Human BCR/ABL1 induces chronic myeloid leukemia-like disease in zebrafish

Abstract

Chronic myeloid leukemia (CML) is induced by the BCR/ABL1 oncogene, which encodes a protein tyrosine kinase. We examined the effect of direct overexpression of the human p210^BCR/ABL1 oncoprotein in zebrafish. Humanized p210^BCR/ABL1 protein was detectable in Tg(hsp70: p210^BCR/ABL1) transgenic zebrafish embryos and adult kidney marrow. Transgenic zebrafish developed CML, which could be induced via cells transplanted into recipients. The expression of human BCR/ABL1 promoted myeloid lineages in Tg(hsp70:p210^BCR/ABL1) transgenic embryos. A total of 77 of 101 (76.24%) Tg(hsp70:p210^BCR/ABL1) adult transgenic zebrafish (age 6 months-1 year) developed CML. CML in zebrafish showed a triphasic phenotype, similar to that in humans, involving a chronic phase predominantly characterized by neutrophils in various degrees of maturation, an accelerated phase with an increase in blasts and immature myeloid elements, and a blast phase with >90% blasts in both the peripheral blood and kidney marrow. Tyrosine kinase inhibitors, as the standard drug treatment for human CML, effectively reduced the expanded myeloid population in Tg(hsp70:p210^BCR/ABL1) transgenic embryos. Moreover, we screened a library of 171 compounds and identified ten new drugs against BCR/ABL1 kinase-dependent or -independent pathways that could also reduce lcp1⁺ myeloid cell numbers in Tg(hsp70:p210^BCR/ABL1) transgenic embryos. In summary, we generated the first humanized zebrafish CML model that recapitulates many characteristics of human CML. This novel in vivo model will help to elucidate the mechanisms of CML disease progression and allow high-throughput drug screening of possible treatments for this disease.

Introduction

Chronic myeloid leukemia (CML) is a malignant bone marrow proliferative tumor originating from hematopoietic stem cells (HSC), with an annual incidence of 1-2/100,000 and accounting for 15-20% of all adult leukemias.1 CML is characterized by uncontrolled proliferation of myeloid cells and their progenitors in the peripheral blood (PB) and bone marrow (BM).2 The development of CML progresses from a chronic phase (CP) to an accelerated phase (AP), and finally to a blast phase (BP). Most patients in the CML-CP are clinically asymptomatic, but are diagnosed with leukocytosis characterized by mature granulocytes in the PB and BM. Disease progression to AP and BP is accompanied by a severe reduction in cellular differentiation, with immature blasts displacing mature cells.3 The final transformation phase can result in both lymphoblastic (25%) and myeloblastic (50%) subtypes, with a further 25% manifesting biphenotypic or undifferentiated phenotypes.4

The presence of the Philadelphia chromosome (Ph) is an important diagnostic indicator for CML.5 It is generated by a reciprocal translocation between chromosomes 9 and 22, referred to as t(9;22)(q34;q11).6 This translocation results in the BCR/ABL1 fusion gene, which is translated to the p210^BCR/ABL1 oncoprotein in almost all patients with CML.87 This fusion protein is a constitutively active tyrosine kinase that persistently activates various signaling pathways regulating cell proliferation, transformation, and survival, thereby promoting leukemogenesis.9 Further research and exploration are needed to recognize the blast crisis of CML since the specific mechanism leading to it is not yet fully understood.

The therapeutic use of tyrosine kinase inhibitors (TKI), such as imatinib, dasatinib, and bosutinib, has transformed the management of CML, largely turning a lethal disorder into a chronic condition. However, conventional TKI therapy for CML still presents challenges, including the appearance of TKI-resistant BCR/ABL1 mutants10 and the relative resistance of CML leukemia stem cells (LSC)11 to TKI. In addition, all TKI have a similar spectrum of toxic effects4 that can negatively affect the patient’s quality of life. Furthermore, CML and other malignancies include a population of cancer stem cells (CSC) that is able to regenerate or self-renew, resulting in therapeutic resistance and disease progression, and the inability to eradicate these CSC remains a significant obstacle to curing these diseases.

Biomedical research requires suitable animal disease models in which to study the mechanisms responsible for the cellular and molecular pathologies, and for testing certain therapeutic methods. There are high levels of conservation in terms of genomics, histoembryology, physiology, cardiac electrophysiology, and drug metabolic pathways between zebrafish and humans,12 and zebrafish thus represent a possible model for studying hematopoietic development and for high-throughput drug screening. However, there is currently no zebrafish CML model. The construction of a zebrafish CML model would expand our ability to study this disease and to develop new drugs that could benefit CML patients.

