The RNA/DNA-binding protein PSF relocates to cell membrane and contributes cells' sensitivity to antitumor drug, doxorubicin


  • Simei Ren,

    1. Department of Hematology, National Center for Clinical Laboratories and Beijing Hospital of the Ministry of Health, Beijing, China
    Search for more papers by this author
  • Ming She,

    1. Department of Pharmacy, State Key Laboratory of Experimental Hematology, Institute of Hematology and Hospital of Blood Diseases, Chinese Academy of Medical Sciences & Peking Union Medical College, Tianjin, China
    Search for more papers by this author
  • Min Li,

    1. Department of Hematology, National Center for Clinical Laboratories and Beijing Hospital of the Ministry of Health, Beijing, China
    Search for more papers by this author
  • Qi Zhou,

    1. Department of Hematology, National Center for Clinical Laboratories and Beijing Hospital of the Ministry of Health, Beijing, China
    Search for more papers by this author
  • Rong Liu,

    1. Department of Pharmacy, State Key Laboratory of Experimental Hematology, Institute of Hematology and Hospital of Blood Diseases, Chinese Academy of Medical Sciences & Peking Union Medical College, Tianjin, China
    Search for more papers by this author
  • Hong Lu,

    1. Department of Hematology, National Center for Clinical Laboratories and Beijing Hospital of the Ministry of Health, Beijing, China
    Search for more papers by this author
  • Chunzheng Yang,

    1. Department of Pharmacy, State Key Laboratory of Experimental Hematology, Institute of Hematology and Hospital of Blood Diseases, Chinese Academy of Medical Sciences & Peking Union Medical College, Tianjin, China
    Search for more papers by this author
  • Dongsheng Xiong

    Corresponding author
    1. Department of Pharmacy, State Key Laboratory of Experimental Hematology, Institute of Hematology and Hospital of Blood Diseases, Chinese Academy of Medical Sciences & Peking Union Medical College, Tianjin, China
    • Correspondence to: Dongsheng Xiong, Department of Pharmacy, Institute of Hematology and Hospital of Blood Diseases, Chinese Academy of Medical Sciences, 288 Nanjing Road, Tianjin 300020, China. E-mail:

    Search for more papers by this author


Cell surface proteins play an important role in multidrug resistance (MDR). However, the identification involving chemoresistant features for cell surface proteins is a challenge. To identify potential cell membrane markers in hematologic cancer MDR, we used a cell- and antibody-based strategy of subtractive immunization coupled with cell surface comparative screening of leukemia cell lines from sensitive HL60 and resistant HL60/DOX cells. Fifty one antibodies that recognized the cell surface proteins expressed differently between the two cell lines were generated. One of them, the McAb-5D12 not only recognizes its antigen but also block its function. Comparative analysis of immunofluorescence, flow cytometry, and mass spectrum analysis validated that the membrane antigen of McAb-5D12 is a nucleoprotein—polypyrimidine tract binding protein associated splicing factor, PSF. Our results identified that PSF overexpressed on the membrane of sensitive cells compared with resistant cells and its relocation from the nuclear to the cell surface was common in hematological malignancy cell lines and marrow of leukemia patients. Furthermore, we found that cell surface PSF contributed to cell sensitivity by inhibiting cell proliferation. The results represent a novel and potentially useful biomarker for MDR prediction. The strategy enables the correlation of expression levels and functions of cell surface protein with some cell-drug response traits by using antibodies. © 2013 International Society for Advancement of Cytometry

The cell surface protein may play an important role in the cancer multidrug resistance (MDR), which has been one of the main causes of failure of the primary chemotherapy for hematologic neoplasms [1, 2]. In many studies, the MDR phenotypes have been associated with the changes of various cell-surface proteins, which serve as active transporters and control the cellular distribution of drugs. The most common changes of cell-surface protein include varied expression levels [3], post-translational modifications [4], and appearance of abnormal proteins [5]. The growing evidences also suggest that a single membrane molecular changes cannot explain all the clinical and laboratory MDR, even for those classical membrane mechanisms such as P-glycoprotein and other ABC transporters [6, 7]. In addition, membrane proteins are fit for therapeutic and diagnostic targets because of their accessible feature. Therefore, discovering novel and potential membrane targets seems primary, which will contribute not only to novel MDR mechanism researches but also to screening of candidate targets.

However, the isolation and characterization of the cell-surface proteins which are usually hydrophobic and low-abundant are challenging tasks even for the experienced researchers using the direct differential proteomics analysis [8]. To identify the membrane proteins which could serve as the potentially therapeutic and prognostic targets, the cell- and antibody-based subtractive immunization coupled with novel proteomic mass spectrometry (MS) sequencing was applied to compare the differentially expressed cell-surface proteins involved in the MDR [9, 10]. In addition, with the combination of the immunosuppressant cyclophosphamide and the sequential immunization of two cell types with highly similar genetic backgrounds but different phenotypes with respect to MDR potential, this approach is able to generate monoclonal antibodies capable of distinguishing cell-surface proteins that may play a functional role in MDR [11-13]. Especially, the monoclonal antibodies and their immunogens generated by this approach may provide novel clinically relevant reagents and/or tools for cancer immunotherapy and diagnosis, because these antibodies normally are not commercially available [14-16].

