was, it’s fundamentally descriptive. The researchers acquired these samples, sequenced them and made associations or inferred conclusions from the data. Some of the findings should prompt more mechanistic and basic questions. The field is already studying in genetically engineered mice the serial acquisition of mutations and what particular order and pattern of mutations lead to disease presentation, which is one direction that could corroborate these findings.
Also, we should address some of the researchers’ observations about how specific mutations influenced the speed at which additional mutations were acquired, which implies that somehow on a molecular level, the acquisition of the first driver mutation influences the likelihood or the rate at which additional mutations occur. Studying that mechanism will be interesting and, if understood, may present additional opportunities for intervention.
References:
Van Egeren D, et al. Abstract 714. Presented at: ASH Annual Meeting and Exposition (virtual meeting); Dec. 5-8, 2020.
Wiemels JL, et al. Lancet. 1999;doi:10.1016/s0140-6736(99)09403-9.
Harvard Medical School
Collapse
Williams N, et al. Abstract LBA-1. Presented at: ASH Annual Meeting and Exposition (virtual meeting); Dec. 5-8, 2020.
Click Here to Manage Email Alerts