Methods

Zebrafish husbandry

All experiments involving zebrafish were carried out in accordance with the guidelines set by the Institutional Animal Care and Use Committee of Southern Medical University, Guangzhou, China. Zebrafish were raised, bred, and staged according to standard protocols.1413 The following strains were used: AB (wild-type strain, WT) and Tg(lyz:DsRed).15

Generation of the pToL hsp70:p210BCR/ABL1 construct and of Tg(hsp70:p210BCR/ABL1) transgenic zebrafish

The transgenic construct consisted of the zebrafish heat shock protein (Hsp) 70 promoter, human BCR/ABL1 (hBCR/ABL1) (b3a2) cDNA, Tol2 elements, and the SV40 polyA sequence. We cloned hsp70 promoter elements by polymerase chain reaction (PCR) using hsp70-specific primers 5′-GTATCGATTCAGGGGTGTCGCTTGGT-3′ and 5′-CCGATATCACCGGTCTGCAGGAAAAAAAAAC-3′. The hBCR/ABL1 (b3a2) cDNA fragment was isolated from the plasmid NGFR P21016 (Addgene) after digestion with EcoRI. The hsp70 promoter sequence was then placed upstream of the hBCR/ABL1 (b3a2) cDNA and subcloned into the pToL vector with minimal Tol2 elements and an SV40 polyA sequence to form the pToL hsp70:p210^BCR/ABL1 construct. The transgenic line was generated by injecting 50 pg of the pToL hsp70:p210^BCR/ABL1 construct together with Tol2 transposase mRNA into zebrafish embryos at the one-cell stage. Founders were identified by PCR confirmation of the transgene.

Western blot

Protein was extracted from whole embryos at 6 days post-fertilization (dpf) or from blood cells from the kidney marrow (KM) of 1-year-old Tg(hsp70:p210^BCR/ABL1) and age-matched WT controls. Proteins were quantified, and assessed by western blot analysis. Protein lysates were probed with rabbit anti-c-Abl antibody (1:1000 dilution, Cell Signaling Technology). Mouse anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) antibody (1:5000 dilution, Cell Signaling Technology) was included as an internal control.

Cell proliferation assay

Wild-type (WT) and Tg(hsp70:p210^BCR/ABL1 embryos at 3 dpf and 1-year old adults were incubated in 10 mM BrdU (Sigma-Aldrich) for 2 hours (h) and 4 h, respectively. The embryos and KM blood cells were stained with mouse-anti-BrdU antibody (Roche) and rabbit-anti-Lcp1 antibody (gift from Dr. Zilong Wen),17 followed by Alexa Fluor 555-anti-mouse antibody (Invitrogen) and Alexa Fluor 488-anti-rabbit antibody (Invitrogen) for fluorescent visualization.

Terminal deoxynucleotidyl transferase dUTP nick end labeling assay

Transferase dUTP nick end labeling (TUNEL) assay was carried out using an In Situ Cell Death Detection Kit (TMR red, Roche), followed by rabbit anti-Lcp1 antibody and Alexa Fluor 488-anti-rabbit antibody (Invitrogen) for fluorescent visualization.

Transplantation

Whole KM cell suspensions were prepared from Tg(lyz:DsRed) and Tg(hsp70:p210^BCR/ABL1-lyz:DsRed) (CML-like) fish. Three days after receiving a sublethal dose of radiation (25 Gy), 0.2 million cells were injected intracardially into irradiated WT recipients using a glass capillary needle (World Precision Instruments).

Drug treatment

Embryos were soaked in egg water containing 1‰ dimethylsulfoxide (DMSO) (Sigma-Aldrich), 20 μmol/L imatinib (Selleck), 5 μmol/L dasatinib (Selleck), 10 μmol/L bosutinib (Selleck), 20 μmol/L LY364947 (MedChemExpress), 2.5 μmol/L FTY720 (MedChemExpress), 0.5 μmol/L BEZ235 (MedChemExpress), or compounds from a compound library (TargetMol) for drug treatment.

Statistical analysis

Data were analyzed using SPSS software (version 20). Differences between two groups were analyzed using Student t-tests and differences among multiple groups by one-way analysis of variance (ANOVA) with Tukey’s adjustment. Significance was accepted when P<0.05. Data were expressed as mean±Standard Error of Mean (SEM).

Details of other methods used are available in the Online Supplementary Appendix.

Results

Transient expression of humanized BCR/ABL1 increased the number of myeloid cells in zebrafish larvae

The BCR/ABL1 fusion gene is present in nearly all cases of CML. Protein sequence comparisons revealed that zebrafish Bcr and Abl1 shared around 71% and 73% identities, respectively, with their human counterparts and contained a highly conserved kinase domain on Abl1 (Ensembl GRCh37 release 92). We evaluated the function of the hBCR/ABL1 oncoprotein in zebrafish by overexpressing hBCR/ABL1 mRNA encoding the p210^BCR/ABL1 oncoprotein. We then detected the numbers of myeloid cells during embryonic hematopoietic development by lcp1, lyz, mpx whole-mount in situ hybridization (WISH), and Sudan Black B (SB) cytochemical staining (Figure 1). lcp1, also named l-plastin, is a pan-myeloid marker that identifies all myeloid subsets, including macrophages and neutrophils. Numbers of lcp1⁺ cells were significantly increased after overexpression of hBCR/ABL1 compared with the control group. Expression levels of markers of more mature neutrophils, such as lyz, mpx, and SB, also increased significantly. Patients with CML typically develop a highly characteristic differential white blood cell (WBC) count with high concentrations of myelocytes and segmented neutrophils. The current results implied that hBCR/ABL1 expression in zebrafish may promote myelocytes and may be capable of inducing myeloid leukemia in vivo.

