C/EBPβ induces B‐cell acute lymphoblastic leukemia and cooperates with BLNK mutations

Abstract BLNK (BASH/SLP‐65) encodes an adaptor protein that plays an important role in B‐cell receptor (BCR) signaling. Loss‐of‐function mutations in this gene are observed in human pre‐B acute lymphoblastic leukemia (ALL), and a subset of Blnk knock‐out (KO) mice develop pre‐B‐ALL. To understand the molecular mechanism of the Blnk mutation‐associated pre‐B‐ALL development, retroviral tagging was applied to KO mice using the Moloney murine leukemia virus (MoMLV). The Blnk mutation that significantly accelerated the onset of MoMLV‐induced leukemia and increased the incidence of pre‐B‐ALL Cebpb was identified as a frequent site of retroviral integration, suggesting that its upregulation cooperates with Blnk mutations. Transgenic expression of the liver‐enriched activator protein (LAP) isoform of Cebpb reduced the number of mature B‐lymphocytes in the bone marrow and inhibited differentiation at the pre‐BI stage. Furthermore, LAP expression significantly accelerated leukemogenesis in Blnk KO mice and alone acted as a B‐cell oncogene. Furthermore, an inverse relationship between BLNK and C/EBPβ expression was also noted in human pre‐B‐ALL cases, and the high level of CEBPB expression was associated with short survival periods in patients with BLNK‐downregulated pre‐B‐ALL. These results indicate the association between the C/EBPβ transcriptional network and BCR signaling in pre‐B‐ALL development and leukemogenesis. This study gives insight into ALL progression and suggests that the BCR/C/EBPβ pathway can be a therapeutic target.


| INTRODUC TI ON
Acute lymphoblastic leukemia (ALL) is the most common malignancy among children. Although the 5-y overall survival is approximately 90% in children with optimal diagnosis and treatments, it remains the most frequent cause of death in patients younger than 20 y of age. 1,2 The majority of childhood ALL originates from B-cell precursors such as pre-B-cells and pro-B-cells in which the B-cell receptor (BCR) signaling pathway plays an important role in cellular differentiation and survival. 3 BCR signaling is frequently suppressed in ALL and its recovery induces cell death, indicating a tumor suppressor role for this pathway. [4][5][6] BLNK/SLP65/BASH is an adaptor protein involved in BCR signaling, and it is required for B-cell maturation. 7 Recurrent lossof-function mutations in BLNK have been reported in childhood pre-B-ALL, 8,9 and a subset of BLNK homozygous knock-out (KO) mice develop pre-B-cell leukemia with a relatively long latency. 10,11 These data indicate that BLNK mutations represent important molecular aberrations for the pre-B component, but are not sufficient for complete leukemogenesis. Therefore, additional mutations might be required as cooperative genetic events for the genesis of pre-B-ALL.
In addition, a specific combination of oncogenic signaling is required to promote pre-B leukemogenesis. For example, Stat5b oncogenic signaling is enhanced by the deletion of Blnk in B-cell transformation. 12 Although genetic mutations that affect BCR signaling such as PAX5, EBF1, or IKZF1 deletions have been identified, 13 it is important to identify additional genes that cooperate with BLNK mutations and/ or aberrations of BCR signaling for understanding the leukemogenic mechanisms and to identify novel target candidates for therapies.
In this study, we identified cooperative genes for Blnk deletion using Moloney murine leukemia virus (MoMLV)-based retroviral tagging. Retroviral tagging is a powerful tool to identify important genetic events that are responsible for conferring a growth advantage to leukemic cells and facilitating disease progression. 14 The technique has been efficiently utilized to identify unknown genetic interactions for the certain mutations. Using this technique, we identified a common insertion site (CIS) in Cebpb, which encodes a CCAAT/enhancer-binding protein β that acts as a sequence-specific transcription factor. There are 2 major isoforms of the encoded protein, namely liver activating protein (LAP) and liver inhibitory protein (LIP). 15 We identified that the Blnk KO and Cebpb LAP overexpression significantly cooperated to accelerate leukemogenesis. Therefore, this study demonstrates the important role of the C/EBPβ LAP isoform in the development of B-cell malignancy and the early stages of B-cell differentiation.

