Leukemia Research 30 (2006) 469–476

Sorcin, an important gene associated with multidrug-resistancein human leukemia cells

Yuan Zhou, Yuanfu Xu, Yaohong Tan, Jing Qi, Ying Xiao, Chunzheng Yang,Zhenping Zhu, Dongsheng Xiong∗

State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital,Chinese Academy of Medical Sciences and Peking Union Medical College, 288 Nanjing Road, Tianjin 300020, PR China

Received 21 July 2005; received in revised form 22 August 2005; accepted 23 August 2005Available online 6 October 2005

Abstract

Sorcin, or soluble resistance-related calcium-binding protein, is a 22 kD calcium-binding protein initially identified in many mutlidrugresistant (MDR) cell lines. We previously observed by gene profiling that sorcin is significantly up-regulated in a doxorubicin-induced MDRl ia patientsc gnosis.I 562 cells.O to a varietyo of sorcine extent ofr kemia cells.©

K

1

lcsdciap

vamt

s is,var-DR

anthells.andwideandical

f thee

nti-nsity

0d

eukemia cell line, K562/A02, over its parent cells. We have also demonstrated that the level of sorcin expression in leukemorrelates not only directly with that of themdr1 gene, but also inversely with patients’ response to chemotherapies and overall pron this report, we have carried out experiments to dissect out the contribution of sorcin by itself to drug resistant phenotype in Kverexpression of sorcin protein by gene transfection in K562 cells resulted in increased drug resistance, from 4.1- to 22.5-fold,f chemotherapeutic agents, including doxorubicin, etoposide, homoharringtonine and vincristine. On the other hand, inhibitionxpression in both MDR K562/A02 and the sorcin-transfected K562 cells with sorcin-targeting small interfering RNA led to varyingeversal of drug resistance. These results confirm that sorcin is an important gene associated with the development of MDR in leu

2005 Elsevier Ltd. All rights reserved.

eywords: Multidrug resistance; Sorcin; Transfection; Doxorubicin; siRNA; Leukemia

. Introduction

Multidrug resistance (MDR) in cancer, especially ineukemia, represents a major obstacle to successfulhemotherapies[1–3]. Under experimental conditions, expo-ure of tumor cells to a single chemotherapeutic agent, e.g.,oxorubicin (DOX), vincristine (VCR), or etoposide (VP16),an lead to development of resistance not only to the agenttself, but also to a broad range of structurally and function-lly unrelated compounds. The mechanisms of MDR areolygenetic and yet to be completely defined[4–6]. Suc-

Abbreviations: MDR, multidrug resistance; DOX, doxorubicin; VCR,incristine; VP16, etoposide; HHT, homoharringtonine; MTX, methotrex-te; AML, acute myeloid leukemia; CR, complete response; MTT,ethylthiazolyldiphenyl-tetrazolium bromide; RT-PCR, reverse transcrip-

ase polymerase chain reaction; siRNA, small interfering RNA∗ Corresponding author. Tel.: +86 22 27230740; fax: +86 22 27230740.

E-mail address: dsxiong@public.tpt.tj.cn (D. Xiong).

cessful development of effective MDR reversal agenttherefore, heavily dependent on our understanding to theious molecules/pathways involved in the emergence of Mphenotype.

In our previous study, we established a MDR humleukemia cell line K562/A02 by stepwise increasingDOX concentration in the culture medium of K562 ceThe K562/A02 cells display a classical MDR-phenotype,demonstrate resistance not only to DOX but also to arange of unrelated compounds including VCR, VP16homoharringtonine (HHT). On molecular and biochemanalysis, K562/A02 cells showed both amplification ofmdr1and over-expression of P-glycoprotein, up-regulation oactivities of glutathione-S-transferase-� and protein kinasC, along with decreased level of topoisomerase II[7,8]. Bygene profiling using a cDNA microarray analysis we idefied 12 genes that showed significantly altered signal debetween K562/A02 and its parent K562 cells[9]. One of the

145-2126/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.oi:10.1016/j.leukres.2005.08.024

470 Y. Zhou et al. / Leukemia Research 30 (2006) 469–476

genes significantly up-regulated in K562/A02 cells is sorcin(soluble resistance-related calcium-binding protein). Sorcin(also known as VP19 and CP22), a 22 kDa Ca2+ binding pro-tein, was first identified in a VCR-resistant Chinese hamsterlung cell line[10], and implicated to be involved in the devel-opment of MDR phenotype in a variety of human cancercell lines[11–15]. The exact role and/or contribution of sor-cin overexpression in MDR in cancer, especially in humanleukemia, are, however, yet to be fully elucidated. In our pre-viously study we showed that over-expression of sorcin inpatients with acute myeloid leukemia (AML) was, similar tothat ofmdr1, associated with low complete response (CR) tochemotherapies as well as poor overall clinical prognosis, andpatients with over-expression of bothmdr1 and sorcin had theworst outcome[16]. In this study we carried out experimentsto dissect out the role of sorcin in drug resistant phenotypein leukemia cells by transfecting and over-expressing sorcinin non-resistant K562 cells. Further, we examined the effectof sorcin-targeting small interfering RNA (siRNA) on DOX-sensitivity of K562/A02 and sorcin-transfected K562 cells.Our results indicate that sorcin may play an important role inthe emergence of drug resistance in human leukemia cells.

