CD56brightCD16+ NK Cells: A Functional Intermediate Stageof NK Cell Differentiation
Vivien Beziat,*,† Darragh Duffy,*,† Stephanie Nguyen Quoc,*,†,‡
Magali Le Garff-Tavernier,*,† Julie Decocq,*,† Behazine Combadiere,*,† Patrice Debre,*,†
and Vincent Vieillard*,†
Human NK cells comprise two main subsets, CD56bright and CD56dim cells, which differ in function, phenotype, and tissue
localization. To further dissect the differentiation from CD56bright to CD56dim cells, we performed ex vivo and in vitro experiments
demonstrating that the CD56brightCD16+ cells are an intermediate stage of NK cell maturation. We observed that the maximal
frequency of the CD56brightCD16+ subset among NK cells, following unrelated cord blood transplantation, occurs later than this of
the CD56brightCD162 subset. We next performed an extensive phenotypic and functional analysis of CD56brightCD16+ cells in
healthy donors, which displayed a phenotypic intermediary profile between CD56brightCD162 and CD56dimCD16+ NK cells. We
also demonstrated that CD56brightCD16+ NK cells were fully able to kill target cells, both by Ab-dependent cell cytotoxicity
(ADCC) and direct lysis, as compared with CD56brightCD162 cells. Importantly, in vitro differentiation experiments revealed that
autologous T cells specifically encourage the differentiation from CD56brightCD162 to CD56brightCD16+ cells. Finally, further
investigations performed in elderly patients clearly showed that both CD56brightCD16+ and CD56dimCD16+ mature subsets were
substantially increased in older individuals, whereas the CD56brightCD162 precursor subset was decreased. Altogether, these data
provide evidence that the CD56brightCD16+ NK cell subset is a functional intermediate between the CD56bright and CD56dim cells
and is generated in the presence of autologous T CD3+ cells. The Journal of Immunology, 2011, 186: 000–000.
Natural killer cells were originally described for theirability to kill target cells without prior stimulation (1). Itis currently recognized that these cells have potent an-
tileukemic and antiviral activities (2, 3). Human NK cells areidentified by the CD32CD56+ phenotype. Different NK cellsubpopulations are determined according to the presence anddensity of CD56 and CD16 (FcgRIII) surface molecules. Low-density CD56 (CD56dim) subsets constitute .90% of peripheralblood NK cells, which also express perforin and killer Ig-likereceptors (KIR). These cells, which express the CD16 marker,are involved in Ab-dependent cellular cytotoxicity (ADCC). Thesubset of CD56bright NK cells, which are rare in blood but pre-dominant in lymph nodes and other tissues, do not expressperforin and KIR (4). This CD56bright subset exhibits immuno-regulatory functions through the secretion of various cytokines(i.e., IFN-g, TNF-a, or IL-10) in response to monokine stimula-tion. In contrast, the CD56dim cells are highly cytotoxic and
preferentially produce cytokines after recognition of target cells(5, 6).It is currently accepted that CD56bright NK cells are the pre-
cursors of the CD56dim NK cells, a concept originally hypothe-sized by Lanier et al. (7) in 1986. Although in vitro differentiationassays could not directly confirm this theory, many observationssupported it. Indeed, the CD56bright population displays longertelomeres than the CD56dim NK cells, suggesting that they haveproliferated less (8). In contrast with CD56dim cells, CD56bright
cells express high levels of CD117 but do not express CD57, tworeceptors expressed on progenitor and senescent cells, respectively(9–11). NK cell differentiation from hematopoietic stem cell pre-cursors primarily gives rise to CD56brightCD162KIR2 NK cells(12–14). The CD56bright NK cells are the first lymphocytes, whichappear following hematopoietic stem cell transplantations (HSCT)(15–18). Thus, HSCT is a good in vivo model to study hemato-poietic cell differentiation, in particular NK cells. Dulphy et al. (15)have recently described an unusual CD56brightCD16+ subpopulationthat was transiently increased 3 mo after identical HSCT. Thissubset was originally described in healthy individuals (19).The aim of the current study was to describe more precisely the
CD56brightCD16+ NK subset and determine its role during NK celldifferentiation. Our analysis of their function, phenotype, andfrequencies during aging or after unrelated cord blood trans-plantation (UCBT), together with in vitro NK differentiationstudies, strongly suggest that the CD56brightCD16+ NK cell subsetis a functional intermediate between the CD56bright and CD56dim
cells. Furthermore, we demonstrate that CD16 acquisition occursin the presence of autologous T CD3+ cells.
Materials and MethodsPatients and donors
Twenty-five patients (median age 43.5 y) underwent UCBT between 2005and 2008 at either the Pitie-Salpetriere or Hotel Dieu hospitals (Paris,
*INSERM Unite Mixte de Recherche-S 945, Hopital Pitie-Salpetriere, 75013 Paris,France; †Universite Pierre et Marie Curie, 75013 Paris, France; and ‡Service d’Hem-atologie Clinique, Hopital Pitie-Salpetriere, 75013 Paris, France
Received for publication January 31, 2011. Accepted for publication April 12, 2011.
This work was supported by a grant from INSERM under the Programme Hospitalierde Recherche Clinique Minicord (P060206-AOM06206) and by the AssociationCent pour Sang la Vie and the Agence de la BioMedecine.
Address correspondence to Dr. Vincent Vieillard, INSERM Unite Mixte de Re-cherche-S 945, Laboratoire Immunite et Infection, Hopital Pitie-Salpetriere, 83 Bou-levard de l’Hopital, 75013 Paris, France. E-mail address: [email protected]
The online version of this article contains supplemental material.