Figure 1.Expression levels of myeloid markers (lcp1, lyz, mpx and SB) increased after transient expression of humanized BCR/ABL1 (hBCR/ABL1) in zebrafish larvae. Whole mount in situ hybridization (WISH) of lcp1 (A and B), lyz (C and D), and mpx (E and F) expressions in wild-type (WT) zebrafish larvae after transient overexpression of hBCR/ABL1 mRNA (right) were higher than controls (left) at 3 dpf. (G and H) The number of SB⁺ cells in WT zebrafish larvae after transient overexpression of hBCR/ABL1 mRNA (right) was higher than controls (left) at 3 dpf. n/n: number of zebrafish larvae showing representative phenotype/total number of zebrafish larvae examined. Original magnification ×32 (A-H). Red rectangles in each panel indicate the signals in PBI region, and the regions were enlarged at the lower right (original magnification ×200). (I-L) Statistical analysis. lcp1⁺ signals (I), lyz⁺ signals (J), mpx⁺ signals (K), and SB⁺ signals (L) in PBI region after injection were calculated and compared at 3 dpf (Student t-tests, mean±Standard Error of Mean, **P<0.01; ****P<0.0001).

hBCR/ABL1 was inherited in Tg(hsp70:p210BCR/ABL1) transgenic zebrafish

Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish were created using a construct (Figure 2A) expressing hBCR/ABL1 under the control of the zebrafish heat shock-inducible hsp70 promoter.1918 The construct was designed to integrate the complete coding sequence into the host genome using the Tol2 transposition system, allowing the generation of multiple lines of transgenic zebrafish. Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish founders were confirmed by PCR (Figure 2B). Stable F1 Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish were obtained by intercrossing founder fish and were confirmed by sequencing (data not shown). F2 and the offsprings were obtained by mating F1 fish with WT fish. The temporospatial expression of hBCR/ABL1 in Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish was evaluated by WISH (Figure 2C). hBCR/ABL1 expression was apparent throughout the body of Tg(hsp70:p210^BCR/ABL1) embryos at 3 dpf after heat shock treatment. Further detection by real-time-quantitative (RT-q)-PCR showed that levels of hBCR/ABL1 mRNA were significantly elevated after heat shock treatment in both coro1a:GFP blood cells from Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish embryos and hematopoietic progenitors and myelocytes in KM blood cells from Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish adults (Online Supplementary Figure S1). The hBCR/ABL1 oncogene encoding the p210^BCR/ABL1 protein was also detected in Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish embryos and adult kidneys (Figure 2D). The molecular weight of p210^BCR/ABL1 measured in vitro (Online Supplementary Figure S2) confirmed that the weight of the fusion protein generated was as expected. The p210^BCR/ABL1 protein was highly expressed in Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish after heat-shock treatment.

Figure 2.Expression of hBCR/ABL1 in Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish. (A) Structure of pToL hsp70:p210^BCR/ABL1. (B) Detection of the hBCR/ABL1 cDNA sequence integrated into the wild-type (WT) zebrafish genome using the specific polymerase chain reaction (PCR) amplifying a 466 bp fragment within the hBCR/ABL1 fusion section. Tg: Tg(hsp70:p210^BCR/ABL1) founder individuals; PC: pToL hsp70:p210^BCR/ABL1 plasmid as the positive control; NC1: ddH₂O as the negative control; NC2: genomic DNA of WT fish as the negative control. (C) Whole mount in situ hybridization (WISH) of hBCR/ABL1 mRNA temporospatial expression in 3 dpf WT and Tg(hsp70:p210^BCR/ABL1) zebrafish embryos without (a-f) or with (g-l) heat-shock treatment. Original magnification ×32 (a, d, g and j). Signals in head (b, e, h, k) and tail (c, f, i, l) region under 100× and 200× magnification, respectively. n/n: number of zebrafish embryos showing representative phenotype/total number of zebrafish embryos examined. (D) Expression of p210^BCR/ABL1 expressed in embryos (n=200) and kidney marrow (KM) blood cells (approx. 2×10⁶ cells) of Tg(hsp70:p210^BCR/ABL1) adult zebrafish assessed by western blot. GAPDH as the loading control. Tg: Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish; No-HS: TTg(hsp70:p210^BCR/ABL1) transgenic zebrafish without heat-shock treatment; HS: Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish with heat-shock treatment. (E) WISH of lcp1 (a-b), lyz (c-d), and mfap4 (g-h) expressions in HS Tg(hsp70:p210^BCR/ABL1) (right) were higher than WT controls (left) at 3 dpf. (e-f) The number of SB positive cells in HS Tg(hsp70:p210^BCR/ABL1) (right) was higher than WT controls (left) at 3 dpf. n/n: number of zebrafish larvae showing representative phenotype/total number of zebrafish larvae examined. Original magnification ×32 (a-h). Red rectangles in each panel indicate the signals in PBI region and the regions were enlarged at the lower right (original magnification ×200). Black rectangles in the panel indicate the signals in the brain region (vertical view, original magnification ×50). (i-l) Statistical analysis. lcp1⁺ signals (i), lyz⁺ signals (j), mfap4⁺ signals (l), and SB⁺ signals (k) in PBI region were calculated and compared at 3 dpf. Student t-tests, mean±Standard Error of Mean; *P<0.05; ****P<0.0001.