2. Materials and methods

2

ums ifeTfi intD .A.( dm rep

2

As react 02c viaR wedb aP51f R:f(A )( anka vec

reverse: 5′ CAGGTCCAGACGCAGGATGGC 3′. The PCRproducts were electrophoresed on a 2% (w/v) agarose geland the sorcin gene was purified using the QIAGEN purifica-tion kit (QIAGEN GmbH, Hilden, Germany). The gene wasdigested with Hind III and Xba I, cloned into the pcDNA3.1vector (Invitrogen Life Technologies, Grand Island, NY),followed by sequence confirmation by DNA sequencing anal-ysis.

2.3. Transfection and expression of the full-length sorcingene in K562 cells

Sub-confluent K562 cells were transfected withpcDNA3.1/sorcin or pcDNA3.1/control vector using lipo-fectamine 2000 following the manufacturer’s protocol(Invitrogen Life Technologies, Gaithersburg, MD). Briefly,6× 105 K562 cells, seeded in an individual well of a 24-wellculture plate, were transfected with 1�g of DNA, followedby G418 selection 48 h after transfection (700�g/ml) incomplete RPMI-1640 media for 2 weeks. The cells wereharvested and re-plated (2000 cells/well) in methylcellulosemedia (1% methylcellulose in complete RPMI-1640 and700�g/ml of G418). After 10 days culture, individualG418-resisitant colonies were picked, propagated andscreened for the expression of sorcin by RT-PCR using theprimers: forward: 5′-TCCGCTGTATGGTTACTTTG-3′;ru

2

antsw thet MT %s SF,1l r-n on at1 ono say[p (w/v)S tion,t cel-l anesw fer( enw een2 anti-s at a1 u-b -goata red Inc.,T

.1. Cell lines and chemicals

K562 cells were maintained in RPMI 1640 mediupplemented with 10% fetal calf serum (Invitrogen Lechnologies, Grand Island, NY), at 37◦C in a humidi-ed, 5% CO2 atmosphere. K562/A02 cells were culturedhe complete RPMI 1640 media containing 1�g/ml DOX.OX was purchased form Pharmacia & Upjohn S.p

Milan, Italy). VCR, VP16, HHT, methotrexate (MTX), anethylthiazolyldiphenyl-tetrazolium bromide (MTT) weurchased from Sigma (St. Louis, MO).

.2. Amplification and cloning of full-length sorcin gene

The full-length sorcin gene was amplified from cDNynthesized by reverse transcriptase polymerase chainion (RT-PCR) using the total RNA extracted from K562/Aells as the template. Briefly, cDNA were synthesizedT-PCR using the random hexamers as primers, folloy amplification of the full-length sorcin gene usingCR procedure comprising of: an initial step at 94◦C formin followed by 30 cycles of 94◦C for 1 min, 58◦C formin, and 72◦C for 1 min, and one last cycle at 72◦C

or 10 min. The following primers were used for PCorward: 5′ AAAAAGCTT ATGGCGTACCCGGGGCAT 3′Hind III site underlined); reverse: 5′ GGGTCTAGATT-AACACTCATGACACATTG 3′ (Xba I site underlined

sorcin sequence information was obtained from GenBccession no. BC011025).�-Actin was used as the positiontrol: forward: 5′ CTACAATGAGCTGCGTGTGGC 3′;

-

everse: 5′-GGTGATCTTTCCATTGGTG-3′. �-Actin wassed as an internal control.

.4. Western blotting analysis of sorcin expression

Sorcin protein expression levels in K562 transfectere determined by Western blot analysis. In brief,

ransfected cells were lysed in RIPA lysis buffer (50 mris–HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 0.5odium deoxycholate, 1 mM EDTA, 0.1% SDS, 1 mM PMmM sodium orthovanadate, 1�g/ml aprotinin and 1�g/ml

eupeptin) at 4◦C with sonication. The cytosolic supeatant was separated from cellular debris by centrifugati2,000× g for 10 min at 4◦C and the protein concentratif the cytosol fraction was determined by the Lowry as

17]. Aliquots of the cytosol fraction containing∼200�groteins were subjected to electrophoresis on a 12%DS-polyacrylamide gel. After electrophoretic separa

he proteins from the gels were transferred onto nitroulose membranes with a semidry blotter. The membrere blocked with 5% (w/v) nonfat dried milk in TBS buf

2.5 mM Tris–HCl, 150 mM NaCl, pH 7.4) for 1 h and thashed three times with TBS buffer containing 0.05% Tw0 (TBST). The membranes were probed with a goatorcin antibody (Santa Cruz Biotechnology, CA, USA):250 dilution in TBST buffer for 2 h at RT, followed by incation with a horseradish peroxidase-conjugated antintibody in TBST for additional 2 h. All the signals weetected by using the DAB detection system (Bio Basicoronto, Canada).