Two isogenic cell lines with different MDR phenotypes were used in this study. The sensitive HL60 cell line was derived from the peripheral blood leukocytes of a patient with acute promyelocytic leukemia and displayed distinct morphological and histochemical commitment toward myeloid differentiation [13, 17]. The MDR cell line, HL60/DOX established by Koizumi et al. [18], was cultured by continuous exposure of doxorubicin (DOX) and characterized by a significant resistance to DOX and cross-resistance to several agents such as vincristine and etoposide. The overexpression of MRP and the enhanced drug efflux are considered as the primary but not the only mechanism of DOX resistance in HL60/DOX subclone [18].

This study was designed to compare the differentially expressed cell surface proteins in the pair of cell lines and followed by the functional analysis of target proteins involved in MDR with the antibody generated by subtractive immunization. One of the antibodies generated by our methodology, 5D12, bound to a relocated nuclear protein, polypyrimidine tract binding protein associated splicing factor (PSF), which is first found to be located on cell surface of many hematological malignances cell lines, and overexpressed on cell surface of sensitive cell lines compared to the resistant cell lines. Blocking of cell surface PSF through monoclonal antibody (McAb) 5D12 resulted in increased cell proliferation and more distinct resistance phenotype toward DOX. Our results suggest that membrane PSF plays a significant role in supporting cell sensitivity. PSF and the antibody 5D12 are of great value to potential prediction in MDR and crucial to adjust the medication regimens to avoid the fail treatment of cancer chemotherapy.

Materials and Methods

Cell Lines

HL60 and the DOX MDR subline HL60/DOX were kindly provided by Dr. Masahiro Imamura (Hokkaido University School of Medicine, Hokkaido). All cell lines were maintained in RPMI1640 medium (Life Technologies, Grand Island, NY) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (GIBCO-BRL, Beijing, China). Cells were grown in a humidified 5% CO2 atmosphere at 37°C. The HL60/DOX MDR cell line has significant drug resistance after long-term DOX induction and cloning selection.

Subtractive Immunization and Hybridoma Production

BALB/c female mice, 5–6-week-old, were obtained from the animal center of the Institute of Hematology (Tianjin, China). Experiments with animals were approved by Ethical Committee of the Chinese Academy of Medical Sciences, and measure taken to protect animals from pain or discomfort. Subtractive immunization was performed as follows: On day 1, HL60 single cell suspensions were washed and resuspended in PBS. HL60 cells (5 × 106) in 300 mL sterile PBS were inoculated intraperitoneally (i.p.) into the BALB/c mice. After 20 min, 24 h, and 48 h, mice were injected i.p. with 100 mg/kg cyclophosphamide (Sigma-Aldrich, St Louis, MO) in sterile PBS. On days 15 and 29, immunization and tolerization were repeated twice as described above. On day 41, HL60/DOX cells (5 × 106) were harvested and injected i.p. into the mice as described above. The above procedures were repeated three times at a 14-day internal. On day 83, the mice received an i.p. boost of 1 × 107 HL60/DOX cells, and on day 87, the mice were sacrificed for the production of hybridomas. The hybridoma culture supernatants were screened by a FACSCalibur equipped with CellQuest software (BD Bioscience, San Jose, CA).

Flow Cytometry and Comparative Screening of Hybridomas

In total, 5 × 105 cells were incubated for 1 h with 70 μL of hybridoma tissue culture supernatant or with antibody in 200 μL of PBS containing 2 μg/mL mAb. Cells were then washed twice and incubated with FITC-conjugated goat antimouse IgG (Institute of Hematology, Tianjin, China) for 30 min at room temperature. The stained cells were washed twice in PBS, and resuspended in 500 μL PBS, and analyzed using flow cytometry. Sera from nonimmunized mice were used as negative controls. Hybridoma clones producing antibodies that showed preferential binding to HL60 or HL60/DOX cells were selected and subcloned three times by limiting dilution.

Monoclonal Antibody Purification

For large-scale antibody production, hybridoma clones of interest were expanded and intraperitoneally injected into BALB/c mice. Ascites were harvested, and antibodies were affinity-purified using a Protein G-coupled Sepharose column according to the manufacturer's protocol (Amersham Bioscience, Piscataway, NJ).

MTT Growth Inhibition Assay

HL60 and HL60/DOX cells were seeded in 96-well formats (Corning, NY) and treated the next day with increasing concentrations of DOX (3.5–35 μM, HL60/DOX; 0.04 μM–0.4 μM, HL60); 50 ng antibody was added as well. After 72 h of incubation, the amount of viable cells was determined using MTT assay (Life Technologies). Percentage of cell survival as a function of DOX concentration was then plotted to determine the IC50 value.

