This article is mentioned in:

Introduction

Rearrangements of lysine methyltransferase 2A (KMT2A, also known as MLL) gene at chromosome 11q23 account for ~10% of all acute leukemia (AL) cases, but are also present in most infant ALs and therapy-associated acute myeloid leukemia (AML), which were previously treated with topoisomerase II inhibitors for other cancers (1).

Although >94 fusion partner genes have been found to fuse with MLL, AF4 (AFF1), AF9 (MLLT3), ENL (MLLT1), AF10 (MLLT10), ELL and AF6 (MLLT4) are the most frequent fusion partners found in Als (2). MLL translocations (MLL-t) confer a poor prognosis in AL, especially MLL/AF6 and MLL/AF10 in AML (3,4). MLL-t alters MLL methyltransferase activity and leads to dysregulation of MLL downstream genes, such as HOXA7-A10, which subsequently impairs hematopoietic lineage commitment and induces leukemia development (5,6). In addition to Hoxa7-Hoxa10 genes, sustained Hoxa11 expression has been detected in the MLL/ENL immortalized myeloid cell line (7). Chromosomal translocation t(7;11)(p15;p15) encoding NUP98/HOXA11 fusion has been recurrently detected in chronic myeloid leukemia and juvenile myelomonocytic leukemia (8,9). HOXA11, HOXA10, HOXA7 and HOXA4 are downregulated during monocyte-macrophage differentiation in a human leukemic THP-1 cell line (10). In contrast to HOXA7-HOXA10, the leukemogenic potential of HOXA11 is not well characterized. In addition to participating in leukemogenesis, the HOX family of genes are involved in organ development. In a homeobox swap experiment, it was discovered that Hoxa10 could partially replace the role of Hoxa11 in regulating skeletal phenotypes and reproductive tract development (11). However, whether these different Hoxa genes are functionally interchangeable or complementary in leukemogenesis is not clear.

Cases of AL with MLL-t are frequently found to harbor RAS pathway mutations, including N-/K-RAS and tyrosine-protein phosphatase non-receptor type 11 (PTPN11) activating mutations. The mutation rate of KRAS ranges from 7.2~42.4%, whereas that for NRAS is 5.3~24.7% and that for PTPN11 is 1~4.8% (12–15). The impact of RAS pathway mutations on MLL-t AL is controversial. This is likely due to varied mutant allele frequencies of RAS pathway mutations in patients (16,17). We and others have established mouse models with results supporting that cooperation of MLL-t with activating N-/K-RAS or PTPN11 mutations accelerate leukemia progression (18–22). Activating N-/K-RAS mutations constitutively activate the downstream signaling cascades controlling cell proliferation, apoptosis, differentiation and cell cycle progression (23). Activating PTPN11 mutations can induce myeloid cell hypersensitivity to growth factors, including granulocyte monocyte-colony stimulating factor and interleukin (IL)-3, and enhance cell cycle progression (22,24,25). We also previously demonstrated that cooperation of MLL-t with activating PTPN11 mutations increased cytarabine (Ara-C) sensitivity in leukemia cells (22).

The underlying mechanism and key downstream players that accelerate leukemia development and Ara-C sensitivity by cooperating mutations have not been clearly illustrated. Thus, in the present study, transcriptomic profiles were compared between mouse MLL/AF10(OM-LZ) leukemia cells carrying wild-type and activating KRAS or PTPN11 to identify differentially expressed genes (DEGs) involved in survival and drug sensitivity. One such upregulated DEG, Hoxa11, was further investigated to characterize its roles in leukemia cell differentiation, proliferation, survival and Ara-C sensitivity.

Materials and methods

Cell culture

The mouse MLL/AF10(OM-LZ) leukemia cell line (12G) and MLL/AF10(OM-LZ) cells harboring wild-type or activating KRAS (KRASG12C) and wild-type or activating PTPN11 (PTPN11G503A) were generated by the retroviral transduction of genes to 5-fluorouracil (5-FU)-enriched C57BL/6J (B6) mouse bone marrow (BM) cells. The mice were purchased from the National Laboratory Animal Center. These different cell types were either generated in the current study (AKw1G) or in previous studies (AK2G, AK3G, APw1 and APm1) (21,22,26) (Fig. 1A). All of these cell lines expressed the myelomonocytic markers. Mouse leukemia cells were cultured in the RPMI-1640 medium (Gibco; Thermo Fisher Scientific, Inc.) with 20% fetal bovine serum (Hyclone; Cytiva), 2 mM L-glutamine, 100 µM 2-mercaptoethanol and 10 ng/ml IL-3 (R&D Systems, Inc.) for maintenance and proliferation analysis.

Figure 1. Identification of DEGs between paired MLL/AF10(OM-LZ) leukemia cells harboring wild-type and RAS pathway mutations. (A) Establishment of immortalized cell lines by retroviral transduction of 5-FU-enriched bone marrow cells with MLL/AF10(OM-LZ) alone (12G) or in combination with wild-type KRAS (AKw1G) or oncogenic KRASG12C (AK2G, AK3G), and wild-type PTPN11 (APw-1) or oncogenic PTPN11G503A (APm-1). (B) Heat map representing relative gene expression levels of 23 DEGs between paired cell lines based on cDNA microarray data. Different Affymetrix chips were used for the following paired cell lines: 12G vs. AK3G and 12G vs. AK2G (430A); 12G vs. AK2G and APw-1 vs. APm-1 (430_2), APw-1 vs. APm-1 and AKw1G vs. AK3G (Clariom D). Raw values were log2-transformed and centered relative to the median. A heat map was obtained using Cluster version 3.0 and Java TreeView version 1.1.6r4. The color bar depicts the color contrast level of the heat map, in which red and green indicates high and low expression, respectively. Grey indicates genes absent in the Affymetrix 430A chip. (C) Enriched GO terms of ‘BP’ and ‘MF’ for the 23 identified DEGs. All GO terms listed in the table show significant enrichment (all P<0.05). Red and black indicate genes that are upregulated or downregulated, respectively, in MLL/AF10 cell lines harboring RAS pathway mutations. DEGs, differentially expressed genes; MLL, lysine methyltransferase 2A; GO, Gene Ontology; BP, biological process; MF, molecular function; PTPN11, tyrosine-protein phosphatase non-receptor type 11; BM, bone marrow; 12G, cells with MLL/AF10(OM-LZ) alone; AK3G, cells with MLL/AF10(OM-LZ) and oncogenic KRASG12C; AKw1G, cells with MLL/AF10(OM-LZ) and wild-type KRAS; APw-1, cells with MLL/AF10(OM-LZ) and wild-type PTPN11; APm-1, cells with MLL/AF10(OM-LZ) and oncogenic PTPN11G503A; PTPN11, tyrosine-protein phosphatase non-receptor type 11.