Y. Zhou et al. / Leukemia Research 30 (2006) 469–476 471

Table 1Nucleotide sequences of the sorcin-targeting siRNA

SiRNA duplex Target sequence Sense strand (5′ → 3′) Antisense strand (5′ → 3′) Position in sorcin genecoding sequence

SI AAAACGATACAGCACCAATGG AACGAUACAGCACCAAUGGtt CCAUUGGUGCUGUAUCGUUtt 435–455SII AATGCTGGATAGAGATATGTC UGCUGGAUAGAGAUAUGUCtt GACAUAUCUCUAUCCAGCAtt 240–260

2.5. siRNA preparation and transfection

Two 21-nucleotide siRNA (SI, SII) duplexes targeted dif-ferent parts of the human sorcin gene were designed usingthe siRNA Target Finder and Design Tools (available athttp://www.ambiom.com). The sequences of the two siRNAduplexes and their targets are listed inTable 1. The siRNAswere prepared using the SilencerTM siRNA Construction Kit(Ambion, Austin, TX) following the manufacturer’s instruc-tion. The scrambled sequences (control siRNA), comprisingof the same nucleotides as in SI and SII but in random-ized order, were used as the negative controls. These scram-bled sequences have no significant homology to any knownhuman gene sequences in the GenBank database. The siRNAduplexes were transfected into leukemia cells using lipofec-tamine 2000 according to the recommended protocol by themanufacturer. The effects of siRNA on sorcin mRNA andprotein expression were examined 24, 48 and 72 h after trans-fection by RT-PCR and Western blot, respectively. A siRNAsequence targeting human GAPDH gene expression was usedas a positive control of RNA interference.

2.6. Chemo-sensitivity assay

The transfected leukemia cells were subjected to a chemo-sensitivity assay to investigate their sensitivity to various

chemotherapeutic agents. The assay was performed follow-ing the method of Mosmann[18] with minor modifica-tions. Briefly, 2× 104 leukemia cells were seeded in eachwell of 96-well tissue culture plates (Costar, Charlotte, NC)and exposed to different concentrations of chemotherapeuticagents for 72 h at 37◦C in a 5% CO2 atmosphere. Twentymicroliters of MTT (5 mg/ml) was added to each well andincubated for additional 4 h, followed by addition of 200�lDMSO/well. The absorbance of the wells was then read at570 nm using a spectrophotometer (Model A-5082, SLT LabInstruments, Grodig, Austria). The IC50 value (defined asthe concentration of drugs that inhibited cell growth by 50%)for each cell line to different chemotherapeutic agents wasdetermined.

The two-tailed Student’st-test was used for statistic anal-ysis using the SPSS software package (SPSS, Chicago, IL,USA).

3. Results

3.1. Clone and expression of sorcin in K562 cells

A 616bp gene fragment encoding the full-length humansorcin was amplified from K562/A02 leukemia cells(Fig. 1A), and cloned into pcDNA3.1 eukaryotic expression

F 2 cells levels ofs pcDNA 2c lone K5 C2) wes agaro esults sa size m 62/C2. (CM K562/S /S7; lane 8,K

ig. 1. (A) Amplification of the full-length sorcin gene from K562/A0orcin expression in K562/A02 as well as K562 cells transfected withells, K562/A02 cells, K562 cells transfected with pcDNA3.1/sorcin (cubjected to RT-PCR analysis. The PCR products were separated byre the representative of three similar experiments. (B) M, molecular, molecular size marker; lane 1, K562/S1; lane 2, K562/S2; lane 3,562/S8; lane 9, K562/S9; lane 10, K562/S10.

by RT-PCR. (B and C) Semi-quantitative RT-PCR analysis of the3.1/sorcin or pcDNA3.1/control plasmid. Total RNA (2�g) from parental K5662/S1 to K562/S10) or pcDNA3.1/control (clone K562/C1 and K562/re

se gel electrophoresis and visualized by ethidium bromide staining. Rhownarker; lane 1, K562; lane 2, K562/A02; lane 3, K562/C1; lane 4, K5)3; lane 4, K562/S4; lane 5, K562/S5; lane 6, K562/S6; lane 7, K562

472 Y. Zhou et al. / Leukemia Research 30 (2006) 469–476

vector using the Hind III/Xba I sites. DNA sequencing con-firmed that the cloned gene segment was 100% homologousto the published sorcin sequence.

