Quantitative expression of regulatory and differentiation-related genes in the key steps of human hematopoiesis: The LeukoStage Database

Differentiation ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Differentiation journal homepage: www.elsevier.com/locate/diff

Quantitative expression of regulatory and differentiation-related genes in the key steps of human hematopoiesis: The LeukoStage Database K. Polgárová a, M. Vášková a, E. Froňková a, L. Slámová a,b, T. Kalina a,b, E. Mejstříková a,b, A. Dobiášová a, O. Hrušák a,b,n a

CLIP – Childhood Leukemia Investigation Prague, Department of Pediatric Hematology and Oncology, 2nd Faculty of Medicine, Charles University Prague, Czech Republic b University Hospital Motol, Czech Republic

ar t ic l e i nf o

a b s t r a c t

Article history: Received 15 April 2015 Received in revised form 23 October 2015 Accepted 13 November 2015

Differentiation during hematopoiesis leads to the generation of many cell types with specific functions. At various stages of maturation, the cells may change pathologically, leading to diseases including acute leukemias (ALs). Expression levels of regulatory molecules (such as the IKZF, GATA, HOX, FOX, NOTCH and CEBP families, as well as SPI-1/PU1 and PAX5) and lineage-specific molecules (including CD2, CD14, CD79A, and BLNK) may be compared between pathological and physiological cells. Although the key steps of differentiation are known, the available databases focus mainly on fully differentiated cells as a reference. Precursor cells may be a more appropriate reference point for diseases that evolve at immature stages. Therefore, we developed a quantitative real-time polymerase chain reaction (qPCR) array to investigate 90 genes that are characteristic of the lymphoid or myeloid lineages and/or are thought to be involved in their regulation. Using this array, sorted cells of granulocytic, monocytic, T and B lineages were analyzed. For each of these lineages, 3–5 differentiation stages were selected (17 stages total), and cells were sorted from 3 different donors per stage. The qPCR results were compared to similarly processed AL cells of lymphoblastic (n ¼ 18) or myeloid (n ¼6) origins and biphenotypic AL cells of B cell origin with myeloid involvement (n ¼ 5). Molecules characteristic of each lineage were found. In addition, cells of a newly discovered switching lymphoblastic AL (swALL) were sorted at various phases during the supposed transdifferentiation from an immature B cell to a monocytic phenotype. As demonstrated previously, gene expression changed along with the immunophenotype. The qPCR data are publicly available in the LeukoStage Database in which gene expression in malignant and non-malignant cells of different lineages can be explored graphically and differentially expressed genes can be identified. In addition, the LeukoStage Database can aid the functional analyses of next-generation sequencing data. & 2015 International Society of Differentiation. Published by Elsevier B.V. All rights reserved.

Keywords: Hematopoiesis Gene expression B and T lymphocytes Differentiation plasticity Lineage promiscuity Acute leukemia

1. Introduction Hematopoiesis is a complex process during which undifferentiated progenitor cells continuously lose part of their Abbreviations: A(L)L, acute (lymphoblastic) leukemia; AML, acute myeloid leukemia; B-ly, B-lymphoid lineage; BM, bone marrow; Cq, quantitation cycle; FACS, fluorescence activated cell sorting; Ly, lymphoid lineage; MIQE, minimum information for publication of quantitative real-time PCR experiments; My, myeloid lineage; ncRNA, non-coding RNA; NGS, next-generation sequencing; PB, peripheral blood; pB-ALL, precursor B-cell acute ymphoblastic leukemia; PCA, principal component analysis; qPCR, quantitative real time PCR; RG, reference gene; swALL, switching acute lymphoblastic leukemia; T-ly, T-lymphoid lineage; T-ALL, T-cell acute lymphoblastic leukemia; WTA, whole transcriptome amplification n Corresponding author at: CLIP, Childhood Leukemia Investigation Prague, Department of Pediatric Hematology and Oncology, 2nd Faculty of Medicine, Charles University Prague, V Uvalu 84, 150 06 Prague, Czech Republic. E-mail address: [email protected] (O. Hrušák).

multilineage potential, leading to the formation of mature cells of different lineages. Acute leukemias (ALs), the most common group of childhood malignancies, occur mostly during the early stages of differentiation. The AL cells maintain some characteristics of the presumed cell of origin, and their differentiation potential may be different. Precise comparisons, however, are hampered by the paucity of in-depth investigations of sorted, non-malignant blood cells. Several differentiation models have been published, the most common being the classical tree representation with strict definitions of individual branches and restrictive pathways (Larsson and Karlsson, 2005). However, recently, unexpected plasticity of hematopoietic cells has been reported (Graf, 2002), and several new models have been proposed (Ceredig et al., 2009). It is unclear whether the lineage infidelity and promiscuous behavior described in some AL cells may also occur under normal physiological conditions or whether they occur only in malignant or manipulated cells.

http://dx.doi.org/10.1016/j.diff.2015.11.003 Join the International Society for Differentiation (www.isdifferentiation.org) 0301-4681/& 2015 International Society of Differentiation. Published by Elsevier B.V. All rights reserved.

Please cite this article as: Polgárová, K., et al., Quantitative expression of regulatory and differentiation-related genes in the key steps of human hematopoiesis: The LeukoStage Database. Differentiation (2015), http://dx.doi.org/10.1016/j.diff.2015.11.003i

2

K. Polgárová et al. / Differentiation ∎ (∎∎∎∎) ∎∎∎–∎∎∎

To account for any kind of alternative behavior, the physiological regulation of differentiation must be described precisely. In the last decade, many studies have described the changes in mRNA expression that accompany the differentiation of particular hematopoietic lineages (Liu et al., 2007; Novershtern et al., 2011). Although microarrays are considered the best choice for genomewide analysis and have been extensively used to analyze gene expression during hematopoietic differentiation, the interpretation of the results can be troublesome. Data processing is known to be a potential source of bias, and the reliability of quantification data also remains an intensively discussed problem because the signal strength is probe dependent and cross hybridization can occur (Draghici et al., 2006). Analysis of freely available data shows different levels of mRNA expression when using different genespecific probes depending on probe length and other factors (Chou et al., 2004; Steger et al., 2011). Moreover, the level of correlation between microarray results and those from other confirmative methods depends, in addition to other factors, on the microarray platform and sample processing protocols used. New approaches, particularly next-generation sequencing (NGS), are becoming available that allow whole genome or whole transcriptome sequencing. Despite the initial excitement over these technologies, it is clear that substantial biases may again arise from the sample processing and data analysis methods. Even basic steps of the procedure, such as nucleic acid isolation or fragmentation methods, may be a source of misleading results (van Heesch et al., 2013). Several studies have also described sequence-dependent biases (specifically effects due to GC content) (Benjamini and Speed, 2012) and sensitivity issues resulting from high read counts from a restricted subset of highly expressed genes, thus leaving the remaining genes under-represented (Raabe et al., 2014). Describing the exact changes in mRNA expression that occur during hematopoietic differentiation may help to reveal unknown levels of regulation. Studies with this aim are limited by low cell numbers of particular populations (i.e., less than 1000 cells), such as most early hematopoietic progenitors. Methods to optimize the detection of mRNA expression in rare cell populations include whole transcriptome amplification (WTA) and gene-specific preamplification. Neither gene expression profiling nor NGS can clearly assess the bias introduced by these techniques. Quantitative real-time PCR (qPCR) can provide very accurate results and thus is generally considered the gold standard for confirmation of microarray and NGS results. Moreover, the dynamic range of qPCR exceeds microarrays as well as NGS performed at the commonly used depth of sequencing. Reports evaluating expression changes of individual genes during the differentiation of specific hematopoietic lineages have already been published. However, in most of the studies where qPCR was used, only one lineage (monocytes to macrophages, differentiation into Th17 cells, etc.) or a limited number of genes were included (Lehtonen et al., 2007; Tuomela et al., 2012). In addition, mouse immune cells are characterized in the Immunological Genome (ImmGen) Project, in which several databases are available online (Robinette et al., 2015). In this study, we aimed to develop a qPCR array that would precisely quantify the expression of key genes during different stages of B-lymphoid (B-ly), T-lymphoid (T-ly), monocytic and granulocytic lineage development and their corresponding malignant counterparts. The resulting gene expression data illustrate the behavior of particular genes during hematopoiesis but can also serve as a tool to optimize newly developed NGS-based gene expression platforms. Moreover, the resulting LeukoStage Database of gene expression can be used to better identify the normal physiological counterparts of hematopoietic malignancies.