Immunofluorescence and Confocal Microscopy

Cells were washed with PBS and then fixed for 10 s with methanol–acetone (1:1) on eight-well glass slides (Labteks, NUNC, Naperville, IL). To stain the cell membrane, the cells were not penetrated and incubated with 2 μM CellTracker CM-Dil (Life Technologies) for 15 min at a minimum of 37°C and then for an additional 15 min at 4°C. After having been washed twice with PBS, the cells were blocked for 20 min with PBS containing 10% FBS. The cells were then further incubated with either clone 5D12 antibody or B92 (a commercially available anti-PSF monoclonal antibody; Sigma) for 1 h at 37°C, followed by incubation with an FITC-conjugated goat anti-mouse secondary antibody for 30 min at room temperature. Finally, DAPI (Sigma) was applied at a final concentration of 1.5 μM in PBS for 10 min to counterstain the nuclei. All images were obtained using a confocal fluorescence microscope (Leica, Nussloch, Germany).


Cells were washed in cold PBS and resuspended in RIPA buffer (Sigma). The cell lysates were pretreated by isotype control IgG and protein A agarose as described previously [19]. Then, 10 μg McAb 5D12 was first incubated with 600 μL of total cell lysate (about 800 μg) with gentle shaking overnight at 4°C, followed by 10 μL of protein A agarose slurry (Life Technologies) at 4°C for 2 h with gentle shaking. The beads were pelleted and washed four times, and the bound proteins were then eluted by adding 50 μL of SDS-PAGE sample buffer and boiling for 5 min. These samples were separated by 12% SDS-PAGE, and Western blotting was performed.

Membrane Protein Extraction and Western Blotting

Cells were washed in cold PBS and resuspended in 10 mM Tris-HCl (pH 7.5) with the protease inhibitor cocktail (Roche, Lewes, UK). Membrane proteins were extracted with Triton X-114 extraction buffer as described [20, 21] For immunoblot analysis, lysates containing 10 μg of membrane protein were resolved by SDS-PAGE, transferred to nitrocellulose membranes (Millipore, Billerica, MA), and blocked overnight at 4°C with 5% nonfat dry milk. The membrane was incubated with primary antibody 5D12 at room temperature, followed by horse radish peroxidase (HRP)-conjugated secondary antibodies. All signals were visualized using a standard 3,3-diaminobenzidine tetrahydrochloride (Sigma) staining according to the instructions of the manufacturer.

MALDI-TOF Mass Spectrometry

The immunoprecipitated proteins were separated by 12% SDS-PAGE gel and visualized with Coomassie Brilliant Blue staining. Bands corresponding to a prominent differentiation between HL60 and HL60/DOX cells were excised. The excised gel pieces were prepared as described previously [19]. Trypsinization was stopped with 5% trifluoroactic acid, and the generated tryptic peptides were analyzed by a 4700 Proteomics Analyzer (Applied Biosystems, Foster, CA) operated in positive ion reflector mode with a m/z range of 700–4,000. α-Cyano-4-hydroxycinnamic acid (Sigma) was used as the matrix. Spectra were obtained by averaging 1,000 acquired spectra in MS mode. The peak list of the spectra was created by the Peak-to-MASCOT script of the 4000 series explorer software 3.0 (Applied Biosystems). The samples were analyzed by peptide mass fingerprinting by comparing to the SwissProt 56.6 database (released on December 28, 2008; 405,506 sequences, 146,166,984 residues) using the Mascot algorithm of the GPS Explorer 3.5 software (Applied Biosystems). The search criteria permitted one missed cleavage, at least four matching peptide masses, a mass tolerance of <100 ppm, and a minimum signal-to-noise ratio of 10.0. Variable modifications, including carbamidomethyl modification and oxidation, were assigned to cysteine and methionine residues, respectively. Significant matches were identified based on an expected Mascot value of <5% for protein identification, which is equivalent to statistically significant search scores of >95%. Finally, the molecular weights of the proteins were compared to the values estimated by gel electrophoresis.

RNA Interference and RT-PCR

PSF (NCBI ID: gi│1709851) siRNA with specific sequence 5′-GACGACAGGAAGAAUUAAGTT-3′ and another nonspecific control siRNA 5′-UUCUCCGAACGUGUCACGUTT-3′ were designed on iTTM RNAi Designer (Life Technologies) and synthesized by GENE PHARMA (Shanghai, China). HL60 cells (2.5 × 105 per well) were cultured in six-well plates for 24 h and transfected with 100 nM of the PSF siRNA per well using Lipofectamine 2000 (Life Technologies) according to the manufacturer's instructions. Control cells were incubated with transfection reagent and the control nonspecific siRNA. To confirm subsequent PSF knockdown, RNA was isolated from samples using TRIzol reagent (Life Technologies) and reversely transcribed using SuperScript III reverse transcriptase (Life Technologies). Human psf gene-specific primers were used: F, 5′-GAGATCCCTATGGTTCAGGAG-3′ and R, 5′-CTTCTCTCCCTCTACCATATCCT-3′. The thermal cycling conditions were as follows: 4 min at 95°C and 30 cycles of denaturation at 95°C for 20 s, primer annealing at 59°C for 30 s, and extension at 72°C for 30 s.