After transfection of K562 cells with the pcDNA3.1/sorcinconstruct and subsequent G418 selection, individual cloneswere isolated and examined by RT-PCR for sorcin insert.As previously described, expression of sorcin in K562/A02was significantly up-regulated compared to that in the par-ent K562 cells (Fig. 1B). Transfection with the controlpcDNA3.1/control plasmid did not change the expressionof sorcin in K562 cells (Fig. 1B). All 10 randomly pickedpcDNA3.1/sorcin-transfected clones (K562/S1 to K562/S10)yielded positive PCR sorcin product with expected size(Fig. 1C). The level of sorcin gene integration and expres-sion among these clones appeared, however, to be different,for example, K562/S1 and K562/S4 yielded much strongerPCR signals than K562/S6 and K562/S7 did (Fig. 1C). Thiswas further confirmed by Western blotting analysis – sor-cin protein expression in K562/S4 was significantly higherthan that in K562/S7 (Fig. 2A). Based on these obser-vations, K562/S4 and K562/S7 were selected for furtherexperiments.

F 2/S4a sw nitro-c clonala eduibsotNrc

3.2. Sorcin expression in K562 cells confers resistanceto chemotherapeutic agents

The sorcin-transfected K562 clones, along withK562/A02 and the parent K562 cells, were examinedfor their sensitivity to various chemotherapeutic agents.As shown inTable 2, the IC50 to DOX of the 10 indi-vidual sorcin-transfected clones (K562/S1 to K562/S10)ranged from 1.7 to 5.0�M, which are 2.7- to 7.9-foldhigher than that of K562 cells (IC50, 0.63�M). Consistentwith the levels of sorcin gene and protein expression(Figs. 1C and 2A), K562/S4 was∼3-fold more resistantto DOX than K562/S7 (IC50, 5.0�M versus 1.7�M). Asshown previously, K562/A02 demonstrated the strongestresistance to DOX, with an IC50 of 70.8�M, representing a112-fold increase over the parent K562 cells. As expected,the two pcDNA3.1/control plasmid-transfected clones,K562/C1 and K562/C2, were equally sensitive to DOX asthe parent K562 cells.

We next investigated the capability of cross-resistance toother chemotherapeutic agents by K562/S4 and K562/S7.Similar to K562/A02, both sorcin-transfected clones showedsignificant cross-resistance to VP16, HHT and VCR(Table 3). As seen with DOX, K562/S4 was more resistantto all the three agents than K562/S7 did. Compared to K562cells, the IC50 values of K562/S4 to VP16, HHT and VCRi oseo ly( C1,w K562c did

TD

C

KKKKKKKKKKKKKK

nts

ig. 2. (A) Western-blot analysis of soricn protein expression in K56nd K562/S7 cells. Equal amounts (200�g proteins) of cell lysate proteinere separated by a 12% SDS-polyacrylamide gel and transferred ontoellulose membrane. The membranes were immunoblotted with polynti-sorcin antibodies or an anti-�-actin antibody. All signals were detect

sing DAB detection system. Results are the representative of three sim-

lar experiments. (B) Cross-resistance to various chemotherapeutic agentsy K562/S4 and K562/S7. The resistance fold was calculated as: IC50 oforcin-transfected K562 cells/IC50 of the parental K562 cells. Transfectionf sorcin led to∼4- to 22-fold IC50 increase in K562/S4 clone, and∼2-

o 7-fold IC50 increase in K562/S7 clone, to DOX, VP16, HHT and VCR.either K562/S4 nor K562/S7 showed cross-resistance to MTX. Data rep-

esent the mean± S.D. of triplicate determinations.* P < 0.05, ** P < 0.01,ompared to K562/C1.

( 2/C1a istantK

s,a

entalc

ncreased by∼4-, 7-, and 22-fold, respectively; whereas thf K562/S7 increased by∼2-, 2-, and 7-fold, respectiveFig. 2B). The control plasmid-transfected clone, K562/as equally sensitive to all the three agents as the parentells. Both K562/S4 and K562/S7, as well as K562/A02,

able 2OX sensitivity of K562 cells transfected with pcDNA3.1/sorcina

ells DOX, IC50 (�M)b Resistance-foldc

562 0.63± 0.04 1.00562/S1 4.7± 0.23** 7.4562/S2 1.8± 0.22* 2.9562/S3 3.1± 0.09** 4.9562/S4 5.0± 0.21** 7.9562/S5 3.2± 0.34** 5.0562/S6 1.8± 0.06* 2.8562/S7 1.7± 0.15* 2.7562/S8 3.1± 0.18** 4.9562/S9 4.4± 0.15** 7.0562/S10 3.0± 0.14** 4.7562/C1 0.62± 0.05 0.98562/C2 0.58± 0.04 0.92562/A02 70.8± 3.6** 112.4a DOX sensitivity of ten individual pcDNA3.1/sorcin transfecte

K562/S1 to K562/S10) and two pcDNA3.1/control transfectants (K56nd K562/C2), along with the parent K562 cells and the multidrug res562/A02 cells, were determined using MTT assay.b Each number represents the mean± S.D. of triplicate determinationnd is representative of at least three separate experiments.c Resistance-fold was defined as: IC50 value of the experimlone/IC50 value of the parental K562 cells.* P < 0.05, compared with the DOX IC50 of K562 cells.