Microarray analysis and Gene Ontology (GO) enrichment analysis

The total RNA of leukemia cells was prepared using TRIzol® reagent (Gibco; Thermo Fisher Scientific, Inc.). The RNA was amplified, labeled and hybridized to the mouse genome 430A Array chip (12G vs. AK3G and 12G vs. AK2G), 430 2.0 Array chip (12G vs. AK2G and APw-1 vs. APm-1), or Clariom D Array chip (APw-1 vs. APm-1 and AKw1G vs. AK3G) (Affymetrix; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions (this procedure was performed by staff of the Genomic Medicine Research Core Laboratory at Chang Gung Memorial Hospital, Linkou, Taiwan). Microarray data are available at the NCBI Gene Expression Omnibus (GEO) website (https://www.ncbi.nlm.nih.gov/geo/; accession nos. GSE82156 and GSE134586) or can be downloaded from Chang Gung University website (21). Differential Expression Analysis was performed using Transcriptome Analysis Console software version 4.0 (Affymetrix; Thermo Fisher). A heat map was obtained using Cluster version 3.0 (http://bonsai.hgc.jp/~mdehoon/software/cluster/) and Java TreeView version 1.1.6r4 (http://jtreeview.sourceforge.net). DEGs with >1.5-fold-change in paired MLL/AF10 leukemia cells harboring wild-type and oncogenic KRAS or PTPN11 were then functionally annotated with GO enrichment analysis using online Database for Annotation, Visualization and Integrated Discovery (DAVID) version 6.8 annotation tools (https://david.ncifcrf.gov/). Statistical significance was evaluated using Fisher's exact test and corrected by Bonferroni correction for multiple testing. The GO ‘Molecular Function’ (MF) or ‘Biological Process’ (BP) categories with P<0.05 were considered statistically significant.

Validation of HOXA11 expression in patients with AML

To validate the differential expression and drug responsiveness of HOXA11 in MLL-t AML, a meta-analysis was performed using the Oncomine™ database (https://www.oncomine.org/), including Valk leukemia, Wouters leukemia, Balgobind leukemia, and Haferlach leukemia (27–30).

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR)

To evaluate the expression levels of target genes, the total RNA of mouse leukemia cells or AML patient BM cells was extracted using TRIzol® reagent (Invitrogen; Thermo Fisher). The total RNA was reverse transcribed into complementary DNA (cDNA) using random hexamers and SuperScript™ II reverse transcriptase (Thermo Fisher) according to the manufacturer's protocol. qPCR was performed using SYBR-Green PCR master mix (Applied Biosystems; Thermo Fisher Scientific, Inc.) and analyzed by ABI Prism 7900 system (Applied Biosystems; Thermo Fisher Scientific, Inc.). Primer sets were as follows: Mouse Hoxa11, 5′-GAAAACCTCGCTTCCTCCGA-3′ and 5′-ATAAGGGCAGCGCTTTTTGC-3′; mouse NF-κB inhibitor α (Nfkbia), 5′-GAGACCTGGCCTTCCTCAAC-3′ and 5′-TCTCGGAGCTCAGGATCACA-3′; mouse transcription factor p65 (Rela), 5′-TGGCTACTATGAGGCTGACCT-3′ and 5′-TGGTCTGGATTCGCTGGCTA-3′; mouse transformation-related protein p53 (Trp53), 5′-CCTCTCCCCCGCAAAAGAAA-3′ and 5′-GGCCCTTCTTGGTCTTCAGG-3′; mouse Gapdh, 5′-TTCACCACCATGGAGAAGGC-3′ and 5′-GGCATGGACTGTGGTCATGA-3′; human HOXA11, 5′-CGTCTTCCGGCCACACTGA-3′ and 5′-AGACGCTGAAGAAGAACTCCC-3′; and human ABL, 5′-TGGAGATAACACTCTAAGCATAACTAAAGG-3′ and 5′-GATGTAGTTGCTTGGGACCCA-3′. The thermocycling conditions were as follows: 95°C for 2 min; then 40 cycles of 95°C for 15 sec and 65°C for 30 sec. The gene expression levels of mouse genes and human HOXA11 were normalized against the housekeeping genes Gapdh and ABL, respectively. Fold-change was calculated using the 2−∆∆Cq method (31). In the present study, leftover BM samples of clinical examination for initial diagnostic work-up of AML were used for gene expression analysis. Samples were obtained from patients admitted to Chang Gung Memorial Hospital (Taipei, Taiwan) between January 2002 and December 2010. The inclusion criteria were as follows: Adults and children with AML with MLL-t. The exclusion criteria were as follows: Non-MLL-t AML specimens. Among the 114 cases with MLL-t AML (56 men and 58 women, age range 0–84 years, median 27 years), it was determined that eight cases with MLL/AF10 or MLL/AF9 had sufficient remaining specimens available for HOXA11 expression level analysis. Of these eight cases, five had KRAS mutations and one had a PTPN11 mutation.

