Received: 12 January 2022 Revised: 7 April 2022 Accepted: 17 April 2022 DOI: 10.1111/tan.14640 ORIGINAL ARTICLE Characterization of KIR+ NK cell subsets with a monoclonal antibody selectively recognizing KIR2DL1 and blocking the specific interaction with HLA-C Raffaella Meazza1 | Michela Falco2 | Paolo Canevali1 | Fabrizio Loiacono1 | Natalia Colomar-Carando 1,3 | Aura Muntasell4,5 | Anna Rea6 | Maria Cristina Mingari1,3 | Franco Locatelli7,8 | 9 4,6 Lorenzo Moretta | Miguel Lopez-Botet | Daniela Pende1 1 Laboratory of Immunology, IRCCS Ospedale Policlinico San Martino, Genoa, Italy 2 Laboratory of Clinical and Experimental Immunology, IRCCS Istituto Giannina Gaslini, Genoa, Italy 3 Department of Experimental Medicine, University of Genoa, Genoa, Italy 4 Immunity and Infection Laboratory, Hospital del Mar Medical Research Institute (IMIM), Barcelona, Spain 5 Department of Cell Biology, Physiology and Immunology, Universitat Autònoma de Barcelona, Bellaterra, Spain 6 Department of Medicine and Life Sciences, Universitat Pompeu Fabra, Barcelona, Spain 7 Department of Hematology/Oncology and Cell and Gene Therapy, IRCCS Ospedale Pediatrico Bambino Gesù, Rome, Italy 8 Department of Gynecology/Obstetrics and Pediatrics, Sapienza University, Rome, Italy 9 Tumor Immunology Unit, IRCCS Ospedale Pediatrico Bambino Gesù, Rome, Italy Correspondence Daniela Pende, Laboratory of The phenotypic identification of different NK cell subsets allows more in-depth Immunology, UO Immunologia, IRCCS characterization of KIR repertoire and function, which are of potential interest Ospedale Policlinico San Martino, L.go R. Benzi 10, 16132 Genoa, Italy. in KIR and disease association studies. KIR genes are highly polymorphic, but Email: daniela.pende@hsanmartino.it a great homology exists among the various sequences and few monoclonal antibodies (mAbs) specifically recognize a single KIR. This is the case of HP- Miguel Lopez-Botet, MELIS, Universitat Pompeu Fabra, c/Doctor Aiguader DM1 which was demonstrated by analysis of cell transfectants and epitope 83, 08003 Barcelona, Spain. mapping to be exclusively KIR2DL1-specific, covering all allotypes identified to Email: miguel.lopez-botet@upf.edu date, except for KIR2DL1*022 and *020, and also to react with KIR2DS1*013. Funding information Here, we compared in immunofluorescence analyses the staining of HP-DM1 European Commission, Grant/Award with other available mAbs to precisely identify KIR2DL1+ NK cells in potential Number: H2020-MSCA-ITN-765104- donors for αβT/B-depleted haplo-HSCT, with known KIR genotype. HP-DM1 MATURE-NK; Fondazione AIRC 5 per mille 2018, Grant/Award Number: 21147; mAb was used in combination with EB6 or 11PB6 (anti-KIR2DL1/S1 and anti- Italian Ministry of Health KIR2DL3*005), 143211 (anti-KIR2DL1/S5), and HP-MA4 (anti-KIR2DL1/S1/ S3/S5) mAbs, allowing the accurate identification of different KIR+ NK cell subsets. These phenotypic evaluations appeared useful to dissect the expression Raffaella Meazza and Michela Falco share first authorship. Miguel Lopez-Botet and Daniela Pende share last authorship. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2022 The Authors. HLA: Immune Response Genetics published by John Wiley & Sons Ltd. HLA. 2022;100:119–132. wileyonlinelibrary.com/journal/tan 119 120 MEAZZA ET AL. pattern of various KIR2D in NK cells from KIR2DL3*005+ individuals, particularly if KIR2DS1 is present. HP-DM1 mAb remarkably refined NK cell phenotyping of donors carrying KIR2DS5, either in the centromeric or telomeric region. Func- tional assays with KIR2DL1+/S1+/S5+ NK cells confirmed that only HP-DM1 exclusively reacts with KIR2DL1. Finally, we demonstrated that HP-DM1 mAb blocked KIR2DL1 recognition of C2+ HLA-C. Altogether, the data support that HP-DM1 is a unique reagent valuable for characterizing KIR+ NK cell subsets. KEYWORDS HLA-C, killer immunoglobulin-like receptors, KIR/KIR-ligand interaction, monoclonal antibodies, natural killer cells 1 | INTRODUCTION though with lower affinity,17,18 whereas the ligands of the other KIR2DS are incompletely characterized.9,19–22 Natural Killer (NK) cells are considered cytotoxic compo- KIR genes are inherited as haplotypes, comprising cen- nents of innate lymphoid cells (ILCs) that provide the tromeric (Cen) and telomeric (Tel) regions, separated by a first line of defense against viruses and tumors in periph- recombination site. These regions are bordered by con- eral blood and tissues.1,2 In uterine mucosa, tissue- served framework genes (i.e., 3DL3, 3DP1, 2DL4, and resident NK cells and ILCs exert an important role in reg- 3DL2). Based on the variety of number and type of the KIR ulating normal placental development.3,4 NK cells keep present, different centromeric (i.e., Cen-A, Cen-B1, and in check the health of neighboring cells through an array Cen-B2) and telomeric (i.e., Tel-A and Tel-B) regions have of germline-encoded receptors, upon engagement with been identified.23–25 Two groups of KIR haplotypes have specific ligands.5 The balance between activating and been defined.26,27 KIR A haplotypes are composed by inhibitory signals transmitted by these receptors finely Cen-A and Tel-A, and have a fixed and limited number of regulates NK cell function.1 Major NK cell receptors are genes. They include genes encoding for iKIR recognizing KIR, a family of transmembrane proteins characterized HLA-C (2DL1, 2DL3), HLA-B (3DL1), HLA-A (3DL2), and by 2 or 3 Ig-like extracellular domains (KIR2D or carry only one aKIR (2DS4) that, in European populations, KIR3D), which include both inhibitory (iKIR) and acti- often codes for a truncated/no-functional receptor.28 Differ- vating receptors (aKIR).6–9 While iKIR have a “long” ently, KIR B haplotypes (i.e., Cen-A/Tel-B, Cen-B/Tel-A, (L) cytoplasmic tail containing ITIMs (i.e., KIR2DL and and Cen-B/Tel-B) are characterized by a great gene content KIR3DL), aKIR have a “short” (S) cytoplasmic tail and diversity and by various genes encoding aKIR.8 Restricted transduce the activating signal through KARAP/DAP12 to KIR B haplotypes are the following genes: 2DS2, 2DL2, adaptor molecule. The most relevant iKIR recognize epi- and 2DL5B located in Cen-B, 3DS1, 2DL5A, and 2DS1 topes shared by distinct groups of HLA-A, -B, or -C allo- located in Tel-B, while 2DS3 and 2DS5 can be present in types.9 The KIR family consists of four distinct lineages, both Cen-B and Tel-B. Regarding 2DS5, ethnic differences which differ in structural characteristics and specific have been observed. In populations of European origin, HLA binding capacity. Lineage III includes all KIR recog- there is a single 2DS5 allele (i.e., 2DS5*002), which is in nizing HLA-C.10 Two epitopes are defined by a dimor- strong linkage disequilibrium with 2DS1 and located in phism at position 80 of HLA-C, where asparagine (N80) Tel-B region. In Africans and African Americans, 2DS5 and lysine (K80) identify the C1 and C2 groups of can be found in both Cen-B and Tel-B regions, and is poly- KIR ligands (KIR-L), respectively.8,11 KIR2DL1 through morphic. In