Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Vol. 4, 2859-2868, November 1998 Clinical Cancer Research 2859
Antileukemia Activity of a Natural Killer Cell Line against
Human Leukemias1
Ying Yan,2 Peter Steinherz,
Hans-Georg Klingemann, Dieter Dennig,
Barrett H. Childs, Joseph McGuirk, and
Richard J. O’Reilly
Bone Marrow Transplantation Service [Y. Y., B. C., J. M., R. J. 0.]and Department of Pediatrics [P. S., D. D., R. J. 0.], Memorial Sloan-
Kettering Cancer Center, New York, New York 1002 1 , and The
Terry Fox Laboratory and University of British Columbia,Vancouver, British Columbia V5Z 1L3. Canada IH-G. K.]
ABSTRACTWe describe here the in vitro and in vivo antileukemia
activity of a recently described natural killer (NK) cell line
(NK-92), which has features of human activated NK cells.
The cytotoxic activity of rhIL2-dependent cultured NK-92cells against primary patient-derived leukemic target cells
[12 acute myelogenous leukemias (AMLs), 7 T acute lym-
phoblastic leukemias (T-ALLs), 14 B-lineage-ALLs, and 13
chronic myelogenous leukemias (CMLs)], human leukemic
cell lines (K562, KG!, HL6O, Raji, NALM6, TALL-!04,
CEM-S, and CEM-T) and normal bone marrow cells was
measured in 5tCr-release assay (CRA). The patient-derived
leukemias could be subdivided into three groups based on
their sensitivity to NK-92 cells: insensitive (�!9% lysis),
sensitive (20-49% lysis), and highly sensitive (�50% lysis)
at an E:T ratio of 9:!. Of 46 patient-derived samples, 24
(52.2%) were sensitive or highly sensitive to NK-92-medi-
ated in vitro cytotoxicity (6 of 12 AMLs, 7 of 7 T-ALLs, 5 of
14 B-lineage-ALLs, and 6 of 13 CMLs). NK-92 cells werehighly cytotoxic against all of the eight leukemic cell lines
tested in a standard 4-h CRA. Normal human bone marrow
hematopoietic cells derived from 18 normal donors were
insensitive to NK-92-mediated cytolysis. In comparison with
human lymphokine-activated killer cells, normal NK cells,
and T cells, NK-92 cells displayed more powerful antileuke-
mia activity against a patient-derived T-ALL as well as
K562 and HL6O cells, both in in vitro CRA and in a xe-nografted human leukemia SCID mouse model. The NK-92
cells did not induce the development of leukemia in SCID
Received 5/12/98; revised 8/17/98; accepted 8/26/98.
The costs of publication of this article were defrayed in part by the
payment of page charges. This article must therefore be hereby marked
advertisement in accordance with 18 U.S.C. Section 1734 solely to
indicate this fact.
I Supported by NIH Grant CA23766 and grants from the Lisa Bilotti
Foundation, the Andrew Gaffney Foundation, the Zelda Radow Wein-
traub Cancer Foundation, and the Guy M. Stewart Foundation.
2 To whom requests for reprints should be addressed, at Cancer Institute!Hospital, Beijing Union Medical College, Chinese Academy of Medical
Sciences, Beijing 100021, China. Phone: 86-10-67718927; Fax: 86-10-
67715058.
mice after i.v., i.p., or s.c. inoculation. In adoptive transfer
experiments, SCID mice receiving i.p. inoculations of human
leukemias derived from a T-ALL (TA27) and an AML(MA26) that were highly sensitive to the cytolysis of NK-92
cells in vitro, as well as a pre-B-ALL (BA3I) that was
insensitive to the in vitro cytohysis of NK-92 cells, were
treated by administration of NK-92 cells with or without
rhLL2 (2 x iO� NK-92 cells i.p.; one dose or five doses).Survival times of SCID mice bearing the sensitive TA27 and
MA26 leukemias were significantly prolonged by adoptive
cell therapy with NK-92 cells. Some of the animals who
received five doses of NK-92 cells with or without rhIL2
administration were still alive without any signs of leukemia
development 6 months after leukemia inoculation. In con-
trast, survival of mice bearing the insensitive BA31 leukemia
were not affected by this treatment. This in vitro and in vivo
antileukemia effect of NK-92 cells suggests that cytotoxic
NK cells of this type may have potential as effectors of
leukemia control.
INTRODUCTION
Cytotoxicity mediated by NK3 cells, has been hypothesized
to play an important role in a host’s defenses against several
forms of cancer ( 1 , 2). Human NK cells are an immunopheno-
typically distinguishable subset of lymphocytes that exhibit a
striking and, as yet, poorly understood capacity to distinguish
and destroy tumor cells or virally infected cells. This cytotoxic
activity does not require prior sensitization and is not restricted
by MHC antigens (1-4). Much research has been focused on
ways to manipulate the functions of NK cells and other immune
cells for therapeutic purposes (5-8). However, the inability to
generate sufficient numbers of such cells in vitro, which main-
tam their in vivo tumoricidal and tumor-targeting capabilities,
remains a major obstacle to their use in clinical applications of
adoptive cell immunotherapy (6, 9, 10). Studies of the mecha-
nisms whereby NK cells exert their tumoricidal effects are also
limited by both difficulties in enriching the NK cell fractions
without compromising their biological functions and obtaining
pure NK cells without contaminating T cells or other immune-
response effector cells. To avoid these problems, many investi-
gators have used established NK-like cell lines to explore the
biological mechanisms of natural cytotoxicity to target cells
(11-14).
NK cells as well as CTL clones established from patients
3 The abbreviations used are: NK, nature killer; AML, acute myeloge-nous leukemia; ALL. acute lymphoblastic leukemia; CML. chronic
myelogenous leukemia: CRA, 51Cr-release assay; MHC, major histo-
compatibility; CTL. cytotoxic T cell; BM, bone marrow; FACS, flu-
oroescence-activated cell sorting; LAK, hymphokine-activated killer;
MST. median survival time; MTX, methotrexate.
Association for Cancer Research. by guest on September 4, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
2860 Antileukemia Activity of a NK Cell Line
with NK- or T-cell leukemias may not only retain the pheno-
typic features (morphology and immunological phenotypes) that
are displayed by their normal counterparts, but they also exhibit
similar functional characteristics with regard to their cytolytic
activity toward target cancer cells (12-14). Studies have shown
that these clones are able to consistently maintain their in vitro
cytotoxicity to a wide variety of tumor cells, without toxic
effects on cells from normal tissue (12-16). Recently, preclin-
ical studies have also demonstrated promising antitumor activity
ill 5110 with a lethally irradiated, MHC-unrestricted, cytotoxic
T-cehh leukemic clone (TALL-104; Refs. 15 and 16). Adoptive
transfer of these cells was able to eliminate human leukemic cell
lines in xenografted SCID mice and to induce remissions of
spontaneous lymphomas in dogs without inducing the develop-
ment of T-cell leukemia in the animal models (15-18). These
observations suggest that the study of the selective cytotoxic
characteristics of the effectors may not only be useful for the
elucidation of the mechanisms of MHC-unrestricted immune
surveillance in human cancer but, it may also provide a potential
alternative approach for clinical adoptive immunotherapy.
