Generation of activated natural killer (A-NK) cells in patients with chronic myelogenous leukaemia...

Generation of activated natural killer (A-NK) cells in patientswith chronic myelogenous leukaemia and their rolein the in vitro disappearance of BCR/abl-positive targets

LUCIA M. R. SILLA,7,8 STEVEN M. PINCUS,1,2,6 JOSEPH D. LOCKER,3,6 JANICE GLOVER,3 ELAINE M. ELDER,6

ALBERT D. DONNENBERG,1,2,6 N. B. NARDI,7,8 JOHN BRYANT,5,6 EDWARD D. BALL1,2,6

AND THERESA L.WHITESIDE

3,4,6 Departments of 1Hematology/Bone Marrow Transplantation, 2Medicine, 3Pathology, 4Otolaryngology,5Mathematics and Biostatistics, University of Pittsburgh School of Medicine and 6Pittsburgh Cancer Institute, Pittsburgh,Pennsylvania, U.S.A., and 7Department of Internal Medicine, Faculty of Medicine, Pontificia Universidade Catolica, RS,and 8Department of Genetics, Federal University of Rio Grande do Sul, Brazil

Received 19 April 1995; accepted for publication 15 January 1996

Summary. Activated natural killer (A-NK) cells, a subset ofCD56dimCD3– lymphocytes, are obtained from PBMC ofnormal donors by adherence to plastic and culture in thepresence of IL2. In this study we tested the feasibility ofgenerating A-NK cells in patients with Ph+ chronic myeloidleukaemia (CML). Cultures obtained from patients with earlychronic phase (ECP; n=7) contained a mean (�SD) of 83�7%of CD56+CD3– cells, and those from patients with advancedchronic phase (ACP; n=7) contained 27�33% CD56+CD3–

cells. In three patients with leukaemia in a blastic phase (BP) itwas only possible to obtain one culture enriched inCD56+CD3– cells (81%). Cellular aggregates of myeloid cellsand large granular lymphocytes were observed in early A-NKcell cultures. Paired freshly-adherent and cultured A-NK cellswere tested for the presence of BCR/abl mRNA by RT-PCR. TheBCR/abl+ cells were detected in all 12 preparations of thefreshly adherent A-NK cells tested. In 6/12 the BCR/abl+ cellswere no longer detectable by RT-PCR on day 14 of culture.Both proliferation and antileukaemic cytotoxicity were

significantly higher (P = 0.002 and P = 0.029, respectively)in the BCR/abl– cultures than those in the six BCR/abl+

cultures. 5/6 BCR/abl– cultures were highly enriched in A-NKcells on day 14, and 1/6 contained predominantlyCD56+CD3+ cells. Only 2/6 BCR/abl+ cultures were enrichedin A-NK cells on day 14, but they had poor cytotoxicity and alow proliferative index. Myeloid cells (CD33+) were morefrequently detected in the BCR/abl+ than BCR/abl– A-NK cellcultures (P = 0.028). These observations suggest that: (1)populations of benign A-NK cells can be generated from theperipheral blood of CML patients; (2) the ability to generate A-NK cells is impaired in patients with advanced CML; and (3)the ability to generate A-NK cells with antileukaemic activitycorrelates with the disappearance of BCR/abl+ cells from thesecultures.

Keywords: chronic myelogenous leukaemia, BCR/abl, acti-vated natural killer (A-NK) cells, elimination of BCR/abl+

targets, generation of A-NK cells in CML.

Chronic myelogenous leukaemia (CML) is a clonal disorder inwhich neoplastic transformation of a stem cell results in theproliferation and accumulation of myeloid cells and theirprecursors. The hallmark of CML, found in >95% of patients,is the Philadelphia chromosome (Ph1) derived from areciprocal translocation between chromosome 9 and 22,t(9:22)(q34:q11) (Nowell & Hungerford, 1960; Rowley,

1973), which gives rise to a fused BCR/abl gene (Shtivelmanet al, 1985). The disease may have a biphasic or triphasiccourse, with an initial chronic phase, which lasts approxi-mately 3 years (although it may last considerably longer) andis easily controlled with therapy. It eventually progresses to apoorly-defined accelerated phase, lasting less than 1–1.5years but sometimes longer, and then to a blastic phase,leading to death generally within 3–6 months. 20–25% ofpatients die from complications of accelerated phase, andanother 20–25% may develop a blastic phase without theintermediate accelerated phase (Kantarjian et al, 1993).

British Journal of Haematology, 1996, 93, 375–385

375# 1996 Blackwell Science Ltd

Correspondence: Dr Theresa L. Whiteside, Pittsburgh CancerInstitute, W1041 Biomedical Science Tower, 211 Lothrop Street,Pittsburgh, PA 15213-2582, U.S.A.

The only curative treatment for CML is allogeneic bonemarrow transplantation (BMT), available to a minority ofpatients. Alpha-interferon (�-IFN) leads to durable remissionin selected patients, and its utilization, alone or in combina-tion with other drugs, may have an important role in theincreased survival of CML patients observed in the last fewyears (Kantarjian et al, 1993). Nevertheless, the treatment ofCML remains a challenge. Although autologous BMT (ABMT)may prolong survival (Buttirini et al, 1990), a high relapserate remains the major problem of this procedure. Reinfusionof clonogenic tumour cells with marrow (Deisseroth et al,1994), and the lack of graft-versus-leukaemia (GVL) effect(Ringden & Horowitz, 1989), may be responsible for thehigher relapse rate seen with autologous BMT than withallogeneic transplantation. Although the role of allogeneic Tcells in graft-versus-host disease (GVHD) has been firmlyestablished, GVL responses might be mediated, in part, byother effector cells.

Recently, several mechanisms have been proposed toexplain the GVL effect (Buttirini et al, 1987; Antin, 1993),including the involvement of NK cells ( Jian et al, 1993). NKcells are among the first cells to appear in the circulation afterBMT (Reittie et al, 1989) or ablative chemotherapy (Laughlinet al, 1993). This early appearance of NK cells might beimportant both for engraftment and GVL response, becauseNK cells have been shown to be able to promote allogeneicbone marrow engraftment as well as mediate antitumoureffects without producing GVHD (Murphy et al, 1993).

We have recently described a subset of human NK cells(named A-NK cells) that have potent antileukaemia activityin vitro (Sedlmayr et al, 1992). A-NK cells are derived fromprecursor cells with the ability to adhere to solid surfaceswithin minutes of activation in the presence of interleukin 2(IL2). This rapid early adherence of A-NK cells is a usefuldiscriminating characteristic between subpopulations of IL2-responsive NK cells. A-NK cells have higher expression ofcytokine genes (Vitolo et al, 1993), higher in vitro cytotoxicityagainst both NK-sensitive and NK-resistant tumour celltargets (Vujanovic et al, 1993), and higher in vivo antitumouror antimetastatic activities than non-adherent, IL2-activatedNK cells (Yasumura et al, 1994).