| Mice
Blnk KO mice have been described previously. 16 To generate Cebpb transgenic mice, Cebpb cDNA, covering the entire coding regions of either LAP or LIP, were amplified using PCR and subcloned into the p3xFLAG-CMV-10 vector (Sigma). After sequence verification, FLAG-tagged LAP or LIP were inserted into the Lck proximal promoter and the intronic enhancer of the Ig heavy chain (Lck/EµH). 17 Transgenic mice were generated using microinjecting constructs as previously described. 18 Three independent transgenic lines for each construct were generated, and each 1 line showed protein expression ( Figure S1). These animals were mated with Blnk KO mice. Transgenic founders were identified using Southern blot hybridization using an hGH sequence as the probe. Genotyping was performed using PCR.
The sequences of all PCR primers are listed in Table S1.
Mice were monitored daily for evidence of disease, and all of the diseased animals were subjected to necropsy. Tissues were analyzed morphologically and by flow cytometry using a FACSCalibur flow cytometer (BD Biosciences). All experiments involving mice were approved by the Institutional Review Board of the Cancer Institute, Japanese Foundation for Cancer Research.

| Retroviral infection and Isolation of retroviral integration sites (RIS)
MoMLV was produced by infecting murine SC1 cells with retroviral stock. The virus-containing medium was harvested and a high titer (>1 × 10 4 pfu/mL) was confirmed by performing an XC cell assay. 19 Newborn Blnk homozygous KO, heterozygous KO or wild type mice were inoculated ip with 100 μL of viral medium.
Southern blot analysis was carried out to assess the clonal insertion and copy numbers of the retrovirus. Genomic DNA was digested with appropriate restriction enzymes, subjected to agarose gel electrophoresis and transferred to a Hybond-N nylon filter (GE Healthcare). The filter was hybridized with the env sequence of MoMLV as a probe.
RISs were identified using the inverse polymerase chain reaction (IPCR) approach as described previously. 18 Briefly, genomic DNA was digested with EcoRI, BamHI, BglII, NcoI, HindIII, or SacI, self-ligated, and subjected to nested PCR. The PCR primers for each restriction digestion are available on request. The PCR products were analyzed using agarose gel electrophoresis, subcloned into the pGEM T-easy plasmid (Promega), and subjected to sequence analysis.

| Flow cytometry
Single-cell suspensions of 1 × 10 6 bone marrow cells were incubated with specific antibodies, as indicated in Table S2 and analyzed using a FACSCalibur flow cytometer (Becton Dickinson).

| Cell separation
CD3-positive (CD3 + ) T cells and CD19 + B-cells were purified from whole spleen by magnetic separation over columns using the MACS Multisort kit in accordance with manufacturer's instructions (Miltenyi Biotec Inc).

| Immunoblot analysis
Western blot analysis was performed using total cell lysates as described previously. 20 Primary antibodies used were anti-C/EBPβ

| Preparation of RNA and real-time quantitative PCR
RNA was extracted from fresh-frozen lymph node or spleen samples using an RNeasy Mini Kit (Qiagen). Reverse transcription and RNA quantification were performed in accordance with methods described previously. 21

| Microarray and database analysis
The Mouse Genome 430 2.0 Array and HT MG-430 PM Array (Affymetrix) were hybridized with cRNA probes generated from

B220-positive bone marrow B-cells and leukemia samples from
Blnk KO/LAP tg mice, and BKO418 cells in accordance with methods described previously. 22,23 The data were analyzed using GeneSpring v.12.6 software (Agilent Technologies). Pathway analyses were performed using gene set enrichment analysis (GSEA) software. 24

| Statistical analysis
Significant differences were demonstrated for the quantitative analysis of mRNA expression using SPSS statistical software. The P-values were calculated using the Student t test, chi-square test, Scheffé method, Mann-Whitney U-test, and log-rank test. P-values < .05 were considered significant.

