From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
Human Lymphokine Activated Killer (LAK) Cells Suppress Generation of
Allospecific Cytotoxic T Cells: Implications for Use of LAK Cells To Prevent
Graft-Versus-Host Disease in Allogeneic Bone Marrow Transplantation
By Joseph Uberti, Frank Martilotti, Ta-Hsu Chou, and Joseph Kaplan
We have found that murine lymphokine activated killer (LAK)
cells have veto and natural suppressor activities in vitro, and
prevent graft-versus-host disease (GVHD) in vivo. To determine whether human LAK cells mediate veto and natural
suppression we measured their ability to inhibit generation
of allospecific cytotoxic T cells (CTL) in mixed lymphocyte
culture (MLC). When added to MLCs at low concentrations
LAK cells caused veto-type inhibition: stimulator-type LAK
cells inhibited generationof CTL but responder or third-party
LAK cells did not. At higher concentrations LAK cells caused
nonspecific inhibition: all three LAK cell types inhibited
generation of CTL. LAK cell veto and natural suppressor
activities were largely eliminated by irradiation with 30 Gy
and by depletion of CD56’ cells, but increased after depletion
of CD3’ cells. LAK cell veto activity is not likely an artifact of
w
HAVE PREVIOUSLY shown that murine lymphokine-activated killer (LAK) cells have potent veto
and nonspecific suppressor activity.’ Thus, in addition to
developing the characteristic ability of LAK cells to kill
both natural killer (NK)-sensitive as well as NK-resistant
target cell lines: murine bone marrow and spleen cells
cultured with interleukin-2 (IL-2) also develop the ability to
specifically and nonspecifically inhibit the in vitro generation of allospecific cytotoxic T lymphocytes (CTL). At low
concentrations in mixed lymphocyte cultures (MLCs) murine LAK cells exhibit what has been termed3 veto activity:
they block generation of allospecific cytotoxic T lymphocytes (allo-CTL) against their own but not against unrelated major histocompatibility complex (MHC) antigens.
When present at higher concentrations in MLCs, they
exhibit what has been termed4 natural suppressor activity:
they nonspecifically inhibit generation of allo-CTL against
both their own and unrelated MHC antigens. Studies we
have conducted in murine bone marrow transplant model
systems indicate that LAK cells also block alloreactions in
vivo. Thus, adoptively transferred LAK cells inhibit the
resistance of lethally irradiated mice to short-term growth
of allogeneic marrow stem cells, prevent lethal graft-versushost disease (GVHD), and permit long-term engraftment
of allogeneic bone marrow: These findings suggest that if
human LAK cells, like murine LAK cells, exhibit veto
and/or natural suppressor activity they might prove useful
in preventing graft rejection and GVHD in human recipients of mismatched allogeneic bone marrow transplants.
With that in mind, we have tested human LAK cells for in
vitro veto and natural suppressor activity. The results of
these experiments show that human LAK cells, like murine
LAK cells, inhibit generation of allospecific CTL against
both self and nonself MHC antigens.
MATERIALS AND METHODS
Isolation ofperipheralblood lymphocytes. Peripheral blood mononuclear cells (PBMC) were obtained by Ficoll-Hypaque (Pharmacia, Piscataway, NJ) separation from heparinized blood of healthy
adult donors.
Blood, Vol79, No 1 (January 1). 1992: pp 261-268
cold-target inhibition by the LAK cells themselves or by
proliferation of T cells contaminating LAK cell preparations:
(1)veto only occurred when LAK cells were added to MLC on
days 0 through 2, but not when added on day 5; (2) addition
of saturating numbers of labeled targets to fixed numbers of
allo-CTL effectors failed to overcome the inhibitory effects of
adding stimulator-type LAK cells at the onset of MLC; and (3)
CD3-depleted LAK cells showed greater veto activity than
threefold greater numbers of control LAK cells. In light of our
previous findings in mice, the current results imply that
adoptive immunotherapy with LAK cells may be useful in
preventing GVHD in human bone marrow transplant recipients.
o 1992 by The American Society of Hematology.
Preparation of LAK cells. PBMC were cultured for 4 days in
T-25 tissue culture flasks at 1 x lo6 cells/mL in complete medium
(RPMI-1640 with 4 mmol/L L-glutamine, 5 mmol/L HEPES, 5 x
lo-’ mmol/L 2-mercaptoethanol, Pen/Strep, and 10% human AI3
serum) together with 1,500 U/mL IL-2 (Cetus Corp, Emeryville,
CA).
LAKassay. The Daudi cell line used as target cells for the LAK
assays was maintained in suspension culture in RPMI-1640 with
10% fetal bovine serum. Two to four million target cells were
labeled with 200 pCi of Na,CrO, ”Cr for 90 minutes at 37°C. After
two washes the labeled cells were brought to 5 x 10, cells/mL and
distributed in 0.1 mL aliquots to 96-well round-bottom microculture plates containing in triplicate 0.1 mL of various dilutions of
PBMC effector cells such that the final effector to target (E:T)
ratios were 40:1,20:1, lO:l, and 5:l. Plates were then centrifuged at
40g for 5 minutes and incubated for 4 hours at 37°C. At the end of
incubation the plates were centrifuged at 4% for 5 minutes and
aliquots of the supernates were collected and counted in a gamma
counter. Spontaneous release was determined from the cpm of
supemates of target cells incubated alone without added effector
cells. Total counts per minute (cpm) incorporated was determined
from the cpm of known numbers of labeled target cells. Percent
spontaneous release (spontaneous cpm/total cpm) was less than
20% in all instances. Percent specific lysis was calculated by the
formu 1a :
(E - S)
(T - S)
% Specific Lysis = 100 x -
From the Division of Hematology/Oncology, Department of Medicine, Wayne State University School of Medicine, Detroit, MI.