The NK-92 cell line is a recently established human NK
cell clone, originally derived from a human non-Hodgkin’s
lymphoma with the morphology of large granular lymphocytes
and a CD56�CD3CDl6 immunophenotype (19). As a rhIL2-
dependent cell line, NK-92 cells retain the characteristics of
activated human NK cells and are highly cytotoxic against
human leukemic cell lines in vitro (19, 20). To evaluate the
tumoricidal activity of NK-92 cells against human leukemias,
we have examined the antileukemia effects of this NK cell clone
against different types of primary patient-derived leukemias and
established human leukemic cell lines both in vitro by cytotox-
icity assays and in vivo by evaluating the effects of adoptive
transfer of such cells on the growth and dissemination of human
heukemias in xenografted SCID mice.
MATERIALS AND METHODSPatient-derived Leukemic Samples. Samples were ob-
tamed, with informed consent, during routine diagnostic blood
studies or BM aspirates from patients with newly diagnosed or
relapsed leukemias. Cells from patients with AML (n 12),
T-ALL (a = 7), B-lineage ALL (n = 14: 13 pre-B-ALLs and 1
B-ALL) and CML (n 13: 6 in chronic phase, 1 in accelerated
phase, and 6 in blast crisis) were studied (Table 1). Blast-
enriched mononuclear cells were isolated by Ficoll Hypaque
density gradient separation and washed in RPMI 1640.
Leukemic Target Cell Lines. The following cell lines
were obtained from the American Type Culture Collection
(Manassas, VA): K562 (CML in blast crisis), HL6O (acute
promyelocytic leukemia), KG! (erythroleukemia), NALM6
(acute pre-B-ALL), and RAJI (Burkitt’s lymphoma). CEM-S
(acute T lymphoblastic leukemic cell line sensitive to MTX) as
well as CEM-T (MTX-resistant subline of CEM-S) were pro-
vided by Dr. T. Trippett (Memorial Sloan-Kettering Cancer
Center). Their characteristics have been described previously
(2 1 ). All cell lines were cultured at 37#{176}Cin 5% CO, in RPMI
1640 supplemented with 10% heat-inactivated FCS (Sigma
Chemical Co.), L-glutamine, and antibiotics.
Normal BM Hematopoietic Cells. Heparinized BM col-
lected from normal donors was separated by Ficoll Hypaque
density gradient isolation to produce the mononuclear cells.
Enrichment of hematopoietic cells and depletion of T cells was
achieved by soybean lectin agglutination (Vector Laboratories,
Inc., Burlingame, CA) of mature marrow elements and removal
of residual T cells by rosetting with sheep RBCs, as described
previously (22).
Effector Cells. NK-92 cells were cultured and main-
tamed in a-MEM medium supplemented with 12.5% FCS,
12.5% horse serum (Gemini Bioproducts, Calabasas, CA), and
rhIL2 (500 IU/ml; Chiron, Emeryville, CA). Another human NK
cell clone, YT cells, were maintained in RPMI 1640 with 10%
FCS and rhIL-2 ( 100 lU/mi; Ref. 12). TALL- 104 cells (a
MHC-unrestricted human CTL clone, generously provided by
Drs. D. Santoli and A. Cesano, The Wistar Institute, Philadel-
phia, PA) were maintained in Iscove’s modified Dulbecco’s
medium supplemented with 10% FCS and rhIL-2 (100 lU/mi;
Ref. 16).
For the generation of LAK cells, mononuclear cells were
isolated from heparmnized peripheral blood from healthy donors.
After depletion of monocytes by a 2-h plastic adherence, non-
adherent cells (5-10 X l0� ce!ls/ml) were resuspended and
cultured in medium consisting of RPMI 1640 with 10% human
heat-inactivated serum (Gemini Bioproducts), L-glutamine, an-
tibiotics (penicillin and streptomycin), and 500 IU/ml rhIL2 at
37#{176}Cand 5% CO2 for 5-7 days. The cell concentration was
adjusted to 10 x 105/ml at each feeding every 3 days.
To isolate the NK-cell populations, a Ceprate cell separa-
tion system based on avidin-biotin immunoaffinity (CelhPro,
Inc., Bothell, WA) was used to purify a CD56� cell fraction
from cultured LAK cells. Briefly, the harvested cells were
washed and resuspended in PBS with 1% BSA (CellPro, Inc.).
To each 1-2 x l0� cells/ml, 40 p.1 of primary monoclonal
antibody (mouse antihuman CD56; CellPro, Inc.) was added,
and the cells were incubated at 4#{176}Cfor 25 mm. After incubation,
the cells were washed and resuspended to a concentration of 1 X
108 cells/ml in PBS with 1% BSA. Then, to each l-ml cell
suspension, 20 �.tl of Biotin-habeled rat antimouse IgGI antibody
(CellPro, Inc.) was added, and the cells were incubated again at
4#{176}Cfor 25 mm. After incubation, the cells were washed and
resuspended at a concentration of 1 X 108 cells/mI in PBS with
5% BSA and slowly passed through the avidin column. The
CD56� cells were captured, and other cells, including the T cell
fraction, were eliminated from the column. After washing the
column, the adherent cells were then disassociated from the
column by agitation and elimination. After separation, the NK-
cell-enriched populations contained >85% CD56�CD3 cells.
The majority of the cells in the eliminated fraction were >95%
CD3�CD56 T cells.