| Blnk mutation accelerates the induction of MoMLV-induced pre-B-ALL
To identify the genes that cooperated with BLNK mutations during B-lymphoid leukemogenesis, MoMLV was injected intraperitoneally

| Identification of Cebpb as a common MoMLV retroviral integration site
Southern blot analysis of MoMLV-induced leukemia using an env probe showed clonal integrations of the retrovirus in ALL with an average copy number of 7.6 ( Figure 2A). RISs were identified using inverse PCR followed by sequencing and sequence mapping to the mouse genome as previously described. 20 In total, 97 RISs were identified in 16 pre-B-ALL and 1 null-type ALL samples developed from Blnk homozygous KO mice (Table S3). Cebpb, Ahi1, Gfi1b, Myb, Myc, Sos1, Sfpi1, and Bmi1 were identified as CIS (Table 1). Most CIS had been already identified as CIS in other systems (http://varia tion.osu. edu/rtcgd/ index.html) 27 ; however, Cebpb and Sos1 had not been identified as CIS in B-cell malignancies. Moreover, the Cebpb locus was identified at one of the highest frequencies among 8 CIS, and most integrations at the locus were found at the 3′ end of the Cebpb gene ( Figure 2B). Although distant locations of integration sites at 3′ regions are not very common, previous studies have reported a similar integration tendency at 3′ sites. 28,29 Real-time quantitative RT-PCR showed that Cebpb mRNA was highly expressed in all pre-B-ALL samples compared with that in CD19-positive bone marrow cells derived from wild type and Blnk homozygous KO mice regardless of retroviral integration ( Figure 2C). High expression of the C/EBPβ protein was also confirmed by immunoblotting, and both LAP and LIP isoforms were expressed in MoMLV-induced pre-B-ALL ( Figure 2D).

| Lymphoid-specific expression of the LAP isoform of C/EBPβ cooperates with the Blnk mutation in pre-B-ALL development
To confirm the cooperative activity between the Blnk mutation and candidate genes identified as CIS, Cebpb transgenic mice were  Figure 3A). Expression of the LIP isoform did not have such a cooperative effect, and LIP expression by itself did not induce ALL, whereas 73% and 37% of Blnk homozygous KO and LAP transgenic mice, respectively, developed ALL within 600 days ( Figure 3A).

| Lymphoid-specific expression of the LAP isoform of C/EBPβ enhances differentiation block of pre-B-cell in BLNK homozygous KO mice
To understand the significance of LAP overexpression in Blnk KOassociated leukemogenesis, its effect on B-cell differentiation was examined. The number of mature B-cells in the bone marrow was significantly decreased in both LAP transgenic and Blnk KO mice ( Figures 4A and S2). Moreover, an accumulation of IL-7Rα/

| D ISCUSS I ON
In this study, we identified the LAP isoform of C/EBPβ as a B-cell oncoprotein and a cooperative factor for a Blnk mutation that accelerates pre-B leukemogenesis ( Figure 5D Human and murine C/EBPβ consists of 2 isoforms, LAP and LIP, which are produced using alternative start codons. 31  share an identical DNA binding domain, however LIP lacks the transactivation domain and its function is antagonistic to LAP. 32 A previous report showed that LIP can induce cell proliferation and collaborates with Evi1 in acute myeloid leukemia. 28 In addition, LIP can induce Myc expression in hematopoietic progenitor cells. 33 An appropriate cellular context and epigenetic status might be   Nevertheless, our present study underscores the important role of the C/EBPβ LAP isoform in the malignant progression of pre-B-ALL, and suggests that the BCR/C/EBPβ pathway could represent a potential therapeutic target.

D I SCLOS U R E
The authors declare no conflict of interest.

O RCI D
Takuro Nakamura https://orcid.org/0000-0002-0419-7547 F I G U R E 5 Correlation between BLNK and CEBPB expression in human pre-B-ALL. A, Gene expression of CEBPB in human patients with B-ALL (n = 73). The patients were divided into BLNK-low (n = 8) and BLNK-high (n = 65) expression groups. B, Inverse correlation between BLNK and CEBPB expression in human pediatric pre-B-ALL in TCGA database by cBioportal. N = 203 cases, r = −0.35, P < .01 using Spearman's rank correlation coefficient. C, High expression of CEBPB and low expression of BLNK (BLCH) is associated with poor prognosis among all types of pediatric pre-B-ALL. In total, 72 patients in the TARGET Phase II cohort for whom survival information was available from the cBioPortal were divided into 8 BLCH and 64 others. Hazard ratio, 2.75, using log-rank test. D, Schematic representation of BLNK and CEBPB changes in leukemogenesis