Establishment of a Novel CD34+ Hematopoietic Stem Cell-derived Xenograft Model of Hyperleukocytic Acute Myeloid Leukemia

Background Hyperleukocytic acute myeloid leukemia (HLL) with high early mortality is dicult to therapy. The related research on HLL is still in its infancy, and the establishment of a patient derived xenograft (PDX) model of HLL has not been fully reported, especially the CD34 + hematopoietic stem cell-derived xenograft model of HLL. + hematopoietic stem cell-derived xenograft model. In this study, we use the NOD.CB17-Prkdc scid Il2rg tm1 /Bcgen (B-NSG) mice with severe immunodeciency for HLL model establishment and evaluation. gene forward (CAGGCTGCAATAAGAGATATTTTAAGCT) and reverse (GAAGTCACACTGGTATGGTTTCTCA), and (2) human reference gene for ABL gene forward (TGGAGATAACACTCTAAGCATAACTAAAGGT) and reverse (GATGTAGTTGCTTGGGACCCA). The positive rate for the WT1 gene in the mice was calculated.


Introduction
Hyperleukocytic leukemia, a high-risk type, is commonly de ned as peripheral blood leukocyte counts exceeding 50-100×10 9 /l [1]; an abnormal increase in white blood cells (WBCs) causes an increase in blood viscosity and leads to leukocyte stasis [2] and leukemic organ in ltration in hyperleukocytic leukemia. Patients with hyperleukocytic acute myeloid leukemia (HLL) usually have higher mortality with severe life-threatening complications [3], for example, severe respiratory failure [4], disseminated intravascular coagulopathy and tumor lysis syndrome and [1,5]. The early mortality of hyperleukocytic leukemia patients is notable, with a mortality rate as high as 40% within a week [6,7]. Moreover, Routine clinical treatment is implemented to decrease the number of WBCs with leukapheresis, which is an effective therapeutic method to rapidly reduce leukocytes counts with a more rapid reduction in the leukemic cellular burden [5].
To study the pathogenesis of HLL and explore effective treatment, it is urgent to establish a suitable model of HLL. Traditional mouse models of leukemia are constructed by transplanting cell lines; however, immortalized cell lines may change their comparative functional phenotype and cannot represent the real state of disease [8]. The patient derived xenograft (PDX) model mainly transplants primary cells from fresh tumor cells or tissues of patients to immunode cient mice, and is widely used to model of nonsmall cell lung cancer, bladder cancer, and triple-negative breast cancer [9][10][11] as well as acute myeloid leukemia (AML) [12]; the PDX model maintains the primary tumor microenvironment and original characteristics of patient in mice, making it more valuable in vivo than cell lines when constructing a leukemia model [13,14]. Therefore, the PDX model is a powerful research tool to explore the mechanism of tumor development and evaluate the potential therapeutic effects of disease. Primary patient-derived tumors require higher immunode ciency levels for xenograft implantation, such as nonobese diabetic (NOD).Cg-Prkdc scid Il2rg tm1Wjl (NSG) mice that lack T, B, functional NK cells and IL2Rγ provide advantages and are ideal hosts for in vivo engraftment studies regardless of their sensitivity to environmental status [10,15].
At present, related research on HLL is still in its infancy, and establishment of a PDX model of HLL has not been reported, especially the construction of a CD34 + hematopoietic stem cell-derived xenograft model. In this study, we use the NOD.CB17-Prkdc scid Il2rg tm1 /Bcgen (B-NSG) mice with severe immunode ciency for HLL model establishment and evaluation.