particular, among the Cen-B 2DS5 alleles, methionine 44 (M44) stringently recognizes C2, whereas 2DS5*006 is highly represented in sub-Saharan Africans KIR2DL2/L3 through K44 mainly recognize C1 and with and encodes for a receptor that recognizes C2.29 Studying low affinity C2.12–15 Notably, KIR2DL1*022 represents an pre-eclampsia in pregnant women, when the fetus carries exception being characterized by K44 and displaying C1 a C2 epitope, maternal KIR AA genotypes are risk factors, recognition.16 For the sake of brevity, hereafter “KIR” whereas the Tel-B 2DS1 in Europeans and the Cen-B acronym will be omitted from genes and proteins. Line- 2DS5*006 in Africans are protective.30 age III also includes five aKIR (2DS1, 2DS2, 2DS3, 2DS4, Circulating NK cells of different donors show extremely and 2DS5). The activating receptor 2DS1, sharing M44 variegated KIR repertoires,31 primarily because of the high with 2DL1, has the same specificity for C2 epitope, polymorphism of KIR and HLA class I genotypes.8 MEAZZA ET AL. 121 Importantly, KIR expression primarily results from stochas- KIR gene profiles were performed using Olerup SSP-PCR tic events during NK cell differentiation but is also (sequence-specific primer-PCR) KIR genotyping kit influenced by interaction with self-HLA class I molecules, (CareDx, Stockholm, Sweden) following the manufac- following an education process.32–35 Besides genetic factors, turer's instruction. KIR-L were evaluated analyzing high- environmental stimuli (e.g., pathogen exposure) may also resolution HLA class I typing with KIR-ligand calculator contribute to the diversity.31,36,37 program (http://www.ebi.ac.uk/ipd/kir/ligand.html).40 In The combined KIR genotype and phenotype analysis case of low-resolution HLA class I typing, analysis of can be informative to precisely evaluate the actual fre- KIR-L was also performed by SSP-PCR using KIR HLA quency of specific NK cell subsets with potential clinical ligand kit (CareDx). KIR genotypes and KIR-L of the indi- interest.38,39 A relevant application aims to select suitable viduals described in this study are reported in Figure S1. donors for haploidentical hematopoietic stem cell trans- plantation (haplo-HSCT) in leukemia patients. In the case of KIR/KIR-L mismatch in graft versus host direction, the 2.3 | Isolation and culture of NK cells size of the alloreactive NK cell subset identified in potential alternative donors can be compared.14,40 In addition, KIR Peripheral blood mononuclear cells (PBMC) were isolated phenotyping can help to distinguish iKIR from aKIR, but by Ficoll gradient centrifugation from heparinized blood the mAb specificity and combined staining strategies should of healthy donors. NK cells, purified using the RosetteSep be precisely defined. Indeed, we previously demonstrated method (StemCell Technologies, Vancouver, BC), were that EB6B (hereafter called EB6) and 11PB6 mAb recognize cultured on irradiated feeder cells in the presence of not only 2DL1/S1 but also the 2DL3*005 allotype, which 2 μg/ml phytohemagglutinin (Sigma-Aldrich, Irvine, UK) shares E35 and R50, two residues involved in these mAb and 600 IU/ml rIL-2 (Proleukin, Chiron Corp., Emeryville, epitopes.41 There is a great homology among the various USA) to obtain proliferation and expansion of activated KIR aminoacidic sequences, and very few anti-KIR mAb polyclonal NK cells. NK cell clones have been obtained produced so far specifically recognize only one KIR. Exam- culturing upon limiting dilution either purified NK cells or ples are anti-KIR3DL1 DX9 mAb, anti-KIR2DS4 FES172 CD3 GL-183 11PB6 HP-MA4+ cells sorted from PBMC mAb, and anti-KIR2DL3 (except *005 and *015 allotypes) to enrich in 2DS5+/other KIR2D clones. ECM-41 mAb.42–44 In this article we characterized the use- fulness of a unique mAb (HP-DM1), which exclusively reacts with KIR2DL1.45 The use of HP-DM1 mAb in combi- 2.4 | Monoclonal antibodies and nation with other selected anti-KIR2D mAbs allowed the cytofluorimetric analysis identification of distinct KIR+ subsets. Moreover, functional assays showed that HP-DM1 specifically blocked the inter- All anti-KIR mAb used in this study are described with the action of 2DL1 with HLA-C carrying C2 epitope. updated specificity and the fluorochrome conjugation in Table 1. The anti-CD3-BV510 (UCHT1, IgG1) and the anti- CD56-BV421 (NCAM16.2, IgG2b) were provided by BD 2 | METHOD S A N D MAT E R I AL S Biosciences (San José, CA). In multi-color fluorescence ana- lyses using un-conjugated HP-DM1 mAb, NK cells were 2.1 | Donors first incubated with HP-DM1 followed by anti-IgG1-PE (Southern Biotechnology, Birmingham, AL), washed twice, All individuals described in this article are healthy potential and then incubated with the other fluorochrome-labeled donors for αβT/B-depleted haplo-HSCT (NCT01810120) ana- mAb. More recently, HP-DM1-PE (Biolegend, San Diego, lyzed for KIR gene profile and NK cell receptor phenotype.40 CA) has been available, allowing co-incubation with other This study was approved by the Ethical Committee of IRCCS anti-KIR-APC mAb. The HP-DM1-PE/EB6-PC7/HP-MA4- Ospedale Pediatrico Bambino Gesù (OPBG, Rome, Italy), APC and HP-DM1-PE/EB6-APC/CH-L-FITC combinations Prot. No. 424/2011. Written informed consent was obtained were also used. Importantly, in co-stainings including EB6 from all donors in accordance with the Helsinki declaration. and HP-MA4 mAbs, EB6 was always added 10 min before the other mAb(s). Flow cytometric analysis of resting NK cells (gating on CD3CD56+ cells of PBMC) or polyclonal 2.2 | KIR gene profile and KIR-ligand activated NK cells were performed on either MACSQuant analyses Analyzer (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) or Gallios flow-cytometer (Beckman Coulter, DNA of the tested samples was extracted using QIAamp Brea, CA). Data were analyzed using FlowJo Version 10.7 DNA Blood Mini kit (QIAGEN, GmbH, Germany). The (TreeStar, Ashland, OR). 122 MEAZZA ET AL. TABLE 1 Anti-KIR monoclonal antibodies used in this study, with updated specificity Clone Specificity Isotype Fluorochrome Vendor/Reference HP-DM1 KIR2DL1 IgG1 Un-conjugated, PE Biolegend (San Diego, CA), reference [45] 143211 KIR2DL1/S5 IgG1 APC R&D Systems (Minneapolis, MN) EB6B KIR2DL1/S1, KIR2DL3*005a IgG1 APC, PC7, PE Beckman Coulter (Brea, CA) 11PB6 KIR2DL1/S1, KIR2DL3*005a IgG1 Un-conjugated, FITC Miltenyi Biotec GmbH (Bergisch Gladbach, Germany) HP-MA4 KIR2DL1/S1/S3/S5 IgG2b APC Biolegend GL-183 KIR2DL2/L3/S2b IgG1 PE Beckman Coulter CH-L KIR2DL2/L3/S2b IgG2b FITC BD Biosciences (San José, CA) ECM-41 KIR2DL3 (not E35 allotypes) IgM Un-conjugated Reference [43] Z27 KIR3DL1/S1 IgG1 PE Beckman Coulter DX9 KIR3DL1 IgG1 PE-Vio770 Miltenyi Biotec a EB6B and 11PB6 mAb also recognize KIR2DL3 allotypes characterized by E35 and R50, including KIR2DL3*005. b GL-183 and CH-L mAbs also recognize KIR2DL3 allotypes characterized by E35, including KIR2DL3*005. 