2. Materials and methods 2.1. Samples Cell-type subsets separated from the peripheral blood (PB) of healthy donors included naive B-ly, mature monocytes and granulocytes. Earlier developmental stages of B-ly and myeloid (My) lineages were obtained from the bone marrow (BM) of healthy donors without infection, immune system activation, systemic disease, hematological malignancies or other possibly considerable diseases. T-lymphocytes and their immature precursors were obtained from thymi removed during cardiosurgery for congenital heart disease from otherwise healthy patients. At least 3 specimens for every cell-type subset from physiological samples were included. Malignant cells were separated from diagnostic samples of B-cell precursor acute lymphoblastic leukemia (pB-ALL, n ¼14), T-lymphoblastic leukemia (T-ALL, n¼ 4), acute myeloid leukemia (AML, n¼ 6) and acute biphenotypic leukemia (pB/My-AL, n ¼5) that were sent to our laboratory. All malignant and non-malignant specimens were separated after written informed consent was obtained. All procedures were performed in accordance with the ethical standards of the Institutional Review Board and the Helsinki declaration. 2.2. Cell lines pB-ALL (REH, SP-B15, RS4;11), T-ALL (Jurkat) and AML (NB4, Molm13, K562, Kasumi) cell lines were used; they were obtained from the German Tissue Culture Collection (DSMZ; Jurkat, REH, SUP-B15, RS4;11, NB4, and Kasumi) or the American Tissue Culture Collection (ATCC; Molm13, respectively). Cell lines were cultivated at 37 °C in 5% CO2 in RPMI1640 supplemented with fetal bovine serum, streptomycin and penicillin. For gene expression analysis, cells between 4 and 10 passages after defrosting were used. 2.3. Gene selection Ninety five genes were selected using publicly available microarray gene expression data (Abbas et al., 2005; van Zelm, 2012) along with our own preliminary or published data (Slamova et al., 2014). The genes included known hematopoietic cell fate regulators, such as PU.1, IKZF1, PAX5, the CEBP family of genes, etc.; genes with known involvement in leukemogenesis, e.g., genes from the HOX family; genes encoding cell surface markers such as CD19 or CD3e to confirm appropriate cell sorting; genes with uncertain functions that have a possible regulatory role in hematopoiesis, such as RAM; and finally internal control or reference genes (RGs: GAPDH, HPRT1, ABL1, GUSB, and B2M). The complete list including assay numbers is available in the supplementary material. 2.4. Cell-type subset separation Mature B-ly and mature monocytes from PB were separated using magnetic beads (the Human Naive B-cell Enrichment Kit based on negative selection and the CD14 Positive Selection Kit, both EasySep™ from STEMCELL Technologies) following the manufacturer’s instructions. The rest of the malignant and non-malignant cell-type subsets were separated by fluorescence-activated cell sorting (FACS). The sorting strategy, including gating panels and the antibodies used, is described in Table 1 and the Supplementary material. 2.5. RNA extraction, reverse transcription, preamplification Cells were subjected to RNA extraction (RNeasy Micro Kit,

Please cite this article as: Polgárová, K., et al., Quantitative expression of regulatory and differentiation-related genes in the key steps of human hematopoiesis: The LeukoStage Database. Differentiation (2015), http://dx.doi.org/10.1016/j.diff.2015.11.003i

K. Polgárová et al. / Differentiation ∎ (∎∎∎∎) ∎∎∎–∎∎∎ Table 1 Immunophenotypical definition of sorted non-malignant specimens. Stage Immunophenotype

B2 B3 B4

CD22posCD34posCD19negCD10pos (B1a) and CD22posCD34posCD19posCD10neg (B1b) CD34posCD19posCD10posCD20neg CD34negCD19posCD10pos CD19posCD10negCD20posCD27neg

T1 T2 T3 T4 Mo1 Mo2 Mo3

CD34posCD1anegCD3negCD7pos CD34posCD1aposCD3negCD7pos CD3posCD4posCD8pos CD3posCD4negCD8pos CD33highCD4posCD14neg CD33highCD4posCD14pos CD33highCD4posCD14pos

Gr0 Gr1 Gr2

CD13posCD16negCD117pos CD13highCD16negCD117neg CD13negCD15highCD16neg (Gr2a); CD13lowCD15highCD16low (Gr2b) CD13highCD16highCD117neg CD13highCD15high CD16pos

B1

Gr3 Gr4

Tissue Bone marrow Bone marrow Bone marrow Peripheral blood Thymus Thymus Thymus Thymus Bone marrow Bone marrow Peripheral blood Bone marrow Bone marrow Bone marrow Bone marrow Peripheral blood

3

3. Results and discussion 3.1. qPCR expression array We assembled a qPCR expression array based on commercially available TaqMans assays. For location of the primers and probes and further information search the manufacturer’s website. Reference genes were chosen using the NormFinder tool, according to which the combination of RGs with the lowest interand intra-group variability was HPRT1 and GUSB. Thus, the mean of these two RGs was used for normalization. We also tested whether possible differences induced by the preamplification step would interfere with our statistical methods. Our analysis showed a good correlation between the results from preamplified and non-amplified samples (p ¼1.8  10  28, r ¼0.861, bias ¼0.327) according to the Pearson correlation coefficient and Bland-Altman comparison (supp. Fig. 3). In conclusion, the gene-specific preamplification step did not introduce further bias to the results, in contrast to the bias that occurs during whole transcriptome amplification. This approach allows the use of fewer than 1000 cells per sample, as opposed to the at least 105 cells necessary for reliable expression data without a preamplification step. 3.2. Clustering and lineage-associated differences

Qiagen). DNase-treated RNA was then transcribed into cDNA (iSCRIPT, BioRad) and preamplified using the Whole Transcriptome Amplification Kit 2 (Sigma-Aldrich) and the gene-specific TaqMan PreAMP system (Life Technologies) according to the manufacturers’ instructions (for further details, see the Supplementary material). Preamplified and diluted cDNA was then used as a template for qPCR. The possibility of bias caused by the preamplification step was ruled out by comparing expression data from the preamplified and non-amplified samples as described in the results section. 2.6. Real-time PCR, gene expression analysis The qPCR experiments are described according to MIQE recommendations (Bustin et al., 2009). The qPCR system was based on commercially available hydrolytic probes (TaqMans gene expression assays, ThermoFisher scientific-Life Technologies). The probe list and the qPCR protocol are available in the supplementary material (Suppl. Tables 1 and 3). For Cq value assessment, the LinReg software was used to avoid bias resulting from subjective evaluation (Ruijter et al., 2009). Normalized gene expression was then calculated using the ΔCq method. An appropriate combination of internal control genes was obtained by performing intraand intergroup variation analysis using the NormFinder tool (Andersen et al., 2004). 2.7. Data analysis Expression data were analyzed using the ΔCq and 2  ΔCq methods. Based on the NormFinder results, a combination of HPRT1 and GUSB was used as the internal control. Hierarchical clustering and principal component analysis (PCA) were performed using the MultiExperiment Viewer (MeV) and R software (Lucent Technologies). We then aimed to identify individual differences between different lineages and their developmental stages. 2.8. Database preparation The database was prepared using the R-project (http://www.rproject.org).

