Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
American Journal of Reproductive Immunology
ORIGINAL ARTICLE
Natural Killer Cell Education in Women With Recurrent
Pregnancy Loss
Amber E. M. Lombardi1,2DeniseH.J.Habets
1,2Salwan Al-Nasiry1,3Marc E. A. Spaanderman1,3,4Lotte Wieten1,2
1GROW Research Institute for Oncology and Reproduction, MaastrichtUniversity, Maastricht, The Netherlands 2Department of Transplantation Immunology,
Maastricht University Medical Centre+, Maastricht, The Netherlands 3Department of Obstetrics and Gynecology, Maastricht University Medical Centre+,
Maastricht, The Netherlands 4Department of Obstetrics and Gynecology, Radboud University Medical Centre, Nijmegen, The Netherlands
Correspondence: Lotte Wieten (l.wieten@mumc.nl)
Funding: This study was supported by the Academic Incentive Maastricht UMC+2016.
Keywords: cytomegalovirus | education | killer-cell immunoglobulin-like receptor | natural killer cells | NKG2A | recurrent pregnancy loss
ABSTRACT
Problem: Natural killer (NK) cells undergo education for full functionality via interactions between killer immunoglobulin-
like receptors (KIRs) or NKG2A and human leukocyte antigen (HLA) ligands. Presumably, education is important during early
pregnancy as insufficient education has been associated with impaired vascular remodeling and restricted fetal growth in mice.
NK cell education is influenced by receptor co-expression patterns, human cytomegalovirus (CMV), the HLA-ER107Gdimorphism,
and HLA-B leader peptide variants. We hypothesized altered NK cell education status and differences in frequencies of HLA-E
genotypes and HLA-B leader peptide variants in women with recurrent pregnancy loss (RPL) compared to women with previously
uncomplicated pregnancies, and between CMV seropositive and seronegative RPL women.
Methods of Study: Peripheral blood mononuclear cells were analyzed by flow cytometry. HLA-ABC was typed by sequence-
specific oligonucleotide PCR, and HLA-E by Sanger sequencing. CMV status was determined by anti-CMV IgG immunoassay. NK
cells were considered “educated” if the HLA ligand to a KIR or NKG2A was present.
Results: KIR/NKG2A co-expression patterns and percentages of educated NK cells were similar between RPL and controls, and
between seropositive and seronegative RPL women. Frequencies of HLA-E genotypes and HLA-B leader peptide variants were
comparable. RPL women with the HLA-B T/T variant had a lower percentage of NKG2A-educated NK cells (47.8%) compared to
controls (66.4%) (p=0.025).
Conclusions: HLA-B leader peptide variants might impact NKG2A-specific NK cell education in RPL, warranting validation in
larger studies. Follow-up studies are needed to investigate the education status of uterine NK cells and their role in pregnancy.
1 Introduction
Natural killer (NK) cells play an important role in the first
line of defense against several types of pathogens and can-
cer. In addition, they may play a supportive role throughout
pregnancy. Their activation is regulated by a combination of
inhibitory and activating signals, mediated by many different
receptors including killer immunoglobulin-like receptors (KIRs)
and C-type lectin-like receptors such as NKG2, both receptor
families can interact with human leukocyte antigen (HLA) class I
molecules found on all nucleated cells [1, 2]. A process known as
“education” ensures tolerance to healthy self-cells and increases
killing potential against diseased cells. The interaction between
inhibitory receptors and their HLA class I ligands generate a
signaling cascade and metabolic changes, whereby the density
of intracellular effector molecules like perforin and granzyme
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work
is properly cited, the use is non-commercial and no modifications or adaptations are made.
© 2025 The Author(s). American Journal of Reproductive Immunology published by John Wiley & Sons Ltd.
American Journal of Reproductive Immunology, 2025; 93:e70045
https://doi.org/10.1111/aji.70045
1of11
increases, and activating receptors are mobilized to make the NK
cell more receptive to activating stimuli and the NK cell becomes
better equipped and capable to recognize and respond to diseased
cells. In contrast, NK cells that do not engage their inhibitory
receptors with HLA class I molecules do not undergo these
changes and are regarded as hyporesponsive [3, 4]. Therefore, NK
cell education can result in two distinct types of NK cells; edu-
cated NK cells that are effective sensors of HLA class I negative
targets and uneducated NK cells that remain unresponsive.
Diversity in NK cell education status can be attributable to various
factors. KIRs are expressed stochastically, creating a diverse pool
of NK cells that can be either educated or not [5]. Moreover, KIRs
are inherited independently of highly polymorphic class I HLA-
A, -B, and -C molecules. Therefore, it could be that an individual
is lacking the corresponding HLA class I ligand for a specific
KIR, hereby generating uneducated NK cells for this HLA-KIR
combination. NKG2A can contribute to NK cell education by
interacting with HLA-E [6–8]. Stabilization of HLA-E on the cell
surface requires binding of HLA-E with peptides derived from
the leader sequences of HLA class I molecules and HLA-G [9,
10]. In contrast to classical HLA class I molecules, only two
main HLA-E protein variants are found, namely, HLA-E*01:01
and HLA-E*01:03 [11]. They can be distinguished by having
either an arginine (HLA-ER) or glycine (HLA-EG) at position
107 that results in HLA-E*01:01 and HLA-E*01:03, respectively.
HLA-E*01:03 has been shown to have a higher peptide binding
affinity resulting in higher HLA-E protein surface expression
on the cell surface compared to HLA-E*01:01 [11], which may
affect their potential to educate NK cells. Moreover, a dimor-
phism at position −21 in the leader sequence of HLA-B has
been shown to impact NKG2A-mediated NK cell education by
lowering the availability of HLA-E on the cell surface [12]. An
additional factor possibly affecting the overall status of NK cell
education is the exposure to viruses, for instance, the human
cytomegalovirus (HCMV) has been shown to have a stable and
durable imprint on KIR repertoires hereby inducing a pool of
“adaptive” NKG2C+NKG2A−self-specific KIR-educated NK cells
with potent effector functions [13]. Furthermore, expression of
increasingly higher numbers of self-specific inhibitory KIRs has
been correlated with an increased effector function [14].
During pregnancy, NK cells should remain tolerant against
the semi-allogeneic trophoblasts, while remaining responsive to
pathogens [15, 16]. NK cells are in close contact with extravillous
trophoblast (EVT) cells that invade into the decidua during early
pregnancy and have been shown to express HLA-C, -E, and -
G molecules [17]. These NK cells have been shown to regulate
trophoblast invasion and direct vascular spiral artery remodeling,
and the interactions between KIR/NKG2A and HLA contribute
to the regulation of NK cell effector functions [18–20]. Given
their apparent importance for supporting pregnancy and their
ability to directly interact with trophoblasts in the decidua, NK
cell education status could be relevant during early pregnancy.
