Characterization of a novel glucocorticoid-resistant human B-cell acute lymphoblastic leukemia cell line, with AMPK, mTOR and fatty acid synthesis pathway inhibition

Acquired glucocorticoid (GC) resistance remains the main obstacle in acute lymphoblastic leukemia (ALL) therapy. The aim of the present study was to establish a novel GC-resistant B-ALL cell line and investigate its biological characteristics. A cell culture technique was used to establish the GC-resistant cell line from the parental cell, NALM-6. Molecular and cellular biological techniques including flow cytometry, MTT assay, western blotting, DNA fingerprinting analysis and whole transcriptome sequencing (WTS) were used to characterize the GC-resistant cell lines. Nude mice were used for xenograft studies. The GC-resistant cell line, NALM-6/HDR, was established by culturing NALM-6 cells under hypoxia for 5 weeks with a single dexamethasone (Dex) treatment. We subcloned the NALM-6/HDR cell lines, and got 6 monoclone Dex-resistant cell lines, NALM-6/HDR-C1, C3, C4, C5, C6 and C9 with resistance index (RI) ranging from 20,000–50,000. NALM-6/HDR and its monoclone cell line, NALM-6/HDR-C5, exhibited moderate (RI 5–15) to high resistance (RI > 20) to Ara-c; low or no cross-resistance to L-Asp, VCR, DNR, and MTX (RI < 5). STR analysis confirmed that NALM-6/HDR and NALM-6/H were all derived from NALM-6. All these cells derived from NALM-6 showed similar morphology, growth curves, immunophenotype, chromosomal karyotype and tumorigenicity. WTS analysis revealed that the main metabolic differences between NALM-6 or NALM-6/H (GC-sensitive) and NALM-6/HDR (GC-resistant) were lipid and carbohydrates metabolism. Western blotting analysis showed that NALM-6/HDR cells had a low expression of GR and p-GR. Moreover, AMPK, mTORC1, glycolysis and de novo fatty acid synthesis (FAS) pathway were inhibited in NALM-6/HDR when compared with NALM-6. NALM-6/HDR cell line may represent a subtype of B-ALL cells in patients who acquired GC and Ara-c resistance during the treatment. These patients may get little benefit from the available therapy target of AMPK, mTORC1, glycolysis and FAS pathway.


Background
Although contemporary chemotherapy regimens have improved the long-term survival rate of childhood acute lymphoblastic leukemia (ALL) to nearly 90% [1], the outcome of relapsed/refractory pediatric ALL is still extremely poor. Acquired drug resistance to chemotherapy, especially resistance to glucocorticoids (GCs), remains the main obstacle in ALL therapy [2,3]. Many efforts were made to seek the mechanism of GC resistance and restore GC sensitivity, but showed little benefit to patients.
Drug-resistant cell lines that are established from drug-sensitive cell lines are the most useful models in tumor research, which have been used to investigate the mechanisms of drug resistance for nearly 50 years [4]. Traditionally, two methods, stepwise dose-escalation continuous exposure and high-concentration pulsatile exposure, are available to acquire drug resistance in cell lines [5]. There is a controversy regarding the traditional drug-exposure method for that the resistant phenotype obtained after a long time (6 ~ 16 months, or more) drug exposure and can only remain for several weeks or months after discontinuation of drug exposure, whether or not they can reflect the clinical setting [4]. Therefore, we constructed a novel and convenient method to establish GC resistant ALL cell lines from GC sensitive ALL cells by mimicking the hypoxic bone marrow (BM) microenvironment [6]. Compared to the traditional methods, the novel method has a high success rate, a single drug exposure, a short duration (5 ~ 6 weeks), and a long maintenance of resistant phenotype without drug re-induction [6].
Hypoxia plays a central role in ALL progression and resistance to therapy [7]. Glycolysis-based metabolic reprogramming enhances multidrug resistance [8]. Previously, we established a T-ALL GC-resistant cell line, CEM-C7/HDR, and reported its biological characteristics [6]. T-ALL and B-ALL are highly heterogeneous disease and may have distinct mechanisms of GC resistance. To clarify the underlying mechanisms and therapeutic targets of hypoxia in relapsed/refractory B-ALL, we established a novel B-ALL GC-resistant cell line, NALM-6/HDR, derived from NALM-6 cell line, and studied the distinct biological characteristics of the cell line.

