Received: 12 January 2022 Revised: 6 April 2022 Accepted: 8 April 2022 DOI: 10.1111/tan.14630 ORIGINAL ARTICLE Epitope characterization of a monoclonal antibody that selectively recognizes KIR2DL1 allotypes Michela Falco1 | Raffaella Meazza2 | Claudia Alicata3 | 2 4,5 1,6 Paolo Canevali | Aura Muntasell | Cristina Bottino | 3 Lorenzo Moretta | Daniela Pende2 | Miguel Lopez-Botet4,7 1 Laboratory of Clinical and Experimental Immunology, IRCCS Istituto Giannina Killer immunoglobulin-like receptor (KIR) genes code for a family of inhibitory Gaslini, Genoa, Italy and activating receptors, finely tuning NK cell function. Numerous studies 2 Laboratory of Immunology, IRCCS reported the relevance of KIR allelic polymorphism on KIR expression, ligand Ospedale Policlinico San Martino, Genoa, Italy affinity, and strength in signal transduction. Although KIR variability, including 3 Tumor Immunology Unit, IRCCS gene copy number and allelic polymorphism, in combination with HLA class I Ospedale Pediatrico Bambino Gesù, polymorphism, impacts both KIR expression and NK cell education, only a pre- Rome, Italy cise phenotypic analysis can define the size of the different KIRpos NK cell subsets. 4 Immunity and Infection laboratory, In this context, reagents recognizing a limited number of KIRs is essential. In this Hospital del Mar Medical Research Institute (IMIM), Barcelona, Spain study, we have characterized the specificity of an anti-KIR mAb termed HP-DM1. 5 Department of Cell Biology, Universitat Testing its binding to HEK-293T cells transfected with plasmids coding for differ- Autònoma de Barcelona, Bellaterra, Spain ent KIRs, we demonstrated that HP-DM1 mAb exclusively reacts with KIR2DL1. 6 Department of Experimental Medicine, Using site-directed mutagenesis, we identified the four amino acids relevant for DIMES, University of Genoa, Genoa, Italy 7 HP-DM1 recognition: M44, S67, R68, and T70. HP-DM1 mAb binds to a confor- Department of Medicine and Life Sciences, Universitat Pompeu Fabra, mational epitope including M44, the residue crucial for HLA-C K80 recognition Barcelona, Spain by KIR2DL1. Based on the HP-DM1 epitope characterization, we could extend its Correspondence reactivity to all KIR2DL1 allotypes identified except for KIR2DL1*022 and, most Michela Falco, Laboratory of Clinical and likely, KIR2DL1*020, predicting that it does not recognize any other KIR with the Experimental Immunology, IRCCS only exception of KIR2DS1*013. Moreover, by identifying the residues relevant Istituto Giannina Gaslini, 16147 Genoa, Italy. for HP-DM1 binding, continuously updating of its reactivity will be facilitated. Email: michelaemma.falco@gmail.com KEYWORDS Funding information killer immunoglobulin-like receptors, monoclonal antibodies, natural killer cells, Instituto de Salud Carlos III, Grant/Award polymorphism, site-directed mutagenesis Numbers: AIRC 5  1000 2018 id. 21147, FIS00/0181, H2020-MSCA-ITN- 2017-765104-MATURE-NK; Italian Ministry of Health Michela Falco and Raffaella Meazza shared credit. Miguel Lopez-Botet and Daniela Pende shared credit. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2022 The Authors. HLA: Immune Response Genetics published by John Wiley & Sons Ltd. HLA. 2022;100:107–118. wileyonlinelibrary.com/journal/tan 107 108 FALCO ET AL. 1 | INTRODUCTION least one KIR2DL2/L3 allele coding for a surface receptor. Differently, KIR2DL1 is present only in a fraction of the Natural killer (NK) cells are potent effectors of the innate centromeric regions, and the percentage of KIR2DL1 nega- immune system relevant for controlling viral infections tive individuals varies among the different ethnic groups and tumor development.1–3 Moreover, NK cells play ranging from 3% to 5% in Caucasian populations.31,37,38 a significant role in reproduction by contributing to KIR expression on the NK cell surface is determined maternal-fetal immune tolerance.4–7 NK cell activities are by genetic (i.e., KIR gene variability and HLA polymor- finely regulated by an array of signals transduced by phism), environmental (e.g., exposure to viral infections), germ-line encoded surface molecules with either inhibi- and stochastic factors.39–50 Moreover, phenotypical ana- tory or activating functions.8–13 During the education/ lyses revealed that the clonally distributed expression of licensing