selective expansion of human natural killer cells leads to enhanced alloreactivity
TRANSCRIPT
RESEARCH ARTICLE
Selective expansion of human natural killer cells leads toenhanced alloreactivity
Diana N Eissens1,3, Clive M Michelo1,3, Frank WMB Preijers2, Bram van Cranenbroek1, Kjeld vanHouwelingen1, Arnold van der Meer1 and Irma Joosten1
In allogeneic stem cell transplantation (SCT), natural killer (NK) cells lacking their cognate inhibitory ligand can induce
graft-versus-leukemia responses, without the induction of severe graft-versus-host disease (GVHD). This feature can be
exploited for cellular immunotherapy. In this study, we examined selective expansion of NK cell subsets expressing
distinct killer immunoglobulin-like receptors (KIRs) within the whole human peripheral blood NK cell population, in the
presence of HLA-Cw3 (C1) or Cw4 (C2) transfected K562 stimulator cells. Coculture of KIR1 NK cells with C1 or C2
positive K562 cells, in the presence of IL-21IL-15, triggered the outgrowth of NK cells that missed their cognate ligand.
This resulted in an increased frequency of alloreactive KIR1 NK cells within the whole NK cell population. Also, after
preculture with K562 cells lacking their cognate ligand, we observed that this alloreactive NK population revealed higher
numbers of CD1071 cells when cocultured with the relevant K562 HLA-C transfected target cells, as compared to
coculture with untransfected K562 cells. This enhanced reactivity was confirmed using primary leukemic cells as target.
This study demonstrates that HLA class I expression can mediate the skewing of the NK cell repertoire and enrich the
population for cells with enhanced alloreactivity towards leukemic target cells. This feature may support future clinical
applications of NK cell-based immunotherapy.
Cellular & Molecular Immunology (2014) 11, 160–168; doi:10.1038/cmi.2013.56; publishedonline 18 November2013
Keywords: alloreactivity; cytotoxicity; immunotherapy; KIR; NK cells; stem cell transplantation
INTRODUCTION
Reducing the severity of graft-versus-host disease (GVHD),
while boosting graft-versus-leukemia responses after hemato-
poietic stem cell transplantation (SCT) is one of the major
challenges within the field of transplantation today.1–4 In this
respect, natural killer (NK) cells are of interest for use in immu-
notherapeutic strategies, as they form the first line of defense in
mediating immunity against microbial pathogens, and are effi-
cient effectors in eradicating tumor cells without inducing
severe GVHD.5–8
NK cells survey potential target cells for the absence or loss of
expression of human leucocyte antigen (HLA) class I classical
(HLA-A,B,C) and non-classical (E,G) molecules through killer
cell immunoglobulin-like receptors (KIRs) and the inhibitory
CD94/NKG2A receptor complex.9,10 Overall, HLA class I
expression is surveyed by expression of the lectin-like receptor
CD94/NKG2A through its recognition of the ubiquitously
expressed HLA-E molecule.11 The expression of KIR allows
for a more subtle surveillance, as these receptors recognize
specific epitopes present on HLA-A, -B, or -C molecules. The
receptors KIR2DL1 and KIR2DL2/3 have as their ligands the
HLA-C2 group alleles and HLA-C1 group alleles, respectively.
Furthermore, HLA-A3 and -A11 are recognized by KIR3DL2,
while HLA-Bw4 is recognized by KIR3DL1. HLA-A3 and -A11
are recognized by KIR3DL2, while HLA-Bw4 is recognized by
KIR3DL1.12
In allogeneic SCT, anti-leukemic NK cell alloreactivity can
be facilitated by allowing mismatches for specific KIR ligands,
i.e., HLA-B and/or HLA-C, between donor and recipient. The
introduction of certain HLA mismatches has been shown to
induce NK cell-mediated graft-versus-leukemia responses,
without inducing severe GVHD, and to contribute to decreased
3 These authors contributed equally to this work.
1Department of Laboratory Medicine—Laboratory of Medical Immunology, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands and2Department of Laboratory Medicine—Laboratory of Hematology, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands
Correspondence: Professor I Joosten, Radboud University Nijmegen Medical Centre, Department of Laboratory Medicine—Laboratory of MedicalImmunology, PO box 9101, 6500 HB Nijmegen, The Netherlands.E-mail: [email protected]: 4 February 2013; Revised: 7 October 2013; Accepted: 8 October 2013
Cellular & Molecular Immunology (2014) 11, 160–168� 2014 CSI and USTC. All rights reserved 1672-7681/14 $32.00
www.nature.com/cmi
relapse, better engraftment and improved overall survival.13–15
Other studies show that NK cell alloreactivity can also be trig-
gered by the presence of an inhibitory KIR in the donor’s geno-
type in the absence of the corresponding KIR ligand in the
recipient’s HLA repertoire.16–18 In SCT, the KIR repertoire of
NK cells was shown to be important for the induction of allor-
eactive NK cell responses.19–22.
In this study, we used an in vitro culture system to investigate
whether we could induce the selective expansion of human
peripheral blood derived NK cells with enhanced alloreactivity.
We cocultured human mature NK cells together with K562
cells transfected with either a single HLA-C1 or C2 gene, and
in the presence of cytokines Results demonstrate that the
absence of a specific KIR ligand (HLA-C1 or C2) on the
K562 cells favored the outgrowth of NK cells expressing the
KIR that lacks their cognate KIR ligand, resulting in an increase
of alloreactive NK cells, exhibiting improved cytolytic ability.