Cell growth and viability assay
Exponentially growing cells were cultured in a 6-well sterile plastic culture plates for seven days. The number of viable cells was counted using heamocytometer and trypan blue staining every day. Doubling time (Td) of cells was calculated according to the following formula: Td (h) = t × lg2/lg(N t /N 0 ), where t is the time of continuous culture, N t is the final number of cells, and N 0 is the initial number of cells.
Cell viability was evaluated by MTT assay. Briefly, cells were seeded in 96-well plates. Next, MTT with a final concentration of 0.5 mM was added to the wells and incubated for 4 h at 37 °C. Then acidified SDS solution (10% SDS in 0.01 M HCl) were added to each well and incubated overnight at 37 °C. Finally, the optical density was measured at 570 nm using a multiplate reader (Multiskan Spectrum, Thermo Electron Co., Waltham, MA, USA). Values were obtained by comparing the experimental cells with their respective controls. Mean values were calculated from triplicate cultures.

Chemosensitivity assays
The chemotherapeutic drugs, Dex, asparaginase (L-Asp), daunorubicin (DNR), vincristine (VCR), arabinoside (Ara-c), and methotrexate (MTX), were purchased from Sigma. All cells were treated with increasing concentrations of different drugs for 48 h, followed by assessment of cell viability by MTT assay. The IC 50 was calculated by linear interpolation.

Cell cycle analysis
Cells were harvested 48 h after treatment and fixed overnight in 70% ethanol at 4 °C. The cells were then washed twice with PBS and stained with 5 μg/ml PI in the presence of DNAse-free RNAse (Sigma) for 30 min at room temperature. Then, the cell cycle distribution was analyzed by flow cytometry (Cytomics FC 500 and CXP & Multicycle software, Beckman Coulter Inc., Miami, FL, USA).

Whole transcriptome sequencing (WTS) analysis
Total RNA was extracted by Trizol reagent (Invitrogen, Thermo Fisher, Carlsbad, CA, USA) according to the manufacturer's instructions. RNA quality and purity were assessed by RNA 6000 Nano assay on Agilent 2100 Bioanalyzer (Agilent, Santa Clara, CA, USA). RNA-sequencing libraries were acquired on the BGISEQ-500 Platform in total and generated approximately 10.98 Gb per sample. After reads filtering, clean reads were mapped to the reference genome using HISAT. Then, gene expression levels for each sample were calculated with RSEM. Based on the gene expression level, the differentially expression genes (DEGs) between samples can be identified. Pos-sionDis algorithms were used to detect the DEGs. With DEGs, we perform Gene Ontology (GO) classification, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway classification and functional enrichment.

Western blotting analysis
Briefly, cells (10 6 ) were washed twice in cold PBS and then lysed by RIPA lysis buffer (Bio-Rad). Equivalent amounts of proteins were separated by 8-15% SDS-polyacrylamide gel electrophoresis and transferred to PVDF membranes (0.22 or 0.45 μm, Millipore). Immunoblots were assayed by using primary antibodies. HRP conjugated secondary antibodies were used for detection. The level of the β-Actin protein was used as a control for the amount of protein loaded into each lane.Proteins were visualized by incubation with ECL plus (Millipore). All experiments were conducted independently at least 3 times.

Animal experiments
Cultured 5 × 10 6 exponentially growing cells were subcutaneously injected into the right flanks of 6-week-old female BALB/c (nu/nu) nude mice (GemPharmatech Co., Ltd, Nanjing, China), and 0.1 ml of PBS was injected into the left flanks as the control. Tumor size was measured by calipers every 2 days. The approximate tumor volume was calculated using the equation V = (length × width × width)/2. All animals were eartagged and monitored individually throughout the experiment. All animal care was in compliance with the guidelines established by the internal Institutional Animal Care and Use Committee and Ethics Committee of Sichuan University. Mice were euthanized by cervical dislocation after anesthetizing with a mixture of 100 mg/kg ketamine and 10 mg/kg xylazine via intraperitoneal injection. Then, the tumor mass was excised, fixed in 10% formalin, and routinely processed for paraffin embedding. Five-millimeter-thick sections were obtained and prepared for standard histopathological examination.