To evaluate the overall differences in the expression profiles of different hematopoietic non-malignant lineages, we used hierarchical clustering. The analyses revealed strong clustering of B-ly, T-ly and My lineages (Fig. 1A). Repeating the analyses after excluding the defining markers (CD19, CD22, CD79A, CD3, CD8, CD4 and CD14) did not markedly affect the results (Fig. 1B). This result suggests that the chosen genes really appropriately describe the changes that occur during differentiation into the particular lineages, thus confirming the accuracy of our gene selection algorithm. The lineage-specific genes are listed in Table 2. In addition, we combined the same data with the results of diagnostic specimens of patients with leukemias (ALL, AML and biphenotypic pB/My AL) using hierarchical clustering and PCA. These analyses have revealed that all ALLs clustered together with their respective B or T lineage, all AMLs clustered with the My lineages (Suppl. Fig. 2). Data of all biphenotypic pB/My AL cases (who were all directed to ALL treatment as the prevailing immunophenotype (Mejstrikova et al., 2010) was B lineage) clustered together with B-Ly and pB ALL. 3.3. Developmental changes in gene expression – B lineage The expression of CD19, CD22, CD79A, IKZF1, IKZF2, IKZF3, PAX5, FOXO1, PAWR, POU2AF1, and S100A10 increased along with the progression of B-ly differentiation in the analyzed populations. The role of several of these genes in B-ly development is already well known (Kim et al., 1996; Lin et al., 2011; Zandi et al., 2008). Basically, some of the genes tend to increase during the differentiation (e.g., the IKZF family and S100A10), some show an opposite trend – decreasing expression during maturation (e. g. FLT3, HOXA10, and HOXA9); finally, some exhibited more complicated expression dynamics throughout differentiation (such as IRF4 and MYB: expression increased with maturation that takes place in the BM, from the B1 to B3 stage, and was then downregulated in peripheral cells of B4 stage). Particular interest lies on S100A10, PAWR and CCDC26. S100A10, with highest expression in monocytes, was described as part of the annexin A2 heterotetramer with a significant role in plasmin-mediated activation of peripheral monocytes (Laumonnier et al., 2006). In other lineages, it probably functions as a mediator of cell-cycle or apoptosis regulation (Hsu et al., 1997; Shan et al., 2013). Several studies have

Please cite this article as: Polgárová, K., et al., Quantitative expression of regulatory and differentiation-related genes in the key steps of human hematopoiesis: The LeukoStage Database. Differentiation (2015), http://dx.doi.org/10.1016/j.diff.2015.11.003i

4

K. Polgárová et al. / Differentiation ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 1. (A) Hierarchical clustering of non-malignant populations based on mRNA expression data from all collected genes (presented as ΔCT). Red color represents high expression, green color represents low expression. Samples/genes presented with gray were below detection sensitivity of the approach. (B) Hierarchical of non-malignant populations based on mRNA expression data (presented as ΔCT) with exclusion of defining genes (CD19, CD22, CD79A, CD3, CD8, CD4 and CD14). Red color represents high expression, green color represents low expression. Samples/genes presented with gray were below detection sensitivity of the approach. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

also described its upregulation in different malignancies, including lymphomas (Rust et al., 2005). Another gene, PAWR, generally had higher expression in the lymphoid lineage (Ly) than in the My. In B-ly, there is a clear trend of increasing expression at more differentiated stages. Until now, there have not been any reports on the role of PAWR in B-ly development. An interesting difference in PAWR expression between healthy B-ly and their malignant counterparts was observed, thus this gene is discussed in more detail below together with CCDC26 which was not detectable in healthy B-ly at all. The expected dynamics of RAG1 could not be confirmed as we had to exclude this gene (the only gene out of 95 total) from our analysis due to an inappropriate assay design, later revealed by the manufacturer. 3.4. Developmental changes in gene expression – T lineage In T-ly development, increasing trends in the expression of CD8a, LAT, ITK, ID2, IKZF3, IRF4, and S100A10 were observed. The opposite trend was observed for HOXA10, HOXA9 and NOTCH3 expression, with low expression in the mature peripheral T lymphocytes (T4 stage). Our data also confirmed T-ly- associated expression of BCL11B (Li et al., 2013). During T-ly differentiation, ID2

(inhibitor of DNA binding 2) expression gradually increased from the T1 to the most mature stage. ID2 acts as a repressor of the E2Amediated trans-activation of the IL-10 locus and plays a key regulatory role in effector T cell differentiation because it limits IL-10 production by activated T cells and thus minimizes their suppressive activity during the effector phase of the immunological response (Masson et al., 2014). The highest expression of ID2 was detected in mature granulocytes, which is in line with already published experimental data showing that ID2-deficient mice exhibit increased cellularity in the granulocyte/myeloid progenitor compartment and show significantly higher numbers of maturing neutrophils (Ko et al., 2008). An opposite trend in ID2 expression was observed in the B-ly, where expression was downregulated in the most mature stage. The suggested function of ID2 in the B-ly is blocking E2A activity via Ig locus contraction, which leads to reduced bone marrow B cell output in adults (Jensen et al., 2013). 3.5. Developmental changes in gene expression – My lineages During My differentiation (both in monocytic and granulocytic differentiation), AIF1, CD14, CSF3R, FCGR2A, FOXO1, LILRA2, MAFB, MNDA, NOTCH1, ID2, S100A10, RNF130 and KLF4 expression

Please cite this article as: Polgárová, K., et al., Quantitative expression of regulatory and differentiation-related genes in the key steps of human hematopoiesis: The LeukoStage Database. Differentiation (2015), http://dx.doi.org/10.1016/j.diff.2015.11.003i

K. Polgárová et al. / Differentiation ∎ (∎∎∎∎) ∎∎∎–∎∎∎ Table 2 Lineage-associated genes. Myeloid lineage-associated genes

AIF1 CEBPD CEBPB RNF130 SPI1 FCGR2a CEBPA CSF3R IL6R MAFB MNDA NFIL3 RBM47 CD14 Monocyte-associated Granulocyte-assogenes ciated genes KLF4 CEBPE LTF CEACAM6

