Ruxolitinib/nilotinib cotreatment inhibits leukemia-propagating cells in Philadelphia chromosome-positive ALL

Six patients with de novo Ph⁺ALL diagnosed at Peking University Institute of Hematology from January 1, 2015 to May 31, 2015 were enrolled for this in vitro and in vivo study. The patient clinical characteristics are shown in Table 1. The inclusion criteria were (1) 18–60 years of age, (2) a diagnosis of ALL based on the 2008 World Health Organization (WHO) criteria, and (3) the detection of the Ph-chromosome and/or BCR–ABL mRNA. Bone marrow mononuclear cells (BMMNCs) from the patients at diagnosis were rapidly isolated by density centrifugation using a lymphocyte separation medium (GE Healthcare, Milwaukee, WI, USA). The BMMNCs were immediately cryopreserved in 10% dimethyl sulfoxide (Sigma, St. Louis, MO, USA) with 90% fetal bovine serum (FBS, Gibco, Gaithersburg, MD, USA). The BMMNCs were stored in liquid nitrogen until the cell sorting procedure was performed. The study was approved by the Ethics Committee of Peking University People’s Hospital, and written informed consent was obtained from all patients before study-entry in accordance with the Declaration of Helsinki.

The frozen BMMNCs of de novo Ph⁺ALL patients (N = 6) were thawed and stained with mouse anti-human CD58-FITC (Beckman-Coulter, Brea, CA, USA) and CD34-PE, CD19-APC-Cy7, CD45-PerCP, CD38-APC monoclonal antibodies (MoAbs, Becton–Dickinson, San Jose, CA, USA). The LPCs (CD34⁺CD38⁻CD58⁻) and other cell fractions (including CD34⁺CD38⁻CD58⁺, CD34⁺CD38⁺CD58⁻ and CD34⁺CD38⁺CD58⁺) in the viable BMMNCs were defined and sorted using a FACS Aria II (Becton–Dickinson) as previously reported¹² (Fig. 1). The purity of each fraction was >97%. Fluorescence-minus-one controls were used to identify positive events for CD34, CD38 and CD58. The data were analyzed using BD LSRFortessa software (Becton–Dickinson).

To search for the potential molecular basis involved in LPC-mediated Ph⁺ALL progression, we performed RNA-Seq with the sorted LPCs and other cell fractions from the patients with de novo Ph⁺ALL (N = 2) to analyze their gene expression profiles. Total RNA was isolated from pellets using the RNeasy Mini Kit (Qiagen, Valencia, CA, USA). Three micrograms of RNA per sample was used as input material, and sequencing libraries were generated using the NEBNext Ultra RNA Library Prep Kit for Illumina (NEB, Ipswich, MA, USA). The library sequencing was performed on an Illumina HiSeq 2500 platform, and 125-bp paired-end reads were analyzed. Downstream analysis was performed using a combination of programs, including Bowtie2, Tophat2, HTseq, Cufflink and our wrapped scripts. The DESeq R package (1.10.1) was used to analyze the differential expression between the LPCs and other cell fractions. In all statistical analyses, P-values were analyzed using the Benjamini-corrected modified Fisher’s exact test, and a P < 0.05 was considered statistically significant.

To confirm the RNA-Seq results, the relative JAK2 mRNA levels (forward primer: 5′-TCTGGGGAGTATGTTGCAGAA-3′; reverse primer: 5′-AGACATGGTTGGGTGGATACC-3′) between the LPCs and other cell fractions sorted from the patients with de novo Ph⁺ALL (N = 6) were analyzed using a SYBR green-based qRT-PCR technique. Normalized levels of the JAK2 ratios in the qRT-PCR assays were evaluated through comparisons with the GADPH levels (forward primer: 5′- GCACCGTCAAGGCTGAGAAC -3′; reverse primer: 5′- TGGTGAAGACGCCAGTGGA -3′).