Submitted January 28,1991; accepted August 29,1991.
Supported by an American Cancer Society Institutional Research
Grant (J. U.),grants from the Children’s Leukemia Foundation of
Michigan ( J K , J.U.), the Elsa Pardee Foundation (J.U.), and Grant
No. HL16008 (J.K.)from the NHLBI.
Address reprint requests to Joseph Kaplan, MD, Children’s Hospital
of Michigan, 3901 Beaubien Blvd, Detroit, MI 48201.
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked
“advertisement” in accordance with 18 U.S.C. section 1734 solely to
indicate this fact.
0 I992 by The American Society of Hematology.
0006-4971I921 7901-0008$3.00/0
26 1
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
UBERTI ET AL
262
where E is experimental cpm, S is spontaneous cpm, and T is total
cpm incorporated. As previously described: an exponential fit
equation was applied to this data using a computer program for
least squares parameter estimation to calculate lytic units (LU)/107
effector cells where 1LU is the number of effector cells required to
lyse 20% of 5 x 103target cells.
Cell marker analysis. LAK cells were stained with a mixture of
fluorescein isothiocyanate (FITC)-conjugated anti-CD3 (BectonDickinson San Jose, CA) and phycoerythrin (PE)-conjugated
anti-CD56 (NKH-I, Coulter, Hialeah, FL). Control cells were
stained with FITC-IgG and PE-IgG to provide an estimate of
nonspecific labeling. After correcting for nonspecific staining, the
percentage of cells staining positively for one or both markers was
determined by flow microfluorimetry using a Coulter EPICS C flow
cytometer (Coulter, Hialeah, FL) gated on lymphocytes by 90” and
forward angle light scatter.
Generation of allospecific cytotoxic lymphocytes. Mixed lymphocyte cultures were performed by culturing 2 million responder
PBMC and 2 million 30-Gy irradiated stimulator PBMC in 2 mL
complete medium for 6 days in the presence or absence of various
numbers of added LAK cells. The cells were then harvested,
washed twice, and tested for cytotoxicity against stimulator or
responder-type target cells (3-day PHA-stimulated lymphocytes) in
a 4-hour chromium release assay identical to that described
previously.
Suppression assays. LAK cells were tested for specific and
nonspecific suppression of generation of allo-(JTL by adding
stimulator-type, responder-type, and/or third-party-type LAK cells
to MLC and then testing each MLC for allo-CTL against stimulatortype targets. The relative level of allospecific cytotoxicity generated
in each MLC culture was calculated as follows:
LU/107Added
LAK Cells
Relative Cytotoxicity = 100 x
LU/107Cells Without
Added LAK Cells
Kinetic analysis. In experiments designed to determine whether
competitive or noncompetitive inhibition accounted for any observed suppression of generation of allo-CTL mediated by added
responder-type LAK cells, tests for allospecific cytotoxicity were
performed using an approach similar to that described by Callewaert et al’ based on the model of enzyme kinetics. According to
this approach, use of saturating concentrations of labeled targets
should overcome competitive or cold-target inhibition but not
noncompetitive inhibition. ABx MLCs were performed as described previously in the absence or presence of stimulator-type
“B” LAKcells at a ratio of 1 LAK cek4 responder cells. At the end
of 6 days of culture the MLCs were harvested and tested for
allo-CTL using a fixed number of effector cells/well (5 X lo4) and
increasing concentrations of chromium-51 labeled “B” targets
ranging from 2.5 x lo4 to 3 x 10S/well. In addition, ABx MLC
effectors cells were analyzed in this fashion in the absence or
presence of a fixed number (5 x lo4) of unlabeled B targets to
verify that cold target inhibition could in fact be overcome by
saturating concentrations of labeled targets. Supernatants were
collected and counted after 2 or 4 hours incubation. Velocity (v),
the number of target cells lysed per hour, was calculated by the
following formula: v = (E - S)T/(M - S)t. Experimental (E),
spontaneous release (S), and maximal release (M) cpm values were
obtained in triplicate assays at each target cell concentration (T)
and at each interval of time (t). The percent of inhibition was
calculated as 100 x (1 - v experimentah control ABx).
Complement lysis depletion. LAK cells, 5 x lo6cells in 0.5 mL,
were incubated with 25 pg monoclonal anti-CD56 (“-1;
Coulter), 15 pg monoclonal anti-CD3 (OKT3; Ortho Diagnostics,
Raritan, NJ), or medium alone for 30 minutes at 37°C. After one
wash 0.3 mL of a 1:4 dilution of rabbit complement (Accurate
Chemical & Scientific Corp, Westbury, NY)was added and the
cells further incubated for 60 minutes at 37°C in 5% CO,. Cells
were then washed three times in medium. By flow cytometric
analysis cells remaining after complement lysis with anti-CD56 or
anti-CD3 contained less than 5% CD56’ or CD3’ cells, respectively. Because these experiments were designed to give a quantitative estimate of the relative contribution of antibody-positive cells
to the observed LAK effects we did not correct for cell losses after
treatment. This approach avoids a potential artifact encountered
when one corrects for cell losses by cell reconstitution, ie, artificial
enrichment of any antibody-negative cells that might mediate the
same effects as antibody-positive cells. Thus, after cell counts, each
of the LAK cell suspensions were resuspended to the same volume,
and equal volumes were then added to labeled Daudi target cells in
tests for LAK activity, and to MLCs in tests of veto and natural
suppressor activity.
