interaction of soybean agglutinin with leukemic t-cells and its use for their in vitro separation...

��

��

�� !�� "#$ % ��&� �� &� "'% ##()� *��

Received 22 August 2002; revised 4 October 2002; accepted 30 October 2002

ABSTRACT: A procedure for separation of leukemic T-cells from normal lymphocytes, using lectin-affinity column chromatogra-phy, is described. CNBr-activated Sepharose 6MB was used as a non-mobile phase. The gel was covalently coupled with soybeanagglutinin (SBA), then served as an affinity probe for fractionation of mixture of normal lymphocytes and leukemic cells. Leukemiccell lines, derived from acute lymphoblastic leukemia (Jurkat, MOLT-4, RPMI-8402), were tested. The elution of normallymphocytes was carried out by PBS(�). The leukemic T-cells, interacting with SBA, were removed by N-acetyl-D-galactosamine orlow-concentration acetic acid. The type and viability of the separated cell fractions were analyzed by flow cytometry and fluorescentmicroscopy, using adequate fluorescent antibodies. The interaction of leukemic T-cells with free SBA, as well as with SBA-conjugated Sepharose beads, was examined fluorimetrically and visualized by fluorescent microscopy, using FITC-SBA as a marker.The rate of cell elution on SBA-affinity column decreased in order: normal � leukemic T-cells. Both normal lymphocytes andleukemic T-cells were removed in a mixture from SBA-free Sepharose 6MB by PBS(�) and were not fractionated discretely. Theleukemic T-cells specifically interacted with SBA as well as with SBA-affinity adsorbent. In contrast, the normal lymphocytes did notinteract with free SBA as well as with SBA-conjugated Sepharose beads in the concentrations applied. The method potentiallycombines a discrete cell fractionation with manifestation of a specific target cytotoxicity of SBA against leukemic T-cells, withoutany influence on normal lymphocytes. Copyright � 2003 John Wiley & Sons, Ltd.

KEYWORDS: acute lymphoblastic leukemia; soybean agglutinin; lectin-affinity chromatography; fraction of cells

�!�� "#$�� !

Leukemia is an increasingly recognized health problem.At present, the research demands are directed to adevelopment of a new strategy for disease control,attending to the requirements of autologous transplanta-tion. Many anticancer drugs are also applied duringchemotherapy followed by bone marrow implantation,until leukemic cells are not observed in the blood.However, these drugs induce severe side-effects inpatients (Santos, 1988, 1990; Petros and Evans, 1990;Hozumi, 1994; Sweetenham, 1995; Reed, 2000; Solary etal., 2000). Sometimes a complete disease remission isinduced, but relapse is usual. The efforts of clinicians and

researchers are directed to reducing the risk of relapse byselective in vitro removal and utilization of tumor cellsfrom autologous graft (Santos, 1988, 1990; Marmont,1998). In this context, the separation of leukemic cellsfrom normal ones, based on their different affinity andsusceptibility to leukemia-cytotoxic drugs or respectiveantibodies, can form the basis for development of a newgeneration of methods for effective control and therapyof leukemia.

In the last 20 years some lectins (plant-derivedproteins) are approved as promising anti-leukemia agentswith their unique biological activities as cytoagglutina-tion, mitogenic activity and cytotoxicity, manifestingwith high target-selectivity for leukemic cells, withoutany significant influence on normal ones (Gabius et al.,1988; Mody et al., 1995; Gabius, 1997; Moriwaki et al.,2000; Sallay et al., 2000). The prerequisite step of theseactivities seems to be initiated by binding of lectins to thecell surface carbohydrate chains of tumor cells (Gabius etal., 1988; Mody et al., 1995; Gabius, 1997, 2001; Chayand Pienta, 2000). These lectin characteristics are broadlyuseful for demonstration of differences in the composi-tion of cell surface and intracellular glycoproteins andglycolipids in tissues at various stages of differentiation,in malignancy and in functional subsets of cells, as well

BIOMEDICAL CHROMATOGRAPHYBiomed. Chromatogr. 17: 239–249 (2003)Published online in Wiley InterScience (www.interscience.wiley.com).DOI: 10.1002/bmc.218

*Correspondence to: R. Bakalova, Natural Substance–ComposedMaterials Group, Institute for Structural and Engineering Materials,Independent Administrative Institution, National Institute of AdvancedIndustrial Science and Technology, AIST-Kyushu, 807-1 Shuku, Tosu,Saga-ken 841-0052, Japan.E-mail: [email protected]

Abbreviations used: ALL, acute lymphoblastic leukemia; FITC,fluorescent isothiocyanate; NAGal, N-Acetyl-D-galactosamine;PBS(�), phosphate buffered saline (Ca2� and Mg2� free); PE,phycoerythrin; SBA, soybean biflorus agglutinin.

Contract/grant sponsor: JSPS Invitation Fellowship Program forResearch in Japan; contract/grant number: L01505.

Copyright 2003 John Wiley & Sons, Ltd.

ORIGINAL RESEARCH

as for isolation, identification and characterization oftumor cell surface receptors (Gabius et al., 1988; Gabius,1989, 1997, 2001; Mody et al., 1995; Chay and Pienta,2000). The lectins proved to be applicable for diagnosticpurposes, especially for the differential diagnosis ofanaplastic tumors (Gabius, 1989; Gabius et al., 1986,1988; Mody et al., 1995; Roth et al., 1996).

In leukemic cells a reduced molecular weight andchanges in the saccharide composition of surface orintracellular glycoproteins were found to be the maindistinguishing marks (Vaickus et al., 1991; Neame et al.,1994; Misra et al., 2000). Antibodies, produced againstlectin affinity-isolated glycoproteins, allowed the demon-stration of structural relationship of glycoproteins withidentical or diverse lectin binding pattern (Forsberg andMacher, 1987; Fischer et al., 1988; Lee et al., 1990).Therefore, the lectins with their specific affinities forsimple and complex sugars on tumor cell surface canrecognize fine differences between leukemic and normalcells. This characteristic may be a potentially useful toolfor separation of leukemic leukocytes from normal oneswith high potential for selective in vitro removal of tumorcells from the autologous graft. A separation wasreported of osteoclasts, erythrocytes and immunogenictumor cells based on their lectin affinity (Killion andKollmorgen, 1976; Pereira and Kabat, 1979; Itokazu etal., 1991). However, the use of lectin-affinity chroma-tography for separation of leukemic T- and B-cells fromnormal lymphocytes is only beginning to be exploited.