Animal studies have described a possible role for uterine NK
(uNK) cell education during early pregnancy. Education of uNK
cells by major histocompatibility complex (MHC) class I has been
shown to be important for fetal growth and arterial transfor-
mation [15, 16]. Moreover, unresponsiveness of uneducated NK
cells has been implied to reduce uterine vascular remodeling
[15]. Hence, altered numbers of educated NK cells or differences
in receptor repertoire might be important for stratifying women
with pregnancy complications, such as immune-compromised
women with unexplained recurrent pregnancy loss (uRPL).
In this study, we aimed to investigate KIR/NKG2A co-expression
and percentages of educated NK cells in women with uRPL and
women who had previously uncomplicated pregnancies. Since
CMV infection can have a durable imprint on KIR expression,
the impact of CMV seropositivity on both KIR/NKG2A co-
expression and NK cell education was investigated. In addition,
frequencies of HLA-E single-nucleotide polymorphisms (SNPs)
and of different HLA-B leader peptide amino acids at position −21
were investigated as these can influence the potential for NK cells
to become educated via NKG2A.
2 Methods
2.1 Study Population
In this retrospective study, 45 women with unexplained RPL and
16 women with previously uncomplicated pregnancies (controls)
were included. RPL was defined as two or more pregnancy losses
before 20 weeks of gestation according to the ESHRE guidelines
[21]. If the clinical RPL evaluation revealed an abnormal parental
karyotype, thrombophilia, endocrine abnormalities, or uterine
anomalies, women were excluded from the study. Written consent
was obtained according to the Medical Ethical Committee of the
Maastricht University Medical Centre (MUMC+) NL6736.068.18.
Thereafter, baseline characteristics and peripheral blood (PB)
samples were obtained from 45 women with RPL and 16 controls
for determining NK cell receptor repertoire and NK cell education
status. Blood plasma was only available from 27 women with RPL
for further CMV analyses. DNA samples from 15 controls and 11
women with RPL, who were included under identical criteria,
were available for HLA-E genotyping.
2.2 PB Leukocyte Isolation
PB of 45 RPL women and 16 controls was collected in ethylenedi-
aminetetraacetic acid (EDTA) tubes. Whole blood samples were
incubated with red blood cell lysis buffer (1:5 ratio) (0.155 mol/L
NH4Cl, 10 mmol/L KHCO3, and 0.1 mol/L EDTA (Na2)withan
adjusted pH of 7.6) followed by centrifugation. The remaining
pellet containing the leukocytes was washed twice with PBS. The
viability of the cells was assessed by trypan blue staining in a
hemocytometer.
2.3 Antibodies and Flowcytometric Analysis
PB leukocytes were stained with Live/Dead Fixable Aqua
Dead Cell Stain Kit (Thermo Fisher Scientific, Waltham, MA,
USA) together with conjugated antibodies (anti-CD56-PE-Vio770
[REA196], anti-CD3-APC-Vio770 [REA6130], anti-CD158a-FITC
[REA284], anti-CD158b-PE [DX27], anti-CD158e-PerCP [DX9],
and anti-CD159a-REA110) (Miltenyi Biotec, Bergisch Gladbach,
Germany) for 30 min at 4◦C to determine KIR and NKG2A
receptor expression. Flow cytometric data analysis was conducted
using the BD FACSDiva software v8.0.2 (BD Biosciences, San Jose,
2of11 American Journal of Reproductive Immunology,2025
CA, USA). Figure S1 shows the gating strategy for determining cell
populations and KIR expression. KIR expression was determined
on both CD3−CD56+NKG2A−and CD3−CD56+NKG2A+NKcells.
To determine whether an individual was positive for a specific
KIR, at least 2% of the NK cells needed to be single positive for
any given KIR.
2.4 CMV Status
Anti-CMV IgG antibody levels in 27 blood plasma samples from
women with RPL were determined by the Elecsys CMV IgG
immunoassay according to the manufacturer’s protocol (Roche,
Mannheim, Germany).
2.5 HLA Typing
Genomic DNA of 45 RPL women and 16 controls was isolated
from lymphocytes using the QIAamp DNA blood mini kit
(Qiagen, Hilden, Germany). HLA typing of HLA-A, HLA-B, and
HLA-C alleles were performed using the Luminex technology for
sequence-specific oligonucleotide (SSO) PCR (Luminex xMAP,
Luminex, Austin, TX, USA). The HLA-B alleles were divided into
either Bw4 or Bw6 epitopes based on amino acid residue positions
76, 82, and 83. HLA-C alleles were categorized as either C1 or C2
epitopes according to the amino acid residues on positions 76,
77, and 80. HLA-A*23, -24, and -32 alleles were classified as Bw4.
Two different amino acids can be identified at position −21 of the
leader sequence of HLA-B, namely, threonine (T) or methionine
(M). The amino acid on position −21 was derived from the HLA-B
typing [22], and was assessed for 45 women with RPL and 16
controls. The complete HLA-E sequence was amplified using a
forward and a reverse primer and sequenced using a forward and
reverse primer specific for exon 3 as published previously [23]for
56 RPL women and 15 controls, see Table S1. Data were analyzed
using DNASTAR Lasergene 12 Core Suite software (DNASTAR
Lasergene, Madison, WI, USA).
2.6 Education Status
KIR surface protein expression was determined by flow cytometry
and was aligned to the presence of the HLA epitopes determined
by sequencing. In theory, NK cells can be educated when the KIR
interacts with its corresponding ligand. Educated NK cells have a
stronger effector function compared to uneducated NK cells, see
Figure S2 for a visual representation. An NK cell was classified
as “educated” if the corresponding HLA epitope that can interact
with the corresponding KIR was present. A NK cell was classified
as “uneducated” if the corresponding HLA epitope to a KIR
was not present. Single KIR educated: NK cells expressing
the combination of KIR2DL1-C2, KIR2DL2/3-C1, or KIR3DL1-
Bw4; double KIR educated: NK cells expressing two KIRs in
combination with their corresponding ligand; educated ≥1KIR:
NK cells expressing one or more KIRs in combination with their
HLA ligand; uneducated by KIR: NK cells that express no KIRs
or KIRs for which the HLA ligand was not genotypically present.
