selective expansion of human natural killer cells leads to enhanced alloreactivity

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

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.


DN Eissens et al

162


(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


DN Eissens et al

163


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.


DN Eissens et al

164


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.


DN Eissens et al

165


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.


DN Eissens et al

166


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.


DN Eissens et al

167


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.


DN Eissens et al

168


selective expansion of human natural killer cells leads to enhanced alloreactivity

Documents