Transcriptional dysregulation of the multifunctional zinc finger factor 423 in acute lymphoblastic leukemia of childhood.

Differentiation arrest is a hallmark of acute lymphoblastic leukemia (ALL). Among a variety of structural and chromosomal alterations, especially mutations in genes encoding for regulators of B cell differentiation are common. The objective of this study was a comprehensive assessment of transcriptional dysregulation and high-resolution genomic profiling of B cell differentiation factors. Here we provide extended materials and methods regarding transcriptome and genome-wide copy number variation analyses published by Harder et al. [1]. Our data provide a resource for the identification of yet undefined factors that play a putative functional role in leukemogenesis such as ZNF423, whose aberrant expression interferes with B-cell differentiation.

Direct link to deposited data

Deposited data can be found here: http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE42221 and http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE42056.

Experimental design, materials and methods

Patient samples

All primary human samples were obtained upon approval by Institutional Ethics Boards. Patients were recruited by the COALL multicenter clinical trial group (Germany) and enrolled in trials COALL 97 and 03. For patient characteristics and clinical data refer to Table 1. Mononucleated cells (MNC) from bone marrow or peripheral blood were processed by gradient centrifugation (Biocoll separation solution, Biochrom) and cryopreserved.

Table 1

Patient characteristics. M, male; F, female; C-ALL, common ALL; NA, not available.

ID	Sex	Immuno-phenotype	BCR–ABL	MLL–AF4	ETV6–RUNX1	Hyper-diploidy	ZNF423 mRNA expression
6646	M	C-ALL	0	NA	0	1	36.23
6787	M	preB-ALL	NA	NA	NA	0	17.00
6822	M	C-ALL	0	0	1	0	24.97
6845	M	C-ALL	0	NA	0	NA	51.38
6869	M	C-ALL	0	0	0	1	7.91
6923	M	preB-ALL	NA	NA	NA	0	22.20
6924	M	C-ALL	0	0	0	0	9.79
6965	F	C-ALL	0	0	1	0	19.08
6992	F	preB-ALL	0	0	1	NA	18.81
7021	M	C-ALL	0	0	1	0	39.56
7065	M	C-ALL	0	0	0	0	8.71
7077	M	C-ALL	0	0	1	NA	27.48
7115	M	C-ALL	0	0	1	0	47.38
7118	M	C-ALL	0	0	0	0	20.88
7137	F	proB-ALL	0	0	0	0	13.54
7191	F	C-ALL	0	0	1	0	69.07
7293	F	C-ALL	0	0	1	0	6.52
7360	M	C-ALL	0	0	1	0	17.42
7503	M	C-ALL	0	0	1	0	27.27
7523	F	C-ALL	0	0	1	0	26.80

Gene expression array

For a comparative matched pair expression analysis, single primary leukemic cells and normal lymphoblasts were isolated from the same individual, using a MoFlow Cytomation instrument by CD34, CD10 and CD19 staining. Please note that primary ALL I4 had a blast fraction of greater than 95% and was included without single cell sort due to experimental circumstances. Total RNA was isolated from a minimum of 4 × 103 mononucleated cells. Quality and concentration of isolated RNA were determined using an Agilent RNA 6000 Nano kit on an Agilent Bioanalyzer. Total RNA underwent linear amplification in a two-step procedure and was labeled and hybridized to each array according to the manufacturer's instructions using the small sample protocol. Human GeneChip U133A arrays (Affymetrix) were then washed using the Affymetrix Fluidics Station 400 and scanned using a HP GeneArray scanner. The array image was acquired using the Affymetrix GeneChip Microarray Suite 5.0. Software. The samples were normalized using the MAS5 method and scaled to a target value of 100.

Quality control

We used Bioconductor packages (simpleaffy [2] and arrayQualityMetrics [3]) to assess the quality of the array experiment. As shown in Fig. 1, samples are affected by slight degradation likely due to the elaborate cell selection and material isolation procedures followed by two amplification steps. In addition, a slight batch effect can be observed. Nevertheless, good sample comparability is given, because the degree of degradation is almost identical across all samples and the batch effect, which manifests as an overall higher signal intensity in one group, is readily eliminated during the normalization process.

Fig. 1

Quality metrics for gene expression dataset GSE42221. (A) The rate of present calls (%), scaling factors (blue bars) as well as 3′ to 5′ ratios for ß-ACTIN and GAPDH were calculated by the simpleaffy package. (B) The dissimilarity matrix of the arrays ahead of normalization shows a batch effect that is linked to hybridization date. (C) This effect is eliminated by the background and normalization procedure. The scores displayed in the dissimilarity matrices (blue: high similarity, yellow: low similarity) reflect the distance between each pair of arrays. They are computed as the mean absolute difference between the array data.

Basic microarray analysis

The expression data of primary ALL at diagnosis (I1s, I2s, I3s, and I4; s, sorted cellular material) were independently filtered based on defined cutoff criteria such as present call, signal intensity (SI) ≥ 20, increase (I) or decrease call (D), change P value ≤ 0.003 for I/≥ 0.997 for D and signal log ratio (SLR) ≥ 0.5849 for I/≤− 0.415 for D (equivalent to 1.5 × up- or down-regulation), which arose from the comparison with the corresponding remission material (E1s, E2s, E3s, and E4s) as control. In all comparisons initial scaling was set to a target signal of 100. All genes meeting these cutoff criteria were considered to be differentially expressed, as depicted in the heatmap.

