University of Groningen
Genetic defects in myeloid malignancies and preleukemic conditions Berger, Gerbrig
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record
Publication date: 2019
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA): Berger, G. (2019). Genetic defects in myeloid malignancies and preleukemic conditions. Rijksuniversiteit Groningen.
Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Download date: 26-09-2021 (Submitted) 3. AB Mulder, S Ogawa, JHA Martens, JH Jansen and E Vellenga and JH Jansen JHA Martens, S Ogawa, AB Mulder, risk and result from an aberrant an aberrant from risk and result G Berger, M Gerritsen, TN Koorenhof-Scheele, G Yi, LI Kroeze, G Yi, LI Kroeze, TN Koorenhof-Scheele, M Gerritsen, G Berger, heme-metabolism gene program Ringsideroblasts in acute myeloid myeloid in acute Ringsideroblasts M Stevens-Kroef, K Yoshida, E van den Berg, H Schepers, G Huls, G Huls, H Schepers, den Berg, E van K Yoshida, M Stevens-Kroef, leukemia are associated with adverse adverse with associated are leukemia
Chapter 3
Abstract Ringsideroblasts (RS) emerge from aberrant erythroid differentiation, resulting in excessive mitochondrial iron accumulation. This is a characteristic feature of myelodysplastic syndromes (MDS) with mutations in spliceosome gene SF3B1, but RS is also observed in patients with acute myeloid leukemia (AML). We therefore characterized the presence of RS in a cohort of AML patients. The RS-AML subgroup was enriched with patients in the ELN adverse-risk category (55%). In line with this finding, 35% of all RS-AML cases had complex cytogenetic aberrancies, and TP53 was most recurrently mutated in this cohort (42%), followed by DNMT3A (29%), RUNX1 (21%) and ASXL1 (19%). In contrast to RS-MDS, the incidence of SF3B1 mutations was low (8%). Whole-exome sequencing and SNP array analysis on a subset of patients showed that the RS phenotype did not result from a single-gene defect. Shared genetic defects between erythroblasts and total mononuclear cell fraction within the same 3 patient indicate common ancestry for the erythroid lineage and the myeloid blast cells in RS-AML patients. RNA sequencing of CD34+ AML cells revealed differential gene expression between RS-AML and a separate AML cohort, including genes involved in megakaryocytic/erythroid differentiation and mRNA splicing. Furthermore, several heme-metabolism-related genes were found to be upregulated in RS-AML, as was observed in SF3B1mut MDS. These results demonstrate that erythroblasts share ancestry with malignant myeloid blast cells in RS-AML. Although the genetic background of RS- AML differs from that of RS-MDS, downstream effector pathways may be comparable, providing a possible explanation for presence of RS in AML.
ingsideroblasts (RS) are erythroid in low-risk MDS, which is characterized precursor cells that accumulate by a stable clinical course and a low risk R excessive mitochondrial iron of leukemic transformation.(8,13) As a and can be observed in bone marrow core component of the U2 small nuclear smears associated with multiple medical ribonucleoprotein particle (snRNP), conditions.(1) Acquired presence of RS is a SF3B1 is essential for pre-RNA splicing. characteristic feature in myelodysplastic (14) The molecular mechanism by which syndrome (MDS) subtypes, including SF3B1 mutations result in RS formation MDS with single lineage dysplasia is not yet fully understood. A proposed (MDS-RS-SLD), multilineage dysplasia mechanism is that specific patterns of (MDS-RS-MLD) and in combination missplicing result in altered expression with the presence of thrombocytosis of genes that are essential for correct (MDS/MPN-RS-T).(2) Non-malignant programming of erythropoiesis.(15,16) causes of RS include several drugs, The relationship between toxins, alcohol, copper deficiency and genetic defects in SF3B1 and the RS- congenital sideroblastic anemia.(3) This phenotype is not one-to-one; in 10-20% latter group comprises conditions of the MDS-RS patients no mutation caused by inborn defects in genes in the SF3B1 gene is detected.(8-12) that operate in several mitochondrial Moreover, RS can also be present in a pathways, including ALAS2, ABCB7, subset of AML patients(17), while SF3B1 SLC25A38 and HSPA9.(4-7) mutations are infrequent findings in In MDS, the RS phenotype is this disease.(10,17,18) Besides SF3B1, strongly correlated with mutations in the only other correlation between splicing factor 3B subunit 1 (SF3B1), a gene defect and the RS phenotype with an incidence higher than 80%.(8-12) was described for PRPF8, for which SF3B1 mutations are usually observed mutations are reported in ~3% of the