This study was designed to test the feasibility ofgeneration of A-NK cells from the peripheral blood ofpatients with CML and to determine antileukaemia activityof these effector cells. Our results indicate that the ability togenerate A-NK cells in patients with CML diminishes withprogression of the disease. Aggregates or ‘rosettes,’ formedby myeloid and lymphoid cells and seen in the early A-NKcell cultures, served as an index of autologous antitumouractivity. These rosettes were seen only in cultures of patientswith CML, and they disappeared as A-NK cells proliferatedand became the predominant cell type in culture. The initialplastic-adherent cells as well as A-NK cells harvested at theend of culture were assayed by RT-PCR for the presence ofBCR/abl+ cells. All cultures tested at establishment wereBCR/abl+; however, after 14 d the cultures enriched inhighly active A-NK cells became BCR/ablÿ, whereas thosecontaining poorly-proliferating or poorly-cytotoxic A-NKcells remained BCR/abl+. Our results suggest that A-NK

cells obtained from the peripheral blood of patients withearly CML and expanded in vitro develop powerful antileu-kaemic activity and appear to be able to eliminate BCR/abl+cells.

MATERIALS AND METHODS

Patient population. Peripheral blood mononuclear cells(PBMC) were obtained from 17 patients with Ph1-positiveCML and eight healthy volunteers (NC). Seven patients werestudied within 1 year after diagnosis and were designatedECP, and seven were studied from 14 months to 5 years afterdiagnosis (ACP). Three patients were in the blastic phase(BP). There were no patients in the accelerated phase ofCML in this study. Of the patients in BP, two had their blooddrawn immediately before starting the conditioning regi-men for BMT, while in untreated blast crisis. The third wason daily prednisone and weekly vincristine, as a part of aninduction treatment for lymphoid blast crisis. At the timethe sample was collected, the patient was recovering froman aplastic period and her WBC was 1.4 � 109/l with noblasts. Three of the ECP patients were treated with �-IFNand four with hydroxyurea. Among the patients in ACP, onewas treated with �-IFN and hydroxyurea, and six withhydroxyurea only. The WBC at the time of study werecomparable in eight patients with ECP (15.1�14.4 � 109/l), seven with ACP (8.6�3.4 � 109/l) and three with BP(9.1� 0.4 � 109/l). All patients were seen for pre-BMTevaluation at the BMT Unit of the Pittsburgh CancerInstitute. All patient and control samples were obtainedwith informed consent, using guidelines approved by theIRB of the University of Pittsburgh Medical Center,Pittsburgh, Pa.

Generation of plastic-adherent NK cells (A-NK). A-NK cellswere obtained by a modification of the method described byus previously (Vujanovic et al, 1993). PBMC were depletedof monocytes by incubation in 5 mM solution of L-phenylalanine methyl ester hydrochloride (Terumo MedicalCorporation, Elkton, Md.) for 30 min at room temperature(RT). The monocyte-depleted PBMC were then depleted of Tcells by panning on anti-CD3 antibody-coated flasks (AISMicroCellector, Applied Immune Sciences Inc., Santa Clara,Calif.) for 1 h at RT. Monocyte- and T-cell-depleted PBMCwere next adjusted to a concentration of 1.5�106 cells/mlof the complete tissue culture medium (TCM) supplementedwith 6000 IU/ml (22 nM) of human recombinant IL2 (rIL2,Chiron/Cetus, Emeryville, Calif.). The medium consisted ofRPMI 1640 supplemented with 10% (v/v) heat-inactivatedpooled normal human AB serum (NABI, Miami, Fla.), 2 mM

L-glutamine, 50 U/ml penicillin, 50 �g/ml streptomycin,and 25 mM Hepes buffer (Gibco, Grand Island, N.Y.)supplemented with 6000 IU/ml (22 nM) of human rIL2.The cells were incubated in the medium for 3–5 h at 378Cin an atmosphere of 5% CO2 in air. Next, nonadherent NK(NA-NK) cells were decanted, A-NK cells were washed fourtimes with warm RPMI and cultured for 14 d in fresh TCMsupplemented with IL2 (6000 IU/ml). Irradiated allogeneicfeeder cells (Con A-activated allogeneic PBMC) prepared asdescribed earlier (Rabinowich et al, 1991) were added to the

# 1996 Blackwell Science Ltd, British Journal of Haematology 93: 375–385

376 Lucia M. R. Silla et al

377A-NK Cells in Chronic Myelogenous Leukaemia


culture at a concentration of 106/ml on day 0 or at theinitiation of the culture. A small proportion of cells in eachPBMC preparation was captured by 3–5 h adherence in aseparate T25 flask, and the plastic-adherent cells weredetached by incubation in cold phosphate-buffered saline(PBS) containing 1 mM EDTA at 48C for 10 min, andsubmitted for RT-PCR analysis. Additional samples of theinitial plastic-adherent population were obtained from threecontrols and three patients for phenotypic studies.

Proliferation. The proliferative potential of the cultures wascalculated as previously described (Vujanovic et al, 1993).Briefly, the total number of cells on day 14 of culture wasdivided by the number of initial plastic-adherent cells. Sinceonly a proportion of the initial adherent cells wasCD56+CD3–, the fold of A-NK cell proliferationwas calculatedby dividing the total number of CD56+CD3– cells obtained atthe culture termination by the estimated number of initial A-NK cells.

Phenotype. Samples of the initial adherent or cultured cellswere stained for expression of the following surface antigens:CD56, CD3, CD14, CD16, CD34 and CD15 using fluorescein-or phycoerythrin-labelled monoclonal antibodies (BectonDickinson, Mountain View, Calif.), and anti-CD33 purchasedfrom AMAC Inc. (Westbrook, Maine). Phenotypic analyseswere done with a FACS IV flow cytometer (BD) equipped witha Consort 40 computer. Fluorescein- or phycoerythrin-labelled mouse isotype-matched immunoglobulins werealways used as controls. The data are presented aspercentages of the total viable cell population in the widegate, encompassing the lymphocyte, monocyte and granulo-cyte populations.