Western blot analysis

Total cell lysate from 5×106 leukemia cells was prepared by direct lysis of cells with RIPA buffer [20 mM Tris-Cl (pH 7.5) 150 mM NaCl, 1% Triton X-100, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 0.5 mM PMSF]. The amount of total protein was assayed using a Pierce® BCA Protein Assay (Pierce; Thermo Fisher Scientific, Inc.). The lysates (20 µg/lane) were electrophoresed on 10% polyacrylamide gel, and subsequently transferred to an Immobilon membrane (EMD Millipore). The membrane was blocked in 5% bovine serum albumin at 4°C (Sigma-Aldrich; Merck KGaA) for 1 h, and then incubated with primary antibodies at 4°C overnight against mouse Hoxa11 (1:5,000; cat. no. NBP1-80228; Novus Biologicals, Ltd.), β-actin or Gapdh (1:10,000; cat. nos. sc-47778 or sc-32233; Santa Cruz Biotechnology, Inc.), followed by incubation for 2 h at room temperature with horseradish peroxidase-conjugated anti-mouse or anti-rabbit IgG secondary antibodies (1:5,000; cat. nos. C04001 or C04003; Croyez Bioscience Co., Ltd.). Western blots were developed with a Western Lightning Plus ECL kit (PerkinElmer, Inc.) and the images were visualized by Analytik Jena™ UVP ChemStudio PLUS and VisionWorks™ software (version 9.0; Analytik Jena US LLC).

In vivo leukemogenesis

The in vivo leukemogenic potential of the leukemia cell lines was determined by BM transplantation assay using male B6 mice (n=45, age, 6–8 weeks; weight, 20–24 g). Mice were maintained in pathogen-free devices under a controlled animal housing conditions (temperature 20±3°C, humidity 60–70% with a 12-h light /dark cycle and access to food and water ad libitum). Briefly, leukemia cells (APm1-shV, APm1-shH11-2, 12G-V, and 12G-H11) were intraperitoneally (i.p.) injected into mice (1×106 cells/100 µl 1X PBS/mouse; n=10 mice for each leukemia cell line) that had received a sublethal dose of γ-irradiation (5.25 Gy) unless otherwise stated. Mice i.p. injected with normal saline (n=5) served as controls. To monitor leukemia development, peripheral blood (100 µl) was collected weekly for cytologic analysis and complete blood count measurement using Hemavet 950 (Drew Scientific; Erba Diagnostics, Inc.) or BC-5000 (Mindray Medical International Limited) hemocytometers. Mice were sacrificed when moribund (7–15 weeks post-injection). The moribund state was defined as mice displaying leukocytosis, with hunched posture, weakness, shortness of breath and 20% weight loss. Mice were euthanized by an i.p. injection of Zoletil (50 mg/kg) and Rompun (xylazine, 10 mg/kg) or by inhalation of isoflurane (3–5%), followed by cervical dislocation (32). BM, peripheral blood, ascites, organs and tumor masses were collected and weighed.

Gene knockdown by short hairpin RNA (shRNA)

To generate stable gene knockdown cell lines, AK3G or APm-1 cells were infected with lentivirus expressing shRNA against Hoxa11 [The RNAi Consortium (cat. nos. TRCN0000413738 and TRCN0000417739)] at a multiplicity of infection of 1 and selected in RPMI-1640 medium containing puromycin (2.5 µg/ml) for a total of 2 weeks. A third generation system was used. Cells stably transfected with blank lentiviral vector pLKO_025 were used as negative controls. All lentiviruses were obtained from the National RNAi Core Facility at the Institute of Molecular Biology/Genomic Research Center, Academia Sinica (Taiwan).

Ectopic expression of Hoxa11

The full-length Hoxa11 gene (~1 Kb) was amplified from cDNA of AK3G cells by PCR using the following primers: 5′-GAAGATCTCCCAAGGTAGCCCAATGATG-3′ and 5′-CCGCTCGAGCCAGTAGGCTGGAGCCTTAG-3′. The PCR product was digested with restriction enzymes BglII and XhoI, and was subsequently cloned into the BglII and XhoI sites of the retroviral vector pMSCVpuro (Clontech Laboratories, Inc.). The fidelity of nucleotide sequences of the Hoxa11 gene was confirmed by Sanger sequencing. Plasmid DNAs of pMSCV-Hoxa11 and pMSCVpuro were transfected into EcoPack2-293 cells (Invitrogen; Thermo Fisher Scientific, Inc.) to package retroviruses. Viral titer was determined by infection of NIH/3T3 (American Type Culture Collection; ATCC® CRL-1658™), a murine fibroblast cell line, with serial diluted supernatant to generate puromycin-resistant colonies. To generate 12G cell lines with ectopic expression of Hoxa11 (12G-H11-1 and 12G-H11-2), 12G cells were mixed with retroviruses at a 1:1 ratio and selected in RPMI complete medium containing puromycin (2.5 µg/ml) for a total of 2 weeks. The cells transduced with MSCVpuro retroviruses were used as negative controls (12G-V1 and 12G-V3). The presence of Hoxa11 gene was confirmed by PCR amplification of the 1-Kb product from the genomic DNA of 12G-H11 cell lines using cloning primers. The fidelity of nucleotide sequences of Hoxa11 was confirmed by Sanger sequencing.

Phenotypic and Ara-C resistance analyses

For cytologic analysis, cells were cytospinned at 700 × g for 3 min or smeared, air-dried, and stained with Liu reagents (Tonyar Biotech, Inc.) at room temperature. For immunophenotypic analysis, cells were stained at 4°C for 15 min with phycoerythrin-macrophage-1 antigen (Mac-1), phycoerythrin-CD115 and allophycocyanine-Ki-67 antibodies (cat. nos. RM2804-3, 12-1152-82 and 17-5698-82; eBioscience; Thermo Fisher) followed by flow cytometric analysis using FACSCanto II Cell Analyzer and FACSDiva software version 5.0 (BD Biosciences). For cell proliferation analysis, cells were assessed at indicated time points using Cell Counting Kit-8 (CCK-8/WST-8; Dojindo Molecular Technologies, Inc.) according to the manufacturer's instructions. To determine Ara-C resistance, cells were cultured in RPMI-1640 complete medium and a gradient concentration of Ara-C (0, 128, 320, 640, 1,600, 3,200, 16,000 or 40,000 ng/ml) for 24 h. Cell viability was measured using CCK-8 (incubation time 2 h). To determine apoptotic cell rate, cells were treated with Ara-C (0 or 160 ng/ml) for 24 h, followed by Annexin-V/propidium iodide (Sigma-Aldrich; Merck KGaA) staining in the dark and flow cytometric analysis using a FACSCanto II Cell Analyzer and FACSDiva software version 5.0 (BD Biosciences).