66 Ringsideroblasts in acute myeloid leukemia
cases.(19) Other spliceosomal genes that lymphoprep (PAA, Cölbe, Germany) are more frequently mutated in myeloid according to standard procedures. malignancies, including SRSF2, U2AF1, To separate cell fractions, following ZRSR2, have not been implicated in the thawing of viably frozen MNCs, cells RS-phenotype.(9) were washed and stained with a panel In the present study we deter- of antibodies and sorted for purification mined the prevalence of ringsideroblasts of different cell fractions (antibodies in various ontogenic AML subtypes and against CD3, CD34, CD71 and the association of the RS phenotype in CD235a surface markers (conjugates AML with adverse risk characteristics. CD3-FITC (cat. 345763), CD34-APC To identify the landscape of genomic (cat. 555824) or CD34-PE Cy7 (cat. defects that underlies the RS phenotype 348811) or CD34-FITC (cat. 345801, in AML, we performed whole exome BioLegend, Uithoorn, the Netherlands), sequencing, targeted sequencing and CD71-BV786 (cat. 563768) or CD71- SNP-array analysis. Finally, to identify APC (cat. 551374) or CD71-PE (cat. 3 differential expression of genes asso- 555537) and CD235a-BV421 (cat. ciated with the RS phenotype in AML, 562938) or CD235a-APC (cat. 551336), we performed RNA sequencing on antibodies were obtained from BD CD34+-selected AML cells. Bioscience (Breda, The Netherlands) unless otherwise indicated)). Analysis and Material & methods sorting of various cell fractions was performed on MoFlo XDP or Astrios Patients and data collection For this (Dako Cytomation, Carpinteria, CA, USA). study, we collected data from 126 AML Single viable cells were selected based and high-risk MDS (≥10% bone marrow on forward and side scatter profiles in blasts) who were diagnosed between combination with negativity for DAPI January 2000 and April 2018 at the or PI (both Sigma-Aldrich, Saint Louis, University Medical Center Groningen. MO, USA). Blast fraction was defined as The inclusion criterion was the reported CD34 positive, T cell fraction as CD3 presence of ringsideroblasts in the positive and erythroblast fraction as diagnostic bone marrow smear. Patients CD71/CD235a positive. Sorting purity with previously reported MDS-RS was defined as ≥95% and confirmed by were excluded. Diagnosis and risk reanalysis. classification was revised based upon DNA isolation and World Health Organization classification amplification. Genomic DNA from (2016)(2) and European Leukemia Net various cell fractions was extracted with (ELN) recommendations (2017)(20). Bone the NucleoSpin Tissue kit (Macherey- marrow (BM) and/or peripheral blood Nagel, Düren, Germany) according to (PB) from patients were biobanked after the manufacturer’s instructions. In informed consent for investigational use. case of insufficient yield, a maximum The study was conducted in accordance of 70ng DNA was amplified using the with the Declaration of Helsinki and Qiagen REPLI-g kit (Qiagen, Venlo, the institutional guidelines and regulations. Netherlands) in one reaction, according Morphologic and cytogenetic analyses to the manufacturer’s protocol. were accomplished following standard Targeted deep sequencing procedures. using a myeloid gene panel. Targeted Sorting of cell fractions. sequencing of DNA derived from BM The mononuclear cell (MNC) fraction or PB samples obtained at diagnosis was from BM and/or PB was obtained by carried out using the myeloid TruSight density gradient centrifugation using sequencing panel (Illumina, San Diego,
67 Chapter 3