Cytotoxicity Assay

A Cell Culture Kit-8 (CCK8; Dojindo, Kumamoto, Japan) assay was performed to evaluate the change of HL60 and HL60/DOX cells' sensitivity to DOX. Briefly, cells were seeded at a density of 2 × 104 cells per well in 96-well plates and maintained overnight. 5D12 was then added to each well at different final concentrations (3, 5, and 8 μg/mL, respectively). After 5D12-treatment for 1 h, each well of cells were loaded with DOX (IC50 and IC30 concentrations, respectively) for 24 h at 37°C. Samples treated with DOX alone at IC50 and IC30 concentrations were used as positive controls, while samples treated with neither DOX nor 5D12 were used as negative controls. Cell cytotoxicity was assayed according to the CCK8-Kit's instructions. Absorbance at 450 nm was recorded in a microtiter plate reader (Model A-5082, SLT lab instrumenta, Grodig, Austria). Cells survival were represented as relative survival percentages and appeared to reflect the function of 5D12-McAb. The relative survival percentage was calculated as the following formula: relative survival (%) = OD samples/OD negative control. The experiments were done in pentaplicate and repeated three times.

Proliferation and Colony Formation Assay

HL60 cells were seeded at a density of 6,000 cells per well in 96-well formats and treated the next day with increasing final concentrations (1, 3, 5, 8, and 12 μg/mL) of 5D12-McAb. At 96 h post-treatment, 10 μL of CCK8 was added to each well for the cell proliferation assay. Abosorbance at 450 nm was recorded in a microtiter plate reader. Control cells were incubated with PBS instead of 5D12. For the colony formation assay, cells were preincubated with 5D12 for 1 h. After the incubation period, cells were plated on 35-mm dish (Costa, CA) with RPMI 1640 medium containing 10% FBS and 0.9% methyl cellulose (M0512, Sigma) at a density of 1,000 cells/dish. After 16–18 days, the cells were stained with 4-nitro blue tetrazolium chloride (BBI, Cocopaxi, CO) and scanned by Epson Perfection SJ4070 scanner as described previously [22].

Statistical Analysis

All data are presented as mean ± SD. The Student's t-test was used to determine statistically significant differences between groups. P values of ≤0.05 were considered statistically significant.


Generation of McAbs Using Subtractive Immunization and Selective Screening

To identify cell-surface proteins that may play important roles in drug resistance of HL60/DOX cells, immune tolerance to the cell surface antigens on live HL60 cells was induced in mice that were subsequently immunized with live HL60/DOX cells. The resulting hybridomas were screened for selective binding to the target cells by flow cytometry. Of 87 hybridoma supernatants, 39 clones showed exclusive binding to MDR HL60/DOX cells and 12 showed a preferential binding to HL60 cells (Fig. 1). Nine clones were selected and subcloned for further studies based on their strong specificity and high affinity to either cell line. The antibodies entitled 5F6, 6F5, 6A2 6B3, 9C11, 3B10, 2D8, 7E7 to HL60/DOX cells and 5D12 to HL60 were purified from mice ascites.

Figure 1.

Binding percentage of hybridoma antibody-based flow cytometry. Results for individual hybridoma clones are displayed. Thirty nine clones preferentially react with HL60/DOX cells and 12 clones show exclusive binding to HL60 cells.

McAb 5D12-Mediated Resistance to DOX in HL60 was Greater than that in HL60/DOX

To further characterize the chemoresistant activity of these isolated McAbs, MTT assay was performed to validate the function of antibodies to HL60 and HL60/DOX cell lines. As shown in Table 1, 72 h-IC50 DOX concentration of HL60/DOX cells is approximately 74.3-fold higher than its parental HL60 cell line. After the pretreatment of 5F6, 6B3, 9C11, 3B10, 2D8, and 7E7 for 1 h, the IC50 concentration of both HL60/DOX and HL60 cell lines showed no obvious changes, whereas 6F5 and 6A2 pretreatment showed that the DOX resistance factor (IC50 HL60/DOX/IC50 HL60) decreased from 74.3- to 57.3- and 66.2-fold, respectively. Interestingly, the 5D12 pretreatment resulted in a more remarkable decline of resistance factor (from 74.3- to 22.3-fold). 5D12 caused an increase of IC50 folds which was greater in HL60 (almost fivefold) than in HL60/DOX (almost 1.5-fold). The observations mean a greater enhancement of DOX-resistant phenotype in sensitive cells. The results suggest that the target of McAb 5D12 may play an important role in cell drug response.

Table 1. Effects of McAbs on the sensitivity of HL60 and HL60/DOX cell line to doxorubicin
GroupsIC50aResistance factor (RF)b
HL60 (μg/mL)HL60/DOX (μg/mL)
  1. Percentage of 72 hr-cell survival as a function of DOX concentration was plotted to determine the IC50 value.