To generate leukemia-reactive allocytotoxic T lympho-
cytes, peripheral blood mononuclear cells isolated from a nor-
mal donor were cultured with irradiated leukemic-stimulating
target cells and irradiated autologous peripheral blood mononu-
clear cells as feeder cells. Cultures were started in 60-well plates
at 1000 responder cells/well in RPMI 1640 containing 15%
human serum and rhIL2 50 IU/ml at 37#{176}C,5% CO.,. The ratio
of stimulator cells and feeder cells to responder cells were 5:1
and 10: 1, respectively. After 10-12 days of culture, the leuke-
Association for Cancer Research. by guest on September 4, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
Table 1 Cytotoxicity of NK-92 against patient-derived leukemia cells
Patients Disease status Blast (%) in sample
NK-92
Specific hysis (%) at E:T Ratio
9: 1 3: 1 1 : 1
AML1. (M4)D” Relapse PB (66%) 66 ± 2* 58 ± 3 5 1 ± 6
2. (Ml) Relapse PB (50%) 54 ± 3 46 ± 6 39 ± 2
3. (M3) Relapse PB (50%) 47 ± 9 35 ± 6 37 ± 1 1
4. (M2) New BM (28%) 35 ± 4 27 ± 4 20 ± 2
5. (M4) Refractory PB (90%) 39 ± 10 34 ± 7 29 ± 10
6. (M4) New BM (97%) 4 ± 4 3 ± 2 2 ± 2
7. (M4) New PB (39%) 9 ± 5 4 ± 1 5 ± 2
8. (M3) New PB (55%) 3 ± 0.2 1 ± 0.4 3 ± 1
9. (M3) New BM (32%) 6 ± 0.4 2 ± 0.2 0.8 ± 1
10. (M7) Rel BM (30%) 19 ± 0.4 9 ± I 3 ± 2
11. (Ml) New BM (90%) 44 ± 4 28 ± 3 22 ± 912. (M2) Rd PB (55%) 12 ± I 1 1 ± 0.4 0.3 ± 1
T-ALL1. Relapse BM(98%) 66±7 51 ±3 48±11
2. Relapse PB (85%) 67 ± 3 57 ± 5 36 ± 6
3. Relapse PB (77%) 74 ± 15 65 ± 13 54 ± 16
4. Relapse PB (60%) 56 ± 2 43 ± 4 42 ± 3
5. Relapse BM (75%) 54 ± 3 34 ± 4 26 ± 2
6. New BM(40%) 33 ± 11 21 ± 5 13 ± 10
7. New BM (66%) 30 ± 7 26 ± 2 12 ± 2
B-lineage-ALL
1. Relapse BM (78%) 50 ± 2 42 5 3 21 ± 4
2. New BM (30%) 46 ± 2 39 ± 3 3 1 ± 0.4
3. Relapse BM (75%) 35 ± 4 24 ± 2 16 ± 2
4. New BM (97%) 43 ± 13 31 ± 12 21 ± 13
5. Relapse BM (60%) 8 ± 3 7 ± 3 7 ± 2
6. Relapse BM (80%) 0 0.8 ± 1 2 ± 0.3
7. Relapse PB (80%) 5 ± 5 9 ± 8 3 ± 1
8. New BM (68%) 5 ± 1 5 ± 1 5 ± 0.4
9. New BM (33%) 0 0 010. Relapse BM (87%) 0 0 011. Relapse BM(75%) 2 ± 2 2 ± 1.4 2 ± 0.4
12. New BM (30%) 5 ± 0.4 3 ± 0 4 ± 0.4
13. New PB (90%) 29 ± 14 31 ± 11 24 ± 8
14. New BM (81%) 0 0 0CML
1. BC PB (45%) 65 ± 8 59 ± 4 56 ± 4
2. AC PB (22%) 60 ± 9 48 ± 11 32 t 163. BC PB (93%) 56 ± 14 50 ± 8 40 ± 5
4. CP PB (l5%)� 49 ± 3 36 ± 4 25 ± 13
5. CP PB (8%)� 24 ± 5 22 ± 3 15 ± I1
6. BC BM(49%) 24±3 16±3 13±57. BC BM (58%) 9 ± 2 6 ± 2 5 ± 1
8. CP BM(l2%)z2� 14±2 7±3 3±2
9. CP BM(l0%)L� 14±3 10±7 12±310. BC PB(60%) 19±4 15±7 10±7
11. BC BM (48%) 10 ± 0.4 4 ± 1 2 ± 1
12. CP PB(2l%)L� 18±7 11±5 4±4
13. CP PB(ll%)L1� 6±3 5±3 4±2
a 0, FAB classification; New, newly diagnosed; & blast and promyelocyte; PB, peripheral blood; �, � ± SD; BC, blast crisis; AC, accelerated
phase; CP, chronic phase.
Clinical Cancer Research 2861
mia-reactive allocytotoxic T cells were harvested from growth-
positive wells and specific lysis toward leukemic target cells and
K562 cells was quantitated by CRA. The cells were continu-
oushy cultured and fed with stimulator and feeder cells in flasks.
After 2-3 weeks of culture, a monoclonal antibody (OKT3) was
added to the culture for rapid expansion of the leukemia-reactive
allocytotoxic T cells.
Cytotoxicity Assays. The cytotoxic activity of NK-92
(radiated and nonradiated) and responding T cells against leu-
kemic targets was measured in a standard 4-h CRA. A fixed
number � ‘ Cr-labeled target cells (5 X l03/well) was tested for
susceptibility to four effector cell concentrations. The percent-
age of specific cytotoxicity in triplicate specimens was calcu-
lated as: percentage of 51Cr release = (average experimental
Association for Cancer Research. by guest on September 4, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
2862 Antileukemia Activity of a NK Cell Line
Tabl e 2 Specific lysis of human leukemia cel 1 lines by NK cell clones NK-9 2, YT, and a CTL clone (TALL-104)
Specific lysis (%)
NK92 TALL-104 YT
E:T ratio
9:1 3:1 1:1 9:1 3:1 1:1 9:1 3:1 1:1
K562
HL6OKG1NALM6
RAJITALL-l04
CEM/SCEMIF
94.1
87.964.672.6
86.057.3
56.657.5
91.2
75.3
53.8
56.8
75.4
53.2
48.842.1
82.1
79.6
43.7
52.4
70.0
44.1
34.7
39.1
88.5
43.0
2.7
67.8
22.2
-
2.7
1.5
85.2
16.0
0.5
55.6
10.2
-
1 .6
0.6
72.5
6.9
0
33.3
0.3
-
0.9
0.3
34.2
2.1
0.1
1.0
25.1
3.2
0.9
1.2
28.2
1.1
0
0.5
18.0
1.4
0.4
0.1
18.4
1.5
0
0
14.2
0.9
0.3
0.2
cpm-average spontaneous cpm)/(average maximum cpm-aver-
age spontaneous cpm X 100). 51Cr release of target cells alone
(spontaneous release) was always <25% of maximal 5tCr re-
lease (target cells in 5% Triton). CRA data were expressed as
specific lysis (%) at various E:T ratios. The degree of sensitivity
of patient-derived leukemic cell targets to each effector was
defined as insensitive (< 19% lysis), sensitive (20-49% lysis),
and highly sensitive (�S0% lysis) at a 9: 1 E:T ratio.