2.5 | Cytotoxicity assay kit (Invitrogen, Carlsbad, CA). DNA sequencing was per- formed using d-Rhodamine Terminator Cycle Sequencing Reverse antibody dependent cellular cytotoxicity (R- kit and a 3100 ABI automatic sequencer (PerkinElmer, ADCC) against FcγR+ P815 target cells was performed Wellesley, MA). either in the absence or presence of the indicated mAbs (0.5 μg/ml) using polyclonal activated NK cells as effec- tors (E:T ratio 4:1) in a 4 h 51Cr-release assays. In R- 2.7 | Detection of 2DL3*004, *005, *010, ADCC, killing is enhanced when the mAb is reacting and *036 alleles with an NK activating receptor, while it is decreased by an anti-inhibitory receptor mAb.14 In masking experi- To identify donors characterized by at least one 2DL3*004, ments using C1R target cells, mAbs were used at a final *005, *010, or *036 alleles, genomic DNA of 2DL3 positive concentration of 20 μg/ml. donors was tested using the set of primers 2DL3*005 for: 50 -CAGAAAACCTTCCCTCCG and 2DL2/L3 rev: 50 - TGGGCCCTGCAGAGAA. These four 2DL3 alleles code 2.6 | Analysis of KIR2DL1, 2DS1, and for allotypes characterized by E35, and therefore not rec- 2DS5 transcripts ognized by ECM-41 mAb. PCR was performed for 30 cycles: 30 s at 95 C, 30 s at 60 C, and 1 min at 72 C. Total RNA was extracted from NK cell bulk population All PCR reactions also included a set of primers amplify- or clone derived from donor TB17B using RNeasy ing a conserved fragment of DRΑ gene (internal control), micro kit (Qiagen), according to the manufacturer's to avoid false negative results.47 In this article these four instructions. The cDNA synthesis was performed on 2DL3 alleles have been indicated as 2DL3*005. The detec- 1 μg RNA using oligo(dT) oligo nucleotides. The pres- tion of 2DL3*005 allele is shown in Figure S2. Additional ence of 2DL1, 2DS1, and 2DS5 transcripts was analyzed 2DL3 alleles coding for allotypes characterized by E35 are using the sets of primers Fcg622 50 -CCATCAGTCG 2DL3*014, *015, *017, *018, *033, and *035. In all 2DL3+ CATGACG, Ra957 50 -CCACTCGTATGGAGAGTCAT, donors analyzed in this study, when an unexpected GL- Ra872 50 -AATGTTCCGTTGACCTTGGT; Fr621 50 - 183+ ECM-41 NK cell subset was present (as in samples TCTCCATCAGTCGCATGAR, Ra899 50 -AGGGCCCAGA analyzed in Figure S3), this PCR analysis always revealed GGAAAGTT; Fc551 50 -AGAGAGGGGACGTTTAACC, the presence of 2DL3*005. R939 50 -GGAAAGAGCCGAAGCACT, respectively.46 The primers used to amplify complete 2DS5 transcripts were 2DS5 ORF up: 50 -CATGTYRCTCATGGTCATC and 2.8 | Statistical analysis C: 50 -AAAACACAGTGATCCAATTA. PCR was performed for 30 cycles: 30 s at 95 C, 30 s at 60 C, and 30 s at 72 C; Statistical analyses were performed using Graphpad the amplification product was cloned into pcDNA3.1/V5/ software Version 6.0. The utilized test was Ordinary His TOPO vector using the Eukaryotic TOPO TA Cloning one-way ANOVA for repeated measures followed by MEAZZA ET AL. 123 Bonferroni multiple comparison test. Not significant (n.s.); HP-DM1 with 143211, EB6, 11PB6, and HP-MA4 mAb. ****p < 0.0001; ***p < 0.001; **p < 0.01; and *p < 0.05. HP-DM1 and 143211 mAb co-stained the 2DL1+ subset, while 11PB6 or EB6 (sharing the same KIR specificity, see Table 1) and HP-MA4 mAb also recognized 2DS1+ 3 | R E SUL T S cells and, differently from HP-DM1 and 143211, triggered activation in R-ADCC. In 2DL1+/S1+ donors, HP-DM1 3.1 | HP-DM1 mAb recognizes KIR2DL1 discriminated in the EB6+, 11PB6+, or HP-MA4+ on NK cells populations two subsets according to their staining inten- sity, most likely corresponding to KIR2DL1+/S1 (dim) Based on the knowledge of the HP-DM1 mAb reactivity and KIR2DL1+/S1+ (bright) cells consistent with a previ- on HEK-293T cells transiently transfected with plasmids ous report.48 coding for different lineage III KIR2D,45 we analyzed its Taken together, these data indicated that HP-DM1 binding to resting NK cells derived from donors with mAb exclusively stains 2DL1 on NK cells. known KIR gene repertoires (Figure 1 and Figure S1). HP-DM1/EB6 double staining of NK cells derived from 2DL1+/S1, 2DL1/S1+, or 2DL1+/S1+ donors confirmed 3.2 | Use of HP-DM1 mAb for NK cell the unique 2DL1 specificity of HP-DM1 mAb in contrast phenotyping of 2DL3*005+ donors to the ability of EB6 mAb to recognize both 2DL1 and 2DS1 (Figure 1A). Next, we analyzed polyclonal Since HP-DM1 can selectively recognize 2DL1 allotypes NK cell populations derived from 2DL1+/S1+/S5 donors (except 2DL1*022 and 2DL1*020) as demonstrated by epi- by flow-cytometry (a representative case is shown in tope mapping,45 the inclusion of HP-DM1 mAb in the Figure 1B) and by R-ADCC assay (Figure 1C), comparing panel of anti-KIR2D reagents could contribute to more F I G U R E 1 HP-DM1 mAb selectively recognizes 2DL1 on NK cells. (A) Resting NK cells derived from three donors (TB336A, TB336B, and TB521A) characterized by different 2DL1 and 2DS1 gene profiles (positive and negative KIR genes are reported as gray and white boxes, respectively) were stained with HP-DM1-PE in combination with EB6-APC. (B) Activated NK cells derived from a donor (TB98A), characterized by all lineage III KIR2D genes except 2DS5, were stained with HP-DM1 mAb followed by anti-IgG1-PE and, subsequently, with either 143211-FITC, EB6-APC, 11PB6-FITC or HP-MA4-APC mAb. Percentages of positive cells are indicated. (C) In R-ADCC assay, 51Cr- labeled P815 were incubated with NK cells derived from different 2DL1+/S1+ donors, including the representative donor described in (B), in the absence or presence of the indicated mAb. The KIR genotype and KIR-L of these donors are reported in Figure S1. Data, obtained by nine independent experiments, are shown as mean + SEM. E:T ratio was 4:1. ****p < 0.0001 124 MEAZZA ET AL. F I G U R E 2 HP-DM1 mAb allows to discriminate 2DL1+ from other KIR2D+ NK cells in 2DL3*005+ donors. Immunofluorescence analysis was performed on resting NK cells derived from two different 2DL3*005+ donors, TB520B donor (A), and TB344B donor (B). Lineage III KIR2D gene profiles are shown, positive and negative genes are identified as gray and white boxes, respectively. The whole KIR genotype and KIR-L of TB520B and TB344B donors are reported in Figure S1. HP-DM1-PE was used in combination with either 143211-APC, HP-MA4-APC, or EB6-APC. Different NK cell subsets are identified by gates, which are numbered and defined for their KIR composition and percentage of NK cells included (table on the right side). In HP-DM1-PE/EB6-PC7/HP-MA4-APC combined staining (B, lower panels), the comparison of HP-MA4/EB6 dot plots before and after HP-DM1+ cell exclusion allows to detect the localization of 2DL1+ cells within the different populations: the whole gate no. 5 and right part of gate no. 6 (indicated within brackets after HP-DM1+ cell exclusion) precisely define NK cell subsets. To approach this issue, shown in Figure 2A). Thus, HP-DM1/HP-MA4, like we tested HP-DM1 staining on NK cells from 2DL3*005+ HP-DM1/143211 identified only 2DL1+ NK cells, while donors because this allele encodes a receptor character- HP-DM1/EB6 detected 2DL3*005+ cells, co-expressing or ized by an unusual antibody reactivity being recognized not 2DL1. If we