Statistical analysis. Statistical tests were performed using
Statview I1 software (Abacus Concepts, Berkely, CA). Levels of
allospecific cytotoxicity generated by different MLCs were compared by the Mann-Whitney U test. Relative cytotoxicity values
were compared by ANOVA using the Scheffe F test. One-sample
and two-sample paired t-tests were used to assess the effects of
complement lysis-depletion of CD3’ and CD56’ cells on LAK
activity. In all cases significance levels were set at .05.
RESULTS
Human LAK cells exhibit both veto and natural suppressor
activities. To determine whether human LAK cells, like
murine LAK cells, exhibit veto and/or natural suppressor
activity, several different approaches were taken. In the first
approach, increasing doses of LAK cells from donor “B”
were added on day 0 to A(Bx Cx) MLCs containing
responder cells from subject “A,” and irradiated fresh
stimulator cells from LAK donor “B” and an unrelated
subject “C.” MLCs were harvested on day 6 and tested for
specific cytotoxicity against stimulator-type PHA blasts.
The results of a single experiment are shown in Fig 1.
(Similar results were obtained in three additional experiments.) They show that LAK cells are highly efficient
inhibitors of generation of allo-CTL directed against their
own MHC antigens, and at lower efficiency also inhibit
generation of allo-CTL directed against unrelated MHC
antigens. At the LAK cel1:MLC responder cell ratio of 1:8
added “B” LAK cells significantly inhibited by 67% the
generation of allo-CTL by “A” responder cells directed
against homologous “B” targets (P = .02), but failed to
significantly inhibit the generation of allo-CTL by “ A ’
responder cells against unrelated third-party “C” targets
(P = .11). At higher LAK cel1:MLC responder cell ratios of
1:2, and 1:1, added “B” LAK cells caused significant
inhibition of the reaction against both homologous “B”
targets and allogeneic “C” targets (P = .04). However, even
at this higher concentration, added “B” LAK cells inhibited
the reaction against homologous B targets significantly
more than they inhibited the reaction against third-party C
targets (P = .02).
These results implied that LAK cell veto-type “selfspecific” inhibition is more efficient than LAK cell nonspe-
+
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
263
HUMAN LAK CELL VETO AND NATURAL SUPPRESSION
lZO
100
.-2.
.-”
1
“Y
s
c
..o
x
80
0
c
&’
60
0
.-P
-
40
60
0
c
c
e
t
40
.-a>
4-
20
Q
20
0
Anti-B
0
0
1: 8
1: 4
1: 2
1: 1
Anti-C
Specificity of Allo-CTL
B LAK Cell : A Responder Cell
Fig 1. Dose-response comparison of the ability of human LAK cells
t o inhibit generation of allospecific CTL against (1)homologous target
cells and (2) target cells from an unrelated donor. Increasing numbers
of LAK cells from donor ”B” were added on day 0 t o an A(Bx + Cx)
MLC containing responder cells from subject “A’and irradiated fresh
stimulator cells from LAK donor ”B” and unrelated subject “C.“
Cultures were harvested on day 6 and tested for specific cytotoxicity
against stimulator-type PHA blasts. Control MLC without added LAK
cells generated 21.2 and 10.6 LU/10 cells against B and C targets,
respectively. (m), Anti-B; (M), antLC.
cifie inhibition. However, it remained possible that in these
experiments lymphocytes from third party subjects “C”
were by chance either intrinsically more efficient in generating allo-CTL or more susceptible as targets for cytotoxicity
by allo-CTL than thc lymphocytes from subjects “B.” This
possibility was ruled out by the results of crisscross experiments in which LAK cells from subjects “B” or “C” were
added to A(Bx + Cx) MLC containing responder cells of
subject “A” and a mixture of equal numbers o f “B” and
“C” stimulator cells (Fig 2). “B” LAK cells added at a ratio
of 1 LAK c e k 2 responder cells significantly inhibited
generation of anti-B allo-CTL ( P = .02) but not generation
of anti-C allo-CTL, whereas the opposite was true of “C”
LAK cells.
In the experiments shown above in Figs 1 and 2, subjects
B and C , although unrelated, may have shared one or more
HLA antigens. Therefore, some of the observed inhibition
by B LAK cells of MLC-induced anti-C allo-CTL could
have been due to veto-type inhibition rather than nonspecific inhibition. To provide an unequivocal measure of
nonspecific inhibition wc tested the ability of LAK cells
generated from the responder “ A ’ or the stimulator “B” to
inhibit generation of anti-B CTL when added to an ABx
MLC. Because the allo-CTLgencrated in an A anti-B MLC
recognize only those MHC antigens on 8 targets not shared
by A and B, veto could not contribute to any inhibition that
might be exerted by responder-type A LAK cells on the
generation of anti-B-specific allo-CTL. As shown in Fig 3,
at the LAK ce1l:MLC responder cell ratio of 1:2 only
stimulator-typc B LAK cells inhibited generation of cytotox.