Experimental Animals. SCID mice (CB 17 scid/scid and
pfp/Rag-2), 6-8 weeks of age, were purchased from Taconic
Farms (Germantown, NY) and maintained in the Memoria
Sloan-Kettering Cancer Center Animal Laboratory in microiso-
lator cages under sterile conditions with a specific pathogen-free
environment. To determine the potential of NK-92 cells to
induce leukemia in vivo, 2 X l0� viable NK-92 cells in 0.3 ml
of PBS were administrated by either i.p. or iv. route every other
day for five injections in each animal. For s.c. inoculations, 2 X
l0� NK-92 cells were injected in the right flank of SCID mouse,
as described previously (23). Thereafter, rhIL2 (5 X l0� IU)
was administered, every other day, to all of the experimental
animals for 2 weeks by i.p. injection. Survival of the animals
was followed for at least 6 months after inoculation. From each
group, two SCID mice were sacrificed at the end of observation,
and peripheral blood and tissues from BM, spleen, liver, kidney,
lung, and brain were collected for histopathological and/or
FACS analysis.
To examine the in vito effects of NK-92 cells and other
effector cells on the growth of human leukemia xenografts, 5 X
106 leukemic target cells alone or mixed with 2 X l0� NK-92 or
other effector cells (E:T ratio, 4: 1 ) were injected s.c. into SCID
mice. rhIL2 was administrated as described above. The size of
s.c. growth of leukemic nodules was measured once a week after
inoculation, and survival of the animals was followed.
For study of the in vito tumoricidal capacity of NK-92
cells, leukemic cells derived from a T-ALL patient (TA27), an
AML patient (MA26), and a pre-B-ALL patient (BA31) were
adoptively grown and expanded in SCID mice by s.c. inocula-
tion. Recovered human leukemic cells from leukemic nodules
grown in SCID mice (first passage) were used for the experi-
ments. The SCID mice in each group received i.p. inoculations
of 5 X 10#{176}leukemic cells in 0.2 ml of PBS, and at the indicated
time after inoculation, 2 X l0� NK-92 cells in 0.4 ml of PBS
were administered by i.p. injection. The animals received either
one dose or a series of five doses of NK-92 cells administered
with or without rhIL2 administration, as described in “Results.”
Analysis of Tissue from Xenografted Animals. Histopa-
thology. Tissue sections from sacrificed SCID mice were
fixed in 10% neutral-buffered formalin, dehydrated and embed-
ded in paraffin, sectioned, and stained according to standard
histological techniques.
Flow Cytometry Analysis. Viable cells recovered from
various tissues were stained by FITC-conjugated or phyco-
erythrin-conjugated Mab, as described (24). A FACS scan flow
cytometer (Becton Dickinson) was used for analysis. The fol-
lowing Mabs were used for characterization of human antigens:
CD2, CD3, CD5, CD7, HLA-DR, CD45, and CD56 (Becton
Dickinson). An FITC-conjugated Rat antimouse Mab mCD45
(Boehringer Mannhein, Indianapolis, IN) was used for charac-
terization of murine leukocyte common antigen.
Statistical Analyses. The Log-rank test and Wilcoxon
test were used for the comparison of the survival of leukemia-
bearing SCID mice.
RESULTS
Cytolysis of Human Primary Leukemic Cells by NK-92
Cells. The sensitivity of patient-derived leukemic cells to the
cytotoxic effect of NK-92 cells is summarized in Table 1. CRAs
were performed two or three times for the leukemia targets and
the mean value of the percentage lysis of each target is presented
in Table 1. The pattern of cytotoxic sensitivity of these leuke-
mias was characterized into three types: insensitive, sensitive,
and highly sensitive according to the percentage of lysis at a
lower E:T ratio (9: 1). Of the 46 patient-derived heukemic sam-
ples, 24 (52.2%) were sensitive or highly sensitive to NK-92-
mediated in vitro cytotoxicity. Leukemia blasts derived from 6
of 12 (50%) AMLs, 7 of7 (100%) T-ALLs, and S of 14(35.7%)
B-hineage-ALLs were either sensitive or highly sensitive to
NK-92-mediated lysis. All of the eight acute leukemia samples,
which demonstrated high sensitivity to the cytotoxic effect of
NK-92 cells, were derived from relapsed patients. Of 13 CML
samples, 6 (46.2%) were sensitive (2 in chronic phase and 1 in
blast crisis) or highly sensitive (1 in accelerated phase and 2 in
blast crisis) to NK-92-mediated cytolysis (Table 1). In compar-
Association for Cancer Research. by guest on September 4, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
A. K562 B. HL6O
NK-92
C
0
0ti)
0.C’)
110
1:1
E:T Ratio
100 -
80 -
60 -
40CD56#{247}
20±�ii:i
0 CD3;:i
E:T Ratio
D. HL6OC. K562
12
0)C
0
0a)0.
C/)
C\J
E0
ci)
ci)0cci
�0
:�(I)
0.3:1
-�--K562
-�0-- K562+LAK-&. K562+CD56#{247}
-ci- K562+NK-92
t
4
2
-4-- HL6O-0-- HL6O+LAK-�1x-. HL6O#{247}CD56+
-0-- HL6O+NK-92
001 2345678
Weeks post-inoculation
9 0123456789
Weeks post-inoculation
Clinical Cancer Research 2863
Fig. 1 In vitro and in vivo antileukemic efficacy of NK-92 cells against K562 and HL6O leukemias as compared with human LAK cells and other
effectors. 51CR-labeled K562 (A) and HL6O (B) cells were tested for susceptibility to cytolysis by NK-92 cells as compared with various other
effectors [including LAK, NK (CD56�CD3), and T cells (CD3�CD56)] at indicated E:T ratios in a 4-h CRA. Results are means ± SD of threeseparate tests for NK-92 cells. and two tests for each of the same two donor- derived effectors for LAK, CD56�, and CD3� cells. SCID mice receive
s.c. inoculations of K562 cells (C) or HL6O cells (D; 5 X 106 each mouse) alone or combined with NK-92, LAK, or NK cells at a 4: 1 E:T ratio. The
tumor sizes were measured once a week after inoculation (n = 5).
ison, with the exception of three samples (a CML in blast crisis,
a T-ALL, and a pre-B-ALL), the majority of leukemic samples
were resistant to the other NK-like cell line tested (YT; data not
shown).
Cytotoxicity of NK-92 toward Human Leukemic Cell
Lines. NK-92 cells were highly cytotoxic to all of the eight
leukemic cell lines tested in a 4-h standard CRA (Table 2). The
MTX-sensitive T-ALL cell line CEM-S, as well as its MTX-
transport resistant subline CEM-T, displayed a similar sensitiv-
ity to the NK-92 cells. TALL-l04 was cytotoxic to K562,
NALM6, and HL6O cells; however, Raji cells exhibited only
22.2% lysis at a 9: 1 E:T ratio, and KG1 cells (CEM-S as well as
CEM-T) were resistant. The YT cell line did not exhibit signif-
icant cytotoxic activity to most leukemia cell lines, with the
exception of K562 cells and Raji cells, which showed a 32% and
25% lysis at a 9: 1 E:T ratio, respectively.