Abbreviations used in this article: ADCC, Ab-dependent cell cytotoxicity; AP-HP,Assistance Publique-Hopitaux de Paris; DC, dendritic cell; HSCT, hematopoieticstem cell transplantation; iDC, immature dendritic cell; ILT-2, Ig-like transcript 2;KIR, killer Ig-like receptor; mDC, mature dendritic cell; UCBT, unrelated cord bloodtransplantation.
Copyright� 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00
www.jimmunol.org/cgi/doi/10.4049/jimmunol.1100330
Published May 9, 2011, doi:10.4049/jimmunol.1100330 by guest on D
France) for high-risk hematopoietic malignancies. The cohort of thesepatients, mainly of Caucasian origin (24 out of 25), was previously de-scribed (18). Briefly, acute myeloid leukemia was the most common di-agnosis (14 out of 25). All patients received a reduced-intensityconditioning regimen; 20 patients received the Minneapolis protocolcontaining cyclophosphamide 50 mg/kg at day 6, fludarabine 200 mg/m2
for 5 d, and total body irradiation 2 Gy. For the last five patients, total bodyirradiation was replaced by 140 mg/m2 melphalan. Cyclosporine andmycophenolate mofetyl were given 3 d before transplantation to intensifythe immunosuppressive conditioning and prevent graft-versus-host disease.Generic HLA-A, HLA-B, and allelic HLA-DR typing was used in thematching strategy; most of the cord blood pairs were 4 out of 6 matches.Adult controls and cord blood samples were obtained, respectively, fromthe Etablissement Francais du sang and the obstetrics department of thePitie-Salpetriere Hospital. All adult donors used for the phenotyping;functional and differentiation assays were aged between 18 and 60 y old.The rheumatology and gerontology departments of Pitie-Salpetriere orCharles-Foix Hospitals furnished 33 samples from individuals aged$60 y,17 aged 60–80 y (mean age 68.86 7.2 y), and 26.80 y (mean age 87.164.9 y). All volunteers affirmatively stated, and their medical records con-firmed, that they had no infectious, malignant, or autoimmune diseasesduring the 6 mo before the study and were without acute illnesses at thetime of the sampling, but could present treatment for classical age-relatedpathologies as previously described (20). Patients and donors providedinformed consent in compliance with the ethics committee guidelinesbefore peripheral blood samples were collected for the study.
Flow cytometry
Phenotypes were realized on whole blood. Cells were stained using theappropriate Ab mixture: anti-CD3 (UCHT1; Beckman Coulter), anti-CD56(N901; Beckman Coulter), anti-CD16 (3G8; Beckman Coulter; or VEP13;Miltenyi Biotec); anti-CD94 (HP-3D9; BD Biosciences); anti-KIR2DL1/KIR2DS1 (EB6B; Beckman Coulter); anti-KIR2DL2/KIR2DL3/KIR2DS2 (GL183; Beckman Coulter); anti-KIR3DL1/KIR3DS1 (Z27;Beckman Coulter), anti-CD159a/NKG2A (Z199; Beckman Coulter); anti-CD85j/Ig-like transcript-2 (ILT-2) (HP-F1; Beckman Coulter); anti-CD8a(T8; Beckman Coulter), anti-CD117 (104D2D1; Beckman Coulter), anti-CD25 (PC61; BD Biosciences), anti-CD27 (B1.49.9; Beckman Coulter),anti-CD127 (R34.34; Beckman Coulter), anti–CD62-L (MEL-14; BDBiosciences), and anti-CX3CR1 (2A9-1; Biolegend). FACS lysing solution(BD Biosciences) was used to lyse erythrocytes. Intracellular stainingswere performed after permeabilization (0.1% saponin, 0.5% BSA, PBS 13solution) using appropriate Abs: anti-granzyme A (CB9; BD Biosciences),anti-granzyme K (GM6C3; Santa Cruz Biotechnology), anti-perforin(dG9; Abcam), or anti-granzyme B (Gb11; Abcam). Acquisitions wereperformed on FC500 (Beckman Coulter), LSRII, or FACSCanto (BDBiosciences) flow cytometers, depending on the experiments.
Cell sorting
NK cells from freshly isolated PBMCs were negatively sorted using theMACS NK Cells Isolation kit (reference number 130.092.657; MiltenyiBiotec). NK cells were stained with an anti–CD3-ECD (UCHT1; BeckmanCoulter), anti–CD56-PC7 (N901; Beckman Coulter), and a nonblockinganti–CD16-FITC (VEP13; Miltenyi Biotec), and the different cellularsubsets were sorted on an FACSAria (BD Biosciences). Purified cells werethen used for [51Cr] release or in vitro differentiation assays. For differ-entiation assays, T cells were magnetically purified using CD3 microbeads(reference number 130.050.101; Miltenyi Biotec). In some experiments,CD4+, CD8+, and CD56+ T cell subsets were sorted on an FACSAria (BDBiosciences) after staining with CD3-ECD (UCHT1; Beckman Coulter),CD4-PE (13B8.2; Beckman Coulter), CD8-FITC (T8; Beckman Coulter),and CD56-PC7 (N901; Beckman Coulter).
NK degranulation, ADCC, and cytolytic assays
NK cell subset’s cytolytic activity was assessed in standard 4- h [51Cr]release assays against K562 or RAJI target cells with a 5:1 E:T ratio, asdescribed (18). ADCC experiments were performed against RAJI cells inthe presence of 1 mg/ml anti-CD20 (rituximab; Roche). Degranulationassays were tested by CD107a-PC5 (H4A3; BD Biosciences) detection, asdescribed (18). Importantly, staining with the nonblocking anti-CD16 mAbVEP13 clone was performed before the CD107a assays.