All NKG2A+NK cells were classified as “NKG2A-educated” as
HLA-E is evolutionarily preserved in humans. NKG2A-educated:
NK cells expressing NKG2A but completely lacking expression of
any KIR.
2.7 Statistical Analysis
Data were tested for normality using the Shapiro–Wilk test.
Baseline characteristics of women with RPL and controls were
presented as either a percentage and analyzed with a Pear-
son’s chi-square test, as a mean with standard deviation and
analyzed with an independent samples t-test, or as a median
with interquartile range and analyzed with a Mann–Whitney
Utest. Data on KIR/NKG2A frequencies and NK cell edu-
cation were presented as median and interquartile range and
analyzed with a Mann–Whitney Utest when comparing both
RPL versus controls and when comparing CMV seronegative
versus seropositive women with RPL. For the HLA-E and HLA-
B leader peptide sequence analyses, data were presented as
percentages and a Pearson’s chi-square test was performed to
test differences in HLA-E genotype frequencies or subgroup
binding affinity variants, and to test for differences in HLA-
B leader peptide sequence frequencies or subgroup presenting
ability variants between women with RPL and controls. Data on
the potential influences of HLA-E genotypes and HLA-B leader
peptide variants on NK cell education via NKG2A were presented
as median with interquartile range and analyzed with a Mann–
Whitney Utest to assess differences between RPL versus controls,
and seronegative versus seropositive RPL women. p<0.05 were
considered statistically significant. All statistical analyses were
performed using the IBM SPSS Statistics software for Windows,
version 29.0.0.0 (SPSS Inc., Chicago, IL, USA).
3Results
When comparing baseline characteristics, women with RPL had
on average more confirmed pregnancies (5 vs. 2, p=0.001), fewer
births (1 vs. 2, p=0.003), and more pregnancy losses (4 vs. 0,
p<0.001) (Table 1).
3.1 KIR/NKG2A Co-Expression Pattern of
Women With RPL Does Not Differ From That of
Controls
To determine the co-expression pattern of various inhibitory
NK cell receptors irrespective of NK cell education, a flow
cytometric analysis was performed on three different inhibitory
KIR receptors and the inhibitory receptor NKG2A on PB NK
cells in 45 RPL women and 16 controls. No significant differences
were observed in the frequencies of the 16 NK cell subsets that
could be defined by KIR and NKG2A co-expression patterns
between RPL and controls (Figure 1;TableS2). KIR2DL2/3 had
the highest prevalence (NKG2A−C: 10.89% [6.45–14.84], RPL:
9.68% [7.02–14.39], NKG2A+C: 6.65% [3.93–13.83], RPL: 5.22%
[3.77–8.01]) followed by KIR3DL1 (NKG2A−C: 2.42% [0.64–6.23],
RPL: 2.23% [0.10–4.55], NKG2A+C: 3.24% [0.55–4.99], RPL: 1.83%
[0.02–5.27]) and KIR2DL1 (NKG2A−C: 0.51% [0–0.86], RPL:
0.58% [0.04–0.86], NKG2A+C: 0.12% [0.03–0.70], RPL: 0.22%
[0.01–0.92]). KIR2DL1−KIR2DL2/3−KIR3DL1−NK cells had the
highest prevalence in both RPL (NKG2A−: 21.40% [13.48–26.10],
3of11
TABLE 1 Baseline characteristics of the study population.
RPL
(n=45)
Controls
(n=16) pvalue
Age (years) 32 [29–34] 32 [29–35] 0.901
Height (cm) 168 ±8172±6 0.224
Weight (kg) 73 ±16 69 ±11 0.533
BMI (kg/m2)26±523±50.232
Medication use (%) 21 20 0.942
Allergies (%) 38 40 0.920
Menarche (age) 13 ±213±10.711
Regular cycle (%) 90 83 0.658
Gravida (median) 5 [3–5] 2 [1–2] <0.001
Para (median) 1 [0–1] 2 [1–2] 0.003
Pregnancy losses (median) 3 [2–5] 0 [0–0] <0.001
Note: Data are presented as either a percentage (medication use [e.g., anti-
hypertensive and antidepressant medication], allergies [e.g., dust mites and
pollen], regular cycle) or as a mean ±SD (height, weight, BMI, menarche),
or as a median with interquartile range (age, gravida, para, pregnancy losses).
FIGURE 1 Co-expression pattern of different KIRs in women
with RPL and controls. KIR co-expression pattern on NKG2A−(A)
and NKG2A+(B) CD3−CD56+NK cells. Bars depict the median with
interquartile range; light gray bars indicate controls (n=16) and dark gray
bars indicate women with RPL (n=45).
FIGURE 2 The impact of anti-CMV IgG seropositivity on the
co-expression pattern of different KIRs in women with RPL. KIR co-
expression pattern on NKG2A−(A) and NKG2A+(B) CD3−CD56+NK
cells. Bars display median with interquartile range; gray bars represent
the anti-CMV IgG-seronegative individuals (n=12) and red bars the
seropositive individuals (n=15).
NKG2A+: 44.90% [38.11–54.04]) and controls (NKG2A−:15.90%
[10.55–22.95], NKG2A+: 48.43% [35.83–58.93]).
3.2 CMV Seropositivity Is Not Associated With
the KIR or KIR/NKG2A Co-Expression Pattern in
Women With RPL
It has been described that CMV seropositivity is associated with
differences in the NK cell receptor repertoire in women with
RPL [24]. Therefore, the effect of CMV seropositivity on KIR
and NKG2A co-expression patterns was investigated in 27 women
with RPL. In our study, no effect of CMV seropositivity was found
on the frequencies of the 16 NK cell subsets that could be defined
by KIR and NKG2A co-expression patterns (Figure 2;TableS3).
3.3 Percentage of KIR or NKG2A Educated NK
Cells Is Comparable Between Women With RPL and
Controls
Our next aim was to compare percentages of educated NK cells
between controls (n=16) and RPL (n=45). Since NK cell
4of11 American Journal of Reproductive Immunology,2025
education occurs via the interaction between KIR or NKG2A and
their corresponding HLA ligands, we first determined the geno-
typic expression of HLA C1/C2/Bw4 KIR ligands. An overview
of the frequencies of individuals expressing the combination
of KIR2DL1 and C2, KIR2DL2/3 and C1, or KIR3DL1 and
Bw4/A23/A24/A32 in both RPL and controls can be found in
Table S4.
