Frequency and Functional Characterization of <i>RUNX1</i> Germline Variants in Myeloid Neoplasms

Nitschke, Nikolaj Juul; Almosailleakh, Marwa; Niu, Yiyuan; Hansen, Jakob Werner; Raaschou-Jensen, Klas; Jespersen, Jakob Schmidt; Severinsen, Marianne Tang; Roug, Anne Stidsholt; Frödin, Morten; Weischenfeldt, Joachim Lütken; Andersen, Mette Klarskov; Grønbæk, Kirsten

doi:https://doi.org/10.1155/2023/4738660

Human Mutation

On this page

Abstract Introduction Methods Results Discussion Conclusions Data Availability Conflicts of Interest Authors’ Contributions Acknowledgments Supplementary Materials References Copyright Related Articles

Research Article | Open Access

Volume 2023 | Article ID 4738660 | https://doi.org/10.1155/2023/4738660

Frequency and Functional Characterization of RUNX1 Germline Variants in Myeloid Neoplasms

Nikolaj Juul Nitschke,^1,2Marwa Almosailleakh,^1,2Yiyuan Niu,²Jakob Werner Hansen,^1,2,3Klas Raaschou-Jensen,⁴Jakob Schmidt Jespersen,^2,5Marianne Tang Severinsen,^6,7Anne Stidsholt Roug,⁸Morten Frödin,²Joachim Lütken Weischenfeldt,^2,5Mette Klarskov Andersen,⁹and Kirsten Grønbæk^1,2

Academic Editor: Ulf Landegren

Received25 Jan 2023

Revised02 May 2023

Accepted08 May 2023

Published02 Jun 2023

Abstract

Current estimates suggest that up to 10% of patients with myeloid neoplasms (MN) harbor variants associated with a germline predisposition. A pathogenic variant in the runt-related transcription factor 1 gene (RUNX1) is a frequent cause of germline predisposition to MN. RUNX1 variants detected in tumor tissue at a VAF close to 50% are potentially germline and causative of RUNX1 familial platelet disorder with associated myeloid malignancies. Previous studies have found germline RUNX1 variants in 3% of patients with acute myeloid leukemia; however, the frequency of germline RUNX1 variants in less advanced myeloid neoplasms has not been examined. We screened 590 patients suspected of MN, excluding myeloproliferative neoplasms, for germline variants in RUNX1. We found RUNX1 variants in 83 patients (14%) by targeted sequencing of tumor tissue. In 40 patients (6.8%), the VAF of RUNX1 was above 30%. In 32 of the 40 patients, skin biopsies were available and used for Sanger sequencing to assess the germline status. Two of the tested variants (6.3%) were confirmed as germline, and both variants were curated as variants of unknown significance. To further explore the pathogenicity of these variants, we implemented a novel CRISPR-Select functional genetic assay. The assay demonstrated a profound effect on proliferation in K562 cells for a known pathogenic variant but no effect for the two germline variants detected in the study. We therefore propose that both germline variants are classified as likely benign. In this study, we show that RUNX1 germline variants are rare in Danish patients with MN and use a novel assay for functional classification of germline RUNX1 variants.

1. Introduction

Current estimates suggest that 10% of patients with myeloid neoplasms (MN) harbor variants associated with germline predisposition [1, 2]. Germline mutations in the runt-related transcription factor 1 gene (RUNX1) were first described in 1999, as causing inherited thrombocytopenia with a propensity to develop acute myeloid leukemia (AML) [3]. It has since been established to be one of the most common causes of germline predisposition to MN [4]. RUNX1 is a critical transcription factor essential for establishing definitive hematopoiesis, as shown in a mouse model where knockout of RUNX1 was embryonically lethal due to CNS hemorrhage [5]. In adult hematopoiesis, RUNX1 plays a critical role in lymphoid development, megakaryocytic differentiation, and platelet formation. MYH10 and MYL9 regulations by RUNX1 as well as repression of KLF1 in megakaryocytic-erythroid progenitors have been suggested as pathways connecting RUNX1 to platelet formation [6, 7]. Several mechanisms have linked RUNX1 haploinsufficiency to oncogenesis including altered inflammatory signaling [8], reduced ribosomogenesis [9], and an attenuated p53-DNA damage response [10]. Germline mutations in RUNX1 are causal for RUNX1 family platelet disorder (RUNX1-FPD) associated with a ~45% lifetime risk of malignancy [11], a varying degree of thrombocytopenia, and a high risk of eczema. Pathogenic germline mutations are generally located in the RUNT homology domain responsible for DNA binding and dimerization with core binding factor-β with a resulting haploinsufficiency [12].