Intracellular analysis of IFN-g production
PBMCswerefirst stainedwithCD16-FITC(cloneVEP-13)and incubated for6 h in the presence of 10 ng/ml IL-12 and 100 ng/ml IL-18 at 37˚C and 5%CO2. Cells were thereafter stained for CD3 and CD56, fixed (BD Cell Fix;
BD Biosciences), and permeabilized (PBS/0.5% BSA/0.1% saponin) beforestaining for intracellular IFN-g (B27; BD Biosciences) expression. Acqui-sitions were performed on an LSRII (BD Biosciences) flow cytometer.
In vitro differentiation assays
Ten thousand FACS-sorted CD56brightCD162 or CD56brightCD16+ NK cellswere cultured in RPMI 1640 (Life Technologies), supplemented with 10%human serum AB (BioWest), nonessential amino acids (13; Life Tech-nologies), antibiotic/antimycotic (13; Life Technologies), and IL-2 (103
U/ml, proleukin; Roche). Cultures were performed in 96-well U-bottomplates (BD Falcon). Depending on the assay, 105 autologous CD3+ T cellswere added to NK cell cultures. In some assays, CD3+ T cells werereplaced by CD3+CD4+, CD3+CD8+, or CD3+CD56+ lymphocyte T sub-sets. When specified, T cells were physically separated from NK cells withtranswells (eight-well strip insert; Nunc). Cytokines and culture mediawere renewed every 2 or 3 d. NK cells were characterized at indicated timepoints using the same panel of Abs as those used for the cell sorting.
Statistics
All statistical analyses were performed using Prism 5 software (GraphPad,San Diego, CA). Nonparametric Wilcoxon and Mann–Whitney tests wereused for paired and unpaired data comparisons, respectively. Repeated-measures ANOVA with Tukey posttest for p value calculation were per-formed for multiple comparisons of paired data. Kruskal–Wallis test withthe Dunn posttest for p value calculation was performed for multiplecomparisons of independent groups. Significance is defined by a pvalue ,0.05 using two-tailed tests. *p , 0.05, **p , 0.01, ***p , 0.001.
ResultsKinetics of CD56brightCD16+ NK cell repopulation followingUCBT
The phenotypic characterization of NK cells following UCBTrevealed that the CD56brightCD16+ NK cell subset is highlyexpressed among whole NK cells following the transplantation.The CD56brightCD162 subset increases very rapidly just afterneutrophil engraftment, at 1-mo post-UCBT, and then graduallydecreases during the time examined (Fig. 1A). Interestingly,CD56brightCD16+ NK cells were present at 1 mo posttransplan-tation, but increased to their maximum frequency 2 mo afterUCBT and remained stable during the following 12 mo (Fig. 1B).These observations suggested that the CD56brightCD162 cells aregenerated early posttransplantation and may be the precursors ofthe CD56brightCD16+ subset.
Phenotypic characterization of CD56brightCD162, CD56bright
CD16+, and CD56dimCD16+ NK cells in healthy individuals
To precisely determine the role of the CD56brightCD16+ NK cells,we performed an extensive phenotypic comparison with CD56bright
CD162 and CD56dimCD16+ NK cells in healthy donors. We firstlyobserved that CD56brightCD162 and CD56brightCD16+ NK cellsexpressed numerous receptors in common with CD56dimCD16+
cells, such as NKp30, NKp46, DNAM-1, 2B4, LAIR-1, andNKG2D (data not shown). However, some markers discriminatedCD56bright from CD56dim NK cells, regardless of CD16 expression,such as CD94, NKG2A, CD127, CD27, CD62L, ILT-2, granzymeB, granzyme K, and perforin (Supplemental Fig. 1). More impor-tantly, Fig. 2 shows that CD117 and CD25 were highly expressedin CD56brightCD162 NK cells, and their expression progressivelydecreased in CD56brightCD16+ and CD56dimCD16+ NK cells.Conversely, the expression of CD8, granzyme A, CX3CR1, andpan-KIR progressively accumulated when looking sequentially atthe NK cell subsets from CD56bright to CD56dim cells. Of note, highexpression of KIR in the CD56brightCD16+ subset compared withthe CD56brightCD162 subset was also demonstrated individuallyfor each KIR tested: KIR2DL1/DS1, KIR2DL2/DL3/DS2, andKIR3DL1 (Supplemental Fig. 2). Together, these data suggest thatCD56brightCD16+ cells may be an intermediate between CD56bright
CD162 and CD56dimCD16+ NK cells.
2 CD56BRIGHTCD16+ NK CELLS: INTERMEDIATE OF NK DIFFERENTIATION
CD56brightCD16+ NK cells are fully functional for cytolyticfunction
To characterize functional properties of NK cells, regardless ofCD16 expression, preliminary experiments were performed todetermine appropriate experimental conditions. Indeed, CD16 isquickly downmodulated after encounter with K562 or RAJI targetcells, which renders discrimination of the different subsets afterdegranulation assays impossible (Supplemental Fig. 3) (21, 22).Furthermore, the anti-CD16 3G8 clone is a blocking Ab of ADCC,which does not allow cell sorting for [51Cr] release assays (datanot shown). For these reasons, we performed all experiments inthe presence of the anti-CD16 VEP13 clone, which does not blockADCC. Fig. 3A shows that in the presence of the anti-CD16 mAbVEP13 clone, the expression of CD16 on NK cells is preservedafter encountering of target cells, which allows analysis of NKsubsets after degranulation assays (Fig. 3A).We next compared the cytolytic capacities of CD56brightCD162,
CD56brightCD16+, and CD56dimCD16+ NK cell subsets in the
presence of VEP13 anti-CD16 mAb. Fig. 3B shows that theCD56bright NK cells, expressing or not CD16, displayed similardegranulation ability against K562 target cells (p , 0.001 in bothcases). However, chromium release assays revealed thatCD56brightCD162 cells were significantly less cytotoxic thanCD56brightCD16+ (p , 0.01) and CD56dimCD16+ (p , 0.05) NKcell subsets against K562 cells, whereas CD56brightCD16+ andCD56dimCD16+ cells had similar direct cytolytic function (Fig.3C). The high lytic ability of CD56brightCD16+ cells was con-firmed by ADCC assays against RAJI cells covered with anti-CD20 mAb (rituximab). Indeed, both in degranulation and chro-mium release assays, CD56brightCD16+ cells were as efficient asCD56dimCD16+ cells (Fig. 3E, 3F). In contrast, CD56brightCD162
cells could neither degranulate nor kill RAJI target cells in thepresence of anti-CD20 mAb, in accordance with their lack ofCD16 expression. Additionally, we demonstrated after IL-12/IL-18 stimulation that both CD56brightCD162 and CD56brightCD16+
NK cell subsets produce large and equivalent amounts of IFN-gcompared with the low production of the CD56dim subset (Sup-plemental Fig. 4). Together, these data show that CD56brightCD16+
NK cells contain more cytotoxic properties than CD56brightCD162
cells but also maintain the full ability to produce IFN-g aftercytokines stimulation.