Next, we compared percentages of educated NK cells, for exam-
ple, NK cells expressing KIR for which the corresponding ligand
is genotypically present and those thus could, at least in theory, be
educated by the interaction between KIR and HLA. This showed
no significant differences between controls and women with RPL
for NKG2A−NK cells (Figure 3A), NKG2A+NK cells (Figure 3B),
or total NK cells (Figure 3C) for the following subgroups: (1) NK
cells that were educated via only one inhibitory KIR (single KIR
educated), (2) NK cells that were educated by two of the KIRs
(double KIR educated), (3) NK cells educated by ≥1 KIRs, and (4)
NK cells that did not express any KIR or NK cells that expressed
KIRs for which the corresponding ligand was not present and
was thus considered uneducated via KIR. Finally, education via
NKG2A was investigated. For this, all NKG2A+KIR−NK cells
were considered educated, as the ligand HLA-E is evolutionarily
conserved in humans [25]. This analysis showed that there was
no significant difference between controls and women with RPL
(Figure 3D). An overview of all percentages of the different
subgroups can be found in Table S5.
3.4 No Association Between CMV Seropositivity
and the Percentage of KIR or NKG2A Educated NK
Cells in Women With RPL
In RPL women, we also compared the frequencies of educated NK
cells between CMV seronegative (n=12) and CMV seropositive
(n=15) individuals. Table S4 shows the absolute numbers and fre-
quencies of KIR-ligand combinations. For NKG2A−(Figure 4A),
NKG2A+(Figure 4B), and total NK cells (Figure 4C), we did not
observe a difference in the percentage of NK cells educated via
one KIR, via ≥1 KIR, or not via KIR.
The percentage of total NK cells educated via NKG2A
(NKG2A+KIR−) was also comparable between CMV seronegative
and seropositive women with RPL (Figure 4D;TableS6). An
overview of percentages for the different subgroups is provided
in Table S6. Double KIR-educated NK cell subgroups were not
analyzed due to low prevalence.
3.5 No Differences in the HLA-E*01:01 and
*01:03 Allele Frequencies in Women With RPL
Compared to Controls
For the HLA-E genotype analysis, DNA samples were available
from 56 women with RPL and 15 controls. HLA-E genotypes based
on the presence of an arginine (HLA-ER) or a glycine (HLA-
EG) on position 107 showed no differences in frequencies when
comparing controls and women with RPL (p=0.287) (Table 2). In
addition, genotypes were grouped into having at least one allele
with higher protein expression/binding affinity (HLA-EGG and
HLA-EGR) or having the lower protein expression/lower binding
TABLE 2 Frequencies of HLA-E alleles in women with RPL and
controls.
Genotype
Controls
(n=15)
RPL
(n=56) pvalue
HLA-ERR 5 (33.3%) 9 (16.1%) 0.287
HLA-EGG 4 (26.7%) 15 (26.8%)
HLA-EGR 6(40.0%) 32(57.1%)
HLA-EGG or HLAGR 10 (66.7%) 47 (83.9%) 0.136
HLA-ERR 5 (33.3%) 9 (16.1%)
Note: Frequencies of different HLA-E genotypes in both controls (n=15) and
women with RPL (n=56). HLA-EGG and HLA-EGR were grouped into the
higher protein expression variant group, HLA-ERR was grouped into the lower
protein expression variant group. Data are depicted as absolute numbers and
as a percentage of individuals positive for the genotype of the total population.
affinity (HLA-ERR). No significant differences were observed
when comparing controls and women with RPL (p=0.136)
(Table 2).
3.6 HLA-E Genotypes Do Not Affect the
Percentage of NKG2A-Educated NK Cells When
Comparing Women With RPL vs. Controls, and
When Comparing CMV Seropositive vs. CMV
Seronegative Women
To investigate whether the different HLA-E genotypes affect
NKG2A-educated NK cells, we compared percentages of
NKG2A+KIR−NK cells in women with RPL (n=44) versus
controls (n=15) within the following subgroups: individuals
with a HLA-ERR genotype, HLA-EGG genotype, or HLA-EGR
genotype. We also compared percentages of NKG2A+KIR−
NK cells between CMV seronegative (n=12) versus CMV
seropositive RPL (n=14) women within the same subgroups.
In none of the HLA-E subgroups, we observed a significant
difference between controls and RPL women (Figure 5A;Table
S7) or between CMV seronegative and seropositive RPL women
(Figure 5B;TableS7). Importantly, this might also be due to the
small sample size in some groups. Moreover, percentages could
not be compared for the HLA-ERR and HLA-EGG genotypes due
to a low number of individuals.
3.7 No Difference in HLA-B Leader Peptide
Sequence Variants in Women With RPL Compared
to Controls
For analyses of HLA-B leader peptide sequence position −21
variants, DNA was available from 45 RPL women and 16 controls.
No significant differences in frequencies were observed between
womenwithRPLandcontrols(p=0.776) (Table 3). Moreover,
individuals with an −21 M/M and M/T were grouped together,
as having a methionine at position −21 results in having a HLA-E
molecule that can better present peptides compared to individuals
with the −21 T/T HLA-B variant. No differences were observed
when comparing women with RPL and controls (p=0.570)
(Table 3).
5of11
FIGURE 3 NK cell education by KIR and NKG2A in women with RPL and controls. Percentages of CD3−CD56+NKG2A−(A),
CD3−CD56+NKG2A+(B), and total NK cells (CD3−CD56+NKG2A−plus CD3−CD56+NKG2A+NK cells) (C). Percentages of CD3−CD56+NK cells
educated by NKG2A (NKG2A+KIR−) (D). Single KIR educated: NK cells expressingthe combination of KIR2DL1-C2, KIR2DL2/3-C1, or KIR3DL1-Bw4;
double KIR educated: NK cells expressing two KIRs in combination with their corresponding ligand; Educated ≥1 KIR: NK cells expressing one or more
KIRs in combination with their HLA ligand; Uneducated by KIR: NK cells that express no KIRs or KIRs for which the HLA ligand is not genotypically
present. NKG2A-educated: NK cells expressing NKG2A but completely lacking expression of KIR. Dots/boxes depict individual controls (n=16) and
RPL patients (n=45). Bars indicate median with interquartile range.