These results may facilitate future clinical applications of NK
cell-based immunotherapy, especially within the field of trans-
plantation
MATERIALS AND METHODS
Cell isolation and genotyping
Buffy coats from healthy human donors were purchased from
Sanquin Blood Bank, Nijmegen, The Netherlands, upon writ-
ten informed consent with regard to scientific use according to
Dutch law. Peripheral blood mononuclear cells were isolated
by density gradient centrifugation (Lymphoprep; Nycomed
Pharma, Roskilde, Denmark). NK cells were negatively selected
(Miltenyi Biotec, Bergisch Gladbach, Germany), resulting in a
purity of more than 95%. Parallel to the experiments, HLA-C
and KIR genotyping were performed on small samples from all
buffy coats (Table 1). Genomic DNA was isolated using
the QIAamp DNA blood mini kit (Qiagen, Venlo, The
Netherlands). A polymerase chain reaction–sequence-specific
primers typing protocol as described by Gagne et al.23 (with
minor modifications) was used for KIR typing, with the exclu-
sion of pseudo genes (KIR2DP and KIR3DP). A limited HLA-C
locus typing was performed as previously described,24 using a
polymerase chain reaction–sequence-specific primer protocol
solely aimed at discriminating between the polymorphisms
located at exon 2 encoded amino acid positions 77 and 80,
relevant for KIR recognition. The HLA-C alleles were thus
divided into two groups: HLA-C1 (HLA-Cw [Ser77Asn80] alleles
with serine at position 77 and asparagine at position 80), and
HLA-C2 (HLA-Cw [Asn77Lys80] alleles with asparagine at posi-
tion 77 and lysine at position 80). Data analysis was focused on
NK cell donors having KIR2DL1, 2 and 3 in their KIR repertoire.
Cell lines
Single cell-derived clonal K562 cell lines expressing HLA-C*03:01
(K562-C1posC2neg) and HLA-C*04:03 (K562-C1negC2pos) were
used as stimulatory cells during culture and as target cells in
functional assays. Briefly, HLA class I-deficient K562 cells were
transfected with full-length HLA-C*03:01 cDNA inserted into
the EcoRI and HindIII sites of expression vector pcDNA3.1(2)
(Invitrogen, Paisley, UK) and HLA-C*04:03 cDNA ligated into
the pEF6/V5-His expression vector (pEF6/V5-His TOPO TA
Expression Kit; Invitrogen). The cell lines K562-C*03:01 and
K562-C*04:03 showed stable and strong HLA class I expression
(.95%) analyzed by flow cytometry using the HLA class I
antibody W6/32 (Sigma, Steinheim, Germany). The expression
of HLA-C*03:01 and HLA-C*04:03 was confirmed by mRNA
profiling (sequencing of HLA-Cw cDNA). As a control, HLA
class I-negative K562 cells transfected with empty vector
pcDNA3.1 were used (K562 Cneg). For surface protein ana-
lysis, we used the monoclonal antibody (mAb) DT9 (a gift
from V. Braud), which has been reported to recognize HLA-
E and HLA-C but not HLA-A or HLA-B allotypes, MEM-E/06
(Monosam, Uden, The Netherlands), specific for HLA-E, and
Table 1 KIR genotype and HLA-C type of NK cell donors
KIR genotype for KIR2DL/S1/2/3a
HLA-Ca
Donorb
2DL1 2DL2 2DL3 2DS1 2DS2 2DS3 C1 C2
1 1 0 1 1 0 1 1 1
2 1 1 1 1 1 1 1
3 1 1 1 1 1 1 1 1
4 1 1 1 0 1 1 1 1
5 1 1 1 0 1 1 1 1
6 1 1 1 1 1 1 2 0
7 1 1 1 1 1 1 1 1
8 1 0 1 0 0 1 1 1
9 1 1 1 1 1 1 2 0
10 1 1 0 1 1 1 1 1
11 1 1 1 1 1 0 2 0
12 1 1 0 1 1 1 1 1
13 1 1 1 1 1 1 2 0
14 1 1 1 0 1 ND 1 1
a Coding for KIR2DL/S1/2/3 genotype and HLA-C type; 15present, 05not present, ND5not able to determine.b The following donors were selected for detailed analysis based on the presence of KIR2DL1/2/3; 2, 3, 4, 5, 6, 7, 9, 11, 13 and 14.
Skewing NK cell alloreactivity
DN Eissens et al
161
Cellular & Molecular Immunology
W6/32 (Sigma), which recognizes HLA class 1, HLA-A, -B, and
-C allotypes.
In vitro culture system
Freshly isolated 13106 NK cells were cocultured with irra-
diated K562-C1posC2neg, K562-C1negC2pos and K562-Cneg sti-
mulatory cells in a 1 : 1 ratio, in the presence of low dose rhIL-2
(10 U/ml; Chiron, Amsterdam, The Netherlands) and rhIL-15
(1 ng/ml; BioSource International, Camarillo, CA, USA) in
culture medium consisting of RPMI 1640 medium supplemen-
ted with pyruvate (0.02 mM), glutamax (2 mM), penicillin
(100 U/ml), streptomycin (100 mg/ml) and 10% human pooled
serum, in a 37 uC, 95% humidity, 5% CO2 incubator.
Carboxyfluorescein succinimidyl ester (CFSE)-based
division analysis
Cell division was studied by CFSE dilution patterns. Freshly
isolated NK cells were labeled with 0.1 mM CFSE (Molecular
Probe, Eugene, OR, USA), aliquoted in CFSE labeling buffer
(phosphate-buffered saline containing 0.02% human pooled
serum), for 10 min at room temperature in the dark. The reac-
tion was stopped by addition of equal volumes of cold human
pooled serum. Subsequently, cells were washed three times
with CFSE labeling buffer and resuspended in culture medium.
CFSE-labeled NK cells were cultured as described above and
were analyzed using flow cytometry.