In our previous paper we already described aseparation of leukemic T-cells from normal lymphocytes,using Sepharose 6MB, conjugated with dolichos biflorusagglutinin (DBA) (Ohba et al., 2002). However, based onthe fact that the fine differences in quaternary structure ofthe lectins relate directly to the difference in theircarbohydrate specificity and the strength of binding withtumor cell surface receptors (Bouckaert et al., 1999), weexamined and compared the degree of lectin-cell bindingfor several lectins (DBA, SBA and WGA, wheat germagglutinin, and its isolectins) and its influence on theutilization of leukemic T-cells from normal ones, usingdifferent lectin-affinity adsorbents. It was established thatthe degree of lectin-cell binding increased in the orderDBA � SBA � WGA, however WGA and its threeisolectins interacted not only with leukemic cells, butalso with normal lymphocytes (unpublished data). Thismade WGA unsuitable for adsorbent saturation andseparation procedure. SBA was found to have a higherdegree of binding with leukemic cells, as well as a betterexpressed cytotoxic effect against leukemic cells, thanDBA. Thus, we gave preference to SBA as the bestcandidate for affinity probe among another lectins used,because of the potential to combine a discrete cellfractionation with specific cytotoxic effect of SBA on theleukemic T-cells, without any influence on the viabilityof normal lymphocytes.

In the present study, using SBA-conjugated Sepharose6MB as an affinity adsorbent, we separated discretelynormal lymphocytes from ALL-derived leukemic T-cells(Jurkat, MOLT-4, and RPMI-8402) and characterizedtheir type and viability after passing through the lectin-affinity column.

%&'%��(%!�)*

'�� The human leukemic T-cell lines (Jurkat,MOLT-4, RPMI-8402; Hayashibara Biochemical Laboratories,Inc., Okayama, Japan) were cultured in RPMI-1640 mediumsupplemented with 10% heat-inactivated fetal bovine serum (FBS),100 �g/mL streptomycin, and 100 U/mL penicillin in a humidifiedatmosphere at 37°C with 5% CO2. The cell lines were a generousgift of Dr J. Minowada (Hayashibara Biochemical LaboratoriesInc., Okayama, Japan). Normal lymphocytes were purified fromheparinized peripheral blood obtained from normal adults (aged38–40 years) by Lymphosepar I. The cells used for assay were in alogarithmic phase. They were sedimented by centrifugation (1000rpm, 10 min) and washed three times by PBS(�) before experi-ments.

*�� The followingbuffers were used to prepare CNBr-activated Sepharose 6MB forcolumn chromatography: coupling buffer (0.1 M NaHCO3/0.5 M

NaCl, pH = 8.5), washing buffer (0.1 M CH3COONa/0.5 M

CH3COOH, pH = 4.5), and blocking buffer (0.1 M NaHCO3/0.2 M glycine, pH = 8.5). The gel was washed consecutively withthe buffers as described in gel certificate, and then added to SBA(2.5 mg SBA/mL, dissolved in a coupling buffer). The mixture wasincubated 24 h at 4°C for conjugation of SBA to the gel particles.The coupling procedure was repeated three times (the total SBAconcentration added to the Sepharose 6MB was 15 mg SBA/g gel).The non-binding SBA was measured in supernatant spectro-photometrically at � = 280 nm. The concentration of SBAconjugated to the Sepharose beads was calculated using acalibration coefficient and was found to be more than 95%. The‘free’ non-SBA-saturated active sites of adsorbent were blocked bywashing several times with blocking buffer at room temperature(RT).

Cells (normal lymphocytes, Jurkat, MOLT-4, RPMI-8402) werewashed by PBS(�), re-suspended in the same buffer to a con-centration of 2 � 106 cells/mL, mixed 1:1 (v:v, normal:tumor) andthe cell suspension (1 mL) was added to SBA-conjugated CNBr-activated Sepharose 6MB (2 mL) for 15 min at RT. The cells,interacting with gel beads, were removed by competitive inhi-bition, using stepwise elution with aliquots of PBS(�), containingsaccharide (0–0.4 M N-acetyl-D-galactosamine, in the case ofJurkat or RPMI-8402), or low concentrated acetic acid (1–2.5%, inthe case of MOLT-4). A flow rate was 1.2 mL/min and the elutioncontinued until the eluate was cell-free. The same chromatographicprocedure was carried out on SBA-free Sepharose 6MB. The cellfractions were collected by ATTO Minicollector (SJ-1410).

+�� The cell suspensions were detectedspectrophotometrically at 600 nm before and after separation onSBA-affinity or SBA-free columns. The adjacent cell fractions,

Copyright 2003 John Wiley & Sons, Ltd. Biomed. Chromatogr. 17: 239–249 (2003)

240 ORIGINAL RESEARCH R. Bakalova and H. Ohba

showing a high OD at 600 nm (OD600nm � 0.050), were combined.The cells were sedimented by centrifugation (1800 rpm for 10 min)and re-suspended in PBS(�) to concentrate the cell fraction and toprepare it for flow cytometric and microscopic analyses. Thenumber of cells in each fraction was determined microscopically.

,�� The separated cell fractions wereincubated with specific fluorescent antibodies—FITC-CD90 forleukemic T-cells and PE-CD44 for normal lymphocytes. Fivemicroliters of each antibody were added to 0.5 mL of each cellfraction, and incubated for 30 min at RT before flow cytometricassay. As controls, 5 �L PE-CD44 or FITC-CD90 antibodies wereadded respectively to normal or tumor cell suspensions (0.5 mL)and the samples were treated in the same manner. Suspensions ofnormal lymphocytes, leukemic T-cells and their mixture withoutantibodies were used as negative controls to determine thespontaneous cell fluorescence.