6of11 American Journal of Reproductive Immunology,2025
FIGURE 4 The impact of anti-CMV IgG seropositivity on NK cell education by KIR and NKG2A in women with RPL. Percentages of
CD3−CD56+NKG2A−(A), CD3−CD56+NKG2A+(B), and total NK cells (CD3−CD56+NKG2A−plus CD3−CD56+NKG2A+NK cells) (C). Percentages of
CD3−CD56+NK cells educated by NKG2A (NKG2A+KIR−) (D). Single KIR educated: NK cells expressing the combination of KIR2DL1-C2, KIR2DL2/3-
C1, or KIR3DL1-Bw4; Educated ≥1 KIR: NK cells expressing one or more KIRs in combination with their HLA ligand; uneducated by KIR: NK cells that
express no KIRs or KIRs for which the HLA ligand is not genotypically present. NKG2A-educated: NK cells expressing NKG2A but completely lacking
expression of KIR. Every dot represents one individual; gray dots indicate anti-CMV IgG-seronegative individuals (n=12) whereas red dots indicate
IgG-seropositive individuals (n=15). Bars represent median with interquartile range.
7of11
FIGURE 5 The impact of different HLA-E alleles on the percentage of NKG2A-educated NK cells in women with RPL versus controls, and in anti-
IgG CMV seropositive versus seronegative women with RPL. Percentages of NKG2A-educated NK cells according to the different HLA-E genotypes.
Gray dots represent controls (n=15) and gray boxes depict RPL women (n=44) (A). Gray dots represent seronegative women with RPL (n=12) and
red dots depict seropositive women with RPL (n=14) (B). Bars indicate median with interquartile range.
TABLE 3 Frequencies of HLA-B leader peptide position −21 vari-
ants in women with RPL and controls.
Variant
Controls
(n=16)
RPL
(n=45) pvalue
−21 M/M HLA-B 1 (6.3%) 4 (8.9%) 0.776
−21 M/T HLA-B 8 (50%) 18 (40%)
−21 T/T HLA-B 7 (43.8%) 23 (51.1%)
−21 M/M or M/T HLA-B 9 (56.3%) 21 (46.7%) 0.570
−21 T/T HLA-B 7 (43.8%) 24 (53.3%)
Note: Frequencies of different HLA-B leader peptide position −21 variants
in women with RPL (n=45) and controls (n=16). −21 M/M and M/T
were grouped into the group with HLA-E that can better present peptides
compared to the −21 T/T group. Data are depicted as absolute numbers and
as a percentage of individuals positive for the HLA-B variants of the total
population.
Abbreviations: M, methionine; T, threonine.
3.8 HLA-B Leader Peptide Sequence Position
−21 T/T Is Associated With a Lower Percentage of
NKG2A Educated-NK Cells in Women With RPL
Compared to Controls
As it has been described that position −21 of the HLA-B leader
peptide sequence might impact NK cell education via NKG2A
[12], percentages of NKG2A-educated NK cells (NKG2A+KIR−
NK cells) were compared between women with RPL and controls
within the T/T, M/T, and M/M variant subgroups. In the subgroup
with individuals having a T/T leader peptide sequence at position
−21, we observed a lower percentage of NK cells educated via
NKG2A in RPL women (47.80% [39.29–52]) compared to the
controls (66.40% [45.95–74.06]) (p=0.025) (Figure 6A;TableS8).
No difference was observed in the M/T subgroup (p=0.311)
(Figure 6A;TableS8). Moreover, no differences were observed
in the percentage of NKG2A-educated NK cells when comparing
variants subgroups of T/T variants (48.09% [38.58–54.76] for
CMV−, 45.27% [36.25–48.43] for CMV+,p=0.537) and M/T
variants (47.52% [41.72–50.72] for CMV−, 44.53% [28.05–57.42] for
CMV+,p=1.000) in CMV seronegative versus CMV seropositive
RPL women (Figure 6B;TableS8). The prevalence of women
with a M/M variant at position −21 was too low for both
the RPL/control group and CMV seronegative/seropositive RPL
group to perform comparative analyses.
4 Discussion
In our study, we investigated NK cell receptor co-expression
patterns and theoretical education status in women with RPL
and women with previous healthy pregnancies (controls). In
addition, we examined whether receptor patterns and theoretical
education status differed between CMV seropositive and CMV
seronegative women with RPL. KIR and NKG2A co-expression
patterns were comparable and no differences in NK cell education
status were found between women with RPL and controls, or
between CMV seropositive and CMV seronegative women with
RPL. In addition, we investigated the HLA-ER107Gdimorphism
and the HLA-B leader peptide sequence position −21 variants in
RPL women and controls. We observed no differences in HLA-E
genotype frequencies or in the frequencies of the HLA-B variants.
Given the potential influence of both HLA-E genotypes and HLA-
B leader peptide variants on NK cell education via NKG2A, we,
therefore, examined their effects on the percentage of NKG2A-
educated NK cells. The HLA-E genotype did not influence
the percentage of NKG2A-educated NK cells when comparing
women with RPL and controls, nor when comparing CMV
seronegative and seropositive women with RPL. However, having
aT/Tatposition−21 of the HLA-B leader peptide sequence was
associated with a lower percentage of NKG2A-educated NK cells
in women with RPL compared to controls (p=0.025).
Despite including a well-defined cohort of uRPL women in this
study, we were unable to distinguish differences in KIR protein
expression patterns. Some studies have reported an increasedMFI
of KIR2DL1 and KIR2DL2/3 [26] and an increased percentage [26]
and number of KIR2DL1+[27] and KIR2DL2/3+[27] NK cells in
women with RPL compared to controls. Others have found no
differences [28, 29]. A possible explanation for these conflicting
results could be that most studies only analyze the presence or
absence of a KIR at the gene or molecule level. It has been shown
that allelic variation and copy number variation can determine
the expression levels of KIRs [30]. Moreover, there are indications
8of11 American Journal of Reproductive Immunology,2025
FIGURE 6 The impact of different HLA-B leader peptide variants on the percentage of NKG2A-educated NK cells in women with RPL versus
controls, and in anti-CMV IgG seropositive versus seronegative women with RPL. Percentages of NKG2A-educated NK cells according to the amino
acid on position −21 of the leader peptide sequence of HLA-B; T/T, M/M, or M/T (T =threonine, M =methionine). Gray dots represent controls (n=16)
and gray boxes depict RPL women (n=45) (A). Gray dots represent seronegative women with RPL (n=12) and red dots depict seropositive women with
RPL (n=15) (B). Bars indicate median with interquartile range. *p<0.05.
that allelic variation of KIR can impact the education status of
NK cells [31]. Therefore, in future studies, it would be important
to also consider allelic variation and copy number variation. In
addition, it would be relevant to study the expression of additional
KIRs on NK cells, that we now consider KIR negative, by using a
pan-KIR binding antibody.