Methods
Patient-derived leukemia cell collection and culture Human cells were collected from newly diagnosed HLL patients enrolled from Zhongnan Hospital of Wuhan University, and the information for the collected clinical samples is provided in Table 1. Patient diagnosis with AML was based on standard morphological and cytochemical examinations of peripheral blood and marrow smears according to the French-American-British (FAB) and World Health Organization (WHO) criteria. The HLL cells were collected using a Fresenius COM.TEC machine at Zhongnan Hospital; the procedures were described in our previous reports [16]. Peripheral blood mononuclear cells (PBMCs) were separated with a Ficoll kit according to the manufacturer's instructions (TBD, lot: LDS1075, Tianjin, China). The separated cells were maintained in RPMI 1640 media containing 10% FBS (Gibco, Invitrogen, Carlsbad, CA, USA) with 100 U/ml penicillin and 100 µg/ml streptomycin at 37°C and 5% CO 2 . Bone marrow-derived cells were collected from HLL patients during bone marrow biopsy and separated using a red blood cell lysis buffer.
Isolation of CD34 + cells CD34 + cells were separated using the CD34 MicoBeads Kit (Miltenyi Biotec, Germany) according to the manufacturer's instructions. In brief, PBMCs from HLL patients (Patient #2) with WT1 and NRAS mutations were magnetically labeled with CD34 microbeads for 30 min at 4 ℃, and then centrifuged at 300 g for 10 min. The cell suspension was loaded onto a MACS® column with a MACS separator. The magnetically labeled CD34 + cell subpopulations were retained within the column, and the unlabeled cells were washed. After removing the column from the magnetic eld, the magnetically retained CD34 + cells were eluted as the positively selected cell fraction.

Animal experiments
Five-to six-week-old female B-NSG mice (NOD.CB17-Prkdc scid Il2rg tm1 /Bcgen) were obtained from Beijing Biocytogen Co., Ltd. (permission number: SCXK 2016-0004) and maintained in a pathogen-free animal facility in laminar air ow cabinets with a 12 h light/12 h dark schedule. The animals were fed an autoclaved rodent diet ad libitum. All procedures followed the institutional and national guidelines for the care and use of laboratory animals.
For HLL PDX model establishment with PBMCs of patients, patient-derived xenografts (PDXs) were constructed as described below. On day 0, NOD-SCID-IL2rγ null (B-NSG) mice were treated with 0 Gy (n = 5) or 2.5 Gy (n = 5) χ-rays for total body irradiation and intravenously injected with harvested human HLL cells from Patient #1 mixed with granulocyte colony-stimulating factor (G-CSF). Every mouse was injected with 1.5 × 10 7 cells within 24 h of irradiation.
For HLL PDX model establishment with BM-derived cells from patients, PDXs were constructed as described below. On day 0, B-NSG mice were treated with 0 Gy (n = 5) or 1.5 Gy (n = 6) χ-rays for total body irradiation and intravenously injected with harvested BM-derived cells from a human HLL patient (Patient #3) at day 1. The mice in the 1.5 Gy irradiated group were injected with 2.0 × 10 6 cells within 24 h of irradiation, and the mice without irradiation were injected with 1.5 × 10 7 cells. Blood samples were collected at intervals to test whether the model was successfully constructed.
On day 9, peripheral blood was collected through the tail vein in mice to validate whether these two models were successfully constructed. Routine blood tests were performed and body weights were determined at regular intervals. The mice were anesthetized with intraperitoneal (i.p.) injection of sodium pentobarbital administration at 120 mg/kg, then the harvested blood, bone marrow and tissue samples were collected for further analysis. The experimental procedures were approved by Experimental Animal Ethics Committee of Wuhan University.

Routine blood examination
Approximately 50 µl of blood was used for routine blood examination with a Coulter STKS automated blood cell analyzer (Beckman Coulter, USA).

Smear analysis and bone marrow biopsy
Blood smears and bone marrow smears were stained via Wright's staining and observed with a microscope under an oil immersion lens (OLYMPUS BX41) to measure the ratio of leukemia cells.
For the bone marrow biopsy, bone marrow collected from mouse thighs was xed in formalin and embedded in para n for bone marrow examination. The sections were sliced using a slicing machine (Leica RM2016) and incubated with anti-CD34 + (1:200, BD Biosciences). Finally, the images were observed with an OLYMPUS BX53 microscope.