process, NK cells become fully functional by the individual KIRs leads to a great diversity of NK cell subsets expression of at least one inhibitory receptor recognizing displaying different KIR combinations. Most studies on self HLA class I molecule(s).14–16 Through this process, KIR phenotype had been primarily performed using anti- NK cells spare healthy cells, expressing physiological KIR mAbs and cytofluorimetric approaches.40,43,48,51–53 levels of HLA class I molecules (“self-tolerance”), and However, because of the high similarity between KIRs eliminate pathological cells that down-regulate HLA class encoded by different genes, the high level of polymor- I (“missing self-recognition”) and/or overexpressing or phism (especially among KIR genes coding for iKIRs), and expressing de novo ligands for activating receptors the elevated amino acid identity between iKIR and aKIR (“induced self-recognition”).16–21 NK cells sense HLA class recognizing the same ligands, the number of reagents spe- I expression via killer Ig-like receptors (KIR) and the cific for a single KIR is somewhat limited. CD94/NKG2 heterodimer, recognizing the highly poly- In this study, we characterized the reactivity of HP- morphic HLA class I classical allotypes and the non- DM1, an anti-KIR mAb selectively recognizing KIR2DL1 classical HLA-E molecules, respectively.9,11,22 The best allotypes. Using the site-directed mutagenesis approach, characterized inhibitory KIRs (iKIRs) include KIR2DL1 we have identified the four residues relevant for HP-DM1 and KIR2DL2/L3, which bind HLA-C molecules carrying binding. The characterization of this reagent paves the Lysine or Asparagine at position 80 (K80 or N80) respec- way for its use, alone or in combination with other anti- tively; KIR3DL1 recognizing both HLA-B and HLA-A allo- KIR antibodies, leading to more accurate identification of types sharing the Bw4 epitope, and KIR3DL2 sensing the different NK subpopulations. These studies may be HLA-A*03, -A*11, and HLA-F in an open conforma- relevant in basic research and clinical applications such tion.8,9,11,23–25 Thus, iKIRs specific for HLA-C sense virtu- as hematopoietic stem cell transplantation, autoimmune ally all allotypes that, except for five, are characterized by diseases, and pregnancy complications. K80 or N80 [Immuno Polymorphism Database (IPD)/HLA release 3.46, 2021]. On the contrary, most HLA-B and HLA-A allotypes are not ligands of iKIR molecules. The 2 | M E T HO DS AND M A T E R I ALS characterization of activating KIR (aKIR) ligands needs to be further clarified. Among them, KIR2DS1, KIR2DS2, 2.1 | Immunization and fusion and KIR3DS1 have extracellular domains similar to their inhibitory counterparts (i.e., KIR2DL1, KIR2DL2/L3, and BALB/c mice were immunized with a human NK clone KIR3DL1), with which they share the HLA ligands.25–27 expressing CD94, ILT2, and KIR2DL1 to obtain mAbs KIRs are coded by a gene family located on chromo- specific for receptors involved in the regulation of NK cell some 19q13.4 consisting of 13 genes and 2 pseudogenes cytotoxicity. Hybridomas were selected by their ability to closely packed in a head-to-head orientation and inherited modulate NK cell functions in reverse antibody- as a haplotype.28–30 More than 80 different KIR haplotypes, dependent cellular cytotoxicity (R-ADCC) assays against differing in gene content variability and allelic polymor- the P815 mastocytoma cell line, as previously phism, have been identified.28,30–33 Despite this, a con- described.54,55 Their specificity was assessed by flow cyto- served characteristic of almost all KIR haplotypes is the metry of polyclonal NK cell populations. presence of two regions, referred to as centromeric and telomeric based on their orientation, separated by a hot spot of recombination.30–36 HLA-C specific iKIR genes are 2.2 | Polyclonal NK cell populations located in the centromeric region, whereas those recogniz- ing HLA-B and HLA-A are in the telomeric one. KIR2DL2 Peripheral blood mononuclear cells were isolated and KIR2DL3 are alleles of the same locus (i.e., KIR2DL2/ by Lympholyte-H density gradient centrifugation L3), and virtually all individuals are characterized by at (Cedarlane, Burlington, Canada) from buffy coats of FALCO ET AL. 109 healthy donors, provided by the blood transfusion center KIRs were