Statistical analysis
All assays were performed in triplicate, and data are expressed as mean values ± SD. One-way ANOVA was used to compare two groups. A p-value < 0.05 was considered to be significant.

Establishment of Dex-resistant NALM-6/HDR cell line and its mono-clone cell line
NALM-6/HDR cell line was constructed following the method we reported previously [6]. Logarithmically growing NALM-6 cells were cultured under hypoxia condition for 5-6 weeks with or without dexamethasone (Dex) treatment. Then, the cells were cultured continuously under normoxic condition and constructed a stable resistant cell line, NALM-6/HDR (Fig. 1a). The IC 50 and RI of Dex in the NALM-6/HDR cells were 55-70 µM and 25,000-30,000, respectively [6]. All NALM-6/HDR subclones were resistant to Dex with IC 50 and RI ranging from 50 to 125 µM and 25,000-50,000, NALM-6/HDR-C5 cells were the most resistant cells to Dex among the monoclonal cell lines (Fig. 1b). Unlike the CEM-C7/H [6], the control group, NALM-6/H, which was cultured under the hypoxic condition with no Dex treatment, did not get Dex resistance. The IC 50 of Dex in the NALM-6, NALM-6/H and its monoclonal cells was 0.002-0.005 µM. GC-resistant cells may get cross-resistance to other drugs [9]. Our results showed that NALM-6/HDR and its monoclonal cell line, NALM-6/HDR-C5, exhibited very high resistance to Dex, moderate (RI 5-15) to high resistance (RI > 20) to Ara-c (Fig. 1f, h); low or no cross-resistance to L-Asp, DNR, VCR, and MTX (RI < 5) ( Fig. 1c-h). Although IC 50 of L-Asp, DNR, VCR and MTX were higher in NALM-6/H than in NALM-6, the RI only ranged from 1.43 to 2.04, which indicated that hypoxia treatment only would not induce drug resistance in NALM-6.

Expression of proteins associated with fatty acid metabolism
Changes in lipid metabolism, in particular increased de novo fatty acid synthesis (FAS), are recognized as one of the key hallmarks in several cancer cells [11,12]. Besides, FASN overexpression has shown to be associated with poor prognosis and resistance to chemotherapy [12,13]. However, in accordance with the results of transcriptome analysis, the results of western blotting showed that the expression of ACLYp-ACLY, ACC, p-ACC, FASN, ACSL1 and AceCS1 were all lower in NALM-6/HDR than in NALM-6 ( Fig. 4a, b), which means that NALM-6/HDR cells got a relative low FAS metabolism phenotype accompanying with GC-resistance feature. NALM-6/H showed a relative similar FAS metabolism phenotype with NALM-6, except for a lower expression of ACSL1 and AceCS1 and a higher expression of FASN (Fig. 4a, b). The results hinted that only the GC-resistant NALM-6/HDR cells got a relative low FAS metabolism phenotype.