added an anti-KIR2DL2/L3/S2 mAb by the anti-KIR2DL1/S1 (EB6 and 11PB6) mAb and not (e.g., CH-L or GL-183) to HP-DM1/EB6 staining, we found by the anti-KIR2DL3 (ECM-41 and 180701) mAbs.41,49 that all EB6+/HP-DM1 cells were also CH-L+, displaying We evaluated whether the HP-DM1/EB6 combination a diagonal staining pattern typical of co-recognizing the would allow distinguishing 2DL1+ from 2DL3*005+ NK same molecule (i.e., 2DL3*005) (Figure S3A, left panel). In cell fractions. To this aim, NK cells from 2DL3*005+ the strategy described by Beziat et al.,49 based on a refined donors with KIR A/A genotype were stained with either 15-color flow cytometry panel and a flowchart with HP-DM1/EB6, HP-DM1/HP-MA4 or HP-DM1/143211. sequential quality controls (QCs), EB6/GL-183 represented Since HP-MA4 does not recognize 2DL3*005 (Table 1),45 the first QC, and the diagonal staining pattern was the HP-DM1/HP-MA4 and HP-DM1/EB6 dot plots were hallmark of the expressed 2DL3*005 allotype (Figure S3, clearly different (the representative case TB520B is middle panels). MEAZZA ET AL. 125 Notably, the described mAb combinations can be very panel could also be performed, providing a better identifica- informative for 2DL3*005+/2DS1+ donors. In the represen- tion of 2DL1+, 2DS1+, and 2DL3*005+ subsets (Figure 2B). tative case TB344B shown in Figure 2B, 143211 and HP- Consistent data were obtained with the GL-183/ECM-41 DM1 co-stained 2DL1+ NK cells, HP-MA4+/HP-DM1 mAb combination. In the 2DL3*005/*x donor TB520B, identified 2DS1+ cells, while EB6+/HP-DM1 subset con- 2DL3*005+ NK cells were identified as the GL-183+/ECM- tained both 2DL3*005+ and 2DS1+ cells, as also docu- 41 subset (Figure S3A, right panel). In donor TB344B, car- mented by the additional staining with CH-L mAb, rying CenA/CenB (i.e., having only one 2DL3 allele), no detecting CH-L+ (i.e., 2DL3*005+) and CH-L (i.e., 2DS1+) ECM-41 staining was observed (Figure S3B, right panel). subsets (Figure S3B, left panel). Finally, the simultaneous Notably, the presence of 2DL3*005 allele in these donors staining with HP-DM1/EB6/HP-MA4 mAb in the same was confirmed by SSP-PCR analysis (Figure S2). F I G U R E 3 HP-DM1 mAb uniquely recognizes 2DL1 in NK cells from donor carrying 2DS1 and telomeric 2DS5. Immunofluorescence analyses were performed on resting NK cells derived from TB62B donor (A), characterized by the presence of telomeric 2DS5, and the respective recipient at 6 months after the transplantation (B). The lineage III KIR2D gene profile is reported (positive and negative genes are identified as gray and white boxes, respectively). The whole KIR genotype and KIR-L of TB62B donor and patient TB62 pre-transplant are shown in Figure S1. HP-DM1-PE was used in combination with either 143211-APC, HP-MA4-APC, or EB6-APC; HP-MA4-APC/EB6-PE staining is also shown. Different NK cell subsets (numbered gates) are identified by the indicated mAb combinations. KIR composition and percentage of NK cells included in each gate are shown in the table on the right side. In gate 3, cells express either 2DS1 or 2DS5 or both (the slash indicates these possibilities). (C) Polyclonal activated NK cells derived from the patient at 6 months after transplantation, shown in panel (B), were tested in R-ADCC assay against 51Cr-labeled P815, in the absence or presence of the indicated mAb. Data are shown as mean + SEM and are pooled from three independent experiments. E:T ratio was 4:1. ****p < 0.0001 126 MEAZZA ET AL. F I G U R E 4 HP-DM1 mAb allows to dissect 2DL1+ from 2DS5+ NK cells in donors carrying centromeric 2DS5. (A and B) Immunofluorescence, transcript and functional analyses performed using NK cells derived from two donors (TB512B, and TB17B), characterized by centromeric 2DS5. The lineage III KIR2D gene profiles are shown (positive and negative genes are reported as gray and while boxes, respectively), while the whole KIR genotype and KIR-L are reported in Figure S1. (A) Resting NK cells derived from donor TB512B were stained with HP-DM1-PE in combination with 143211-APC, HP-MA4-APC, or EB6-APC; HP-MA4-APC/EB6-PE staining is also shown. Different NK cell subsets are identified by numbered gates. Definition of KIR subsets and percentage of NK cells included in each gate are reported in the table on the right side. (B) Activated polyclonal (upper panels) or clonal (lower panels) NK cells derived from donor TB17B were analyzed by immunofluorescence. Percentages of positive cells, RT-PCR of the indicated KIR2D transcript (the positive transcripts are identified as gray, while negative as white boxes), and R-ADCC against P815 in the absence or presence of the indicated mAb are reported. ****p < 0.0001; ***p < 0.001; **p < 0.01 3.3 | Use of HP-DM1 mAb on NK cell the activating 2DS1 and/or 2DS5 molecules (HP-DM1/ phenotyping of telomeric 2DS5+ donors HP-MA4+). Staining performed using EB6 in combina- tion with HP-DM1 or HP-MA4 allowed the visualization In our cohort, mainly consisting of Italian individuals, of additional NK cell subsets as detailed in Figure 3A. 2DS5 was located primarily in Tel-B region, in agreement Apparently, EB6 did not recognize 2DS5 on NK cells with studies analyzing KIR polymorphism in European since HP-MA4+/EB6 cells (i.e., gate 5) were detectable, populations.24 In 2DL1+/2DS1+/2DS5+ donors, the com- as previously reported.50 Also in this case, the HP-DM1/ bined staining of HP-DM1 with 143211 mAb allowed the EB6/HP-MA4 mAb combination was useful to identify discrimination of 2DL1+ (HP-DM1+/143211+) from the 2DL1+ subset, providing more precise discrimination 2DL1/2DS5+ (HP-DM1/143211+) NK cell subsets between 2DS1+ (recognized by both EB6 and HP-MA4 (Figure 3A). In addition, with HP-DM1/HP-MA4 it was mAb) and 2DS5+ (stained by HP-MA4 mAb) subsets possible to discriminate the inhibitory 2DL1+ (HP- (Figure S4A). In this donor, as in most studied cases, the DM1+/HP-MA4+) NK cell subset from that expressing frequency of the 2DS5+ NK cell subset was low (2%), at MEAZZA ET AL. 127 least among 2DL1 cells. KIR genotype and KIR-L of the 2DL1+/2DS1+/2DS5+ TB17B donor (of Cuban origin) donor (TB62B) and the related patient before transplanta- (Figure 4B). Notably, the profiles obtained using 143211 tion (TB62) are shown in Figure S1 panel A and B, and HP-MA4 revealed both dull and bright staining, with respectively. Interestingly, testing the NK cell phenotype the weak reactivity suggesting 2DS5 recognition. Again, of the reconstituted repertoire in the leukemia patient R-ADCC assay provided a functional demonstration that following αβT/B-depleted haplo-HSCT from this donor, only HP-DM1 mAb was reacting exclusively with the we observed a marked expansion of the donor-derived inhibitory receptor 2DL1, while 143211 (also recognizing 2DS5+ subset (from 2% to 23%) and a reduction of 2DL1+ 2DS5), 11PB6 mAb (also recognizing 2DS1), and HP- NK cells (from 9% to 1%) (Figure 3B). In R-ADCC assay, MA4 (also recognizing 2DS1 and 2DS5) triggered lysis. all the mentioned mAb except HP-DM1 triggered lysis of These findings were further