Inducible hBCR/ABL1 expression in Tg(hsp70:p210BCR/ABL1) transgenic zebrafish promoted myeloid lineage in zebrafish embryos

Expression of p210^BCR/ABL1 induces leukemia and myeloproliferative disorders, indicating a direct, causal role of BCR/ABL in CML.23209 We established Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish with expression of p210^BCR/ABL1 and stable inheritance. To further explore the function of p210^BCR/ABL1 in zebrafish, we observed its influence on hematopoietic development in zebrafish embryos using WISH and cytochemical staining with lineage-specific markers (Figure 2E). The numbers of lcp1⁺ pan-myeloid cells, lyz⁺ neutrophils, SB neutrophils, and mfap4⁺ macrophages were significantly increased in Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish larvae at 3 dpf compared with WT controls. This suggested that hBCR/ABL1 expressed in zebrafish could either promote the production of HSC or their differentiation into each hematopoietic lineage. To distinguish between these possibilities, we detected the HSC marker (cmyb), erythrocyte marker (βe1), and lymphocyte marker (rag1). There was no difference in the number of cmyb⁺ HSCs between Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish and WT controls at 36 hpf, but the number was significantly increased in transgenic zebrafish at 60 hpf (Online Supplementary Figure S3A-D, I and J). Numbers of βe1⁺ erythrocytes and rag1⁺ lymphocytes were significantly decreased in the transgenic zebrafish compared with the WT control zebrafish at 5 dpf (Online Supplementary Figure S3E-H). These findings suggest that hBCR/ABL1 may promote myeloid differentiation.

Inducible hBCR/ABL1 expression in Tg(hsp70:p210BCR/ABL1) adult zebrafish created phenotype resembling human CML

The natural progression of untreated CML is bi- or triphasic, with the initial CP followed by AP, BP, or both. CP is characterized by leukocytosis in both the PB and BM, and a preponderance of granulocytes in various degrees of maturation. However, blasts account for <2% of the peripheral WBC and <5% of the nucleated cells in the BM.24 As the disease progresses, patients enter the AP followed by the BP, during which there is hematopoietic differentiation arrest, allowing immature blasts to accumulate in the BM and spill into the circulation. A level of 10-19% of blasts in the PB or BM marks the transition from CP to AP, along with a predominance of promyelocytes. A level of at least 20% PB or BM blasts indicates the progression to the BP.24 To explore the possibility of developing leukemia-like hematologic disorders in Tg(hsp70:p210^BCR/ABL1) adult fish, PB and KM cells were collected from Tg(hsp70:p210^BCR/ABL1) and WT fish at 6 months to 1-year old and subjected to cytological and WBC analyses (Table 1 and Figure 3A and B). Seventy-seven of 101 (76.24%) Tg(hsp70:p210^BCR/ABL1) adult zebrafish developed CML-like disease, including 68 with a CML-CP phenotype, marked by massive leukocytosis in the PB or KM, including increased percentages of myelocytes and myeloid precursors. In the early stage of CML-CP, the increased leukocytes were primarily neutrophils in various degrees of maturation. Myelocytes accounted for >15% in the PB or >50% in the KM, with blasts usually accounting for <2% of the PB and <5% of the KM during this phase. We referred to this period as CML-CP1 (Table 2). Differentiation was then interrupted in the late stage of CML-CP as the condition progressed towards CML-AP. The increased leukocytes were primarily myeloid precursors and blasts, with myeloid precursors >10% and blasts >2% in the PB, and myeloid precursors >15% and blasts >5% in the KM. We referred to this period as CML-CP2 (Table 2). Eight of the 77 CML-like Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish showed CML-AP phenotype including significant 2- to 10-fold increases in the percentages of blasts and myeloid precursors, with blasts >10% in the PB or KM. Amongst the 77 CML-like Tg(hsp70:p210^BCR/ABL1) adult zebrafish, one progressed to CML-BP with >90% blasts expanding in both the PB and KM. We also identified some phenotypes accompanying these CML-like Tg(hsp70:p210^BCR/ABL1) adult zebrafish, including eosinophilia, lymphocytosis and thrombocytosis (Figure 3C). Six of 77 (7.79%) CML-like Tg(hsp70:p210^BCR/ABL1) adult zebrafish presented with eosinophilia, with eosinophils accounting for >0.1% of the PB cell count compared with approximately 0.01±0.00% in WT adult zebrafish (n=55) (Online Supplementary Table S1). This was similar to the “Ph-positive eosinophilic/basophilic CML” described by Goh et al.24 Large numbers of lymphocytes accumulated in the PB in 8 of 77 (10.39%) CML-like Tg(hsp70:p210^BCR/ABL1) adult zebrafish, accounting for >5% of the PB cell count com-pared with around 1.80±0.23% in WT adult zebrafish (n=55) (Online Supplementary Table S1), similar to the lymphocytosis25 observed in CML patients. Thrombocytosis26 is present in approximately half of all newly diagnosed CML patients. Thirteen of the 77 (16.88%) CML-like Tg(hsp70:p210^BCR/ABL1) adult zebrafish presented with thrombocytosis, with platelets accounting for >0.5% of the PB cell count compared with around 0.11±0.04% in WT adult zebrafish (n=55) (Online Supplementary Table S1). Histological examination of the spleen in CML-like Tg(hsp70:p210^BCR/ABL1) demonstrated expansion of the splenic red pulp, predominantly by granulocytic myeloid cells (Figure 3D). In addition, the morbidity of CML-like disease in heat-shock-treated Tg(hsp70:p210^BCR/ABL1) adult zebrafish was higher than in non-induced Tg(hsp70:p210^BCR/ABL1) adult zebrafish. The ratios of individuals in CML-CP2 and CML-AP were increased among heat-shock-treated Tg(hsp70:p210^BCR/ABL1) compared with untreated Tg(hsp70:p210^BCR/ABL1) adult zebrafish (Figure 3E). This result suggests that overexpression of BCR/ABL1 is an important factor in accelerating the course of CML.