Discussion
Over the last century, studies on patient-derived cancer cell lines have provided crucial insights into the biological processes in tumorigenesis and high-throughput screening for drug development [19]. Molecular alterations that are detected in a drug-resistant cell line, when compared to its drug-sensitive counterpart, can suggest molecular mechanisms underlying drug resistance. Drug resistant cell lines, constructed by the traditional drugexposure method, have been used in the research on the mechanisms of drug resistance for 50 years, which created a new problem whether or not they can reflect the clinical setting and resole the clinical problems [4]. Furthermore, drugs at a high concentration can kill all parental cells, which make construction of a highly resistant cell line difficult and make construction of a resistant cell line from a highly drug-sensitive cell line more difficult. Therefore, we established a method that matches the manner in which chemotherapy is administered to patients [6]. NALM-6 is highly sensitive to GC, and it is one of the 5 pre-B-ALL cell lines in 100 authenticated leukemia-lymphoma cell lines [20,21]. Meanwhile, NALM-6 was widely used in the researches on ALL and was recommended as a classic/reference cell line of BCP-ALL [21]. Fortunately, via the novel method, we established the GC-resistant B-ALL cell line, NALM-6/HDR, which is highly resistant to GC with a moderate to high cross-resistant to Ara-C, and shows similar morphology, growth curves, immunophenotype, chromosomal karyotype and tumorigenicity with its parental cell line, NALM-6, and the control cell line, NALM-6/H. GCs specifically induce apoptosis in malignant lymphoblasts and are thus pivotal in the treatment of lymphoid malignancies, especially in ALL [22]. However, GC resistance remains one of the strongest predictors of relapse in ALL and accounts for most of the treatment failures [2,3]. Despite decades of clinical practice and laboratory investigation, the exact molecular mechanisms contributing to GC resistance has not been fully uncovered. GCs induce apoptosis by combined with GR and translocated to nucleus and then interaction with GC response elements (GRE), and then activating or suppression transcription of target genes [23,24]. Initially, most studies showed that low GR protein levels plays a crucial role in GC resistance [15,16,[23][24][25][26], but then the conclusion was rebutted by the results that high GR expression was found also in GC-resistant ALL cells [23,24,27,28]. Anyway, almost all of the GC sensitive ALL cells had a high expression of GR [23]. Recently, loss of GR expression renewed attention on the exploration of GC resistance [15,16,23]. Our results showed that NALM-6/ HDR expressed a lower GR and p-GR (Ser211) compared to NALM-6 and NALM-6/H in vivo and in vitro. The same, CEM-C7/HDR cells, which were constructed using the same method, showed a low GR expression also [6]. These data implicate reduced GR expression as an important cause of acquired GC resistance.
However, the mechanism of GC resistance is much more complicated than a GR alternation. The level of the GRs expression among ALL patients and derived cell lines appears to be highly heterogeneous [23]. There are several additional mechanisms, triggered by alterations of different signaling pathways, including Notch, Il7R/JAK/ STAT, RAS/MEK/ERK, PTEN/PI3K/AKT/mTOR and Ca 2+ signaling, which cause the metabolic reprogramming, with an enhanced level of glycolysis and oxidative phosphorylation, apoptosis resistance, and multidrug resistance [23]. According to the clinical set, hypoxic stress play a pivotal role in inducing GC resistance in ALL cells [6,7,29,30]. Generally, tumor cells respond to hypoxic stress by upregulating various genes involved in glucose uptake, glycolysis, and angiogenesis, to maintain the cellular metabolic demands; at the same time, translation of "unnecessary" proteins are deregulated to prevent the accumulation of stress-induced unfolded and/ or misfolded proteins [29,31]. Leukemia cells were originated from BM and grown in the hypoxic microenvironment, and then released to the peripheral blood with adequate oxygen. Therefore, leukemic cells, living in both hypoxic and normoxic microenvironments, may have distinct metabolic response to hypoxia from solid tumor cells. A real-time metabolism study on chronic lymphocytic leukemia (CLL) cells showed that circulating CLL cells were primed for hypoxia and showed the same adaptive response upon secondary exposure to hypoxia [32]. Similarly, our results showed that NALM-6/H had a little difference in metabolism when compared to NALM-6. The expression of p-AMPK even decreased in NALM-6/H when compared to NALM-6, which means that there was almost no energy metabolic problem for NALM-6 upon exposure to hypoxia. However, the GCresistant NALM-6/HDR showed an obvious difference in metabolism, especially in lipid metabolism, when compared to NALM-6 and NALM-6/H. Cancer cells reprogram lipid metabolism during their malignant progression [12]. Enzymes in de novo FAS pathway, including ACLY, ACC and FASN, are upregulated in numerous cancers [12]. ACLY