CD3+ T cells drive CD16 acquisition and subsequent ADCCability of CD56brightCD162 NK cells
We next performed in vitro differentiation assays to furthercharacterize the CD56brightCD162 to CD56brightCD16+ differen-tiation. These experiments were performed with highly purifiedCD56brightCD162 cells. Fig. 4A (left panels) shows that IL-2 alonewas not able to drive CD16 expression on CD56brightCD162 NKcells after 7 or 14 d of culture. By contrast, in the presence ofautologous purified CD3+ T cells, a significant fraction ofCD56brightCD162 cells acquired CD16 (Fig. 4A, middle panels).To determine the role of the cell–cell contact requirements, similarexperiments were performed in transwell plates. As shown in Fig.4A (right panels), in the absence of cellular contacts, theCD56brightCD162 NK cells showed reduced CD16 acquisition.We next evaluated the efficacy of CD4+, CD8+, and CD3+CD56+
T cell subsets to induce CD16 acquisition by CD56brightCD162
NK cells. In the presence of all of these CD3+ T cell subsets,similar proportions of NK cells expressing CD16 were observed(Fig. 4B). Importantly, kinetic studies revealed that in the presenceCD3+ T cells, expression of CD16 increased until day 14, when itreached a maximum level of between 25 and 30% (Fig. 4C). In anattempt to increase CD16+ expression on NK cells, we next per-formed experiments in the presence of autologous dendritic cells(DC). Immature DCs (iDC) were derived from purified CD14monocytes in the presence of IL-4 and GM-CSF, whereas matureDCs (mDC) were obtained from iDC pulsed with LPS or bacillusCalmette–Guerin. Both iDC and mDC were cultured with orwithout CD3+ T cells to induce CD16 expression on CD56bright
CD162 cells. Unfortunately, the presence of iDC or mDC had noeffect on CD16 expression (data not shown).To determine the differentiation state of CD16+ cells generated
in vitro in the presence of CD3+ T cells, we studied the expressionof major differentiation markers including KIR, NKG2A, ILT-2,and CD62L (Fig. 4D). We demonstrated that CD16+ NK cellsexpressed a significantly higher level of KIR than the CD162
subset at all culture times examined, although this level remainslow (∼10%). Concomitantly, NKG2A remained highly expressedon both subsets, whereas CD62L and ILT-2 decreased and in-creased, respectively, during the culture, independently of CD16expression.
FIGURE 1. Frequency of CD56brightCD16- and CD56brightCD16+ among
NK cells after UCBT. Box and whiskers plots of CD56brightCD162 (A) or
CD56brightCD16+ (B) frequencies among NK cells 1 (M1), 2 (M2), 3 (M3),
6 (M6), and 12 (M12) mo after UCBT compared with healthy donors (Ctl)
and cord blood samples (CB). C, Representative dot plots gated on CD32
ressively increased to ∼80% of the cells after 28 d of culture. Whendifferentiation markers were examined on purified CD56bright
CD16+ or CD56brightCD162 NK cells (from matched healthydonors) in the presence of CD3+ T cells, similar profiles wereobtained for NKG2A, CD62L, and ILT-2 (Fig. 5B). In contrast,
higher KIR expression was observed in in vitro-culturedCD56brightCD16+ cells compared with CD56brightCD162 NKcells (Fig. 5B).
Aging is associated with the accumulation of CD16+ NK cellsamong the CD56bright subset
Finally, we performed an analysis of CD16 expression among NKcells during aging. It is already known that the aging process isassociated with fewer CD56bright NK cells and an accumulation ofCD56dim NK cells in peripheral blood (20, 23). We showed thatthe percentage of CD56bright NK cells expressing CD16 increaseslinearly with aging (Fig. 6A). Indeed, when comparing people’saging with ,60 and .80 y old, we observed that older personsexpressed 2–3-fold more CD16 in their CD56bright compartment(p , 0.001). Thus, younger donors expressed CD16 on ∼25% ofCD56bright NK cells, whereas elderly people .80 y old expressedthis marker on ∼50–75% of these cells. Fig. 6B and 6C show thatthis change is principally due to a decrease of the CD56bright
CD162 (p , 0.01) and an increase of CD56brightCD16+ cellcounts (p , 0.01). These data clearly show that old age is asso-ciated with fewer CD56brightCD162 cells and an accumulation ofboth CD56brightCD16+ and CD56dimCD16+ NK cells in the pe-riphery.