Table 1.Hemogram and classification of Tg(hsp70:p210BCR/ABL1) fish at 6-12 months

Table 2.Hemogram and classification of Tg(hsp70:p210^BCR/ABL1) fish in chronic myeloid leukemia chronic phase.

Figure 3.Tg(hsp70:p210^BCR/ABL1) adult fish display abnormal myeloid cell expansion resembled chronic myeloid leukemia (CML)-like phenotypes. (A) May-Grunwald-Giemsa staining of peripheral blood (PB) cells (top) and kidney marrow (KM) blood cells (bottom) were obtained from wild-type (WT) (a, e) (n=55) and Tg(hsp70:p210^BCR/ABL1) (b-d, f-h) (n=101) adult zebrafish. Blue arrows indicate accumulated promyelocytes in CML-chronic phase (CP)-like fish. Yellow arrows indicate neutrophils in CML-CP-like fish. Red arrows indicate myeloid precursors in CML accelerated phase (AP)-like fish. Black arrows indicate blasts in both CML-AP-like fish and CML-BP-like fish. The percentages indicate the ratio that number of fish with leukemia-like phenotype in total number of fish. Original magnification ×400. (B) Statistical analysis. Blood cell counts of PB and KM were calculated manually based on their morphology. ANOVA; mean±Standard Error of Mean; *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001. (C) Large numbers of eosinophils (b, yellow arrow), lymphocytes (c, blue arrow), and platelets (d, green arrow) accumulated in CML-like Tg(hsp70:p210^BCR/ABL1) adult fish PB compared with WT controls (a, n=55). The percentages indicate the ratio that number of fish with abnormal cells accumulated in total number of CML-like Tg(hsp70:p210^BCR/ABL1) adult fish. Red arrows indicate neutrophils. Original magnification ×400. Stains were May-Grunwald-Giemsa. (D) Histology of the spleen from CML-like Tg(hsp70:p210^BCR/ABL1) adult fish showing disruption of the spleen architecture and massive invasion by hematopoietic cells compared with WT controls. Original magnifications ×200 for each panel. Inset (red rectangles): granulocytic myeloid cells at higher magnification ×1000. Stains were Hematoxylin & Eosin. (E) The morbidity rate of CML was higher in 1 Y-HS Tg(hsp70:p210^BCR/ABL1) adult fish (n=35) compared with 1 Y-No HS Tg(hsp70:p210^BCR/ABL1) adult fish (n=18), 97.14% and 44.44%, respectively. 1Y-No HS: 1-year-old Tg(hsp70:p210^BCR/ABL1) adult fish without heat shock treatment. 1Y-HS: 1-year-old Tg(hsp70:p210^BCR/ABL1) adult fish with heat shock treatment. The morbidity rates of Tg(hsp70:p210^BCR/ABL1) transgenic fish in the HS group developing into CML-CP 1 phase was 17.14%, while that in the No-HS treatment group was 33.33%. The morbidity rate of Tg(hsp70:p210^BCR/ABL1) transgenic fish in the heat shock treatment group developing into CML-CP 2 phase and CML-AP phase were 42.86% and 37.14%, compared with 0.00% and 11.11%, respectively, in the No-HS Tg(hsp70:p210^BCR/ABL1) adult fish.

Tg(hsp70:p210BCR/ABL1) transgenic fish displayed abnormal myeloid cell expansion resulting from increased proliferation and inhibition of apoptosis

The above results indicated that myeloid cells accumulated in Tg(hsp70:p210^BCR/ABL1) fish from the embryonic stage to the adult, which could be caused by accelerated proliferation or reduced apoptosis. To clarify the cellular mechanisms responsible for myeloid cell expansion in Tg(hsp70:p210^BCR/ABL1) fish, we monitored myeloid cell proliferation and death by BrdU incorporation and TUNEL assay, respectively. BrdU incorporation was significantly increased in Tg(hsp70:p210^BCR/ABL1) larvae and adult KM compared with WT controls, indicating that myeloid cell expansion in Tg(hsp70:p210^BCR/ABL1) fish was the result of increased proliferation (Figure 4A, B, E and F). However, myeloid cell apoptosis was also significantly decreased in Tg(hsp70:p210^BCR/ABL1) larvae and adult KM compared with WT controls, suggesting that the expansion of myeloid cells Tg(hsp70:p210^BCR/ABL1) was also caused by reduced apoptosis (Figure 4C, D, G and H).