RUNX1 mutations are identified in 10% of patients with AML and myelodysplastic syndrome (MDS) [13–15]. While most mutations are assumed to be acquired, growing evidence suggests that a substantial fraction are of germline origin. In patients with AML, Simon et al. showed that 30% of variants in RUNX1 with a variant allele frequency (VAF) greater than 30% were germline [16]. Bąk et al. identified three patients with AML with germline RUNX1 variants in a cohort of 100 patients [17], and Ernst et al. showed that 8% of patients with AML who achieved complete remission had a persistent RUNX1 variant throughout baseline and after achieving complete remission [18]. Distinguishing between somatic mutations with a high VAF and germline mutations can be challenging but crucial as it impacts clinical management, e.g., hematopoietic stem cell donor selection and genetic counselling. The frequency of germline RUNX1 mutations in preleukemic conditions, e.g., MDS and premalignant phenotypes such as clonal cytopenia of unknown significance (CCUS), remains unknown. In this paper, we uncover the germline status of RUNX1 variants in patients with a spectrum of MN (ICUS, CCUS, MDS, chronic myelomonocytic leukemia (CMML), and AML) and examine factors associated with germline RUNX1 variants. Furthermore, we present a novel technique based on CRISPR-Cas9 for the functional classification of RUNX1 germline variants.

2. Methods

2.1. Patient Cohort and Material

This study is part of the Danish research program: the program for translational hematology (PTH), approved by the Danish National Ethics Committee. Patients referred with suspicion of MN or with relapse/progression of MN, excluding myeloproliferative neoplasms (MPN), were included in PTH between 2018 and 2021. On the first visit to a Danish hematological department, written consent was obtained, and a skin biopsy, peripheral blood samples, and a bone marrow aspirate were drawn. Skin biopsies were taken from the area of the skin above the posterior superior iliac spine and frozen dry at -80°C. We excluded patients from the study if the clinical work-up provided sound evidence of a nonmyeloid disease as the cause of the patient’s symptoms, e.g., acute lymphoblastic leukemia.

2.2. Next-Generation Sequencing (NGS)

As part of their clinical workup, 450 patients were sequenced with one of the following amplicon-based platforms: the Ion AmpliSeq™ AML Research Panel covering 19 genes, Sophia Genetics™ Myeloid Panel covering 30 genes, and Ion Torrent™ Oncomine™ Myeloid Assay covering 72 genes. Filtering and curation of variants were done at the local departments of clinical genetics, pathology, or hematology. Information on variants was obtained from patient’s electronic records along with clinical information. The indication for NGS is not specified on a national level but relies on the individual physician/department; therefore, the date of sequencing in some patients differed from the date of inclusion in the PTH study. Furthermore, 237 patients were sequenced as part of the PTH research program using DNA extracted from live frozen mononuclear cells separated with LeucoSep (Greiner Bio-One) from either peripheral blood or bone marrow aspirates drawn at inclusion. Sequencing was performed on an Illumina NextSeq 500 using a targeted sequencing panel (Illumina TruSeq Custom Amplicon, Illumina) covering 145 genes previously related to myeloid neoplasms. We used BWA [19] to align reads to GRCh38, Picard tools [20] to sort the reads and mark duplicates, and GATK tools [21] to recalibrate base quality scores. We then used VarDict [22], SNVer [23], and LoFreq [24] to call variants and Funcotator [21] for annotation. Variants with less than 4 reads, read depths of less than 200 or more than 3000, not assigned to chromosome 1-22 or X or Y, and VAFs of less than 2% were filtered out. The remaining variants were manually reviewed. In 98 patients, we had sequencing data from both the clinical workup and the PTH study.

All RUNX1 variants were annotated to the RUNX1c transcript (NM_001754.4). RUNX1 variants classified as benign or likely benign in ClinVar [25] were excluded from further analysis. RUNX1 variants with VAF greater than or equal to 30% were classified as high VAF RUNX1 variants (hVAF-RUNX1). For patients with sequencing data from multiple time points, only RUNX1 variants with a consistent VAF greater than 30% were deemed hVAF-RUNX1. Variants with a VAF less than 30% after allogeneic hematopoietic stem cell transplantation (alloHSCT) were included for germline analysis. Patients with a known myeloid germline predisposition syndrome were also excluded for further analysis.

2.3. Germline DNA Extraction and Sanger Sequencing

Germline DNA was isolated from frozen skin biopsies using the DNeasy® Blood and Tissue Kit (Qiagen #69504), according to the manufacturer’s protocol. Before isolation, bloody tissue was removed, and biopsies were washed in PBS (Gibco™ 10010023). Polymerase chain reaction (PCR) was performed with EconoTaq PLUS GREEN 2X Master Mix (Lucigen #30033-1) on a Veriti 96-Well Thermal Cycler (Applied Biosystems®) as per manufacturer’s protocol. PCR conditions were used in initial denaturing for 1 min at 94°C, then 30 cycles of 94°C for 10 s, 60°C–62°C for 30 s, 72°C for 15 s, and a final post-PCR extension for 5 min at 72°C. Primers were designed using Primer3 to amplify exons 4-8, including intronic regions close to the exons. List of primer properties is available upon request. Bidirectional Sanger sequencing (Eurofins Genomics) was performed in PCR-amplified regions of RUNX1, depending on the location of the variant detected with NGS. Chromatograms were analyzed using the genome analysis server [26].