DiscussionThe present study presents evidence that CD56brightCD16+ NK cellsare a functional differentiation intermediate between CD56bright
and CD56dim cells. A recent paper of Dulphy et al. (15) showedthat 3 mo after matched HSCT, the frequency of CD56brightCD16+
NK cells was largely increased among NK cells. In this study, weconfirm this observation and additionally show that the maximumlevel of CD56brightCD16+ into NK cells is observed later than thatof CD56brightCD162. This suggests that the CD56brightCD16+ cellsmay be an intermediate between CD56brightCD162 and CD56dim
CD16+ NK cell subsets. Concomitantly, in elderly individuals,more CD56dimCD16+ cells associated with aging were previouslyreported (20, 23, 24). More importantly, in this study, we havedetermined the percentage of CD56bright cells expressing CD16and observed that it is inversely associated with aging. These datasuggest that old age favors an accumulation of more mature NKcell subsets, such as CD56brightCD16+ and CD56dimCD16+ cells, indetriment to the CD56brightCD162 precursor cells, as previouslydescribed for T and B cells (25). It was observed in older indi-viduals both increasing of soluble IL-2R, which downregulated IL-2 activity (26), and IL-12p40 homodimers, which could act as anIL-12 antagonist (27). This could suggest that the disturbance ofspecific cytokine signaling could partially block the final matura-tion of NK cells at a stage of CD56brightCD16+ NK cells, which areaccumulated in older subjects.To minimize this effect due to particular clinical situations, the
main experiments of this study were performed in healthy donors.The CD56brightCD16+ NK cell subpopulation represents 1.7 61.6% of NK cells. An extensive phenotypic analysis revealed fewdifferences between CD56brightCD162 and CD56brightCD16+
cells. CD56brightCD16+ NK cells showed intermediate expressionlevels of CD25, CD117, CD8, CX3CR1, KIRs, and granzyme Abetween the CD56brightCD162 and the CD56dimCD16+ cells. Thiswas in accordance with Caligiuri’s model (28), which predictsCD117 loss and KIR acquisition during stage 4 (CD56bright) tostage 5 (CD56dim) transition. Specific cytolytic molecules alsorevealed the maturity of CD56brightCD16+ NK cells. Indeed, theCD56brightCD162 NK cells expressed almost exclusively gran-zyme K, whereas CD56brightCD16+ additionally overexpressedgranzyme A. Finally, CD56dimCD16+ NK cells lost granzyme K
FIGURE 3. CD56brightCD16+ NK cells efficiently kill target cells. A,
Representative samples of CD107a expression on CD56bright, and CD56dim
NK cells in absence (w/o target) or in the presence of K562 or RAJI (plus
rituximab) target cells. Analysis is gated on CD32CD56+ NK cells in the
presence of anti-CD16 VEP-13 Ab clone. Box and whiskers plots of
CD107 expression (B, D) or chromium release assay (C, E) using purified
CD56brightCD162 (B16-), CD56brightCD16+ (B16+), or CD56dimCD16+
(Dim) NK cell subsets from 11 healthy donors against K562 (B, C) or
and expressed abundant levels of granzyme A, granzyme B, andperforin. This succession of granzymes in NK cell subsets issimilar to what is observed during memory CD8+ T cells differ-entiation (29).Importantly, we show that CD56brightCD16+ and CD56dim
CD16+ cells have similar cytotoxic functions, greater than those of
CD56brightCD162 precursor cells. Furthermore, the progressiveincrease of CX3CR1 expression on CD56brightCD162, CD56bright
CD16+, and CD56dimCD16+ NK cell subsets suggests an evolutionin the acquisition of migratory properties to inflammations sites(30). Remarkably, we also found that both CD56brightCD162 andCD56brightCD16+ cells displayed equivalent high abilities to
FIGURE 4. T cells permit CD16 expression and subsequent acquisition of ADCC ability by CD56brightCD162 NK cells. A, Purified CD56brightCD162
NK cells were cultured with IL-2 alone (IL-2) in the presence of IL-2 plus autologous T cells (IL-2 + CD3) or in a Transwell assay (IL-2 + TW-CD3).
Expression of CD16 was monitored 7 and 14 d after the beginning of culture and gated on CD32CD56+ NK cells. B, Representative patterns of three
independent experiments realized on purified CD56brightCD162 NK cells cultured in the presence of autologous, purified CD4+CD562, CD8+CD562, or
CD3+CD56+ T cells in direct contact (Mix) or in a Transwell assay. Experiments were performed at 14 d postculture and gated on CD32CD56+ NK cells. C,
Box and whiskers plots of 12 different cultures (left panel) and representative pictures (right panel) of CD16 expression on CD56brightCD162 NK cells
coculture with CD3+ T cells at different time points postculture. D, Box and whiskers plots of differentiation markers expression on CD162 (white box) and
CD16+ (gray box) NK cell subsets occurring after cocultures of CD56brightCD162 with CD3+ T cells during 7 (d7), 14 (d14), and 28 d (d28). Analyses were
based on CD32CD56+ gated NK cells of 12 different cultures. E, CD107a degranulation assays of cultured CD56brightCD162 NK cells against K562 or
RAJI 6 1 mg/ml rituximab target cells. Degranulation assays were performed after 28 d of coculture with autologous CD3+ T cells. These data are
representative of three independent experiments. *p , 0.05, **p , 0.01.