Figure 4.The abnormal myeloid cell expansion in Tg(hsp70:p210^BCR/ABL1) fish caused by proliferation and apoptosis perturbation. Lcp1 and BrdU antibody immunofluorescence double staining indicates a significant increase in the myeloid cells proliferation in the PBI region of 3 dpf HS Tg(hsp70:p210^BCR/ABL1) larvae (n=28) compared with the wild-type (WT) controls (n=18) (A and B) and kidney marrow (KM) Lcp1⁺ cells in 1-year Tg(hsp70:p210^BCR/ABL1) adults (n=3) compared with the WT controls (n=3) (E and F). Lcp1 antibody and transferase dUTP nick end labeling (TUNEL) assay co-staining were used to detect the apoptosis. The number of Lcp1⁺ cells in the PBI region (C and D)/KM (G and H) in 3 dpf/1-year HS Tg(hsp70:p210^BCR/ABL1) (n=27/n=4) significantly decreased compared with WT (n=19/n=5). Arrows indicate Lcp1BrdU/Lcp1TUNEL double-positive cells. Original magnification ×100 (A, C, E, G). Percentage of the PBI region and KM localized Lcp1⁺ myeloid cells that incorporate BrdU/TUNEL in Lcp1+ myeloid cells in 3 dpf or 1-year HS WT and Tg(hsp70:p210^BCR/ABL1) fish. Student t-test; mean±Standard Error of Mean; *P<0.05; ****P<0.0001.

Tg(hsp70:p210BCR/ABL1) transgenic cells with induced CML-like disease were transplantable

To determine the aggressiveness of the leukemia induced by Tg(hsp70:p210^BCR/ABL1) activity, whole KM blood cells from Tg(hsp70:p210^BCR/ABL1) fish were trans-planted into γ-irradiated WT adult hosts and the resulting fish were tested to determine if the CML-like phenotype developed in the Tg(hsp70:p210^BCR/ABL1) fish could be transplanted into the WT fish. We used 1-year old Tg(hsp70:p210^BCR/ABL1-lyz:DsRed) CML-like donors and Tg(lyz:DsRed) control donors, in which the granulocytes were marked by red fluorescence. Each irradiated WT fish received 0.2 million KM blood cells from donors and were then raised under normal conditions. All four surviving recipients of Tg(hsp70:p210^BCR/ABL1-lyz:DsRed) CML-like donor cells developed CML-like disease within 2-3 weeks of transplantation with whole KM blood cells, indicated by infiltration of DsRed granulocytes into the periphery (Figure 5A) and the robust expansion of myeloid cells in both the PB and KM (Figure 5B). In contrast, no control fish showed signs of a CML-like phenotype. We collected leukemia cells from the PB and KM and showed that these inflated cells were BCR/ABL1⁺ by PCR (Figure 5C). We concluded that the myeloid cells that accumulated in Tg(hsp70:p210^BCR/ABL1) fish could proliferate autonomously and could cause CML-like disease in a WT host.

Figure 5.Wild-type (WT) fish transplanted with chronic myeloid leukemia (CML)-like leukemic cells show myeloid expansion in the early stage. Of 43 and 65 recipients transplanted with Tg(lyz:DsRed) control and CML-like Tg(hsp70:p210^BCR/ABL1-lyz:DsRed) kidney marrow (KM) cells, three and four survived, respectively. (A) DsRed positive cells repopulated in recipients within 2-3 weeks after transplantation of Tg(lyz:DsRed) control (left) and CML-like Tg(hsp70:p210^BCR/ABL1-lyz:DsRed) (right) KM blood cells. (B) All survived recipients were stained by May-Grunwald-Giemsa. Myelocytes (d, blue arrows) increased in KM of WT fish after transplanted with CML-like Tg(hsp70:p210^BCR/ABL1-lyz:DsRed) KM cells. Myeloid precursors (b and d, red arrows) increased in both peripheral blood (PB) and KM of WT fish after being transplanted with CML-like Tg(hsp70:p210^BCR/ABL1-lyz:DsRed) KM blood cells. Original magnification ×400. (C) hBCR/ABL1⁺ cDNA fragments detected in PB and KM cells of WT fish after being transplanted with Tg(lyz:DsRed) control (WT-PB/WT-KM) and CML-like Tg(hsp70:p210^BCR/ABL1-lyz:DsRed) KM cells (Tg-PB/Tg-KM) assessed by poly-merase chain reaction. PC: pToL hsp70:p210^BCR/ABL1 as the positive control. NC: ddH₂O as the negative control.

Tg(hsp70:p210BCR/ABL1) transgenic leukemic model responded to chemotherapeutic drug treatment

Recent studies demonstrated that zebrafish shares 82% of disease-associated targets and numerous drug metabolism pathways with humans.12 To determine if the pharmacological mechanism in Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish was also conserved compared with CML patients, we treated the WT and Tg(hsp70:p210^BCR/ABL1) embryos with the widely used anti-CML drugs, imatinib, dasatinib, and bosutinib, to the maximum teratogenic doses, with DMSO as a placebo (Online Supplementary Figure S4). After incubation with these TKI for 48 h, we calculated the numbers of lcp1⁺ myeloid cells in WT and Tg(hsp70:p210^BCR/ABL1) larvae in the posterior blood island (PBI) region at 5 dpf (Figure 6A and B). All the TKI significantly reduced the number of lcp1⁺ myeloid cells in Tg(hsp70:p210^BCR/ABL1) larvae compared with the DMSO control group. In addition, lower concentrations (20 and 40 μmol/L) of imatinib significantly reduced the expanded lcp1⁺ myeloid population in Tg(hsp70:p210^BCR/ABL1) larvae, but the number of lcp1⁺ myeloid cells was also significantly reduced in WT larvae at higher concentrations (80 μmol/L) compared with DMSO (Online Supplementary Figure S5). These results suggest that high doses of imatinib may affect normal myelopoiesis, which may be associated with more adverse events or unpredictable off-target effects. Further studies are needed to clarify these effects and to support the clinical treatment of patients with CML.