2.4. CRISPR-Select

We used a CRISPR-Select cell-based variant knock-in assay to functionally assess the pathogenicity of germline variants. As we have recently reported the principle of the method [27], we will only describe the setup briefly. K562 cells were cultured in Roswell Park Memorial Institute 1640 medium (ATCC #30-2001) supplemented with 10% (v/v) fetal bovine serum (Thermo Fisher Scientific #12389802) and penicillin-streptomycin (Gibco™ #15140122).

GuideRNAs (gRNAs) were designed for S. pyogenes Cas9 with the online software Benchling (https://benchling.com). The base pairs to be mutated were located as close as possible to the genomic cut site to increase knock-in efficiency and within the protospacer adjacent motif (PAM) or the 1-10 PAM proximal nucleotides within the gRNA target site for the mutations to effectively destroy the Cas9 target site. Single-stranded oligodeoxynucleotide (ssODN) repair templates encoding mutations to be knocked in were designed such that the synonymous WT´ control mutation was placed within the same codon as the variant of interest to promote knockin at similar frequencies. gRNAs were used in the form of crRNA:tracrRNA duplexes purchased from IDT™ and reconstituted in nuclease-free duplex buffer at 100 μM. For ribonucleoprotein (RNP) generation, Alt-R S. pyogenes Cas9 Nuclease V3 from IDT™ (#1081059) was used. ssODN repair templates were purchased from IDT™ as unmodified Ultramer DNA oligonucleotides at 100 μM in IDTE, pH 8.0.

Briefly, for a nucleofection of 10⁶ k562 cells, 500 pmoles each of crRNA and tracr-RNA were mixed and allowed to complex by incubation for 10 min at room temperature. Next, 96 pmoles of Cas9 proteins were mixed with the crRNA:tracrRNA duplexes and incubated for further 10 min. Next, cells were resuspended in 80 μl of electroporation solution and added to RNPs and 500 pmol each of variant and WT ssODN. Finally, the cell suspension was transferred to a nucleocuvette and electroporated in a Lonza 4D-Nucleofector device using the T-003 program.

Genomic DNA was extracted on days 2 and 24 after nucleofection from an aliquot of the cell cultures using the Quick-DNA™ Miniprep kit (Zymo #D4069) as per the manufacturer’s protocol. The genomic target site was amplified for NGS analysis using 100 ng of genomic DNA as a template and a two-round PCR [28]: Primer pairs for amplification of the target site in the first PCR were designed to anneal 40-120 nt outside the region covered by the ssODN repair template and to generate PCR products of 230-350 bps, using a Primer-BLAST3 from NCBI (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). The second-round PCR and relevant primer design and PCR conditions were performed as described [27]. After mixing roughly equal amounts of second-round PCR products, the amplicon sequencing library was made using the MiSeq Reagent Kit v2 (Illumina #MS-102-2002) and finally sequenced in a MiSeq instrument from Illumina, according to the manufacturer’s instructions. Sequencing depths ranged from 20,000-200,000 reads per sample. NGS data were analysed by the CRISPResso2 online tool using default settings (https://crispresso.pinellolab.partners.org/submission) [29].

2.5. Statistics

Statistical analysis included the calculation of odds ratios from two-by-two tables and differences between groups using Fisher’s exact test. Platelet levels in different groups were analyzed using a linear model that adjusted for sex, age, and baseline diagnosis. Variant/WT’ ratios were compared with a two-tailed paired -test. was considered significant. All statistical analyzes were performed in R using version 3.6.1.

3. Results

3.1. Frequency of Germline RUNX1 Variants

In 590 unselected patients referred to a hematological department with suspected MN, excluding MPN, and without evidence of a nonmyleoid cause of symptoms, we identified a RUNX1 variant in 83 patients (14%); see Figure 1(a). MDS was the most frequent diagnosis in our cohort, followed by AML and CCUS; see Supplemental Figure 1. CMML patients had the highest frequency of RUNX1 variants including a high frequency of hVAF-RUN1; see Supplemental Figure 2. In 40 patients (6.8%) we found a hVAF-RUNX1. Of these, one patient was excluded due to a diagnosis of Shwachman-Diamond syndrome; 7 patients were excluded as no skin biopsy was available. In 2 of 32 patients (6.3%), the hVAF-RUNX1 was confirmed as germline by bidirectional Sanger sequencing of DNA extracted from frozen skin biopsies; see Figure 1(b). None of the RUNX1 variants with a VAF between 30 and 40% were germline. In our cohort, we observed a significant association between RUNX1 variants and a lower platelet count () after adjusting for age and diagnosis; see Figure 1(c); furthermore, patients with RUNX1 variants had a significantly higher blast count; see Table 1. Clinical characteristics of patients with hVAF-RUNX1 are provided in Supplemental Table 2.

