As many hematology/oncology trainees will learn during their fellowship, certain germline mutations give rise to hematologic disorders. Some of these relationships are seemingly simple — a single base pair alteration leads to a GLU to VAL switch in the β-globin gene, for example. But others are more complex, including the germline mutations that confer leukemia risk or diseases of the megakaryocytes or platelets. There is a growing understanding that to ensure that genetic diagnoses are accurate and patient-centered, the scientific and clinical communities require a solid interpretative infrastructure to safeguard the many stakeholders of these data: What are the true genetic and phenotypic associations? Are all variants equally causative? Are all tests performed in a valid manner? Are the findings actionable?
This is the mission of the National Institutes of Health (NIH) Clinical Genome Resource (ClinGen), a federally funded project aimed at building shared knowledge repositories for genes and their variants. In early 2018, ASH partnered with the University of North Carolina (UNC) at Chapel Hill, an NIH ClinGen grantee, to develop a broad and accessible collection of genomic data aimed at improving the diagnosis of myeloid malignancies and hereditary platelet disorders. The two expert review panels have been working for nearly two years on this and have begun publishing their findings.1 Drs. Lucy Godley (Section of Hematology/Oncology and Center for Clinical Cancer Genetics, The University of Chicago, Chicago, IL) and David Wu (Department of Laboratory Medicine, University of Washington, Seattle, WA) co-lead the panel developing curation rules for variants in genes that confer risk for hereditary myeloid malignancies, and Drs. Jorge Di Paola (Department of Pediatric Hematology Oncology, Department of Pediatrics, Washington University, St. Louis, MO) and Wolfgang Bergmeier (Department of Biochemistry and Biophysics; UNC Blood Research Center, University of North Carolina at Chapel Hill, Chapel Hill, NC) co-lead the team developing similar rules for gene variants that cause inherited platelet disorders.
Editor-in-Chief Dr. Laura Michaelis recently interviewed Drs. Godley and Di Paola about the panels’ work.
Laura Michaelis: Let’s start by talking briefly about why ASH got involved with ClinGen. What need was being addressed by the collaboration?
Lucy Godley: Several years ago, the ASH Task Force on Precision Medicine recognized that germline mutations are responsible for many hematologic conditions. Also, increasingly commonly, we are performing molecular profiles of tumors. In performing these studies, however, we are inadvertently finding germline variants. If you look at the way those are categorized by clinical laboratories in the United States and internationally, there is no consistency. ASH sought to bring consistency to genes that are important for leukemia risk and platelet function, so they funded two ClinGen committees. The committees’ intent is to specify the American College of Medical Genetics’ rules by which gene variants are classified for diseases that are relevant for ASH members. And so ASH forged partnerships with ClinGen to populate two committees — the one that I co-chair with Dr. David Wu from the University of Washington to classify variants that confer risk for myeloid malignancies, and the other committee to develop rules for genes that cause inherited thrombocytopenia.
Jorge Di Paola: There is this proliferation of genetic panels now, not only for cancer genes, but also for germline mutations that cause platelet disorders, for example. If you are a hematologist and you send a gene panel for platelet disorders because you suspect your patient has a congenital thrombocytopenia, you will often get a report back that reads: “We found a genetic variant of uncertain significance in this particular gene in which mutations have been reported to cause thrombocytopenia, but the variant itself has not been previously reported.” But if you’re going to make patients face disease-based decisions, we’ve got to catalog this in a systematic way, and that’s what we are doing.
We started this a year-and-a-half ago, and I think, looking back, we’ve made a lot of progress. These two pilot programs (myeloid malignancies and platelet disorders) were chosen because there have been a lot of discoveries over the last two decades in those particular areas and because there is, again, a proliferation of genetic panels that are available to order in the clinical setting.
Michaelis: It seems like the impact of variants might be different if one were focusing on different tissues. I take it that your two panels are, for now, specifically interested in the bone marrow or megakaryocyte lineage?
Godley: Exactly right. Yes, we are classifying variants that are conferring risk to the bone marrow to develop myeloid malignancies. We started with RUNX1 because it had the greatest number of variants in publicly available ClinVar, which is where the variant data are first deposited. And we also thought it would be helpful, since RUNX1 mutation carriers also have thrombocytopenia and the other ASH-sponsored ClinGen committee is working on germline predisposition to thrombocytopenia. So we thought that would be a nice starting place. It took us about a year-and-a-half to generate these rules for RUNX1. It’s quite a rigorous process I have to say, and it’s been really interesting and eye-opening.