Competitive engraftment and clonal expansion assay

The competitive engraftment and clonal expansion assay was described previously (22,33). Briefly, paired cells (AK3G-shV vs. AK3G-shH11 and 12G-V vs. 12G-H11) were mixed at a ratio of 1:1 and then i.p. injected into B6 mice (1×106 cells/mouse). The mice were sacrificed at days 43 and 57. Mice BMs and spleens were collected and used to extract genomic DNA. To amplify 300-bp DNA fragments of the pMSCVpuro vector (from 12G-V) or the region spanning the pMSCVpuro-Hoxa11 junction (from 12G-H11) from genomic DNA, PCR was performed using the following primers: MSCV 5′ primer 5′-CCCTTGAACCTCCTCGTTCAGCC-3′ in combination with MSCV 3′ primer 5′-GAGACGTGCTACTTCCATTTGTC-3′ or Hoxa11 primer 5′-GAGTAGCAGTGGGCCAGATTGC-3′. The PCR products were sequenced using MSCV 5′ primer, and the peak height of the 62th nucleotide (C for 12G-H11 and T for 12G-V) was measured. The (C/C+T) peak height ratio was converted to the cell ratio (12G-H11/12G-H11+12G-V) by aligning to the standard curve. The standard curve was generated by assessing the relationship between nucleotide peak height ratios and cell ratios from cell mixtures with mixed paired cells in ratios of 10:0, 9:1, 8:2, 7:3, 6:4, 5:5, 4:6, 3:7, 2:8, 1:9 and 0:10.

Ethics statement

All animal experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals published by National Institutes of Health (publication no. 85–23, revised 1996) and were carried out according to the protocol approved by the Animal Research Committee of Chang Gung Memorial Hospital (IACUC No. 2014092403; Taoyuan, Taiwan). Human sample collection was conducted in accordance with the Declaration of Helsinki and was approved by the Chang Gung Memorial Hospital Research Ethics Committee (IRB No. 96-1748B). Written informed consent was obtained from all patients. For patients under the age of 18, consent/permission was obtained from the parents/guardians.

Statistical analysis

Statistical analyses were performed using SPSS software version 20.0 (SPSS, Inc). The statistical significance of differences in gene expression levels of the two groups was compared using a Mann-Whitney test. Survival analysis was conducted according to the Kaplan-Meier method, and differences in survival were assessed using the log-rank test. Drug sensitivity was compared using an unpaired two-sample Student's t-test. P<0.05 was considered to indicate a statistically significant difference.

Results

Identification of DEGs and GO enrichment analysis

Our previous studies demonstrated that mice transplanted with AK2G, AK3G or APm-1 had shorter survival than those transplanted with the 12G and APw-1, respectively (21,22,26). Moreover, APm-1 cells have been reported to be more resistant to daunorubicin, but more sensitive to Ara-C than APw-1 cells (22). To identify genes involved in the acceleration of leukemia development and drug resistance by cooperation of MLL/AF10 with RAS signaling mutations, transcriptomic profiling were compared between 6 paired MLL/AF10(OM-LZ) leukemia cells harboring wild-type and oncogenic KRAS (12G vs. AK3G, 12G vs. AK2G and AKw1G vs. AK3G) or PTPN11 (APw-1 vs. APm-1) using cDNA microarray data. A total of 23 DEGs (seven upregulated and 16 downregulated) with >1.5-fold-change were identified in AK3G and APm-1 cell lines compared with 12G/AKw1G and APw-1 cells, respectively (Fig. 1B). The GO terms ‘BP’ and ‘MF’ enriched in the 23 DEGs included ‘innate immune response’, ‘immune system process’, ‘actin filament binding’, ‘cellular response to interferon-alpha’, ‘sequence-specific DNA’ and ‘regulation of gene expression’ (Fig. 1C). Among these enriched GO terms, three terms (‘innate immune response’, ‘immune system process’ and ‘cellular response to interferon-alpha’) are related to the immune response; ‘actin filament binding’ is related to cell protrusion and migration, while the last two terms (‘sequence-specific DNA’ and ‘regulation of gene expression’) are related to the regulation of gene expression by transcription factors.

Expression of transcription factors in patients with AML with MLL-t and/or RAS signaling mutations

A total of four genes, Hoxa11, PR domain zinc finger protein 5 (Prdm5), Iroquois-class homeodomain protein IRX-5 (Irx5) and homeobox protein PKNOX2 (Pknox2), were mapped to the GO ‘BP’ term ‘sequence-specific DNA’; other than Pknox2, these genes were upregulated in the AK3G and APm-1 cell lines according to cDNA microarray data (Fig. 1B). To investigate whether these genes were upregulated in the patients with AML with MLL-t, KRAS or PTPN11 mutations, meta-analysis was performed using datasets deposited by Valk (285 AML cases), Wouters (503 AML cases), Haferlach (542 AML cases), and Balgobind (237 childhood AML cases) in the Oncomine™ clinical research data repository for gene expression change. The results revealed that two (Balgobind and Haferlach) of the four AML series showed significant differences (P<0.01) in HOXA11 expression for patients with AML with or without MLL-t (11q23-r) using reporter ID 208493 (Fig. 2A). One (Valk) of the two AML series showed significant differences (P<0.05) in HOXA11 expression for patients with AML with or without KRAS activating mutations (KRASm) (Fig. 2B). In the Balgobind series, there was a trend of higher HOXA11 expression in patients with AML with PTPN11 mutations than those without PTPN11 mutations, but the difference was not statistically significant (Fig. 2C). This is probably due to the low case numbers (n=5). On the other hand, there were no significant differences in IRX5 and PRDM5 expression in leukemia cells of patients with AML with or without MLL-t, KRAS or PTPN11 mutations (Fig. S1A-F).

