Introduction

Genetic alterations in leukemic cells play major roles in leukemogenesis and drug sensitivity. Thus, better characterization of leukemia genetics will be critical to the improvement of therapeutic strategies. Recent progress in technologies used for biological investigations has enabled characterization of patient primary samples to generate sufficient genetic data to facilitate in vitro and in vivo experiments [1,2,3]. Thus, knowledge of the genetic alterations present in leukemia cell lines is essential for these cells to be used in place of primary leukemia samples for in vitro and in vivo studies.

The aim of this study was to characterize genetic alterations in a large panel of B cell precursor acute lymphoblastic leukemia (B-ALL) cell lines. For this purpose, we investigated copy number abnormalities (CNAs) of genes involved in B-ALL leukemogenesis in a large number of B-ALL cell lines.

It is also important to identify cell lines harboring specific clinically relevant genetic alterations, including Philadelphia chromosome-like acute lymphoblastic leukemia (Ph-like ALL), comprising patients who exhibit a gene expression profile similar to that of Ph + B-ALL [4, 5]. In this study, we also identified putative Ph-like ALL cell lines, performed detailed genetic analysis of them.

Materials and methods

Samples and multiplex ligation-dependent probe amplification

Genomic DNA samples from 83 B-ALL cell lines stocked in the department of pediatrics Yamanashi University were analyzed using the SALSA multiplex ligation-dependent probe amplification (MLPA) kit, P335-B2 (MRC Holland, Amsterdam, the Netherlands) to detect the copy number abnormality of eight genes, namely IKZF1, PAX5, CDKN2A, CDKN2B, ETV6, RB1, BTG1, and EBF1, according to the manufacturer’s instructions, as previously described [6,7,8,9,10,11,12,13,14,15,16,17]. These cell lines included 15 Philadelphia chromosome positive (Ph+) ALL, 11 KMT2A-rearranged (KMT2A+) ALL, 17 t(1;19)(q23;p13)/TCF3-PBX1-ALL, 4 t(17;19)(p22;q13)/TCF3-HLF-ALL, four t(12;21)(p13.1;q22)/ETV6-RUNX1-ALL, and 33 other B-ALL (without the above five translocations, classified as “B-other”) cell lines. No sister cell line was included in this analysis.

Real-time quantitative-PCR analysis

Total RNA was extracted from leukemic cells using an RNeasy Mini Kit (Qiagen, Venlo, the Netherlands) according to the manufacturer’s instructions. The cDNA for reverse transcription (RT)-PCR was synthesized using the SuperScript First-Strand Synthesis System (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s instructions. Real-time quantitative-PCR (q-PCR) was performed on human ALL cell lines to determine the expression levels of CRLF2 using a 7300 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) with SYBR Green 1 (Takara Bio, Tokyo, Japan). Relative expression levels of target mRNA molecules were determined using the comparative threshold (ΔCT) method. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control.

Cell culture and flow cytometric analysis

Three ALL cell lines, KOPN49, KOPN75, and YCUB-5, were cultured in suspension in RPMI-1640 medium supplemented with 10% fetal bovine serum, penicillin (100 U/ml), and streptomycin (10 mg/ml) at 37 °C in a 5% CO2 humidified atmosphere.

Cells (1.0 × 106) were harvested, washed twice with PBS, incubated for 20 min with PE anti-human TSLPR (TSLP-R) antibody (BioLegend, San Diego, CA, USA), and analyzed using a FACS Calibur cytometer (BD Biosciences) with FlowJo software (Treestar, SanCarlos, CA, USA). Immunophenotypic analysis was also performed for cell lines with IKZF1 deleted.

Reverse transcription-PCR and mRNA sequencing

To determine unknown genetic alterations and identify Ph-like ALL-specific fusion genes, reverse transcription-PCR (RT-PCR) was performed in three B-other cell lines harboring IKZF1 deletions, according to previously reported protocols [18]. The primer pairs are listed in supplementary Table 1 (sTable 1). Messenger RNA sequencing (mRNA-seq) was performed using KOPN49 and KOPN75 cell samples. An NEBNext® mRNA Library Prep Reagent Set for Illumina was used for construction of sequencing libraries, according to the manufacturer’s instructions, and then sequenced using a HiSeq2500 (Illumina Inc., San Diego, CA, USA). Fusion transcripts were detected using Genomon-fusion (http://genomon.hgc.jp/rna/).

