Identification of galectin-1 as a novel mediator for chemoresistance in chronic myeloid leukemia cells

Multidrug resistance protein-1 (MDR1) has been proven to be associated with the development of chemoresistance to imatinib (Glivec, STI571) which displays high efficacy in treatment of BCR-ABL-positive chronic myelogenous leukemia (CML). However, the possible mechanisms of MDR1 modulation in the process of the resistance development remain to be defined. Herein, galectin-1 was identified as a candidate modulator of MDR1 by proteomic analysis of a model system of leukemia cell lines with a gradual increase of MDR1 expression and drug resistance. Coincidently, alteration of galectin-1 expression triggers the change of MDR1 expression as well as the resistance to the cytotoxic drugs, suggesting that augment of MDR1 expression engages in galectin-1-mediated chemoresistance. Moreover, we provided the first data showing that NF-κB translocation induced by P38 MAPK activation was responsible for the modulation effect of galectin-1 on MDR1 in the chronic myelogenous leukemia cells. Galectin-1 might be considered as a novel target for combined modality therapy for enhancing the efficacy of CML treatment with imatinib.


INTRODUCTION
Formation of the BCR-ABL fusion gene coding for a constitutively active BCR-ABL tyrosine kinase via t(9;22)(q34;q11) reciprocal translocation initiates 95% of chronic myelogenous leukemia (CML), and 25% of adults and 5% of children acute lymphoblastic leukemia (ALL) [1]. Imatinib provides a promising treatment for CML by high selectively binding to the ATP-binding site of BCR-ABL and inhibiting BCR-ABL activation [2]. However, the occurrence of drug resistance was reported in CML patients with advanced stages treated with imatinib [3]. Amplification of the BCR-ABL gene and mutations of the kinase domain of ABL have been described as the molecular mechanisms for the development of imatinib resistance [4][5][6][7][8][9]. However, overexpression or mutations of BCR-ABL could not explain all drug resistance to imatinib in CML patients, implying that the alternative mechanisms may exist [10][11][12].
MDR1, the ABCB1 gene product, is an ABC transporter at the cell surface responsible for extruding the compounds out of the cell, and has the potentials of mediating multiple drug resistance (MDR) by reducing intracellular drug concentrations [13]. Previous investigations showed that imatinib is the substrate of MDR1 and considered drug efflux mediated by MDR1 as a causal role for imatinib drug resistance in CML [14,15]. But the precise mechanisms of MDR1 modulation in chronic myelogenous leukemia cells remain to be unclear.
In the current study, we first applied proteomic approach to identify galectin-1 as a candidate of MDR1 modulators for mediating drug resistance in CML cells by comparison of the protein profiles among a model system of leukemia cell lines with a gradual increase of MDR1 expression and drug resistance, and further explored the mechanisms of galectin-1 acting as a novel MDR1 modulator contributing to functional resistance against the cytotoxic drugs.

Characterization of the MDR phenotype in K562, K562/ADM and the revertant K562/ADM cells
Initially, the sensitivity profiles against adriamycin and imatinib were explored in a model system of cell lines including K562, K562/ADM and the revertant K562/ADM cells. As shown in Table 1, the resistant cell line K562/ADM displayed higher resistance against adriamycin and imatinib, with 50-fold and 5-fold increase of IC 50 for adriamycin and imatinib respectively, than its sensitive counterpart K562 cells. We observed that the revertant K562/ADM cell line showed less resistance than K562/ADM cell line but higher resistance than K562 cell line against both adriamycin and imatinib, suggesting the resistant cells gradually lose the resistant character when cultured in the absence of the chemical compound.
Quantitative PCR (q-PCR) analysis showed that the expression level of MDR1 in the revertant K562/ ADM cells is less than the resistant K562/ADM cells, but higher than the sensitive ones, suggesting that the expression level of MDR1 gradually decreases during the course of cultivation in the absence of adriamycin ( Figure  1). The results imply that a gradual increase of resistance against adriamycin and imatinib is accompanied by a gradual increase of MDR1 level during the course of development of drug resistance in K562 cells exposed to adriamycin.