Flow cytometry
Non-CFSE labeled NK cells were phenotypically analyzed on
the FC500 (Beckman Coulter, Miami, FL, USA) using the fol-
lowing conjugated mAbs: NKAT2-FITC (KIR2DL2/DL3/DS2)
and NKB1-PE (KIR3DL1) from BD Bioscience (Breda, The
Netherlands) and CD158a, h-PE (KIR2DL1/S1), CD3-PC5,
and CD56-PC7 from Beckman Coulter (Woerden, The
Netherlands). CFSE-labeled NK cells were analyzed on the
Gallios using the following conjugated mAbs: CD16-FITC
HLA-C neg environment
HLA-C environment
30**** *
25
20
15
10
0Cneg C1 C2
HLA-C environment
Cneg C1 C2
5
HLA-C1 environment
Single KIR2DL1/S1
6.99% 8.15% 5.74%
CD
56
HLA-C2 environmenta
b c
Sin
gle
KIR
2DL/
S1+
cel
ls (%
)
30
25
20
15
10
0
5S
ingl
e K
IR2D
L/S
2/3+
cel
ls (%
)
Figure 1 Distribution of KIR2DL/S1 and KIR2DL/S2/3 NK cells at day 7 of culture in a nonspecific (HLA-Cneg), HLA-C1posC2neg and HLA-C1negC2pos environment. Freshly isolated NK cells, from donors 6, 7, 9, 11, 13 and 14, were cultured in the presence of irradiated HLA-Cneg,HLA-C1posC2neg and HLA-C1negC2pos K562 stimulatory cells and their phenotype was analyzed at day 7 of culture. (a) Representative dot plots forthe cumulative data shown in new b. The open plots (#) in b/c are for heterozygous (C1C2) donors and the closed plots (N) are for donors lackingC2 (C1C1). (b) The percentages of single KIR2DL/S11 cells within the total NK cell population after nonspecific (HLA-Cneg), HLA-C1posC2neg andHLA-C1negC2pos stimulation at day 7 of culture. Differences between the different stimulation settings were analyzed using repeated measuresANOVA analysis; *P,0.05, ***P,0.001. (c) The percentages of single KIR2DL/S2/31 cells within the total NK cell population after nonspecific(HLA-Cneg), HLA-C1posC2neg and HLA-C1negC2pos stimulation at day 7 of culture. Differences between the different stimulation settings wereanalyzed using repeated measures ANOVA analysis; *P,0.05. HLA, human leucocyte antigen; KIR, immunoglobulin-like receptor; NK, naturalkiller.
Skewing NK cell alloreactivity
DN Eissens et al
162
Cellular & Molecular Immunology
(Dako, Heverlee, Belgium), CD3-ECD, CD56-APC-A750,
CD158b1/b2, j-PC7 (KIR2DL2/DL3/DS2), CD158e1/e2-APC
(KIR3DL1/S1), CD158a, h-APC-A700 (KIR2DL1/S1) and
CD45-PO (all provided by Beckman Coulter). For 10-color
flow cytometry, fluorochrome combinations were balanced
to avoid antibody interactions, sterical hindrance and to detect
also dimly expressing populations. Before 10-color analyses
were performed, all conjugates were titrated and individually
tested for sensitivity, resolution and compensation of spectral
overlap. Isotype controls were used to define marker settings.
Live gating was performed based on forwards scatter vs. side
scatter.
Functional analysis
The functional capacity of NK cells was studied by degranula-
tion patterns (CD107a expression) upon target encounter.
CD107a is a functional marker for the identification of natural
killer cell activity that correlates with NK cell degranulation
and target cell lysis.25 NK cells were harvested from cul-
ture and viable NK cells were plated in 96-well V-bottom
plates in culture medium supplemented with fluorochrome-
conjugated anti-CD107a (1 : 200; BD Biosciences, Erembodegem,
Belgium). Subsequently, non-irradiated K562-C1posC2neg, K562-
C1negC2pos, K562 Cneg or Kasumi-1 (HLA-C1 homozygous,
acute myeloid leukemia cell line) target cells were added at an
E/T ratio of 1 : 2 and incubated at 37 uC, 95% humidity, 5% CO2.
After 4 h, the degranulation of NK cells was phenotypically ana-
lyzed by flow cytometry.
Statistical analysis
Statistical analysis was performed using GraphPad Prism 5.0.
Results of different conditions were compared using paired
Student’s t-tests or repeated measures ANOVA analysis with
the Tukey multiple comparison test for post-testing. P values
f0.05 were considered statistically significant.
RESULTS
To study the phenotypical changes and cytolytic response of
mature peripheral NK cells when cocultured in the presence of
irradiated stimulatory cells lacking a specific KIR ligand, we set
up an in vitro culture system using single HLA-C (either C1 or
C2) transfected K562 cells. To this end, freshly isolated NK cells
were cultured in the presence of low-dose IL-2 and IL-15
together with irradiated K562 stimulator cells transfected with
HLA-C1 (K562-C1posC2neg), HLA-C2 (K562-C1negC2pos) or
an empty vector (K562 Cneg) in a 1 : 1 ratio. Subsequently,
NK cells were phenotypically and functionally analyzed.
From our mRNA profiling and staining with the DT9 and
W6/32 mAb, we concluded that the transfected cells did not
express any classical HLA class I molecules, other than the
transfected HLA-C1 or HLA-C2 (results not shown); hence,
we confined our analysis to KIR2DL1, KIR2DL2, KIR2DL3,
KIR2DS1, KIR2DS2 and KIR2DS3. MEM-E/06 staining
revealed that HLA-E was expressed on both transfected, as well
as untransfected K562 cells, in all cases to a similar level (not
shown).
NK cell phenotype is skewed in vitro in the absence of
specific KIR ligands
At day 7 of culture, NK cells were phenotypically analyzed for
KIR2DL1/S1 and KIR2DL2/DL3/DS2 expression; we focused
analysis on NK subsets that were KIR2DL1 and/or KIR2DS11,
but KIR2DL2/DL3/DS22 (single KIR2DL/S11) or KIR2DL1/
S12, but KIR2DL21 and/or KIR2DS21 and/or KIR2DL31
(single KIR2DL2/DL3/DS21). As in each donor tested, the
percentage of KIR3DL11 cells never exceeded the 2% (data
not shown), precluding meaningful analysis, we omitted this
marker from further analysis. CD531/CD32 NK cells were
gated from a live gate on a forward-side scatter dot plot.