documented at the clonal activated NK cells, consistent with the engagement of level. We selected HP-MA4+/143211+/EB6/HP-DM1 2DS5 by 143211, 2DS1 by EB6, and both 2DS1 and 2DS5 (with dull staining of HP-MA4 and 143211) NK cell by HP-MA4 mAb (Figure 3C). clones, in which HP-MA4 and 143211 mAb triggered lysis in R-ADCC, revealing the activating function of the engaged receptor (i.e., 2DS5). By RT-PCR, we detected 3.4 | Impact of HP-DM1 mAb on NK cell the presence of 2DS5 transcript and the absence of 2DL1 phenotype of centromeric 2DS5+ donors and 2DS1 (a representative clone is shown in Figure 4B). These 2DS5+ 2DL1 NK cell clones appeared to be edu- We had the opportunity to also analyze donors from Africa cated either through the expression of self-reactive iKIRs and Central America, finding two characterized by a cen- as 2DL2/L3 and 3DL1, or through NKG2A (Figure S5B). tromeric 2DS5 allele; both donors have C1, C2, and Bw4 KIR-L (Figure S1). While the presence of a 2DS5 centro- meric allele in donor TB512B was quite evident (this 3.5 | HP-DM1 recognizes the 2DL1 donor is 2DL5B+, 2DL5A, 2DS3, and 2DS5+, Figure S1), epitope relevant for binding to HLA-C in donor TB17B (2DL5B+, 2DL5A+, 2DS3+, and 2DS5+, Figure S1) the identification of the 2DS5 allele was Epitope mapping of HP-DM1 suggested that this mAb required. To this end we sequenced 2DS5 transcripts from binds to residues involved in ligand recognition.45 To ver- polyclonal activated NK cells. This analysis revealed the ify this hypothesis, we analyzed whether pre-treatment of presence of 2DS5*006, an allele that has been described NK cells with HP-DM1 mAb could mask the interaction only in the centromeric region.29 Notably, both individuals between 2DL1 and HLA-C allotypes with C2 epitope. To were characterized by exceptionally high percentages of this end, we tested in cytolytic assays activated NK cells 2DS5+ NK cells. Indeed, in the 2DL1+/2DS1/2DS5+ derived from two donors (TB372B and D9) characterized by TB512B donor (of Ugandan origin), the 2DS5+/2DL1 an expansion of 2DL1+ subset (82% and 98%, respectively) subset was quantified as 35% or 46% using 143211/HP- DM1 or HP-MA4/HP-DM1 combinations, respectively, showing a better 2DS5 staining by HP-MA4 than 143211 mAb (Figure 4A). This donor was also characterized by the presence of 2DL3*005 (Figure S2), leading to a further complication in KIR subset definition. Thus, the presence of 2DL3*005+ subset in TB512B NK cells was detected by the bright staining of EB6, partially co-expressing 2DL1 (in EB6/HP-DM1) and/or 2DS5 (in EB6/HP-MA4) (Figure 4A). As expected, EB6+/HP-DM1 cells were also CH-L+, and the staining with GL-183/ECM-41 mAb did not show any ECM41+ cells (Figure S3C). The HP-DM1/ EB6/HP-MA4 mAb combination further defined various F I G U R E 5 HP-DM1 mAb blocks 2DL1/HLA-C2 interaction. KIR+ subsets (Figure S4B). In addition to the evidence that Polyclonal activated NK cells with a large 2DL1+ cell subset were 2DS5+ 2DL1 NK cells could co-express 2DL3*005 (gate derived from two donors (TB372B and D9, whose KIR genotypes 6 in Figure 4A), we also found that they co-expressed and KIR-L are reported in Figure S1), and tested in 51Cr-release 2DL2, 3DL1 and/or NKG2A, supporting the notion of assays against C1R B-EBV cell line (HLA-C2) in the absence or in their licensing state (Figure S5A). the presence of the indicated mAb. E:T ratio was 5:1. Data are A highly represented 2DS5+ NK cell subset was also shown as mean + SEM and are pooled from duplicates in two detected in the polyclonal population derived from the independent experiments. ****p < 0.0001. 128 MEAZZA ET AL. against C1R B-EBV cell line, expressing HLA-C*04:01 whenever combinations of mAbs react with different epi- (C2 epitope) (Figure 5). These target cells were resistant to topes of the same KIR, the staining procedure should be lysis, and, upon 2DL1 masking by HP-DM1 mAb, the kill- carefully assessed to prevent binding interferences. In our ing could be efficiently restored at the same level detected experience, when EB6 and HP-MA4 mAbs were combined, using 11PB6 mAb and by blocking HLA class I molecules a sequential staining was required to preserve the optimal using 6A4 mAb.51 The anti-2DL2/L3/S2 GL-183 mAb was reactivity of both mAbs, with a procedure (see Methods) used as a negative control. These data demonstrated that different from that described by Czaja et al.50 In the other HP-DM1 mAb blocks the HLA-C binding site of 2DL1. mAb combinations including HP-DM1, simultaneous incu- bation with different reagents was acceptable. Notably, only HP-DM1 mAb precisely identified 2DL1+ NK cells, 4 | DISCUSSION and its use in combination with other mAb allowed the identification of distinct subsets co-expressing or not The high diversity of NK cell phenotypic repertoire may 2DL3*005, 2DS1, and/or 2DS5. Moreover, this approach depend on different factors including KIR gene content, contributed to clarify potentially ambiguous staining pat- allelic polymorphism, transcriptional regulation, copy num- terns. For instance, the use of HP-DM1 enabled to precisely ber variation, self HLA class I polymorphism, and cytomeg- define the KIR2DL1+ subset in the presence of 2DL3*005, alovirus (CMV) infection.36,52–59 Hence, in addition to KIR which is recognized by anti-KIR2DL1/S1 EB6 and 11PB6 and HLA class I genotyping, phenotypic characterization of mAbs. Adding HP-MA4 mAb to these two mAb allowed KIR repertoire becomes an important tool for basic and to dissect 2DL1+, 2DS1+, and 2DL3*005+ subsets. In the clinical studies. In this context, given the frequent reactivity same line, as predicted by detailed epitope mapping,45 the of mAbs with different KIRs, the combined use of well- combination of anti-KIR mAb employed in this study characterized reagents is required to correctly identify cer- would allow the discrimination of canonical C2-specific tain KIR expressing subsets. To this end, knowledge of the 2DL1 allotypes (HP-DM1+) from the atypical C1-specific residues relevant in their reactivity is helpful. Regarding 2DL1*022 (HP-DM1 but 143211+, EB6+, and HP-MA4+). anti-KIR2DL1 mAbs, epitope mapping was previously In the case of TB512B donor from Uganda, characterized reported for EB6 mAb (E35 and R50),41 being as well by centromeric 2DS5 and 2DL3*005 (Figure 4A), the pres- achieved for HP-DM1 (M44, S67, R68, and T70) and 143211 ence of 2DL1*022 could be excluded because EB6+/HP- (R68 and T70) mAb.45 Definition of KIR residues relevant DM1 cells were all CH-L+ and thus 2DL3*005+. for HP-DM1 binding and analysis of KIR sequences allowed HP-DM1 mAb appears of particular interest for to predict its exclusive reactivity with 2DL1, covering all immunophenotyping 2DS5+ donors, because the other allotypes so far identified excepting 2DL1*022 (character- available anti-KIR2DL1 143211 mAb does also bind ized by K44) and, probably, 2DL1*020 (carrying G67), 2DS5. Thus, HP-DM1/143211 co-staining allows the dis- described in sub-Saharan African populations which were tinction of 2DL1 from 2DS5, which