Cytotoxicity assays. NK activity of unseparated PBMC or A-NK cells was tested against NK-sensitive K562 and NK-resistant Daudi cell lines (see below) in a 4 h 51Cr releaseassay, as described by us previously (Whiteside et al, 1990).Briefly, target cells were labelled with 100 �Ci of 51Cr (specificactivity, 5 �Ci/ mM; NEN, Boston, Mass.) for 1 h at 378C.Washed target cells (T) were mixed and incubated witheffector cells (E), at the E:T ratios ranging from 50:1 to 6.25:1for NK and from 6:1 to 0.75:1 for A-NK cells. Spontaneousrelease (SR) and maximal release (MR) were determined byincubating target cells in medium alone or in 5% Triton X-100, respectively. The assay was always performed intriplicate. Radioactivity was counted in a gamma counter,and the percentage of specific lysis determined according tothe formula:

% specific lysis �

mean cpm experimental releaseÿ mean cpm SR� 100mean cpm MRÿ mean cpm SR

Lytic units (LU) were calculated using a computer programbased on the formula developed by Pross et al (1981). Onelytic unit was defined as the number of effector cells needed tolyse 20% of 5� 103 target cells, and the number of lytic unitsper 107 effector cells were calculated.

Tumour cell lines. K562, a human myeloid leukaemia cellline and Daudi, a human Burkitt’s lymphoma-derived cellline, were maintained in culture in RPMI-1640 medium

supplemented with 10% (v/v) fetal calf serum (FCS) and usedfor cytotoxicity assays, when the cultures were in the logphase of growth.

Cell aggregates and morphological studies. The formation ofcell aggregates in early cultures of A-NK cells was observedand photographed through an inverted microscope. In threecultures a small proportion of the initial monocyte- and T-cell-depleted PBMC were incubated for adherence in slidechambers (Nunc Inc., Naperville, Ill.). The slides were stainedwith Wright-Giemsa, and the morphology of the cells in theaggregates was recorded.

Co-incubation of A-NK cells with autologous bone marrow cellsand CFU assays. A-NK cells were generated in 14 d culturesfrom PBMC obtained from normal donors. On day 14,autologous bone marrow (BM) was obtained, processed torecover MNC and cultured alone or co-cultured with A-NKcells at the E (A-NK):T (BM) ratios of 10:1, 1:1 and 1:10 inmedium consisting of RPMI-1640 supplemented with 10%(v/v) human AB serum and IL2 (22 nM). The cultures wereincubated for 7 d at 37�C and 5% CO2 in humidified air. At theend of co-incubation period, cell mixtures or control cultureswere washed and plated in methylcellulose culture medium,containing IL3 (10 ng/ml), GM-CSF (10 ng/ml) anderythropoietin (3 U/ml).

The plates were incubated for 14 d and the colonies werescored for CFU-GM, BFU-E and CFU-GEMM, as describedearlier (Pincus et al, 1994). In some experiments, 14 d co-cultures of A-NK cells with BM cells were tested for NKactivity in 4 h 51Cr-release assays and for the presence ofCD56+CD3– cells by flow cytometry, as described above.

Detection of BCR/abl gene rearrangement. Total RNA from theinitial plastic-adherentcells of 12 patients and from the pairedcultured A-NK cells was obtained, using RNAzol (Cinna/Biotecx Laboratories, Houston, Tex.), essentially as describedby the manufacturer. Synthesis of cDNA from approximately1 mg of the resultant RNA was performed with MMLV reversetranscriptase (Gibco-BRL), using an oligonucleotide primer,NCMLB, complementary to the third exon of the abl mRNA.Oligonucleotides were prepared on an Applied BiosystemsInc., model 381A DNA synthesizer; their sequences are givenbelow. DNA sequences used for selection of primers werepublished by Heisterkamp et al (1985), and Shtivelman et al(1986).

Oligonucleotides:

NCMLA: CAA CAT CCC CCT GGT GCC CGANCMLAA: CTC AGC GGC ATT GCG GGA CACNCMLBB: GCT TCT CCC TGA CAT CCT TGGNCMLB: ATG CTA CTG GCC GCT GAA GGG

PBGDA: TGT CTG GTA ACG GCA ATG CGGPBGDAA: AGA CGG ACA GTG TGG TGG CAAPBGDBB: AAC TTT CTC TGC AGC TGG GCTPBGDB: CAG GCC AGC TGT TGC TAG GAT

Half of the reverse transcription product was amplified in aPerkin-Elmer Cetus model 9600 thermocycler for 40 cycles attemperatures of 948C for 30 s, 558C for 20 s, and 728C for30 s, using primers NCMLA and NCMLB. 2�l of the first PCRreaction were added to a nested PCR reaction and amplified

with primers NCMLAA and NCMLBB under the samethermocycling parameters as before. 20�l of the nestedPCR product were fractionated on a 2% agarose gel andvisualized using ethidium bromide. To control for RNAintegrity, parallel amplification of non-erythropoietic porpho-bilinogen deaminase, PBGD (Grandchamp et al, 1987) wasperformed using the primers indicated above. As with theBCR/abl primers, the AA-BB set of PBGD primers are internalto the A-B set. The thermocycling parameters were identicalto those for BCR/abl.

Statistical analyses. Nonparametric methods of hypothesistesting were used throughout the analysis. Two-groupcomparison (e.g. comparison of the phenotype or functionof BCR/abl– or BCR/abl+ cultures) were based on theWilcoxon test, with stratification to control for possibledifferences due to phase of disease. Because of the limitedsample size, P values were determined by exact permutationmethods. Association of cell phenotype, proliferative activityand cytotoxicity with disease phase were tested by means ofJonckheere’s method. For hypothesis tests of effects that havebeen previously described or were predicted prior to testing,one-sided P values are reported; otherwise, two-sided P valuesare given.

RESULTS

A-NK cell enrichment in 14 d cell culturesResults of phenotypic analyses of day 14 A-NK culturesestablished from controls (NC) as well as patients with CMLare shown in Fig 1. Enrichment in A-NK cells (CD56+CD3–)was comparable in NC (85 � 16%) and ECP patients(83�7%). In contrast, ACP patients generated cultureswith fewer CD56+CD3– cells (27�33%), and in 4/7 culturesthe percent of CD56+CD3– cells was <10% (Fig 1). From twoof the BP patients it was not possible to obtain any adherentcells to start the cultures. From the third BP patient,recovering from a chemotherapy-induced aplasia, it waspossible to establish a culture and obtain enrichment in A-NKcells (81%). The percentages of A-NK cells generated in thesecultures declined significantly with the disease duration. Thedisease was highly significant when ACP patients arecompared to NC (P = 0.00005, Jonckheere’s test, one-sided)(Fig 1). Correspondingly, the total content of T cells (CD3+) in14 d cultures significantly increased with the disease phase (P= 0.0002, Jonckheere’s test, one-sided; data not shown).Thus, whereas all NC generated highly-enriched CD56+CD3–

cultures with few CD3+ T cells, patients with CML differedsignificantly from NC in the ability to generate pure culturesof A-NK cells. Only ECP patients generated A-NK cells incultures with a purity comparable to that of NC.

