Repeat to gene expression ratios in leukemic blast cells can stratify risk prediction in acute myeloid leukemia

Abstract

Background: Repeat elements constitute a large proportion of the human genome and recent evidence indicates that repeat element expression has functional roles in both physiological and pathological states. Specifically for can‐ cer, transcription of endogenous retrotransposons is often suppressed to attenuate an anti‐tumor immune response, whereas aberrant expression of heterochromatin‐derived satellite RNA has been identified as a tumor driver. These insights demonstrate separate functions for the dysregulation of distinct repeat subclasses in either the attenuation or progression of human solid tumors. For hematopoietic malignancies, such as Acute Myeloid Leukemia (AML), only very few studies on the expression/dysregulation of repeat elements were done. Methods: To study the expression of repeat elements in AML, we performed total‐RNA sequencing of healthy CD34 + cells and of leukemic blast cells from primary AML patient material. We also developed an integrative bioin‐ formatic approach that can quantify the expression of repeat transcripts from all repeat subclasses (SINE/ALU, LINE, ERV and satellites) in relation to the expression of gene and other non‐repeat transcripts (i.e. R/G ratio). This novel approach can be used as an instructive signature for repeat element expression and has been extended to the analy‐ sis of poly(A)‐RNA sequencing datasets from Blueprint and TCGA consortia that together comprise 120 AML patient samples. Results: We identified that repeat element expression is generally down‐regulated during hematopoietic differentia‐ tion and that relative changes in repeat to gene expression can stratify risk prediction of AML patients and correlate with overall survival probabilities. A high R/G ratio identifies AML patient subgroups with a favorable prognosis, whereas a low R/G ratio is prevalent in AML patient subgroups with a poor prognosis. Conclusions: We developed an integrative bioinformatic approach that defines a general model for the analysis of repeat element dysregulation in physiological and pathological development. We find that changes in repeat to gene expression (i.e. R/G ratios) correlate with hematopoietic differentiation and can sub‐stratify AML patients into low‐risk and high‐risk subgroups. Thus, the definition of a R/G ratio can serve as a valuable biomarker for AML and could also provide insights into differential patient response to epigenetic drug treatment. © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Background Repeat elements constitute around 50% of the mammalian genome and, in addition to their role as insertional mutagens, have been involved in functions for genome evolution and stability [34, 42], embryonic development [22, 31, 43], immune response [12] and in fine-tuning of gene regulatory networks [17, 45, 51]. While the Open Access *Correspondence: jenuwein@ie‐freiburg.mpg.de 1 Max Planck Institute of Immunobiology and Epigenetics, Freiburg, Germany Full list of author information is available at the end of the article Page 2 of 20 Onishi‐Seebacher et al. BMC Med Genomics (2021) 14:166 majority of repeat elements are permutated and silent, it has been estimated that around 10–15% maintain transcriptional competence [7, 53]. In addition, several subclasses of repeat elements have also been shown to be deregulated in cancer [8]. Derepression of centromeric satellite repeats (ALR, GSAT, HSATII), for example, have been found in multiple solid human cancers [57, 66]. Deregulated satellite RNA can induce repeat expansion at pericentric heterochromatin [4], generate genomic instabilities [66, 67] and aberrantly sequester some key epigenetic factors, such as YBX1 [35] or the Polycomb complex PRC1 [28]. By contrast, therapeutic activation of endogenous retroviruses (ERV) has been found to trigger an innate immune response in human prostate and breast cancer cell lines [11, 47], suggesting that some types of cancer maintain a suppressed level of ERV expression in order to evade immune surveillance. The analysis of repeat element expression/dysregulation has recently been extended to several hematological malignancies, particularly Acute Myeloid Leukemia [13, 14]. Acute Myeloid Leukemia (AML) is a heterogeneous cancer of the myeloid lineage of blood cells, which, in addition to prevalent genetic lesions, also exhibit epigenetic alterations in DNA methylation and histone modifications [63, 65]. Treatments against AML with ‘epigenetic drugs’, such as the DNMT inhibitors Azacytidine and Decitabine [41], or combination therapies with DNMTi and ATRA [26] or DNMTi and HDACi [5] are effective and are under clinical investigation. However, the mechanism behind these epigenetic alterations and how these epigenetic drugs target leukemic cells are only now starting to become apparent. In addition, no clear biomarkers that can predict AML treatment response have yet been identified. Therefore, we set out to determine repeat element expression in AML, by high-throughput sequencing of total-RNA of AML patient samples compared to healthy blood controls. The repeat expression analysis was further extended to poly(A)-RNA sequencing data sets of AML patient samples from the Blueprint [10] and TCGA consortia [9]. Our integrative analysis indicates that repeat element expression is dynamically regulated during hematopoiesis and that differences in repeat to gene expression ratios associate with distinct subgroups of AML patients and correlate with overall prognosis. Through this approach, we identified repeat to gene expression ratios (R/G ratios) as a signature in AML that may serve as a prognostic biomarker through which epigenetic drug efficacy may also be tested. Methods Purification of primary human CD34 + cells Unused human blood transfusions were obtained from the University of Freiburg Medical Center (Transfusionsmedizin des Universitätsklinikums Freiburg). 50 ml blood was diluted 1:1 in PBS and separated on a Ficoll gradient (Pancoll, Ibian Technologies). The interphase (buffy coat) was transferred to a 50 ml Falcon tube, the cell suspension was washed in PBS and cells were processed for MACS sorting which was done with anti-CD34 MicroBeads (Miltenyibiotec). Around 0.5–1 × 106 CD34 + cells are typically obtained from one buffy coat. Cells were resuspended in 1 ml Trizol (Sigma) and kept at − 80 °C until RNA extraction. Samples of pre-therapeutic and newly diagnosed AML patients were obtained from the Hematology, Oncology and Stem cell Transplantation department of the University of Freiburg, Medical Center. Leukemic blast cells were isolated from peripheral blood of 7 patients and only for one patient were the leukemic blast cells isolated from bone marrow. For 6 patients, both bone marrow and peripheral blood blast percentages were analyzed (in the remaining 2 patients, no bone marrow puncture could be performed), indicating that the median bone marrow and blood blast percentages are highly comparable (61% and 59%) (see Additional file 11: Table 1). Leukemic blast cells were purified by MACS sorting with both anti-CD34 and anti-CD117 MicroBeads (Miltenyibiotec). Isolation of total RNA from primary human CD34 + cells Total RNA from around 0.5–1 × 106 cells was isolated with Trizol (Sigma), digested with TurboDNAse (Ambion), washed and resuspended in H2O. The RNA integrity was confirmed with an RNA Bioanalyzer (Agilent). HiSeq RNA sequencing of non‐poly(A) selected total RNA Comparable amounts (250 ng–1 ug) of total RNA was converted into non-poly(A) selected, ribosomal RNA depleted (TruSeq RNA Library Prep Kit v2) cDNA libraries following Illumina protocols. The cDNA libraries were sequenced on a NextSeq 500 (Illumina) or HiSeq 2500 (Illumina) platform using a 75 bp paired-end approach to give a coverage of 26–78 million reads per sample. Blueprint data sets In the Blueprint-HSC dataset [10], an ultra-low RNA input kit (Clontech SMARTer Ultra Low RNA Kit) for library preparation was used [52], which may also enable the detection of repeat element transcripts that are only weakly poly-adenylated. The Blueprint-HSC cDNA libraries were done in an unstranded manner using 100 bp paired-end reads (100 bp paired–end reads also for Blueprint-MPP and 75 bp paired-end reads for Blueprint-CMP) and had a coverage of 42–103 million reads per sample. The Blueprint-AML dataset enriched for poly(A)-RNA using standard mRNA library generation Page 3 of 20 Onishi‐Seebacher et al. BMC Med Genomics (2021) 14:166 protocols that can analyze transcripts in a strand-specific manner. The Blueprint-AML libraries used 75 bp pairedend reads and had a coverage of 53–79 million reads per sample. In accordance with the Ft Lauderdale agreement, these data are available for additional analysis under https:// europ epmc. org/ artic les/ PMC63 63099. TCGA data sets For the TCGA-AML dataset, mRNA purification by oligo-dT was supplemented by the use of random hexamer primer