Figure 6.Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish model responds to chemotherapy. (A and B) TKI treatment in zebrafish larvae. 3 dpf HS WT (left panels) and Tg(hsp70:p210^BCR/ABL1) (right panels) larvae treated for 48 hours with 1‰ DMSO control, 20 μmol/L imatinib, 5 μmol/L dasatinib and 10 μmol/L bosutinib. (A) WISH of lcp1 expression in the drug treated larvae at 5 dpf. n/n, number of zebrafish larvae showing representative phenotype/total number of zebrafish larvae examined. Original magnification, ×200. (Bottom): statistical analysis. Enumeration of lcp1⁺ signals shown in (A). Average numbers of lcp1⁺ cells per larva with drug treatment. ANOVA; mean ± SEM; ****P<0.0001. (C) Drug screening. Upper: WISH of lcp1 expression in Tg(hsp70:p210^BCR/ABL1) (right) larvae and WT controls (left) at 4 dpf after treatment (1‰ DMSO, 20 μmol/L CC-223, 2.5 μmol/L FTY720) for 24 hours. n/n, number of zebrafish larvae showing representative phenotype/total number of zebrafish larvae examined. Original magnification, ×200. Under: Statistical analysis. 3 dpf HS WT and Tg(hsp70:p210^BCR/ABL1) larvae (n=25 and n=20, n=23 and n=25, n=20 and n=15, n=22 and n=24, n=20 and n=21, n=14 and n=24, n=19 and n=19, n=20 and n=19, n=19 and n=18, n=23 and n=24, n=23 and n=17, n=16 and n=15, respectively) were treated for 24 hours with drugs (1‰ DMSO, 2.5 μmol/L FTY720, 20 μmol/L CC-223, 0.5 μmol/L BEZ235, 40 μmol/L icaritin, 20 μmol/L AZD3759, 80 μmol/L icotinib, 40 μmol/L DB07268, 80 μmol/L NQDI-1, 40 μmol/L selonsertib, 20 μmol/L LY364947, 1 μmol/L ciliobrevin A, respectively). Average numbers of lcp1⁺ cells per larva with drug treatment. ANOVA; mean ± SEM; **P<0.01; ***P<0.001; ****P< 0.0001. (D) Sketch map of inhibitors target BCR/ABL1 kinase-dependent or kinase-independent pathways. JNK: the c-Jun N-terminal kinase. ASK: apoptosis signal-regulating kinase. TGF-β, the transforming growth factor β. EGFR: the epidermal growth factor receptor.

We screened a library of 171 compounds in 3 dpf WT and Tg(hsp70:p210^BCR/ABL1) embryos to examine their ability to reverse the disease phenotype. We reduced the incubation time to 24 h to speed up the screening process, and then calculated the numbers of lcp1⁺ myeloid cells in WT and Tg(hsp70:p210^BCR/ABL1) larvae in the PBI region at 4 dpf. Ten inhibitors, including the natural compound, icaritin, as well as CC-223, BEZ235, AZD3759, icotinib, DB07268, NQDI-1, selonsertib (GS-4997), LY364947 and ciliobrevin A (HPI-4) effectively reduced the expanded lcp1⁺ myeloid population in Tg(hsp70:p210^BCR/ABL1) embryos compared with DMSO-treated controls (Figure 6C).

Discussion

We constructed a new germline of transgenic zebrafish expressing the hBCR/ABL1 fusion protein. Expression of hBCR/ABL1 in Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish altered hematopoiesis by up-regulating myeloid genes, as detected in larvae at 3 dpf. Adult Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish developed CML characterized by clonal myelocytic blasts, representing the first zebrafish model of hBCR/ABL1-induced CML. As the disease progressed, hematopoietic differentiation was interrupted, and immature blasts and myeloid precursors accumulated in the BM and spilled into the circulation in this zebrafish model, closely resembling the natural course of human CML progression without treatment. The most accurate CML animal model to date is the SCLtTA/BCR-ABL mouse line21 established by Koschmieder et al. in 2005. However, these mice only survive for 4-17 weeks, while adult Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish could survive for from 12 to >18 months, with or without heat shock, which was longer than all previous mouse models. The incidence of CML in the Tg(hsp70:p210^BCR/ABL1) transgenic model was increased by hBCR/ABL1 heat shock. This Tg(hsp70:p210^BCR/ABL1) transgenic model may thus provide insights into the mechanism that drives the transition from CML-CP to CML-AP or CML-BP.