In our study, we did not observe differences in NK cell education
status in women with RPL compared to the controls. However, a
limitation of this study is that we were only able to investigate
the theoretical possibility of NK cells becoming educated. To
draw firm conclusions about NK cell education in RPL, future
studies should also include functional data, demonstrating the
subsequent impact on the magnitude of the effector response.
Another limitation is that we have now focused on three different
inhibitory KIRs that can facilitate NK cell education, namely,
KIR2DL1, KIR2DL2/3, and KIR3DL1. However, several studies
have shown that activating receptors such as KIR2DS1 can
also contribute to calibrating NK cell effector functions [8, 32].
Therefore, to gain a more comprehensive understanding of NK
cell education, activating KIRs and other activating receptors
such as NKG2E and NKG2C should also be investigated.
In addition to KIR, NKG2A is an important mediator of NK cell
education. In peripheral NKG2A−or NKG2A+NK cell subsets,
we did not observe differences in co-expression patterns of KIRs
between women with RPL versus controls. Also, the percentage of
NKG2A-educated NK cells was comparable between both groups
of women. We did observe a trend towards a lower percentage
of NKG2A+NK cells that were educated by at least one KIR in
women with RPL, however, this needs to be confirmed in a larger
cohort. It has been previously shown that NKG2A-educated NK
cells are more resilient and have enhanced responses compared to
KIR-educated NK cells [33]. In addition, a study in mice showed
that the absence of the NKG2A receptor in pregnant mice led to
the inability of the NK cells in the uterus to modulate vascular
changes needed during early pregnancy [16]. Together, these
results indicate that it would be interesting to further investigate
the role of NKG2A in the pathogenesis of RPL.
When studying NK cell education status, we investigated NK cells
derived from PB (pNK). However, increasing evidence showsthat
decidual NK (uNK) cells are phenotypically different from pNK
cells and exert different functions. Whereas pNK cells are mainly
CD56dimCD16+and are involved in controlling viral infections and
eliminating cancer cells, dNK cells are mainly CD56brightCD16−
and play an important role in regulating trophoblast invasion
and spiral artery remodeling during early pregnancy [18–20, 34].
dNK cells are found in high abundance in close proximity to
invading trophoblast cells in the decidua, where they do not only
interact with fetal cells, but also with decidual stromal cells and
myeloid cells that all express maternal HLA class I molecules [35].
As these stromal cells are in close proximity to dNK cells, they
seem excellent candidates for educating dNK cells. In a mouse
model, Depierreux et al. showed that NK cells isolated from the
uterus could only be educated via interactions with maternal
MHC and not via fetal MHC [15], emphasizing the relevance of
studying maternal HLA repertoires. They also demonstrated the
importance of education status supporting early pregnancy, as the
fetuses from female mice with uneducated NK cells were growth
restricted and uterine vasculature was insufficiently remodeled
[15]. Given the importance of dNK cells during early pregnancy,
it would be of great value to further investigate human dNK cell
education by sampling NK cells through biopsies or menstrual
shed [35, 36] in women with RPL.
One factor that can impact NK cell education status is CMV
infection. Multiple studies have indicated that there is an associa-
tion between CMV seropositivity or reactivation and an increased
percentage of NKG2C+and KIR+NK cells [13, 37] and a higher
expression of NKG2C [38]. Moreover, CMV infection leads to
the clonal expansion of a specific subset of long-lived adaptive
NKG2C+CD57+self-KIR+educated NK cells [13, 39]. In our study,
we did not find differences in CMV seropositive versus CMV
seronegative women when investigating either KIR and NKG2A
receptor co-expression patterns or the percentage of educated NK
cells. In a previous study, using the same cohort of women, it
was shown that the percentage of peripheral NKG2C+NK cells
only significantly differed between seropositive and seronegative
RPL women, not in controls [24]. Unfortunately, co-expression
of NKG2C, CD57, and KIR or NKG2A could not be analyzed
in this patient cohort as the markers were not included in the
same flow cytometry panel. Therefore, would be relevant to study
the imprinting of CMV on NK cell phenotype and function in
9of11
more extensive flow cytometry panels. Moreover, it would be
interesting to analyze additional factors that have been suggested
to influence the impact of CMV infection on the immune system,
including host and virus genetics [40, 41].
Another factor that can influence NK cell education status is
the binding affinity of HLA class I molecules, such as HLA-
E. A meta-analysis on HLA-E genotypes revealed that women
having the HLA-E*01:03 genotype, meaning they only have HLA-
E molecules with higher protein levels and peptide binding
affinity, have a decreased risk of RPL [42]. As we did not find any
differences in frequencies of the genotype between our cohorts
of women, it is important that future studies investigate the
association of the HLA-E genotypes in larger cohorts of women
as now we cannot draw any firm conclusions. In addition, not
only HLA-E genotypes influence the capacity to educate NK cells.
A dimorphism at position 21 in the leader sequence of HLA-B
has also been shown to impact the ability of NKG2A to educate
NK cells [12]. A methionine residue at position −21 provides
a good anchor and allows for a stable cell surface expression
of HLA-E. However, a threonine residue at position −21 does
not bind effectively to HLA-E. Consequently, these NK cells
with a homozygous −21 threonine residue are mainly educated
through HLA-A, -B, and -C [12]. We could not observe differences
in frequencies of different HLA-B leader peptide amino acid
residues at position −21 between women with RPL and controls.
However, the presence of a −21 T/T HLA-B variant seemingly
decreased the percentage of NKG2A-educated NK cells in women
with RPL. A limitation of this study is that we did not have enough
cells to further investigate the functional impact, for example, by
performing degranulation or cytotoxicity assays of these HLA-
E genotypes and HLA-B variants, which is important to better
understand the effects of HLA-E genotypes and NK cell education
status in RPL.
In conclusion, our results show that PB NK cell education by
KIR is not valuable for stratifying immunological aberrations
in women with RPL. However, it would be informative to
further investigate peripheral NK cell education by NKG2A in
combination with the HLA-B leader peptide sequence at position
−21 in larger cohorts. In addition, future studies investigating the
process of NK cell education more locally in the uterus or more
functionally could be valuable to further elucidate its role during
early pregnancy.
Acknowledgments
This study was supported by the Academic Incentive Maastricht UMC+
2016.
Ethics Statement
All authors affirm that the procedures in this study were in accordance
with the ethical standards of the Medical Ethical Committee of the
Maastricht University Medical Centre (MUMC+) (NL6736.068.18), and
with the 1964 Helsinki Declaration and its later amendments. Informed
consent was obtained from all participants involved in the study.