** P < 0.01, compared with the DOX IC50 of K562 cells.

Y. Zhou et al. / Leukemia Research 30 (2006) 469–476 473

Table 3Effects of pcDNA3.1/sorcin transfection on sensitivity of K562 cells to other chemotherapeutic agentsa

Cells Chemotherapeutic agents, IC50 (�M)b

VP16 HHT VCR MTX

K562 3.4± 0.3 0.034± 0.003 0.023± 0.004 0.068± 0.017K562/S4 13.7± 1.2** 0.23± 0.01** 0.53± 0.12** 0.062± 0.006K562/S7 6.9± 1.0* 0.067± 0.003* 0.17± 0.06* 0.069± 0.015K562/C1 3.4± 0.2 0.037± 0.004 0.025± 0.007 0.064± 0.002K562/A02 229.6± 29.3** 4.02± 0.10** 5.46± 0.62** 0.070± 0.017

a Sensitivity to various chemotherapeutic agents was assayed by MTT assay.b Each number represents the mean± S.D. from at least three separate experiments.* P < 0.05, compared with the IC50 of K562 cells.

** P < 0.01, compared with the IC50 of K562 cells.

not show any cross-resistance to MTX, a folate analogue thatinhibits the activity of dihydrofolate reductase.

3.3. Inhibition of sorcin expression by siRNA leads toincreased chemo-sensitivity in K562/A02 and K562/S4

Next we performed studies to further dissect out the contri-bution of sorcin to DOX resistance in K562/A02 and K562/S4

Feo(eae

cells. In these experiments, the expression level of sorcinin K562/A02 and K562/S4 cells was down-regulated viatransfection with two different siRNA duplexes targeting thecoding region 435-455 (SI) and 240-260 (SII) of sorcin gene,respectively. Compared to the control siRNA, both SI andSII siRNA significantly decreased the mRNA level of sor-cin at 24, 48 and 72 h after transfection in K562/A02 cells(Fig. 3A). Consequently, sorcin protein expression was alsogreatly reduced in these siRNA-transfected K562/A02 cellsas assayed by Western-blot analysis (Fig. 3B and C). Similaroutcome was also observed in K562/S4 cells after transfec-tion with SI and SII siRNA (not shown).

The SII siRNA-transfected K562/A02 and K562/S4 cellswere then assayed for their DOX sensitivity. As shown inFig. 4A and B, SII siRNA transfection led to increased DOXsensitivity in a dose-dependent manner in both K562/A02and K562/S4 cells. The IC50 of K562/A02 to DOXwere 40.2± 3.8�M, 37.2± 4.2�M, and 27.1± 0.4�M aftertransfection with 10, 20, and 40 nM SII siRNA, comparedto that of 72.9± 2.7�M in untransfected K562/A02 cells,representing a∼1.8-, 2.0-, and 2.7-fold increase in drug sensi-tivity (Fig. 4C), respectively. Similarly, The IC50 of K562/S4to DOX were 2.3± 0.1�M, 1.7± 0.1�M, and 1.1± 0.2�Mafter treated with 10, 20, and 40 nM SII siRNA, compared tothat of 5.0± 0.4�M in untransfected K562/S4 cells, repre-senting a∼2.2-, 3.0-, and 4.8-fold increase in drug sensitivity( fec-t ndK

4

fec-

ig. 3. Sorcin-targeting siRNA suppress both sorcin mRNA and proteinxpression in K562/A02 cells. K562/A02 cells were treated with 40 nM SIr SII siRNA or the scrambled control siRNA (SIc or SIIc). Sorcin mRNAA) and protein levels (B and C) in siRNA-treated K562/A02 cells werexamined at 24, 48, and 72 h post treatment via RT-PCR and Western-blotnalysis, respectively. Results shown are the representative of three similarxperiments.

t rugr umorc teins[ mpst ells.C es too lockt ars,

Fig. 4D), respectively. As expected, control siRNA transion had no effect on DOX sensitivity of both K562/A02 a562/S4 cells (Fig. 4).