  2. a

    IC50 values are shown as mean ± SD calculated from five independent experiments performed in triplicates.

  3. b

    RF, resistance factors were calculated as IC50 for resistant cells/IC50 for sensitive cells.

DOXControl0.194 ± 0.03214.41 ± 2.4274.3
McAb5F60.188 ± 0.04413.991 ± 2.1174.4
McAb6A20.211 ± 0.03413.979 ± 2.5866.2
McAb6B30.207 ± 0.04714.49 ± 3.0470.0
McAb9C110.192 ± 0.05114.07 ± 4.1873.3
McAb3B100.186 ± 0.03613.845 ± 2.7974.4
McAb2D80.201 ± 0.02914.872 ± 3.7174.0
McAb7E70.179 ± 0.04013.416 ± 3.6575.0
McAb6F50.156 ± 0.0348.937 ± 1.8557.3
McAb5D120.954 ± 0.13121.33 ± 4.2622.3

McAb 5D12 Preferentially Binds to the HL60 Cell Membrane

Flow cytometry showed that the McAb 5D12 exhibited strong binding to HL60 cells and weak reactivity to HL60/DOX cells (Fig. 2A). It is not surprised to generate the antibodies against HL60 cells in contrast to HL60/DOX in subtractive immunization under the condition of cyclophosphamide induced tolerance. In fact, the reported percentages of clones obtained from subtractive immunization do not always have the desired reactivity as theoretically expected. As we more prefer the antibodies with functional activity, the following studies with McAb 5D12 were conducted. To further validate the cell-surface expression of the antigen recognized by 5D12, CM-Dil was used to locate the cell membranes and the nuclei were indicated by DAPI. HL60 and HL60/DOX cells were stained with the 5D12-antibody, immunofluorescence staining confirmed the distinct membrane staining pattern between the two cell lines. HL60 exhibited much stronger cell membrane staining while slight staining was observed on HL60/DOX (Fig. 2B), which further confirmed the surface binding specificity of 5D12-McAb on HL60 compared to HL60/DOX cells.

Figure 2.

Binding of McAb 5D12 to HL60 and HL60/DOX cell lines were determined by flow cytometry (A). HL60 cells exhibited dramatically high expression of the antigen recognized by 5D12 compared with HL60/DOX cells. The left graph represents the background staining measured without 5D12, while the right graph represents the binding with the 5D12 followed by FITC-conjugated secondary antibody. Three experiments were performed, and only a representative figure is shown. Immunofluorescence staining of cell surface with 5D12 was assessed by laser scanning confocal microscope (B). Intense membrane staining on HL60 cell surface was observed for McAb 5D12 (green), while weaker staining was observed on DOX-resistant HL60/DOX cells. The plasma membrane is confirmed with a membrane-located photostable fluorescent dye, CM-Dil (red), and the nuclei are stained with DAPI (blue). Images were acquired using a Nikon ×100 oil-immersion objective lens. Scale bar represents 20 μm. Image processing was done with Adobe Photoshop CS version 8.0.1 Isotype-matched control McAbs were included in all experiments. Western blotting shows that a 100 kDa band to HL60 cell membrane lysate is stronger than that for HL60/DOX cells (C). [Color figure can be viewed in the online issue, which is available at]

To confirm the binding specificity and to determine approximate molecular mass of antigens recognized by McAb-5D12, equal amount of membrane extracts from HL60 and HL60/DOX cells were analyzed by Western blotting. An approximately 100 kDa band was shown in both cell lines. The different band intensity appeared to reflect a higher expression of 5D12 target protein on HL60 cell surface than HL60/DOX (Fig. 2C).

Identification of PSF as the Target of McAb 5D12 by Immunoprecipitation and MALDI-TOF/TOF MS

It is difficult to perform immunoprecipitation (IP) for cell membrane lysates because the membrane proteins are usually hydrophobic and low-abundant. Therefore, we performed IP and gel analysis with whole cell lysates to enrich the amount of antigen, improve the result of IP, and give facilities for MS identification. Commassie blue staining of gel revealed an unique protein band of approximately 100 kDa (Fig. 3A), consistent with the one in Western blot analyses from membrane extracts (Fig. 2C). The protein band was then analyzed by MALDI-TOF/TOF MS, the data obtained for 11 peptides from the 100-kDa protein matching PSF (NCBI ID: gi│1709851), which is also known as the splicing factor proline/glutamine-rich based on a total of 114 ion scores (SwissProt Protein Knowledgebase ID: P23246; Fig. 3B; Table 2).

Figure 3.

Identification of proteins coimmunoprecipitating with 5D12 by MS. Commassie blue stained 12% SDS-PAGE gel of 5D12 immunoprecipitates from whole lysate of HL60 and HL60/DOX cells, purified McAb 5D12 alone (A). 5D12 exhibited stronger reactivity to HL60 cell line than HL60/DOX. The band (∼100 kDa) was excised from the gel, trypsin-digested, and analyzed by MALDI-TOF MS. The PSF identification is shown in (B), mass spectra were acquired on a 4700 Proteomics Analyzer (Applied Biosystems) in the positive ion mode. MS/MS sequencing of peptides identifying PSF (GenBank ID: P23246). Reciprocal IP and Western blotting (WB) experiments were performed using 5D12 and B92, an anti-PSF commercial antibody on cell lysates derived from HL60 cells (C).