Lymphoid lineage-associated genes

MSH6 MYB IKZF3 PAWR B-ly-associated genes CD19 POU2AF1 PAX5 CD22 FOXO1 BLNK EBF1

T-ly-associated genes CD8a HOXB4 CD3e ITK LCK TCF7 CD2

BCL11A

BCL11B

CD79A TCF3

GATA3 NOTCH3 LAT ITG6A UBASH3A

increased. As for the Ly subsets, several of these genes are already known to play a role in hematopoietic differentiation of My lineages. The functional importance of CSF3R in My differentiation is evident from the documented CSF3R mutations in severe congenital neutropenia (Triot et al., 2014). MNDA was shown to be involved in programmed cell death regulation in granulocyte-macrophage progenitors. Its downregulation in patients with myelodysplastic syndrome is thought to contribute to the pathogenesis of this disease (Briggs et al., 2006). The immunoglobulin Fc gamma receptor genes FCGR2A (CD32) and FCGR3A (CD16) are functionally associated with phagocytosis, and their upregulation during the maturation of granulocytic and monocytic lineages is thus logical. A decrease in MYB expression was observed during differentiation. Compared to more differentiated stages, MYB gene expression was higher in AML samples. It has been shown to block monocytic differentiation in an AML cell line (Zhao et al., 2013). In

5

our data set, MYB expression was generally higher in Ly cells than in the My lineage. The gene is known to regulate the G1-S and G2M cell cycle transitions in early hematopoietic progenitors. Several experiments have also shown that inhibition of MYB function leads to absence of lymphocytes and other hematopoietic lineages (Nakata et al., 2007; Soza-Ried et al., 2010). At the appropriate stages of Ly differentiation, MYB inhibits RAG and thus prevents unwanted changes in DNA (Timblin and Schlissel, 2013). 3.6. Aberrant expression of molecules associated with other lineages When pathological cells of a certain lineage express a molecule that is otherwise considered restricted to another lineage, this immunodiagnostic situation is called aberrant expression (Kalina et al., 2005). We discovered such lineage-aberrant mRNA expression of several molecules in our non-malignant cell-type subsets that were not contributable to contamination. One such exceptional gene was CD79A, which was expressed also in the T-ly, although with a lower expression level than observed in B-ly (Fig. 2A). The lack of simultaneous expression of other B-ly-specific markers, such as CD19 (Fig. 2B) or CD22, showed that this phenomenon could not be simply explained by possible contamination of the T-ly specimens with B-ly. Expression of this antigen was previously observed on the surface of T-ly malignancies (Lai et al., 2000), but no data describing its expression and possible function in healthy T cells have been published. However, aberrant expression of this molecule is thought to play a role in the promotion of tumorigenesis mediated by myeloid-derived suppressor cells (Luger et al., 2013). Interestingly, the FCGR3A (CD16a) transcript was detected in mature B cells, which, although capable of phagocytosis, are characterized as CD16neg at the protein level. The contamination by My cells should be considered in this case as the mature B-ly were separated from peripheral blood, where CD16high populations are frequent. In such case, however, the expression of others My genes should also be increased, which was not observed. 3.7. Comparison of malignant specimens with their physiologically normal counterparts We observed several differences in the expression of particular genes when comparing malignant specimens with healthy cell

Fig. 2. comparison of differences in expression of CD79a (A) between non-malignant populations showing moderate levels of mRNA expression of a B-ly antigen CD79a in T-ly (monocytes, Mono, and granulocytes, Gra represent negative cells). The lack of expression of other B-ly specific genes, such as CD19 (B) excludes the possible contamination being the only explanation of this phenomenon (data presented as “Negative” were below the detection sensitivity). Note that both B-ly and T-ly represent all stages of physiological development defined in Table 1. Horizontal lines represent medians.

Please cite this article as: Polgárová, K., et al., Quantitative expression of regulatory and differentiation-related genes in the key steps of human hematopoiesis: The LeukoStage Database. Differentiation (2015), http://dx.doi.org/10.1016/j.diff.2015.11.003i

6

K. Polgárová et al. / Differentiation ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 3. Comparison of particular genes expression in malignant and non-malignant populations. CCDC26 was expressed in pB-ALL but not in non-malignant B-ly (p o 0.001 using Mann-Whitney-Wilcoxon nonparametric test) (A). PAWR was found to be expressed in healthy B-ly on the contrary to their malignant counterparts, where it was not expressed at all (po 0.001) (B). Another example are Flt3 (p o 0.001) (C), and then HOXA genes, particularly HOXA9, HOXA10 and differences in their expression comparing malignant an non-malignant T-cell (with p o 0.05) and B-cell populations (p o 0.001) (D, E). Horizontal lines represent medians.

Please cite this article as: Polgárová, K., et al., Quantitative expression of regulatory and differentiation-related genes in the key steps of human hematopoiesis: The LeukoStage Database. Differentiation (2015), http://dx.doi.org/10.1016/j.diff.2015.11.003i

K. Polgárová et al. / Differentiation ∎ (∎∎∎∎) ∎∎∎–∎∎∎

subsets. Some of these genes are responsible for regulation of the proliferation of immature cells and, not surprisingly, were upregulated in early differentiation stages as well as in malignant cells. Other molecules are functionally important and are expressed in the most mature effector stages, with typically lower expression in immature non-malignant stages and malignant specimens. 3.8. CCDC26 and FLT3 as markers of malignant B precursors One of the most interesting differentially expressed genes, CCDC26, was almost always under the detection limit in healthy B-ly; however, it was detected in all cases of pB-ALL (Fig. 3A) at a level comparable that of healthy monocytes. CCDC26 expression was also detected in non-malignant T-ly (mainly T1-T3) specimens. In addition to pB-ALL, AML specimens also exhibited clearly higher expression of CCDC26 than their non-malignant counterparts (which was not true for T-ALL compared with non-malignant T-ly). The gene was previously shown to regulate retinoic acidinduced maturation in the HL60 (myeloid) cell line (Hirano et al., 2008). Its polymorphism is correlated with glioma risk (Li et al., 2013) and it has a possible function in leukocyte development (Yin et al., 2006). Furthermore, abnormalities of CCDC26 (deletion or amplification) were found in pediatric ALL patients with poor prognosis (Mullighan et al., 2007), and one study showed amplification of the genomic region around this gene in 14% of pediatric AML cases (Radtke et al., 2009). Whether CCDC26 expression is related to regulation by retinoid acid or connected to an ncRNA, as it was described for AML, remains unknown, as well as its exact role during differentiation or its impact on the survival of malignant cells. In pB-ALL, FLT3 expression is generally higher than in nonmalignant B lineage cells (Fig. 3C), especially when more mature B lineage stages are considered. FLT3 expression at the protein level was observed especially in the most immature pB-ALL stage (proBALL, Vaskova et al., unpublished observation) with or without MLL fusions, which is in line with the decline observed during normal B-ly differentiation (Fig. 3C).