. Discussion

The emergence of MDR is a major obstacle to efive leukemia treatment. A common mechanism of desistance to chemotherapy is the over-expression by tells of one or more plasma membrane drug efflux pro19,20], such as P-glycoprotein, that function as efflux puo actively transport cytotoxic agents out of the tumor conceivably, one of the most heavily pursued strategivercome MDR is to develop effective agents that bhe pump activity of these efflux proteins. In past ye

474 Y. Zhou et al. / Leukemia Research 30 (2006) 469–476

Fig. 4. Effects of SII siRNA on DOX sensitivity of K562/A02 (A) and K562/S4 (B) cells. Four hours after transfection with various concentrations of SIIsiRNA (SII) or the control siRNA (SIIc), cells were incubated for another 72 h with various doses of DOX. Cell proliferation was determined by a MTT-basedassay. IC50 values of DOX for both the K562/A02 (C) and K562/S4 cells (D) were calculated at each concentration of siRNA. Data shown are the mean± S.D.of triplicate determinations.** P < 0.01, compared to control siRNA-transfected cells.

a number of efflux-pump inhibitors have been extensivelystudied in both experimental and clinical settings for theirMDR-reversal activity. Despite certain promising clinicalbenefits have been achieved with some of these compounds,it has become increasingly clear that simple inhibition of the“pumping” function of the efflux proteins may not be effica-cious enough to completely reverse MDR in patients[21,22].Further development of efficacious MDR reversal agents is,therefore, dependent on our complete understanding of themultiple molecular pathways/mechanisms involved in emer-gence of the MDR phenomenon.

Sorcin was first identified in a VCR-resistant Chinesehamster lung cell line, DC-3F/VCRd-5L[10], and laterdemonstrated to be overexpressed in over 50% of MDR celllines, including those of human, hamster and mouse origin,derived from selection against a variety of compounds suchas colchicines, actinomycin D, taxol, vinblastine, teniposide,VP16, and DOX[12]. The exact role and the underlyingmechanisms of sorcin in drug resistance remain, however,mostly elusive. Additional studies in a number of otherMDR cell lines showed that sorcin was not always ampli-fied and over-expressed[23]. Even in those MDR cells withsorcin-amplification, its level of expression was not nec-essarily correlated directly with the degree of drug resis-

tance[24]. For example, Lee et al., established a series ofMDR lymphoma cell lines with varying degree of resistanceto VCR: HOB1/VCR0.01, HOB1/VCR0.1. HOB1/VCR0.5,HOB1/VCR0.75, and HOB1/VCR1.0, and showed that thelevel of mdr1 amplification correlated directly to the degreeof VCR resistance and reached plateau in HOB/VCR0.5line, whereas sorcin was only amplified in the most resistantline, HOB1/VCR1.0. Further, transfection of the sorcin genecloned from the HOB1/VCR1.0 line into the parent HOB1cells failed to confer significant MDR phenotype[25]. Theseobservations suggest that amplification/over-expression ofsorcin in MDR cells is unlikely a phenomenon secondar-ily to that of mdr1 gene, but rather a direct and independentresponse/outcome to increasing drug selection pressure.

By gene profiling analysis we previously identifiedsorcin as one of the genes significantly up-regulated inDOX-induced human leukemia cell line, K562/A02. In asubsequent study, we further observed that sorcin expressionnot only correlated directly to that ofmdr1 in primaryhuman AML specimen, but also often indicated poor patientresponse to chemotherapies. For example, CR rate was21.7% (5 out of 23 patients) in sorcin+ patients versus thatof 83.3% (35 out of 42) in sorcin̄patients, and the responsedropped to 7.1% in patients who are both sorcin+ andmdr1+,

Y. Zhou et al. / Leukemia Research 30 (2006) 469–476 475

compared to that of 91.4% in patients of sorcin/̄mdr1¯[16].As a comparison, patients in the same study who aremdr1+had a CR rate of 19% (4/21 patients), andmdr1 ,̄ of 81.8%(36/44 patients). It is pertinent to note that sorcin ampli-fication and over-expression was also observed in anotherDOX-resistant leukemia cell line, HL60/ADR, over itsparent HL60 cells. In contrast, in three MDR solid tumor celllines we studied, including a DOX-resistant MCF-7/ADR(breast carcinoma), acis-platinum-resistant A549DDP (lungcarcinoma), and a VCR-resistant KBv200 (epidermoidcarcinoma line) line, no sorcin amplification/overexpressionwere observed, albeit the first two lines and their parentnon-resistant cells express sorcin at a low basal level (ourunpublished observations). These results suggest that sorcinmay represent a significant and independent indicator ofleukemia resistance to chemotherapies.

In this study, we demonstrated that sorcin over-expressionby DNA transfection conferred a MDR-like phenotype tothe non-resistant K562 human leukemia cells. The sorcin-transfected K562 cells were 2.7- to 7.9-fold more resistantto DOX, and like themdr1-amplified K562/A02 cells, werecross-resistant to several other chemotherapeutic agents. Thelevel of drug resistance in sorcin-transfected K562 cellsseems to correlate directly with that of the expression of sor-cin proteins in these cells. These results are in good consistentwith those reported by Parekh et al., who demonstrated thats morcT rcin-t X,c at utet ayr iona ilarD ellst ldl DRc

n inm andu cor-r fer-e callyi lls.C es-s dentD n ofu asei ec-tt cedt m-p 2c moree her

concentrations of the sorcin-targeting siRNA. These results,nevertheless, provide further support to the role that sorcinplays in the emergence of drug resistance.