(a)

(b)

(c)

3.2. Cooccurring Mutations

In line with previous findings, we observed frequent cooccurring mutations in SRSF2 (18/32), TET2 (16/32), and ASXL1 (10/22); see Figure 2(a). In patients with confirmed germline RUNX1 variant, we did not observe any cooccurring mutations with a similar high VAF, whereas 5 out of 30 patients with somatic hVAF-RUNX1 had cooccurring mutations at a similar VAF; see Figure 3(b).

(a)

(b)

(c)

Figure 3

(a) Distribution of high VAF RUNX1 variants found in our cohort. Germline variants are depicted on top whereas somatic are depicted in the bottom. (b) Distribution of cooccurring mutations in patients with and without germline RUNX1 variants. ; MDS: myelodysplastic syndrome; AML: acute myelogenous leukemia; CCUS: clonal cytopenia of unknown significance; CMML: chronic myelomonocytic leukemia; VAF: variant allele frequency. (c) Curation of germline RUNX1 variants. MM-VCEP: Myeloid Malignancy Variant Curation Expert Panel; RUNX1-FPD: RUNX1 familial platelet disorder; gnomAD: Genome Aggregation Database; CCUS: clonal cytopenia of unknown significance; MDS: myelodysplastic syndrome.

3.3. Curation of Germline RUNX1 Variants

The hVAF-RUNX1 were either frameshift, splice-site, nonsense, or missense, and the majority resided in the RUNT homology domain (RHD); see Figure 3(a). The germline variants (c.649G > A; p.G217R; and c.668A > G; p.E223G) were located between the Transactivation Domain (TAD) and the RHD.

The RUNX1 variant p.E223G was identified in a 22-year-old woman with thrombocytopenia. Thrombocytopenia was diagnosed during an episode of deep vein thrombosis following a luxation of the right patella. The patient had no family history of hematological malignancies or thrombocytopenia. Hematological workup revealed no malignancy and only mild symptoms of bleeding (menorrhagia), and therefore she was initially diagnosed with CCUS due to the RUNX1 variant. Subsequent examination revealed abnormalities of the aortic valve requiring surgery, periventricular nodular heterotopia detected with MR-cerebrum and a de novo missense VUS in the FLNA gene. Variants in FLNA are associated with periventricular nodular heterotopia, congenital heart disease, thrombocytopenia, among other things, and therefore, she was diagnosed with FLNA deficiency [30]. The RUNX1 variant is classified as a variant with conflicting results in ClinVar (a report of VUS and likely benign). In gnomAD, it is identified at a frequency of 0.012%. In silico prediction with REVEL, using a recommended threshold of 0.75 from the ClinGen Myeloid Malignancy Variant Curation Expert Panel (MM-VCEP) [11], indicates a deleterious effect, score 0.799.

The p.G217R variant was identified in a 64-year-old male with MDS, a normal platelet count, and cooccurring mutations in IDH1 and SRSF2, but family history was unavailable. The p.G217R variant is classified as a VUS in ClinVar and reported with a frequency of 0.003% in gnomAD. Using the same cutoff as mentioned above, REVEL does not predict the variant to be deleterious according to the recommended threshold, score 0.624. When applying ClinGen MM-VCEP RUNX1-specific curation, both variants are classified as VUS. See Figure 3(c) for curation details.

3.4. Cell-Based Functional Characterization of Germline RUNX1 Variants

To assess pathogenicity, we established CRISPR-Select for the functional analysis of RUNX1 variants in relation to MN. As assay cells, we chose the human lymphoblast K562 cell line, which has previously been reported to express RUNX1 [31]. As a CRISPR-Select readout for RUNX1 function, we chose proliferation, as Cai et al. [32] reported that loss of RUNX1 was associated with decreased ribosome biogenesis and slow growth in HSPC. For each variant, we nucleofected a culture of K562 cells with CRISPR-Cas9 ribonucleoprotein targeting the variant genomic site along with two ssODN repair template coding for either the variant of interest or a synonymous control variant in the same codon, thereby producing a culture containing cells with the variant of interest (variant) and the corresponding control cells (WT´), see Figure 4(a). We examined the two germline RUNX1 variants p.G217R and p.E223G in our cohort, and as a positive control, we used a known pathogenic variant in the same region, i.e., c.679G > T, p.Glu227Ter. To assess a relative change in Variant and WT´ cells over time as a readout for variant effect on cell proliferation and/or survival, we first extracted DNA from an aliquot of the cell populations at day two and day 24 after nucleofection. We then performed NGS on PCR products covering the region of knockin. As an outcome, we examined the variant and WT´ number of reads and compared the changes in the variant:WT´ ratio over time, using day two after nucleofection as the baseline. The pathogenic p.Glu227Ter control showed a large significant decrease in the variant:WT´ ratio at day 24, as depicted in Figure 4(b), revealing the expected loss of RUNX1 function. For the germline variant p.G217R, we observed no significant change over time, and for the germline variant p.E223G, we observed a significant change in variant:WT’, but the mean difference over time was small compared to the known pathogenic control. The observed lack of effect could be a false negative due to a lack of selective pressure for various reasons in the specific experiments. However, the complete NGS characterization of CRISPR-Select editing outcomes allowed us to analyze the ratio of frameshift insertions and deletions (InDels) to WT´, which demonstrated strong negative selection against cells with frameshift InDels in the same cell culture dishes, where the variants exhibited no selective disadvantage; see Figure 4(c).