A-NK cell expansionTo determine the fold proliferation of each A-NK cell culture,the total number of cells recovered in 14 d cultures wasdivided by the number of initial plastic-adherent cells after 3–5 h activation in the presence of IL2. As shown in Fig 2A, NCcultures consistently showed significantly better proliferationthan cultures obtained from patients, and this difference in

proliferation was greatest in more advanced disease (P=0.03one-sided Jonckheere’s test). The phenotypic analysis of theinitially adherent cells showed that only a small percentagewere CD56+CD3– A-NK cells in patients with CML (seebelow). Since it was not possible to analyse the initial plasticadherent cells by flow cytometry without compromising theestablishment of the culture itself, we resorted to calculatingthe ‘A-NK cell proliferation index’ by dividing the totalnumber of A-NK cells in culture by an estimated number ofthe initial adherent cells. We determined this number byshowing that 5 � 1.5% of total adherent cells wereCD56+CD3– in the initial three patients studied. In contrast,69�6% of the initial adherent cells were CD56+CD3– insamples obtained from NC (Table I). There was no statisticallysignificant difference in the average proliferationindexes of A-NK cells from patients versus NC (Fig 2B). However, incontrast to NC, whose proliferation index ranged from 80 to500, that of CML patients varied widely, e.g. from 20 to30 000 for the ECP group. Three of the patients with ACP hada very low A-NK proliferation index, as opposed to none ofECP patients.

Since the percent of total plastic-adherent cells was knownfor all NC and CML patients studied, it was possible toexamine the possibility of an association between the initialadherence to plastic and disease phase or prior treatment.However, no such association could be demonstrated(P=0.38, two-sided).



Fig 1. Two-colour FACS analysis of day 14 A-NK cells generated fromthe blood of eight normal controls (NC), seven early chronic phase(ECP) and seven advanced chronic phase (ACP) patients with CML.The percentage of A-NK cells generated in these cultures declinedsignificantly with disease progression (P=0.00005, Jonckheere’stest).



Cell aggregates and morphological studiesThe microscopic observation of early patient cultures revealedthe presence of cell aggregates formed by activated lympho-cytes and large, presumably myeloid, cells (data not shown).Activated lymphocytes were characteristically elongated andbranched with the morphology similar to that of the cells inenriched A-NK cultures obtained from normal donors. A-NKcells appeared to surround large round cells, forming rosette-like clusters. These ‘rosettes’ could be observed as early as48 h after the initiation of the culture, and they graduallydisappeared as the culture proliferated. On Wright-Giemsa-stained slides, obtained from slide chambers 48 h after theinitiation of the culture (Figs 3A and 3B), it was possible toconfirm the presence of myeloid cells among the initialadherent cell population. The myeloid cells were surroundedby typical large granular lymphocytes.

Phenotypic studies of adherent cells captured on plasticAs expected from the morphological characteristics of theinitial plastic-adherent cells, the flow cytometry analysisconfirmed the presence of myeloid cells among them (Table I).In three patient samples a mean of 44�15% cells wereCD33+, whereas only 8�4% of adherent cells in NCexpressed this marker. The proportion of T cells among the

adherent cells of CML patients (63�22%) was alsosignificantly increased over that of NC (18�12%). In con-trast, only 5�1.5% of the cells expressed the CD56+CD3–

phenotype in the patient samples compared to 69� 6% ofadherent cells in NC. Except for CD34+ cells, present in asimilar number in both patients and NC, the differences in theproportion of myeloid cells, T lymphocytes, and A-NK cellsbetween the initial adherent cells obtained from NC and CMLpatients were all statistically significant (Table I). In each casethere was no overlap between the percent expressions of thethree patients’ samples with those of the three controls(P=0.05 one-sided, Wilcoxon test; use of Student t-test wouldresult in much lower P values, but the Gaussian assumption isat best questionable here).

Detection of BCR/abl gene rearrangementIn all 12 cultures tested by RT-PCR (six were from ECP and sixfrom ACP patients), the newly-adherent cell populations were

BCR/abl+. This finding provided an opportunity to follow thefate of malignant cells in A-NK cell cultures. On day 14, six ofthese cultures became BCR/abl– and all but one of theseshowed substantial enrichment in A-NK cells. The one non-enriched culture had a significant expansion of predomi-nantly CD56+CD3+ cells. In contrast, among the BCR/abl+

cultures, only two were enriched in A-NK cells; however, theyshowed a very low proliferation index. Both total proliferationand A-NK proliferation index were significantly higher inBCR/abl– than in BCR/abl+ cultures (Fig 4; P = 0.013 and P =0.007, one-sided Wilcoxon‘s test), even if one controls fordisease phase. One representative RT-PCR experiment isshown in Fig 5. Amplification of BCR/abl+ cells results inthe production of 195 or 120 bp amplimers, depending uponthe location of the breakpoint in the BCR gene, whereas cellslacking a Philadelphia chromosome produce no visibleamplimer. A 373 bp amplimer was observed following

Table I. Surface markers on adherent cells captured on plastic andstudied prior to culture of A-NK cells in patients with CML andnormal controls.*

No. of Positive cells (%)samples

CD56+CD3– CD3+ CD33+ CD34+

CML† 3 5�1.5 63�22 44�15 1.3�0.6NC 3 69�6 18�12 8�5 1.9�1.9

P value 0.05 0.05 0.05 NS

* Adherent cells were detached from the plastic surface and stainedfor flow cytometry, as described in Materials and Methods. The Pvalues were obtained using the Wilcoxon’s test. In each case therewas no overlap between the percent expressions of the three patients’samples with those of the three controls (P = 0.05 one-sided,Wilcoxon test; use of Student t-test would result in much lower Pvalues, but the Gaussian assumption is at best questionable here). NS= not significant.

† All of these patients were in an early phase of CML (ECP).

Fig 2. Proliferative capacity of A-NK cultures generated from eightnormal controls (NC), seven ECP and seven ACP patients with CML.The normal controls (NC) culture fold expansion (A) was significantlygreater than that in cultures obtained from CML patients, and thisdifference increased with disease progression (P=0.03, Jonckheere’stest). Utilizing the A-NK cell proliferation index, as described inMaterials and Methods, A-NK cells of patients, although varyingwidely, proliferated similarly to A-NK cells of NC (B), with theexception of three cultures in patients with ACP (advanced chronicphase) ECP-early chronic phase.