Quantitative real-time PCR

Ahead of high-resolution genomic profiling ZNF423 expression was evaluated by quantitative real time PCR (qPCR) in 200 primary B-precursor ALL samples. For this purpose RNA isolation was performed after TRIZOL lysis (Invitrogen) according to manufacturer's instructions. TRIZOL-isolated RNA was reverse transcribed to cDNA by M-MLV reverse transcriptase for 1 h at 37 °C using random primers (Promega). qPCR was carried out with SYBR Green I (Roche) according to manufacturer's instructions using following primers: ZNF423 (NM_015069.2) fw: 5′-gca gac ctg acg gac cac-3′ and rev-5′-agg cca ccc agg aga gtt-3′; and beta-2-microglobulin (B2M, NM_004048) fw: 5′-ttc tgg cct gga ggc tatc-3′ and rev: 5′-tca gga aat ttg act ttc cat tc-3′. Relative mRNA levels were depicted after normalization to B2M (2− ΔCt*1000) as a reference gene. For genomic profiling ALL samples from 20 patients that showed a significantly higher ZNF423 expression than hematopoietic/lymphopoietic progenitors at various stages of development, were selected.

High resolution genomic profiling

Genomic DNA was isolated from 20 initial ALL samples and intraindividually matched mononuclear cells (MNC) from remission bone marrow using the QIAGEN DNA Blood Mini Kit following the manufacturer's instructions. Sample preparation, hybridization and staining were performed according to the manufacturer's standard protocol for Affymetrix Genome-Wide Human SNP 6.0 microarrays. Arrays were scanned on an Affymetrix 3000 7G GeneChip scanner using the Affymetrix Command Console Software. Processing of the raw signals and raw copy-number calculation were performed using the CRMA (version 2) procedure (Aroma, Affymetrix Package Version 2.0.0) for R statistical platform (version 2.12.1) according to the Aroma Affymetrix vignette for paired total copy-number analysis. CBS algorithm (DNA copy package version 1.24.0) was then applied to perform segmentation. Quality was accessed using the Contrast QC metric recommended by Affymetrix and displayed in Table 2.

Table 2

Affymetrix SNP array quality metrics.

GEO accession	Sample name	Contrast QC	QC call rate
GSM1031509	ALL 6869 initial	1.00	94.04
GSM1031510	ALL 6845 initial	1.43	94.01
GSM1031511	ALL 6646 initial	1.77	95.33
GSM1031512	ALL 7021 initial	1.24	91.99
GSM1031513	ALL 6923 initial	2.14	95.43
GSM1031514	ALL 7077 initial	2.29	95.2
GSM1031515	ALL 7191 initial	2.04	96.13
GSM1031516	ALL 7115 initial	1.21	91.53
GSM1031517	ALL 7360 initial	2.02	97.15
GSM1031518	ALL 7065 initial	1.93	94.97
GSM1031519	ALL 6965 initial	1.67	95.96
GSM1031520	ALL 7503 initial	2.14	95.47
GSM1031521	ALL 7523 initial	2.18	96.03
GSM1031522	ALL 6992 initial	2.02	96.03
GSM1031523	ALL 6787 initial	2.32	97.25
GSM1031524	ALL 6924 initial	1.68	93.28
GSM1031525	ALL 7137 initial	2.49	97.88
GSM1031526	ALL 6822 initial	2.79	95.57
GSM1031527	ALL 7118 initial	4.26	95.6
GSM1031528	ALL 7293 initial	2.07	97.29
GSM1031529	ALL 6869 remission	1.75	93.81
GSM1031530	ALL 6845 remission	1.21	94.57
GSM1031531	ALL 6646 remission	1.66	94.71
GSM1031532	ALL 7021 remission	1.87	94.11
GSM1031533	ALL 6923 remission	1.95	96.23
GSM1031534	ALL 7077 remission	2.29	95.43
GSM1031535	ALL 7191 remission	1.72	94.08
GSM1031536	ALL 7115 remission	1.98	94.34
GSM1031537	ALL 7360 remission	0	93.58
GSM1031538	ALL 7065 remission	2.62	97.58
GSM1031539	ALL 6965 remission	1.57	94.57
GSM1031540	ALL 7503 remission	1.7	94.61
GSM1031541	ALL 7523 remission	1.97	95.6
GSM1031542	ALL 6992 remission	1.99	95.93
GSM1031543	ALL 6787 remission	2.12	97.32
GSM1031544	ALL 6924 remission	2.17	96.29
GSM1031545	ALL 7137 remission	2.18	96.62
GSM1031546	ALL 6822 remission	2.41	95.9
GSM1031547	ALL 7118 remission	1.91	92.55
GSM1031548	ALL 7293 remission	1.42	94.28

Discussion

In this report we provide an extended description of materials and methods for two datasets deposited in the GEO database containing gene expression and high-resolution genomic profiling data. The corresponding biologic material was isolated from fluorescence activated sorted primary leukemic cells (except one case) and normal lymphoblasts (gene expression data) as well as from 20 initial ALL and intraindividually matched MNC from remission bone marrow (genomic data). Based on these data we identified a potentially causative role for ZNF423 in ALL by its interference with B-cell differentiation and modulation of Smad1–Smad4 dependent transcription [1]. We encourage the use of our deposited datasets for further investigation into mechanistic underpinnings of ALL. We will provide further information if needed.

Specifications
Organism/cell line/tissue	Homo sapiens
Strain(s)	Primary human ALL specimens
Sequencer or array type	human GeneChip U133A Affymetrix, human Genome-Wide SNP 6.0 Array Affymetrix
Data format	Raw data: TAR, normalized data: SOFT, MINiML, TXT
Experimental factors	Expression profiling and genome-wide copy number variation profiling in B-precursor ALL of childhood
Consent	All primary ALL samples were obtained with written informed consent of patients' parents or their legal guardians