Single KIR2DL/S11 and single KIR2DL2/DL3/DS21 cells were
subsequently gated on these NK cells. Skewing of the NK cell
phenotype was seen for both KIR subsets (Figure 1). The per-
centage of single KIR2DL/S11 NK cells was significantly higher
in cultures containing HLA-C1posC2neg stimulatory cells, as
compared to cultures containing HLA-C1negC2pos stimulator
cells or HLA-C negative stimulator cells (Figure 1b). Visa versa,
percentages of single KIR2DL2/DL3/DS21 NK cells were ele-
vated in the presence of HLA-C1negC2pos stimulatory cells, as
compared to cultures containing HLA-C1posC2neg stimulatory
cells or, although not in all cases, HLA-C negative stimulator
cells (Figure 1c). Thus, in vitro data suggest that an envir-
onment in which a specific KIR ligand is absent, leads to a
favored NK cell phenotype in which NK cells preferably express
the KIR receptor for which the cognate ligand is lacking.
Coculture with either C1 or C2 transfected K562 cells did not
lead to a difference in the number of cells expressing NKG2A
(results not shown).
Absence of a specific KIR ligand induces oligoclonal division
of specific KIR-positive NK cells
To investigate whether the change in KIR expression pattern,
observed upon coculture with transfected K562 cells that
lacked the cognate KIR ligand, was due to clonal expansion
of specific KIR1 subsets, we performed CFSE-based division
analyses. To this end, CFSE-labeled NK cells were cocultured
with irradiated stimulator cells lacking a specific KIR ligand. At
day 7, the level of cell division of the cultured CFSE-labeled NK
cells was analyzed (Figure 2). In the presence of both HLA-
C1posC2neg and HLA-C1negC2pos stimulator cells, KIR-negative
NK cells were able to divide . However, in the presence of HLA-
C1posC2neg stimulator cells, data from two out of three donors
showed that single KIR2DL/S11 NK cells had a higher division
rate than single KIR2DL2/DL3/DS21 NK cells (Figure 2). The
opposite effect was seen in the presence of HLA-C1negC2pos
stimulatory cells, as single KIR2DL2/DL3/DS21 NK cells
revealed a higher division rate than single KIR2DL/S11 NK
cells (Figure 2b). Thus, KIR1 NK cells that lack their cognate
ligand are more prone to expand in culture, as compared to the
KIR1 NK cells that are able to recognize their proper ligand in
Skewing NK cell alloreactivity
DN Eissens et al
163
Cellular & Molecular Immunology
vitro. This suggests that the lack of a KIR ligand guides the
proliferation of allospecific KIR1 NK cells.
Overall, these results suggest that the introduction of NK
cells into an environment, in which a specific KIR ligand is
absent, triggers the outgrowth of KIR1 NK cells that lack their
cognate KIR ligand, and thus, induces a shift in the overall KIR
expression pattern in the whole NK cell population. This shift
may well have functional consequences, and thus, we hypothe-
sized that the lack of a specific KIR ligand in vitro may lead to an
increase of specific cytolytic KIR1 NK cells within the whole
NK cell population.
A selective outgrowth of KIR expressing NK cell subsets is
associated with increased functional alloreactivity
To investigate whether lack of a specific KIR ligand leads to an
increase of specific cytolytic KIR1 NK cells within the whole
alloreactive (CD1071) NK cell population, HLA-C1posC2neg
and HLA-C1negC2pos stimulated NK cells were analyzed for
their degranulation potential against K562-C1posC2neg and
K562-C1negC2pos target cells at day 7 of culture (Figure 3).
Results showed that stimulation with HLA-C1posC2neg K562
cells led to an increase of cytolytic KIR2DL/S11 NK cells
against K562-C1posC2neg target cells as compared to the per-
centage of KIR2DL/S11 NK cells stimulated by HLA-
C1negC2pos K562 cells (Figure 3a and b). Vice versa, stimulation
of NK cells with HLA-C1negC2pos K562 cells increased the per-
centage of KIR2DL2/DL3/DS21 NK cells within the alloreac-
tive CD1071 NK cell pool against K562-C1negC2pos target cells
as compared to the percentage of KIR2DL2/DL3/DS21 NK
cells stimulated by HLA-C1posC2neg K562 cells (Figure 3a and
c). These results show a clear trend, albeit not significant, that
stimulation of NK cells in the absence of a specific KIR ligand
may lead to an increase of specific cytolytic KIR1 NK cells
within the whole alloreactive NK cell pool against target cells
lacking the same KIR ligand. This trend was observed for both
NKG2A1 and NKG2A2 NK cells as long as they expressed the
relevant KIR. As the percentage of the specific cytolytic KIR1
NK cells increased, we further investigated whether this would
lead to an enhanced cytolytic alloresponse within a specific
alloreactive KIR1 subset.
Increased specific cytolytic alloresponse after culture in the
absence of a specific KIR ligand
As we were interested to know whether a given alloreactive
KIR1 NK cell subset showed enhanced alloreactivity when
prestimulated by K562 cells lacking the inhibitory KIR ligand,
we tested the degranulation potential of KIR2DL/S11 NK cells
against K562-C1posC2neg target cells after 7 days of prestimula-
tion with HLA-C1posC2neg stimulatory cells (K562-C1posC2neg)
and compared this with prestimulation with HLA-negative
K562 cells (K562-Cneg) or unstimulated NK cells (D0)
(Figure 4a). Stimulating the cells for 7 days with K562-Cneg
cells led to an enhanced cytolytic alloresponse (% degranulat-
ing cells) of KIR2DL/S11 NK cells, as compared to freshly
isolated NK cells (D0) against K562-C1posC2neg target cells.