Fig 3. Wright-Giemsa stained slides, obtained from slide chambers 48 h after the initiation of the cultures, show myeloid cells (arrow) surroundedby typical large granular lymphocytes (A and B) (magnification �1000).



amplificationof PBGD. The RT-PCR assay was found to be ableto detect 1 BCR/abl+ cell/105 normal cells.

The finding of the BCR/abl translocation in A-NK cellcultures most likely demonstrates the presence of residualmyeloid leukaemic cells at the time these cultures wereterminated. This appeared to be the case, because CD33+ cellswere more frequently present at the end of those cultures inwhich BCR/abl+ cells persisted, as compared to the ones thatbecame BCR/abl– (P = 0.028). The mean percentage ofCD33+ cells in cultures which failed to successfully expand A-NK cells was 18 � 22 (SD). The mean percentage of CD34+

cells was 0.4 � 0.8 (SD) in these cultures. The phenotypiccharacteristics of cells in two representative 2-week culturesare shown in Table II. Both cultures were BCR/abl+ at the timeof initial adherence and became BCR/abl- by the time ofharvest 2 weeks later. Both were highly-enriched inCD56+CD3– cells and were depleted of CD14+, CD15+ orCD33+ myeloid cells (Table II).

Antitumour cytotoxicity in 14 d A-NK cell culturesIn view of the observation that the enrichment in and theproliferation of A-NK cells correlated with the efficiency ofelimination of BCR/abl+ cells in culture, it was of interest todetermine antitumour activity in these cultures. Cytotoxicity,measured against K562 and Daudi targets in 4 h 51Cr releaseassays, was significantly higher in BCR/abl– cultures, for bothtargets (P = 0.029), as compared to the BCR/abl+ cultures(Fig 4). The limited number of samples tested for cytotoxicityprecluded the ability to control for disease phase in thisanalysis. Although there was a tendency for a decreasedcytotoxicity in cultures obtained from patients with advancedCML (Figs 6A and 6B), this difference was not statisticallysignificant. The culture obtained from the single BP patientdisplayed a high level of activity against K562 and Daudi andstrongly influenced the statistical analysis.

NK cytotoxicityNK activity measured in peripheral blood of patients withCML was uniformly low (mean � SD: 20 LU � 18; range

Fig 4. Culture fold expansion (n=12), A-NK cell proliferation index(n=12) (see text), and cytotoxicity against K562 (n=7) and Daudi(n=7) as measured by lytic units (LU). All 12 cultures were obtainedfrom CML patients (six from ECP and six from ACP) and weresubmitted to RT-PCR for the presence of the BCR/abl rearrangement.The open circles (k) represent cultures that became BCR/abl– and thefilled circles (l) cultures that remained BCR/abl+ at day 14 of culture.Culture fold expansion, A-NK cell proliferation index, and cytotoxicityagainst both K562 and Daudi cell lines were significantly higher inBCR/abl– than in BCR/abl+ cultures (P = 0.002, P = 0.001 and P =0.029, respectively).

Fig 5. Detection of BCR/abl gene rearrangement. Amplification ofBCR/abl is presented in lanes 2–6; PBGD control amplifications are inlanes 7–11. Lanes 2 and 7: K562, Ph+ cell lines. Lanes 3 and 8:normal controls. Lanes 4 and 9: freshly adherent A-NK cells. Lanes 5and 10: cultured A-NK cells. Lanes 6 and 11: nucleic acid-freecontrols of the BCR/abl and PBGD amplifications, respectively. Lanes 1and 12: molecular size standards,�X174 RF DNA digested with HaeIII (Gibco-BRL).

0–48 LU) irrespective of the phase of their disease ascompared to NK activity of NC (range 50–300 LU).

In vitro effects of A-NK cells on haemopoietic progenitor cellsIn order to identify possible cytotoxic effects of A-NK cellson haemopoietic progenitor cells (HPC), cultured A-NKcells obtained from the peripheral blood of normal donorswere co-incubated with autologous donor marrow mono-nuclear cells for up to 1 week in medium containing IL2(22 nM). At the end of co-incubation, semi-solid colonyassays were performed to measure HPC frequency. Asshown in Fig 7, we observed preservation of HPC, withactual enhancement of colony formation at 10:1 ratios ofeffector (A-NK):target (HPC) cells in comparison to culturesnot containing A-NK cells. These results clearly demon-strate the lack of adverse effects of A-NK cells on in vitrogrowth of autologous HPC and suggest that A-NK cells mayactually be beneficial. The same effect was observedfor various types of colonies (BFU-E, CFU-GM, CFU-Mix).In addition, assays of NK activity in 51Cr-release assays andof phenotype by flow cytometry at the end of 14 d co-culture period consistently demonstrated the presence ofCD56+CD3– large granular lymphocytes, which mediatedNK activity (data not shown).


382 Lucia M. R. Silla et alTable II. Phenotypic characteristics of cells in two patient culturesstudied at the time of cell adherence and of cell harvest.*

Positive cells (%)

Patient 8 (APC) Patient 12 (EPC)

Markers ADHy Culturedz ADHy Culturedz

CD56�CD3ÿ 7 84 5 79CD56ÿCD3� 49 5 42 4CD56�CD3� 4 7 5 4CD3� 53 12 47 8CD14� 44 0 24 0CD15� 47 1 21 0CD16� 18 88 23 82CD33� 60 2 40 0CD34� 1 0 1 0

* Following adherence in the presence of IL2, cells were recoveredfrom the plastic surface, washed, stained with labelled mAbs andanalysed by two-colour flow cytometry. The same phenotypic analysiswas repeated 2 weeks later, at the time of termination of each culture.y BCR/abl� culture at the time of initial adherence to plastic.z BCR/ablÿ culture at time of harvest.

Fig 6. Cytotoxicity of day 14 A-NK cells obtained from six normalcontrols (NC), five ECP and three ACP patients with CML testedagainst NK-sensitive K562 (A) and NK-resistant Daudi (B). There wasa tendency (not statistically significant) for a decreased cytotoxicity incultures obtained from patients with advanced CML.

Fig 7. Effects of A-NK cells on autologous haemopoietic colonyformation. Initially, A-NK cells and autologous bone marrow(BM) cells obtained from normal donors were co-cultured at theE (A-NK):T (BM) ratios shown for 14 d under conditions describedin Materials and Methods. Next, cell mixtures or control cultureswere plated in methylcellulose medium and cultured. Colonies werescored on day 14 for CFU-GM, CFU-GEMM and BFU-E. Results areshown as total colony counts. The data are means �SD from fiveexperiments.