CA, USA). Library preparation was per- (Agilent Technologies, Santa Clara, CA, formed according to the manufacturer’s USA), massively parallel sequencing protocol (Illumina). Aligning and filter- was performed on enriched exome ing of sequence data was performed fragments using the HiSeq 2500 using NextGENe version 2.3.4.2 (Soft- platform (Illumina, San Diego, CA, USA). Genetics, Pennsylvania, US). Cartagenia Alignment of sequences and calling of Bench Lab NGS (Agilent, Santa Clara, mutations was executed our previously CA, USA) was used for analysis of the described in-house pipelines(9,21), with resulting vcf file. Sequencing artifacts minor modifications. The resultant data were excluded using a threshold of 5%. file was analyzed for the presence of A minimal variant read depth of 20 germline variants in genes that have been reads was set as criterion. Variants that previously implicated in sideroblastic frequently occur in the general healthy anemia (including ALAS2, SLC25A38, population (>2% 1000 Genome Phase 1, FECH, MFRN1, ISCU, ISCA1/2, 3 ESP6500 and dbSNP, and >5% Genome NIFS, HSP70, HSPA9, HSCB, GLRX5, of the Netherlands) were excluded from ABCB7, IRBP1, GLRX5, PUS1, further analysis. YARS2).(23) Subsequently, DNA isolated Microarray-based genomic from sorted autologous T cells was used profiling. Microarray-based genomic as a constitutive reference to exclude profiling on MNC and erythroblast germline variants. Filtering strategy fractions was carried out with the and variant calling was performed as CytoScan HD array platform (Affymetrix, previously described.(24) Inc., Santa Clara, CA, USA) in agreement Confirmation of mutations in with the manufacturer’s reference. erythroblasts. For a subset of patients, Data analysis was performed using the presence of somatic variants in Chromosome Analysis Suite software TP53 and SRSF2 identified by WES package (Affymetrix) and annotations were validated and quantified in the of reference genome build GRCh37 erythroblast fraction using amplicon- (hg19). Comprehensive analysis based deep sequencing on an Ion and interpretation of the obtained Torrent Personal Genome Machine microarray genomic profiling data was (Thermo Fisher Scientific, Waltham, performed using a previously described MA, USA). The sequencing procedure filtering pipeline and according to using an automated robotic workflow criteria that have been described was performed as previously described previously.(21) Aberrations meeting (Sandmann, PlosOne, 2017). The obtained these criteria were included for genomic sequencing data was mapped to the profiling and described in accordance reference genome build GRCh37 (hg19). with the standardized ISCN 2016 Variant calling was performed using the nomenclature system.(22) Visualization SeqNext module of the Sequence pilot of the resultant genomic profiles was software, version 4.2.2 (JSI Medical performed using NEXUS software Systems, Ettenheim, Germany). (Nexus Copy Number 8.0, BioDiscovery, RNA Extraction and Illumina El Segundo, CA, USA). high-throughput sequencing. RNA Whole exome sequencing. was isolated by separation of the Whole exome sequencing (WES) aqueous phase by TRIzol Reagent to an average depth of 143x was (Thermo Fisher) according to the performed on DNA isolated from manufactures protocol. The aqueous diagnostic BM-MNC (n=13) or PB- phase was subsequently mixed 1:1 with MNC (n=3) samples. Following exome 70% ethanol and isolation was continued capturing using Human All Exon V5 using the RNeasy mini kit (Qiagen)
68 Ringsideroblasts in acute myeloid leukemia
Table 1 – Clinical characteristics of AML with RS phenotype
All RS cases 1-4% RS 5-14% RS ≥15% RS
Total 126 53 42% 28 22% 45 36% Age (years) Median 67 67 63 70 Range 32-87 40-81 32-78 43-87 Sex Male 84 67% 32 60% 20 71% 32 71% Female 42 33% 21 40% 8 29% 13 29% Type of disease de novo AML 70 56% 35 66% 12 43% 23 51% 3 sAML 20 16% 6 11% 8 29% 6 13% t-MN 17 14% 7 13% 3 11% 7 16% Other 19 15% 5 9% 5 18% 9 20% WHO diagnosis AML with MDS- 47 37% 14 26% 16 57% 17 38% related changes AML NOS 24 19% 13 25% 3 11% 8 18% t-MN 17 14% 7 13% 3 11% 7 16% MDS-EB2 17 14% 5 9% 4 14% 8 18% AML with recurrent 16 13 % 13 25 % 1 4% 2 4% abnormalities Other 5 4% 1 2% 1 4% 3 7% ELN risk score Favorable 12 10% 11 21% 1 4% 0 0% Intermediate 6 5% 4 8% 2 7% 0 0% Intermediate* 34 27% 17 32% 6 21% 11 24 % Adverse 71 56% 21 40% 18 64% 32 71% Unknown 3 2% 0 0% 1 4% 2 4% BM blasts, % Median 32% 34% 24% 32% Range 10%-91% 10-88% 11-88% 10-91% Erythroblasts, % Median 21% 23% 21% 17% Range 1-64% 4-64% 2-50% 1-59%