Figure 2. Overexpression of HOXA11 in AML carrying MLL-t, oncogenic KRAS or PTPN11 mutations. (A-C) Meta-analysis of HOXA11 expression in AML using leukemia database deposited by Valk (n=285), Wouters (n=503), Haferlach (n=542) and Balgobind (n=237, childhood AML) in Oncomine™. Box plots are data of HOXA11 (reporter ID. 208493) based on cDNA microarray data in patients with AML (A) with 11q23-r or No 11q23-r, (B) patients with KRASm or KRASwt, and (C) patients with PTPN11m or PTPN11wt. Numbers listed at the bottom are case numbers. Center line in box plot represents median value, box limits are 10 and 90th percentiles, and dots represent minimum and maximum values. P-values were determined using an unpaired two-sample Student's t-test. *P<0.05 and **P<0.01. (D and E) Reverse transcription-quantitative PCR analyses were performed to determine relative HOXA11 expression level in patients with AML carrying (D) MLL/AF10 or (E) MLL/AF9 with or without oncogenic KRASm or PTPN11m mutations. Assays were performed in triplicate and data are representative of three independent experiments. Error bars indicate SD of mean. AML, acute myeloid leukemia; MLL, lysine methyltransferase 2A; MLL-t, MLL translocations; PTPN11, tyrosine-protein phosphatase non-receptor type 11; 11q23-r, MLL rearrangements; No 11q23-r, wild-type MLL; KRASm, oncogenic KRAS mutations; KRASwt, wild-type KRAS; PTPN11m, oncogenic PTPN11 mutations; PTPN11wt, wild-type PTPN11.

HOXA11 expression was also determined in the patients with MLL-t AML included in the present study

The results showed that MLL/AF10+ patients with AML with KRAS or PTPN11 mutations had higher levels of HOXA11 expression than patients with wild-type KRAS and PTPN11 genes (Fig. 2D). Similarly, MLL/AF9+ patients with AML with KRAS mutations expressed higher levels of HOXA11 compared with the patient with wild-type KRAS (Fig. 2E). These results supported that cooperation of MLL-t with oncogenic KRAS or PTPN11 mutation induces HOXA11 expression in patients with AML.

Differential expression of Hoxa genes in mouse MLL/AF10 leukemia cells with different RAS signaling mutations

Based on the cDNA microarray data, Hoxa10 and Hoxa11 were upregulated (>1.5-fold) in AK3G compared with 12G cells (Fig. 3A). Compared with AKw1G, AK3G cells had higher expression levels (>1.5-fold) of Hoxa5, Hoxa6, Hoxa10 and Hoxa11 (Fig. 3B). Compared with APw-1, only Hoxa11 was upregulated in APm-1 cells (Fig. 3C). These data suggested that KRAS and PTPN11 mutations had overlapping, but not exactly the same effects on Hoxa gene expression in MLL/AF10 leukemia cells. Conversely, no significant differences in the expression levels of Hoxb, Hoxc, Hoxd and Meis1 genes were observed in these three pairs of mouse leukemia cell lines (Fig. 3A-C).

Figure 3. Expression levels of Hox clusters A, B, C, D and Meis1 genes in paired MLL/AF10 leukemia cells harboring wild-type or oncogenic RAS pathway mutations. (A-C) Heat maps of genes based on cDNA microarray data between paired cell lines: (A) 12G and AK3G, (B) AKw1G and AK3G, and (C) APw-1 and APm-1. Raw values are log2-transformed. Red and black indicate high and low levels of gene expression, respectively. Color bar depicts log2-transformed value of genes. (D) Reverse transcription-quantitative PCR analysis was performed to determine the level of Hoxa11 expression in 12G, AKw1G, AK3G, AK2G, APw-1 and APm-1 cells. Assays were performed in triplicate and data are representative of three independent experiments. Error bars indicate SD. (E) Western blotting was performed to determine the protein level of Hoxa11 in cells of the different cell lines. The detection of β-actin and Gapdh served as a loading control for immunoblot analysis. MLL, lysine methyltransferase 2A; 12G, cells with MLL/AF10(OM-LZ) alone; AK3G, cells with MLL/AF10(OM-LZ) and oncogenic KRASG12C; AKw1G, cells with MLL/AF10(OM-LZ) and wild-type KRAS; APw-1, cells with MLL/AF10(OM-LZ) and wild-type PTPN11; APm-1, cells with MLL/AF10(OM-LZ) and oncogenic PTPN11G503A; PTPN11, tyrosine-protein phosphatase non-receptor type 11.

RT-qPCR and western blotting confirmed that the transcriptional and translational levels of Hoxa11 were increased in MLL/AF10(OM-LZ) mouse leukemia cell lines harboring KRAS mutation (AK2G, AK3G) and PTPN11 mutation (APm-1) compared with cells harboring wild-type KRAS (12G, AKw1G) or wild-type PTPN11 (APw-1) (Fig. 3D and E).