The sorted LPCs (1 × 10⁵/well) were cultured in StemSpan (Stem cell Technologies, Vancouver, BC, Canada) supplemented with four growth factors (20 ng/mL recombinant human (rh) interleukin-3 (rhIL-3), 20 ng/mL rhIL-7, 20 ng/mL rh Flt3-ligand (rhFlt3-L), and 50 ng/mL rh stem cell factor (rhSCF)) (PeproTech, Locky Hill, NJ, USA). After 48 h of suspension culture with the vehicle (DMSO, Sigma), imatinib (5 μM, Novartis, Basel, Switzerland) [24, 33], nilotinib (5 μM, Novartis) [33], and/or ruxolitinib (300 nM, Novartis) [34, 35] treatments, an apoptosis assay was performed on the LPCs using the Annexin-V and 7-amino-actinomycin D (7-AAD) Apoptosis Detection Kit (Becton–Dickinson) as described in our previous reports [36, 37]. The late apoptotic cells (Annexin-V⁺/7-AAD⁺) were analyzed using the BD LSRFortessa software (Becton–Dickinson). Aliquots of isotype-identical antibodies served as negative controls.

The anti-mouse CD122 [interleukin-2 receptor β (IL-2Rβ)]-conditioned NOD/SCID xenograft assay was performed with intra-bone marrow injection (IBMI) as previously reported [10, 38‐40]. Briefly, 5- to 6-week-old NOD/SCID mice (Vital River Laboratories, Beijing, China) were sub-lethally irradiated (2.1 Gy total body irradiation from a ⁶⁰Co source) followed by treatment with 200 μg of anti-mouse CD122 monoclonal antibody generated from hybridoma TM-β1 (provided by Dr. T. Tanaka of Hyogo University of Health Sciences, Kobe, Japan) [38]. The mice then received an IBMI of the LPCs fraction sorted from de novo Ph⁺ALL patients (N = 6) < 24 h post-irradiation and CD122 treatment. The injected LPCs doses were 1 × 10⁴ cells/mouse (6 mice for each treatment condition). All animal experiments were approved by the Ethics Committee of Peking University People’s Hospital.

To evaluate the effects of selective BCR–ABL and/or JAK2 inhibition therapy on Ph⁺ALL LPCs, the recipient mice were treated with vehicle (10% 1-methyl-2-pyrrolidone and 90% polyethylene glycol 300; Sinopharm Chemical Reagent Co., Ltd, Shanghai, China) or with a single or different dual combination oral gavage treatments with imatinib (100 mg/kg/day, Novartis), nilotinib (75 mg/kg/day, Novartis), and/or ruxolitinib (30 mg/kg/day, Novartis) simultaneously once or twice a day starting at 2 weeks post-transplant and lasting for 2 weeks. The mice were observed daily for body weight changes and survival during and after treatment.

Engraftment of human Ph⁺ALL cells was defined based on the frequency of hCD45⁺ cells. Moribund mice were euthanized at 8 and 12 weeks post-transplant, and a mixture of cells from tibias, femurs and the spleen were obtained. To assess human Ph⁺ALL cell engraftment, flow cytometric analyses were performed using mouse anti-hCD58-FITC (Beckman-Coulter), CD34-PE, CD19-APC-Cy7, CD45-PerCP, and CD38-APC MoAbs (Becton–Dickinson).

To evaluate the engraftment levels of BCR/ABL-expressing cells, BCR/ABL transcripts were detected in BM cells from murine recipients using a TaqMan-based qRT-PCR assay performed as previously described [10]. BCR/ABL primers and probes that amplify b3a2 and b2a2 junctions have been reported [41, 42]. The primers and probes that amplify the ABL and e1a2 BCR/ABL junctions came from data reported by the Europe against Cancer Program [43, 44]. Normalized levels of the BCR/ABL ratios in the qRT-PCR assays were evaluated by comparisons to the ABL levels [10, 41, 42].

For a histopathological analysis of leukemic infiltration, femur, liver, spleen, and kidney tissues of recipients were fixed with 4% paraformaldehyde for 1 h, dehydrated with 70% ethanol, and embedded in paraffin, followed by the preparation of 5-µm sections. Hematoxylin–eosin (HE) staining was performed on each tissue section. Immunohistochemistry (IHC) with rabbit anti-hCD19 (Abcam, Cambridge, MA, USA) was performed after a graded alcohol dehydration and antigen retrieval using heated citrate buffer. Sections were studied by light microscopy (Axiovert 200; Carl Zeiss, Jena, Germany).