DISCUSSION

NK cells are known to play an important role in the control ofsolid tumour metastasis (Whiteside & Herberman, 1995). Inleukaemia, the presence and activity of NK cells maycontribute to the control of disease, based on the well-documented low or absent NK activity during leukaemiarelapse and its recovery in remission (Adler et al, 1988). InCML, NK activity is significantly reduced in peripheral bloodas compared to that of NC. The decrease in NK activity hasbeen consistently observed in early as well as advanced phasesof CML by us and several other investigators (Fujimiya et al,1986; Lotzova, 1988). An argument could therefore be madethat increased numbers and/or activity of NK cells might betherapeutically beneficial in CML. Adoptive transfer ofautologous NK cells activated in vitro has been proposed asone therapeutic approach in neoplastic diseases unresponsiveto conventional therapies (Lister et al, 1995; Rosenberg et al,1985). However, in CML and other leukaemias, generation ofautologous antileukaemic effector cells from the patients’peripheral blood could be a problem for at least two reasons:(a) the possibility that NK-cell precursors are low in numberor functionally deficient; and (b) the likelihood that malignantcells might be present or induced to proliferate in NK-cellcultures. Both these possibilitieswere considered and tested inthis study.

A subset of NK cells called A-NK cells, because of theproperty of these cells to adhere to solid surfaces withinminutes after IL2 activation, contains potent antitumoureffector cells, as demonstrated recently in in vitro and in vivoexperiments in our laboratory (Vujanovic et al, 1993, 1995).A-NK cells cultured for 14 d in the presence of IL2 consist ofCD3-CD56+CD16dim or – cells which mediate strong anti-leukaemic activity in vitro (Vujanovic et al, 1993). Recently,using a modified adherence protocol, we have been able togenerate almost pure (>90% CD3–CD56+ cells) cultures of A-NK cells from peripheral blood of NC or patients with solidtumours (Elder et al, 1993). It was therefore of interest todetermine if similarly pure A-NK cell cultures could begenerated from PBMC of patients with early or advanced CML.A short period of adherence (3–5 h) in the presence of 6000IU/ml of IL2, which resulted in capture of A-NK cells butrelatively few T cells or CD33+ cells from PBMC of NC, wassubstantially less effective with the patients’ PBMC (Table I).Clearly, a considerable number of T cells and myeloid cellsadhered to plastic in spite of the preceding T-cell- andmonocyte-depleting steps and the short period of adherenceused. Nevertheless, the captured A-NK cells proliferated well,and 14 d cultures of these cells were enriched in aCD3–CD56+ subset in all patients with early CML and aproportion of those with advanced CML. This is in agreementwith results of Verfaillie et al (1989, 1990) who were also ableto generate cultures of A-LAK cells, in which CD56+CD3–

cells were >80%, from PBMC of CML patients, including thosewith advanced disease. A considerable amount of publisheddata indicate that A-NK cells selected by the property of rapidIL2-induced adherence to solid surfaces represent a uniqueand homogenous subset of NK cells, endowed with theproperty of eliminating tumour cell targets (Vujanovic et al,

1993, 1995; Yasumura et al, 1994; Whiteside & Herberman,1995). A-NK cells have CD56dimCD3– phenotype, and theydiffer from CD56brightCD3– subset of NK cells described byothers (Caligiuri et al, 1990) by a variety of functionalproperties (Vitolo et al, 1993; Vujanovic et al, 1993, 1995). Itis of great clinical significance that we were able to captureand expand this subset of effector cells in patients with earlyCML, irrespective of ‘contamination’ of the initial adherentpopulation with either T or myeloid cells. Our data indicatethat precursors of A-NK cells are present in peripheral bloodof patients with CML, and that in spite of the prominentpresence of T and myeloid cells among cells captured byadherence to plastic, it is feasible to expand A-NK cells usingthese patients’ PBMC as a starting population.

The presence of BCR/abl+ cells was observed in all 12samples of freshly-captured A-NK cells, presumably becauseof ‘contamination’ with myeloid cells adherent to plastic.Using a sensitive RT-PCR technique, which is able to detect 1BCR/abl+ cell/105 normal cells, we found BCR/abl+ cells in 6/12 A-NK cell cultures on day 14. Irrespective of the phase ofCML or the extent of the initial ‘contamination’ with myeloidor T cells, the A-NK cell cultures became devoid of malignantcells during 14 d of expansion, as long as A-NK cellsproliferated well in vitro. On the other hand, poorlyproliferating cultures of A-NK cells remained ‘contaminated’with cells expressing the BCR/abl markers. These culturesalso contained myeloid cells detectable by flow cytometry,indicating that leukaemic cells could survive in the absence ofa sufficient number of antileukaemic effector (A-NK) cells.Furthermore, antileukaemic cytotoxicity of A-NK cells in theBCR/abl+ cultures was significantly lower than that of BCR/abl– A-NK cell cultures. These observations suggest that A-NK cells generated from patients with CML can developantileukaemic activity in vitro, and that the ability of thesecells to purge the culture of malignant cell targets depends, toa large extent, on the proliferation and function of the subsetof activated antileukaemic effector cells. Also, the presence ofaggregates or rosettes containing a myeloid cell surroundedby several A-NK cells in early A-NK cell cultures indicatesthat interactions between autologous leukaemia cells andimmune effectors occur, probably leading to the elimination ofleukaemic cell targets. The rosettes were observed in allcultures, but were especially prominent in the six CMLcultures in which A-NK cells proliferated best. Rosetteformation between tumour cell targets and activated Tlymphocytes has been described in IL2-containing culturesof tumour-infiltrating lymphocytes (TIL) obtained from solidtumours (Rosenberg et al, 1986; Muul et al, 1987; Yasumuraet al, 1993) but not in cultures of A-NK cells generated fromPBMC of patients with leukaemia. By analogy with TIL, it isreasonable to expect that cultures containing rosettes becomedepleted of BCR/abl+ cells.