Thus, stimulation of NK cells by coculture with irradiated
HLA-C1 environmentKIR negative
CFSE CFSE CFSE
Single KIR2DL1/S1 Single KIR2DL2/S2/L3
27%
FS FS FS
34% 23%
HLA-C2 environment
= HLA-C1 environment
= HLA-C2 environment
KIR negative
CFSE
100
** **
0KIR
negativeSingle
KIR2DL/S1
Div
isio
n w
ithin
NK
cel
l sub
set (
%)
SingleKIR2DL/S2/3
80
60
40
20
CFSE CFSE
Single KIR2DL1/S1 Single KIR2DL2/S2/L387%
FS FS FS
58% 83%
a
b
Figure 2 Division of single KIR2DL/S11, single KIR2DL/S2/31 andKIRneg NK cell subsets at day 7 of culture in an HLA-C1posC2neg andHLA-C1negC2pos environment. CFSE-labeled NK cells, from donors 11,13 and 14, were cultured in the presence of irradiated HLA-C1posC2neg
and HLA-C1negC2pos K562 stimulatory cells and analyzed at day 7 ofculture. (a) Representative dot plots for the cumulative data shown in b.The open plots (#/%) in b are for heterozygous (C1C2) donors and theclosed plots (N/&) are for donors lacking C2 (C1C1). (b) The results forthree donors of whom NK cells were cultured in an HLA-C1 (#) and anHLA-C2 (%) environment. Depicted are the percentages of dividedcells within the KIR negative, single KIR2DL/S11 and single KIR2DL/S2/31 NK cell subsets at day 7 of culture. Differences in divisionbetween the different NK cell subsets were analyzed using repeatedmeasures ANOVA analysis; **P,0.01. HLA, human leucocyte antigen;KIR, immunoglobulin-like receptor; NK, natural killer.
Skewing NK cell alloreactivity
DN Eissens et al
164
Cellular & Molecular Immunology
HLA-C-negative K562 cells (Cneg) leads to increased alloreac-
tivity. More importantly, stimulation of NK cells in the absence
of the specific KIR ligand for KIR2DL/S11 (C1posC2neg)
showed an even stronger cytolytic response of KIR2DL/S11
NK cells towards K562-C1posC2neg target cells as compared
to nonspecifically stimulated NK cells (Cneg).
As priming of NK cells towards their target is of interest for
future NK cell therapies against hematological malignancies or for
treatment of solid tumors, we further investigated if stimulation
of NK cells in the absence of a specific KIR ligand would trigger
the cytolytic KIR1 NK cell subset towards increased killing of a
primary tumor cell line. To this end, we tested freshly isolated NK
cells before culture (D0) and after 7 days of culture with HLA-
negative stimulatory cells (Cneg) and HLA-C1posC2neg stimula-
tory cells (C1posC2neg) against a primary HLA-C1 homozygous
tumor cell line (Kasumi) (Figure 4b). Comparable to the allor-
eactivity towards K562-C1posC2neg target cells, results show that
overall nonspecific stimulation of NK cells within culture (Cneg)
led to enhanced alloreactivity of KIR2DL/S11 NK cells towards
the primary HLA-C1 homozygous Kasumi tumor cell line. In
addition, the alloreactivity of the KIR2DL/S11 NK cells was fur-
ther improved when NK cells were specifically stimulated with
HLA-C1posC2neg K562 cells. Thus, in the absence of a specific KIR
ligand within the culture system, specific KIR1 NK cells can be
primed to give an enhanced cytolytic alloresponse towards non-
previously encountered primary tumor cell lines lacking the same
KIR ligand.
DISCUSSION
Regulation of cytolytic responses of alloreactive NK cells
through interactions between inhibitory KIR and HLA class I
ligands has been well described.12 In this study, we show that
the phenotype of mature human NK cells from healthy donors
can be skewed through stimulation with cells lacking a specific
KIR ligand (KIR vs. HLA-C), resulting in an enrichment of
specific alloreactive NK cells bearing higher cytolytic responses
towards specified targets. A simplified model summarizing our
findings is shown in Figure 5.
Previous research by Rose et al.26 showed that the KIR rep-
ertoire of human NK cells can be shaped mainly by the HLA
class I environment. They showed that when feeder cells were
mismatched with NK cells for a specific KIR ligand (HLA mis-
match), the frequency of the cognate KIR in the NK cell popu-
lation would increase as compared to autologous or allogeneic
HLA-matched cultures. Our results demonstrate that the NK
cell KIR repertoire can be skewed in both HLA-matched, as well
HLA-C environmentC1
30
20
10
40
0
Sin
gle
KIR
2DL/
S1+
cel
lsw
ithin
CD
107+
allo
reac
tive
pool
(%)
Sin
gle
KIR
2DL/
S2/
3+ c
ells
with
in C
D10
7+ a
llore
activ
e po
ol (%
)
C1HLA-C environment HLA-C environment
C2
30
20
10
40
0C1 C2
23% 7%
16%
KIR2DL/S1
KIR2DL/S2/3
KIR2DL/S2/3
KIR2DL/S1
KIR
2DL/
S1+
with
in C
D10
7-K
IR2D
L/S
2/3+
with
in C
D10
7+K
IR2D
L/S
2/3+
with
in C
D10
7-K
IR2D
L/S
1+w
ithin
CD
107+
11%
22% 27%
25% 31%
CD
56
CD
56
Targ
et: K
562-
C1
Target:K562-C1
Target:K562-C2
Targ
et: K
562-
C2
CD
56
CD
56C
D56
CD
56
CD
56C
D56
C2
a
b c
Figure 3 Distribution of single KIR2DL/S11 and single KIR2DL/S2/31 NK cells within the alloreactive (CD1071) NK cell population upon targetencounter. Freshly isolated NK cells, from donors 2, 3, 4 and 5, were cultured in an HLA-C1posC2neg and HLA-C1negC2pos environment andanalyzed for degranulation (CD1071) at day 7 of culture against K562-C1posC2neg and K562-C1negC2pos target cells. (a) Representative dot plots forthe cumulative data shown in b. The upper two panels depict the percentages of alloreactive (CD1071) and non-alloreactive (CD107a2) singleKIR2DL/S11 NK cells within the whole CD1071 and CD107a2 NK cell pool respectively upon K562-C1posC2neg target encounter at day 7 of culturein an HLA-C1posC2neg and HLA-C1negC2pos environment. The lower panels depict the percentages of alloreactive (CD1071) and non-alloreactive(CD107a2) single KIR2DL/S2/31 NK cells within the whole CD1071 and CD107a2 NK cell pool respectively upon K562-C1negC2pos targetencounter at day 7 of culture in an HLA-C1posC2neg and HLA-C1negC2pos environment. (b) The percentages of single KIR2DL/S11 NK cells withinthe alloreactive (CD1071) NK cell population upon K562-C1posC2neg target encounter at day 7 of culture in an HLA-C1posC2neg and HLA-C1negC2pos environment. Differences between the different stimulation settings were analyzed using paired t-tests. (c) The percentages of singleKIR2DL/S2/31 NK cells within the alloreactive (CD1071) NK cell population upon K562-C1negC2pos target encounter at day 7 of culture in an HLA-C1posC2neg and HLA-C1negC2pos environment. Differences between the different stimulation settings were analyzed using paired t-tests. HLA,human leucocyte antigen; KIR, immunoglobulin-like receptor; NK, natural killer.