HE staining and immunohistochemistry
The tissues were xed with formalin and embedded in para n to test the in ltration of leukemic cells by HE staining or immunohistochemistry. For IHC, the tissues were stained with antibodies against human anti-CD45 + (1:200, BD Biosciences) or human anti-CD34 + (1:200, BD Biosciences). The images were observed with an OLYMPUS BX41 microscope.

Flow cytometry
Harvested mouse peripheral blood and bone marrow were treated with RBC lysis buffer, HLL cells from Patient #1 were stained with human anti-CD45-PerCPcy5.5, anti-CD15-FITC, and anti-CD33-PE antibodies (BD Biosciences), and HLL or CD34 + cells from Patient #2 were stained with human anti-CD45-PerCPcy5.5 and anti-CD34-APC antibodies (BD Biosciences). The stained cells were analyzed by ow cytometry (FACS Canto II, FACS Verse, and FACS Diva software, BD Biosciences), and data were analyzed with FlowJo Software.

RNA extraction and quantitative real-time PCR (qPCR) detection
The collected fresh liver and spleen tissues from mice were treated and prepared as single cell suspensions. Then, total RNA from the liver, spleen and bone marrow was extracted with TRIzol reagent (Vazyme, Lot: R401-01). The cDNAs were obtained by Evo M-MLV reverse transcript kits (Accurate Biotechnology Co., Ltd, Lot: AG11705).
All qPCR reactions were performed with the SYBR® Premix Ex Taq™ kit (Takara) for 40 cycles using a uorescence quantitative PCR instrument (Applied Biosystems ABI 7500). The primer sequences were synthesized by Tsingke Biological Technology (Wuhan, China) and were as follows: (1) human WT1 gene forward (CAGGCTGCAATAAGAGATATTTTAAGCT) and reverse (GAAGTCACACTGGTATGGTTTCTCA), and (2) human reference gene for ABL gene forward (TGGAGATAACACTCTAAGCATAACTAAAGGT) and reverse (GATGTAGTTGCTTGGGACCCA). The positive rate for the WT1 gene in the mice was calculated.
NRAS mutation analysis DNA was extracted from the liver, spleen and bone marrow using a DNA Extraction System (TIANGEN, Lot: DP304). The PCR primer sequences for the NRAS gene were synthesized by Tsingke Biological Technology (Wuhan, China) and were as follows: forward (TAGCTAAGGATGGGGGTTGC) and reverse (ACTGGGCCTCACCTCTATGG). PCR was performed with 2 × Hieff™ PCR Master Mix (YEASEN, Lot: H2001071), and the products were sequenced by Tsingke Biological Technology (Wuhan, China) and analyzed by Chromas software (version 1.62).

Construction of the HLL model with PBMCs from patients
The individual information of enrolled HLL patients were collected from the electronic patient records and provided in Supplementary Table 1. PBMCs were collected from HLL patients who underwent leukapheresis by using a Fresenius COM.TEC blood cell separator. The collected leukocytes were isolated by density gradient centrifugation and cultured. The cell viability of separated leukemic cells was still over 90% after culture for four days (Additional le: Figure S1).
Harvested HLL cells were intravenously injected into B-NSG mice to construct the HLL PDX model, and the integrated work ow strategy is shown in Fig. 1a. The body weights of mice in the 2.5 Gy irradiated group gradually decreased (Fig. 1b). The nonirradiated mice survived longer than the mice in the 2.5 Gy irradiated group, and the survival rate was 66.67% (Fig. 1c). On the 9th day, the immunophenotype ratios of hCD45+, hCD15 + and hCD33 + cells in the 2.5 Gy irradiated group were higher than that of nonirradiated group (Fig. 1d). This shows that the HLL model can be successfully established under these two conditions. Because of the rapid death and poor survival of the mice in the 2.5 Gy irradiated group, we performed related experiments on the nonirradiated group. Routine blood examination showed that WBC counts rst increased and gradually decreased in the peripheral blood (PB) of B-NSG mice, and the maximum leukocyte count was above 37.35 × 10 9 /L. The red blood cell counts, platelet counts and hemoglobin content decreased gradually (Fig. 1e-h). Leukemic cells were found on the 9th, 21st and 25th days by blood smear analysis (Fig. 1i), and the immunophenotype ratio of hCD45 + cells increased up to 73.8% gradually in PB (Fig. 1j). Many leukemic cells in ltrated the bone marrow (BM) on the 25th day and even caused erythrocyte lysis, which was observed on a bone marrow smear (Fig. 1k). The immunophenotype ratio of hCD45 + cells in bone marrow were 43.62% (9.5%-75.2%), the hCD15 + cells were 54.07% (3.3%-79.9%) and the hCD33 + cells were 48.4% (5.7%-80.4%) (Fig. 1l). In addition, the histopathological analysis indicated that a large number of human leukemic cells in ltrated the liver and spleen tissue of the mouse (Fig. 1m).