To understand how the different treatments affected the molecular pathways, western blot analyses using anti-human antibodies to phosphorylated CrkL and JAK2 were performed on the BM cells harvested from the different groups of recipient mice. The antibodies were obtained from Cell Signaling Technologies (Danvers, MA, USA).

Statistical analyses were performed using the χ² test for categorical variables and the Mann–Whitney U test for continuous variables. Analyses were performed using the SPSS 22.0 (IBM, Armonk, NY, USA) and GraphPad Prism 6.0 software packages (GraphPad Software, La Jolla, CA, USA), and differences with P < 0.05 were considered statistically significant.

Approximately 62–76 million clean reads were obtained to establish four RNA-Seq libraries. As shown in Table 2, high percentages of reads (83.6–84.6%) were mapped. A total of 77.0–91.3% of the mapped reads were located within exons, whereas less than 19.6% of the mapped reads were located within introns and intergenic regions. These data indicated that we established four libraries with high quality, which allowed us to compare the transcriptomes between the LPCs and other cell fractions from the de novo Ph⁺ALL patients.

A Venn diagram shows that 3722 genes were differentially expressed between the LPC₁ and Other Cell₁ fractions sorted from patient No. 1 with de novo Ph⁺ALL, whereas 4162 genes were differentially expressed between the LPC₂ and Other Cell₂ fractions from patient No. 2, and 2800 differentially expressed genes were co-expressed in both patients (Fig. 2a). A scatter plot shows the differentially expressed genes (Fig. 2b left panel, LPC₁ vs. Other Cell₁; Fig. 2b right panel, LPC₂ vs. Other Cell₂) with |fold change| >1.5 and Padj < 0.05. There were 3722 genes differentially expressed between the LPC₁ and Other Cell₁ fractions. Among them, 1825 genes were up-regulated and 1897 genes were down-regulated. Moreover, 1954 genes were up-regulated and 2208 genes were down-regulated in the LPC₂ fraction compared with their expression levels in the Other Cell₂ fraction. Based on the RNA-Seq data from patients with de novo Ph⁺ALL, JAK2 was differentially expressed and was more highly expressed in the LPC₁ fraction than in the Other Cell₁ fraction (1.60-fold change, P = 0.03) and was more highly expressed in the LPC₂ fraction than the Other Cell₂ fraction (1.79-fold change, P = 0.004).

The relative mRNA expression levels of BCR/ABL and JAK2 between the LPCs and other cell fractions from the patients with de novo Ph⁺ALL (N = 6) were further analyzed using qRT-PCR assays. No significant difference was found in the BCR/ABL mRNA levels between the LPCs and other cell fractions from the patients with de novo Ph⁺ALL as measured by qRT-PCR (Fig. 2c, 1.07 ± 0.17-fold, P > 0.99). Consistent with the RNA-Seq results, JAK2 mRNA levels in the LPCs fraction were significantly higher than those in the other cell fractions (Fig. 2c, 2.05 ± 0.19-fold, P = 0.002). Likewise, the significantly higher levels of JAK2 mRNA detected by qRT-PCR were further validated in the protein level of phospho-JAK2 in the LPCs fraction when compared to the other cell fractions by western blot (Fig. 2d).

To investigate anti-LPCs effects in vitro, the levels of drug-induced cell apoptosis were analyzed in the different treatment groups. As shown in Fig. 3, imatinib or ruxolitinib alone had no significant anti-LPCs effect in vitro compared with the effect in the vehicle-treated control group. Among the different treatment groups, cotreatment with nilotinib and ruxolitinib induced significantly higher levels of early apoptosis (Fig. 3b) and late apoptosis (Fig. 3c) in LPCs than did the other treatment groups. Alternatively, similar high levels of alive LPCs were found in the ruxolitinib-treated group (59.3 ± 1.2% vs. 62.6 ± 1.9%, P = 0.15) or imatinib-treated group (58.9 ± 1.4% vs. 62.6 ± 1.9%, P = 0.15) compared with the vehicle group. No significant difference was found between ruxolitinib-treated group and imatinib-treated group (59.3 ± 1.2% vs. 58.9 ± 1.4%, P = 0.98). In contrast, nilotinib (35.8 ± 0.6% vs. 59.3 ± 1.2%, P < 0.0001), imatinib combined with ruxolitinib (41.1 ± 0.8% vs. 59.3 ± 1.2%, P < 0.0001), or nilotinib combined with ruxolitinib (19.4 ± 0.9% vs. 59.3 ± 1.2%, P < 0.0001) resulted in significant reduction in alive LPCs compared with those in the ruxolitinib-treated group. Among these treatment groups, the levels of alive LPCs in the nilotinib combined with ruxolitinib-cotreated group exhibited the most marked reduction (Fig. 3d).