The viability and the type of cells were analyzed by flowcytometer Beckman Coulter-Epics XL. The flow cytometer wasoperated in accordance with the manufacturer’s recommendationsafter fine adjustments for optimization. The forward- and side-scatter parameters were adjusted to accommodate the inclusion ofboth leukemic T-cells and normal lymphocytes within theacquisition data. No cells excluded from the analysis, and 10,000cells were counted. Data were collected and analyzed by using ‘XLSystem II’ software.

The results were presented as a dot plot of FITC-fluorescenceand PE-fluorescence with quadrant markers drawn to distinguishFITC- and PE-labeled cells. Quadrants A in Fig. 2 contain viablenormal lymphocytes, quadrants B contain viable leukemic T-cells,quadrants C contain all viable cells, and quadrants D contain alldead cells.

The percentage lysis of cells was calculated before (sponta-neous lysis) and after their passing through the column from thefollowing equations:

� for leukemic T-cells: percentage lysis = [1 � quadrant Bevents/(quadrant C � quadrant D)] � 100;

� for normal lymphocytes: percentage lysis = [1 � quadrant Aevents/(quadrant C � quadrant D)] � 100.

To estimate the effect of SBA-affinity column on the viability ofnormal and leukemic leukocytes, we applied different cell lines tothe column separately and incubated for 15 min at RT or 24 h at4°C. To estimate the effect of incubation on the spontaneous lysis,the respective cell lines were incubated in PBS(�) for 15 min at RTor 24 h at 4°C.

,�� -�� Identification of the type of separatedcell fractions was examined also by fluorescent microscopy, usingFITC-CD90 (green light), and PE-CD44 (red light) antibodies. Tenmicroliters of each antibody were added to 0.5 mL of separated andconcentrated cell fractions, and were incubated for 30 min at RT.The cell fractions were washed twice by PBS(�) for elimination offree fluorescent antibodies and were analyzed by fluorescentmicroscopy, detecting the fluorescence of cell–antibody conju-gates. As controls, 10 �L of PE-CD44 and 10 �L of FITC-CD90were added to normal or leukemic cells, respectively (0.5 mL,

2 � 106 cells/mL) and the samples were treated and analyzed at thesame conditions.

,�� .��-�� +�)� FITC-SBA (0.3 �M)was added to Jurkat, MOLT-4, RPMI-8402 or normal cells(2 � 106 cells/mL) and the mixtures were incubated 15 min at RT.The cells were sedimented by centrifugation at 1800 rpm for10 min, washed twice with PBS(�), re-suspended in PBS(�) andFITC–SBA–cell conjugates were analyzed by fluorescent micro-scopy. In parallel, the degree of lectin–cell interaction wasestimated spectrofluorimetrically at �ex = 485 nm, and �em =538 nm, using BioRad Fluoromark� (JASCO Co., Japan).

,�� -�� +�)��/�� +�� 0(�� Jurkat, MOLT-4,RPMI-8402 or normal cells (1 mL, 2 � 106 cells/mL) wereincubated with FITC-SBA (0.3 �M) for 15 min at RT. The cellswere sedimented by centrifugation at 1800 rpm for 10 min, washedtwice with PBS(�) and added to SBA-conjugated CNBr-activatedSepharose 6MB for 15 min at RT. The gel was filtrated by PBS(�)and then FITC–SBA–cell conjugates, bound to the lectin-saturatedSepharose beads, were analyzed by fluorescent microscopy. Thecontrol experiment consisted of gel particles incubated withnormal lymphocytes and treated in the same manner as leukemicT-cells. The samples were analyzed by Olympus IX70 microscope.

�� All reagents of analytical grade were obtained fromAmersham Pharmacia Biotech AB, Becton Dickinson Co., GibcoGRL, Seikagaku Co. or Wako Pure Chem. Industries Ltd.

�%+#*�+

+�� +�)��/�� +�)��+�� 0(�

Figure 1(A) represents a chromatogram of fractionationof mixture of normal lymphocytes and Jurkat cells bySBA-conjugated Sepharose 6MB. This chromatogram istypical for separation of mixture of normal lymphocyteswith another leukemic T-cells, used in this study (MOLT-4, RPMI-8402). There were two well-defined peaks,consisted of cell fractions with different affinity to SBA-conjugated Sepharose beads. The fractions no. 5–7 (peak1) on Fig. 1(A) were eluted by PBS(�) in the first 1–3 min after beginning of the elution. The fractions no. 24–26 (peak 2) were removed by N-acetyl-D-galactosamine(in the case of Jurkat and RPMI-8402) or low-concentration acetic acid (in the case of MOLT-4 cells)in increasing concentrations—a well-defined peak wasdetected at high concentration of the respective eluent.The cells in fraction no. 41 (peak 3) were strongly boundto SBA-Sepharose 6MB and was removed by mechanicalmixing and subsequent filtration by PBS(�).

In the case of separation of normal lymphocytes fromleukemic T-cells on SBA-free Sepharose 6MB there was


Lectin-affinity chromatography and separation of leukemic T-cells ORIGINAL RESEARCH 241

only one well-defined peak in the chromatogram and thecell elution was carried out by PBS(�), regardless of thecell line used [Fig. 1(B)]. The sum total of optical densityat 600 nm of all fractions no. 5–10 in peak 1 indicatedthat the cells were removed almost completely in a

mixture (normal � leukemic) in the first 2–3 min after thebeginning of the elution. Peak 2 (fraction no. 30) was notwell-defined and it consisted of one cell fraction with lowoptical density at 600 nm and therefore with lowconcentration of cells.