Effect of NK-92 Cells on Normal Human BM Hemato-
poietic Cells. Hematopoietic cell-enriched fractions of normal
BMs from 18 normal donors were tested by standard CRA to
determine their susceptibility to lysis by NK-92 cells. All of the
normal BM samples were insensitive to the NK-92-mediated
cytolysis. The median values of the cytolysis mediated by
NK-92 at a 90:1 and 9:1 E:T ratio were 5.5% (range, 1.9-
13.5%) and 3.2% (range, 0.6-12.1%), respectively.
In Vivo Leukemogenesis of NK-92 cells in SCID Mice.
CB-l7 scid/scid mice as well as pfp-Rag-2 mice received inoc-
ulations of NK-92 cells by iv. (n = 3, each group), s.c. (n = 2,
each group), and i.p. (CB-l7, n = 8; Rag-2, n = 3) injection.
Survival of the animals was followed at least 6 months after
inoculation. All of the animals appeared healthy, and there was
no hepatosplenomegaly, lymphadenopathy, or leukemic nodular
growth, which would indicate leukemia development in these
animals during the period of follow up. Leukemic cellular
infiltration was not detected in any of the tissues of the sacri-
ficed animals by histopathology, and there were no cells of
human origin detectable in the tissues by FACS analysis.
Comparison of Antileukemic Effect of NK-92 Cells with
LAK, NK, and T Cells against Human Leukemic Cell Lines.The antileukemic effects of NK-92 cells, human LAK cells, NK
cells (CD56�CD3), and T cells (CD3�CDS6), were assessed
Association for Cancer Research. by guest on September 4, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
A.
NK-92
0)C
0
0ci)0.
ci)
C’J
E0
ccici)
a)0
:sC’)
E:T Ratio
B.
�-G-�.-&. TA27 1L2 ip-0-- TA27 NK-92 5.2 p.-v-- TA27 TA27-T IL2 p
- 0 1 2 3 4 5 6 7 8 9 10
Weeks post-inoculation
2864 Antileukemia Activity of a NK Cell Line
by measuring in vitro cytolytic activity in standard CRA and by
measuring inhibition of heukemic cell xenograft growth in vivo
when the effector cells and targets were coinoculated s.c. into
SCID mice. NK-92 cells displayed high in vitro cytotoxicity
against K562 and HL6O with a mean specific lysis of 89% and
78% at a 9: 1 E:T ratio, respectively. This was superior to the
killing mediated by human LAK (52% and 1 1%), NK (72% and
28%), and T (12% and 1.2%) cells (Fig. 1, A and B). Correspond-
ingly, the NK-92 cells demonstrated a more intense in vivo inhibi-
rion of the growth of K562 and HL6O leukemic cells xenografts
than did the human LAK and NK cells (Fig. 1 , C and D).Comparison of Antileukemic Effect of NK-92 Cells with
Leukemia-reactive Allocytotoxic T Lymphocytes cells. To
compare the antileukemia effect of NK-92 cells with leukemia-
reactive allocytotoxic T lymphocytes, we generated allogeneic
leukemia-reactive T cells against leukemia derived from a pa-
tient with T-ALL (TA27). More than 98% of the leukemia-
reactive allocytotoxic T cells stained CD3�CD56 by FACS
analysis and cytolysis of the T cells against an NK-sensitive cell
line K562 at a 9: 1 E:T ratio was <3% in a CRA. Both NK-92
and the leukemia-reactive allocytotoxic T cells (TA27-T) dis-
played a significantly higher specific cytolysis (70% and 58% at
a 9: 1 E:T ratio, respectively) than the other effectors [LAK cells
(22%), NK cells (38%), and T cells (1.5% specific lysis),
respectively] against the TA27 leukemic cells (Fig. 14). Corre-
spondingly, the s.c. growth of TA27 leukemic cells was inhibited
by coinjection of either NK-92 cells or TA27-T cells (Fig. 2B).
Antileukemia Effect of NK-92 Cells in a Human Leu-
kemia Xenografted SCID Mouse Model. SCID mice who
received i.p. inoculations of human heukemic cells were treated
by i.p. injections of NK-92 cells with or without rhIL2. Leuke-
mic cells derived from a patient (TA27) with T-ALL and a
patient (MA26) with AML M4 leukemia were highly sensitive
to the NK-92 cells (73% and 66% specific killing at a 9:1 E:T
ratio in CRA, respectively), whereas cells from a patient with
pre-B-ALL (BA3 1 ) were insensitive to the NK-92 cells (4%
specific killing at a 9: 1 E:T ratio in CRA). All these human
leukemias grew aggressively in SCID mice. The survival of
mice bearing TA27 leukemia was significantly prolonged by the
NK-92 cell administration (Fig. 3). The MST of the animals
with no treatment or 1L2 alone was 72 days (n 5) and 63 days
(n 5; P >0.05), respectively. All of these animals died of
leukemia. In contract to the animals without treatment or 1L2
alone, treatment with NK-92 cells alone, or with addition of
rhIL2 significantly increased the MST to 102 days (n 6; P
<0.05) and 1 14 days (n = 6; P <0.05), respectively, for the
one-dose injection schedule (2 X l0� NK-92 cells, day 1). The
MST increased to 160 days (n = 6) and 129 days (n 6),
respectively, with five doses of NK-92 with or without rhIL2
injection (2 x l0� NK-92 cells, days 1, 3, 5, 7, and 9; Fig. 3).
Three animals who received five doses of NK-92 cell injections
with or without rhIL2 administration survived without any signs
of leukemia development 6 months after inoculation. Survival
was significantly prolonged for all of the groups that received
NK-92 cells (Fig. 3). There was no significant survival differ-
ence between the treatments with or without rhIL2 administra-
tion either in the one-dose group (P = 0.75), or the five-dose
group of NK-92 cell administration (P = 0.45) schedule. Com-
pared with the one dose of NK-92 cell injection with or without
Fig. 2 Antiheukemic effect of NK-92 cells, leukemia-reactive allocy-totoxic T cells, and other effectors against a patient-derived T-ALL(TA27). A, specific killing of TA27 target cells by NK-92, leukemia-
reactive allocytotoxic T cells (TA27-T), as well as other effectors, were
determined by a 4-h CRA using the indicated E:T ratios. Results aremeans ± SD of two or three separate tests. B, SCID mice received s.c.inoculations of TA27 cells (5 X 106 each mouse) alone or with NK-92,
TA27-T, or other effector cells at a 4: 1 E:T ratio. rhIL2 was adminis-tered to mice i.p. for 2 weeks at the dose of 5 X l0� IU every other day.
Leukemic tumor sizes were measured once a week after inoculation
(n 5).
rhIL2 treatment, survivals were significantly extended in ani-
mals that received five doses of NK-92 cells without rhIL2
treatment (P = 0.009 and P = 0.009, respectively).