sequenced using a d-Rhodamine Terminator of IRCCS Ospedale Policlinico San Martino (Genoa, Cycle Sequencing kit and a 3100 ABI automatic sequencer Italy), following approved internal operational proce- (PerkinElmer, Wellesley, MA) to confirm the required sub- dures (IOH78). The donor KIR genotypes were analyzed stitutions. The produced 2DL1*002 mutants included the as previously described.52 NK cells were purified using following substitutions: M44K, M44T, R50H, S67G, R68P, the RosetteSep method (StemCell Technologies, Vancou- T70M, T70R, T70K, V90L, and M44T+S67G. Moreover, ver, BC) and cultured on irradiated feeder cells in the we generated 2DS1*001 R70T and 2DS1*002 K70T. presence of 2 μg/ml phytohemagglutinin (Sigma-Aldrich, Irvine, UK) and 600 IU/ml rIL-2 (Proleukin, Chiron Corp., Emeryville, CA) to obtain activated polyclonal NK 3 | RESULTS cells. NK cells were stained first with HP-DM1 mAb followed by anti-IgG1-PE (Southern Biotechnology, Bir- 3.1 | HP-DM1 mAb selectively mingham, AL), washed twice, and incubated with EB6B- recognizes KIR2DL1 APC (anti-KIR2DL1/S1 and anti-KIR2DL3 allotypes carrying E35 and R50, Beckman Coulter, Brea, CA). To obtain mAbs specific for receptors involved in the 143211-PE mAb (anti-KIR2DL1/S5, R&D Systems, regulation of NK cell cytotoxicity, we generated hybrid- Minneapolis, MN) in combination with EB6B-APC was omas from BALB/c mice immunized with a phenotypi- also used. Flow cytometric analysis was performed on cally characterized human NK clone. A mAb, hereafter Gallios flow-cytometer (Beckman Coulter), and data were termed HP-DM1 (IgG1), was selected for inhibiting the analyzed using FlowJo Version 10.7 (BD Biosciences, San NK cell clone cytotoxicity in R-ADCC and its reactivity José, CA). on NK cells appeared comparable to that of the anti- KIR2DL1/S1 EB6B mAb. Further studies with poly- clonal NK cells derived from donors with known KIR 2.3 | Transfection repertoires showed that HP-DM1 mAb displayed a reac- tivity similar to the anti-KIR2DL1/S5 143211 mAb. Dou- HEK-293T cells were transiently transfected using the ble stainings, combining HP-DM1 with the anti- linear polyethylenimine derivative jetPEI transfection KIR2DL1/S1 EB6B mAb or 143211 with EB6B mAb, reagent (Polyplus, New York, NY) following the manu- were performed (Figure 1A; hereafter, KIR will be omit- facturer's instruction. The pcDNA 3.1 plasmids used in ted by the acronyms). In particular, in 2DL1pos/2DS1neg/ the study code for: 2DL1*002, *003, *004, *012; 2DS5neg donors, 2DL1pos NK cells were co-stained by 2DL2*001; 2DL3*002, *005; 2DS1*001, *002; 2DS2*001, these mAbs; in 2DL1neg/2DS1pos/2DS5neg individuals, 2DS4*001, and 2DS5*002. Forty-eight hours after the no staining of HP-DM1 and 143211 was observed, while transfection, the cells were tested with a panel of anti- EB6B identified the 2DS1pos NK cells; in 2DL1pos/ KIR2D mAb followed by anti-IgG-PE or anti-IgM-PE 2DS1pos/2DS5neg donors, both 2DL1neg 2DS1pos (HP- (Southern Biotechnology) by immunofluorescence and DM1neg or 143211neg and EB6Bpos) and pos pos flow cytometric analysis (FACSCalibur and Cell Quest 2DL1 ± 2DS1 (co-stained by all the three mAbs) software, BD Biosciences). The mAbs used in these populations were detected (Figure 1A). experiments were: 143211 (R&D Systems, Minneapolis, To better define the HP-DM1 recognition pattern, we MN), 11PB6 (anti-KIR2DL1/S1 and anti-KIR2DL3 allo- tested its reactivity on HEK-293T cells transiently trans- types carrying E35 and R50, Miltenyi Biotec, Bergisch fected with plasmids coding for different lineage III Gladbach, Germany), HP-MA4 (anti-KIR2DL1/S1/S3/ KIR2D (namely, the KIR2D receptors composed of D1 S5, Biolegend, San Diego, CA), GL-183 (anti-KIR2DL2/ and D2 domains; Figure 1B and Figure S1). For compari- L3/S2, Beckman Coulter), and FES172 (anti-KIR2DS4, son, we selected three different mAbs recognizing 2DL1 Beckman Coulter). molecules (i.e., 143211, 11PB6, and HP-MA4) whose reac- tivity has already been characterized.51–53,56 Remarkably, HP-DM1 mAb stained exclusively 2DL1*002 transfected 2.4 | Mutagenesis cells, showing a pattern different from the other three mAbs (Figure 1B). Indeed, HP-DM1 mAb did not stain Mutagenesis was performed using the Gene Tailor Site- 2DS5*002, which conversely is recognized by 143211 and Directed Mutagenesis System (Invitrogen, Carlsbad, CA), HP-MA4 mAbs. 11PB6 mAb brightly stained transfected according to the manufacturer's instructions. The primers cells expressing 2DL1*002, 2DS1*001, 2DL3*005, as previ- include sequence coding for the amino acid we wanted to ously reported.53,56 Using 11PB6, we also obtained unex- replace (Table S1). The plasmids encoding the mutated pected dull staining of 2DS5*002pos transfectants, 110 FALCO ET AL. F I G U R E 1 HP-DM1 mAb selectively recognizes KIR2DL1 transfected cells. (A) Double fluorescence analyses were performed using activated NK cell populations derived from representative donors with defined KIR gene profiles (positive genes are identified as gray, negative as white boxes). Cells were stained with HP-DM1 and EB6B-APC (upper panels) or with EB6B-APC and 143211-PE mAbs (lower panels). (B) HEK-293T cells, transiently transfected with plasmids coding for the indicated KIR2D alleles, were stained with HP-DM1, 143211, 11PB6, and HP-MA4 mAb, followed by the PE-conjugated goat anti-mouse IgG (filled profiles). Empty profiles represent staining with the second reagent alone. A representative of five independent experiments is shown FALCO ET AL. 111 probably because of the high level of 2DS5 expression. extracellular regions of 2DL1 and 2DS1, we firstly Our data confirmed that HP-MA4 mAb recognizes focused our analysis on the amino acid positions that dif- 2DL1*002, 2DS1*001, and 2DS5*002,53 and extended its fer between these two KIRs. The alignment of the characterization showing that it does not recognize sequences of four tested HP-DM1 positive 2DL1 allotypes 2DL3*005. None of the selected mAbs reacted with (namely 2DL1*002, *003, *004, and *012) and two HP- 2DL2*001, 2DL3*002, 2DS2*001, and 2DS4*001 DM1 negative 2DS1 (i.e., 2DS1*001 and *002) revealed (Figure S1). that the HP-DM1 epitope should include residue(s) These data indicated that HP-DM1 mAb is a unique located in the D1 domain (Figure 2). Indeed, 2DL1*004, reagent selectively recognizing 2DL1. strongly recognized by HP-DM1 mAb, has a D2 domain identical to 2DS1*001 and 2DS1*002 (Figure 2). The com- parison led us to identify Threonine 70 (T70) and Valine 3.2 | Amino acid residue T70 is crucial 90 (V90) as possible relevant residues for HP-DM1 bind- for HP-DM1 reactivity ing. Thus, we introduced in the 2DL1*002 backbone the amino acid present at these positions in 2DS1 allotypes To better define the HP-DM1 mAb reactivity, we investi- and determined the effect of the mutations on HP-DM1 gated the residues relevant for its binding. Since a mAb reactivity. The replacement of T70 with either high level of amino acid identity characterizes the Lysine (K) (present in most of the 2DS1 allotypes) or F I G U R E 2 HP-DM1 epitope includes residue(s) located in the first KIR2DL1 domain. (A) Alignment of KIR2D extracellular regions of three 2DL1 and two 2DS1 allotypes of known HP-DM1 mAb reactivity. The numbers indicate the amino acid positions present in the mature proteins. Arrows indicate the amino acid positions considered for HP-DM1 mAb epitope characterization. (B) HEK-293T cells transiently transfected with the plasmid coding for the 2DL1*003, *004, and *012 allotypes were surface stained with the indicated mAbs followed by the PE-conjugated goat anti-mouse IgG second reagent (filled profiles). Empty profiles represent staining with the second reagent alone. A representative of five independent experiments is reported 112 FALCO ET AL. Arginine (R) (found in 2DS1*001) completely abrogated we used 11PB6 as control mAb to confirm the surface HP-DM1 mAb reactivity, while the replacement of V90 expression of the 2DL1 mutants. Conversely, the replace- with Leucine did not (Figure 3A). In these experiments, ment in 2DS1 backbones of either R70 or K70 with T determined a strong HP-DM1 mAb recognition (Figure 3B). The data showed that T70, a residue present in all the 2DL1 allotypes identified so far, is crucial for HP-DM1 binding. 