(a)

(b)

(c)

4. Discussion

Germline RUNX1 variants have previously been examined in patients with AML, but the frequency of germline RUNX1 mutations in pre-/less malignant phenotypes such as MDS and CCUS is unknown. Identification and classification of germline RUNX1 variants is important as they have implications for clinical decision-making, e.g., donor selection. To estimate the frequency of germline RUNX1 variants, we examined 590 patients and found that in 2 of 32 patients, hVAF-RUNX1 was of germline origin, suggesting an absolute frequency of 0.3%. Previous studies reported absolute frequencies of germline RUNX1 variants ranging from 1-3% in patients with AML, suggesting that the germline RUNX1 frequency is higher in AML. In this study, we were not able to technically detect previously reported germline structural variants in RUNX1 [33]. The two observed germline variants are classified as VUS, suggesting an even lower incidence of true disease-causing variants. In previous studies of germline RUNX1 frequencies, no functional testing was performed [17, 34], potentially resulting in an overestimation of the frequency of disease-causing RUNX1 variants. We had a relatively high frequency of RUNX1 variants in our cohort compared to previous reports in MDS and AML [15, 35]; this could be due to the inclusion of CMML with a higher frequency of RUNX1 variants.

Our data is real-world-based with variants being identified at multiple institutions using different sequencing tools, panels of genes, computational methods, and different approaches to variant curation. This makes the findings directly applicable to a clinical setting. However, there is a risk that variants are overlooked. To address this, we examined hVAF-RUNX1 variants identified in patient samples which were sequenced using both our research panel and a clinical routine panel as part of the clinical workup (). Only two of ten hVAF-RUNX1 were identified exclusively by the research panel (Supplemental Table 1). Unfortunately, no skin biopsy was available to examine the germline status of the c.e7-6TAAGC > C variant, but the c.1287_1341dup variant was confirmed as a somatic variant.

The pathogenicity of RUNX1 variants is difficult to assess. In silico predictions are unreliable, especially in RHD as this is a highly conserved region and most prediction tools rely on conservatism. The ClinGen Myeloid Malignancy Variant Curation Expert Panel has proposed criteria for classifying variants, but many variants remain classified as VUS. We designed a cell-based assay for RUNX1 variants, which determines the effect of CRISPR-Cas9-introduced variants on cell proliferation and/or survival. When introducing a known pathogenic variant, we observed an expected relative decrease in the proliferative capacity of variant cells compared to WT´ cells. With this assay, we were able to show that the two germline variants p.G217R and p.E223G had no functional effect in vitro on proliferative capacity, and we therefore reclassify them as likely benign variants. It should be emphasized that the CRISPR-Select format produces data of very high reliability. For instance, (i) CRISPR off-target actions and several other experimental artifacts are normalized out by the internal WT´ control, (ii) the data are based on hundreds of independent knockin cells for variant and WT´, (iii) results are based on analysis of the variant of interest in a proper genomic context without overexpression, and several additional features support the data reliability, as discussed by [27].

A previous study by Decker et al. used a set of three different assays to investigate the functional effect of RUNX1 variants: heterodimerization ability with CBFB measured with a flow cytometry-based FRET assay, phosphorylation of RUNX1 quantified with western blot and the ability of RUNX1 to activate transcription using a luciferase reporter assay [36]. Only one of the tested variants resided in the region where we detected two germline variants. Another study by Decker et al. analyzed different variants by applying a luciferase reporter assay to measure active transcription of three RUNX1 target genes, CSF1R, ETV1, and MYL9 [37]. The assay was able to reclassify two VUSs as likely pathogenic variants but were only able to reclassify one out of four VUSs in the region where our two germline variants resided. In total, the two studies could reclassify two out of five variants in our region of interest, demonstrating the need for further functional assays such as the current. Furthermore, none of the known assays measure the impact of RUNX1 variants on cell phenotypes, whereas our assay offers a direct way of testing whether a variant influences cell proliferation and/or survival. It is furthermore easy to modify the assay culture conditions with the aim of evaluating the effect of different exposures, e.g., drug screening or differentiation stimuli. A limitation of our assay is that we only used a truncating variant as a pathogenic control, but at this point, no known pathogenic missense variants have been reported in the region between RHD and TAD. Further studies are needed to investigate the performance of the assay in different regions of RUNX1.