Targeted capture sequencing

Targeted sequencing of 163 genes known to be mutated in hematological diseases, customized to detect cancer-related genes, as previously described [19], was performed on a MiSeq system (Illumina, San Diego, CA) using HaloPlex custom panels (Agilent, Santa Clara, CA). Read alignment and variant calling were performed using the MiSeq reporter (Illumina), and variants were annotated using SureCall software (Agilent). Synonymous or noncoding variants were excluded, and single nucleotide polymorphisms (SNPs) reported to dbSNP, unless they were found in the COSMIC database. Among the remaining variants, with allele frequencies > 0.1, recurrently mutated genes were considered candidate genes and validated by conventional Sanger sequencing. In addition, IgH rearrangements were captured using 662 baits tiling the entire IgH enhancer locus and sequenced using the HiSeq2500 platform. Detection of structural variations was performed using Genomon-SV (https://github.com/Genomon-Project/GenomonSV). Briefly, Genomon-SV uses information from chimeric reads (containing breakpoints) and discordant read pairs. For each candidate structural variation, reads are realigned to the assembled contig sequence containing the structural variation breakpoint (variant sequence).

Statistical analyses

Statistical analysis was performed using the Student’s t test. P values < 0.05 were considered statistically significant. Other comparisons were performed using the 2 × 2 Chi square test or Fisher’s exact test, as appropriate.

Results

Copy number abnormalities in 83 B-ALL cell lines

To identify CNAs of eight genes related to B-ALL leukemogenesis, including IKZF1, PAX5, CDKN2A, CDKN2B, ETV6, RB1, BTG1, and EBF1, in 83 B-ALL cell lines we performed MLPA analysis. The CNA results for the 83 cell lines are summarized in Fig. 1 and sTable 2, according to genetic subgroups. Overall, 68 of the 83 cell lines had at least one CNA. No CNAs of the eight genes were determined in 4 of 11 (36%) KMT2A(+) ALL and 8 of 17 (47.1%) t(1;19) positive ALL cell lines. In contrast, all of the Ph + B-ALL and 29 of 33 (87.9%) B-other cell lines harbored at least one CNA of the eight genes included in this analysis. Deletion of CDKN2A and/or CDKN2B was the most frequent CNA (60 of 83, 72.3%), while EBF1 and RB1 deletions were rarely detected (2 of 83, 2.4%, and 5 of 83, 6.0%, respectively). Interestingly, MLPA analysis also revealed that three of the B-other ALL cell lines, YCUB-5, KOPN49, and KOPN75, had IKZF1 deletions (YCUB-5, deletion of exons 2–8; KOPN49, deletion of exons 2–3; and KOPN75, deletion of exons 1–8), indicating that they may be derived from Ph-like ALL patient samples.

Fig. 1
figure 1

Copy number abnormalities in 83 B-cell precursor acute lymphoblastic leukemia cell lines, according to genetic subtypes. The bar graph represents frequency of the deletion of eight genes, namely IKZF1, PAX5, RB1, CDKN2A, CDKN2B, ETV6, EBF1 and BTG1, according to the six genetic subtypes, namely Ph+, TCF3-PBX1 positive, ETV6-RUNX1 positive, KMT2A rearranged, TCF3-HLF positive and B-others

Full size image

Distinct pattern of CNA according to the genetic subtypes

Next, we compared CNAs in B-ALL cell lines according to their specific genetic abnormalities. In Ph + B-ALL cell lines, IKZF1, CDKN2A/2B, BTG1, and PAX5 deletion were more enriched than in Ph negative B-ALL cell lines (Fig. 2). On the other hand, the frequency of CDKN2A/2B deletion in KMT2A rearranged cell lines was significantly lower than that in non-KMT2A rearranged cell lines (Fig. 2). These findings suggested that CNA was related to genetic subtype in B-ALL cell lines, which was similar to that in primary patients’ samples [17, 20].