3.3 | HP-DM1 epitope also includes M44, S67, and R68 Taking into account the relevance of T70 for HP-DM1 mAb binding, we focused on the amino acids present in the other KIRs belonging to lineage III that are not recognized by HP-DM1. As all 2DL2/L3, 2DS2, and 2DS4 allotypes are characterized by a Methionine at position 70 (M70), we examined if it negatively affects HP-DM1 mAb binding by the generation of the 2DL1*002 T70M mutant. As shown in Figure 3C, replacing T70 with M in 2DL1 partially reduced HP- DM1 mAb reactivity, suggesting that other residues may contribute to the HP-DM1 epitope. This possibility was further supported by the observation that, although all the 2DS5 allotypes are characterized by T70, 2DS5*002 transfected cells were not stained by HP-DM1 mAb (Figure 1B). First, we investigated why 2DL2/L3, 2DS2, and 2DS4 molecules did not react with HP-DM1 mAb. Comparing the D1 amino acid sequences of the KIRs mentioned above with 2DL1, we identified four additional residues (M44, R50, S67, and R68) present only in 2DL1 (Figure S2). To analyze the possible involvement of these positions, we generated four novel 2DL1*002 mutants (i.e., M44K, R50H, S67G, and R68P). As shown in Figure 4A, M44K and R68P abrogated HP-DM1 recognition, while S67G sharply decreased HP-DM1 staining. On the contrary, the replacement of R50 with H did not interfere with HP- DM1 mAb binding. These data demonstrated that M44, S67, and R68 are relevant for HP-DM1 binding, explaining why this mAb does not react with 2DL2, 2DL3, 2DS2, and 2DS4 allotypes. In these experiments, we used 11PB6 mAb to control the surface expression of the transfected mutants. As expected, since 11PB6 F I G U R E 3 Threonine 70 is part of HP-DM1 epitope. (A, B, and C) HEK-293T cells transiently transfected with the indicated plasmids were surface stained with the HP-DM1 and 11PB6 mAb followed by the PE-conjugated goat anti-mouse IgG second reagent (filled profiles). Empty profiles represent staining with the second reagent alone. The residues present in wild-type KIR are indicated in brackets. A representative of five independent experiments is shown FALCO ET AL. 113 Then we examined the four residues critical for HP- DM1 binding (i.e., M44, S67, R68, and T70) in 2DS5 allo- types (Figure S2). Since all 2DS5 allotypes, except for 2DS5*014, are characterized by R68 and T70 (that are rel- evant for HP-DM1 binding), G67 (that marginally affects the HP-DM1 staining), and Threonine 44 (T44), we gen- erated 2DL1*002 M44T mutant to test if T44 could abro- gate HP-DM1 mAb recognition. As shown in Figure 4B, M44T replacement decreased, but not abolished, HP- DM1 mAb staining. Finally, we tested if the combined presence of T44 and G67 (the two amino acid residues that reduce HP-DM1 mAb binding) had an additive effect. To this end, we generated the 2DL1*002 M44T S67G double mutant, which was no longer recognized by HP-DM1 mAb (Figure 4B). Taken together, our data indi- cated that M44, S67, R68, and T70 are the four amino acids relevant for HP-DM1 recognition (Table 1). The identification of the residues relevant for HP- DM1 mAb binding allows predicting that this mAb exclu- sively recognizes all 2DL1 allotypes present in the IPD- KIR database (release 2.10.0, 2020) except for 2DL1*022 and, most likely, 2DL1*020. Indeed, 2DL1*022 codes for a mature protein identical to the HP-DM1 negative 2DL1*002 M44K mutant (Figure S3 and Figure 4), and 2DL1*020 encodes for a receptor that carries G67 that we know to decrease the HP-DM1 binding (Figure S3 and Figure 4). 3.4 | The HP-DM1 and 143211 mAb epitopes partially overlap Finally, we investigated the residues relevant for 143211 mAb specific binding to 2DL1 and 2DS5. To this end, we used an approach similar to that previously described for HP-DM1. We focused on T70 (Figure S4), a residue pre- sent in the extracellular portions of 2DL1 and 2DS5 but absent in the other lineage III KIRs. We tested the 143211 mAb reactivity on 2DL1*002 T70M, T70R, and T70K mutants. As shown in Figure 5, the replacement of T70 with either R or K virtually abrogated the 143211 mAb recognition, as also occurred for HP-DM1 mAb (see Figure 3A), while the substitution with M had no effect. F I G U R E 4 HP-DM1 recognition is also influenced by M44, S67, and R68. (A, B) HEK-293T cells transiently transfected with Since all the tested KIRs characterized by M70 have P68 the indicated plasmids were surface stained with the HP-DM1 and (Figure S4), we further investigated if position 68 11PB6 mAbs followed by the PE-conjugated goat anti-mouse IgG influenced the 143211 binding. The results obtained ana- second reagent (filled profiles). Empty profiles represent staining lyzing the reactivity of 