Role of Hoxa11 in survival

To determine the role of Hoxa11 in the leukemogenesis of MLL/AF10 leukemia cells harboring RAS pathway mutations, two lentivirus-based shRNAs that targeted mouse Hoxa11 gene (shH11-1 and shH11-2) were stably transduced into APm-1 cells. RT-qPCR analysis of the Hoxa11 knockdown APm-1 (APm-1-shH11-1 and APm-1-shH11-2) cell lines showed that the expression levels of Hoxa11 were reduced to 86 and 46%, respectively, compared with the control cell line (APm-1-shV) (Fig. 4A). By combining the fold changes in Hoxa11 between APw-1 and APm-1 and between APm-1-shV and APm-1-shH11-1, the expression level of Hoxa11 in APm-1-shH11-2 cells was estimated to be 2.4-fold higher than in APw-1 cells (Figs. 3D and 4A). Western blot analysis further confirmed that the protein levels of Hoxa11 were decreased in APm-1-shH11 cell lines compared with the APm-1-shV control cell line (Fig. 3E). Mice i.p. injected with APm-1-shH11-2 cells, which had lower Hoxa11 expression, had significantly longer survival than those injected with APm-1-shV cells (median 64 days vs. 50 days; P<0.01; Fig. 4B and Table I). To further support the role of Hoxa11 in the survival of leukemic mice, Hoxa11 overexpressed 12G cells (12G-H11-1 and 12G-H11-2) were generated by transduction of retrovirus-based full-length Hoxa11 gene into 12G cells. The 12G cells transduced with empty retroviruses (12G-V1 and 12G-V3) served as controls. Overexpression of Hoxa11 in the 12G-H11-1 and 12G-H11-2 cell lines compared with the control 12G-V1 and 12G-V3 cell lines was confirmed by western blotting and RT-qPCR (Figs. 3E and 4C). BM transplantation assay data showed that 12G-H11-1 mice had significantly shorter survival than control 12G-V1 mice (median 76 days vs. 91 days; P<0.001; Fig. 4D and Table I). These results indicated that Hoxa11 plays a critical role in the survival of leukemic mice induced by MLL/AF10 leukemia cells and MLL/AF10 cells harboring activating PTPN11 mutation.

Figure 4. Expression of Hoxa11 affects survival of MLL/AF10 leukemia mice. (A and C) Reverse transcription-quantitative PCR analyses were performed to determine Hoxa11 expression in (A) Hoxa11-knockdown APm-1 (APm-1-shH11-1, APm-1-shH11-2) and control (APm-1-shV) cell lines, or in (C) Hoxa11-overexpression 12G (12G-H11-1, 12G-H11-2) and control (12G-V1 and 12G-V3) cell lines. Assays were performed in triplicate and data shown are representative of three independent experiments. Error bars indicate SD. (B and D) Survival curves of mice i.p. injected with (B) APm-1-shV or APm-1-shH11-2, and (D) 12G-V1 or 12G-H11-1 cells. Survival analysis was conducted according to the Kaplan-Meier method. **P<0.01 and ***P<0.001 vs. APm-1-shV or 12G-V1 cells. (E) Flow cytometry analyses was performed to determine Ki-67 and Mac-1 expression in the BM cells obtained from 12G-V1 and 12G-H11-1 leukemia mice at moribund stage. Data shown are representative of three mice with similar results. (F) Competitive engraftment and clonal expansion ability between 12G-V1 and 12G-H11-1 cells in vivo. 12G-V1 and 12G-H11-1 cells were mixed in a 1:1 ratio and i.p. injected into recipient mice (n=9). The mice were sacrificed at 43 and 57 days post-transplantation (n=2 and n=7, respectively). Allele burden of the 12G-H11-1 clone in BM (left column) or spleen (right column) was determined by PCR-DNA sequencing. MLL, lysine methyltransferase 2A; APm-1, cells with MLL/AF10(OM-LZ) and oncogenic PTPN11G503A; sh, short hairpin RNA; 12G, cells with MLL/AF10(OM-LZ) alone; i.p. intraperitoneally; Mac-1, macrophage-1 antigen; BM, bone marrow; PTPN11, tyrosine-protein phosphatase non-receptor type 11.

Table I. Phenotypic characteristics of the mice transplanted with the Hoxa11-knockdown APm-1 and Hoxa11-overexpression 12G leukemia cells.

Table I.

Phenotypic characteristics of the mice transplanted with the Hoxa11-knockdown APm-1 and Hoxa11-overexpression 12G leukemia cells.

A, Hoxa11-knockdown APm-1 cells
Features	APm-1-shV	APm1-shH11	P-value
Number of mice, n	10.0	10.0
Survival median, days, n (range)a	50.0 (47.0–56.0)	64.0 (53.0–86.0)	0.00543
WBC, 1×109/ml, median (range)a	64.2 (58.0–97.6)	85.5 (41.2–207.2)	0.25160
Anemia, n/n (%)b	4/4 (100.0)	6/6 (100.0)	1.00000
Thrombocytopenia, n/n (%)b	1/4 (25.0)	3/6 (50.0)	0.57100
Ascites, n/n (%)b	3/8 (37.5)	0/8 (0.0)	0.20000
Hepatosplenomegaly, n/n (%)	8/8 (100.0)	9/9 (100.0)	1.00000
Myeloid sarcoma, n/n (%)	1/8 (12.5)	1/8 (12.5)	1.00000
B, Hoxa11-overexpression 12G cells
Features	12G-V	12G-H11	P-value
Number of mice, n	10.0	10.0
Survival median, days, n (range)a	91.0 (80.0–102.0)	76.0 (67.0–82.0)	0.00023
WBC, 1×109/ml, median (range)a	151.6 (61.1–354.6)	58.8 (23.3–149.1)	0.07290
Anemia, n/n (%)b	1/8 (12.5)	2/5 (40.0)	0.51049
Thrombocytopenia, n/n (%)b	2/8 (25.0)	4/7 (57.1)	0.31469
Ascites, n/n (%)b	2/9 (22.2)	0/8 (0.0)	0.47059
Hepatosplenomegaly, n/n (%)	8/8 (100.0)	9/9 (100.0)	1.00000
Myeloid sarcoma, n/n (%)	1/9 (22.2)	0/8 (0.0)	1.00000

a Student's t-test

b Fisher's exact test. MLL, lysine methyltransferase 2A; APm-1, cells with MLL/AF10(OM-LZ) and oncogenic PTPN11^G503A; sh, short hairpin RNA; 12G, cells with MLL/AF10(OM-LZ) alone; PTPN11, tyrosine-protein phosphatase non-receptor type 11; WBC, white blood cells.