LAK
icity for B-target cells’ However, at
responder cell ratio of l: l both stimulator-type B LAK cells
and responder-type A LAK cells inhibited generation of
Fig 2. Crisscross comparison of ability of LAK cells t o specifically
and nonspecifically inhibit generation of allo-CTL. Identical numbers
of LAK cells from donors “ B or “C” were added on day 0 to an
A(Bx Cx) MLC containing responder cells from subject “A’ and
irradiated fresh stimulator cells from LAK donors ” B and ”C.” The
ratio of added LAK cells t o ” A responder cells was 1:2. Cultures were
harvested on day 6 and tested for specific cytotoxicity against
stimulator-type PHA blasts. Control MLC without added LAK cells
generated 13.4 and 11.8 LUIIO’ cells against B and C targets,
respectively. LAK cells added t o A(Bx + Cx) MLC: (M), B LAK cells;
(U),C LAK cells.
+
anti-B CTL, although, as shown previously, responder-type
A LAK cells were less inhibitory than stimulator-type B
LAK cells. These results confirm that in addition to their
veto-type inhibitory activity LAK cells also exhibit natural
suppressor-type nonspecific inhibitory activity.
Table 1 provides a summary of the results of eight
experiments comparing the effects of responder-type, stimulator-type, or third-party-type LAK cells on MLC genera-
lZO
100
l
40
20
0
0
1:2
L
1: 1
LAK Cells : Responder Cells
Fig 3. Human LAK cells inhibit generation of allo-CTL against both
their o w n and unrelated MHC antigens. Two different doses of LAK
cells from responder “ A and stimulator ”B“ were added on day 0 t o
ABx MLC. Cultures were harvested on day 6 and tested for specific
cytotoxicity against stimulator-type PHA blasts. Control MLC generationof anti-B CTL was 3.7 LU/107 cells,
LAK cells
added to AB^ MLC:
(.),A
LAK; (B), B LAK.
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
UBERTI ET AL
264
Table 1. Comparative Ability of Stimulator-Type, Responder-Type,
and Third-Party-Type LAK Cells To Prevent LMC-Induced Generation
of Allo-CTL
LAK Cells Added To MLC
(LAK cel1:responder cell = 1.2)
Relative Cytotoxicity
(mean
?
SD)
1 4 ? 16
56 5 26
46 ? 26
Stimulator type
Responder type
Third-party type
P
Value”
.0001
.0091
.0263
“One sample t-test comparing relative cytotoxicity with added LAK
cells to 100% control value.
tion of allo-CTL. On average, all three types of LAK cells
significantly inhibited generation of allo-CTL when added
at 1 LAK cell:2 MLC responder cells. However, addition to
MLC of stimulator-type LAK cells was significantly more
inhibitory than addition of similar numbers of either
responder-type or third-party LAK cells ( P = .008).
Veto-ype L A K cell inhibition a$ects an early event in
MLC-generation ojallo-CTL. As shown in Fig 4, although
stimulator-type LAK cells specifically inhibited generation
of allo-CTL against homologous targets when added to
MLC on day 0 or day 2 ( P = .03), they did not inhibit when
added on day 5, 1 day before harvesting and testing the
MLC for ah-cytotoxicity. These results indicate that LAK
cell veto-type inhibition affects an early event in the
MLC-induced generation of allo-CTL. They also provide
evidence against the possibility that cold-target inhibition
accounts for the observed veto-type inhibition by stimulatortype LAK cells added on days 0 through 2 to MLC.
Radiation sensitivity of LAK suppressor activity. As shown
in Fig 5, the ability of both stimulator-type and respondertype LAK cells to inhibit MLC-generated CTL was eliminated by exposure to 30-Gy irradiation, a dose that failed to
affect LAK cytolytic activity against Daudi target cells (data
not shown).
2oo
100
1
-
0 .
Day 0
Day 2
Day 5
Day LAK Cells Added To MLC
Fig 4. Human LAK cells inhibit an early event in MLC-generation of
allo-CTL. MLCs consisting of responder ”A‘ and stimulator ”B“ were
performed in the presence or absence of LAK cells from ”B” (LAK cell:
MLC responder cell = 1 :2)added on days 0 , 2 , or 5 of culture. Cultures
were harvested on day 6 and tested for specific cytotoxicity against
stimulator-type PHA blasts. Control MLC generation of anti-B CTL
was 4.3 LU/ IO’ cells.
100 -
80
-
60
-
40
-
20
-
0 GY
30 Gy
Dose of Irradiation
Fig 5. Radiation sensitivity of LAK suppressor activity. Irradiated
(30 Gy) or unirradiated LAK cells from responder “ A and stimulator
”B” were added on day 0 to and ABx MLC (LAK cell: MLC responder
cell = 1:l). Cultures were harvested on day 6 and tested for specific
cytotoxicity against stimulator-type PHA blasts. Control MLC generation of anti-B CTL was 4.7 LU/ 10’ cells. LAK cells added to ABx MLC:
(W), A LAK; (B),B LAK.
IL-2-stimulated PBMC that mediate veto and nutural
suppression express the CD3- CD56’ L A K cell phenotype.