Skewing NK cell alloreactivity
DN Eissens et al
165
Cellular & Molecular Immunology
as in HLA-mismatched situations, in the presence of a KIR–
KIR ligand mismatch. Notably, in our culture set-up, we used
purified and freshly isolated NK cells and cultured them
directly, without pre-activation, in the presence of low-dose
IL-2 and IL-15. This set-up may allow for a more bona fide
reaction of mature human NK cells to missing KIR ligands,
as compared to the a situation where NK cells have been pre-
activated. Remarkably, in our CFSE-based division analyses,
the skewing effect obtained through HLA-C1negC2pos stimu-
lation tended to be stronger than HLA-C1posC2neg stimulation,
which may be explained by the HLA-C type of the donors. Of
the three donors used for these experiments (Table 1; donors
11, 13 and 14), two out of three were HLA-C1 homozygous,
whereas only one was HLA-C1/C2 heterozygous. According to
the NK cell licensing/education model,27–29 KIR2DL/S11 NK
cells of HLA-C1 homozygous donors may be hyporesponsive,
as these cells lacked their cognate ligand in vivo. This would
result in a damped response within an HLA-C1posC2neg envi-
ronment. The KIR2DL2/DL3/DS21 NK cells from these same
donors would be properly armed as their ligand, HLA-C1, was
present in vivo and thus, these cells may be perfectly able to
respond within a HLA-C1negC2pos environment. The KIR2DL/
S11 and KIR2DL2/DL3/DS21 NK cell subsets from the HLA-C
heterozygous donor would have been properly armed and thus
able to respond to either HLA-C1posC2neg or HLA-C1negC2pos
environments (Figure 2b). Thus, our results suggest that the
self-HLA-C background of NK cells may be an important
determinant for the strength of their response to missing KIR
ligands ex vivo. However, our experimental set-up was not
designed to allow for conclusive statements on the effect of
the HLA background or indeed the licensing/education of
NK cells on KIR responses to their respective ligands.
Concerning adoptive transfer of mature NK cells for immu-
notherapeutic purposes, the results of this study suggest that
the presence of inhibitory KIR on donor NK cells in absence of
its cognate ligand in the recipient as well as the HLA back-
ground of the donor NK cells could be two key factors that
determine the alloreactive NK cell response within the recipi-
ent.
During NK cell maturation, the KIR repertoire is predomi-
nantly formed by the KIR genotype of the cells and is only
mildly influenced by the HLA class I type.30,31 The minor role
of HLA class I in shaping the KIR repertoire is also reflected in
the setting of allogeneic SCT, as reconstitution of the KIR re-
pertoire mainly reflects that of the donor and not the recipi-
ents.19,32 Our in vitro study, however, suggests that the KIR
repertoire of human mature NK cells can be reshaped by the
absence of the relevant ligands. Moreover, in vitro reshaping of
the KIR repertoire of mature NK cell populations, within a
specific KIR ligand lacking environment, is likely to increase
the cytolytic response of the alloreactive KIR1 NK cells against
target cells lacking the cognate ligand. Thus, as the KIR geno-
type is envisaged to be important in the formation of the basic
KIR repertoire during NK cell maturation, the HLA class I
environment could be a dominant factor in reshaping the
KIR repertoire and determining the strength of the cytolytic
NK cell response after NK cell maturation.
Currently, NK cells are already being exploited in the setting
of hematological malignancies to induce an antileukemic
effect.33 An ex vivo coculture protocol in which NK cell skewing
can be induced by the presence of specific KIR ligands (i.e.,
1000
c
600
400
200
FS
800
0
CD107a
15.83%
1000
600
400
200
FSTa
rget
: K56
2-C
2Ta
rget
: K56
2-C
1
800
0100 102101
CD107a103 100 102101 103
100 102101 103100 102101 103
1000
600
400
200
FS
800
0
CD107a
9.15% 17.82%
1000
600
400
200
FS
800
0
CD107a
9.75%
HLA-C1 environmentexpanded single
KIR2DL1/S1+ cells
HLA-C2 environmentexpanded single
KIR2DL2/S2/L3+ cells
30
20
10
% d
egra
nula
tion
with
inK
IR2D
L1/S
1+ c
ells
Target:K562-C1
40
*
**
0D0
a
Cneg C1
8
6
4
2
% d
egra
nula
tion
with
inK
IR2D
L1/S
1+ c
ells
Target:Kasumi
10 *
0D0
b
Cneg C1
Figure 4 Degranulation of single KIR2DL/S11 NK cells upon targetencounter. Freshly isolated NK cells, from donors 6, 7 and 9, werecultured in a nonspecific (HLA-Cneg) and HLA-C1posC2neg envir-onment. NK cells were analyzed for degranulation before culture (D0)and at day 7 of culture upon the encounter of (a) K562-C1posC2neg
target cells and (b) the HLA-C1 homozygous primary tumor cell lineKasumi. Differences in degranulation within the single KIR2DL/S11 NKcell subset before and after culture were analyzed repeated measuresANOVA analysis; *P,0.05, **P,0.01. (c) Inhibition of degranulationof in vitro expanded single KIR2DL1/S11 and single KIR2DL2/S2/L31
NK cells from donor 4 when cultured with K562-C2 and K562-C1 targetcells, respectively. HLA, human leucocyte antigen; KIR, immuno-globulin-like receptor; NK, natural killer.