5. Conclusions

Germline RUNX1 variants are rare in this large Danish cohort of randomly sampled MN. In patients with a VAF greater than 30%, the frequency of germline variants was 6.3%. Here, we introduce a robust functional assay to evaluate RUNX1 variants and were able to reclassify two RUNX1 variants from VUS to likely benign.

Data Availability

It is not possible to deposit the data in public repositories, since these data are considered sensitive personal data according to Danish law and the European Union General Data Protection Regulation (GDPR) and thus cannot be shared with third parties without prior approval. To access the data, please contact the corresponding author at kirsten.groenbaek@regionh.dk. Access can only be granted for research purposes and only if a data processor or data transfer agreement can be made in accordance with Danish and European law at the given time. The expected timeframe from response until access is granted is ~6 months.

Conflicts of Interest

Anne Stidsholt Roug received an honorarium of teaching from AbbVie and Pfizer and a congress fee from Jazz.

Authors’ Contributions

N.N., M.A., Y.N., J.S.J., M.F., J.L.W., M.K.A., and K.G. conceived and designed the experiments. N.N., M.A., Y.N., and J.S.J. performed the experiments. N.N., M.A., M.K.A., and K.G. interpreted the data. J.W.H., K.R.J., M.T.S., A.S.R., and K.G. handled patient material and collected clinical data. N.N., M.K.A., and K.G. wrote the manuscript. All authors critically reviewed the final version of the manuscript.

Acknowledgments

The study was supported by a center grant from the Novo Nordisk Foundation (Novo Nordisk Foundation Center for Stem Cell Biology, Dan Stem; grant NNF17CC0027852) and the Greater Copenhagen Health Science Partners (Clinical Academic Group in Translational Hematology). The project is part of the Danish Research Center for Precision Medicine in Blood Cancers, funded by the Danish Cancer Society grant R223-A13071. Further funding was provided by the Rigshospitalet’s Research Foundation (to N.N.), Danish Cancer Society (R302-A17259), and the European Union’s Horizon 2020 Research and Innovation Programme (under the Marie Skłodowska-Curie grant agreement No. 801481 to Y.N.).

Supplementary Materials

Supplementary Figure 1: distribution of diagnosis in the cohort. Ranked from highest to lowest frequency. MDS: myelodysplastic syndrome; AML: acute myelogenous leukemia; CCUS: clonal cytopenia of unknown significance; CMML: chronic myelomonocytic leukemia. Supplementary Figure 2: proportion of patients with RUNX1 variants among the different myeloid neoplasms included in the cohort. MDS: myelodysplastic syndrome; AML: acute myelogenous leukemia; CCUS: clonal cytopenia of unknown significance; CMML: chronic myelomonocytic leukemia. Supplementary Table 1: variants in RUNX1 with VAF greater than 30% in patients sequenced both in a clinical setting and as part of the PTH study. VAF: variant allele frequency; PTH: program for translational hematology. Supplementary Table 2: clinical characteristics of patients with hVAF-RUNX1 variants. hVAF-RUNX1: RUNX1 variants with persistent VAF greater than or equal to 30%; VAF: variant allele frequency. (Supplementary Materials)