Identification of galectin-1 as a significantly up-regulated protein in resistant K562 cells by 2D-PAGE and MALDI-TOF-TOF mass spectrometry
In order to elucidate the mechanisms of MDR1 modulation in K562 cells, a proteomic approach was initially applied to identify differentially expressed proteins among three types of K562 cells with different MDR1 expression and potentials of drug resistance. As shown in Figure 2, the well-resolved, reproducible 2D-PAGE patterns of K562 (Figure 2A), K562/ADM ( Figure 2B), and the revertant K562/ADM cells ( Figure 2C) were established, and yielded about 1000 protein spots each. In total, 11 protein spots were found to be differentially expressed among the investigated cells. All of these were excised and analyzed by MALDI-TOF-TOF mass spectrometry and a subsequent search in the IPI databases for protein identification. The identification information was summarized in Table 2.
Among the identified differentially expressed proteins, six of them were significantly up-regulated in K562/ADM cells compared with K562 cells, and significantly down-regulated in the revertant K562/ ADM cells compared with K562/ADM cells. Whereas, the expression levels of the other five identified proteins in the revertant K562/ADM cells were less than those in K562 cells, and higher than those in K562/ADM cells ( Figure 3). The differentially expressed proteins related to cell behaviors, metabolism, calcium-binding, proteolysis, cellular transcription, and signal transduction maybe valuable for further elucidating the chemoresistant mechanisms in CML.
Among the identified proteins, galectin-1 (Gal-1), which exerts effects on cell apoptosis, proliferation and differentiation, increased 4.85 folds in K562/ADM cells compared with the revertant K562/ADM cells, and upregulated 22.3 folds compared with K562 cells, implying that galectin-1 may contribute to augment of MDR1 expression and drug resistance in CML.

Verification of galectin-1 expression
In order to confirm the trends of the expression levels of galectin-1 identified by the proteomic approach, q-PCR method was first applied to measure the mRNA levels of galectin-1. The results present in Figure 4A indicated that the mRNA level of galectin-1 in K562/ADM was 1.87 folds higher than the revertant K562/ADM cells, and 4.4 folds than K562 cells, respectively. Further western blot analysis revealed a significant up-regulation of galectin-1 expression in K562/ADM cells compared with the other two types of cells with less drug resistance, which were paralleled to the protein level changes observed in the proteomic analysis ( Figure 4B).
As the expression levels of galectin-1 increased with the augment of drug resistance, and the resistant cell lines were selected by exposure to adriamycin, we speculated it may be the chemical drugs that trigger the increase of galectin-1 product. Further investigations indicated that both adriamycin and imatinib up-regulated galectin-1 expression in a dose dependent manner ( Figure 4C and 4D).
We next evaluated the effects of galectin-1 siRNA targeting treatment on the chemoresistance in K562/ ADM cells. As shown in Figure 5C and 5D, treatment of K562/ADM cells with 50 nM galectin-1 siRNA induced down-regulation of galectin-1 expression at the mRNA and protein levels. Knockdown of galectin-1 in K562/ADM cells could increase the chemosensitivity to both adriamycin and imatinib treatment, with the IC 50 of adriamycin decreasing from 26.56±0.41 μg/ml to 14.76±0.45 μg/ml, and that of imatinib from 2.5±0.214 μM to 1.1±0.12 μM as present in Table 3.

Galectin-1 is a modulator of MDR1
As shown by the above results, the trends of galectin-1 expression change are parallel to those of MDR1 in three types of cells with a gradual increase of resistance against adriamycin and imatinib. In the light of MDR1 directly contributing to the drug resistance by exporting drugs out of the cell, combined with our results showing that galectin-1 enhances the chemoresistance in CML, we proposed that galectin-1 may decrease the chemosensitivity via increase of MDR1 expression. To address this issue, q-PCR was applied to compare the mRNA levels of MDR1 between K562/pc and K562/Gal-1 cells, and between K562/ADM/sc siRNA and K562/ADM/ Gal-1 siRNA. The results present in Figure 6A indicated that overexpression of galectin-1 up-regulated MDR1 transcription, while knockdown of galectin-1 decreased the mRNA levels of MDR1. Alteration of galectin-1 expression had little effect on the other well-studied drug transporter MRP1 ( Figure 6B). The results suggest that augment of MDR1 expression involves in galectin-1mediated chemoresistance.