Results of flow cytometric analysis of cell surface markers
showed, as expected, that most of the cells in 4-day LAK
cell cultures expressed the CD3 ’ CD56- T-cell phenotype
(range = 60% to 75%), a smaller population expressed the
CD3 CD56’ NK cell phenotype (range = 10% to 25%),
and a very small population expressed the CD3’ CD56’
phenotype of “non-MHC-restricted cytotoxic T cells”
(range = 0% to 5%). The results of analysis o f a representative LAK cell preparation are shown in Fig 6.
To determine the phenotype of the cells responsible for
the veto and natural suppressor activities of human LAK
cell cultures we tested them for LAK cytotoxicity and
suppressor activity after complement-lysis depletion of
either CD3’ or CD56’ cells. To avoid artificial enrichment
of antibody-nonreactive cells, n o correction was made for
cell losses after antibody treatments. The results of a
representative experiment testing the effect of depleting
LAK cell suspensions of CD56’ cells are shown in Fig 7. In
keeping with the findings reported previously that 10% to
25% of cells in a 4-day LAK cell culture are CDSh’, LAK
cell suspensions “A” and “B” treated with anti-CD56 + C
contained 18% and 13% fewer cells, respectively, than
identical LAK cell suspensions treated with C alone. As
expected from the fact that LAK activity is mediated by
CD56’ cells,* the CDS6-depleted LAK cell suspensions
“A” and “B” exhibited 74% and 91% less LAK activity,
respectively ( P = .02), than control LAK cells (Fig 7A).
The effect of antibody-depletion on the veto and natural
suppressor activities of these cells is shown in Fig 7B.
Control “A” and “B” LAK cells treated with C alone both
inhibited ABx MLC-induced anti-B allo-CTL, but “B”
LAK cells, as expected, inhibited more efficiently than “A”
LAK cells. In parallel with the reduction in their LAK
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
HUMAN LAK CELL VETO AND NATURAL SUPPRESSION
f
- .-
“”i’
-
~
’ . .----
- -
265
”“7
cell-mediated veto-type inhibition we applied a previously
described model for analysis of cell-mediated cytotoxicity
that is based on the classical analysis of enzyme kinetics.’ In
this model timed cytotoxicity assays are performed using
fixed concentrations of effector cells and varying concentrations of labeled target cells. This analysis requires that a
maximal rate of cytotoxicity, Vmax, is achieved by using
increasing numbers of labeled target cells that eventually
saturate the relatively small number of cytotoxic lymphocytes. By analogy to classical studies of enzyme inhibitors,
this approach yields information regarding the mechanism
of LAK cell inhibition of allo-cytotoxicity. Thus, in simple
terms, if addition of LAK cells to MLC causes cold-target
1
FITC=CD3
Fig 6. Flow cytometric analysis of CD3/CD56 phenotype of LAK
cells. LAK cells generated by culturing PBMC for 4 days with 1,500
U/mL recombinant IL-2were stained with FITC-CD3and PE-CD56 and
analyzed by two-color immunofluorescence with an EPICS C flow
cytometer. Results of a typical LAK cell preparation are shown
demonstrating 71% CD3’ CD56-, 16% CD3- CD56’. and 4% CD3‘
CD56’.
activity, the nonspecific and specific inhibitory activities of
“A” and “B’ LAK cells were reduced by 100% and 80%,
respectively, after depletion of CD56’ cells.
In contrast to the effects of removing CD56’ cells from
LAK cell suspensions, depletion of CD3’ cells resulted in a
greater loss in cell number, only a minimal loss of LAK
activity, and an enhancement of both specific and nonspecific inhibitory activity when added to MLCs. In the
experiment shown in Fig 8, following treatment with antiCD3 C “B” LAK cells showed a 64% reduction in cell
yield compared to the control “B” LAK cell suspension
treated with C alone. However, they showed only an
insignificant loss of LAK activity (P = .09) compared to
control LAK cells treated with C alone (Fig 8A). Moreover,
when tested for suppressor activity (Fig 8B), the CD3depleted LAK cells showed a somewhat enhanced vetotype inhibition of generation of allo-CTL by “A” responders against “B” targets, and markedly increased nonspecific
inhibition of generation of allo-CTL by “B” responders
against “A” targets. We have observed (unpublished data,
1991) that anti-CD3 by itself can inhibit MLC-induced
generation of allo-CTL at a concentration of 1 ng/mL but
not at a concentration 100 pg/mL. Therefore, it is important to note that to avoid any possible contribution of
anti-CD3 to inhibition of generation of allo-CTL, care was
taken in this experiment to wash the LAK cells sufficiently
to reduce any anti-CD3 carried over with the treated LAK
cells into the MLC to less than 2 pg/mL.
LAK cells cause noncompetitive rather than competitive
inhibition of allocytotoxicity generated in MLC. To further
determine whether cold-target inhibition accounts for LAK
+
Medium + C
170
.z
.-
100
0
c
80
0
60
0
Anti-CD56
+C
I
X
QI
>
-
40
0
g
n
tJ
20
0
Medium+C
Anti CD56+C
LAK Cell Treatment
Fig 7. IL-2-stimulated PBMC which mediate veto and natural
suppression express the CD56’ LAK cell phenotype. LAK cells from
subjects A and B were treated for 60 minutes with antLCD56 or
medium alone, washed, and then treated with complement for 60
minutes. The cells were resuspended t o the same volume, and were
tested (A) for cytotoxicity against Daudi targets and (B) for specific
and nonspecific suppression of generation of allo-CTL as described in
legend of Fig 2. Compared t o control “ A and ” B LAK cells treated
C
with C alone, LAK cells “ A ’ and “ B treated with antLCD56
contained 18% and 13% fewer cells, respectively. Allocytotoxicity
generated in control MLC without added LAK cells was 13.4 LU/107
cells. (A) (m), A LAK; (@), B LAK. (B) LAK cells added t o ABx MLC: (W),
A LAK; (El), B LAK.