Skewing NK cell alloreactivity
DN Eissens et al
166
Cellular & Molecular Immunology
HLA molecules) can be an augmentation to these protocols to
achieve desired KIR-mediated alloreactivity. Obviously, this
should be performed under GMP conditions and the relevant
HLA molecules should preferably be expressed in a cell-free
system.
Overall, these findings hold promise for future transplanta-
tion strategies using mature NK cells as effectors, and future
research is warranted to optimize and exploit the skewing of
NK cell responses towards specific targets for the immunother-
apeutic treatment of hematological malignancies.
COMPETING FINANCIAL INTEREST
Authors have no conflicting interests.
ACKNOWLEDGEMENTSThe authors would like to thank Sylvie van der Zeeuw-Hingrez for
technical assistance in generating the transfected K562 cell lines.
Technical support on the Gallios, and specified conjugated mAbs were
kindly provided by Beckman Coulter. Clive M. Michelo is funded by
the Dutch kidney foundation.
1 Li JM, Giver CR, Lu Y, Hossain MS, Lonial S, Waller EK. Separatinggraft-versus-leukemia from graft-versus-host disease in allogeneichematopoietic stem cell transplantation. Immunotherapy 2009; 1:599–621.
2 Li N, Chen Y, He W, Yi T, Zhao D, Zhang C et al. Anti-CD3preconditioning separates GVL from GVHD via modulating hostdendritic cell and donor T-cell migration in recipients conditionedwith TBI. Blood 2009; 113: 953–962.
3 Das R, Komorowski R, Hessner MJ, Subramanian H, Huettner CS, CuaD et al. Blockade of interleukin-23 signaling results in targetedprotection of the colon and allows for separation of graft-versus-hostand graft-versus-leukemia responses. Blood 2010; 115: 5249–5258.
4 Kitazawa Y, Li XK, Liu Z, Kimura H, Isaka Y, Hunig T et al. Preventionof graft-versus-host diseases by in vivo supCD28mAb-expandedantigen-specific nTreg cells. Cell Transplant 2010; 19: 765–774.
5 Ruggeri L, Capanni M, Urbani E, Perruccio K, Shlomchik WD, Tosti Aet al. Effectiveness of donor natural killer cell alloreactivity inmismatched hematopoietic transplants. Science 2002; 295:2097–2100.
6 Cook MA, Milligan DW, Fegan CD, Darbyshire PJ, Mahendra P,Craddock CF et al. The impact of donor KIR and patient HLA-Cgenotypes on outcome fol lowing HLA-ident ical s ib l inghematopoietic stem cell transplantation for myeloid leukemia.Blood 2004; 103: 1521–1526.
7 Giebel S, Nowak I, Wojnar J, Markiewicz M, Dziaczkowska J, Wylezol Iet al. Impact of activating killer immunoglobulin-like receptorgenotype on outcome of unrelated donor-hematopoietic celltransplantation. Transplant Proc 2006; 38: 287–291.
8 Pradier A, Passweg J, Villard J, Kindler V. Human bone marrow stromalcells and skin fibroblasts inhibit natural killer cell proliferation andcytotoxic activity. Cell Transplant 2011; 20: 681–691.
9 Papamichail M, Perez SA, Gritzapis AD, Baxevanis CN. Natural killerlymphocytes: biology, development, and function. Cancer ImmunolImmunother 2004; 53: 176–186.
10 Farag SS, Caligiuri MA. Human natural killer cell development andbiology. Blood Rev 2006; 20: 123–137.
IrradiatedHLA-C1+ feeder cells
Expansion of (single)KIR2DL/S1+ NK cells
IrradiatedHLA-C2+ feeder cells
Legend KIR and HLA:
Human peripheralNK cells
= KIR2DL/S1 = KIR2DL/S2/3= HLA-C2 = HLA-C1
Expansion of (single)KIR2DL/S2/3+ NK cells
- Increased amount of cytolytic KIR2DL/S1+ NK cells- Cytolytic KIR2DL/S1+ NK cells are more alloreactive on a per cell basis
- Cytolytic KIR2DL/S2/3+ NK cells are more alloreactive on a per cell basis
Against HLA-C1 target cells:
- Increased amount of cytolytic KIR2DL/S2/3+ NK cells
Against HLA-C2 target cells:
Figure 5 Schematic in vitro model for the selective expansion of alloreactive NK cells. Here, a human peripheral NK cell population is showncontaining KIR-negative, KIR2DL/S11, KIR2DL/S2/31 and KIR2DL/S11/KIR2DL/S2/31 NK cells. In the presence of irradiated HLA-C1posC2neg
stimulator cells (upper scheme), KIR2DL/S11 NK cells expand in culture leading to an increase in cytolytic KIR2DL/S11 NK cells within the wholealloreactive NK cell population upon HLA-C1posC2neg target encounter. In addition, these cytolytic KIR2DL/S11 cells reveal a higher alloreactivecapacity, as compared to freshly isolated or nonspecific stimulated KIR2DL/S11 NK cells. Vice versa, in the presence of irradiated HLA-C1negC2pos
stimulator cells (lower scheme), KIR2DL/S2/31 NK cells expand in culture leading to an increase in cytolytic KIR2DL/S2/31 NK cells within thewhole alloreactive NK cell population upon HLA-C1negC2pos target encounter. In addition, although not analyzed in this study, these cytolyticKIR2DL/S2/31 cells may be more alloreactive, as compared to freshly isolated or nonspecific stimulated KIR2DL/S2/31 NK cells. HLA, humanleucocyte antigen; KIR, immunoglobulin-like receptor; NK, natural killer.