7

non-malignant B precursors Expression of both HOXA9 and HOXA10 was higher in the B1 and B2 stages compared to the more differentiated stages and malignant cells (Fig. 3D and E, respectively). Interestingly, when malignant cells were compared to their non-malignant counterparts, opposite trends were observed in the myeloid and lymphoid lineages: expression of both HOXA9 and HOXA10 was higher in AML but lower or undetectable in ALL when compared with their respective non-malignant counterparts. 3.11. Differences between malignant T cells and their counterparts Several differences were also observed between T-ALL and T-lymphocytes. Remarkably, most differentially expressed genes were downregulated in malignant cells (IRF4, FOXP3, HOXA10, CSF3R, and MNDA). HOXA9, on the other hand, was undetectable in T-ALL but expressed in the corresponding healthy thymic cell subsets up to the double positive stage (T1-T3). ID2 and ITK expression levels were comparable in T-ALL and the T1 stage but higher in the other more differentiated stages. In addition to increased expression in differentiating myeloid cells, we also observed lower CSF3R expression in T-ALL cells compared with their non-malignant counterparts and other leukemia subtypes. Whether there is any connection between this observation and the fact that the biphenotypic T/My-ALs are much less frequent than pB/My-AL remains unclear. 3.12. CEBP genes correspond to the differentiation status of AML cells In My cells, both CEBPB and CEBPD expression levels were low in AML and in the most immature physiological differentiation stages but increased as differentiation proceeded. Generally, the differences between malignant specimens and their non-malignant counterparts were much less marked in the My than in the Ly lineages. 3.13. Comparison of different subtypes of leukemia

3.9. PAWR expression is lost in malignant cells On the contrary, the expression of another gene, PAWR, was remarkably higher in healthy B-ly than in pB-ALL (Fig. 3B). This gene product is a WT1-interacting protein capable of inducing apoptosis through activation of the Fas pathway along with inhibition of the NFkB pathway (El-Guendy and Rangnekar, 2003). PAWR downregulation is related to poor prognosis in breast cancer (Nagai et al., 2010), and it was identified to be a direct suppression target of SCR1/HOXC11(Walsh et al., 2014). In line with its presumed tumor suppressor functions, the levels of expression of PAWR were generally low in malignant pB-ALL specimens. Interestingly, as much as half of patients from the swALL cohort (which is analyzed separately and mentioned in the LeukoStage Database below) exceeded the highest PAWR expression value of the nonswALL pB-ALL cohort. The downregulation of PAWR was not observed in T-ALL. A possible explanation may be the simultaneous expression of Notch3. The PAWR/THAP1 protein complex and Notch3 competitively bind the CCAR1 promoter. The PAWR/THAP1 complex induces full-length CCAR1 isoform expression and cellular apoptosis, whereas Notch3 binding induces splicing of a shorter isoform of CCAR1 that leads to decreased apoptosis in T-ALL cells (Lu et al., 2013). The potentially proapoptotic effect of PAWR may therefore be suppressed by its competitor Notch3, which is known to be highly expressed in T-ALL based on both previous studies and our current data. 3.10. HOXA genes: opposite differentiation trends in malignant and

Several pB-ALL genotype-specific genes were observed; however, with low specimen number per a given genotypic subset the statistical significance was not reached. E. g., S100A10 was overexpressed in MLL þ leukemias compared with other subtypes. The expression levels even reached those observed in mature B-ly or in AML, whereas in other subtypes, they remained low. Interestingly, higher expression of ITGA6 (also known as CD49f) was detected in ETV6/RUNX1 pB-ALL than in their non-malignant counterparts, while expression in the other ALL cases was similar to that observed in differentiating B-ly. ITGA6 was previously suggested as a possible marker for minimal residual disease investigation (Coustan-Smith et al., 2011). From the data presented here (as well as from our unpublished flow cytometry data from diagnostic ALL samples), it is obvious that ITGA6 could possibly be used as marker only in some clearly positive samples. Expression of ITGA6 is associated with the T-ly under normal physiological conditions and has generally low levels of positive expression in other lineages (except for Gr2b, in which its expression is comparable to T-ly levels). Similar to ITGA6, the expression of both NFIL3 and KLF4 was higher in ETV6/RUNX1 ALL. Unlike ITGA6, both genes are normally expressed throughout My (granulocytic and monocytic) development but are virtually absent in the non-malignant T-ly. In the non-malignant B-ly, KLF4 is expressed in part of the specimens, but NFIL3 is mostly absent. NFIL3 was shown to restrict FOXOinduced gene expression and antagonize the proapoptotic effect of PTEN (Keniry et al., 2013). Ikushima et al. (1997) also

Please cite this article as: Polgárová, K., et al., Quantitative expression of regulatory and differentiation-related genes in the key steps of human hematopoiesis: The LeukoStage Database. Differentiation (2015), http://dx.doi.org/10.1016/j.diff.2015.11.003i

K. Polgárová et al. / Differentiation ∎ (∎∎∎∎) ∎∎∎–∎∎∎

8

Fig. 4. Different expression of particular genes dependent on pB ALL subtype – e.g., EPOR is increased not only in ETV6/RUNX1 leukemias but also in a minority of cases with other subtypes. Horizontal lines represent medians.

demonstrated its role in malignant transformation of pro-B cells by inhibiting growth factor deprivation-driven apoptosis. Our data show that increased NFIL3 expression in malignant cells compared with non-malignant cells is specific for the B-ly lineage. Whether this expression pattern is due to an possible role for this gene in B-ly development/malignant transformation or simply reflects the fact that myeloid growth factors and differentiation regulators may increase IL-3 production (Smith et al., 1995) thus stimulating their natural function remains unanswered. Other genes overexpressed in ETV6/RUNX1 malignancies included EPOR (Fig. 4), which is in line with already published data, as this gene is supposed to be regulated directly by the fusion transcript (Inthal et al., 2008). 3.14. Biphenotypic (pB/My) AL differ from “pure” pB ALL We also compared gene expression between pB-ALL and biphenotypic AL. Differentially expressed genes included CEBPE, LCK, CRFL2, NDN and LILRA2, which exhibited higher expression in pB-ALL specimens (p o0.05, Mann–Whitney), and GATA3, which was upregulated in biphenotypic AL specimens. Several other genes exhibited interesting expression patterns, although the differences were not statistically significant. These genes included the previously mentioned ITG6A, which exhibited even higher expression in biphenotypic AL than in AML; NFIL3, S100A10 and MAFB, which all exhibited very low or even undetectable expression in biphenotypic AL; and CEBPA, which was overexpressed in biphenotypic AL at levels similar those observed in AML. 3.15. LeukoStage Database All expression data can be accessed through our web-based LeukoStage Database (http://camelot.lf2.cuni.cz/fiserkar/LSDat/ gens2/index.php). It is possible to choose and compare different lineages and their differentiation stages and visualize expression (represented as 2  ΔCT) of particular genes in dot plots. It is also

possible to use internal control RGs other than the default ones. This feature makes the tool useful also for comparison with users’ already analyzed data without adding a new internal control gene. Leukemia cases are included for comparison with their healthy counterparts and to other ALs. We also included a specific, newly described group of switching leukemias (swALL) that are characterized by a switch from pB-ALL to the monocytic lineage (Slamova et al., 2014). Although many biological features of swALL are known, no single underlying genetic feature has been found so far. The swALLs typically present as CD2pos pB-ALL and appear to transdifferentiate into monocytoid cells, which are phenotypically akin to monocytes, shortly after the start of treatment. The swALL subset overlaps with several other pB-ALL subsets including pBALL with deletion of ERG and/or IKZF1 (Clappier et al., 2014; Zaliova et al., 2014). Genotypic tracking of the various stages of swALL – B-lymphoid, intermediate and monocytoid – has served as an example of how our quantified data can be used to address a practical issue concerning cell differentiation. In the LeukoStage Database, it is now possible to search for differences and similarities between each stage and the various malignant or non-malignant cells. Finally, examples of various acute leukemia cell lines are presented. In addition to graphical comparisons, it is possible to search for differentially expressed genes between any two groups selected from the normal sorted subsets, primary AL and/or cell lines. The existing HemaExplorer database (Bagger et al., 2013), which also provides quantitative information, combines data from several platforms on samples of mature blood cells and AML cells processed in different laboratories. The LeukoStage Database presented here provides additional useful information on the separate stages of cells undergoing differentiation and on both ALL and AML separated into cell subsets. Another advantage is that LeukoStage is based on qPCR data and thus does not have to address the bias caused by inter-platform and inter-laboratory variability. Furthermore, our study provides quantitative data with a wide dynamic range. In addition to providing searchable information on specific genes and/or cell differences, the LeukoStage Database can be used as a tool for optimizing NGS-based platforms (Fiser et al., manuscript in preparation).