References

M. W. Drazer, S. Kadri, M. Sukhanova et al., “Prognostic tumor sequencing panels frequently identify germ line variants associated with hereditary hematopoietic malignancies,” Blood Advances, vol. 2, no. 2, pp. 146–150, 2018.
View at: Publisher Site | Google Scholar
U. Wartiovaara-Kautto, E. A. M. Hirvonen, E. Pitkänen et al., “Germline alterations in a consecutive series of acute myeloid leukemia,” Leukemia, vol. 32, no. 10, pp. 2282–2285, 2018.
View at: Publisher Site | Google Scholar
W.-J. Song, M. G. Sullivan, R. D. Legare et al., “Haploinsufficiency of CBFA2 causes familial thrombocytopenia with propensity to develop acute myelogenous leukaemia,” Nature Genetics, vol. 23, no. 2, pp. 166–175, 1999.
View at: Publisher Site | Google Scholar
A. Rio-Machin, T. Vulliamy, N. Hug et al., “The complex genetic landscape of familial MDS and AML reveals pathogenic germline variants,” Nature Communications, vol. 11, no. 1, 2020.
View at: Publisher Site | Google Scholar
Q. Wang, T. Stacy, M. Binder, M. Marín-Padilla, A. H. Sharpe, and N. A. Speck, “Disruption of the Cbfa 2 gene causes necrosis and hemorrhaging in the central nervous system and blocks definitive hematopoiesis,” Proceedings of the National Academy of Sciences, vol. 93, no. 8, pp. 3444–3449, 1996.
View at: Publisher Site | Google Scholar
O. N. Kuvardina, J. Herglotz, S. Kolodziej et al., “RUNX1 represses the erythroid gene expression program during megakaryocytic differentiation,” Blood, vol. 125, no. 23, pp. 3570–3579, 2015.
View at: Publisher Site | Google Scholar
L. Lordier, D. Bluteau, A. Jalil et al., “RUNX1-induced silencing of non-muscle myosin heavy chain IIB contributes to megakaryocyte polyploidization,” Nature Communications, vol. 3, no. 1, p. 717, 2012.
View at: Publisher Site | Google Scholar
D. C. Bellissimo, C. H. Chen, Q. Zhu et al., “Runx 1 negatively regulates inflammatory cytokine production by neutrophils in response to toll-like receptor signaling,” Blood Advances, vol. 4, no. 6, pp. 1145–1158, 2020.
View at: Publisher Site | Google Scholar
X. Cai, L. Gao, L. Teng et al., “Runx1 deficiency decreases ribosome biogenesis and confers stress resistance to hematopoietic stem and progenitor cells,” Cell Stem Cell, vol. 17, no. 2, pp. 165–177, 2015.
View at: Publisher Site | Google Scholar
D. Wu, T. Ozakis, Y. Yoshiharas, N. Kubos, and A. Nakagawara, “Runt-related transcription factor 1 (RUNX1) stimulates tumor suppressor p53 protein in response to DNA damage through complex formation and acetylation,” Journal of Biological Chemistry, vol. 288, no. 2, pp. 1353–1364, 2013.
View at: Publisher Site | Google Scholar
X. Luo, S. Feurstein, S. Mohan et al., “ClinGen myeloid malignancy variant curation expert panel recommendations for germline RUNX1 variants,” Blood Advances, vol. 3, no. 20, pp. 2962–2979, 2019.
View at: Publisher Site | Google Scholar
M. Sakurai, H. Kunimoto, N. Watanabe et al., “Impaired hematopoietic differentiation of RUNX1 -mutated induced pluripotent stem cells derived from FPD/AML patients,” Leukemia, vol. 28, no. 12, pp. 2344–2354, 2014.
View at: Publisher Site | Google Scholar
T. Haferlach, Y. Nagata, V. Grossmann et al., “Landscape of genetic lesions in 944 patients with myelodysplastic syndromes,” Leukemia, vol. 28, no. 2, pp. 241–247, 2014.
View at: Publisher Site | Google Scholar
R. C. Lindsley, W. Saber, B. G. Mar et al., “Prognostic mutations in myelodysplastic syndrome after stem-cell transplantation,” New England Journal of Medicine, vol. 376, no. 6, pp. 536–547, 2017.
View at: Publisher Site | Google Scholar
E. Papaemmanuil, M. Gerstung, L. Bullinger et al., “Genomic classification and prognosis in acute myeloid leukemia,” New England Journal of Medicine, vol. 374, no. 23, pp. 2209–2221, 2016.
View at: Publisher Site | Google Scholar
L. Simon, J. F. Spinella, C. Y. Yao et al., “High frequency of germlineRUNX1mutations in patients withRUNX1-mutated AML,” Blood, vol. 135, no. 21, pp. 1882–1886, 2020.
View at: Publisher Site | Google Scholar
A. Bąk, K. Skonieczka, A. Jaśkowiec et al., “Searching for germline mutations in the RUNX1 gene among polish patients with acute myeloid leukemia,” Leukemia & Lymphoma, vol. 62, no. 7, pp. 1749–1755, 2021.
View at: Publisher Site | Google Scholar
M. P. T. Ernst, F. G. Kavelaars, B. Löwenberg, P. J. M. Valk, and M. H. G. P. Raaijmakers, “RUNX1 germline variants in RUNX1-mutant AML: how frequent?” Blood, vol. 137, no. 10, pp. 1428–1431, 2021.
View at: Publisher Site | Google Scholar