�� 1��

The viability and the type of separated cell-fractions wereidentified by flow cytometric assay and the characteristichistograms are shown in Fig. 2(A) for SBA-conjugatedSepharose 6MB, and in Fig. 2(B) for SBA-free Sepharose6MB.

Comparing the results in Figs 1(A) and 2(A), it wasobserved that peak 1 in the chromatogram of SBA-conjugated adsorbent corresponded mainly to viablenormal lymphocytes, whereas peaks 2 and 3 corre-sponded to viable leukemic T-cells. Comparing theresults in Figs 1(B) and 2(B) it was found that peak 1in the chromatogram of SBA-free Sepharose 6MBcorresponded to the mixture of normal and leukemiccells, mainly viable. It was difficult to identify the type ofcells in peak 2 of this chromatogram based on the datafrom flow cytometry, because the cells were in quadrantD, corresponding to dead cells [Fig. 2(B-e)].

The results from microscopic analysis of chromato-graphically separated cell fractions, marked by adequatefluorescent antibodies (green light for leukemic T-cells,red light for normal lymphocytes), confirm the conclu-sion, mentioned above, about the type of separated cellfractions (Plate 1). Moreover, it was observed that peak 2in chromatogram on SBA-free Sepharose 6MB corre-sponded predominantly to leukemic T-cells. It was alsofound that the leukemic T-cells interacted selectivelywith FITC-CD90 antibody (green light), and slightly withPE-CD44 (red light), whereas the normal lymphocytesinteracted only with PE-CD44.

The data about separation and identification of anotherleukemic T-cell lines, MOLT-4 and RPMI-8402 wereanalogous to these for Jurkat.

)�� +�)��/�� +�)��+�� 0(� � ��

In several independent chromatographic procedures,using normal lymphocytes or leukemic T-cell linesseparately, we checked the affinity of both columns tothe normal, Jurkat, MOLT-4 or RPMI-8402 cells,respectively.

As it is seen from Table 1, more than 95% of normallymphocytes were easily eluted by PBS(�) on SBA-conjugated Sepharose 6MB. In the case of leukemicT-cell lines, the concentration of cells retained on theSBA-conjugated Sepharose beads was about 80–100%

Figure 1. Separation of normal lymphocytes from leukemicT-cells by SBA-conjugated (A) and SBA-free (B) Sepharose6MB-chromatograms. Chromatographic conditions: a non-mobile phase–CNBr-activated Sepharose 6MB, conjugated ornon-conjugated with SBA; a mobile phase–PBS(�), containinggraduating concentration of N-acetyl-D-galactosamine or lowconcentrated acetic acid; flow rate, 1.2 mL/min. The mixture ofnormal and leukemic T-cells (2 � 106 cells/mL; 1:1, v:v) wasadded to the column. After 15 min incubation at RT the cellswere removed by eluents and the fractions were detectedspectrophotometrically at 600 nm. The chromatograms charac-terize the fractionation of mixture of normal lymphocytes andJurkat cells. In the case of fractionation of normal lymphocytesfrom MOLT-4 or RPMI-8402, respectively, the chromatogramshad a similar profile.



Plate 1. Microscopic analysis of normal lymphocytes and leukemic T-cells before and after separation by SBA-conjugated and SBA-free Sepharose 6MB. (A) Normal lymphocytes before separation; (B) leukemic T-cells before separation; (C) peak 1 after separationon SBA-conjugated Sepharose 6MB; (D) peak 2 after separation on SBA-conjugated Sepharose 6MB; (E) peak 3 after separation onSBA-conjugated Sepharose 6MB; (F) peak 1 after separation on SBA-free Sepharose 6MB; (G) peak 2 after separation on SBA-freeSepharose 6MB. Specific fluorescent antibodies (PE-CD44 for normal lymphocytes, red light; and FITC-CD90 for leukemic T-cells,green light) were used also to identify normal and leukemic T-cells, using fluorescent microscopy. The results characterize the type ofcells corresponding to the chromatograms in Fig. 1 (normal plus Jurkat). In the case of separation of normal lymphocytes fromMOLT-4 or RPMI-8402, respectively, the results from microscopic analysis were similar.

Copyright 2003 John Wiley & Sons, Ltd. Biomed. Chromatogr. 17 (2003)

Lectin-affinity chromatography and separation of leukemic T-cells ORIGINAL RESEARCH

Plate 2. Binding of normal lymphocytes and leukemic T-cells (Jurkat,MOLT-4, RPMI-8402) with free FITC-SBA. (A) Spectrofluorimetricdetection - fluorescence intensity of FITC-SBA-cell conjugates at538 nm (�ex = 485 nm) as a function of FITC-SBA concentration; (B)microscopic detection, in the case of leukemic T-cells manyfluorescent FITC-SBA-cell conjugates were detected (B1, greenlight), whereas in the case of normal cells at the same experimentalconditions fluorescent conjugates were not observed (B2).

Plate 3. Binding of leukemic T-cells (Jurkat, MOLT-4, RPMI-8402) on SBA-conjugated Sepharose beads. Note that in thecase of leukemic T-cells many fluorescent FITC-SBA-cellconjugates were detected on the Sepharose beads after filtrationwith PBS(�). In the case of normal lymphocytes at the sameexperimental conditions we did not detect such fluorescentconjugates as well as cells on Sepharose beads (data notshown).

Copyright 2003 John Wiley & Sons, Ltd. Biomed. Chromatogr. 17 (2003)

ORIGINAL RESEARCH R. Bakalova and H. Ohba

from their initial concentration, added to the column. Itwas found also that more than 70% of retained leukemicT-cells could be removed by N-acetyl-D-galactosamineor low-concentration acetic acid.

With SBA-free Sepharose 6MB, more than 95% ofnormal lymphocytes as well as leukemic T-cells (Jurkat,MOLT-4, RPMI-8402) were removed easily by PBS(�),without any retention in the column (Table 2).