is also efficiently rec- scarcely represented in our cohort. In European ognized by HP-MA4.50 On that basis, the HP-DM1/HP- populations, the most common KIR2DL1 alleles on the A MA4/EB6 mAb combination was set up to discriminate haplotype are KIR2DL1*003, *002 or *001, while the domi- 2DL1, 2DS5, 2DS1, and 2DL3*005. Remarkably, through nant allele on B haplotype is KIR2DL1*004.15,57 Regarding this approach we detected higher proportions of 2DS5+ 2DL1*022, this allele has been described in KhoeSan popu- NK cells in centromeric than telomeric 2DS5+ individ- lation of Southern Africa, that are characterized by a uals, to be confirmed by studying a larger cohort. These unique demographic history and are genetically diverse data are consistent with the expression levels detected on from other human populations.16,60 Another potential the surface of NKL cells transfected with different 2DS5 exception is the predicted HP-DM1 binding to 2DS1*013, alleles, showing lower levels of 2DS5*002 compared with which carries M44, S67, R68, and T70, like most 2DL1 and the other allotypes.61 This depends on polymorphic varia- different, at position 70, from the other 2DS1 allotypes. To tion that impacts the N-linked glycosylation, maturation the best of our knowledge, no information is available and/or transport of the protein to the membrane.61 More- about the frequency of 2DS1*013. Altogether these excep- over, the use of this mAb panel for monitoring the NK tions should not be relevant when analyzing 2DL1 expres- cell reconstitution after HSCT allowed us to observe in a sion by HP-DM1 mAb in European populations. patient a marked expansion of the 2DS5+ subset, with a On that basis, HP-DM1 reactivity with NK cells concomitant contraction of the 2DL1+ population. As derived from potential HSCT donors with a known KIR compared with 2DS5+ NK cells (2%) detected in the gene profile was assessed, employing different flow cyto- donor carrying telomeric 2DS5 (Figure 3), their propor- metry staining strategies to define the proportions and tions raised up to 23% at 6 and 9 months post-HSCT, functional activity of distinct KIR+ subsets. Technically, declining at 1 year (5%). It should be noted that MEAZZA ET AL. 129 donor but not patient KIR genotype contained 2DS5 haplo-HSCT setting described above, it is conceivable that (Figure S1A,B), further sustaining that NK cells analyzed selectively blocking 2DL1 instead of all KIR2D molecules in the patient after transplantation were donor-derived. It might prevent dampening of NK cell activity, given the is uncertain which factor(s) promoted these phenotypic preserved expression of 2DL2/L3, recognizing donor but changes, and their putative relation with CMV reac- not recipient cells. In addition, the presence and activity of tivation or GvHD, that occurred in this patient, deserves 2DS1/S2 would be maintained on the NK cell surface. further investigation. The potential influence of KIR-L in the donor (C1 and Bw4) and in the patient (C1, C2, and AUTHOR CONTRIBUTIONS Bw4) remains also undefined, because the telomeric Raffaella Meazza, Michela Falco, and Daniela Pende 2DS5*002, the predominant allotype in Europeans, does designed the study, analyzed and interpreted the data. not bind HLA-A, -B or –C.29 Actually, the more accurate Miguel Lopez-Botet generated and selected HP-DM1 mAb, definition of the 2DS5+ subset achieved with our staining proposing its collaborative characterization. Raffaella protocol paves the way for phenotypic and functional Meazza, Paolo Canevali, Fabrizio Loiacono, Natalia studies which might contribute to better understanding Colomar-Carando, Aura Muntasell, and Anna Rea per- the biological role of this aKIR.22,62 formed phenotypic and functional analyses. Michela Falco Altogether our study shows that HP-DM1 mAb is a performed KIR genotype and analyzed KIR alleles. Raffaella unique reagent that specifically recognizes 2DL1 mole- Meazza, Paolo Canevali, Fabrizio Loiacono obtained poly- cules, becoming valuable to more precisely characterize clonal and clonal NK cells. Franco Locatelli coordinated the KIR repertoire by multicolor flow or mass cytometry. the study of haplo-HSCT donors and patients. Raffaella In the context of haploidentical HSCT, where KIR-HLA Meazza, Michela Falco, Miguel Lopez-Botet, and Daniela class I mismatch in graft versus host direction has been Pende wrote the manuscript. Aura Muntasell, Maria associated to NK-mediated graft versus leukemia Cristina Mingari, and Lorenzo Moretta critically revised the effect,14,40,63,64 the frequencies of 2DL1+ alloreactive NK manuscript. Michela Falco, Lorenzo Moretta, Franco cells might be precisely monitored by multicolour flow Locatelli, Miguel Lopez-Botet, and Daniela Pende provided cytometry, combining HP-DM1 with the other iKIR- fundings. All the authors approved the final version. specific mAbs. Importantly, HP-DM1 mAb reacts with a conformational epitope including positions M44, R68, ACKNOWLEDGMENTS and T70 that have been described to be involved in the This work was supported by a network grant of the European HLA-C binding site of 2DL1.15,65 Indeed, we provide Commission (H2020-MSCA-ITN-765104-MATURE-NK) to functional evidence supporting that HP-DM1 mAb- Miguel Lopez-Botet and Daniela Pende; Natalia Col- mediated masking of 2DL1 efficiently blocked the recog- omar-Carando and Anna Rea are fellows in the project. nition of HLA-C molecules carrying C2 epitope, with the This work was also supported by Italian Ministry of potential advantage with respect to other mAb as 11PB6 Health (Ricerca Corrente G. Gaslini, Ricerca Corrente to avoid concomitant blocking of the activating 2DS1/ and 5xmille-2016 Ospedale Policlinico San Martino), HLA-C2,14 and the inhibitory 2DL3*005/HLA-C1 interac- Fondazione AIRC 5 per mille 2018 id 21147 to Franco tion.41 On that basis, engineering therapeutic mAbs with Locatelli and Lorenzo Moretta. antagonistic effects based on HP-DM1 structure might be of potential translational relevance. An example can be CONFLICT OF INTEREST represented by selectively blocking donor 2DL1 interac- Miguel Lopez-Botet declares that HP-DM1 and HP-MA4 tion with recipient carrying C2+ HLA-C in haplo-HSCT are commercially available through license agreements to possibly enhance NK alloreactivity, particularly in cer- signed by Universitat Pompeu Fabra. The other authors tain settings (e.g., donor C1/C2 and recipient C2/C2). A declare no conflict of interest. caveat would be to prevent potential interferences with 2DL1-dependent “education” following donor NK cell DA TA AVAI LA BI LI TY S T ATE ME NT differentiation, thus requiring to define the suitable The data that support the findings of this study are avail- schedule for administration of the antagonistic mAb. In able from the corresponding author upon reasonable this regard, it was reported that treatment of cancer request. patient by infusing IPH2101 (the fully human antibody specific for both inhibitory 2DL1/L2/L3 and activating ORCID 2DS1/S2) had limited side effects,66 but also reduced anti- Raffaella Meazza https://orcid.org/0000-0003-0242- tumor activity.67 Indeed, loss of free KIR2D surface mole- 3157 cules by trogocytosis correlated with reduction of NK-cell Michela Falco https://orcid.org/0000-0002-0317-9830 function, presumably interfering with licensing.67 In the Paolo Canevali https://orcid.org/0000-0002-0029-1376 130 MEAZZA ET AL. Fabrizio Loiacono https://orcid.org/0000-0001-9370- the ligand-binding site makes KIR2DL2 a stronger receptor for 8910 HLA-C than KIR2DL3. J Immunol. 