Fig. 2
figure 2

Comparison between cell lines of copy number abnormalities of six genes, according to the genetic subtypes. The ratio of [CNA positive cell lines (%) with positive genetic marker]/[CNA positive cell lines (%) with negative genetic marker] is shown. In Ph + B-ALL cell lines, the frequencies of IKZF1, CDKN2A/2B, BTG1, and PAX5 deletion were significantly higher than those in Ph negative B-ALL cell lines. The frequency of CDKN2A/2B deletion in KMT2A rearranged cell lines was significantly lower than that in non-KMT2A rearranged cell lines. KMT2Ar indicates KMT2A rearranged B-ALL cell lines

Full size image

Molecular characterization of three B-other ALL cell lines harboring IKZF1 deletions

As B-other ALL cell lines harboring IKZF1 deletion are rare and considered putative Ph-like ALL cell lines [5, 21], we performed further molecular characterization of the three cell lines, YCUB-5, KOPN49, and KOPN75. Given that genetic alteration of CRLF2 is the most frequent genetic event observed in IKZF1 deleted B-other ALL [5, 21] and high expression of CRLF2 is considered to be one of the surrogate makers for Ph-like ALL [22], we performed real-time qPCR to determine levels of CRLF2 expression in these cell lines. Real-time qPCR analysis revealed high levels of CRLF2 expression in YCUB-5 and KOPN49 cell lines, which was confirmed by FACS analysis (Fig. 3a, b).

Fig. 3
figure 3

CRLF2 mRNA and protein expression levels in three IKZF1 deleted B-other ALL cell lines. a CRLF2 mRNA expression levels in three IKZF1 deleted B-other ALL cell lines, YCUB-5, KOPN49, and KOPN75, determined by quantitative RT-PCR. Expression of CRLF2 mRNA was detected in YCUB-5 and KOPN49, but not in KOPN75 cells. Reh cell lines carrying the ETV6-RUNX1 fusion were used as a negative control. The bars indicate mean ± SD of duplicate PCR assays. b CRLF2 protein expression in the three cell lines, YCUB-5, KOPN49, and KOPN75, was determined by FACS analysis. CRLF2 was expressed on YCUB-5 and KOPN49 cells

Full size image

Next, we performed multiplex PCR to determine whether any Ph-like ALL associated fusion transcripts were expressed in these three cell lines. The results revealed that YCUB-5 was positive for the P2RY8–CRLF2 fusion transcript and this was confirmed by Sanger sequencing (Fig. 4a, b); however, no fusion transcripts were detected in KOPN49 or KOPN75 (Fig. 4a). To screen for other fusion transcripts in the KOPN49 and KOPN75 cell lines, we performed mRNA-seq. These investigations identified a PAX5-ETV6 fusion transcript in KOPN75, which was confirmed by Sanger sequencing (Fig. 5a, b); however, no CRLF2-related fusion transcripts were identified in the KOPN49 cell line.

Fig. 4
figure 4

The P2RY8–CRLF2 fusion transcript was detected only in YCUB-5. a The P2RY8-CRLF2 fusion transcript was detected only in YCUB-5 by reverse transcription polymerase chain reaction (RT-PCR). b Sanger sequencing of the RT-PCR product revealed that P2RY8 exon 1 was fused in-frame to CRLF2 exon 1. PC indicates positive control from patient’s leukemic sample with P2RY8–CRLF2 fusion transcript. NTC no template control

Full size image
Fig. 5
figure 5

The PAX5-ETV6 fusion transcript was detected in KOPN75. a The PAX5-ETV6 fusion transcript was detected by reverse transcription polymerase chain reaction (RT-PCR) amplification in KOPN75. NTC no template control. b Sanger sequencing of the RT-PCR product revealed that PAX5 exon 4 was fused in-frame to ETV6 exon 3. NTC no template control

Full size image

To further molecularly characterize these B-ALL cell lines, we performed targeted capture sequencing (TCS) of 163 genes established as mutated in hematological malignancies. TCS identified the following mutations: JAK2 R683I and KRAS G12D in YCUB-5, and JAK2 R683G and KRAS G13D in KOPN49. These findings suggest that both cell lines not only have activating mutations in the JAK-STAT pathway, but also in the RAS pathway (Table 1 and sTable 4). Finally, we performed TCS using 662 probes covering the entire IgH enhancer locus, resulting in detection of an IgH-CRLF2 fusion in KOPN49 cells. The molecular and genetic characteristics of these three cell lines are summarized in Table 1.