�� +�)

The binding of normal lymphocytes and leukemic T-cellswith free SBA was examined and visualized qualitativelyby fluorescent microscopy, using FITC as a fluorescentmarker (Plate 2). It was observed that normal lympho-cytes did not interact with FITC-SBA in concentrations

applied (up to 0.3 �M) [Plate 2(B2)]. The leukemic T-cells interacted with FITC-SBA and fluorescent SBA-cellconjugates were registered [Plate 2(B1), green light].

The degree of binding between free SBA and normalor leukemic T-cells as a function of FITC-SBA concen-tration was examined also by spectrofluorimetric method[Plate 2(A)]. It was established that the affinity of FITC-SBA to leukemic T-cells increased with increasing lectinconcentration. However, in the case of normal lympho-cytes the fluorescence intensity was at the baseline leveland fluorescent SBA-cell conjugates were not registeredfluorimetrically at 0.15–0.3 �M FITC-SBA.

�� +�)��/��+�� 0(�

The result in Plate 3 provides direct evidence for binding

Table 1. Specific fractionation of normal lymphocytes and leukemic T-cell lines using SBA-conjugated Sepharose 6MB

Cell lineCells applied to

the columnCells eluted by PBS(�) (percentage

from initial concentration)Elution by saccharides or acetic acid,

(percentage from binding cells)

Normal 2 � 106 96.2 � 2.2 ndJurkat 2 � 106 18.5 � 3.7 65–75MOLT-4 2 � 106 22.6 � 6.2 70–80RPMI-8402 2 � 106 nd 70–80

The data are presented as mean � SD.

Table 2. Fractionation of normal lymphocytes and leukemic T-cell lines by SBA-free Sepharose 6MB

Cell lineCells applied to

the columnCells eluted by PBS(�) (percentage

from initial concentration)Elution by saccharides or acetic acid,

(percentage from binding cells)

Normal 2 � 106 96.5 � 2.4 ndJurkat 2 � 106 95.8 � 2.8 ndMOLT-4 2 � 106 95.7 � 1.5 ndRPMI-8402 2 � 106 96.3 � 1.7 nd

The data are presented as mean � SD.

Table 3. Percentage lysis of normal lymphocytes and leukemic T-cells before (spontaneous lysis) and after passing throughSBA-conjugated Sepharose 6MB

Cell lineCell lysis before passingthrough the column (%)

Cell lysis after passing throughthe column (%) p

Normal15 min at RT 11.10 � 4.05 13.74 � 5.85 ns24 h at 4°C 18.24 � 5.43 20.51 � 6.38 ns

Jurkat15 min at RT 35.14 � 7.62 39.83 � 6.30 ns24 h at 4°C 37.51 � 6.37 56.20 � 7.69 ns

RPMI-840215 min at RT 27.25 � 6.12 30.07 � 5.22 ns24 h at 4°C 28.48 � 5.23 42.14 � 8.17 ns

MOLT-415 min at RT 22.87 � 5.53 94.12 � 4.76 �0.00124 h at 4°C nd nd —

The percentage of lysis was calculated from flow cytometric data and is presented as mean � SD; ns, non significant.



Figure 2. (A) Flow cytometric assay of normal lymphocytes and leukemic T-cells before and after separation by SBA-conjugatedSepharose 6MB. (a) Normal lymphocytes before separation; (b) leukemic T-cells before separation; (c) normal lymphocytes �leukemic T-cells before separation; (d) peak 1 after separation; (e) peak 2 after separation; (f) peak 3 after separation. (B) Flowcytometric assay of normal lymphocytes and leukemic T-cells before and after separation by SBA-free Sepharose 6MB. (a) Normallymphocytes before separation; (b) leukemic T-cells before separation; (c) normal lymphocytes � leukemic T-cells before separation;(d) peak 1 after separation; (e) peak 2 after separation. Quadrants A contain viable normal cells, quadrants B contain viable leukemicT-cells, quadrants C contain all viable cells, and quadrants D contain dead cells. Specific fluorescent antibodies (PE-CD44 for normallymphocytes, and FITC-CD90 for leukemic T-cells) were used also to identify fluorimetrically normal lymphocytes and leukemicT-cells. The log red and log green signals were analyzed using Beckman Coulter-Epics XL flow cytometer. The histogramscharacterize the type of cells corresponding to the chromatograms in Fig. 1 (normal plus Jurkat). In the case of separation of normallymphocytes from MOLT-4 or RPMI-8402, respectively, the histograms had a similar profile.



of leukemic T-cells to the SBA-conjugated Sepharose6MB—there were many fluorescent FITC–SBA–cellconjugates adsorbed to the gel beads after filtration byPBS(�). In the case of normal lymphocytes, in the sameexperimental conditions we did not detect such fluor-escent FITC–SBA–cell conjugates as well as non-fluorescent cells on SBA-saturated Sepharose beads.

2��

Comparing the cell viability before and after separationon the SBA-conjugated Sepharose (Table 3), it was foundthat the lysis of normal, Jurkat and RPMI-8402 cells doesnot change significantly after incubation in the column

Figure 2. Continued.



for 15 min at RT or 24 h at 4°C. However, the lysis ofJurkat and RPMI-8402 cells tends to increase afterincubation of cells in the column for 24 h at 4°C. Thelysis of MOLT-4 increased significantly after removingfrom the column, because the elution was carried out byacetic acid. In this case, the decrease in cell viability wasnot an effect of SBA-conjugated adsorbent.

"�+$#++� !

This study describes a simplified chromatographicprocedure for discrete separation of normal lymphocytesfrom leukemic T-cell lines (Jurkat, MOLT-4, RPMI-8402, derived from ALL patients), based on theirdifferent affinity to soybean agglutinin. The methodproceeds from the concept that the lectins with theirmonosaccharide specificity can recognize fine differ-ences in many complex structures on leukemic T-cells,without any influence and interaction with normallymphocytes (Gabius et al., 1988; Mody et al., 1995;Gabius, 1997; Moriwaki et al., 2000; Sallay et al., 2000).