2008;180(6):3969-3979. Natalia Colomar-Carando https://orcid.org/0000-0001- 14. Pende D, Marcenaro S, Falco M, et al. Anti-leukemia activity of alloreactive NK cells in KIR ligand-mismatched haploidentical 9716-181X HSCT for pediatric patients: evaluation of the functional role of Aura Muntasell https://orcid.org/0000-0003-2894-0486 activating KIR and redefinition of inhibitory KIR specificity. Anna Rea https://orcid.org/0000-0003-1307-8045 Blood. 2009;113(13):3119-3129. Maria Cristina Mingari https://orcid.org/0000-0001- 15. Hilton HG, Guethlein LA, Goyos A, et al. Polymorphic HLA-C 6998-5459 receptors balance the functional characteristics of KIR haplo- Franco Locatelli https://orcid.org/0000-0002-7976-3654 types. J Immunol. 2015;195(7):3160-3170. Lorenzo Moretta https://orcid.org/0000-0003-4658-1747 16. Hilton HG, Norman PJ, Nemat-Gorgani N, et al. Loss and gain of natural killer cell receptor function in an African hunter- Miguel Lopez-Botet https://orcid.org/0000-0003-4882- gatherer population. PLoS Genet. 2015;11(8):e1005439. 065X 17. Stewart CA, Laugier-Anfossi F, Vely F, et al. Recognition of Daniela Pende https://orcid.org/0000-0003-1565-451X peptide-MHC class I complexes by activating killer immunoglobulin-like receptors. Proc Natl Acad Sci U S A. R EF E RE N C E S 2005;102(37):13224-13229. 1. Vivier E, Raulet DH, Moretta A, et al. Innate or adaptive 18. Chewning JH, Gudme CN, Hsu KC, Selvakumar A, Dupont B. immunity? The example of natural killer cells. Science. 2011; KIR2DS1-positive NK cells mediate alloresponse against the C2 331(6013):44-49. HLA-KIR ligand group in vitro. J Immunol. 2007;179(2):854-868. 2. Vivier E, Artis D, Colonna M, et al. Innate lymphoid cells: 19. Garcia-Beltran WF, Holzemer A, Martrus G, et al. Open con- 10 years on. Cell. 2018;174(5):1054-1066. formers of HLA-F are high-affinity ligands of the activating 3. Parham P, Moffett A. Variable NK cell receptors and their NK-cell receptor KIR3DS1. Nat Immunol. 2016;17(9):1067- MHC class I ligands in immunity, reproduction and human 1074. evolution. Nat Rev Immunol. 2013;13(2):133-144. 20. Carlomagno S, Falco M, Bono M, et al. KIR3DS1-mediated rec- 4. Huhn O, Zhao X, Esposito L, Moffett A, Colucci F, ognition of HLA-*B51: modulation of KIR3DS1 responsiveness Sharkey AM. How do uterine natural killer and innate lym- by self HLA-B Allotypes and effect on NK cell licensing. Front phoid cells contribute to successful pregnancy? Front Immunol. Immunol. 2017;8:581. 2021;12:607669. 21. Sim MJW, Rajagopalan S, Altmann DM, Boyton RJ, Sun PD, 5. Quatrini L, Della Chiesa M, Sivori S, Mingari MC, Pende D, Long EO. Human NK cell receptor KIR2DS4 detects a con- Moretta L. Human NK cells, their receptors and function. Eur J served bacterial epitope presented by HLA-C. Proc Natl Acad Immunol. 2021;51(7):1566-1579. Sci U S A. 2019;116(26):12964-12973. 6. Moretta A, Bottino C, Vitale M, et al. Receptors for HLA class-I 22. Blunt MD, Khakoo SI. Activating killer cell immunoglobulin- molecules in human natural killer cells. Annu Rev Immunol. like receptors: detection, function and therapeutic use. Int J 1996;14:619-648. Immunogenet. 2020;47(1):1-12. 7. Vilches C, Parham P. KIR: diverse, rapidly evolving receptors 23. Pyo CW, Guethlein LA, Vu Q, et al. Different patterns of evolu- of innate and adaptive immunity. Annu Rev Immunol. 2002;20: tion in the centromeric and telomeric regions of group a and B 217-251. haplotypes of the human killer cell Ig-like receptor locus. PLoS 8. Parham P. MHC class I molecules and KIRs in human history, One. 2010;5(12):e15115. health and survival. Nat Rev Immunol. 2005;5(3):201-214. 24. Jiang W, Johnson C, Jayaraman J, et al. Copy number variation 9. Pende D, Falco M, Vitale M, et al. Killer Ig-like receptors leads to considerable diversity for B but not a haplotypes of the (KIRs): their role in NK cell modulation and developments human KIR genes encoding NK cell receptors. Genome Res. leading to their clinical exploitation. Front Immunol. 2019;10: 2012;22(10):1845-1854. 1179. 25. Cisneros E, Moraru M, Gomez-Lozano N, Muntasell A, Lopez- 10. Wroblewski EE, Parham P, Guethlein LA. Two to tango: co- Botet M, Vilches C. Haplotype-based analysis of KIR-gene pro- evolution of hominid natural killer cell receptors and MHC. files in a south European population-distribution of standard Front Immunol. 2019;10:177. and variant haplotypes, and identification of novel recombi- 11. Mandelboim O, Reyburn HT, Vales-Gomez M, et al. Protection nant structures. Front Immunol. 2020;11:440. from lysis by natural killer cells of group 1 and 2 specificity is 26. Uhrberg M, Valiante NM, Shum BP, et al. Human diversity in mediated by residue 80 in human histocompatibility leukocyte killer cell inhibitory receptor genes. Immunity. 1997;7(6): antigen C alleles and also occurs with empty major histocom- 753-763. patibility complex molecules. J Exp Med. 1996;184(3):913-922. 27. Martin AM, Kulski JK, Gaudieri S, et al. Comparative genomic 12. Winter CC, Long EO. A single amino acid in the p58 killer cell analysis, diversity and evolution of two KIR haplotypes A and inhibitory receptor controls the ability of natural killer cells to B. Gene. 2004;335:121-131. discriminate between the two groups of HLA-C allotypes. 28. Maxwell LD, Wallace A, Middleton D, Curran MD. A common J Immunol. 1997;158(9):4026-4028. KIR2DS4 deletion variant in the human that predicts a soluble 13. Moesta AK, Norman PJ, Yawata M, Yawata N, Gleimer M, KIR molecule analogous to the KIR1D molecule observed in Parham P. Synergistic polymorphism at two positions distal to the rhesus monkey. Tissue Antigens. 2002;60(3):254-258. MEAZZA ET AL. 131 29. Blokhuis JH, Hilton HG, Guethlein LA, et al. KIR2DS5 allo- reaction with sequence-specific primers: an update. Tissue types that recognize the C2 epitope of HLA-C are common Antigens. 2002;59(3):184-193. among Africans and absent from Europeans. Immun Inflamm 47. Vilches C, Castano J, Gomez-Lozano N, Estefania E. Facilita- Dis. 2017;5(4):461-468. tion of KIR genotyping by a PCR-SSP method that amplifies 30. Nakimuli A, Chazara O, Hiby SE, et al. A KIR B centromeric short DNA fragments. Tissue Antigens. 2007;70(5):415-422. region present in Africans but not Europeans protects pregnant 48. Fauriat C, Ivarsson MA, Ljunggren HG, Malmberg KJ, women from pre-eclampsia. Proc Natl Acad Sci U S A. 2015; Michaelsson J. Education of human natural killer cells by acti- 112(3):845-850. vating killer cell immunoglobulin-like receptors. Blood. 