6 CD56BRIGHTCD16+ NK CELLS: INTERMEDIATE OF NK DIFFERENTIATION
produce IFN-g after an IL-12/IL-18 stimulation compared withCD56dimCD16+ cells. Altogether, these data demonstrate thatCD56brightCD16+ NK cells share functional properties of theCD56brightCD162 and CD56dimCD16+ subsets and suggest that
CD16 acquisition could be used as a measure of NK cells’ func-tional maturity.These ex vivo data were confirmed by in vitro NK cell differ-
entiation experiments. Our data show that culture of CD56bright
FIGURE 5. Modulation of NK receptor on CD56brightCD162 or CD56brightCD16+ NK cells cultured in the presence of autologous T cells. A, Box and
whiskers plots of 12 different cultures (left panel) and representative pictures (right panel) of CD16 expression on CD56brightCD16+ NK cells after 0, 7 (d7),
14 (d14), and 28 d (d28) of cocultures with CD3+ T cells. Analysis are based on CD32CD56+ gated NK cells. B, Evolution of differentiation markers on
CD56brightCD162 or CD56brightCD16+ NK cells after 7 (d7), 14 (d14), and 28 d (d28) of coculture of with CD3+ T cells. Data are represented as mean and
SD of three independent cultures. *p , 0.05, **p , 0.01, ***p , 0.001.
FIGURE 6. Old age is associated with a decrease of CD56brightCD162 cells and an accumulation of CD56brightCD16+ and CD56dimCD16+ NK cell
subsets. A, Expression of CD16 on CD56bright NK cell subsets from donors ,60 (n = 29; 18 , x , 60) and between 60 and 80 (n = 22; 60 , x , 80), or
.80 (n = 17; x. 80) y old. B, Representative patterns of two independent donors gated on CD32CD56+ NK cells. C, Absolute values of CD56brightCD162,
CD56brightCD16+, and CD56dimCD16+ NK cell subsets from donors ranged between 18 and 60 (n = 20; 18, x, 60), between 60 and 80 (n = 22; 60, x,80), or .80 (n = 17; x . 80) y old. *p , 0.05, **p , 0.01, ***p , 0.001.
CD162 cells with IL-2 alone is not sufficient to induce CD16expression, as previously reported (19). This contrasts with resultsreported by two other groups, describing an overexpression ofCD16 after culture with IL-2 alone (4, 8). This discrepancy couldbe explained by cell-sorting conditions, as these groups sorted allCD56bright NK cells, regardless of CD16 expression. However, ourresults are in line with the findings of in vitro NK cell differen-tiation from CD34+ stem cells (31). Freud et al. (31) showed thatin medium containing IL-2 or IL-15, it was only possible togenerate a CD56bright-like subset, expressing no CD16 and KIR,but high amounts of CD117.More importantly, we provide solid evidence that CD3+ T cells
are a key component to drive acquisition of functional CD16during the CD56brightCD162 to CD56brightCD16+ transition. Thesedata are in line with Fehniger et al. (32), who showed thatCD56bright NK cells were found in the parafollicular T area oflymph nodes in direct contact with CD3+ T cells. Additionally,Freud et al. (31) demonstrated that autologous activated T cellswere able to induce the differentiation of CD34dimCD45RA+
integrin b7+ stem cells into CD56bright NK cells. More recently,we have shown that after haploidentical HSCT, NK cell re-constitution was positively influenced by the amount of T cellscontained in the graft (33). This work also showed that NK–T cellinteractions seemed to drive CD16 acquisition, although a slightexpression of this marker was also detected in transwells. Thissuggests that CD56brightCD162 cells gave rise to CD56brightCD16+
not only under the influence CD3+ T cell contacts, but also fromT cell-derived cytokines. Our results also indicate that theseCD16+ NK cells expressed significantly more KIR than theCD162 subpopulation. However, in CD56bright CD16+ cells, thislow KIR expression is coupled to high levels of NKG2A, sug-gesting that they are closer to the CD56brightCD16+ than theCD56dimCD16+ NK cell subset. Intriguingly, ILT-2, a markerspecifically expressed on the CD56dimCD16+ subset (34), is sim-ilarly expressed on CD162 and CD16+ NK cells after in vitroculture. This could reflect long-lasting activation/proliferationinstead of a differentiation of cells, as previously shown forCD62L (35). As expected, purified CD56brightCD16+ NK cellsretained a greater ability to express CD16 and KIR after culture inthe presence of autologous IL-2–activated CD3+ T cells. Indeed,despite a decreased expression of CD16 at early time points of theculture, CD16 was expressed by .80% of NK cells after 28 d ofculture. Early decreases in CD16 expression might be explainedby an extracellular metalloprotease activity, as previously de-scribed (21, 22).Recently, we and others suggested that the CD56dim subset is
composed from several differentiation stages, but only few studiesinvestigated the CD56bright subset (36–38). Our data obtainedin vitro and ex vivo from UCBT and elderly individuals stronglysuggested that the CD56brightCD16+ subset might be considered asa functional differentiation intermediate between CD56bright
CD162 and CD56dimCD16+ cells. These CD56brightCD16+ cellspossess equivalent cytolytic ability to the CD56dimCD16+ cellswhile conserving a phenotype closer to the CD56brightCD162
subset. Finally, our work highlights that CD3+ T cells may playa key role in NK cell differentiation, favoring the differentiation ofCD56brightCD162 NK cells into a later stage that expresses CD16and is highly functional.
AcknowledgmentsWe thank Dr. V. Siguret (Assistance Publique-Hopitaux de Paris [AP-HP]
Hopital Charles-Foix, Service de Geriatrie, Ivry-sur-Seine, France) and
Prof. F. Gandjbakhch (AP-HP Hopital Pitie-Salpetriere, Service de Rhu-
matologie, Paris, France) for the recruitment of aging subjects and Dr. B.
Rio (AP-HP Hotel-Dieu, Paris, France) for implication in the Minicord
Program of UCBT. We also thank the personnel from the Etablissement
Francais du Sang for healthy adult blood samples and those from the
Department of Gynecologie-Obstetrique at the Pitie-Salpetriere Hospital
(Paris, France) for cord blood samples.
DisclosuresThe authors have no financial conflicts of interest.