2DL1*002 R68P mutant revealed with the second reagent alone. A representative of four that the presence of P68 almost completely abrogated independent experiments is shown 143211 staining (Figure 5). In this set of experiments, 11PB6 mAb was chosen as transfection control mAb and epitope includes Glutamic Acid and Arginine at posi- the obtained profiles (not shown) were similar to those tions 35 and 50, respectively, this mAb labeled all the reported in Figures 3A,C and 4A. Being both T70 and 2DL1*002 mutants except for 2DL1*002 R50H.52 R68 residues relevant for HP-DM1 binding, we conclude 114 FALCO ET AL. TABLE 1 Anti-KIR mAbs used in the study Clone name KIR specificity Residues relevant for mAb binding References HP-DM1 2DL1 (except for 2DL1*020, *022), 2DS1*013 M44, S67, R68, T70 This study 8,51,52,56,57 EB6B 2DL1, 2DS1, 2DL3 E35 and R50 E35, R50 51,52 11PB6 2DL1, 2DS1, 2DL3 E35 and R50 E35, R50 143211 2DL1, 2DS5 R68, T70 This study51–53,56 53 HP-MA4 2DL1, 2DS1, 2DS3, 2DS5 N.D. 51,56,57 GL-183 2DL2, 2DL3, 2DS2 N.D. 58 FES172 2DS4 N.D. Note: ND, not determined. F I G U R E 5 Identification of the residues relevant for 143211 mAb binding. HEK-293T cells transiently transfected with the indicated 2DL1*002 mutants were surface stained with the 143211 mAb followed by the PE-conjugated goat anti-mouse IgG second reagent (filled profiles). Empty profiles represent staining with the second reagent alone. A representative of five independent experiments is shown that the amino acid positions required for HP-DM1 and of the other available anti-KIR mAbs, identifying the resi- 143211 mAb recognition partially overlap (Table 1). dues relevant for its binding. Our approach was based on assessing HP-DM1 reactivity performed by cyto- fluorimetric analysis of HEK-293T cells transfected with 4 | DISCUSSION plasmids coding for different KIRs or 2DL1 mutants. Introducing into the 2DL1*002 backbone the amino acids Recently, progress has been achieved in understanding present in HP-DM1 negative allotypes, we identified four the impact of allelic KIR variability, KIR gene copy num- residues, namely M44, S67, R68, and T70, relevant for ber, KIR ligand polymorphism (i.e., HLA class I), and HP-DM1 reactivity. The characterization of HP-DM1 epi- viral infections on KIR expression in different individ- tope allowed us to predict the mAb reactivity with all uals.14,40,42–46,48,50 The cytofluorimetric analysis repre- KIR allotypes present in the IPD-KIR database. In partic- sents the leading approach to define the presence and ular, we inferred that HP-DM1 mAb should recognize all proportions of the different NK cell subsets. However, the the currently identified 2DL1 allotypes except for panel of available anti-KIR monoclonal antibodies is lim- 2DL1*022 and, most likely, 2DL1*020. Lack of HP-DM1 ited, and they often recognize allotypes encoded by more binding to 2DL1*022 was demonstrated by the negative than one KIR gene. As far as we know, only anti-KIR staining obtained testing the 2DL1*002 M44K mutant, mAbs selectively recognizing 2DS4, 3DL1, 2DL3, 2DL4, which encodes for a mature protein identical to and 2DL5 molecules are available. The major obstacle in 2DL1*022 (Figure 4 and Figure S3). Since the dimor- KIR-specific antibody production is the high degree of phism M/K at position 44 determines the HLA-C K80 amino acid identity among different KIRs, resulting in a and N80 specificity, respectively,59 the 2DL1*022 allotype low number of specific residues, further complicated by recognizes HLA-C N80 molecules.60–62 2DL1*022 arose the allelic polymorphism that characterizes each locus of by non-synonymous point mutation most likely from this multigene family. In this study, we characterized the 2DL1*001 and was detected in sub-Saharan African specificity of the HP-DM1 mAb, which differs from that populations (ranging 17.2% in KhoeSan), but not in FALCO ET AL. 115 other ethnic groups.60,61 Considering the 2DL1*022 distri- functional studies, allowing the analysis of the 2DL1 con- bution, the inability of HP-DM1 to bind 2DL1*022 would tribution in cell target recognition. This prediction was be relevant whenever sub-Saharan African individuals supported by experimental evidence.66 In this study, are analyzed. Moreover, even in the presence of characterizing the 143211 epitope, we revealed that