Table 1 Summary of genetic analysis of B-other ALL cell lines with IKZF1 deletion
Full size table

Discussion

CDKN2A/2B are well-known tumor suppressor genes that map to chromosome 9p21. CDKN2A encodes p14/ARF and p16/INK4A, and CDKN2B encodes p15/INK4B [23]. All of these proteins regulate the cell cycle and apoptosis through RB or TP53. Deletion is the most common cause of inactivation of CDKN2A/2B in ALL. Although deletions of CDKN2A/2B are reported as associated with poor prognostic features, such as high white blood cell count, older age at diagnosis, IKZF1 deletion, and positive Philadelphia chromosome, the prognostic significance of this genetic alteration in B-ALL remains controversial [24, 25].

In this study, we identified deletion of CDKN2A/2B as frequent in Ph + B-ALL cell lines. It is interesting that CDKN2A and IKZF1 alterations in Ph + B- ALL are reported to synergize in promotion of leukemia development by conferring stem cell-like properties [26]. These findings could explain why CDKN2A/2B deletion is enriched in Ph + B-ALL cell lines.

We also identified deletion of BTG1 was frequent in Ph + cell. BTG1 also has anti-proliferative effects and the deletion of Btg1 was reported to increase the frequency of T cell acute lymphoblastic leukemia development in combination with Ikzf1 alteration in a mouse model, suggesting Btg1 deletion might augment leukemogenesis of Ikzf1 deletion [27, 28]. These findings also suggest that BTG1 deletion contributes to the establishment of Ph + B-ALL cell lines via a similar mechanism to CDKN2A/2B deletion. Taking together with high frequency of deletion of IKZF1 in Ph + B-ALL cell lines, alterations of IKZF1, CDKN2A/2B, and BTG1 may be necessary for the acquired characteristics of Ph + leukemic cell lines, such as high proliferation and augmented self-renewal activity.

The limitation of this study is that the genetic alterations detected in the cell lines are not always events occurring in the patients’ bone marrow. Due to the lack of data on CNA in the original patient samples from which cell lines were established, we were unable to determine whether deletion of CDKN2A/2B and BTG1 was cell line-specific or the alterations were already present in the primary leukemic blasts. Analysis of CNA in leukemic blasts from the patients of origin would answer this question.

In this study, we identified three B-other cell lines with IKZF1 deletion: KOPN75, KOPN49, and YCUB-5. As IKZF1 deletion in B-other ALL is a surrogate marker for Ph-like ALL, we conducted a comprehensive genetic analysis of these three cell lines and mRNA-seq revealed that KOPN75 had PAX5-ETV6 fusion transcript. PAX5-ETV6 was first identified by Cazzaniga and colleagues in a sample from a patient with ALL carrying a t(9;12)(q11;p13) translocation [29]. This fusion protein has also been identified in ALL patients with the recurrent chromosomal abnormality dic(9;12)(q11;p13) [30]; however, PAX5-ETV6 positive ALL is different from Ph-like ALL because, as the PAX5-ETV6 fusion protein retains the PAX5 paired-box domain and both the helix-loop-helix and ETS binding domains, hence this protein could act as an aberrant transcription factor. Consistent with this hypothesis, Fazio and colleagues determined that PAX5-ETV6 induces a B-cell development block via repression of genes activated by PAX5 [31]. Hence, KOPN75 should be a useful cell line for further investigation of the leukemogenic functions of PAX5-ETV6.

We determined that KOPN49 and YCUB-5 carry IgH-CRLF2 and P2RY8-CRLF2, respectively. Considering that both cell lines have an IKZF1 deletion, they can be considered Ph-like ALL cell lines. JAK2 mutation is established as occurring in patients with Ph-like ALL with CRLF2 rearrangement [5, 32]. In accordance with previous finding, TCS revealed JAK2 activating mutation in both cell lines. Interestingly, TCS also revealed that activating mutation of KRAS was present in both cell lines. These findings are consistent with the previous report that describing some Ph-like ALL patients harbored activating mutations in both JAK2 and RAS pathways [5].

In conclusion, we examined genetic alterations in 83 B-ALL cell lines and identified enrichment for CDKN2A/2B and BTG1 deletions particularly in Ph + B-ALL cell lines. We also identified two putative Ph-like ALL cell lines harboring IKZF1 deletion, CRLF2 rearrangement, and activating mutations of JAK2 and KRAS. The genetic data characterizing alterations in these cell lines will be useful for the investigation of B-ALL.