Construction of the HLL model with BM-derived cells of patients
Furthermore, we collected the BM-derived cells of patients (Patient #3) who underwent BM aspiration, and the separated BM-derived cells were injected into B-NSG mice. The body weights were worse and the survival was poorer in the 1.5 Gy group than the 0 Gy group (Fig. 2a, b). Cell morphology analysis showed that human leukemia cells were present in blood smears and detected in blood smear experiments (Fig. 2c, d). Furthermore, the immunophenotype ratios of human leukemia cells in the PB and BM of mice were increased (Fig. 2e, f).

Successful construction of the CD34 + cell-derived xenograft model
To better explore the HLL model, we constructed a CD34 + hematopoietic stem cell-derived xenograft model of HLL. First, the work ow of model establishment is shown in Fig. 3a. At day 9, the immunophenotypes of hCD45 + and hCD34 + cells were detected at higher ratios in the CD34 + group with PB, and leukemia cells were also detected (Fig. 3b, c), which suggested that the model was successfully established. The body weights of mice in the CD34 + group decreased faster due to the severe tumor load (Fig. 3d). In the CD34 + group, routine blood examination indicated changes, especially the WBC counts and PLT counts (Fig. 3e, f, g, h). On the 25th day, leukemic cells were found by blood smear analysis and ow cytometry (Fig. 3i, j), which were more obviously observed in BM, and the immunophenotype ratios of hCD45 + and hCD34 + cells were slightly increased in the CD34 + group (Fig. 3k, l). Further analysis showed that CD34 + cells in ltrated into the BM, liver and spleen tissues on the 25th day, which indicated more obvious model construction using CD34 + hematopoietic stem cells (Fig. 3m). We also detected gene mutations in mice samples and found that the WT1 gene was positive in the liver, spleen and BM in mice (Fig. 3n). Moreover, NRAS mutations were found in the liver, spleen and BM in mice (Fig. 3o).

Summary of the constructed HLL models with B-NSG mice
The HLL models were constructed for multiple repeated experiments, and the patient information, separated cell characteristics, irradiation doses, injection cell counts, and survival time are summarized in Fig. 4a. By comparison analysis, the mice treated with a higher irradiation dose of 2.5 Gy had a shorter survival time (Fig. 4b), and the more injection of cells derived from PB resulted in a shorter survival time (Fig. 4c). In the absence of irradiation, the constructed model mice injected with cells derived from PB had shorter survival times than the mice injected with cells derived from BM of patients; however, there were no signi cant differences in survival time between the 1.5 Gy irradiated model mice receiving cells derived from PB and BM (Fig. 4d). Furthermore, mutations of the NRAS, DNMT3A, FLT3, and NPM1 genes were also analyzed, which indicated that the model mice injected with the NRAS gene mutation had a shorter survival time than those injected with cells lacking the NRAS gene mutation, and the same results were obtained with the DNMT3A, FLT3, and NMPM1 gene mutations (Fig. 4e). In addition, correlation analysis showed that the survival time in mice had a signi cant correlation with the survival status of enrolled patients (Fig. 4f).