Conflicts of Interest
The authors declare no conflicts of interest.
Data Availability Statement
The data that support the findings of this study are available from the
corresponding author upon reasonable request.
References
1. J. Dębska-Zielkowska, G. Moszkowska, M. Zieliński, A. Dukat-
Mazurek, P. Trzonkowski, and K. Stefańska, “KIR Receptors as Key
Regulators of NK Cells Activity in Health and Disease,” Cells 10, no. 7
(2021): 1777.
2. Y. Bartel, B. Bauer, and A. Steinle, “Modulation of NK Cell Function by
Genetically Coupled C-Type Lectin-Like Receptor/Ligand Pairs Encoded
in the Human Natural Killer Gene Complex,” Frontiers in Immunology 4
(2013): 362.
3. M. F. Wu and D. H. Raulet, “Class I-Deficient Hemopoietic Cells and
Nonhemopoietic Cells Dominantly Induce Unresponsiveness of Natural
Killer Cells to Class I-Deficient Bone Marrow Cell Grafts,” Journal of
Immunology 158, no. 4 (1997): 1628–1633.
4. N. C. Fernandez, E. Treiner, R. E. Vance, A. M. Jamieson, S. Lemieux,
and D. H. Raulet, “A Subset of Natural Killer Cells Achieves Self-
ToleranceWithout Expressing Inhibitory Receptors Specific for Self-MHC
Molecules,” Blood 105, no. 11 (2005): 4416–4423.
5. H.-W. Chan, J. S. Miller, M. B. Moore, and C. T. Lutz, “Epigenetic
Control of Highly Homologous Killer Ig-Like Receptor Gene Alleles,”
Journal of Immunology 175, no. 9 (2005): 5966–5974.
6. N. K. Björkström, P. Riese, F. Heuts, et al., “Expression Patterns
of NKG2A, KIR, and CD57 Define a Process of CD56 Dim NK-Cell
Differentiation Uncoupled From NK-Cell Education,” Blood 116, no. 19
(2010): 3853–3864.
7. P. Brodin, T. Lakshmikanth, S. Johansson, K. Kärre, and P. Höglund,
“The Strength of Inhibitory Input During Education Quantitatively Tunes
the Functional Responsiveness of Individual Natural Killer Cells,” Blood
113, no. 11 (2008): 2434–2441.
8. J. B. Le Luduec, J. E. Boudreau, J. C. Freiberg, and K. C. Hsu, “Novel
Approach to Cell Surface Discrimination Between KIR2DL1 Subtypes
and KIR2DS1 Identifies Hierarchies in NK Repertoire, Education, and
Tolerance,” Frontiers in Immunology 10, (2019): 734.
9. N. Lee, D. R. Goodlett, A. Ishitani, H. Marquardt, and D. E. Geraghty,
“HLA-E Surface Expression Depends on Binding of TAP-Dependent
Peptides Derived From Certain HLA Class I Signal Sequences,” Journal
of Immunology 160, no. 10 (1998): 4951–4960.
10. E. J. Petrie, C. S. Clements, J. Lin, et al., “CD94-NKG2A Recognition
of Human Leukocyte Antigen (HLA)-E Bound to an HLA Class I Leader
Sequence,” Journal of Experimental Medicine 205, no. 3 (2008): 725–735.
11. R. K. Strong, M. A. Holmes, P. Li, L. Braun, N. Lee, and D. E.
Geraghty, “HLA-E Allelic Variants: Correlating Differential Expression,
Peptide Affinities, Crystal Structures, and Thermal Stabilities,” Journal of
Biological Chemistry 278, no. 7 (2003): 5082–5090.
12. A. Horowitz, Z. Djaoud, N. Nemat-Gorgani, et al., “Class I HLA
Haplotypes Form Two Schools That Educate NK Cells in Different Ways,”
Science Immunology 1, no. 3 (2016): eaag1672.
13. B. Foley, S. Cooley, M. R. Verneris, et al., “Cytomegalovirus Reacti-
vation After Allogeneic Transplantation Promotes a Lasting Increase in
Educated NKG2C+Natural Killer Cells With Potent Function,” Blood 119,
no. 11 (2012): 2665–2674.
14. J. Yu, G. Heller, J. Chewning, S. Kim, W. M. Yokoyama, and K. C. Hsu,
“Hierarchy of the Human Natural Killer Cell Response Is Determined by
Class and Quantity of Inhibitory Receptors for Self-HLA-B and HLA-C
Ligands,” Journal of Immunology 179, no. 9 (2007): 5977–5989.
15. D. M. Depierreux, J. Kieckbusch, N. Shreeve, et al., “Beyond Maternal
Tolerance: Education of Uterine Natural Killer Cells by Maternal MHC
Drives Fetal Growth,” Frontiers in Immunology 13 (2022): 808227.
10 of 11 American Journal of Reproductive Immunology,2025
16. N. Shreeve, D. Depierreux, D. Hawkes, et al., “The CD94/NKG2A
Inhibitory Receptor Educates Uterine NK Cells to Optimize Pregnancy
Outcomes in Humans and Mice,” Immunity 54, no. 6 (2021): 1231–1244.e4.
17. R. Apps, S. P. Murphy, R. Fernando, L. Gardner, T. Ahad, and A.
Moffett, “Human Leucocyte Antigen (HLA) Expression of Primary Tro-
phoblast Cells and Placental Cell Lines, Determined Using Single Antigen
Beads to Characterize Allotype Specificities of Anti-HLA Antibodies,”
Immunology 127, no. 1 (2009): 26–39.
18. T. Tilburgs, Â. C. Crespo, A. Van Der Zwan, et al., “Human HLA-G+
Extravillous Trophoblasts: Immune-Activating Cells That Interact With
Decidual Leukocytes,” Proceedings of the National Academy of Sciences of
the United States of America 112, no. 23 (2015): 7219–7224.
19. J. Hanna, D. Goldman-Wohl, Y. Hamani, et al., “Decidual NK Cells
Regulate Key Developmental Processes at the Human Fetal-Maternal
Interface,” Nature Medicine 12, no. 9 (2006): 1065–1074.
20. A. E. Wallace, G. S. Whitley, B. Thilaganathan, and J. E. Cartwright,
“Decidual Natural Killer Cell Receptor Expression Is Altered in Preg-
nancies With Impaired Vascular Remodeling and a Higher Risk of
Pre-Eclampsia,” Journal of Leukocyte Biology 97, no. 1 (2015): 79–86.