References1. Lanier, L. L., J. H. Phillips, J. Hackett, Jr., M. Tutt, and V. Kumar. 1986. Natural
killer cells: definition of a cell type rather than a function. J. Immunol. 137:2735–2739.
2. Biron, C. A., K. B. Nguyen, G. C. Pien, L. P. Cousens, and T. P. Salazar-Mather.1999. Natural killer cells in antiviral defense: function and regulation by innatecytokines. Annu. Rev. Immunol. 17: 189–220.
3. Wu, J., and L. L. Lanier. 2003. Natural killer cells and cancer. Adv. Cancer Res.90: 127–156.
4. Ferlazzo, G., D. Thomas, S. L. Lin, K. Goodman, B. Morandi, W. A. Muller,A. Moretta, and C. Munz. 2004. The abundant NK cells in human secondarylymphoid tissues require activation to express killer cell Ig-like receptors andbecome cytolytic. J. Immunol. 172: 1455–1462.
5. Bryceson, Y. T., C. Fauriat, J. M. Nunes, S. M. Wood, N. K. Bjorkstrom,E. O. Long, and H. G. Ljunggren. 2010. Functional analysis of human NK cellsby flow cytometry. Methods Mol. Biol. 612: 335–352.
6. Cooper, M. A., T. A. Fehniger, S. C. Turner, K. S. Chen, B. A. Ghaheri,T. Ghayur, W. E. Carson, and M. A. Caligiuri. 2001. Human natural killer cells:a unique innate immunoregulatory role for the CD56(bright) subset. Blood 97:3146–3151.
7. Lanier, L. L., A. M. Le, C. I. Civin, M. R. Loken, and J. H. Phillips. 1986. Therelationship of CD16 (Leu-11) and Leu-19 (NKH-1) antigen expression on hu-man peripheral blood NK cells and cytotoxic T lymphocytes. J. Immunol. 136:4480–4486.
8. Romagnani, C., K. Juelke, M. Falco, B. Morandi, A. D’Agostino, R. Costa,G. Ratto, G. Forte, P. Carrega, G. Lui, et al. 2007. CD56brightCD16- killer Ig-like receptor- NK cells display longer telomeres and acquire features ofCD56dim NK cells upon activation. J. Immunol. 178: 4947–4955.
9. Matos, M. E., G. S. Schnier, M. S. Beecher, L. K. Ashman, D. E. William, andM. A. Caligiuri. 1993. Expression of a functional c-kit receptor on a subset ofnatural killer cells. J. Exp. Med. 178: 1079–1084.
10. Brenchley, J. M., N. J. Karandikar, M. R. Betts, D. R. Ambrozak, B. J. Hill,L. E. Crotty, J. P. Casazza, J. Kuruppu, S. A. Migueles, M. Connors, et al. 2003.Expression of CD57 defines replicative senescence and antigen-induced apo-ptotic death of CD8+ T cells. Blood 101: 2711–2720.
11. Lopez-Verges, S., J. M. Milush, S. Pandey, V. A. York, J. Arakawa-Hoyt,H. Pircher, P. J. Norris, D. F. Nixon, and L. L. Lanier. 2010. CD57 definesa functionally distinct population of mature NK cells in the humanCD56dimCD16+ NK-cell subset. Blood 116: 3865–3874.
12. Lotzova, E., C. A. Savary, and R. E. Champlin. 1993. Genesis of humanoncolytic natural killer cells from primitive CD34+CD33- bone marrow pro-genitors. J. Immunol. 150: 5263–5269.
13. Mrozek, E., P. Anderson, and M. A. Caligiuri. 1996. Role of interleukin-15 in thedevelopment of human CD56+ natural killer cells from CD34+ hematopoieticprogenitor cells. Blood 87: 2632–2640.
14. Mingari, M. C., C. Vitale, C. Cantoni, R. Bellomo, M. Ponte, F. Schiavetti,S. Bertone, A. Moretta, and L. Moretta. 1997. Interleukin-15-induced maturationof human natural killer cells from early thymic precursors: selective expressionof CD94/NKG2-A as the only HLA class I-specific inhibitory receptor. Eur. J.Immunol. 27: 1374–1380.
15. Dulphy, N., P. Haas, M. Busson, S. Belhadj, R. Peffault de Latour, M. Robin,M. Carmagnat, P. Loiseau, R. Tamouza, C. Scieux, et al. 2008. An unusual CD56(bright) CD16(low) NK cell subset dominates the early posttransplant periodfollowing HLA-matched hematopoietic stem cell transplantation. J. Immunol.181: 2227–2237.
16. Nguyen, S., N. Dhedin, J. P. Vernant, M. Kuentz, A. Al Jijakli, N. Rouas-Freiss,E. D. Carosella, A. Boudifa, P. Debre, and V. Vieillard. 2005. NK-cell re-constitution after haploidentical hematopoietic stem-cell transplantations: im-maturity of NK cells and inhibitory effect of NKG2A override GvL effect. Blood105: 4135–4142.
17. Jacobs, R., M. Stoll, G. Stratmann, R. Leo, H. Link, and R. E. Schmidt. 1992.CD16- CD56+ natural killer cells after bone marrow transplantation. Blood 79:3239–3244.
18. Beziat, V., S. Nguyen, S. Lapusan, B. Hervier, N. Dhedin, D. Bories, M. Uzunov,A. Boudifa, H. Trebeden-Negre, F. Norol, et al. 2009. Fully functional NK cellsafter unrelated cord blood transplantation. Leukemia 23: 721–728.
19. Nagler, A., L. L. Lanier, S. Cwirla, and J. H. Phillips. 1989. Comparative studiesof human FcRIII-positive and negative natural killer cells. J. Immunol. 143:3183–3191.