Discussion
In this study, we used B-NSG mice with an interleukin-2 receptor gamma chain (IL2Rγ) deletion and nearly complete absence of the murine immune system to construct an HLL model, to improve HLL cell engraftment and reduce graft versus host disease [17,18]. The above experimental results showed that we successfully established a CD34 + hematopoietic stem cell-derived xenograft model of human HLL in B-NSG mice. First, a large number of HLL cells could be detected in blood by transplanting cells into the mice, and leukemic cells were observed based on cell morphology in the peripheral blood and bone marrow of mice. Then, a higher ratio of the immunophenotype of leukemic cells was detected by ow cytometry, which was consistent with the immunophenotype of the patients. Finally, the leukemic cells were severely in ltrated into other tissues, and gene mutations were detected in in ltrating tissues in the CD34 + hematopoietic stem cell-derived xenograft model. According to previous reports, 1.5-3.5 Gy doses of χ-ray radiation have been used before leukemia cell [19][20][21] or CD34 + cells [22,23] transplantation into mice, but only a few studies did not use irradiation [24]; for example, Shafat et al. reported the successful construction of a primary AML model with nonirradiated NSG mice [25]. In this study, we explored the radiation dose in model construction. For the 2.5 Gy χ-ray irradiated group, the body weight loss of mice was faster, and the survival time was shorter than that of the 0 Gy irradiated group, which indicated the 0 Gy irradiated mice could better used for drug research. The body weight of mice did not decrease much, and their body condition was good in the 0 Gy irradiated group; however, the mice in the 2.5 Gy group survived approximately 17 days and were in poor condition, dying easily, which is unfavorable for the study of drug e cacy or the mechanism of HLL development. Importantly, irradiation can brie y result in serious myelosuppression [26] and lead to an inability to clearly observe leukemic cells in blood smears and bone marrow smears, making it di cult to assess whether the model was successful. Moreover, myelosuppression seriously affects routine blood tests and cannot truly allow the evaluation of changes in WBCs [27,28]. Our results also showed that myelosuppression caused by irradiation could affect the survival time of mice and possibly lead to no signi cant difference in survival time between the 1.5 Gy irradiated model mice injected with cells derived from the PB and BM of patients.
In our previous reports on drug screening using a PDX model of HLL, B-NSG mice in the 2.5 Gy group survived beyond 20 days when injected with other HLL cells without malignant gene mutations [29], and B-NSG mice without irradiation survived beyond 31 days when injected with human AML cells separated from a hyperleukocytic AML-M5 patient with NPM1 and DNMT3A mutations [30]. In addition, the B-NSG mice with 1.5 Gy irradiation survived beyond 20 days when injected with human AML cells separated from a hyperleukocytic AML-M5 patient with WT1, DNMT3A and FLT3 mutations [31,32]. The mutational status of AML cells is also a key factor in immunode cient mice for the construction of a PDX model [33,34]. In this study, we found that the constructed model mice injected with NRAS gene mutations had shorter survival times than the injected with cells lacking the NRAS gene mutations, and the same results were observed for DNMT3A, FLT3, and NMPM1 gene mutations. The survival time in mice had signi cant correlation with the survival status of enrolled patients. Therefore, our results suggest that using B-NSG mice for HLL PDX construction is dependent on the use of HLL cells from patients who have malignant gene mutations and poor prognosis and is also related to the injected cell counts and the irradiation dosage in mice.
To establish a CD34 + hematopoietic stem cell-derived xenograft model with human hyperleukocytic AML cells, we found that the use of nonirradiated B-NSG mice injected with CD34 + cells from HLL patients without malignant gene mutations was not successful, and until the 50th day, leukemic cells were not found in blood smears or bone morrow smears. In this study, B-NSG mice were treated with 1.5 Gy χ-ray irradiation and injected with CD34 + cells derived from a HLL patient with WT1 and NRAS mutations, two mutations that frequently occur in AML [35]. The CD34 + cell-derived PDX model of HLL was established successfully, which also indicated that B-NSG mice with low-dose radiation were better used for model construction based on injection of cells with moderate malignant mutations.
However, the HLL model still has some limitations; for example, HLL cells from patients show di culty proliferating in vitro, and the mice do not present full clinical characteristics, such as certain immunophenotypes, so the PDX model of HLL is still not standardized and requires further investigation, and we are still unable to replicate the complex immune microenvironments in immunode ciency mice [36], so it is di cult to explore the complex heterogeneity of HLL and the interaction and progression between HLL cells and their microenvironment [14,37]. Further related experiments will be investigated. Even so, the HLL mouse model is still of great value in the study of the pathogenesis and drug e cacy.