In SCID mice who received inoculations of human pre-B-
ALL (BA3 1) leukemia with or without rhIL2 treatment, the
MST was 63 days (n 5) and 64 days (n 5), respectively. For
the animals who received 2 X l0� NK-92 cells for five doses
with or without rhIL2 administration, the MST was increased to
79 days (n = 5) and 76 days (n = 5), respectively. However,
these survival times were not significantly different from the
animals that were not treated by NK-92 cells (P >0.05; Fig. 4).
In animals bearing human AML (MA26) leukemia with or
without rhIL2 treatment, MST was 97 days (n 6) and 100
days (n = 6; Fig. 5). The MST was extended to 173 days in the
animals that received 2 X !0� NK-92 cells 5 times (P < 0.01;
n 6). Three of six animals who received NK-92 cells were
alive 6 months after leukemia inoculation. Two mice appeared
healthy without any signs of leukemia development. One mouse
Association for Cancer Research. by guest on September 4, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
>>‘a.
:,
100
80
60
40
20
0-40
- TA27- TA27IL2ip
_�_---_1 --. TA27 NK-92x1 p; TA27 NK-92x1 p 1L2 p
u.-��.-1 L�, �- TA27 NK-92x5 IP
I � � I �‘ TA27 NK-92x5 p lL2 p
L � I:�i Ii_1__�IL.� #{149}� #{149}� U � I � UII:
60 80 100 120 140 160 180
Days Post-inoculation
(6>>
100
80
60
40
20
040
n=5- BA31
- BA31 1L2 p n=5
�-- BA31 NK-92x5 p n=5
BA31 NK-92x5 ip lL2 p n=5
60 80 100
Fig. 4 Survival of SCID mice bearing pre-
B-ALL (BA3I) after treatment with NK-92cells. Mice received 5 X 106 BA31 cells i.p.
NK-92 cells were injected i.p. (2 X 10�) for
five doses from days 3-1 1 after leukemia
inoculation. Mice in the indicated groups
received rhIL2 every other day from day 3
after leukemia inoculation for 2 weeks.
Days Post-inoculation
Clinical Cancer Research 2865
Fig. 3 Survival of SCID mice
bearing T-ALL (TA27) after
treatment with NK-92 cells.
Mice received 5 x 106 TA27
cells i.p. NK-92 cells were in-
jected i.p. (2 X l0�) either onceor five times beginning one dayafter leukemia inoculation, withor without the addition of rhIL2
every other day for 2 weeks.
n=5n=5
n=6n=6n=6
n=6n=6
had an enlarged abdomen, indicating residual leukemia. The six
animals who received NK-92 cells plus rhIL2 treatment were all
alive 6 months after leukemia inoculation without any signs
suggestive of leukemia development (Fig. 5).
DISCUSSION
There have been numerous studies suggesting that NK cells
are important in immunosurveillance against leukemia in pa-
tients treated with or without the addition of hymphokines and/or
BM transplantation (4, 24, 25). However, the biology of NK
cells has been less well characterized than that of T and B
lymphocytes. The difficulty of isolating large numbers of highly
purified NK cells and the problems of amplifying or maintaining
them in vitro as NK cell lines, which have high cytotoxic
activity against different malignant target cells has been a major
obstacle in the study of the biology of this particular lymphocyte
subset. These obstacles also hamper the application of NK cells
in cancer-adoptive immunotherapy. In this study, we were able
to overcome a number of these obstacles by the use of a well
characterized NK cell line.
In the present study, the NK cell line NK-92 displayed
marked cytotoxic activity against a wide range of human pri-
mary leukemias and leukemic cell lines (Tables 1 and 2). The
patient-derived leukemic cells in our study were susceptible to
NK-92 cells (defined as �20% specific lysis) at a 9: 1 E:T ratio
in a standard 4-h CRA. The mean Cr-release was 47.8 ± 14.2%
in the NK-92-sensitive (n = 24) versus 7.6 ± 6.0% in the
resistant (n = 22) leukemias. This is similar to the finding of a
recent report in which 51% of 150 fresh leukemic cell samples
were found to be susceptible to allogeneic LAK cells (25).
However, in that study, a much higher E:T ratio (50: 1) was
required to detect comparable levels of cytotoxic activity against
leukemic cells by the allogeneic LAK cells. This finding is
consistent with our data. The NK-92 cells displayed higher
cytotoxic activity than LAK cells and normal donor-derived NK
cell (CD56�CD3) populations against T-lymphoblastic leuke-
Association for Cancer Research. by guest on September 4, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
100 -
80
60
40 -
20
(6>>
C/)
I I
n=6
n=6
n=6
lL2 p n=6
Fig. 5 Survival of SCID mice bearing hu-
man AML (MA26) after treatment with
NK-92 cells. Mice received 5 X 106 MA26leukemia cells i.p. NK-92 cells were in-
jected i.p. (2 X h0�) for five doses from
days 3-1 1 after leukemia inoculation. Mice
in the indicated groups received rhIL2 every
other day from day 3 after leukemia inocu-
hation for 2 weeks.
40 60 80 100 120 140 160 180
2866 Antileukemia Activity of a NK Cell Line
4 Y. Yan. B. H. Childs, and R. J. O’Reilly, unpublished data.
Days Post-inoculation
- MA26
- MA26 lL2 ip
--. MA26 NK-92x5 p
MA26 NK-92x5 p
mia (TA27), K562, or HL6O leukemic cells (Figs. 1 and 2) in
both the CRA and coinocuhation assays in SCID mice. These
observations suggest that the cytotoxicity of NK-92 cells against
human leukemias may be superior to that of the heterogeneous
populations of LAK cells or isolated NK cells from the blood of
normal donors.
Susceptibility or resistance to NK-cehl-mediated cytotoxic-
ity may vary with the phenotype and HLA background of the
leukemic cells (4, 25-27). Some studies have suggested that
AMLs and T-ALLs may be more resistant to lysis by LAK or
isolated NK cells from normal donors than CML or B-lineage
ALL. Chronic phase CML cells have been found to be more
resistant than CML cells in blast crisis (25-27). However, in our
study. leukemic samples from T-ALLs and AMLs were more
susceptible to the cytohysis of NK-92 cells than the samples
from B-lineage-ALL and CML. The molecular basis for the
observed resistance or susceptibility of leukemia cells to NK-
mediated cytolysis is not presently understood. Several studies
have shown that modulation of expression of HLA class I alleles
can render cells more or less susceptible to NK cell lysis,
presumptively by HLA class I allele engagement of receptors,
which down-regulate NK-cell-induced cytohysis of cells (28-
30). Another study has demonstrated that the expression of the
class I molecule HLA-G can protect target cells from lysis by
NK cells (3 1 ). A recent report has also provided data suggesting
that the expression of the cytoskeletal-membrane linker protein,
ezrin, by leukemic cells induces concentration of the adhesion
molecule ICAM-2 into uropods on the cell membrane, which
renders the cells susceptible to NK-cell-mediated lysis (32).