The earlier published methods for in vitro separation ofleukemic and normal T-cells are based mainly on the sizeor different electrophoretic mobility of cell populations.However, these methods were found unsatisfactorybecause of a non-well-defined separation between normaland leukemic T-cells, a result of the heterogeneity of themost leukemic cell lines.

The use of lectin-affinity chromatography for separa-tion of cell populations, and especially for separation oftumor cells, has been well known from about threedecades (Pereira and Kabat, 1979). A separation oferythrocytes, osteoclasts, intact tissue culture cells bylectin-affinity chromatography has also been described(Killion and Kollmorgen, 1976; Pereira and Kabat 1979;Itokazu et al., 1991). Recently, monoclonal antibody-coated magnetic beads, based on a similar principle, havealso been applied for separation of tumor cells (Wigzellet al., 1972; Kemmner et al., 1992). In leukemia researchlectin-affinity chromatography is widely useful forseparation and isolation of glycoproteins and glycolipids,specific for leukemic cell surface (Boldt, 1982; Dahmsand Hart, 1986; Volman et al., 1987; Fischer et al., 1988;Reading, et al. 1988). However, despite much researchover several decades, the use of lectin-affinity adsorbentsfor separation of normal lymphocytes from leukemicones, which is potentially applicable for clinical practicein leukemia, is only beginning to be exploited.

There are several problems with the use of lectin-affinity chromatography for separation of cell popula-tions: (i) for the preparation of distinct affinity absorbent,a covalent coupling of lectins is required; (ii) sometimesthe cells with specific receptors to lectins are so stronglybound to the affinity column that they cannot be removedby physiological eluents; (iii) it is necessary to select the

optimal concentration ratio cells/lectin, bearing in mindthe dissociation constants of the complex between them;(iv) it is not clear whether the immobilized lectins have asimilar affinity to tumor cells as free lectins; and (v) thereleased cells have to keep their viability.

Based on all these difficulties, we attempted to developand select a lectin-affinity chromatographic column fordiscrete fractionation of normal lymphocytes fromleukemic T-cell lines, derived from ALL patients. Afterscreening of different adsorbents our choice was onCNBr-activated Sepharose 6MB as a non-mobile phase,because of easy coupling with lectins and easy blockingof non-saturated active sites. The choice of SBA pro-ceeded from our preliminary experiments, demonstratingthat even at high concentrations SBA interact withleukemic T-cells, without any interaction with normalones. SBA also manifested well-defined target cytotoxi-city against leukemic T-cells, without any effect on theviability of normal lymphocytes (data not shown).

The gel was covalently coupled to SBA, then served asan affinity adsorbent for fractionation of normal lympho-cytes from leukemic T-cells, depending on their variousspecificity to SBA. The main advantage of the methoddescribed here is that CNBr-activated Sepharose 6MB canbe used for covalent coupling to a variety of lectins ofvarious specificity to leukemic T-cells. The mobile phaseconsisted of PBS(�), containing N-acetyl-D-galacto-samine or low-concentrated acetic acid in graduatingconcentrations. Using SBA-conjugated Sepharose 6MBwe fractionated discretely normal lymphocytes fromleukemic T-cells, Jurkat, MOLT-4 and RPMI-8402,identified and proved by flow cytometry and fluorescentmicroscopy, using adequate fluorescent antibodies.

It was observed that normal lymphocytes were easilyremoved from the lectin-affinity column in the first2–3 min after beginning of the elution by PBS(�). About3–4% of normal lymphocytes were retained on the SBA-affinity column. It is probably a result of non-specificbound of cells to Sepharose beads, because the samepercentage of normal lymphocytes was retained on SBA-free Sepharose 6MB, too (Tables 1 and 2). In severalcontrol experiments we established that the percentage ofnormal cells retained on SBA-free or SBA-conjugatedSepharose depended on the conditions of the blockingstep during preparation of the columns. It appears that thenon-specific retention of normal cells to SBA-free orSBA-conjugated Sepharose is a result of their interactionwith the active (non-blocked) CNBr-groups of theadsorbent. A similar result was also obtained in the caseof leukemic T-cells and SBA-free Sepharose.

In contrast to normal lymphocytes, more than 80% ofthe leukemic T-cells were retained on SBA-conjugatedSepharose 6MB and were removed in 70–80% by N-acetyl-D-galactosamine (in the case of Jurkat and RPMI-8402) or low concentrated acetic acid (in the case ofMOLT-4). About 20% of Jurkat and MOLT-4 cells were



easily removed from SBA-affinity column by PBS(�) inthe mixture with normal lymphocytes, whereas RPMI-8402 cells were totally retained. It was established alsothat about 20% of leukemic T-cells were strongly boundto the SBA-conjugated Sepharose beads and could bereleased only mechanically. However, this destroyed theintegrity of the gel beads, visualized microscopically.The inability to remove the strongly bound cells from thecolumn may be due to strong secondary interactionsbetween leukemic cells and gel beads. More than 95% ofthe leukemic T-cells were not retained on SBA-freeSepharose 6MB and were released in the mixture withnormal lymphocytes in the first 2–3 min after thebeginning of elution by PBS(�). Therefore, it wasassumed that the retention of leukemic T-cells, Jurkat,MOLT-4 and RPMI-8402, on the SBA-affinity columnwas a result of their specific interaction with SBA.

The binding of leukemic T-cells with SBA-conjugatedSepharose beads was proved directly and visualized byfluorescent microscopy, using FITC-SBA as a fluorescentmarker (Plate 3, green light). Many fluorescent SBA-cellconjugates were observed on the gel beads in the case ofleukemic T-cells, and no conjugates were detected in thecase of normal lymphocytes.

In a separate experimental protocol we established thatthe normal lymphocytes also did not interact with freeSBA in concentrations applied (up to 0.3 �M), whereasthe leukemic T-cells interact with free SBA to a differentextent, depending on the cell line and SBA concentration(Plate 2).