the 2DL1*022, HP-DM1 mAb can be a useful reagent since it two residues involved in the 143211 recognition (i.e., R68 allows the restricted identification of the NK cell subsets and T70) were also relevant for the HP-DM1 binding. expressing iKIRs recognizing an HLA-C K80 ligand. This information should be considered when performing 2DL1*020 was also detected in sub-Saharan African stainings with the simultaneous use of 143211 and HP- populations and codes for a receptor characterized by DM1 mAbs. G67 that we proved to diminish (although not abolish) The more precise detection of 2DL1pos NK cells can the recognition of HP-DM1 on transfectants. Unfortu- be relevant for both basic research and clinical studies. In nately, we did not have the opportunity to confirm this the haploidentical hematopoietic stem cell transplanta- prediction by testing NK cells derived from a 2DL1*020pos tion setting, patients receiving the graft from an NK donor. Being the level of KIR expression on NK cells alloreactive donor (i.e., with a population of NK cells lower than that obtained after plasmid transfection, we expressing only an inhibitory receptor for HLA molecules tentatively included 2DL1*020 allotype among the mole- present in the donor but absent in the recipient) have a cules not recognized by HP-DM1. A similar analysis on better outcome.67,68 Among the alloreactive NK 2DS1 alleles, revealed that the HP-DM1 mAb failed to populations identified using this model (namely, bind all 2DS1 except for 2DS1*013, which is the only one KIR/KIR-L mismatch in graft versus host direction), one carrying T70. The HP-DM1 binding to 2DS1*013 was ver- is characterized by the expression of 2DL1 receptors ified by positive staining of the mutant 2DS1*002 K70T, only.69,70 In this context, HP-DM1 should allow a better which encodes for a mature protein identical to definition of the alloreactive NK population. Moreover, 2DS1*013 (Figure 3 and Figure S3). No information on the use of the HP-DM1 mAb in combination with other the 2DS1*013 frequency or the ethnic origin of the donor available anti-KIR mAbs allows more precise detection of in which this allele was found is available. Furthermore, the different KIRpos NK cell subsets.66 Pregnancy disorder none of the 2DL2, 2DL3, 2DS2, 2DS4, and 2DS5 studies are another field that could take advantage of the described allotypes include all the amino acids relevant availability of HP-DM1. In the first trimester of gestation, for HP-DM1 binding. Although it has been documented uterine NK cells are present in the decidua playing a rele- that 2DS3 does not reach the cell surface,53,63 we also ver- vant role in the placentation. Several studies demon- ified whether HP-DM1 could stain the allotypes encoded strated that women with a high risk of pregnancy by this locus. All 2DS3 alleles in the database are charac- disorders have two KIR A haplotypes (namely two copies terized by R70, which prevented HP-DM1 binding of 2DL1 alleles coding for “strong receptors”) and bear an (Figure 3). The data obtained in this study indicated that HLA-C K80 fetus.71,72 HP-DM1 mAb is a unique reagent. Indeed, the other available 2DL1 binding mAbs also react with other KIRs. AUTHOR CONTRIBUTIONS In particular, EB6B and 11PB6 also recognize 2DS1 and Michela Falco designed the study, analyzed and inter- 2DL3 allotypes carrying E35 and R50, while 143211 also preted the data. Miguel Lopez-Botet generated and binds 2DS5.8,52,53,56 Moreover, since the number of KIR selected HP-DM1 mAb, proposing its collaborative charac- alleles is expected to increase, based on the introduction terization. Michela Falco and Claudia Alicata performed of next-generation sequencing approaches, knowledge on mutagenesis and transfection experiments. Raffaella the residues required for HP-DM1 binding will allow con- Meazza and Paolo Canevali performed FACS analysis of tinuing to predict/revise its reactivity. polyclonal NK cell populations. Michela Falco, Raffaella The positions of the residues relevant for HP-DM1 Meazza, Daniela Pende, and Miguel Lopez-Botet wrote the binding (i.e., 44, 67, 68, and 70) indicated that the epitope manuscript. Claudia Alicata, Aura Muntasell, Cristina Bot- recognized by this mAb is conformational. This hypothe- tino, and Lorenzo Moretta critically revised the manu- sis is supported by