Galectin-1 induces MDR1 expression via P38 MAPK activation and NF-κB translocation
It has been investigated that NF-κB transcription factor regulates MDR1 expression by specifically binding to an intronic response element of MDR1 gene promoter [16]. In the light of MAPK signal pathways mediating nuclear translocation of NF-κB, which is required for its transcription activity, and galectin-1 directly interacting with Ras, which subsequently activates Erk1/2 and P38 mitogen-activated protein kinase (MAPK) signal pathways [17][18][19], we proposed that the mechanisms of galectin-1 inducing MDR1 expression may involve MAPK signal activation and NF-κB translocation. In order to define the signal pathway participating in mediating galectin-1-induced NF-κB translocation, and subsequent MDR1 transcription, we first detected the effects of galectin-1 on the activation of Erk1/2 and P38 MAPK cascades. As shown in Figure 7A, overexpression of galectin-1 increased the phosphorylation levels of P38 rather than Erk1/2, implying that activation of P38 MAPK signal pathway may be responsible for mediating galectin-1 functions in K562 cells. Further suppression of P38 activation with the P38 inhibitor SB202190 attenuated NF-κB translocation ( Figure 7B), MDR1 gene promoter activity ( Figure 7C), and MDR1 mRNA levels ( Figure  7D). Moreover, blockade of P38 activation in K562/Gal-1 cells decreased the chemoresistance to both adriamycin and imatinib treatment, with the IC 50 of adriamycin reducing from 1.636±0.141 μg/ml to 1.29±0.125 μg/ml, and that of imatinib from 0.8±0.083 μM to 0.4±0.037 μM. Similar results were found in K562/ADM cells treated with SB202190 (Table 3). Combined with the above results, we suggest that galectin-1/P38 MAPK/NF-κB/

DISCUSSION
Imatinib has been proved to be an efficient target drug with few adverse consequences for CML treatment. However, patients initially displaying positive response to imatinib become refractory to imatinib treatment due to the development of drug resistance. It has been found that MDR1 expression was more frequent in both advanced CML and acute myeloid leukemia (AML) patients [20,21]. Mahon F et al. reported overexpression of ABCB1 gene in leukemia cells results in resistance to imatinib [14]. Therefore, besides BCR-ABL amplification and mutation, decreased intracellular levels of imatinib caused  by MDR1 up-regulation is also considered as a major reason leading to imatinib resistance [15]. However, the mechanisms by which the leukemia cells increase the expression of this transmembrane export pump are still unknown. We proposed that differentially expressed proteins identified by comparison of the protein profiles among a model system of the leukemia cells with different MDR1 expression levels may provide a hint for revealing the mechanisms modulating the expression of MDR1. K562/ADM cells selected by adriamycin exposure produce more abundant MDR1 than the parental K562 cells. Cultivation of K562/ADM cells in the condition without adriamycin selection for three consecutive months led to obtain the revertant K562/ADM cells with a  Galectin-1 has a wide distribution in tissues and is modulated accompanying with the alteration of the physiological or pathological conditions. Accumulating evidences have demonstrated that galectin-1 is closely relevant to the tumor progression by promoting transformation, angiogenesis, and metastasis [22][23][24][25]. However, there are few reports exploring the role of galectin-1 in tumor chemoresistance. The endogenous expression levels of galectin-1 vary in different kinds of tumors. Some cancerous cells with high levels of endogenous galectin-1 expression, such as nonsmall cell lung cancer (NSCLC) cells, have relative high potentials of anti-chemotherapy. Suppression of galectin-1 sensitizes the NSCLC cells to platinumbased chemotherapy [19]. In our studies, we found the expression levels of the endogenous galectin-1 are pretty low in K562 cells, whereas significant upregulation of galectin-1 is observed in the corresponding chemoresistant K562/ADM cells. Moreover, treatment of K562 cells with adriamycin or imatinib could increase galectin-1 expression, suggesting that tumor cells may gradually develop drug resistance along with the chemotherapy. These results imply that galectin-1 might be a potential biomarker for the chemoresistant ability of tumor cells, and provide a novel target for combined therapy for enhancing the efficiency of the chemical drugs.  Table 3. Data are presented as the mean (±SD) of three independent experiments. *p<0.05, **p<0.01, ***p<0.001. www.impactjournals.com/oncotarget   Promotion of apoptosis by galectin-1 in the activated T cells has been extensively investigated in the past two decades [26,27]. The molecular mechanisms underlying galectin-1 inducing apoptosis involve suppression of anti-apoptotic protein expression and stimulation of caspases [28]. As one mechanism by which tumor cells develop drug resistance is associated with resistance to apoptosis, it seems paradoxical that galectin-1 on one hand induces apoptosis, and on the other hand contributes to drug resistance. We initially presumed that galectin-1 may exert an oppose effect on the apoptosis in K562 cells as compared to the activated T cells. Therefore, flow cytometric analysis was applied to determine the effects of galectin-1 on the apoptosis of K562 cells. Contrary to our speculation, galectin-1 significantly promoted apoptosis of K562 cells treated with or without adriamycin (data not shown). It seems that apoptosis induced by galectin-1 contributes little to the galectin-1-triggered drug resistance. In the light of reports suggesting that galectin-1 does not activate a full apoptotic pathway in some kinds of cells including T cells [29], T leukemia cells, promyelocytic cells and activated neutrophils [30], the possible explanation may be that K562 cells induced by galectin-1 undergo partial apoptotic process, which has less relevance with drug resistance than the full apoptotic process does. However, further investigations are required to clarify the relationship between galectin-1-elicited apoptosis and drug resistance.
In summary, we found that galectin-1 was a novel modulator of MDR1 by proteomic analysis of a model system of leukemia cell lines with a gradual increase of MDR1 expression and drug resistance, and NF-κB translocation induced by P38 MAPK activation was responsible for the enhancing effects of galectin-1 on MDR1, suggesting that galectin-1 might be a novel target for improving the efficacy of CML chemotherapy.