Table 2. Identification of PSF (accession number gi│1709851) by MS
No. matched peptideNo. unmatched peptideCalculated massObserved mass± ppmStart sequenceEnd sequenceSequence
  1. Measured masses from MALDI-TOF, theoretical masses from SwissProt human protein database, their difference (± ppm), and the corresponding sequence in PSF are shown for each matched peptide. After MALDI-TOF MS and database searching, 11 tryptic peptides matched with theoretical masses, 8 peptides did not match, leading to a sequence coverage of 21%.


To confirm the proteomic results described above, IP and Western blotting were repeated using a commercial monoclonal antibody, B92 (Sigma, St. Louis, MO), which can react specifically with PSF. As shown in Figure 3C, both the B92 and 5D12 antibodies yielded identical patterns of immunoreactivity with the 100-kDa protein when blotted against immunoprecipitates from HL60 whole cell lysate. In a parallel experiment, HL60 whole-cell lysate was first immunoprecipitated with 5D12, followed by blot analysis with B92. The 100-kDa protein was also observed in the 5D12-precipitated sample (Fig. 3C, lane 3), which further confirmed the binding specificity of 5D12 for PSF.

Downregulation of PSF Caused the Decreased Cell Surface Expression of PSF

Whole cell lysates were used both for IP and in gel analyses before purification and MS, which may make the results unclear. To this end, we ask if PSF siRNA-silencing treatment is responsible for membrane PSF overexpression. HL60 cells were transfected with siRNA, membrane PSF were then detected by flow cytometry and immunofluorescence with McAb-5D12. The 48-h transfection of anti-PSF siRNA resulted in significant downregulation of mRNA and protein (P < 0.01, Fig. 4A). Flow cytometry and immunofluorescence staining showed that the surface expression of PSF in HL60 cells was downregulated after PSF-siRNA transfection (Figs. 4B and 4C). These results provide further evidence that the cell surface PSF recognized by 5D12 was indeed PSF. We additionally tested psf-siRNA-mediated cell drug response in HL60 cells. The psf-siRNA transfected cells were incubated with medium containing 0.6 μM DOX (IC50 concentration) for 48 h and then performed using CCK8 proliferation assay. As shown in Figure 4D, the specific psf-siRNA-treated cells increased its relative cell survival from 50% to 67%, the same treatment had no effect on nonspecific-siRNA treated cells and wild-type HL60 cells.

Figure 4.

Expression of PSF by transfected siRNAs in HL60 cells. RT-PCR and Western blot analysis of expression of PSF at 48 h after psf-siRNA (100 nM) transfection (A). The intensity of the bands has significance in statistics (P <0.01). PSF expression on HL60 cell surfaces is demonstrated by 5D12 at 48 h of siRNA transfections. Three experiments were performed, only a representative figure was shown. 5D12 binding profile revealed that membrane PSF expression significantly decreased by anti-PSF siRNA, but not by nonspecific siRNA (P <0.01) (B). 5D12 immunofluorescence staining of membrane PSF on HL60 cell surfaces were performed at 48 h after siRNA transfections (C). Transfected cells were incubated with 5D12 for 1 h at 37°C, followed by staining with FITC-conjugated secondary antibody for 30 min at room temperature. The plasma membrane was indicated with CM-Dil (red), the nuclei was stained with DAPI (blue). The psf-siRNA transfection caused an obviously decreased expression of cell surface PSF. HL60 cells were loaded with 0.6 μM DOX (IC50 concentration) after siRNA transfection for 48 h. Cell survival was significantly increased in the psf-siRNA transfected cells compared with nonspecific siRNA-transfected cells and wild type HL60 cells (D). Data are represented as the mean ± SD of three independent experiments. **P < 0.01, versus nonspecific-treated and wild-type HL60 cells. [Color figure can be viewed in the online issue, which is available at]

Membrane PSF Partially Contributed to the Sensitivity of HL60 Cells

Because the expression of PSF in both cell membrane and cytoplasm were reduced by suppression of anti-PSF siRNA, we further investigate whether the increased DOX sensitivity was only affected by cell surface PSF rather than by cytoplasmic PSF. Loading with DOX at IC50 concentration for 24 h, treatment that 5D12 selectively blocks cell surface PSF significantly increased the HL60 relative cell survival in a dose-related fashion. In 3, 5, and 8 μg/mL 5D12, the relative survival of HL60 cells is 54.86%, 74.62%, and 85.06%, respectively, whereas the same treatment caused relatively few changes of HL60/DOX cells' survival (Fig. 5A). Experiments loaded with DOX at IC30 concentration produced a similar result, 5 μg/mL 5D12-treatment increased the relative HL60 cell survival to 88.14% (Fig. 5B). The increased relative cell survival reflects a cell resistance to DOX. Collectively, these data indicate that membrane PSF confers a survival advantage to cells exposed to chemotherapeutic reagents.