When the recipient mice exhibited ruffled fur and lethargy, the engraftment levels of human Ph⁺ALL were analyzed in the BM and spleen. Two months after the transplantation of 1 × 10⁴ LPCs and treatment with vehicle, the recipient mice exhibited splenomegaly. The engrafted human cells were further confirmed by morphologic and cytogenetic analyses (Fig. 4a). Flow cytometry analysis demonstrated that the BMs and spleens of the recipients were efficiently engrafted with human Ph⁺ALL cells with an aberrant phenotype similar to that in the donor Ph⁺ALL patients (Fig. 4b). There was widely disseminated disease, including significant leukemic infiltration into the BM, liver, spleen, and kidneys of the recipient mice, as shown by HE staining and IHC with anti-hCD19 (Fig. 4c). These results indicated that a humanized Ph⁺ALL xenotransplant model was successfully established using anti-CD122-conditioned NOD/SCID mice and IBMI with LPCs from Ph⁺ALL patients.

To further evaluate the effects of selective BCR–ABL and/or JAK2 inhibition therapy on Ph⁺ALL LPCs in vivo, the recipient mice transplanted with LPCs were randomized to different treatment groups on day 15 post-transplant and were treated with various single or dual drug interventions for 14 days (Fig. 5a). At 8 weeks (Fig. 5b) and 12 weeks (Fig. 5c) post-transplant, similar high levels of human Ph⁺ALL engraftment were observed in the imatinib-treated mice and the vehicle-treated control mice. Consistent with the anti-LPCs effects in vitro, ruxolitinib alone had no significant anti-LPCs effect in the recipient mice, but this drug reduced Ph⁺ALL engraftment more significantly when ruxolitinib was administered with imatinib or nilotinib. Notably, treatment with the combination of nilotinib and ruxolitinib, compared with imatinib, nilotinib, or ruxolitinib treatment alone or imatinib combined with ruxolitinib, led to the most significant reduction in human Ph⁺ALL engraftment in the BMs and spleens of the recipients (Fig. 5b, c). Moreover, HE staining and IHC with anti-hCD19 demonstrated that the infiltrating levels of the transplanted LPCs were significantly lower in the spleen tissues of the nilotinib and ruxolitinib-cotreated mice than in those of mice in the other treatment groups (Fig. 6b). Further evidence that the most effective anti-LPCs effect occurred with the combination treatment was derived by the engraftment analysis of BCR/ABL expressing cells using a TaqMan-based qRT-PCR assay (Fig. 6c).

To determine the activities of JAK2 in humanized mice with Ph⁺ALL LPCs following treatment with single agents or a combination of ruxolitinib and imatinib or nilotinib, levels of phospho-JAK2 were examined by western blot analysis (Fig. 6d). Among the different treatment options, the combination of nilotinib and ruxolitinib was the most effective at reducing phospho-JAK2 levels in the recipients, whereas imatinib did not reduce the level of phospho-JAK2.

The phosphorylation levels of CrkL (Fig. 6d), the kinase substrate of BCR–ABL, were evaluated. We found that nilotinib or nilotinib treatment combined with ruxolitinib inhibited BCR–ABL kinase activity completely, whereas the phospho-CrkL levels did not change significantly when the recipient mice were treated with ruxolitinib alone.

These results are consistent with the engraftment levels evaluated using flow cytometry (Fig. 5b, c), histopathological analyses (Fig. 6b), and qRT-PCR assays (Fig. 6c), indicating that the combination of nilotinib and ruxolitinib more effectively reduced the LPCs capacity in immunodeficient mice through a deeper suppression of JAK2 activity than either single agent or the combination of imatinib and ruxolitinib. However, further comparison of the total JAK and CrkL protein is required to illuminate whether the change in phosphorylation was due to reduced amounts of total protein.