20. Le Garff-Tavernier, M., V. Beziat, J. Decocq, V. Siguret, F. Gandjbakhch,E. Pautas, P. Debre, H. Merle-Beral, and V. Vieillard. 2010. Human NK cellsdisplay major phenotypic and functional changes over the life span. Aging Cell9: 527–535.
8 CD56BRIGHTCD16+ NK CELLS: INTERMEDIATE OF NK DIFFERENTIATION
21. Mota, G., I. Moldovan, A. Calugaru, M. Hirt, E. Kozma, C. Galatiuc,L. Brasoveanu, and G. Boltz-Nitulescu. 2004. Interaction of human immuno-globulin G with CD16 on natural killer cells: ligand clearance, FcgammaRIIIAturnover and effects of metalloproteinases on FcgammaRIIIA-mediated binding,signal transduction and killing. Scand. J. Immunol. 59: 278–284.
22. Grzywacz, B., N. Kataria, and M. R. Verneris. 2007. CD56(dim)CD16(+) NKcells downregulate CD16 following target cell induced activation of matrixmetalloproteinases. Leukemia 21: 356–359, author reply 359.
23. Borrego, F., M. C. Alonso, M. D. Galiani, J. Carracedo, R. Ramirez, B. Ostos,J. Pena, and R. Solana. 1999. NK phenotypic markers and IL2 response in NKcells from elderly people. Exp. Gerontol. 34: 253–265.
24. Sansoni, P., A. Cossarizza, V. Brianti, F. Fagnoni, G. Snelli, D. Monti,A. Marcato, G. Passeri, C. Ortolani, E. Forti, et al. 1993. Lymphocyte subsetsand natural killer cell activity in healthy old people and centenarians. Blood 82:2767–2773.
25. Gruver, A. L., L. L. Hudson, and G. D. Sempowski. 2007. Immunosenescence ofageing. J. Pathol. 211: 144–156.
26. McNerlan, S. E., I. M. Rea, H. D. Alexander, and T. C. Morris. 1998. Changes innatural killer cells, the CD57CD8 subset, and related cytokines in healthy aging.J. Clin. Immunol. 18: 31–38.
27. Rea, I. M., S. E. McNerlan, and H. D. Alexander. 2000. Total serum IL-12 andIL-12p40, but not IL-12p70, are increased in the serum of older subjects; re-lationship to CD3(+)and NK subsets. Cytokine 12: 156–159.
28. Freud, A. G., and M. A. Caligiuri. 2006. Human natural killer cell development.Immunol. Rev. 214: 56–72.
29. Bratke, K., M. Kuepper, B. Bade, J. C. Virchow, Jr., and W. Luttmann. 2005.Differential expression of human granzymes A, B, and K in natural killer cellsand during CD8+ T cell differentiation in peripheral blood. Eur. J. Immunol. 35:2608–2616.
30. Gregoire, C., L. Chasson, C. Luci, E. Tomasello, F. Geissmann, E. Vivier, andT. Walzer. 2007. The trafficking of natural killer cells. Immunol. Rev. 220: 169–182.
31. Freud, A. G., B. Becknell, S. Roychowdhury, H. C. Mao, A. K. Ferketich,G. J. Nuovo, T. L. Hughes, T. B. Marburger, J. Sung, R. A. Baiocchi, et al. 2005.A human CD34(+) subset resides in lymph nodes and differentiates intoCD56bright natural killer cells. Immunity 22: 295–304.
32. Fehniger, T. A., M. A. Cooper, G. J. Nuovo, M. Cella, F. Facchetti, M. Colonna,and M. A. Caligiuri. 2003. CD56bright natural killer cells are present in humanlymph nodes and are activated by T cell-derived IL-2: a potential new link be-tween adaptive and innate immunity. Blood 101: 3052–3057.
33. Nguyen, S., M. Kuentz, J. P. Vernant, N. Dhedin, D. Bories, P. Debre, andV. Vieillard. 2008. Involvement of mature donor T cells in the NK cell re-constitution after haploidentical hematopoietic stem-cell transplantation. Leu-kemia 22: 344–352.
34. Colonna, M., F. Navarro, T. Bellon, M. Llano, P. Garcıa, J. Samaridis,L. Angman, M. Cella, and M. Lopez-Botet. 1997. A common inhibitoryreceptor for major histocompatibility complex class I molecules on humanlymphoid and myelomonocytic cells. J. Exp. Med. 186: 1809–1818.
35. Maenpaa, A., J. Jaaskelainen, O. Carpen, M. Patarroyo, and T. Timonen. 1993.Expression of integrins and other adhesion molecules on NK cells; impact of IL-2 on short- and long-term cultures. Int. J. Cancer 53: 850–855.
36. Beziat, V., B. Descours, C. Parizot, P. Debre, and V. Vieillard. 2010. NK cellterminal differentiation: correlated stepwise decrease of NKG2A and acquisitionof KIRs. PLoS ONE 5: e11966.
37. Bjorkstrom, N. K., P. Riese, F. Heuts, S. Andersson, C. Fauriat, M. A. Ivarsson,A. T. Bjorklund, M. Flodstrom-Tullberg, J. Michaelsson, M. E. Rottenberg, et al.2010. Expression patterns of NKG2A, KIR, and CD57 define a process ofCD56dim NK-cell differentiation uncoupled from NK-cell education. Blood116: 3853–3864.
38. Yu, J., H. C. Mao, M. Wei, T. Hughes, J. Zhang, I. K. Park, S. Liu, S. McClory,G. Marcucci, R. Trotta, and M. A. Caligiuri. 2010. CD94 surface densityidentifies a functional intermediary between the CD56bright and CD56dim hu-man NK-cell subsets. Blood 115: 274–281.