Conclusion
In conclusion, we transplanted hyperleukemic leukemia cells into severe immunode cient B-NSG mice, which retained most of the original characteristics in the B-NSG mice, and more accurately simulate the characteristics of tumor growth in clinical patients. we assessed the construction methods used for HLL

Declarations
Ethics approval and consent to participate The study was carried out followed the Declaration of Helsinki and approved by the Research Ethics Committee of Zhongnan Hospital at Wuhan University (license number: 2017048). Informed written consent was obtained from the HLL patients.

Consent for publication
Not applicable.

Data Availability Statement
All data are available in the main text or the Supplementary materials.

Competing interests
The authors declare that they have no competing interests.  Establishment of the model of HLL with PBMCs of patients. a The HLL cells were collected from HLL patients using a Fresenius COM.TEC machine, and then the PBMCs were separated with a Ficoll kit and cultured. After irradiation at day 0, the B-NSG mice were treated with 0 Gy (n=5) or 2.5 Gy (n=5) χ-rays for total body irradiation and intravenously injected with harvested from PBMCs of human HLL cells from Patient #1 mixed with granulocyte colony-stimulating factor. Every mouse was injected with 1.5 × 107 Construction of the HLL model with BM-derived cells of patients a On day 0, B-NSG mice were treated with 0 Gy (n=5) or 1.5 Gy (n=6) χ-rays for total body irradiation and intravenously injected with harvested BM-derived cells from a human HLL patient (Patient #3) at day 1, then tested the change in body weight at intervals. b Survival analysis. c Blood smear analysis. PB, peripheral blood. d Bone marrow smear analysis. BM, bone marrow. e Immunophenotype analysis of leukemia cells in peripheral blood at day 30 by ow cytometry. **P<0.01. f Detection of the immunophenotype in bone marrow by ow cytometry after mice were sacri ced.

Figure 3
Construction of a CD34+ hematopoietic stem cell-derived xenograft model of HLL. a The work ow of model construction. The B-NSG mice were treated with 1.5 Gy χ-rays for total body irradiation on day 0 and intravenously injected with isolated control cells (n=5) or 1.5 × 106 CD34+ cells using magnetic beads (n=5) from Patient #2 (WT1 and NRAS mutation) mixed with G-CSF within 24 h of irradiation. b Detection of immunophenotype in peripheral blood by ow cytometry at day 9. c Blood smear analysis at day 9. d The analysis of body weight. e-h Routine blood tests. e, WBC. f, RBC. g, PLT. h, HGB. i Blood smear analysis at day 25. j Detection of immunophenotype in peripheral blood at day 25. The cells were stained with human anti-CD45-PerCPcy5.5 and anti-CD34-APC antibodies. k Bone marrow smear analysis at day 25. l Detection of immunophenotype in bone marrow at day 25. m Immunohistochemical analysis of leukemia cell in ltration into bone marrow, liver and spleen tissues. The tissues were stained with antibodies against human anti-CD45+. BM, bone marrow. Magni cation: ×400. n Detection of the positivity rate of the WT1 gene using a uorescence quantitative PCR instrument in tissues of mice in the CD34+ group. o Detection of NRAS mutation in tissues of mice in the CD34+ group. Green represents base A, red represents base T, blue represents base C and black represents base G. The green arrow represents a mutation in BM. BM, bone marrow.