occupies a privileged position at the crossroads of the metabolism of cancer cells by linking both the glycolysis and lipid metabolism, having been found overexpressed in many aggressive cancers, and might play an important role in cancer relapse [33][34][35]. The researches on acute myeloid leukemia (AML) showed that low expression of ACLY is associated with favorable prognosis [36,37]. On the contrary, our results showed that the ACLY and p-ALCY reduced in NALM-6/HDR. ACC is the first rate-limiting enzyme for FAS, which plays a critical role in leukemia progression [38]. ACC was initially considered as an oncoprotein [39], but recent papers showed that ACC might act as a tumor suppressor [38,40]. ACC1 may protect mice of AML models from leukemia death [38]. Our results also support that ACC might be a tumor suppressor, for the expression of ACC was reduced in NALM-6/ HDR obviously with no change in NALM-6/H. FASN, the only human lipogenic enzyme available for de novo FAS, has been widely reported to promote cancer progression [41,42]. Upon hypoxic stress, different cancer cells had different cell-type-specific regulation of FASN, increased, decreased or unaffected [42]. A recent paper reported that FASN was upregulated in relapsed ALL patients; moreover, FASN inhibition may help to overcome glucocorticoid resistance [43]. FASN inhibition may sensitize AML cells to differentiation therapy [13]. On the contrary, our results showed that hypoxic stress induced FASN expression in NALM-6/H; however, FASN was reduced in GC resistant NALM-6/HDR cells. To be used in metabolic pathways, fatty acid must be activated by ACSLs, which convert free fatty acid to acyl-CoA [12]. ACSL1 may play a potential oncogenic role in colorectal and breast cancer and play a potential tumor suppressor in lung cancer [44]. AceCS1 catalyzes the conversion of acetate to CoA to acetyl-CoA and promote the FAS. Cancer cells display an up-regulated capability to utilize acetate [45]. Our results showed that ACSL1 and AceC-S1were reduced in both NALM-6/HDR and NALM-6/H, which might be a stress reaction to hypoxia. Therefore, the results on GC resistant NALM-6/HDR and GC sensitive NALM-6/H may raise some new points about the roles of de novo FAS in relapsed/refractory leukemia. The relative low expression of ACLY, ACC, and FASN in GC-resistant NALM-6/HDR cells pointed out that targeting FAS might not a good idea for certain patients of relapsed/refractory leukemia.
Reprogramming of carbohydrate metabolism, particularly increased glycolysis, has been recognized as a hallmark of cancer [46] and plays a key role in cancer progression in different cancers [8,10,47], which forms the basis of tumor imaging by positron emission tomography. Targeting glycolytic metabolism has become a potent therapeutic strategy to suppress cancer progression [14,48]. Hypoxic stress may influence the energy metabolism and then induce glycolysis, which may contribute to chemoresistance [14,31,49]. Our results showed that hypoxic stress did not induce the expression of AMPK and p-AMPK. On the contrary, the expression ratio of p-AMPK and AMPK reduced in NALM-6/HDR and NALM-6/H. AMPK plays a paradoxical dual role in ALL survival and progression [49][50][51]. Here, decreased expression of AMPK and p-AMPK might contribute to GC resistance in NALM-6/HDR. NALM-6/HDR showed a decreased HKII and an increased LDHA in vitro or in vivo. NALM-6/H showed an increased Glut-1, LDHA and p-LDHA. These results suggest that AMPK and glycolysis pathway may not be a therapeutic target in GC resistant NALM-6/HDR. A recent paper reported that cancer cells preferred to uptake glutamine while myeloid cells in tumor microenvironment (TME) took up the most glucose [52]. Therefore, targeting glycolysis may enhance or impair tumor-related inflammation by disrupting the metabolic reprogramming of immune cells in TME, but not cancer cells [52]. A program of mTORC1-driven uptake of glutamine may suppress the expression of glycolysis-related genes and glucose metabolism in cancer cells [52]. We observed mTOR pathway was inhibited in NALM-6/HDR cells. Chronic and prolonged hypoxic exposure results in repression of mTORC1 signaling [53]. Initially, most researchers reported that mTOR signaling activation may contribute to chemoresistance and ALL progression [17,23,54]. On the contrary, a recent paper showed that activation of mTORC1 pathway may help suppress the drug resistance of T-ALL in hypoxic niches [55]. It is intriguing that while AMPK functions as an antagonist of mTOR [56]; both AMPK and mTOR signaling pathways were inhibited in NALM-6/HDR cells, which is consistent with the findings in CEM-C7/ HDR [6]. The orchestration of the cellular metabolism is not conducted by a single leader; the complex interplay between AMPK, mTOR, and other signaling network help to meet the conflicting self-requirement cancer cells [56]. Therefore, the relative low expression of AMPK and mTOR pathway is a distinct feature of NALM-6/HDR, which might be another key point to understand the mechanisms of GC resistance in ALL [6], and NALM-6/ HDR may provide a novel model that represents a subtype of GC-resistant B-ALL cells.