The correlation we observed between the disappearance ofBCR/abl+ cells and successful growth and function of A-NKcells in culture has important therapeutic implications. It isknown that Ph1 chromosome-positive cells disappear whenmarrow from patients with CML is placed in long-term bonemarrow culture (LTBMC) 3–4 (Coulombel et al, 1983).Recently, it has been shown that the addition of IL2 to these

cultures leads to the more rapid disappearance of Ph1-positivemetaphases. The same in vitro conditions without IL2 did notlead to a significant decrease of Ph1-positive metaphasesduring the same time period (Verma et al, 1994). It is possiblethat A-NK cells generated in the presence of IL2 in thesecultures are involved in in vitro purging of leukaemic cellsfrom LTBMC. These observations, together with our data thathuman A-NK cells do not harm normal haemopoieticprogenitors in vitro, suggest a possible role for A-NK cells inin vitro purging of leukaemic cells from the bone marrow ofpatients with leukaemia. In addition, preliminary data fromclinical trials, in which autologous A-NK cells are adoptivelytransferred 2 d after the peripheral stem cell transfer inpatients with advanced lymphoma (Lister et al, 1993, 1995),indicate that both engraftment and antitumour effects arefacilitated by this type of therapy. Since minimal residualdisease is responsible for relapse of CML after ABMT, it mightbe beneficial to consider AIT with A-NK cells in this setting aswell. While our report provides support for the important roleof A-NK cells in antileukaemic effects, it also emphasizes theneed for careful selection of patients with CML who are likelyto benefit from this unconventional approach. Clearly,patients in the ECP of CML who generate A-NK cells as wellas NC, and whose A-NK cells are capable of eliminating BCR/abl+ leukaemic targets, are the best candidates for AIT withautologous A-NK cells.

ACKNOWLEDGMENTS

This work was supported in part by the Pathology Educationand Research Foundation and Blood Science Foundation ofPittsburgh, Pittsburgh, Pa., American Cancer Society(IM696-TLW) and CNPq-Conselho Nacional de Desenvolvi-mento Cientifico e Tecnologico, Brazil.

REFERENCES

Adler, A., Chervenick, P.A., Whiteside, T.L., Lotzova, E. & Herberman,R.B. (1988) Interleukin 2 induction of lymphokine activated killer(LAK) activity in the peripheral blood and bone marrow of acuteleukemia patients. I. Feasibility of LAK generation in adult patientswith active disease and in remission. Blood, 71, 709–716.

Antin, J.H. (1993) Graft-versus-leukemia: no longer an epi-phenomenon. Blood, 82, 2273–22777.

Buttirini, A., Bortin, M.M., Seeger, R.P. & Gale, R.P. (1987) Graftversus leukemia following bone marrow transplantation: a modelof immunotherapy in man. Cellular Immunotherapy of Cancer (ed. byR. L. Truitt, R. P. Gale and N. M. Bortin), p. 371. Liss, New York.

Buttirini, A., Keating, A., Goldman, J.M. & Gale, R.P. (1990)Autotransplants in chronic myelogenous leukaemia: strategiesand results. Lancet, i, 1255–1258.

Caligiuri, M.A., Zmuidzinas, A., Manley, T.J., Levine, H., Smith, K.A. &Ritz, J. (1990) Functional consequences of interleukin 2 receptorexpression on resting human lymphocytes: identification of a novelnatural killer cell subset with high affinity receptors. Journal ofExperimental Medicine, 171, 1509–1526.

Coulombel, L., Kalousek, D.K., Eaves, C.J., Gupta, C.M. & Eaves, A.C.(1983) Long-term marrow culture reveals chromosomally normalhematopoietic progenitor cells in patients with Philadelphia

chromosome-positive chronic myelogenous leukemia. New EnglandJournal of Medicine, 308, 1493–1498.

Deisseroth, A.B., Zu, Z., Claxton, D., Hanania, E.G., Fu, S., Ellerson, D.,Goldberg, L., Thomas, M., Janicek, K., Anderson, F., Hester, J.,Korbling, M., Durett, A., Moen, R., Berenson, R., Heimfeld, S.,Hamer, J., Calvert, L., Tibbits, P., Talpaz, M., Kantarjian, H.,Champlin, R. & Reading, C. (1994) Genetic marking shows thatPh+ cells present in autologous transplants of chronic myelogenousleukemia (CML) contribute to relapse after autologous bonemarrow in CML. Blood, 83, 3068–3076.

Elder, E.M., Kirkwood, J.M. & Whiteside, T.L. (1993) Characterizationof human NK cells obtained from cancer patients and cultured foradoptive immunotherapy. (Abstract). Journal of Immunology, 150,59A.

Fujimiya, Y., Bakke, A., Chang, W.C., Linker-Israeli, M., Udis, B.,Horwitz, D. & Pattengale, P.K. (1986) Natural killer cell immuno-deficiency in patients with chronic myelogenous leukemia. I.Analysis of the defect using the monoclonal antibodies HNK-1 (Leu7) and B73.1. International Journal of Cancer, 37, 639–649.

Grandchamp, B., Verneuil, H.D., Beaumont, C., Chretien, S., Walter,O. & Nardmann, Y. (1987) Tissue-specific expression of porphobi-linogen deaminase. European Journal of Biochemistry, 162, 105–110.

Heisterkamp, N., Stam, K., Groffen, J., de Klein, A. & Grosveld, G.(1985) Structural organization of the bcr gene and its role in thePh1 translocation. Nature, 315, 758–761.

Jian, Y.-Z., Cullis, J.O., Kanfer, E.J., Goldman, J.M. & Barret, A.J. (1993)T cell and NK cell mediated graft-versus-leukemia reactivityfollowing donor buffy coat transfusion to treat relapse aftermarrow transplantation for chronic myeloid leukemia. BoneMarrow Transplantation, 11, 133–138.

Kantarjian, H.M., Deisseroth, A., Kurzrock, R., Estrov, Z. & Talpaz, M.(1993) Chronic myelogenous leukemia: a concise update. Blood,82, 691–703.

Laughlin, M., Loftus, J., Herzig, G.P., Caligiuri, T. & Caligiuri, M.(1993) Natural killer (NK) cell activity following dose intensivechemotherapy (CT). (Abstract). Blood, 82, 242a.

Lister, J., Pincus, S.M., Elder, E.M., Whiteside, T.L., Rybka, W.B. &Donnenberg, A.D. (1993) Adoptive immunotherapy duringperipheral blood stem cell (PBSC) transplantation: amplificationof natural killer cell function early after transplantation. Journal ofImmunology, 150, 218A.

Lister, J., Rybka, W.B., Donnenberg, A.D., de Magalhaes-Silverman,M., Pomcis. S.M., Bloom, E.J., Elder, E.M., Ball, E.D. & Whiteside, T.L.(1995) Autologous peripheral blood stem cell transplantation andadoptive immunotherapy with A-NK cells in the immediate post-transplant period. Clinical Cancer Research, 1, 607–614.

Lotzova, E. (1988) Role of natural killer cells in defense againstleukemia: therapeutic considerations. Natural Immunity and CellGrowth Regulation, 7, 170–179.