Compared with the control cell lines, Hoxa11-knockdown APm-1 cells and Hoxa11-overexpression 12G cells had similar blast-like morphology (Fig. S2A and C, left column), similar percentages of CD115+ cells (Fig. S2A and C, right column), and similar cell proliferation rates in vitro (Fig. S2B and D). These results suggested that the role of Hoxa11 in the acceleration of disease progression in the leukemia mice was not caused by changes in differentiation potential or proliferation rate of the MLL/AF10 leukemia cells. Flow cytometry analysis of Ki-67, which detects proliferating cells, was performed on the leukemia cells (Mac1+) obtained from BM of 12G-V1 and 12G-H11-1 mice. The results showed that 12G-H11-1 cells were more actively proliferating in vivo (Fig. 4E). To investigate whether Hoxa11 enhanced growth advantage of MLL/AF10 leukemia cells in vivo, competitive engraftment and clonal expansion assays were performed. Compared with the control 12G-V1, 12G-H11-1 cells were more competitive to engraft to and expand in BM and spleen (Figs. 4F and S3A-C). These results suggested that Hoxa11 promotes disease progression, at least partly, by promoting leukemia cell recruitment and proliferation in their niche.

On the other hand, AK3G cells were also stably transduced with the two lentivirus-based shRNAs. Although the knockdown efficiencies of the shRNAs were significant (Fig. S4), the Hoxa11 expression levels in AK3G-shH11-2 cells were estimated to be 16.9-fold and 21.7-fold higher than that of 12G and AKw1G cells, respectively, based on the combined RT-qPCR results (Figs. 3D and S4). Moreover, AK3G cells had higher expression levels of Hoxa5, Hoxa6 and Hoxa10 compared with AKw1G (Fig. 3B). Due to the high Hoxa11 expression levels and the possible complementary effect of the Hoxa5-a10 genes (11), no further assays were performed to assess survival and drug sensitivity of AK3G-shV and AK3G-shH11 cells in this study.

Role of Hoxa11 in Ara-C resistance

To determine whether Hoxa11 affects Ara-C resistance of MLL/AF10 leukemia cells harboring PTPN11G503A, APm-1-shV and APm-1-shH11-2 cells were treated with a gradient concentration of Ara-C for 24 h. Compared with APm-1-shV cells, APm-1-shH11-2 cells exhibited significantly higher viability at Ara-C concentrations between 128–1,600 ng/ml (P<0.01; Fig. 5A). The median lethal dose (LD50) of Ara-C for APm-1-shV and APm-1-shH11-2 cells was 235 and 315 ng/ml, respectively. By contrast, Hoxa11-overexpressing 12G-H11-1 cells showed significantly lower cell viabilities at all tested concentrations of Ara-C (P<0.005; Fig. 5B). The LD50 of Ara-C for 12G-V and 12G-H11 cells was 459 and 224 ng/ml, respectively. Further studies revealed that the apoptosis rate of APm-1-shH11-2 cells was lower than that of APm-1-shV cells before (17.6 vs. 25.4%, respectively) and after (41.3 vs. 45.2%, respectively) Ara-C treatment (160 ng/ml) for 24 h (Fig. 5C). The apoptosis rate of 12G-H11-1 cells was higher than that of 12G-V1 cells before (30.5 vs. 14.9%, respectively) and after Ara-C treatment (57.8 vs. 37.2%, respectively) (Fig. 5D). These results indicated that cooperation of MLL/AF10 with PTPN11G503A upregulated Hoxa11, which in turn increased cell apoptosis and rendered cells more sensitive to Ara-C.

Figure 5. Hoxa11 expression affects chemotherapy drug resistance. (A-D) Viability assays of (A and C) Hoxa11-knockdown APm-1 (APm-1-shH11-2) and control (APm-1-shV) cell lines or (B and D) Hoxa11-overexpression 12G (12G-H11-1) and control (12G-V1) cell lines treated with Ara-C at indicated concentrations for 24 h. The viability of leukemia cells was determined by a Cell Counting Kit-8 assay. Assays were performed in triplicate and data are representative of three independent experiments. The error bars indicate SD, and P-values were determined with an unpaired two-sample Student's t-test. (E) The expression levels of the apoptosis-related genes, Nfkbia, Rela and Trp53 in Hoxa11-knockdown or Hoxa11-overexpression MLL/AF10 leukemia cells were evaluated via reverse transcription-quantitative PCR. *P<0.05, **P<0.01, ***P<0.001 vs. APm-1-shV or 12G-V1 cells. (F) Meta-analysis of HOXA11 expression in AML using a leukemia database deposited by Heuser (n=33) in Oncomine™. Box plots are the HOXA11 (reporter ID. AA598674) expression levels based on cDNA microarray data in patients with AML grouped by (F-a) chemotherapy responsiveness and (F-b) AML induction/consolidation response status. Numbers listed at the bottom are the number of cases. The center line in the box plot represents the median, the box limits indicate the 10 and 90th percentiles, and dots represent minimum and maximum values. MLL, lysine methyltransferase 2A; APm-1, cells with MLL/AF10(OM-LZ) and oncogenic PTPN11G503A; sh, short hairpin RNA; 12G, cells with MLL/AF10(OM-LZ) alone; PTPN11, tyrosine-protein phosphatase non-receptor type 11; Ara-C, cytarabine; Nfkbia, NF-κB inhibitor α; Rela, transcription factor p65; Trp53, transformation-related protein p53; AML, acute myeloid leukemia; CR, complete remission; Res, responder; Non-Res, non-responder.