In conclusion, our results in this study, along with those ofprevious observations[9,16], further confirm that sorcin playsan important role in the drug-resistant phenotype of humanleukemia cells. Sorcin may, therefore, represent a good targetfor modulation in our continuous efforts to searching anddeveloping more efficacious MDR reversal agents.

Acknowledgements

This work was supported by grants from the ChineseNational Natural Sciences Foundation (No. 30271515 andNo. 30570772) and the “National High Technology Researchand Development Program (the 863 Program)” of the ChineseGovernment (No. 2002AA2Z346D).

Contributions. Y. Zhou contributed to the concept anddesign, data analysis and interpretation, statistical expertise,drafting of the article, collected and assembled the data andgave final approval. Y. Xu contributed to the concept anddesign, provided critical revisions, provided technical sup-port and gave final approval. Y. Tan contributed to the studydesign, data analysis and interpretation, obtained funding andgave final approval. J. Qi contributed to technical support,h finala d thed con-t f thea avefi udyd ing,g .

R

J,drugemia

ogica

in

ela-cute18.and

cer:8–58.al.

rative

gicin.7.re-urg

orcin DNA transfection into human ovarian and breast tuells also resulted in low level of paclitaxel resistance[15].he seemingly low resistant capacity conferred by so

ransfection in K562 cells, e.g., 2.7- to 7.9-fold to DOompared to that of 112-fold by K562/A02, is unlikelyrue indication of the degree to which sorcin may contribo the overall drug-resistance in MDR tumor cells, but mather be a reflection of the efficiency of DNA transfectnd protein expression in the transfected cells. In a simNA transfection study, Yu et al. observed that MCF-7 c

ransfected with the humanmdr1 gene were 10- to 50-foess resistance to DOX and vinblastine than MCF-7/Aells established from selection against DOX in culture[26].

The use of RNA interference to inhibit gene expressioammalian cells offers a promising new tool for the studynderstanding of the function of a particular gene and itsesponding protein. In this study, two siRNAs targeting difnt sequences in sorcin gene were employed to specifi

nhibit sorcin expression in K562/A02 and K562/S4 ceonsistent with their activity in down-regulating the exprion of sorcin protein, both siRNA showed a dose-depenOX sensitization effect, and at the highest concentratiosed (40 nM), SII siRNA resulted in 2.7- and 4.8-fold incre

n DOX sensitivity in K562/A02 and K562/S4 cells, respively. It is noteworthy that SII siRNA reversed∼80% ofhe DOX resistance in K562/S4 cells: the IC50 was reduo 1.05�M for the siRNA-transfected K562/S4 cells, coared to that of 5.0 and 0.63�M for the K562/S4 and K56ells, respectively. It is reasonable to speculate thatffective drug sensitization may be achieved with hig

elped to analyze the data, obtained funding and gavepproval. Y. Xiao contributed to collected and assembleata, statistical expertise and gave final approval. Z. Zhu

ributed to the data analysis and interpretation, drafting orticle, provided critical revisions to the revision and gnal approval. C. Yang and D. Xiong contributed to the stesign, gave critical input to the revision, obtained fundave final approval and provided administrative support

eferences

[1] Suarez L, Vidriales MB, Moreno MJ, Lopez A, Garcia-LaranaPerez-Lopez C, et al. Differences in anti-apoptotic and multiresistance phenotypes in elderly and young acute myeloid leukpatients are related to the maturation of blast cells. Haematol2005;90:54–9.

[2] Hirose M, Hosoi E, Hamano S, Jalili A. Multidrug resistancehematological malignancy. J Med Invest 2003;50:126–35.

[3] Ivy SP, Olshefski RS, Taylor BJ, Patel KM, Reaman GH. Corrtion of P-glycoprotein expression and function in childhood aleukemia: a children’s cancer group study. Blood 1996;88:309–

[4] Borst P, Elferink RO. Mammalian ABC transporters in healthdisease. Annu Rev Biochem 2002;71:537–92.

[5] Gottesman MM, Fojo T, Bates SE. Multidrug resistance in canrole of ATP-dependent transporters. Nat Rev Cancer 2002;2:4

[6] Filipits M, Stranzl T, Pohl G, Heinzl H, Jager U, Geissler K, etDrug resistance factors in acute myeloid leukemia: a compaanalysis. Leukemia 2000;14:68–76.