4. Conclusion We present a qPCR expression array of 90 genes with regulatory functions and/or lineage-restricted expression patterns that is useful for hematopoietic differentiation evaluation. Our approach, unlike many others, effectively analyzes a low number of cells without introducing additional bias to the results. The raw data can be reanalyzed in our web database (the LeukoStage Database), which shows the selected genes in both normal physiological and malignant contexts. The LeukoStage Database allows users to compare their own results with defined maturation stages of different hematopoietic lineages. To the best of our knowledge, this study is the first report of qPCR analysis of dozens of genes in 4 different hematopoietic lineages, including B-lymphocytes, T-lymphocytes, monocytes and granulocytes, and their precursors at distinct stages of development. Our study showed several genes to be associated with differentiation, lineage and/or malignant status, which makes them good candidates for diagnostics or for the development of treatment strategies.

Conflicts of interest The authors claim no conflicts of interest.

Please cite this article as: Polgárová, K., et al., Quantitative expression of regulatory and differentiation-related genes in the key steps of human hematopoiesis: The LeukoStage Database. Differentiation (2015), http://dx.doi.org/10.1016/j.diff.2015.11.003i

K. Polgárová et al. / Differentiation ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Acknowledgments GACR P301/10/1877; Charles University UNCE 204012; the work of E.F. was supported by GACR P304/12/2214; MH CZ-DRO, UH Motol, Prague, Czech Republic 00064203 We appreciate the expertize of M. Kropacek, with whom we collaborated during the setup of the LeukoStage Database. Use of thymi was enabled due to excellent cooperation with the Children’s Heart Center, Motol.

Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.diff.2015.11.003.

References Abbas, A.R., Baldwin, D., Ma, Y., Ouyang, W., Gurney, A., Martin, F., Fong, S., van Lookeren Campagne, M., Godowski, P., Williams, P.M., Chan, A.C., Clark, H.F., 2005. Immune response in silico (IRIS): immune-specific genes identified from a compendium of microarray expression data. Genes Immun. 6, 319–331. Andersen, C.L., Jensen, J.L., Ørntoft, T.F., 2004. Normalization of real-time quantitative reverse transcription-PCR data: a model-based variance estimation approach to identify genes suited for normalization, applied to bladder and colon cancer data sets. Cancer Res. 64, 5245–5250. Bagger, F.O., Rapin, N., Theilgaard-Mönch, K., Kaczkowski, B., Thoren, L. a, Jendholm, J., Winther, O., Porse, B.T., 2013. HemaExplorer: a database of mRNA expression profiles in normal and malignant haematopoiesis. Nucleic Acids Res. 41, D1034–D1039. http://dx.doi.org/10.1093/nar/gks1021. Benjamini, Y., Speed, T.P., 2012. Summarizing and correcting the GC content bias in high-throughput sequencing. Nucleic Acids Res. 40, e72. Briggs, R.C., Shults, K.E., Flye, L. a, McClintock-Treep, S. a, Jagasia, M.H., Goodman, S. a, Boulos, F.I., Jacobberger, J.W., Stelzer, G.T., Head, D.R., 2006. Dysregulated human myeloid nuclear differentiation antigen expression in myelodysplastic syndromes: evidence for a role in apoptosis. Cancer Res. 66, 4645–4651. http: //dx.doi.org/10.1158/0008-5472.CAN-06-0229. Bustin, S.A., Benes, V., Garson, J.A., Hellemans, J., Huggett, J., Kubista, M., Mueller, R., Nolan, T., Pfaffl, M.W., Shipley, G.L., Vandesompele, J., Wittwer, C.T., 2009. The MIQE guidelines: minimum information for publication of quantitative realtime PCR experiments. Clin. Chem. 55, 611–622. Ceredig, R., Rolink, A.G., Brown, G., 2009. Models of haematopoiesis: seeing the wood for the trees. Nat. Rev. Immunol. 9, 293–300. Clappier, E., Auclerc, M.F., Rapion, J., Bakkus, M., Caye, a, Khemiri, a, Giroux, C., Hernandez, L., Kabongo, E., Savola, S., Leblanc, T., Yakouben, K., Plat, G., Costa, V., Ferster, a, Girard, S., Fenneteau, O., Cayuela, J.M., Sigaux, F., Dastugue, N., Suciu, S., Benoit, Y., Bertrand, Y., Soulier, J., Cavé, H., 2014. An intragenic ERG deletion is a marker of an oncogenic subtype of B-cell precursor acute lymphoblastic leukemia with a favorable outcome despite frequent IKZF1 deletions. Leukemia 28, 70–77. http://dx.doi.org/10.1038/leu.2013.277. Coustan-Smith, E., Song, G., Clark, C., Key, L., Liu, P., Mehrpooya, M., Stow, P., Su, X., Shurtleff, S., Pui, C.-H., Downing, J.R., Campana, D., 2011. New markers for minimal residual disease detection in acute lymphoblastic leukemia. Blood 117, 6267–6276. Draghici, S., Khatri, P., Eklund, A.C., Szallasi, Z., 2006. Reliability and reproducibility issues in DNA microarray measurements. Trends Genet. El-Guendy, N., Rangnekar, V.M., 2003. Apoptosis by Par-4 in cancer and neurodegenerative diseases. Exp. Cell Res. 283, 51–66. Graf, T., 2002. Differentiation plasticity of hematopoietic cells. Blood 99, 3089–3101. Hirano, T., Ike, F., Murata, T., Obata, Y., Utiyama, H., Yokoyama, K.K., 2008. Genes encoded within 8q24 on the amplicon of a large extrachromosomal element are selectively repressed during the terminal differentiation of HL-60 cells. Mutat. Res. Fundam. Mol. Mech. Mutagen. 640, 97–106. Hsu, S.Y., Kaipia, A., Zhu, L., Hsueh, A.J., 1997. Interference of BAD (Bcl-xL/Bcl-2associated death promoter)-induced apoptosis in mammalian cells by 14-3-3 isoforms and P11. Mol. Endocrinol. 11, 1858–1867. Chou, C.-C., Chen, C.-H., Lee, T.-T., Peck, K., 2004. Optimization of probe length and the number of probes per gene for optimal microarray analysis of gene expression. Nucleic Acids Res. 32, e99. Ikushima, S., Inukai, T., Inaba, T., Nimer, S.D., Cleveland, J.L., Look, A.T., 1997. Pivotal role for the NFIL3/E4BP4 transcription factor in interleukin 3-mediated survival of pro-B lymphocytes. Proc. Natl. Acad. Sci. USA 94, 2609–2614. Inthal, A., Krapf, G., Beck, D., Joas, R., Kauer, M.O., Orel, L., Fuka, G., Mann, G., PanzerGrümayer, E.R., 2008. Role of the erythropoietin receptor in ETV6/RUNX1-positive acute lymphoblastic leukemia. Clin. Cancer Res. 14, 7196–7204. Jensen, K., Rother, M.B., Brusletto, B.S., Olstad, O.K., Dalsbotten Aass, H.C., van Zelm, M.C., Kierulf, P., Gautvik, K.M., 2013. Increased ID2 levels in adult precursor B cells as compared with children is associated with impaired Ig locus contraction and decreased bone marrow output. J. Immunol. 191, 1210–1219.