Comparing the degree of binding of different lines ofleukemic T-cells with free SBA (RPMI-8402 = MOLT-4� Jurkat) (i) and the percentage of retention on SBA-conjugated Sepharose 6 MB (MOLT-4 � Jurkat �RPMI-8402) (ii), no significant correlation was foundbetween the parameters. However, this does not affect theconclusion that the retention of leukemic T-cells on SBA-conjugated Sepharose 6MB is a result of cell–lectincomplex formation. Using fluorescent microscopy wefound that about 10–20% of Jurkat and MOLT-4 cells didnot interact with free FITC-SBA. The cell–SBAconjugates also had different fluorescent intensities,depending on the type of leukemic T-cells (data notshown). It appears that the leukemic T-cells are veryheterogeneous with respect to the number and/or kind ofSBA-receptors on their surface, resulting in differentpossibility for interaction with free and/or immobilizedSBA. This can explain at least partially the ease of elutionof about 20% of Jurkat and MOLT-4 cells by PBS(�)from SBA-affinity column (Table 1). Unfortunately, atpresent we cannot discuss whether the easily removedleukemic T-cells have a high carcinogenecity or they areharmless, which is a crucial point for clinicians. Usingflow cytometry and microscopy (tripan blue staining) weestablished that easily removed leukemic cells were dead,which is a promising result.

An important consideration for separation of cellpopulations by affinity methods is the complete recoveryof the cells and maintaining their unique functions afterpassing through the column. In this case, physiologicaleluents are required. Comparing the viability of normallymphocytes and leukemic T-cells before and afterincubation with SBA-conjugated Sepharose 6MB, weobserved that the lysis of normal lymphocytes andleukemic T-cells does not change significantly after15 min incubation. However, there was a tendency toincrease the lysis of Jurkat and RPMI-8402 cells after24 h incubation with the column. It may be speculatedthat this is a result of manifestation of SBA cytotoxicity,but this possibility is disputable and requires verification.In the case of elution of MOLT-4 cells from SBA-conjugated Sepharose 6MB we used acetic acid andobserved that the lysis of T-cells increased markedly afterelution. This is only an effect of eluents on the viability ofMOLT-4 cells. Thus, the problem with the methoddescribed is that not all bound leukemic T-cells can beremoved by physiological eluents and keep their viabilityafter passing through the column.

Comparing the present results with the data aboutfractionation of normal and leukemic T-cells on DBA-affinity column (Ohba et al., 2002), we consider thatSBA-affinity adsorbent is the better choice, because ofthe potential to combine a discrete cell separation with aspecific target cytotoxicity of SBA against leukemiccells, without any influence on the viability of normallymphocytes. As DBA does not express cytotoxic effectagainst leukemic T-cells even in high concentrations(data are not shown), it is probably most suitable as anaffinity probe in the case of separation of embryonic stemcells.

$ !$*#+� !

The present study makes no claim to be original,describing a separation of cell populations using lectin-affinity chromatography. The novelty is in the selectionof affinity adsorbent, resulting in well-defined fractiona-tion of normal lymphocytes from leukemic T-cells,applied in a mixture to the lectin-affinity column, withhigh percentage of cell recovery. The chormatographicprocedure is simple and applicable to many otherleukemic cells, and the affinity adsorbent does not in-fluence the viability of normal lymphocytes. Since manyleukemic T-cell clones are heterogenous in their size andelectrophoretic mobility, the lectin-affinity adsorbentsare more suitable for discrete cell fractionation than flowcytometry. The method described is comparable inprinciple to antibody-coated magnetic beads. However,negative selection of tumor cells by means of beadscoated with a leukocyte binding antibody offers noadvantage since it has been established that dead cells



and erythrocytes remain (Kemmner et al., 1992). It hasbeen found that the lectin-affinity adsorbents interactwith erythrocytes only if they are coupled with bloodgroup glycoproteins (Pereira and Kabat, 1979). Analo-gous to the antibodies the lectins are also applicable inseparation of cell populations, using lectin-coatedmagnetic beads.

The methodology described in this paper allowsidentification of the type and viability of cells in eachfraction, as well as to estimation of the degree ofpurification of the fraction of normal lymphocytes. As theefforts of clinicians and researchers in leukemia aredirected to avoiding relapse by selective in vitro removaland utilization of leukemic cells from the autologousgraft in transplantations, this methodology is potentiallyuseful for clinical practice. Moreover, the lectin-affinitychromatography may be considered as a new generationof method in leukemia research, because it potentiallycombines a well-defined cell separation (i) with specificcytotoxic effect of lectins on leukemic cells, without anyeffect on normal ones (ii).

)��

The technical assistance of Mrs. Kawahara of NationalInstitute of Advanced Industrial Science and Technol-ogy-Kyushu is gratefully acknowledged. This study wassupported in part by the JSPS Invitation FellowshipProgram for Research in Japan.

�%,%�%!$%+

Boldt DH. Chronic lymphocytic leukemia: demonstration of cellsurface structural heterogeneity by lectin affinity chromatography ofcell extracts. Leukemia Research 1982; 6(4): 499.

Bouckaert J, Hamelryck T, Wyns L and Loris R. Novel structures ofplant lectins and their complexes with carbohydrates. CurrentOpinion in Structural Biology 1999; 9: 572.

Chay CH and Pienta KJ. Evidence for lectin signaling to the nuclearmatrix: cellular interpretation of the glycocode. Journal of CellBiochemistry 2000; 35: S123.

Dahms NM and Hart GW. Influence of quaternary structure onglycosylation. Differential subunit association affects the site-specific glycosilation of the common beta-chain from Mac-1 andLFA-1. Journal of Biological Chemistry 1986; 261(28): 13186.

Fischer J, Klein PJ and Uhlenbruck G. Detection and characterizationof glycoconjugates in normal and malignant tissues by lectins. ActaHistochemica 1988(Suppl.); 86: 191.

Forsberg LS and Macher BA. Lectin affinity fractionation ofasparagine-linked oligosaccharides from normal human and chronicleukemic leukocytes. Cancer Biochemistry and Biophysics 1987; 9:155.