Murphy, W.J., Reynolds, C.W., Tiberghien, P. & Longo, D.L. (1993)Natural killer cells and bone marrow transplantation. Journal of theNational Cancer Institute, 85, 1475–1482.

Muul, L.M., Spiess, P.J., Director, E.P. & Rosenberg, S.A. (1987)Identification of specific cytolytic immune responses againstautologous tumor in humans bearing malignant melanoma.Journal of Immunology, 138, 989–995.

Nowell, P.C. & Hungerford, D.A. (1960) A minute chromosome inhuman chronic granulocytic leukemia. Science, 132, 1497

Pincus, S., Silla, L.M.R., Beltz, L., Donnenberg, A.D., Ball, E.D. &Whiteside, T.L. (1994) Autologous activated natural killer (A-NK)cells enhance normal human hematopoiesis. Clinical Research, 41,365A.

Pross, H.F., Baines, M.G., Rubin, P., Shragge, P. & Patterson, M.S.





(1981) Spontaneous human lymphocyte mediated cytotoxicityagainst tumor target cells. IX. The quantitation of natural killer cellactivity. Journal of Cancer Immunology, 1, 51–63.

Rabinowich, H., Sedlmayr, P., Herberman, R.B. & Whiteside, T.L.(1991) Increased proliferation, lytic activity and purity of naturalkiller cells co-cultured with mitogen-activated feeder cells. CellularImmunology, 135, 454–470.

Reittie, J.E., Gottlieb, D., Heslop, H.E., Leger, O., Drexler, H.G.,Hazlehurst, G., Hoobrand, A.V., Prentice, H.G. & Brenner, M.K.(1989) Endogenously generated activated killer cells circulate afterautologous and allogeneic marrow transplantation but not afterchemotherapy. Blood, 73, 1351–1358.

Ringden, O. & Horowitz, M.M. (1989) Graft-versus-host leukemiareactions in humans. For the Advisory Committee of theInternational Bone Marrow Transplant Registry. TransplantationProceedings 21, 2989–2992.

Rosenberg, S.A., Lotze, M.T., Muul, L.M., Leitman, S., Chang, A.E.,Ettinghausen, S.E., Matory, Y.L., Skibber, J.M., Shiloni, E., Vetto, J.T.,Seipp, C.A., Simpson, C. & Reichert, C.M. (1985) Observations on thesystemic administration of autologous lymphokine-activated killercells and recombinant interleukin 2 to patients with metastaticcancer. New England Journal of Medicine, 313, 1485–1492.

Rosenberg, S.A., Spiess, P. & Lafreniere, R. (1986) A new approach toadoptive immunotherapy of cancer with tumor-infiltratinglymphocytes. Science, 233, 1318–1321.

Rowley, J.D. (1973) A new consistent chromosome abnormality inchronic myelogenous leukemia identified by quinacrine fluores-cence and Giemsa staining. Nature, 243, 145–147.

Sedlmayr, P., Rabinowich, H., Winkelstein, A., Herberman, R.B. &Whiteside, T.L. (1992) Generation of adherent lymphokineactivated killer (A-NK) cells from patients with acute myelogenousleukaemia. British Journal of Cancer, 65, 222–228.

Shtivelman, E., Lifshitz, B., Gale, R.P. & Canaani, E. (1985) Fusedtranscript of ABL and BCR genes in chronic myelogenousleukaemia. Nature, 315, 550–554.

Shtivelman, E., Lifshitz, B., Gale, R.P., Roe, B.A. & Canaani, E. (1986)Alternative splicing of RNAs transcribed from the human abl geneand from the bcr-abl fused gene. Cell, 47, 277–284.

Verfaillie, C., Kay, N., Miller, W. & McGlave, P. (1990) Diminished

A-LAK cytotoxicity and proliferation accompany disease progres-sion in chronic myelogenous leukemia. Blood, 75, 401–408.

Verfaillie, C., Miller, W., Kay, N. & McGlave, P. (1989) Adherentlymphokine-activated killer cells in chronic myologenous leukemia:a benign cell population with potent cytotoxic activity. Blood, 74,793–797.

Verma, U.N., Bagg, A., Brown, E. & Mazumder, A. (1994) Interleukin-2 activation of human bone marrow in long-term cultures: aneffective strategy for purging and generation of antitumor cytotoxiceffectors. Bone Marrow Transplantation, 13, 115–123.

Vitolo, D., Vujanovic, N., Rabinowich, H., Schlesinger, M., Herberman,R.B. & Whiteside, T.L. (1993) Rapid interleukin-2 inducedadherence of human natural killer (NK) cells. II. Expression ofmRNA for cytokines and IL2 receptors in adherent NK cells. Journalof Immunology, 151, 1926–1937.

Vujanovic, N.L., Rabinowich, H., Lee, Y.J., Jost, L., Herberman, R.B. &Whiteside, T.L. (1993) Distinct phenotypic and functional char-acteristics of human natural killer cells obtained by rapidinterleukin 2-induced adherence to plastic. Cellular Immunology,151, 133–157.

Vujanovic, N.L., Yasumura, S., Hirabayashi, H., Lin, W.C., Watkins, S.,Herberman, R.B. & Whiteside, T.L. (1995) Antitumor activities ofhuman IL2-activated natural killer (A-NK) cells in solid tumortissues. Journal of Immunology, 154, 281–289.

Whiteside, T.L., Bryant, J., Day, R. & Herberman, R.B. (1990) Naturalkiller cytotoxicity in the diagnosis of immune dysfunction: criteriafor a reproducible assay. Journal of Clinical Laboratory Analysis, 2,102–114.

Whiteside, T.L. & Herberman, R.B. (1995) Role of natural killer cellsin immune surveillance of cancer. Current Opinion in Immunology, 7,704–710.

Yasumura, S., Hirabayashi, H., Schwartz, D.R., Toso, J.F., Johnson, J.T.,Herberman, R.B. & Whiteside, T.L. (1993) Human cytotoxic T-celllines with restricted specificity for squamous cell carcinoma of thehead and neck. Cancer Research, 53, 1461–1468.

Yasumura, S., Lin, W., Hirabayashi, H., Vujanovic, N.L., Herberman,R.B. & Whiteside, T.L. (1994) Immunotherapy of liver metastases ofhuman gastric carcinoma with IL2-activated natural killer cells.Cancer Research, 54, 3808–3816.

Generation of activated natural killer (A-NK) cells in patients with chronic myelogenous leukaemia...

Documents

Transcript of Generation of activated natural killer (A-NK) cells in patients with chronic myelogenous leukaemia...