9

Kalina, T., Vaskova, M., Mejstrikova, E., Madzo, J., Trka, J., Stary, J., Hrusak, O., 2005. Myeloid antigens in childhood lymphoblastic leukemia: clinical data point to regulation of CD66c distinct from other myeloid antigens. BMC Cancer 5, 38. http://dx.doi.org/10.1186/1471-2407-5-38. Keniry, M., Pires, M.M., Mense, S., Lefebvre, C., Gan, B., Justiano, K., Lau, Y.-K.I., Hopkins, B., Hodakoski, C., Koujak, S., Toole, J., Fenton, F., Calahan, A., Califano, A., DePinho, R. a, Maurer, M., Parsons, R., 2013. Survival factor NFIL3 restricts FOXO-induced gene expression in cancer. Genes Dev. 27, 916–927. http://dx. doi.org/10.1101/gad.214049.113. Kim, U., Qin, X.F., Gong, S., Stevens, S., Luo, Y., Nussenzweig, M., Roeder, R.G., 1996. The B-cell-specific transcription coactivator OCA-B/OBF-1/Bob-1 is essential for normal production of immunoglobulin isotypes. Nature 383, 542–547. Ko, J., Patel, N., Ikawa, T., Kawamoto, H., Frank, O., Rivera, R.R., Van Etten, R.A., Murre, C., 2008. Suppression of E-protein activity interferes with the development of BCR-ABL-mediated myeloproliferative disease. Proc. Natl. Acad. Sci. USA 105, 12967–12972. Lai, R., Juco, J., Lee, S.F., Nahirniak, S., Etches, W.S., 2000. Flow cytometric detection of CD79a expression in T-cell acute lymphoblastic leukemias. Am. J. Clin. Pathol. 113, 823–830. Larsson, J., Karlsson, S., 2005. The role of Smad signaling in hematopoiesis. Oncogene 24, 5676–5692. Laumonnier, Y., Syrovets, T., Burysek, L., Simmet, T., 2006. Identification of the annexin A2 heterotetramer as a receptor for the plasmin-induced signaling in human peripheral monocytes. Blood 107, 3342–3349. http://dx.doi.org/10.1182/ blood-2005-07-2840. Lehtonen, A., Ahlfors, H., Veckman, V., Miettinen, M., Lahesmaa, R., Julkunen, I., 2007. Gene expression profiling during differentiation of human monocytes to macrophages or dendritic cells. J. Leukoc. Biol. 82, 710–720. Li, L., Zhang, J.A., Dose, M., Kueh, H.Y., Mosadeghi, R., Gounari, F., Rothenberg, E.V., 2013. A far downstream enhancer for murine Bcl11b controls its T-cell specific expression. Blood 122, 902–911. http://dx.doi.org/10.1182/ blood-2012-08-447839. Li, M., Zhou, Q., Tu, C., Jiang, Y., 2013. A meta-analysis of an association between the XRCC1 polymorphisms and gliomas risk. J. Neurooncol. 111, 221–228. Lin, Y.C., Jhunjhunwala, S., Benner, C., Heinz, S., Welinder, E., Mansson, R., Sigvardsson, C.A., Hagman, J., Espinoza, C.A., 2011. NIH Public Access 11, 635–643. http://dx.doi.org/10.1038/ni.1891.A. Liu, X.-L., Yuan, J.-Y., Zhang, J.-W., Zhang, X.-H., Wang, R.-X., 2007. Differential gene expression in human hematopoietic stem cells specified toward erythroid, megakaryocytic, and granulocytic lineage. J. Leukoc. Biol. 82, 986–1002. Lu, C., Li, J.-Y., Ge, Z., Zhang, L., Zhou, G.-P., 2013. Par-4/THAP1 complex and Notch3 competitively regulated pre-mRNA splicing of CCAR1 and affected inversely the survival of T-cell acute lymphoblastic leukemia cells. Oncogene 32, 5602–5613. http://dx.doi.org/10.1038/onc.2013.349. Luger, D., Yang, Y.-A., Raviv, A., Weinberg, D., Banerjee, S., Lee, M.-J., Trepel, J., Yang, L., Wakefield, L.M., 2013. Expression of the B-cell receptor component CD79a on immature myeloid cells contributes to their tumor promoting effects. PLoS One 8, e76115. Masson, F., Ghisi, M., Groom, J.R., Kallies, A., Seillet, C., Johnstone, R.W., Nutt, S.L., Belz, G.T., 2014. Id2 represses E2A-mediated activation of IL-10 expression in T cells. Blood 123, 3420–3428. http://dx.doi.org/10.1182/blood-2014-03-561456. Mejstrikova, E., Volejnikova, J., Fronkova, E., Zdrahalova, K., Kalina, T., Sterba, J., Jabali, Y., Mihal, V., Blazek, B., Cerna, Z., Prochazkova, D., Hak, J., Zemanova, Z., Jarosova, M., Oltova, A., Sedlacek, P., Schwarz, J., Zuna, J., Trka, J., Stary, J., Hrusak, O., 2010. Prognosis of children with mixed phenotype acute leukemia treated on the basis of consistent immunophenotypic criteria. Haematologica 95, 928–935. http://dx.doi.org/10.3324/haematol.2009.014506. Mullighan, C.G., Su, X., Ma, J., Yang, W., Relling, M.V., Carroll, W.L., Reaman, G., Bowman, P., Devidas, M., Hunger, S.P., Downing, J.R., 2007. Genome-wide profiling of high-risk pediatric Acute Lymphoblastic Leukemia (ALL): the ALL pilot project for the Therapeutically Applicable Research To Generate Effective Treatments (TARGET) initiative. ASH Annual Meeting Abstracts. vol. 110, p. 229. Nagai, M.A., Gerhard, R., Salaorni, S., Fregnani, J.H.T.G., Nonogaki, S., Netto, M.M., Soares, F.A., 2010. Down-regulation of the candidate tumor suppressor gene PAR-4 is associated with poor prognosis in breast cancer. Int. J. Oncol. 37, 41–49. Nakata, Y., Shetzline, S., Sakashita, C., Kalota, A., Rallapalli, R., Rudnick, S.I., Zhang, Y., Emerson, S.G., Gewirtz, A.M., 2007. c-Myb contributes to G2/M cell cycle transition in human hematopoietic cells by direct regulation of cyclin B1 expression. Mol. Cell Biol. 27, 2048–2058. Novershtern, N., Subramanian, A., Lawton, L.N., Mak, R.H., Haining, W.N., McConkey, M.E., Habib, N., Yosef, N., Chang, C.Y., Shay, T., Frampton, G.M., Drake, A.C.B., Leskov, I., Nilsson, B., Preffer, F., Dombkowski, D., Evans, J.W., Liefeld, T., Smutko, J.S., Chen, J., Friedman, N., Young, R. a, Golub, T.R., Regev, A., Ebert, B.L., 2011. Densely interconnected transcriptional circuits control cell states in human hematopoiesis. Cell 144, 296–309. http://dx.doi.org/10.1016/j.cell.2011.01.004. Raabe, C.A., Tang, T.-H., Brosius, J., Rozhdestvensky, T.S., 2014. Biases in small RNA deep sequencing data. Nucleic Acids Res. 42, 1414–1426. Radtke, I., Mullighan, C.G., Ishii, M., Su, X., Cheng, J., Ma, J., Ganti, R., Cai, Z., Goorha, S., Pounds, S.B., Cao, X., Obert, C., Armstrong, J., Zhang, J., Song, G., Ribeiro, R.C., Rubnitz, J.E., Raimondi, S.C., Shurtleff, S.A., Downing, J.R., 2009. Genomic analysis reveals few genetic alterations in pediatric acute myeloid leukemia. Proc. Natl. Acad. Sci. USA 106, 12944–12949. Robinette, M.L., Fuchs, A., Cortez, V.S., Lee, J.S., Wang, Y., Durum, S.K., Gilfillan, S., Colonna, M., 2015. Transcriptional programs define molecular characteristics of innate lymphoid cell classes and subsets. Nat. Immunol. 1, 45–46. http://dx.doi.