However, further study in a large population of patient-derived
leukemia samples will be required to confirm these findings and
to clarify their relative contributions to heukemic target cell
recognition and killing by NK-92 cells.
In assessing the activity of isolated NK cells or NK cell
lines against leukemic cells, it is also important to determine
whether normal BM cells are susceptible to lysis by NK cells. In
the present study, we found that T-cell-depleted normal marrow
cells were resistant to the cytohysis of NK-92 cells. Similarly, a
recent report also demonstrated that NK-92 cells can selectively
kill leukemia cells without inducing direct cytotoxicity against
normal hematopoietic progenitors and marrow cells (20). On the
basis of these observations, it was hypothesized that NK-92 cells
could be used for purging BM for autologous BM transplanta-
tion as well as for adoptive immunotherapy (20).
A series of studies of the MHC-unrestricted, cytotoxic T
cell leukemia cell line, TALL-104, have demonstrated that these
T cells, after irradiation to doses ablating their capacity to
proliferate, can maintain a highly cytotoxic effect against ma-
lignant cells (15, 16). The radiated TALL-l04 cells display a
tumoricidal effect against a wide spectrum oftumor cells both in
vitro and in vivo in experimental animals without inducing the
development of leukemia (15-18). Similarly, in our study,
NK-92 cells irradiated to doses of 500 cOY, which absolutely
ablates their clonogenic activity, maintain their cytolytic activity
against heukemic cell targets.4 The effects of radiation on the
leukemogenic potential of NK-92 cells was difficult to assess
because unirradiated cells failed to induce leukemias in CB 17
SCID mice or in pfplRag-2 knockout mice even when these
animals were treated with rhIL-2. Because the pfp/Rag-2 SCID
mice are NK cell-deficient, it seems unlikely that the murine NK
cells could prevent the engraftment and growth of the NK-92
cells in these animals after administration of rhIL2. Neverthe-
less, lethal irradiation or perhaps other alternative methods, such
as suicide (thymidine kinase) gene transfer, could be used to
ablate the clonogenic potential of these cells and, thereby,
eliminate even the small potential risk of in vivo leukemia
development by allogeneic NK-92 or TALL-104 cells, yet not
effect their highly tumoricidal activity, allowing the possibility
that adoptive transfer of such cells could prove to be practical
and potentially useful in the treatment of NK cell sensitive,
chemotherapy refractory tumors.
As shown in Figs. 2, 3, and 5, adoptive transfer of NK-92
cells was active in the treatment of xenografted human leukemia
in the SCID mouse model. The survival of animals bearing
Association for Cancer Research. by guest on September 4, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
Clinical Cancer Research 2867
5 Y. Yan, B. H. Childs, and R. J. O’Reilly, unpublished observation.
human T-ALL as well as AMLs that were sensitive to in vitro
cytolysis by NK-92 cells, was not only significantly prolonged
by NK-92 cell treatment with or without rhIL2, but some of the
animals did not develop leukemia at all during a long-term
follow-up period (Figs. 3 and 5). During the same follow-up
interval, no animals were curable in this T-ALL SCID mouse
model when treated by maximally tolerated doses of either
MTX or idarubicin.5
Our results indicate that primary leukemias vary in their
susceptibility to NK-92 and other NK-like effector cells, but that
the in vitro susceptibility of a given primary leukemia to the
NK-92-cell-mediated cytolysis is predictive of its in vivo sus-
ceptibility. The sensitivity of such primary leukemias to the
cytolytic activity of NK-92 cells or other NK cells cannot be
predicted on the basis of the activity of these effectors against
established leukemia cell lines. Indeed, in our study, all of the
eight leukemic cell lines displayed marked susceptibility to
cytolysis by NK-92 cells at a 9: 1 E:T ratio (Table 2). Addition-
ally, in a previous report, NK-92 cells were found to be highly
cytotoxic against 14 established leukemia, lymphoma, and my-
eloma cell lines at a 10: 1 E:T ratio (20). Our studies, therefore,
suggest that the determination of the in vitro susceptibility of
patient-derived leukemia samples to the NK-92 cells might be
an informative predictor of activity if NK-92 cells were to be
used for in vitro marrow graft purging or for in vivo adoptive
immunotherapy.
In conclusion, this study represents the first data available
on the effect of NK-92 cells against primary patient-derived
leukemias. The in vitro and in vivo antileukemic activity of
NK-92 cells suggests that this type of cell and other well
characterized NK cells, may have potential as antileukemic
effectors in adoptive immunotherapy.
ACKNOWLEDGMENTS
We are grateful to Sharon Bleau for BM processing assistance, toTeresa Diaz-Barrientos for CU cells culture, to Thoma Delohery andMacthu Menon for flow cytometry technical assistance, and to ThomasWilliams for statistics consultation.
REFERENCES
1. Brittenden, J., Heys, S. D., Ross, J., and Eremin, 0. Natural killer
cells and cancer. Cancer (Phila.), 77: 1226-1243, 1996.
2. Robertson, M. J., and Ritz, J. Biology and clinical relevance ofhuman natural killer cells. Blood, 76: 2421-2438, 1990.
3. Trincheri, G. Biology of natural killer cells. Adv. Immunol., 47.
187-375, 1989.
4. Pawelec, G. MHC-unrestricted immune surveillance of leukemia.Cancer Biother., 9: 265-288,1994.
5. Lotzova, E., Savary, C. A., and Herberman, R. B. Induction of NKactivity against fresh human leukemia in culture with interheukin 2.
J. Immunol., 138: 2718-2727, 1987.
6. Melder, R. J., Whiteside, T. L., Vujanovic, N. L., Hiseridt, J. C., andHerberman, R. B. New approach to generating antitumor effectors for
adoptive immunotherapy using human adherent lymphokine-activatedkiller cells. Cancer Res., 48: 3461-3469, 1988.
7. Yasumura, S., Lin, W., Hirabayashi, H., Vujanovic, N. L., Herber-
man, R. B., and Whiteside, T. L. Immunotherapy of liver metastases of
human gastric carcinoma with IL2-activated natural killer cells. Cancer
Res., 54: 3808-3816, 1994.
8. Vujanovic, N. L., Yasumura, S., Hirabayashi, H., Lin, W. C.,
Watkins, S., Herberman, R. B., and Whiteside, T. L. Antitumor activi-
ties of human IL2-activated natural killer cells in solid tumor tissues.
J. Immunol., 154: 281-289, 1995.
9. Richards, S. J., and Scott, C. S. Human NK cells in health anddisease: clinical, functional, phenotypic and DNA genotypic character-
istics. Leuk. Lymphoma, 7(5-6): 377-399, 1992.