2D-PAGE and image analysis
An equal amount (1 mg) of protein sample was mixed with rehydration buffer complemented with 7 M urea, 2 M thiourea, 4% CHAPS, 65 mM DTT, 0.2% Bio-Lyte 3-10 ampholyte and 0.001% bromophenol blue, and loaded on a 17 cm, pH3-10 nonlinear immobilized pH gradient (IPG) gel strip (Bio-Rad). The IPG strips were then passively rehydrated at 20°C for 13 h, and subjected to isoelectric focusing (IEF) performed at 20 °C at 100 V for 30 min, 150 V for 3 h, 250 V for 1 h, 500 V for 1 h, 1000 V for 2 h, 5000 V for 3 h, 8000 V for 64000 V/h, and 500 V for 24 h. After IEF separation, the strip was equilibrated, and separated on 12% SDS-PAGE gels as previously described [32]. The gel was stained with Coomassie brilliant blue G250 (Bio-Rad), and scanned with a UMAX POWERLOOK 2100XL USB scanner (UMAX, Dallas, TX, USA). PDQuest 8.0 software (Bio-Rad) was used to analyze the images.

Mass spectrometric analysis and database search
In-gel digestion of the differentially expressed proteins was carried out before mass spectrometric analysis. Briefly, the protein spots were excised, destained in 25 mM ammonium bicarbonate 50% NH 4 HCO 3 / acetonitrile (ACN) (v/v), dehydrated in 100% ACN, and then incubated with trypsin at 37 °C overnight. The peptides were extracted, dried in a vacuum concentrator for 3 h, and subjected to tandem time-of-flight mass spectrometry (ABI 4800 TOF-TOF) analysis. Mascot software (Matrix science, London, UK) was applied to search IPI (International Protein Index) databases for protein identification.