+
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
266
UBERTI ET AL
30
.>2
.-
1
20-
4-
2
Y
4
10
A
n
-
0 .
Medium + C
Anti-CDS
+C
targets that was similar to that achieved in their absence.
This shows that cold-target inhibition can in fact be
overcome by saturating levels of homologous labeled targets. By contrast, effector cells derived from the MLCs to
which “B” LAK cells were added at culture initiation
exhibited impaired velocities of cytotoxicity even when
twice the saturating number of labeled targets were added.
As shown another way in Fig 9B, the fixed number of
unlabeled targets inhibited the velocity of cytotoxicity by
roughly 60% at the lowest number of labeled homologous
targets, but as the labeled target cells increased to the
saturating numbers equal to 1.5 times the number of
unlabeled target cells, this inhibitory effect was completely
overcome. By contrast, the inhibition of cytotoxicity medi-
31
A
U
Medium + C
Anti-CDS
7
8
9
10
11
12
13
Hot B Targetdwell (natural log)
+C
LAK Cell Treatment
Fig 8. IL-2-stimulated PBMC that mediate veto and natural suppression are CD3-. LAK cells from subject B were treated for 60
minutes with anti-CD3 or medium alone, washed, and then treated
with complement for 60 minutes. After washing three times with 10
mL medium, the cells were resuspended t o the same volume and
without correction for cell losses were then tested (A) for cytotoxicity
against Daudi targets and (B) for specific and nonspecific suppression
of generation of allo-CTL as described in legend of Fig 2. Compared t o
control LAK cells treated with C alone, LAK cells treated with anti-CD3
t C contained 64% fewer cells. Allo-cytotoxicity generated in control
ABx and BAx MLCs without added LAK cells were 12.7 and 17.8
L U I lo’ cells, respectively. (B) MLC:
ABx; (a),BAx.
1
(m),
(ie, competitive) inhibition in the assay for allo-CTL, this
should be overcome at saturating levels of labeled targets
(analogous to saturating levels of enzyme substrate). By
contrast, if addition of LAK cells reduces the number or
activity of allo-CTL generated in an MLC (ie, causes
noncompetitive inhibition), addition of saturating levels of
labeled targets should not overcome this inhibition.
As shown in Fig 9A, using a fixed number of 5 X lo4
control AEix MLC effector cells per well, estimated Vmax
was achieved at roughly 1.5 x lo5 labeled “B” target
cells/well. At relatively low levels of labeled targets, addition of 5 x lo4 unlabelled “B” targets to the CTL assay
inhibited the velocity of cytotoxicity. However, as the
numbers of labeled “B” targets increased an estimated
Vmax was achieved in the presence of unlabeled “B”
6
7
8
9
10
11
12
13
Hot B Targetshell (natural log)
Fig 9. Responder-type LAK cells noncompetitively rather than
competitively inhibit generation of allo-CTL when added t o MLC. (A)
A maximal velocity of cytotoxicity is achieved by control ABx MLC
effectors at saturating concentrations of labeled targets, a condition
that satisfies the minimum requirements for this type of analysis.
Both (A) and (B) show that the inhibitory effects of added unlabeled B
targets t o these same ABx MLC effectors are overcome at saturating
concentrations of labeled targets, whereas saturating concentrations
of labeled targets fail t o overcome the diminished allo-cytotoxicity
generated in ABx B LAK MLC. (A) Effector cells: (-W),ABx alone;
(--e-), ABx cold B; (-m-),
ABx t B LAK. (B) Effector cells:
(-m-),
ABx cold B; (-+-),
ABx + B LAK.
+
+
+
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
HUMAN LAK CELL VETO AND NATURAL SUPPRESSION
ated by addition of LAK cells at onset of MLC was not
overcome even when the numbers of labeled targets were
equivalent to 48 times the numbers of added LAK cells.
DISCUSSION
The findings presented here indicate that human LAK
cells, like murine LAK cells, exhibit both veto and natural
suppressor activities. The presence in MLCs of low concentrations of LAK cells selectively inhibited the generation of
allo-CTL against homologous targets but did not inhibit
generation of allo-CTL against unrelated allogeneic targets. This specific inhibition of reactivity against "self" but
not "non-self' antigens is characteristic of veto. Higher
concentrations of LAK cells in MLCs caused nonspecific
inhibition of generation of allo-CTL against allogeneic
targets from donors unrelated to the LAK cell donor, albeit
not to the same extent as the inhibition observed for the
generation of allo-CTL against targets homologous to the
LAK cells. This nonspecific inhibition of alloreactivity is
characteristic of natural suppression.