Data denoted as number (percentage), unless otherwise stated. * No evaluation for ASXL1, RUNX1 and TP53. Abbreviations: AML – acute myeloid leukemia, t-MN – therapy-related myeloid neoplasm, NOS – not otherwise specified, sAML- secondary AML, MDS-EB2 – myelodysplastic syndrome with excess blasts type 2, ELN– European Leukemia Net, WHO – World Health Organization, BM – bone marrow, RS - ringsideroblasts
69 Chapter 3
including performing on-column a heatmap. To assess the enriched DNaseI treatment. RNA libraries annotation for deregulated genes, were prepared using the KAPA RNA we performed functional enrichment HyperPrep Kit with RiboErase (HMR) analysis with the Metascape tool, and according to the manufactures protocol those terms showing p-values below (KR1351 – v1.16, Roche Sequencing 0.01 were considered significantly over- Solutions). In short; 25ng -1ug input represented. RNA was depleted from ribosomal Statistical Analysis. Bivariate RNA by oligo hybridization, RNaseH correlations were made using a Pearson treatment and DNase digestion. rRNA- correlation (continuous variables) or depleted RNA was fragmented to ~200 Spearman correlation (categorical bp fragments and first strand synthesis variables). A p-value <0.05 was used to was performed using random primers. define statistical significance. Statistical Second strand synthesis was performed calculations were performed using 3 using dUTP for strand specificity. After Prism version 6.0. adapter ligation, library amplification was performed and the number of Results cycles was dependent on the amount of starting material. Fragment sizes and Clinical Characteristics quality was checked on a bioanalyser In our previous study on clonal evolution using a high sensitivity DNA Chip in therapy-related myeloid neoplasms (Agilent). Samples were sequenced on (t-MNs)(24), ringsideroblasts (RS) were the Illumina HiSeq 2000. Finally, each observed frequently in bone marrow eligible library was subjected to 2×43 smears. To study the RS phenotype in bp paired-end sequencing (PE43) on an more detail in a more comprehensive Illumina NextSeq 500 system. group of myeloid neoplasms (MNs), in RNA-Seq analysis. The hg19 the present study we collected clinical reference genome was first indexed data was collected for on a cohort of by STAR aligner with UCSC gene patients (N = 126) consisting of group annotation. The resulting RNA- of 126 AML and high-risk MDS seq reads were mapped to the hg19 (≥10% BM blasts) patients hereafter genome using STAR with two- also indicated as ‘AML patients’. pass mode, and the gene-level read Patients reported here with RS (≥1%) counts were enumerated at the same in the diagnostic bone marrow smear, time. The DESeq2 tool was used to excluded those with a documented examine differentially expressed genes prior clinical history of MDS-RS. The by conducting pair-wise comparison median blast percentage in this cohort between different groups. Only genes was 32% (range 10-91%), two-third of with a Benjamini-Hochberg-adjusted the patients were male and the median p-value <0.1 and fold change >1.5 were age at diagnosis was 67 years (range considered significantly deregulated. 32-87). The majority of patients were Principal component analysis (PCA) diagnosed with de novo AML (55.6%) and t-Distributed Stochastic Neighbor and AML with myelodysplasia-related Embedding (t-SNE) were used to probe changes was the most common WHO the transcriptomic relationships between (2016) subtype (37.3%, Table 1). these groups. To group differential Although highest RS percentages were genes with similar expression patterns, observed in cases with lower blast counts, the k-means clustering approach the blast count did not significantly was performed based on z-score correlate with RS percentage (Pearson normalization and then displayed as r = 0.16, p = 0.07; Figure 1A). Patients
70 Ringsideroblasts in acute myeloid leukemia