Construction of K562/Gal-1 cell line stably overexpressing galectin-1
The full-length galectin-1 coding sequence was amplified from cDNA synthesized by reverse transcriptase polymerase chain reaction (RT-PCR) using the total RNA extracted from K562/ADM cells as the template. Briefly, cDNA was synthesized via reverse transcription using the oligo dT 18 , and subjected to PCR amplification of the full-length galectin-1 coding sequence (NCBI accession no. NM_002305.3). The PCR procedure comprised of an initial step at 94 °C for 5 min, followed by 40 cycles of 94 °C for 30 s, 58 °C for 15 s, and 72 °C for 1 min, and an additional extension at 72 °C for 30 min. The following primers were used for galectin-1 amplification: 5'-AG CGGATCCATGGCTTGTGGTCTGGTC-3' (forward, BamHI site underlined) and 5'-TATAAGCTTTCAGTCA AAGGCCACACA-3' (reverse, HindIII site underlined). The PCR products were purified by the TIANGEN purification kit (TIANGEN, Beijing, China), digested with BamHI and HindIII, and inserted into the pcDNA3.1(-) vector (Invitrogen, Carlsbad, CA, USA) to obtain the recombinant plasmid pcDNA3.1(-)/Gal-1, which was subsequently sent to Sangon Company (Shanghai, China) for sequencing.
The recombinant plasmid pcDNA3.1(-)/Gal-1 harboring the correct galectin-1 coding sequence and the control vector pcDNA3.1(-) were transfected into K562 cells with lipofectamine LTX (Invitrogen) following the manufacturer's protocol. Briefly, 3×10 4 K562 cells seeded in an individual well of a 24-well culture plate were transfected with 1 μg of plasmid DNA. After transfection for 48 h, cells were selected in complete RPMI-1640 media containing G418 (500 μg/ml) for 2 weeks, and subsequently replated (10 cells/well) in 96-well culture plate for continuous G418 selection. After cultured for 10 days, individual G418-resisitant colonies were picked, propagated and screened for K562/Gal-1 cell clone stably expressing galectin-1 by q-PCR and western blotting. The sequences of galectin-1 and GAPDH primers for q-PCR were shown in Table 4. GAPDH was used as an internal control.

Suppression of galectin-1 expression by small interfering RNA
A small interfering RNA (siRNA) targeting galectin-1 (Gal-1 siRNA) [19] with the sequence of sense 5'-GC UGCCAGAUGGAUACGAAUUdtdt-3', and antisense 5'-AAUUCGUAUCCAUCUGGCAGCdtdt' was synthesized by RiboBio (Guangzhou, China). A scrambled siRNA (sc siRNA) obtained from RiboBio was used as a negative control. For siRNA-mediated inhibition of galectin-1 gene expression, K562/ADM cells were transfected with Gal-1 siRNA or sc siRNA at a final concentration of 50 nM using RNA MAX siRNA Transfection Reagent (Invitrogen) according to the manufacturer's instructions. Silencing efficiency was estimated at mRNA and protein levels by q-PCR and western blotting.

Quantitative PCR analysis
Total RNA was isolated with PureZOL according to the manufacturer's instructions. A first-strand cDNA synthesis kit was applied to produce cDNAs from 1μg of total RNA, which were then used as templates for q-PCR amplification with the SYBR green q-PCR Kit using the CFX96 Touch™ Real-Time PCR Detection System. The primers of galectin-1, MDR1, MRP1 and GAPDH were showed in Table 4. GAPDH was amplified as an internal control. The q-PCR conditions were 94 °C for 5 min followed by 40 cycles of 95 °C for 5 s, 59 °C for 20 s.

Western blotting analysis
Cells were harvested and lysed in 1 × SDS-PAGE loading buffer, then centrifuged at 12,000 g for 30 min at 4 °C to remove the insoluble components. The resultant protein samples were resolved by 10% SDS-PAGE gel and transferred to a PVDF membrane. The membrane was blocked at room temperature for 1 h in TBST (25 mM Tris, pH 7.4, 150 mM NaCl, and 0.1% Tween-20) containing 5% non-fat dry milk, and subsequently incubated with the primary antibodies at 4 °C overnight followed by incubation with goat anti-rabbit or rabbit anti-mouse IgG, HRP-linked antibody for 1 h at room temperature. The blots were detected with an ECL detection kit (Pierce) according to the manufacturer's procedure. GAPDH was used as the reference control. The results were analyzed by Quantity One software to determine the ratio relative to GAPDH.

Extraction of nuclear proteins
Cells were seeded in 6-well culture plates at a density of 2.5×10 5 cells per well. After cultivation for 6 h, K562/Gal-1 cells were treated with P38 inhibitor SB202190 at the concentration of 10 μM for 12 h, 24 h, and 48 h before extraction of the nuclear proteins using the Nuclear and Cytoplasmic Protein Extraction Kit (KeyGEN Biotech, Nanjing, China) according to the manufacturer's instructions. The protein concentration was determined by a BCA protein assay kit (Pierce). The nuclear proteins (NF-κB/p65 and PCNA) were analyzed by western blotting.