Please cite this article as: Polgárová, K., et al., Quantitative expression of regulatory and differentiation-related genes in the key steps of human hematopoiesis: The LeukoStage Database. Differentiation (2015), http://dx.doi.org/10.1016/j.diff.2015.11.003i

10

K. Polgárová et al. / Differentiation ∎ (∎∎∎∎) ∎∎∎–∎∎∎

org/10.1038/ni.3094. Ruijter, J.M., Ramakers, C., Hoogaars, W.M.H., Karlen, Y., Bakker, O., van den hoff, M. J.B., Moorman, A.F.M., 2009. Amplification efficiency: Linking baseline and bias in the analysis of quantitative PCR data. Nucleic Acids Res. 37, e45. Rust, R., Visser, L., van der Leij, J., Harms, G., Blokzijl, T., Deloulme, J.C., van der Vlies, P., Kamps, W., Kok, K., Lim, M., Poppema, S., van den Berg, A., 2005. High expression of calcium-binding proteins, S100A10, S100A11 and CALM2 in anaplastic large cell lymphoma. Br. J. Haematol. 131, 596–608. Shan, X., Miao, Y., Fan, R., Qian, H., Chen, P., Liu, H., Yan, X., Li, J., Zhou, F., 2013. MiR590-5P Inhibits Growth of HepG2 Cells via Decrease of S100A10 Expression and Inhibition of the Wnt Pathway. Int. J. Mol. Sci. 14, 8556–8569. Slamova, L., Starkova, J., Fronkova, E., Zaliova, M., Reznickova, L., van Delft, F.W., Vodickova, E., Volejnikova, J., Zemanova, Z., Polgarova, K., Cario, G., Figueroa, M., Kalina, T., Fiser, K., Bourquin, J.P., Bornhauser, B., Dworzak, M., Zuna, J., Trka, J., Stary, J., Hrusak, O., Mejstrikova, E., 2014. CD2-positive B-cell precursor acute lymphoblastic leukemia with an early switch to the monocytic lineage. Leukemia 28, 609–620. http://dx.doi.org/10.1038/leu.2013.354. Smith, W.B., Guida, L., Sun, Q., Korpelainen, E.I., van den Heuvel, C., Gillis, D., Hawrylowicz, C.M., Vadas, M.A., Lopez, A.F., 1995. Neutrophils activated by granulocyte-macrophage colony-stimulating factor express receptors for interleukin-3 which mediate class II expression. Blood 86, 3938–3944. Soza-Ried, C., Hess, I., Netuschil, N., Schorpp, M., Boehm, T., 2010. Essential role of c-myb in definitive hematopoiesis is evolutionarily conserved. Proc. Natl. Acad. Sci. USA 107, 17304–17308. Steger, D., Berry, D., Haider, S., Horn, M., Wagner, M., Stocker, R., Loy, A., 2011. Systematic spatial bias in DNA microarray hybridization is caused by probe spot position-dependent variability in lateral diffusion. PLoS One, 6. Timblin, G. a, Schlissel, M.S., 2013. Ebf1 and c-Myb repress rag transcription downstream of Stat5 during early B cell development. J. Immunol. 191, 4676–4687. http://dx.doi.org/10.4049/jimmunol.1301675. Triot, A., Järvinen, P.M., Arostegui, J.I., Murugan, D., Kohistani, N., Dapena Díaz, J.L., Racek, T., Puchałka, J., Gertz, E.M., Schäffer, A.A., Kotlarz, D., Pfeifer, D., Díaz de

Heredia Rubio, C., Ozdemir, M.A., Patiroglu, T., Karakukcu, M., Sánchez de Toledo Codina, J., Yagüe, J., Touw, I.P., Unal, E., Klein, C., 2014. Inherited biallelic CSF3R mutations in severe congenital neutropenia. Blood 123, 3811–3817. Tuomela, S., Salo, V., Tripathi, S.K., Chen, Z., Laurila, K., Gupta, B., Äijö, T., Oikari, L., Stockinger, B., Lähdesmäki, H., Lahesmaa, R., 2012. Identification of early gene expression changes during human Th17 cell differentiation. Blood 119, e151–e160. van Heesch, S., Mokry, M., Boskova, V., Junker, W., Mehon, R., Toonen, P., de Bruijn, E., Shull, J.D., Aitman, T.J., Cuppen, E., Guryev, V., 2013. Systematic biases in DNA copy number originate from isolation procedures. Genome Biol. 14, R33. van Zelm, M.C., 2012. Human CD27 þIgM þIgD þ B cells: T-cell or TLR-dependent? Blood 120, 4905–4906. http://dx.doi.org/10.1182/blood-2012-10-459081. Walsh, C.A., Bolger, J.C., Byrne, C., Cocchiglia, S., Hao, Y., Fagan, A., Qin, L., Cahalin, A., McCartan, D., McIlroy, M., O’Gaora, P., Xu, J., Hill, A.D., Young, L.S., 2014. Global gene repression by the steroid receptor coactivator SRC-1 promotes oncogenesis. Cancer Res. 74, 2533–2544. http://dx.doi.org/10.1158/0008-5472. CAN-13-2133. Yin, W., Rossin, A., Clifford, J.L., Gronemeyer, H., 2006. Co-resistance to retinoic acid and TRAIL by insertion mutagenesis into RAM. Oncogene 25, 3735–3744. Zaliova, M., Zimmermannova, O., Dörge, P., Eckert, C., Möricke, A., Zimmermann, M., Stuchly, J., Teigler-Schlegel, A., Meissner, B., Koehler, R., Bartram, C.R., Karawajew, L., Rhein, P., Zuna, J., Schrappe, M., Cario, G., Stanulla, M., 2014. ERG deletion is associated with CD2 and attenuates the negative impact of IKZF1 deletion in childhood acute lymphoblastic leukemia. Leukemia 28, 182–185. http://dx.doi.org/10.1038/leu.2013.282. Zandi, S., Mansson, R., Tsapogas, P., Zetterblad, J., Bryder, D., Sigvardsson, M., 2008. EBF1 is essential for B-lineage priming and establishment of a transcription factor network in common lymphoid progenitors. J. Immunol. 181, 3364–3372. Zhao, L., Ye, P., Gonda, T.J., 2013. The MYB proto-oncogene suppresses monocytic differentiation of acute myeloid leukemia cells via transcriptional activation of its target gene GFI1. Oncogene. 33, 4442–4449.

Please cite this article as: Polgárová, K., et al., Quantitative expression of regulatory and differentiation-related genes in the key steps of human hematopoiesis: The LeukoStage Database. Differentiation (2015), http://dx.doi.org/10.1016/j.diff.2015.11.003i