As Hoxa11 encodes a DNA-binding transcription factor and affects apoptosis of leukemia cells, the present study next determined the expression levels of three apoptosis-related genes in Hoxa11-knockdown APm-1 and Hoxa11-overexpression 12G cells using RT-qPCR analysis. It was found that the silencing of Hoxa11 significantly increased the expression levels of Nfkbia, Rela and Trp53, whereas the overexpression of Hoxa11 significantly reduced the expression level of these genes (Fig. 5E). These results suggested that Hoxa11 induces cell apoptosis, at least partly, via regulation of apoptosis-related gene expression.

To determine whether the expression of HOXA11 in leukemia cells of patients with AML was associated with chemotherapy drug sensitivity, a meta-analysis was performed using a data set deposited by Heuser consisting of 33 AML cases enrolled in the AML-SHG 01/99 trial, in the Oncomine™ clinical research data repository for gene expression change (34). The results of this analysis showed that responders of chemotherapy or AML induction/consolidation had higher HOXA11 expression levels than non-responders using reporter AA598674 (Fig. 5F-a and F-b). Collectively, these findings indicated that patients with AML with higher HOXA11 expression are associated with an improved response to chemotherapy with Ara-C.

Discussion

The present study compared transcriptomic profiling between mouse MLL/AF10 leukemia cells harboring wild-type and activating KRAS or PTPN11, and found that Hoxa7-Hoxa10 were expressed in all MLL/AF10 leukemia cell lines, whereas Hoxa11 was only expressed in MLL/AF10 leukemia cells with activating KRAS or PTPN11 mutations (Fig. 3A-C). Furthermore, a meta-analysis using microarray datasets deposited in Oncomine™ and an analysis of our clinical samples indicated that HOXA11 is upregulated in MLL-t AML with RAS signaling mutations. As 29.4~45.8% of cases with MLL-t AML harbor N-/K-RAS or PTPN11 mutations, this finding may partly explain why HOXA11 expression is less frequently reported in MLL-t AML (12,13).

Data obtained from BM transplantation assay using Hoxa11-knockdown or Hoxa11-overexpression MLL/AF10 leukemia cells revealed that the expression levels of Hoxa11 in leukemia cells was associated with the survival of recipient mice. It was also demonstrated that Hoxa11 overexpression promoted MLL/AF10(OM-LZ) leukemia cells to engraft and proliferate in BM and spleen (Fig. 4E and F). A similar observation was reported by Sun et al (35), which showed that HOXA11 overexpression promoted cell proliferation and migration, and reduced cell apoptosis in breast cancer.

Based on the present study analyses of in vitro cytotoxicity and apoptosis rate of Hoxa11-knockdown and Hoxa11-overexpression MLL/AF10(OM-LZ) leukemia cells showed that Hoxa11 expression was associated with Ara-C sensitivity and apoptotic cell rate (Fig. 5A-D). In addition, gene expression analysis revealed that Hoxa11 induced apoptosis, at least partly, by regulating the expression of apoptosis-related genes, including Nfkbia, Rela and Trp53 (Fig. 5E), which is in line with previous findings by Guo et al (36), which established Ara-C-resistant human AML OCI-AML2 cell lines. Based on a comparative transcriptomic analysis, they identified HOXA11 as a DEG between resistant and parent cell lines, and further determined that HOXA11 promoted Ara-C sensitivity and apoptosis in the cell line (36). Moreover, a meta-analysis of HOXA11 expression using microarray data deposited by Heuser in the present study supported the findings of an association between HOXA11 expression and responsiveness of patients with AML treated with chemotherapy or on an AML induction/consolidation regimen (Fig. 5F). These data provided further support that HOXA11 expression in AML is predictive of an improved response to chemotherapy with Ara-C. The molecular mechanism of Hoxa11 in the regulation of apoptosis-related genes needs further characterization.

The role of HOXA11 varies according to cancer type. Epigenetic inactivation of HOXA11 is a poor prognostic marker and contributes to disease progression in ovarian cancer, non-small cell lung cancer, gastric cancer, urothelial bladder cancer, glioblastoma, renal cell carcinoma and breast cancer (37–43). In these solid tumors, HOXA11 acts as a functional tumor suppressor; however, the present results showed that upregulation of Hoxa11 accelerated leukemia development in MLL/AF10 cooperating RAS pathway mutations, suggesting an oncogenic role of Hoxa11 rather than one of tumor suppression. Further investigation of downstream Hoxa11 targets and biological pathways will provide an improved understanding of the mechanism underlying the different roles of Hoxa11 in AML and solid tumors.

Supplementary Material

Supporting Data

Acknowledgements

We would like to thank Mr. Jun-Wei Huang (Chang Gung Memorial Hospital, Taoyuan, Taiwan) and Mr. Chih-Shien Chuang (Chang Gung Memorial Hospital, Taoyuan, Taiwan) for their technical assistance in all the experiments.

Funding

This work was supported by grants from the Ministry of Science and Technology, Taiwan (grant no. 107-2320-B-182A-013) and Chang Gung Memorial Hospital, Taiwan (grant nos. CMRPG3E0301-3, CMRPG3E1391-3, CMRPG3G1821 and CMRPG3J1331).

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors' contributions

JFF was responsible for the conception and design of the present study, data analysis, funding acquisition and writing the original draft. LYS was responsible for providing the resources, analysis and interpretation of data, and wrote, reviewed and edited the manuscript. THY was responsible for data analysis, interpretation and discussion. JFF and LYS confirm the authenticity of all the raw data. All authors read and approved the final manuscript.

Ethics approval and consent to participate

All animal experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were carried out according to the protocol approved by the Animal Research Committee of Chang Gung Memorial Hospital (IACUC No. 2014092403; Taoyuan, Taiwan). Human sample collection was conducted in accordance with the Declaration of Helsinki and was approved by the Chang Gung Memorial Hospital Research Ethics Committee (IRB No. 96-1748B). Informed consent was obtained from all patients. For patients under the age of 18, consent/permission was obtained from the parents/guardians.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Glossary

Abbreviations

Abbreviations:

References