10. Rosenberg, S. A., Lotze, M. T., Muul, L. M., Chang, A. E., Avis,
F. P., Leitman, S., Linehan, W. M., Robertson, C. N., Lee, R. E., Rubin,J. T., Seipp, C. A., Simpson, C. G., and White, D. E. A. Progress reporton the treatment of 157 patients with advanced cancer using lympho-kine-activated killer cells and interheukin-2 or high-dose interleukin-2
alone. N. EngI. J. Med., 316: 889-897, 1987.
11. Hercend, T., Reinherz, E. L., Meue, S., Schlossman, S. F., and Ritz,J. Phenotypic and functional heterogeneity of human cloned naturalkiller cell lines. Nature (Lond.), 301: 158-160, 1983.
12. Yodoi, J., Teshigawarw, K., Nikaido, T., Fukui, K., Noma, T.,
Honjo, T., Takigawa, M., Sasaki, M., Minato, N., Tsudo, M., Uchiyama,T., and Maeda, M. TCGF(1L2)-receptor inducing factor(s). I. Regulationof IL2 receptor on a natural killer-like cell line (YT cells). J. Immunol.,
134: 1623-1630, 1985.
13. Fernandez, L. A., Pope, B., Lee, C., and Zayed, E. Aggressivenatural killer cell leukemia in an adult with establishment of an NK cellline. Blood, 67: 925-930, 1986.
14. Robertson, M., Cochran, K., Cameron, C., Lee, J. M., Tantravahi
R., and Ritz. J. Characterization of a cell line, NKL, derived from anaggressive human natural killer cell leukemia. Exp. Hematol., 24: 406-415, 1996.
15. Cesano, A., Visonneau, S., Cio#{233},L., Clark, S. C., and Santohi, D.Effects of lethal irradiation and cyclosporin A treatment on the growth
and tumoricidal activity of a T-cehl clone potentially useful in cancer
therapy. Cancer Immunol. Immunother., 40: 139-151, 1995.
16. Cesano, A., Pierson, G., Visonneau, S., Mighiaccio, A. R., andSantohi, D. Use of a lethally irradiated major histocompatibihity complexnonrestricted cytotoxic T-cell line for effective purging of marrows
containing lysis-sensitive or-resistant leukemic targets. Blood, 87: 393-
403, 1996.
17. Cesano, A., Visonneau, S., Cio#{233},L., Clark, S. C., Rovera, G., andSantoli, D. Reversal of acute myelogenous leukemia in humanizedSCID mice using a novel adoptive transfer approach. J. Clin. Invest., 94:
1076-1084, 1994.
18. Cesano, A., Visonneau, S., Jeglum, K. A., Owen, J., Wilkinson, K.,Carner, K., Reese, L., and Santohi, D. Phase I clinical trial with a human
major histocompatibihity complex nonrestricted cytotoxic T-celh line
(TALL-104) in dogs with advanced tumors. Cancer Res., 56: 3021-
3029, 1996.
19. Gong, J., Maki, G., and Khingemann, H-G. Characterization of ahuman cell line (NK-92) with phenotypical and functional characteris-
tics of activated natural killer cells. Leukemia (Baltimore), 8: 652-658,
1994.
20. Klingemann, H-G., Wong, E., and Maki, G. A cytotoxic NK-cell
line (NK-92) for ex vivo purging of leukemia from blood. Biol. Blood
Marrow Transplant., 2: 68-75, 1996.
21. Mini, E., Moroson, B. A., Franco, C. T., and Bertino, J. R. Cyto-
toxic effects of folate antagonists against methotrexate-resistant humanheukemic lymphoblast CCRF-CEM cell lines. Cancer Res., 45: 325-
330, 1985.
22. Reisner, Y., Kapoor, N., Kirkpatrick, D., Pollack, M. S., Dupont,
B., Good, R. A., and O’Reilly, R. J. Transplantation for acute leukaemiawith HLA-A and B nonidentical parental marrow cells fractionated with
soybean agglutinin and sheep red blood cells. Lancet, 2: 327-331, 1981.
23. Yan, Y., Ophira, S., McGuirk, J., Dennig, D., Fernandez, J.,Jagiello, C., Hai, N., Collins, N., Steinherz, P., and O’Reilly, R. J.
Growth pattern and clinical correlation of subcutaneously inoculated
Association for Cancer Research. by guest on September 4, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
2868 Antileukemia Activity of a NK Cell Line
human primary acute leukemias in SCID mice. Blood, 88: 3137-3146,
1996.
24. Dokhelar, M. C., Wiels, J., and Lipinski, M. Natural killer cell
activity in human bone marrow transplantation. Early appearance of
peripheral natural killer activity in graft versus host disease. Transphan-
tation (Baltimore), 31: 61-65, 1981.
25. Teichmann, J. V., Ludwig, W. D., and Thiel, E. Cytotoxicity of
interleukin 2-induced lymphokine-activated killer (LAK) cells against
human leukemia and augmentation of killing by interferons and tumor
necrosis factor. Leuk. Res., 16: 287-298, 1992.
26. Teichmann. J. V., Ludwig, W. D., and Thiel, E. Susceptibility ofhuman leukemia to allogeneic and autologous lymphokine-activated killer
cells activity-Analysis of 252 samples. Nat. Immun., 11: 1 17-132, 1992.
27. Pattengale, P. K., Sundstrom, C., Yu, A. L., and Leving, A. Lysis of
fresh leukemic blasts by interferon-activated human nature killer cells.
Nat. Immun. Cell Growth Regul., 3: 165-180, 1984.
28. Malnati, M. S., Peruzzi, M., Paker, K. C., Biddison, W. E., Ciccone,
E., Moretta, A., and Long, E. 0. Peptide specificity in the recognition of
MHC class I by natural killer cell clones. Science (Washington DC),267: 1016-1018, 1995.
29. Colonna, M., and Samaridis, J. Cloning of immunoghobihin-super-
family members associated with HLA-C and HLA-B recognition by
human natural killer cells. Science (Washington DC), 268: 405-408,
1995.
30. Litwin, V., Gumperz, J., Parham, P., Phillips, J. H., and Lanier,L. L. Specificity of HLA class I antigen recognition by human NK
clones: evidence for chonal heterogeneity, protection by self and non-
self alleles, and influence of target cell type. J. Exp. Med., 178: 1321-
1336, 1993.
31. Pazmany, L., Mandelborim, 0., Val#{233}s-G#{243}mez,M., Davis, D. M.,Reyburn, H. T., and Strominger, J. L. Protection from natural killer
cell-mediated hysis by HLA-G expression of target cells. Science(Washington DC), 274: 792-795, 1996.
32. Helander, T. S., Carp#{233}n,0., Turunen, 0., Kovanen, P. E., Vaheri,A., and Timonen, T. ICAM-2 redistributed by ezrin as a target for killercells. Nature (Lond.), 382: 265-268, 1996.
Association for Cancer Research. by guest on September 4, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from