Gabius HJ. Endogenous lectins in tumors and their possible importancein the diagnosis and therapy of cancers. Onkologie 1989; 12(4): 175.

Gabius HJ. Concepts of tumor lectinology. Cancer Investigations1997; 15: 454.

Gabius H. Probing the cons and pros of lectin-induced immunomo-dulation: Case studies for the mistletoe lectin and galectin-1.Biochimie 2001; 83: 659.

Gabius HJ, Engelhard R and Cramer F. Endogenous tumor lectins: anew class of tumor markers and targets for therapy? MedicalHypotheses 1985; 18(1): 47.

Gabius HJ, Engelhard R and Cramer F. Endogenous tumor lectins:overview and perspectives. Anticancer Research 1986; 6(4): 573.

Gabius HJ, Vehmeyer K, Gabius S and Nagel GE. Clinical applicationof various plant and endogenous lectins to leukemia. Blood 1988;56: 147.

Hozumi M. Fundamentals of chemo-differentiation therapy of myeloidleukemia. Anticancer Research 1994; 14(3A): 1177.

Itokazu M, Tan A and Tanaka S. Separation of osteoclasts by lectinaffinity chromatography. Nippon Seikeigeka Gakkai Zasshi 1991;65: 762.

Kemmner W, Moldenhauer G, Schlag P and Brossmer R. Separation oftumor cells from a suspension of dissociated human colorectalcarcinoma tissue by means of monoclonal antibody-coated magneticbeads. Journal of Immunological Methods 1992; 147: 197.

Killion JJ and Kollmorgen GM. Isolation of immunogenic tumor cellsby cell-affinity chromatography. Nature 1976; 259: 674.

Lee N, Wang WC and Fukuda M. Granulocytic differentiation of HL-60 cells is associated with increase of poly-N-acetyllactosamine inAsn-linked oligosaccharides attached to human lysosomal mem-brane glycoproteins. Journal of Biological Chemistry 1990;265(33), 20476.

Marmont AM. Stem cell transplantation for severe autoimmunediseases: progress and problems. Haematologica 1998; 83(8): 733.

Misra RR, Pinsky PF and Srivastava S. Prognostic factors forhematologic cancers. Hematology and Oncology Clinics of NorthAmerica 2000; 14(4): 907.

Mody R, Joshi S and Chaney W. Use of lectins as diagnostic andtherapeutic tools for cancer. Journal of Pharmacology andToxicology Methods 1995; 33: 1.

Moriwaki S, Ohba H, Nakamura O, Sallay I, Suzuki M, Tsubouchi H,Yamasaki N and Itoh K. Biological activities of the lectin, abrin-a,against human lymphocytes and cultured leukemic cell lines.Journal of Hematotherapy and Stem. Cell Research 2000; 9: 47.

Neame PB, Soamboonsrup P, Quigley JG and Pewarchuck W. The useof monoclonal antibodies and immune markers in the diagnosis,prognosis, and therapy of acute leukemia. Transfusion MedicineReview 1994; 8(1): 59.

Ohba H, Bakalova R, Moriwaki S and Nakamura O. Fractionation ofnormal and leukemic T-cells by lectin-affinity column chroma-tography. Cancer Letters 2002; 184: 207.

Pereira MEA and Kabat EA. A versatile immunoadsorbent capable ofbinding lectins of various specificities and its use for the separationof cell populations. Journal of Cell Biology 1979; 82: 185.

Petros WP and Evans WE. Pharmacokinetics and pharmacodynamicsof anticancer agents: contributions of the therapy of childhoodcancer. Pharmacotherapy 1990; 10(5): 313.

Reading CL, Hickey CM and Yong WB. Analysis of cell surfaceglycoprotein changes related to hematopoetic differentiation.Journal of Cell Biochemistry 1988; 37(1): 21.

Reed CJ. Apoptosis and cancer: strategies for integrating programmedcell death. Seminars in Hematology 2000; 37((4): Suppl.): 9.

Roth J, Zuber C, Komminoth P, Sata T, Li WP and Heitz PU.Applications of immunogold and lectin-gold labeling in tumorresearch and diagnosis. Histochemistry and Cell Biology 1996;106(1): 131.

Sallay I, Moriwaki S, Nakamura O, Yasuda S, Kimura M, Yamasaki N,Itoh K and Ohba H. Interaction of the hemolytic lectin, CEL-III,with cultured human leukemic cell lines. Journal of Hematotherapyand Stem Cell Research 2000; 9: 877.

Santos GW. Problems and strategies for bone marrow transplantationin acute leukemia and chromic myelogenous leukemia. CancerDetection and Prevention 1988; 12((1–6)): 589.

Santos GW. Bone marrow transplantation in hematologic malignan-cies. Cancer 1990; 65(Suppl.): 786.

Solary E, Droin N, Bettaieb A, Corcos L, Dimanche-Boitrel MT andGarrido C. Positive and negative regulation of apoptotic pathwaysby cytotoxic agents in hematological malignancies. Leukemia 2000;14(10): 1833.

Sweetenham JW. The importance of dose and schedule in cancerchemotherapy: haematological cancer. Anticancer Drugs 1995;6(Suppl.): 7.

Vaickus L, Ball ED and Foon KA. Immune markers in hematologicmalignancies. Critical Reviews in Oncology and Hematology 1991;11(4): 267.



Volman GN, Narasimhan S and Letarte M. Analysis of asparagine-linked oligosaccharide structures of chronic lymphocytic leukemiacells. Molecular Immunology 1987; 24(8): 871.

Wigzell H, Sundqvist KG and Yoshida T. Separation of cells accordingto surface antigens by the use of antibody-coated columns.Fractionation of cells carrying immunoglobulins and blood groupantigen. Scandinavian Journal of Immunology 1972; 1: 75.



interaction of soybean agglutinin with leukemic t-cells and its use for their in vitro separation...

Documents