21. The ESHRE Guideline Group on RPL, “ESHRE guideline: recurrent
pregnancy loss: an update in 2022,” in HumanReproductionOpen1 (2023),
hoad002.
22. D. J. Barker, G. Maccari, X. Georgiou, et al., “The IPD-IMGT/HLA
Database,” Nucleic Acids Research 51, no. D1 (2023): 1D1053–1D1060.
23. T. I. Olieslagers, C. E. M. Voorter, M. Groeneweg, Y. Xu, L. Wieten,
and M. G. J. Tilanus, “New Insights in HLA-E Polymorphism by Refined
Analysis of the Full-Length Gene,” HLA 89, no. 3 (2017): 143–149.
24. D. H. J. Habets, A. Schlütter, S. M. J. van Kuijk, M. E. A. Spaanderman,
S. Al-Nasiry, and L. Wieten, “Natural Killer Cell Profiles in Recurrent
Pregnancy Loss: Increased Expression and Positive Associations With
TACTILE and LILRB1,” American Journal of Reproductive Immunology 88,
no. 5 (2022): e13612.
25. L. Kanevskiy, S. Erokhina, P. Kobyzeva,M. Streltsova, A. Sapozhnikov,
and E. Kovalenko, “Dimorphism of HLA-E and Its Disease Association,”
International Journal of Molecular Sciences 20, no. 21 (2019): 1–16.
26. Y. Zhang, C. Huang, R. Lian, et al., “The Low Cytotoxic Activity
of Peripheral Blood NK Cells May Relate to Unexplained Recurrent
Miscarriage,” American Journal of Reproductive Immunology 85, no. 6
(2021): e13388.
27. L. Zhu, M. Aly, H. Wang, et al., “Increased Natural Killer Cell Subsets
With Inhibitory Cytokines and Inhibitory Surface Receptors in Patients
With Recurrent Miscarriage and Decreased or Normal Subsets in Kidney
Transplant Recipients Late Post-Transplant,” Clinical and Experimental
Immunology 193, no. 2 (2018): 241–254.
28. P. M. Emmer, W. Nelen, E. A. P. Steegers, J. C. M. Hendriks, M.
Veerhoek, and I. Joosten, “Peripheral Natural Killer Cytotoxicity and
CD56posCD16pos Cells Increase During Early Pregnancy in Women With
a History of Recurrent Spontaneous Abortion,” Human Reproduction 15,
no. 5 (2000): 1163–1169.
29. E. I. Ntrivalas, C. R. Bowser, J. Kwak-Kim, K. D. Beaman, and A.
Gilman-Sachs, “Expression of Killer Immunoglobulin-Like Receptors on
Peripheral Blood NK Cell Subsets of Women With Recurrent Spontaneous
Abortions or Implantation Failures,” American Journal of Reproductive
Immunology 53, no. 5 (2005): 215–221.
30. S. E. Dunphy, K. J. Guinan, C. N. Chorcora, et al., “2DL1, 2DL2 and
2DL3 All Contribute to KIR Phenotype Variability on Human NK Cells,”
Genes and Immunity 16, no. 5 (2015): 301–310.
31. C. Philippon, S. Tao, D. Clement, et al., “Allelic Variation of KIR and
HLA Tunes the Cytolytic Payload and Determines Functional Hierarchy
of NK Cell Repertoires,” Blood Advances 7, no. 16 (2023): 4492–4504.
32. C. Fauriat, M. A. Ivarsson, L. HG, K. J. Malmberg, and J. Michaëlsson,
“Education of Human Natural Killer Cells by Activating Killer Cell
Immunoglobulin-Like Receptors,” Blood 115, no. 6 (2010): 1166–1174.
33. A. J. Highton, B. P. Diercks, F. Möckl, et al., “High Metabolic Function
and Resilience of NKG2A-Educated NK Cells,” Frontiers in Immunology
11 (2020): 559576.
34. A. Albini and D. M. Noonan, “Decidual-Like NK Cell Polarization:
From Cancer Killing to Cancer Nurturing,” Cancer Discovery 11, no. 1
(2021): 28–33.
35. R. G. Van Der Molen, J. H. F. Schutten, B. Van Cranenbroek,
et al., “Menstrual Blood Closely Resembles the Uterine Immune Micro-
Environment and Is Clearly Distinct From Peripheral Blood,” Human
Reproduction 29, no. 2 (2014): 303–314.
36. K. Vomstein, P. Egerup, A. M. Kolte, et al., “Biopsy-Free Profiling of
the Uterine Immune System in Patients With Recurrent Pregnancy Loss
and Unexplained Infertility,” Reproductive Biomedicine Online 47, no. 2
(2023): 103207.
37. A. Tschan-Plessl, M. Stern, L. Schmied, et al., “Human
Cytomegalovirus Infection Enhances NK Cell Activity in Vitro,”
Transplantation Direct 2, no. 7 (2016): e89.
38. A. R. Manser, N. Scherenschlich, C. Thöns, H. Hengel, J. Timm, and
M. Uhrberg, “KIR Polymorphism Modulates the Size of the Adaptive NK
Cell Pool in Human Cytomegalovirus–Infected Individuals,” Journal of
Immunology 203, no. 8 (2019): 2301–2309.
39. V. Béziat, L. L. Liu, J. A. Malmberg, et al., “NK Cell Responses
to Cytomegalovirus Infection Lead to Stable Imprints in the Human
KIR Repertoire and Involve Activating KIRs,” Blood 121, no. 14 (2013):
2678–2688.
40. M. López-Botet, A. Muntasell, and C. Vilches, “The CD94/NKG2C+
NK-Cell Subset on the Edge of Innate and Adaptive Immunity to Human
Cytomegalovirus Infection,” Seminars in Immunology 26, no. 2 (2014):
145–151.
41. M. López-Botet, A. De Maria, A. Muntasell, M. Della Chiesa, and C.
Vilches, “Adaptive NK Cell Response to Human Cytomegalovirus: Facts
and Open Issues,” Seminars in Immunology 65 (2022): 101706.
42.T.Meuleman,L.Lashley,O.M.Dekkers,J.M.M.vanLith,F.H.J.
Claas, and K. W. M. Bloemenkamp, “HLA Associations and HLA Sharing
in Recurrent Miscarriage: A Systematic Review and Meta-Analysis,”
Human Immunology 76, no. 5 (2015): 362–373.
Supporting Information
Additional supporting information can be found online in the Supporting
Information section.
11 of 11