Skewing NK cell alloreactivity
DN Eissens et al
167
Cellular & Molecular Immunology
11 Braud VM, Allan DS, O’Callaghan CA, Soderstrom K, D’Andrea A, OggGS et al. HLA-E binds to natural killer cell receptors CD94/NKG2A, Band C. Nature 1998; 391: 795–799.
12 Parham P. MHC class I molecules and KIRs in human history, healthand survival. Nat Rev Immunol 2005; 5: 201–214.
13 Ruggeri L, Capanni M, Casucci M, Volpi I, Tosti A, Perruccio K et al.Role of natural killer cell alloreactivity in HLA-mismatchedhematopoietic stem cell transplantation. Blood 1999; 94: 333–339.
14 Ruggeri L, Capanni M, Mancusi A, Urbani E, Carotti A, Aloisi T et al.Alloreactive natural killer cells in mismatched hematopoietic stemcell transplantation. Blood Cells Mol Dis 2004; 33: 216–221.
15 Ruggeri L, Mancusi A, Capanni M, Urbani E, Carotti A, Aloisi T et al.Donor natural killer cell allorecognition of missing self in haploidenticalhematopoietic transplantation for acute myeloid leukemia: challengingits predictive value. Blood 2007; 110: 433–440.
16 Leung W, Iyengar R, Turner V, Lang P, Bader P, Conn P et al.Determinants of antileukemia effects of allogeneic NK cells. JImmunol 2004; 172: 644–650.
17 Hsu KC, Keever-Taylor CA, Wilton A, Pinto C, Heller G, Arkun K et al.Improved outcome in HLA-identical sibling hematopoietic stem-celltransplantation for acute myelogenous leukemia predictedby KIR andHLA genotypes. Blood 2005; 105: 4878–4884.
18 Leung W, Iyengar R, Triplett B, Turner V, Behm FG, Holladay MS et al.Comparison of killer Ig-like receptor genotyping and phenotyping forselection of allogeneic blood stem cell donors. J Immunol 2005; 174:6540–6545.
19 ShillingHG, McQueen KL, Cheng NW, Shizuru JA, Negrin RS, ParhamP. Reconstitution of NK cell receptor repertoire following HLA-matched hematopoietic cell transplantation. Blood 2003; 101:3730–3740.
20 Cooley S, McCullar V, Wangen R, Bergemann TL, Spellman S,Weisdorf DJ et al. KIR reconstitution is altered by T cells in thegraft and correlates with clinical outcomes after unrelated donortransplantation. Blood 2005; 106: 4370–4376.
21 Zhao XY, Huang XJ, Liu KY, Xu LP, Liu DH. Reconstitution of naturalkiller cell receptor repertoires after unmanipulated HLA-mismatched/haploidentical blood and marrow transplantation: analyses ofCD94:NKG2A and killer immunoglobulin-like receptor expressionand their associations with clinical outcome. Biol Blood MarrowTransplant 2007; 13: 734–744.
22 Yu J, Venstrom JM, Liu XR, Pring J, Hasan RS, O’Reilly RJ et al.Breaking tolerance to self, circulating natural killer cells express-ing inhibitory KIR for non-self HLA exhibit effector function after Tcell-depleted allogeneic hematopoietic cell transplantation. Blood2009; 113: 3875–3884.
23 Gagne K, Brizard G, Gueglio B, Milpied N, Herry P, Bonneville F et al.Relevance of KIR gene polymorphisms in bone marrowtransplantation outcome. Hum Immunol 2002; 63: 271–280.
24 van der Meer A, Schaap NP, Schattenberg AV, van Cranenbroek B,Tijssen HJ, Joosten I. KIR2DS5 is associated with leukemia freesurvival after HLA identical stem cell transplantation in chronicmyeloid leukemia patients. Mol Immunol 2008; 45: 3631–3638.
25 Aktas E, Kucuksezer UC, Bilgic S, Erten G, Deniz G. Relationshipbetween CD107a expression and cytotoxic activity. Cell Immunol2009; 254: 149–154.
26 Rose MJJ, Brooks AG, Stewart LA, Nguyen TH, Schwarer AP. Killer Ig-like receptor ligand mismatch directs NK cell expansion in vitro.J Immunol 2009; 183: 4502–4508.
27 Raulet DH, Vance RE. Self-tolerance of natural killer cells. Nat RevImmunol 2006; 6: 520–531.
28 Parham P. Taking license with natural killer cell maturation andrepertoire development. Immunol Rev 2006; 214: 155–160.
29 Vivier E, Tomasello E, Baratin M, Walzer T, Ugolini S. Functions ofnatural killer cells. Nat Immunol 2008; 9: 503–510.
30 Yawata M, Yawata N, Draghi M, Little AM, Partheniou F, Parham P.Roles for HLA and KIR polymorphisms in natural killer cell repertoireselection and modulation of effector function. J Exp Med 2006; 203:633–645.
31 Andersson S, Fauriat C, Malmberg JA, Ljunggren HG, Malmberg KJ.KIR acquisition probabilities are independent of self-HLA class Iligands and increase with cellular KIR expression. Blood 2009;114: 95–104.
32 Shilling HG, Young N, Guethlein LA, Cheng NW, Gardiner CM, Tyan Det al. Genetic control of human NK cell repertoire. J Immunol 2002;169: 239–247.
33 Lim SA, Kim TJ, Lee JE, Sonn CH, Kim K, Kim J et al. Ex vivoexpansion of highly cytotoxic human NK cells by cocultivation withirradiated tumor cells for adoptive immunotherapy. Cancer Res 2013;73: 2598–2607.
Skewing NK cell alloreactivity
DN Eissens et al
168
Cellular & Molecular Immunology