H. Li and R. Durbin, “Fast and accurate short read alignment with Burrows-Wheeler transform,” Bioinformatics, vol. 25, no. 14, pp. 1754–1760, 2009.
View at: Publisher Site | Google Scholar
Institute, B, Picard Toolkit, Github Repository, 2019, https://broadinstitute.github.io/picard/.
G. A. Van der Auwera and B. D. O'Connor, “Genomics in the Cloud,” Tech. Rep., O'Reilly Media, Inc, 2020, https://www.oreilly.com/library/view/genomics-in-the/9781491975183/.
View at: Google Scholar
Z. Lai, A. Markovets, M. Ahdesmaki et al., “VarDict: a novel and versatile variant caller for next-generation sequencing in cancer research,” Nucleic Acids Research, vol. 44, no. 11, article e108, 2016.
View at: Publisher Site | Google Scholar
Z. Wei, W. Wang, P. Hu, G. J. Lyon, and H. Hakonarson, “SNVer: a statistical tool for variant calling in analysis of pooled or individual next-generation sequencing data,” Nucleic Acids Research, vol. 39, no. 19, article e132, 2011.
View at: Publisher Site | Google Scholar
A. Wilm, P. P. K. Aw, D. Bertrand et al., “LoFreq: a sequence-quality aware, ultra-sensitive variant caller for uncovering cell-population heterogeneity from high-throughput sequencing datasets,” Nucleic Acids Research, vol. 40, no. 22, pp. 11189–11201, 2012.
View at: Publisher Site | Google Scholar
M. J. Landrum, J. M. Lee, M. Benson et al., “ClinVar: improving access to variant interpretations and supporting evidence,” Nucleic Acids Research, vol. 46, no. D1, pp. D1062–D1067, 2018.
View at: Publisher Site | Google Scholar
T. Rausch, M. H.-Y. Fritz, A. Untergasser, and V. Benes, “Tracy: basecalling, alignment, assembly and deconvolution of Sanger chromatogram trace files,” BMC Genomics, vol. 21, no. 1, p. 230, 2020.
View at: Publisher Site | Google Scholar
Y. Niu, C. A. Ferreira Azevedo, X. Li et al., “Multiparametric and accurate functional analysis of genetic sequence variants using CRISPR-select,” Nature Genetics, vol. 54, no. 12, pp. 1983–1993, 2022.
View at: Publisher Site | Google Scholar
J. L. Schmid-Burgk, T. Schmidt, M. M. Gaidt et al., “OutKnocker: a web tool for rapid and simple genotyping of designer nuclease edited cell lines,” Genome Research, vol. 24, no. 10, pp. 1719–1723, 2014.
View at: Publisher Site | Google Scholar
K. Clement, H. Rees, M. C. Canver et al., “CRISPResso2 provides accurate and rapid genome editing sequence analysis,” Nature Biotechnology, vol. 37, no. 3, pp. 224–226, 2019.
View at: Publisher Site | Google Scholar
M. H. Chen and C. A. Walsh, “FLNA Deficiency,” Gene Reviews®, M. P. Adam, G. M. Mirzaa, and R. A. Pagon, Eds., University of Washington, Seattle, Seattle (WA), 1993, https://www.ncbi.nlm.nih.gov/books/NBK1213/.
View at: Google Scholar
R. Bera, M. C. Chiu, Y. J. Huang, T. H. Lin, M. C. Kuo, and L. Y. Shih, “RUNX1 mutations promote leukemogenesis of myeloid malignancies in ASXL1-mutated leukemia,” Journal of Hematology & Oncology, vol. 12, no. 1, p. 104, 2019.
View at: Publisher Site | Google Scholar
L. Ai, S. Mu, Y. Wang et al., “Prognostic role of myeloid-derived suppressor cells in cancers: a systematic review and meta-analysis,” BMC Cancer, vol. 18, no. 1, 2018.
View at: Publisher Site | Google Scholar
M. Béri-Dexheimer, V. Latger-Cannard, C. Philippe et al., “Clinical phenotype of germline RUNX1 haploinsufficiency: from point mutations to large genomic deletions,” European Journal of Human Genetics, vol. 16, no. 8, pp. 1014–1018, 2008.
View at: Publisher Site | Google Scholar
R. H. Vonderheide and C. H. June, “Engineering T cells for cancer: our synthetic future,” Immunological Reviews, vol. 257, no. 1, pp. 7–13, 2014.
View at: Publisher Site | Google Scholar
E. Bernard, H. Tuechler, P. L. Greenberg et al., “Molecular international prognostic scoring system for myelodysplastic syndromes,” NEJM Evidence, vol. 1, no. 7, 2022.
View at: Publisher Site | Google Scholar
M. Decker, T. Lammens, A. Ferster et al., “Functional classification of RUNX1 variants in familial platelet disorder with associated myeloid malignancies,” Leukemia, vol. 35, no. 11, pp. 3304–3308, 2021.
View at: Publisher Site | Google Scholar
M. Decker, A. Agarwal, A. Benneche et al., “Validation and clinical application of transactivation assays for RUNX1 variant classification,” Blood Advances, vol. 6, no. 11, pp. 3195–3200, 2022.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2023 Nikolaj Juul Nitschke et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies

Views

664

Downloads

280

Citations