Figure 5.

HL60 and HL60/DOX cells were incubated with various concentrations of 5D12 for 1 h, and followed by exposure to respective IC50 concentration of DOX for an additional 24 h. 5D12 increased membrane PSF-mediated HL60 cell survival in a dose-dependent manner (A). The same treatment was performed at a lower IC30 concentration of DOX (B). The relative cell survivals were obtained in three independent experiments, each based on determinations in pentaplicate. **P < 0.01, ***P < 0.001 versus DOX alone treated cells.

Membrane PSF Inhibited Cell Proliferation in HL60 Cells

To explore the possible role involved in membrane PSF-induced cell drug sensitivity, we considered cell proliferation that is affected by cell surface PSF. We used McAb-5D12 to block the effect of membrane PSF. Equal numbers of HL60 and HL60/DOX cells were incubated with 5D12 at increasing concentration of 1, 3, 5, 8, and 12 μg/mL for 96 h. As show in Figure 6A, membrane PSF blocking in HL60 cells caused a dose-dependent increase in proliferation, while marginally increased in HL60/DOX. We further investigated the colony forming efficiency responsible for cell drug sensitivity induced by overexpression of membrane PSF on HL60 cells. The membrane PSF blocked with McAb-5D12 at any concentration (5, 8, and 12 μg/mL) significantly increased the number of cell colonies compared with that of those treated with PBS (Fig. 6B) (22). McAb-5D12 against PSF on cell surface promotes in vitro cell growth, validating membrane PSF as a potential target for development of MDR prognosis through cell proliferation inhibition.

Figure 6.

5D12-induced cell proliferation and colony formation assay. (A) HL60 and HL60/DOX cells were pretreated with different concentrations of 5D12 and cultured for 96 h. Cell viabilities were determined by CCK8 assay. The proliferation in HL60 cells was distinct, while a moderate response was observed in HL60/DOX cells. The results are expressed as the mean ± SD of three independent experiments, each based on determinations in pentaplicate. *P<0.05, ** P<0.01 versus HL60/DOX. (B) Membrane PSF was blocked with McAb-5D12 at various concentrations. Single-HL60 cell suspensions were plated in soft agar, and the plates were incubated at 37°C for 16 days and photographed. Colonies were counted using Quantity One software and were expressed as mean±SD of three independent experiments. The representative figure was shown (colony number: 5 μg/mL *P<0.05, versus non-5D12 control; 8 μg/mL and 12 μg/mL ** P<0.01, versus non-5D12 control).

PSF Membrane Expression Are Common in Hematological Neoplasms

We tested the percentages of mAb 5D12 binding to some additional hematological malignant cell lines and primary leukemia specimens to further characterize the cell surface expression levels of PSF. All clinical samples were collected at different days, each specimen was stained within 24 h of collection, gated, and analyzed by all cells. Isotype controls were performed to define negative cells for respective specimen. Flow cytometry showed that myeloid leukemia cell lines U937, HL60 and B lymphocytic leukemia cell line Raji exhibited high cell surface expression, Nomalva, Jurkat, and CEM cell lines exhibited moderate cell surface expression of the PSF recognized by 5D12. We also identified the marrow of some primary leukemia patients, similar to the results of sensitive myeloid leukemia cell lines, PSF exhibited high membrane expressions in these patient specimens (Fig. 7).

Figure 7.

Data are obtained from hematological malignant cell lines, healthy persons, and 12 patients with acute myeloid leukemia (AML), chronic myelocytic leukemia (CML), or atypical CML. Each clinical specimen was stained within 24 h of collection and was gated and analyzed by all cells. Values of specimens were calculated as the value minus their respective isotype control value. The groups, cell lines, healthy persons, and AML M2 from which we obtained at least three values were presented as mean ± SD.


Subtractive immunization has been broadly and successfully applied for generating monoclonal antibodies, which can be used to search for the cell surface proteins with altered expression in selected cell lines with similar genetic backgrounds [9, 12, 13, 23]. In this study, this comparative proteomics strategy was combined with a functional screening to identify the potential cell surface proteins involved in the process of DOX resistance of HL60/DOX cells. By whole-cell subtractive immunization with HL60 followed by HL60/DOX cell lines, we obtained 51 discriminatory monoclonal antibodies, including 39 clones to HL60/DOX cells and 12 clones to HL60 cells. Theoretically, all these antibodies should preferentially react with the surface protein of HL60/DOX cells, because HL60 cells are used in the tolerization step with cyclophosphamide. However, antibodies produced from subtractive immunization strategy are not always specific for immunogen but also toward the tolerogen in some cases, which may due to the incomplete or nonabsolute elimination of tolerogen responsive B cells, as reported in studies by Mernaugh et al. [13]. Importantly, this study is not only for demonstrating some cell surface markers but also for identifying the cell surface protein with functional role in cell drug response. The cell growth inhibition assay was next used to screen the function of antibodies. We picked up one antibody designated 5D12 which shows remarkable alteration in the drug sensitivity for further characterization.