Veto-type inhibition mediated by human LAK cells is not
an artifact of cold-target inhibition mediated by added
LAK cells themselves, or by T cells contaminating stimulator-type LAK cells that might proliferate in response to the
alloantigen stimulation by the responder lymphocytes in the
MLC preparations. Rather, it appears to be an active
process mediated by proliferating LAK cells that inhibit a
relatively early event in the generation of allospecific CTL.
This conclusion derives from the following lines of evidence: (1) Specific veto-type inhibition occurred when
stimulator-type LAK cells were added to MLC on days 0
through 2 of culture but not when added on day 5, one day
before testing the MLC for cytotoxicity against stimulatortype target cells. If the stimulator-type LAK cells themselves cause cold-target inhibition one would expect such
cold-target inhibition to occur when LAK cells are added at
day 5 as well as when added at days 0 through 2. (2) Kinetic
analysis showed that LAK cells added at onset of MLC
noncompetitively rather than competitively inhibited allocytotoxicity generated in the MLC. Addition of saturating
numbers of labeled targets to fixed numbers of MLC
effectors overcame inhibition by unlabeled homologous
targets added at the time of the cytotoxicity assay but failed
to overcome the inhibition mediated by LAK cells added at
onset of the MLC. (3) Anti-CD3 + C treated LAK cells
caused even more veto-type inhibition than a threefold
greater number of control LAK cells treated with C alone.
By considerations of cell dose alone this latter finding
provides strong evidence against cold target inhibition in
general. In addition, it specifically eliminates the possibility
of cold-target inhibition mediated by any T cells in the LAK
cell preparation that might proliferate in response to the
MLC "responder" cells.
Although the majority of cells present in 4-day cultures of
IL-2-stimulated peripheral blood lymphocytes are CD3'
CD56- T cells, as shown here and elsewhere*,*the LAK
cytolytic activity of such cultures is mediated by a minor
fraction of cells that express the CD3- CD.56' NK marker
phenotype. This same NK-type LAK cell appears to medi-
267
ate both the veto and natural suppressor activities of
IL-2-stimulated human lymphocytes. Both of these LAK
cell suppressor activities were nearly completely eliminated
after complement lysis depletion of cells expressing CD.56.
By contrast, as described previously, both activities increased after complement lysis depletion of cells expressing
CD3. In addition to showing that human suppressor LAK
cells are CD3-, these latter findings suggest that LAK cell
veto and natural suppressor activities may be downregulated by CD3' T cells and/or upregulated by lymphokines
rapidly released by anti-CD3-activated CD3' T cells. Both
of these possibilities are currently under investigation.
Cloned murine LAK cells can mediate both veto and
natural suppression, suggesting that a single LAK cell is
capable of both suppressor activities.' Whether the same is
true of human LAK cells remains to be determined. If a
single LAK cell does in fact mediate both activities the
different dose requirements for expression by LAK cells of
each suppressor activity could be due to differences in the
mechanisms involved in veto and natural suppression. Veto
requires cell-cell contact whereas natural suppression involves release of soluble suppressor factor^.^-'^ Thus, on a
cell-for-cell basis veto is probably a more efficient mechanism for inhibition of alloreactivity because the target cells
for veto-type inhibition specifically bind to the suppressor
cells, which then directly inactivate them. By contrast, the
targets of natural suppression are inactivated by suppressor
cell-derived factors released at a distance.
We and others have shown that murine LAK cells not
only inhibit alloreactions in vitro but also inhibit bone
marrow graft rejection and lethal GVHD in vivo in class I
and class I1 MHC mismatched transplants.'.'.'' Therefore,
the present demonstration that human LAK cells show the
same ability to inhibit in vitro alloreactions as murine LAK
cells strongly suggests that LAK cell adoptive immunotherapy might be useful in preventing marrow graft rejection
and GVHD in human bone marrow transplant recipients.
Approaches similar to those that have been used to generate large numbers of human LAK cells ex vivo for use in
adoptive immunotherapy trials in patients with cancer"
could readily be used to generate the large number of LAK
cells likely to be needed for adoptive immunotherapy in
BMT recipients.
LAK cell adoptive immunotherapy would provide a
number of potential advantages over donor marrow T-cell
depletion as an approach to the prevention of GVHD in
recipients of allogeneic marrow. Although BMT with T-celldepleted allogeneic marrow has resulted in a decreased
incidence of GVHD it has also been associated with an
increased incidence of graft rejection,'*-'" and an increased
rate of leukemic relapse as a result of impairment of a
graft-versus-leukemia (GVL) effect.2'.'2By contrast, animal
studies indicate that adoptively transferred LAK cells
enhance both short-term' and long-term engraftment" of
allogeneic marrow. Although the effect of LAK cell adoptive immunotherapy on GVL remains to be determined, the
fact that LAK cells are able to kill both autologous and
allogeneic leukemic blastsz3implies that adoptively transferred LAK cells might themselves provide a direct antitumor effect.