[7] Yang CZ, Luan FJ, Xiong DS, Liu BR, Xu YF, Gu KS. Multidruresistance in leukemic cell line K562/A02 induced by doxorubZhongguo Yao Li Xue Bao (Acta Pharmacol Sin) 1995;16:333–

[8] Zhao CT, Yang CZ, Shao XF, Xiong DS, Xu YF, Qi J, et al. Pliminary study on the role of protein kinase C in the multid

476 Y. Zhou et al. / Leukemia Research 30 (2006) 469–476

resistance phenotype of K562/A02. Zhonghua Xue Ye Xue Za Zhi1998;19:261–2.

[9] Tan Y, Zhao C, Yang C. Identification of unknown genes involved inmultidrug resistance in K562/A02 cells by cDNA microarray. ActaAcad Med Sinica 2001;23:272.

[10] Meyers MB, Biedler JL. Increased synthesis of a low molecularweight protein in vincristine-resistant cells. Biochem Biophys ResCommun 1981;99:228–35.

[11] Koch G, Smith M, Twentyman P, Wright K. Identification of a novelcalcium-binding protein (CP22) in multidrug-resistant murine andhamster cells. FEBS Lett 1986;195:275–9.

[12] Meyers MB. Sorcin a calcium-binding protein overproduced in manymultidrug-resistant cells. In: Smith VL, Dedman JR, editors. Stim-ulus response coupling: the role of intracellular calcium-bindingproteins. Boca Raton: CRC Press; 1990. p. 159–71.

[13] Wang SL, Tam MF, Ho YS, Pai SH, Kao MC. Isolation andmolecular cloning of human sorcin a calcium-binding protein invincristine-resistant HOB1 lymphoma cells. Biochim Biophys Acta1995;1260:285–93.

[14] Beyer-Sehlmeyer G, Hiddemann W, Wormann B, Bertram J. Sup-pressive subtractive hybridisation reveals differential expression ofserglycin, sorcin, bone marrow proteoglycan and prostate-tumour-inducing gene I (PTI-1) in drug-resistant and sensitive tumour celllines of haematopoetic origin. Eur J Cancer 1999;35:1735–42.

[15] Parekh HK, Deng HB, Choudhary K, Houser SR, Simpkins H. Over-expression of sorcin, a calcium-binding protein, induces a low levelof paclitaxel resistance in human ovarian and breast cancer cells.Biochem Pharmacol 2002;63:1149–58.

[16] Tan Y, Li G, Zhao C, Wang J, Zhao H, Xue Y, et al. Expression ofsorcin predicts poor outcome in acute myeloid leukemia. Leuk Res2003;27:125–31.

[ wryhem

[18] Mosmann T. Rapid colorimetric assay for cellular growth and sur-vival: application to proliferation and cytotoxicity assays. J ImmunolMethods 1983;65:55–63.

[19] Ambudkar SV, Kimchi-Sarfaty C, Sauna ZE, Gottesman MM. P-glycoprotein: from genomics to mechanism. Oncogene 2003;22:7468–85.

[20] Schinkel AH, Jonker JW. Mammalian drug efflux transporters of theATP binding cassette (ABC) family: an overview. Adv Drug DelivRev 2003;55:3–29.

[21] Baer MR, George SL, Dodge RK, O’Loughlin KL, Minderman H,Caligiuri MA, et al. Phase 3 study of the multidrug resistance mod-ulator PSC-833 in previously untreated patients 60 years of age andolder with acute myeloid leukemia: Cancer and Leukemia Group BStudy 9720. Blood 2002;100:1224–32.

[22] Greenberg PL, Lee SJ, Advani R, Tallman MS, Sikic BI, Letendre L,et al. Mitoxantrone, etoposide, and cytarabine with or without val-spodar in patients with relapsed or refractory acute myeloid leukemiaand high-risk myelodysplastic syndrome: a phase III trial (E2995).J Clin Oncol 2004;22:1078–86.

[23] Hamada H, Okochi E, Oh-hara T, Tsuruo T. Purification of the Mr22,000 calcium-binding protein (sorcin) associated with multidrugresistance and its detection with monoclonal antibodies. Cancer Res1988;48:3173–8.

[24] Meyers MB. A 22 kDa calcium-binding protein, sorcin, is encodedbyamplified genes in multidrug-resistant cells. In: Heizman CW, editor.Novel calcium-binding protein: fundamentals and clinical implica-tions. Heidelberg: Springer; 1991. p. 385–99.

[25] Lee WP. Purification, cDNA cloning, and expression of human sorcinin vincristine-resistant HOB1 lymphoma cell lines. Arch BiochemBiophys 1996;325:217–26.

[26] Yu G, Ahmad S, Aquino A, Fairchild CR, Trepel JB, Ohno S, etmul-ncer

17] Hartree EF. Determination of protein: a modification of the Lomethod that gives a linear photometric response. Anal Bioc1972;48:422–7.

al. Transfection with protein kinase C alpha confers increasedtidrug resistance to MCF-7 cells expressing P-glycoprotein. CaCommun 1991;3:181–9.