Tyrosine kinase inhibitors (imatinib, dasatinib, and bosutinib) effectively reduced the expanded myeloid population in Tg(hsp70:p210^BCR/ABL1) embryos, suggesting that the pharmacological pathways in this model were similar to those in human CML. The discovery of imatinib has greatly improved the longevity and quality of life of patients with CML; however, some patients develop resistance to TKI and may even progress toward CML-AP or CML-BP of the disease despite TKI therapy. Second- and third-generation TKI were developed to treat patients in whom imatinib fails, with up to 40-87% of patients achieving durable complete cytogenetic remission.4 However, more serious side-effects have recently been associated with these second- and third-generation TKI. Understanding the underlying cause of resistance and screening for novel targeted drugs with low toxicity and high efficiency thus remain important steps in combating CML. Further studies are planned to generate site-directed mutations of the ABL1 kinase domain in Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish using gene-editing technology (such as CRISPR/Cas9). The ABL1 kinase domain is frequently mutated in clinical cases, and examination of these mutants may thus help to elucidate the mechanism responsible for TKI resistance.

In the present study, we screened a compound library and discovered 10 new targeted drugs that reduced the expanded myeloid population in Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish embryos. Icaritin is a natural flavonoid derived from the traditional Chinese medicine Epimedium. Icaritin was previously shown to inhibit the growth of leukemic cell lines, including imatinib-resistant BCR/ABL1⁺ blast cells and BCR/ABL1-T315I mutant cells via mechanisms involved in MAPK and JAK/STAT signaling.2927 The current results show that icaritin could reduce the expanded lcp1⁺ myeloid population in Tg(hsp70:p210^BCR/ABL1) embryos, consistent with these previous findings. Overactivation of PI3K/AKT/mTOR is known to play a pivotal role in many human cancers, thus providing strong support for the therapeutic anti-cancer application of PI3K/Akt/mTOR inhibitors.3130 Sadovnik et al. found that escape of CML LSC was disrupted by the addition of PI3K/mTOR blockers.32 Furthermore, the PI3K/mTOR dual inhibitor BEZ235 had beneficial effects on a variety of tumors in vivo and in vitro, including lymphoid malignancies33 and myeloid malignancies.3534 Bendell also identified the active-site inhibitor CC-223, which targets both mTORC1 and mTORC2 through mTOR kinase activity to inhibit activation of AKT and 4EBP1, as a promising therapeutic agent with activity against many non-Hodgkin lymphoma and solid tumor cell lines.36 In the current study, the Tg(hsp70:p210^BCR/ABL1) embryonic zebrafish model responded well to both BEZ235 and CC-223. Overall, these results suggest that targeting the PI3K/Akt/mTOR signaling pathway may be an effective strategy for overcoming CML therapy resistance. Unexpectedly, however, Tg(hsp70:p210^BCR/ABL1) larvae did not respond well to the sphingosine 1-phosphate antagonist, FTY720, and the number of lcp1⁺ myeloid cells in WT zebrafish conversely increased after treatment with FTY720. Previous studies reported that the FTY720-mediated PP2A reactivation could markedly reduce the survival and self-renewal of CML-quiescent HSC through BCR-ABL1 kinase-independent and PP2A-mediated inhibition of JAK2 and β-catenin.37 We therefore hypothesized that, although sphingosine 1-phosphate may play a role in hematopoietic regulation, further studies are needed to determine its precise mechanism. LY364947,3938 ciliobrevin A,40 DB07268,41 selonsertib,42 NQDI-1,43 AZD3759,44 and icotinib45 have recently been shown to target key factors and signaling pathways essential for the survival of CML LSC and other CSC, including transforming growth factor β,39 Hedgehog,46 c-Jun N-terminal kinases,47 apoptosis signal-regulating kinase 1,48 and epidermal growth factor receptor.49 The Tg(hsp70:p210^BCR/ABL1) transgenic zebrafish embryos in the current study also responded well to these compounds. Our findings, therefore, suggest that inhibition of BCR/ABL1 kinase-dependent or kinase-independent pathways (Figure 6D) might offer potential for overcoming resistance to TKI and thus eradicate LSC, thereby paving the way for the development of novel, more effective LSC-eradicating treatment strategies for CML.

In conclusion, the Tg(hsp70:p210^BCR/ABL1) transgenic model represents a phenotype-based, cost-effective, in vivo model of CML suitable for high-throughput chemical screening. This model may improve our understanding of the protein functions, pharmacological mechanisms, and toxicology of novel targeted drugs, thus improving the cure rate for patients with CML.

Acknowledgments

The authors would like to thank Dr. Nathan Lawson and Dr. Koichi Kawakami for providing pTol vector. The authors would like to thank Dr. Zilong Wen for providing Lcp1 antibody. The authors would like to thank Dr. Kuangyu Yen, Dr. Yali Chi, Dr. Shan Xiao, Xiaohui Chen, and Yi Zheng for their helpful suggestion. The authors would like to thank Christine Walsh, Hank Duan and the editor of International Science Editing for correcting the grammar and spelling. This work was supported by the Young Teacher National Natural Science Foundation of China (Grant No. 81700150), the Natural Science Foundation of Guangdong Province, China (Grant No. 2014A030312002), the Department of Science and Technology of Guangdong Province, China (Grant No. 2015B050501006) and Shunde Economy, Science and Technology Bureau, China (Grant No. 2015CXTD06).

Footnotes

• Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/105/3/674

• Received January 4, 2019.
• Accepted July 5, 2019.