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
UBERTI ET AL
268
REFERENCES
1. Azuma E, Kaplan J Role of lymphokine activated killer cells
as mediators of veto and natural suppression. J Immunol141:2601,
1988
2. Herberman RB: Lymphokine-activated killer cell activity.
Immunol Today 8:178,1987
3. Miller RG: An immunologic suppressor cell inactivating
cytotoxic T lymphocyte precursor cells recognizing it. Nature
287:544,1980
4. Oseroff AS, Okada S, Strober S: Natural suppressor (NS)
cells in the spleen of neonatal mice and adult mice given total
lymphoid irradiation (TLI) express the null surface phenotype. J
Immunol132:101,1984
5. Azuma E, Yamamoto H, Kaplan J: Use of lymphokine
activated killer cells to prevent bone marrow graft rejection and
lethal graft-versus-host disease. J Immunol 143:1524,1989
6. Pross HF, Baines MG, Rubin P, Shragge, Patterson MS:
Spontaneous human lymphocyte-mediated cytotoxicity against tumor target cells. IX. The quantitation of natural killer cell activity.
J Clin Immunol 1:51, 1981
7. Callewaert DM, Mahle NH, Dayner S, Earnshaw B, Bowman
S, O'Connor M: Kinetic analysis of the specificityof human natural
cytotoxicity. Scand J Immunol18:429,1983
8. Philips JH, Lanier L L Dissection of the lymphokine activated
killer phenomenon. Relative contribution of peripheral blood NK
cells and T lymphocytes to cytolysis. J Exp Med 164:814,1986
9. Muraoka S, Miller RG: Cells in bone marrow and in T cell
colonies grown from bone marrow can suppress generation of
cytotoxic T lymphocytes directed against their self antigens. J Exp
Med 15254,1980
10. Fink PJ, Rammensee H-G, Benedetto JD, Staerz UD,
Lefrancois L, Bevan MJ: Studies on the mechanism of suppression
of primary cytotoxic responses by cloned cytotoxic T lymphocytes. J
Immunol133:1769,1984
11. Argyris B F Suppressor factor produced by neonatal mouse
spleen cells. Cell Immunol62:412, 1981
12. Jadus MR, Peck AB: Naturally occurring spleen-associated
suppressor activity of the newborn mouse. Biochemical and functional identification of three monokines secreted by newborn
suppressor-inducer monocytes. Scand J Immunol23:35,1986
13. Knaan-Shanzer S, Van Bekkum D W Soluble factors secreted by naturally occurring suppressor cells that interfere with in
vivo graft-vs-host disease and with T cell responsiveness in vitro.
Eur J Immunol17:82,1987
14. Jadus MR, Parkman R: The selective growth of murine
newborn-derived suppressor cells and their probable mode of
action. J Immunol 136:783,1986
15. Argyris B F Further studies on suppressor-activity in the
spleen of neonatal mice. Cell Immunol48:398, 1979
16. Nakamura H, Gress RE: Interleukin 2 enhancement of veto
suppressor cell function in T cell-depleted bone marrow in vitro
and in vivo. Transplantation 49:931,1990
17. Rosenberg SA, Lotze MT, Muul LM, et al: Observations on
the systemic administration of autologous lymphokine-activated
killer cells and recombinant interleukin-2 to patients with metastatic cancer. N Engl J Med 313:1485, 1986
18. OReilly RJ, Collins NH, Kernan N, Brochstein J, Dinsmore
D, Kirkpatrick D, Siena S, Keever C, Jordan B, Shank B, Wolf L,
Dupont B, Reisner Y: Transplantation of marrow depleted of T
cells by soybean lectin agglutination and E-rosette depletion:
Major histocompatibility complex-related graft resistance in leukemic transplant recipients. Transplant Proc 17:455,1985
19. Martin PJ, Hansen JA, Buckner D, Sanders JE, Deeg HJ,
Stewart P, Applebaum FR, Lift R, Fefer A, Witherspoon RP,
Kennedy MS, Sullivan KM, Flournoy N, Storb R, Thomas ED:
Effects of in vitro depletion of T cells in HLA-identical allogeneic
marrow grafts. Blood 66:665, 1985
20. Vriesendorp HM, Klapwijk WM, Heidt DJ, Hogeweg B,
Zurcher C, van Bekkum DW: Factors controlling the engraftment
of transplanted dog bone marrow cells. Tissue Antigens 20:63,1982
21. Goldman JM, Gale RP, Horowitz MM, Biggs JC, Champlin
RE, Gluckman E, Hoffmann RG, Jacobsen SJ, Marmont AM,
McGlave PB, Messner HA, Rimm AA, Rozman C, Speck B, Tura
S, Weiner RS, Bortin MM: Bone marrow transplantation for
chronic myelogenous leukemia in chronic phase: Increased risk for
relapse associated with T-cell depletion. Ann Intern Med 108:806,
1988
22. Horowitz MM, Gale RP, Sonde1 PM, Goldman JM, Kersey
J, Kolb H-J, Rimm AA, Ringden 0, Rozman C, Speck B, Truitt R,
Zwaan FE, Bortin MM: Graft-versus-leukemia reactions after
bone marrow transplantation. Blood 75:555,1990
23. van den Brink MRM, Voogt PJ, Marijt WAF, van Luxemburg-Heys SAP, van rood JJ, Brand A Lymphokine-activated
killer cells selectively kill tumor cells in bone marrow without
compromising bone marrow stem cell function in vitro. Blood
74:354,1989
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
1992 79: 261-268
Human lymphokine activated killer (LAK) cells suppress generation of
allospecific cytotoxic T cells: implications for use of LAK cells to
prevent graft-versus-host disease in allogeneic bone marrow
transplantation
J Uberti, F Martilotti, TH Chou and J Kaplan
Updated information and services can be found at:
http://www.bloodjournal.org/content/79/1/261.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.