Di Paola: Our group opted to start with probably the most well-known platelet function disorder on earth — Glanzmann thrombasthenia, first described by the Swiss pediatrician Eduard Glnazmann in 1918, where there is an absence or deficiency in the platelet fibrinogen receptor αIIbβ3. This receptor is encoded by two different genes: ITGA2B and ITGB3. The reason we started with that is that this disease has been genetically well characterized all over the world. We said, “we have a lot of biological information, we have a lot of publications, we have a lot of genetic data that have already been published, and we have databases on these genes.”
Michaelis: Initial publications from your panels are out or pending, correct?
Godley: Yes, a paper discussing our detailed curation rules is now published in Blood Advances.1 For the most part, I think the end users of our Blood Advances paper are the clinical laboratories that are generating the variant interpretation. So now when a clinical laboratory sees a RUNX1 variant, it will be able to use these rules to say whether this variant is known pathogenic, likely pathogenic, variant of unknown significance, likely benign, or known benign — the five different levels of functional annotation. The variants that are in ClinVar are already being classified by our committee using those rules, so that if a clinical laboratory sees a variant, they’ll first go to the ClinGen/ClinVar website and see whether this variant has been classified already by the committee. Any variant that’s already deposited has essentially been classified by the rules, and that makes their job very easy. If they have a new variant that’s never been described, then they’ll use these rules to classify it. If you see the Haematologica article [publication date TBA], you’ll see that there were conflicting interpretations from different laboratories for some RUNX1 variants, as there are for all of these gene variants. Some laboratories were calling a particular variant benign, and another laboratory has called the exact same variant deleterious. That’s obviously a huge problem! And so, these rules build consensus. We’ve fashioned the Haematologica article almost as a “How I Curate Gene Variants.”
Di Paola: Much like Lucy’s group, we collaborated with ClinGen, and with the people at UNC where ClinGen is mostly based, we started the curation and (sort of) the design of the rules for curating these two genes. We just finished our pilot project, and we curated approximately 60 variants of the genes implicated in Glanzmann. This publication will be submitted to Blood Advances shortly.
Michaelis: Which genes or disorders will you tackle next?
Godley: We started with RUNX1, and we’re working on curation rules for GATA2 right now. After that, we probably need to do DDX41, CEBPα, [and] ETV6. Although ETV6 could be shared with the thrombocytopenia committee.
Di Paola: We’re going to curate the most traditional platelet diseases first, such as Bernard-Soulier syndrome caused by defects on the von Willebrand factor receptor, the GP1bα-IX-V complex on the platelet surface. After that, we’re going to start to go deeper into other disorders, and probably after we finish what we know about platelet function disorders, we’re going to go into other thrombocytopenias including the giant platelet disorders such as those caused by MYH9 mutations.
Michaelis: What does your output mean to patients and, not necessarily geneticists, but the working hematologist?
Godley: The first thing I would say is that we are identifying more and more families and individuals based on molecular profiling of leukemia cells. Because the prognostication of the leukemia is so dependent on molecular testing, and many of these genes are included in those panels, we are very often finding these individuals and families at the time of diagnosis of the first leukemia in the family. Recognition that molecular profiling of a tumor cell yields both germline and somatic data is extremely important. Labeling these tests as “somatic panels” is extremely misleading. It’s critical to recognize that when you see a TP53, a RUNX1, a DDX41, a CEBPα, or a GATA2 mutation in a myeloid leukemia, you have to question: Is this a germline variant? And once you determine that it’s germline, you now have rules that are quite rigorous in terms of the functionality of that variant.
Di Paola: Most of us dream of platforms that are comprehensive enough that when you get your genetic tests back, your doctor or hematologist would be able to check on a website and confirm with other sources and eventually tell you, “This is everything that we know about this variant.”
There is a lot of misinformation out there and a lot of fear. I think that overall if we do our job well, in several years, we’re going to be able to truly diagnose these disorders in a way that the patient will understand if they do have or do not have the disease. Also, physicians and genetic counselors will be absolutely needed to interpret with patients and families the implications of these findings. The fact that ASH is supporting this is fantastic because it has to happen.