McAb 5D12 was used as an immunoprecipitating agent to purify its membrane target antigen from HL60 cell lysates. Peptide sequence information obtained from the trypsin-digested, immunoprecipitated antigen allowed for database profiling that identified a protein named PSF [24, 25]. PSF was initially termed as splicing factor since it is normally associated with pre-mRNA splicing [25]. It is a multifunctional nuclear factor that has been implicated in diverse reactions in the nucleus [26]. With the N-terminal proline/glutamine-rich domain, PSF is able to interact with DNA, RNA, and several proteins including p54nrb [27], RAD51 [28], Hakai [29], Fox-3 [30, 31] topoisomerase I, and protein kinase C to form multiprotein complexes that are involved in pre-mRNA splicing [25, 32], gene transcription [33], DNA repair [28, 31], and recombination [27, 34]. With the C-terminal nuclear localization signals, PSF normally displays in a nuclear localization pattern [35], but recent reports indicated that nuclear PSF could also be relocated to the cytoplasm after being phosphorylated by the NPM/ALK fusion protein [36] or BRK [37].

We found the membrane localization of PSF is common in many hematological malignances cell lines and leukemia specimens. PSF showed a different cell surface expression level in hematological malignances and exhibited unifying feature of high membrane expression in the sensitive myeloid leukemia cell lines and the primary leukemia patients. McAb 5D12 preferentially binds to the surface of HL60 cells compared with HL60/DOX cells, providing direct evidence of PSF relocation on cell surface and overexpression on HL60 cell surface. Blockage of PSF by 5D12 resulted in an increase in DOX resistance in both cell lines. Note that the increase was greater in HL60 than in HL60/DOX, which should be related to different express levels of PSF on cell surface. Analysis of the PSF amino-acid sequence failed to identify motifs that would indicate it to be a functioning cell surface enzyme, or to belong to a specific class of receptor molecules or cell surface ligands. However, the ability of PSF to be tyrosine phosphorylated and the presence of potential binding sites for other factors suggest that it may function in signal transduction. In addition, PSF once was identified as a myoblast cell surface antigen with a monoclonal antibody [38], and this finding was verified on human brain microvascular endothelial cell surfaces [39]. Therefore, we are not surprised to observe that a multifunctional nuclear or cytoplasmic PSF protein may be a cell surface marker and be involved in the drug resistance in leukemia HL60 and HL60/DOX cells.

The specificity of 5D12-McAb was further confirmed by the anti-PSF siRNA transfection, which resulted in a decrease in immunofluorescence staining against membrane PSF by the inhibition of PSF expression in HL60 cells. However, the reduced cell surface PSF expression by PSF-siRNA transfection just partially increased HL60 cells' sensitivity to DOX, not as obvious as the direct effect of 5D12-McAb on the cell surface PSF of HL60 cells. This may be because the HL60 cells are hard-to-transfect human suspension blood cells, and the anti-PSF siRNA are not able to block the PSF function as efficiently as 5D12-antibody [40].

Our siRNA study did not address the question on the exact physiological and pathological roles of cell surface expressed PSF. Therefore, we further examined the role of cell surface PSF in cell sensitivity and proliferation by selective cell surface PSF blocking. The pretreated HL60 cells with 5D12 showed more obvious resistant phenotype than siRNA did. Interestingly, the treatment of 5D12 specifically increased proliferation and colony formation, through a potential membrane PSF blocking, which is consistent with the report of Galietta et al. [36]. However, their findings should be cautiously interpreted, because they focused on the signal transduction via phosphorylation of intracellular PSF. Therefore, additional experiments are required to clarify whether cell surface PSF has the similar functional role to intracellular PSF. In addition, more in-depth mechanistic studies are also warranted to further elucidate the relevance and the roles of cell surface PSF expression in the cell drug response and cell proliferation of HL60 cells.

In conclusion, we have generated a large panel of antibodies with differential reactivity to HL60 and HL60/DOX cells by subtractive immunization, and the functional role of these antibodies in drug resistance were screened subsequently. Furthermore, we applied 5D12 McAb as an efficient tool and probe for immunopurification, permitting the detection, isolation, and identification of the cell surface PSF protein. Downregulation of the PSF using the siRNA approach partially decreases HL60 cells' sensitivity to DOX, and the selective blocking of the cell surface PSF with 5D12 McAb reduces cell sensitivity to doxorubicin by promoting cell proliferation. Further studies are required to elucidate the molecular mechanisms of cell surface PSF involved in the emergence and maintenance of HL60 sensitivity to DOX, so as to develop more effective approaches for cancer therapy.


The authors thank Dr. Masahiro Imamura (Hokkaido University School of Medicine, Hokkaido) for the generous gifts of the human acute myeloid leukemia cell line HL60 and its doxorubicin-selected MDR cell line HL60/DOX.