2010; 31. Freud AG, Mundy-Bosse BL, Yu J, Caligiuri MA. The broad 115(6):1166-1174. spectrum of human natural killer cell diversity. Immunity. 49. Beziat V, Traherne J, Malmberg JA, et al. Tracing dynamic 2017;47(5):820-833. expansion of human NK-cell subsets by high-resolution analy- 32. Kim S, Poursine-Laurent J, Truscott SM, et al. Licensing of nat- sis of KIR repertoires and cellular differentiation. Eur J ural killer cells by host major histocompatibility complex class Immunol. 2014;44(7):2192-2196. I molecules. Nature. 2005;436(7051):709-713. 50. Czaja K, Borer AS, Schmied L, Terszowski G, Stern M, 33. Anfossi N, Andre P, Guia S, et al. Human NK cell education by Gonzalez A. A comprehensive analysis of the binding of anti-KIR inhibitory receptors for MHC class I. Immunity. 2006;25(2): antibodies to activating KIRs. Genes Immun. 2014;15(1):33-37. 331-342. 51. Ciccone E, Pende D, Nanni L, et al. General role of HLA class I 34. Raulet DH, Vance RE. Self-tolerance of natural killer cells. Nat molecules in the protection of target cells from lysis by natural Rev Immunol. 2006;6(7):520-531. killer cells: evidence that the free heavy chains of class I mole- 35. Elliott JM, Yokoyama WM. Unifying concepts of MHC- cules are not sufficient to mediate the protective effect. Int dependent natural killer cell education. Trends Immunol. 2011; Immunol. 1995;7(3):393-400. 32(8):364-372. 52. Andersson S, Fauriat C, Malmberg J-A, Ljunggren H-G, 36. Horowitz A, Strauss-Albee DM, Leipold M, et al. Genetic and Malmberg K-J. KIR acquisition probabilities are independent environmental determinants of human NK cell diversity rev- of self-HLA class I ligands and increase with cellular KIR ealed by mass cytometry. Sci Transl Med. 2013;5(208):208ra145. expression. Blood. 2009;114(1):95-104. 37. Strauss-Albee DM, Fukuyama J, Liang EC, et al. Human NK 53. Beziat V, Traherne JA, Liu LL, et al. Influence of KIR gene cell repertoire diversity reflects immune experience and corre- copy number on natural killer cell education. Blood. 2013; lates with viral susceptibility. Sci Transl Med. 2015;7(297): 121(23):4703-4707. 297ra115. 54. Béziat V, Liu LL, Malmberg J-A, et al. NK cell responses to 38. Khakoo SI, Carrington M. KIR and disease: a model system or cytomegalovirus infection lead to stable imprints in the human system of models? Immunol Rev. 2006;214:186-201. KIR repertoire and involve activating KIRs. Blood. 2013; 39. Kulkarni S, Martin MP, Carrington M. The Yin and Yang of 121(14):2678-2688. HLA and KIR in human disease. Semin Immunol. 2008;20(6): 55. Dunphy SE, Guinan KJ, Chorcora CN, et al. 2DL1, 2DL2 and 343-352. 2DL3 all contribute to KIR phenotype variability on human 40. Meazza R, Falco M, Loiacono F, et al. Phenotypic and func- NK cells. Genes Immun. 2015;16(5):301-310. tional characterization of NK cells in alphabetaT-cell and B-cell 56. Béziat V, Hilton HG, Norman PJ, Traherne JA. Deciphering depleted Haplo-HSCT to cure pediatric patients with acute leu- the killer-cell immunoglobulin-like receptor system at super- kemia. Cancers. 2020;12(8):2187. resolution for natural killer and T-cell biology. Immunology. 41. Falco M, Romeo E, Marcenaro S, et al. Combined genotypic 2017;150(3):248-264. and phenotypic killer cell Ig-like receptor analyses reveal 57. Huhn O, Chazara O, Ivarsson MA, et al. High-resolution genetic KIR2DL3 alleles displaying unexpected monoclonal antibody and phenotypic analysis of KIR2DL1 alleles and their associa- reactivity: identification of the amino acid residues critical for tion with pre-eclampsia. J Immunol. 2018;201(9):2593-2601. staining. J Immunol. 2010;185(1):433-441. 58. Manser AR, Scherenschlich N, Thöns C, Hengel H, Timm J, 42. Litwin V, Gumperz J, Parham P, Phillips JH, Lanier LL. NKB1: Uhrberg M. KIR polymorphism modulates the size of the adap- a natural killer cell receptor involved in the recognition of poly- tive NK cell Pool in human cytomegalovirus–infected individ- morphic HLA-B molecules. J Exp Med. 1994;180(2):537-543. uals. J Immunol. 2019;203(8):2301-2309. 43. Vitale M, Carlomagno S, Falco M, et al. Isolation of a novel 59. Guma M, Angulo A, Vilches C, Gomez-Lozano N, Malats N, KIR2DL3-specific mAb: comparative analysis of the surface Lopez-Botet M. Imprint of human cytomegalovirus infection distribution and function of KIR2DL2, KIR2DL3 and KIR2DS2. on the NK cell receptor repertoire. Blood. 2004;104(12):3664- Int Immunol. 2004;16(10):1459-1466. 3671. 44. Della Chiesa M, Romeo E, Falco M, et al. Evidence that the 60. Nemat-Gorgani N, Hilton HG, Henn BM, et al. Different selected KIR2DS5 gene codes for a surface receptor triggering natural mechanisms attenuated the inhibitory interaction of KIR2DL1 killer cell function. Eur J Immunol. 2008;38(8):2284-2289. with C2(+) HLA-C in two indigenous human populations in 45. Falco M, Meazza R, Alicata C, et al. Epitope characterization of a southern Africa. J Immunol. 2018;200(8):2640-2655. monoclonal antibody that selectively recognizes KIR2DL1 allo- 61. Steiner NK, Dakshanamurthy S, Nguyen N, Hurley CK. Allelic types. HLA. 2022;100(2):107-118. variation of killer cell immunoglobulin-like receptor 2DS5 46. Gomez-Lozano N, Vilches C. Genotyping of human killer-cell impacts glycosylation altering cell surface expression levels. immunoglobulin-like receptor genes by polymerase chain Hum Immunol. 2014;75(2):124-128. 132 MEAZZA ET AL. 62. Della Chiesa M, Falco M, Bertaina A, et al. Human cytomega- 67. Carlsten M, Korde N, Kotecha R, et al. Checkpoint inhibition lovirus infection promotes rapid maturation of NK cells of KIR2D with the monoclonal antibody IPH2101 induces con- expressing activating killer Ig-like receptor in patients traction and Hyporesponsiveness of NK cells in patients with transplanted with NKG2C/ umbilical cord blood. myeloma. Clin Cancer Res. 2016;22(21):5211-5222. J Immunol. 2014;192(4):1471-1479. 63. Ruggeri L, Capanni M, Urbani E, et al. Effectiveness of donor natural killer cell alloreactivity in mismatched hematopoietic SU PP O R TI N G I N F O RMA TI O N transplants. Science. 2002;295(5562):2097-2100. Additional supporting information may be found in the 64. Locatelli F, Pende D, Falco M, Della Chiesa M, Moretta A, online version of the article at the publisher's website. Moretta L. NK cells mediate a crucial graft-versus-leukemia effect in haploidentical-HSCT to cure high-risk acute leukemia. Trends Immunol. 2018;39(7):577-590. How to cite this article: Meazza R, Falco M, 65. Fan QR, Long EO, Wiley DC. Crystal structure of the human Canevali P, et al. Characterization of KIR+ NK cell natural killer cell inhibitory receptor KIR2DL1-HLA-Cw4 com- subsets with a monoclonal antibody selectively plex. Nat Immunol. 2001;2(5):452-460. recognizing KIR2DL1 and blocking the specific 66. Vey N, Karlin L, Sadot-Lebouvier S, et al. A phase 1 study of interaction with HLA-C. HLA. 2022;100(2):119‐132. lirilumab (antibody against killer immunoglobulin-like receptor antibody KIR2D; IPH2102) in patients with solid tumors and doi:10.1111/tan.14640 hematologic malignancies. Oncotarget. 2018;9(25):17675-17688.