the analysis of the 2DL1 crystal struc- script. Michela Falco, Cristina Bottino, Lorenzo Moretta, ture, showing that M44, R68, and T70 (three out of four Daniela Pende, and Miguel Lopez-Botet provided funds. residues involved in the HP-DM1 binding) are in the All the authors approved the final version. solvent-accessible area at close contact with each other.64 Moreover, mutational studies have demonstrated the rel- ACKNOWLEDGMENTS evance of 2DL1 M44 and T70 in HLA-C K80 recogni- This work was supported by Italian Ministry of Health tion.59,65 Consequently, HP-DM1 should mask receptor/ (Ricerca Corrente G. Gaslini, Ricerca Corrente Ospedale ligand contacts and become a suitable reagent also for Policlinico San Martino) (Cristina Bottino and Daniela 116 FALCO ET AL. Pende), Instituto de Salud Carlos III, FIS00/0181 (to Miguel 10. Moretta A, Bottino C, Vitale M, et al. Activating receptors and Lopez-Botet), H2020-MSCA-ITN-2017-765104-MATURE- coreceptors involved in human natural killer cell-mediated NK (to Miguel Lopez-Botet and Daniela Pende), AIRC cytolysis. Annu Rev Immunol. 2001;19:197-223. 11. Parham P. MHC class I molecules and KIRs in human history, 51000 2018 id. 21147 (Lorenzo Moretta). health and survival. Nat Rev Immunol. 2005;5(3):201-214. 12. Long EO, Kim HS, Liu D, Peterson ME, Rajagopalan S. Con- CONFLICT OF INTEREST trolling natural killer cell responses: integration of signals for Miguel Lo pez-Botet declares that HP-DM1 is commer- activation and inhibition. Annu Rev Immunol. 2013;31:227-258. cially available through a license agreement signed by 13. Barrow AD, Martin CJ, Colonna M. The natural cytotoxicity University Pompeu Fabra with Biolegend®. The other receptors in health and disease. Front Immunol. 2019;10:909. authors declare no conflict of interest. 14. Anfossi N, Andre P, Guia S, et al. Human NK cell education by inhibitory receptors for MHC class I. Immunity. 2006;25(2): 331-342. DATA AVAILABILITY STATEMENT 15. Orr MT, Lanier LL. Natural killer cell education and tolerance. The data that support the findings of this study are avail- Cell. 2010;142(6):847-856. able from the corresponding author upon reasonable 16. Boudreau JE, Hsu KC. Natural killer cell education in human request. health and disease. Curr Opin Immunol. 2018;50:102-111. 17. Karre K. NK cells, MHC class I molecules and the missing self. ORCID Scand J Immunol. 2002;55(3):221-228. Michela Falco https://orcid.org/0000-0002-0317-9830 18. Yawata M, Yawata N, Draghi M, Partheniou F, Little AM, Parham P. MHC class I-specific inhibitory receptors and Raffaella Meazza https://orcid.org/0000-0003-0242- their ligands structure diverse human NK-cell repertoires 3157 toward a balance of missing self-response. Blood. 2008;112(6): Claudia Alicata https://orcid.org/0000-0003-2053-626X 2369-2380. Paolo Canevali https://orcid.org/0000-0002-0029-1376 19. Hilton HG, Parham P. Missing or altered self: human NK cell Aura Muntasell https://orcid.org/0000-0003-2894-0486 receptors that recognize HLA-C. Immunogenetics. 2017;69(8–9): Cristina Bottino https://orcid.org/0000-0001-6695-1739 567-579. Lorenzo Moretta https://orcid.org/0000-0003-4658-1747 20. Vitale M, Cantoni C, Della Chiesa M, et al. An historical over- Daniela Pende https://orcid.org/0000-0003-1565-451X view: the discovery of how NK cells can kill enemies, recruit defense troops, and more. Front Immunol. 2019;10:1415. Miguel Lopez-Botet https://orcid.org/0000-0003-4882- 21. Dębska-Zielkowska J, Moszkowska G, Zieli nski M, et al. KIR 065X receptors as key regulators of NK cells activity in health and disease. Cell. 2021;10(7):1777. R EF E RE N C E S 22. Braud VM, Allan DS, O'Callaghan CA, et al. HLA-E binds to 1. Vivier E, Tomasello E, Baratin M, Walzer T, Ugolini S. natural killer cell receptors CD94/NKG2A, B and C. Nature. Functions of natural killer cells. Nat Immunol. 2008;9(5): 1998;391(6669):795-799. 503-510. 23. Goodridge JP, Burian A, Lee N, Geraghty DE. HLA-F and 2. Meza Guzman LG, Keating N, Nicholson SE. Natural killer MHC class I open conformers are ligands for NK cell Ig-like cells: tumor surveillance and signaling. Cancer. 2020;12(4):952. receptors. J Immunol. 2013;191(7):3553-3562. 3. Waggoner SN, Reighard SD, Gyurova IE, et al. Roles of natural 24. Falco M, Moretta L, Moretta A, Bottino C. KIR and KIR ligand killer cells in antiviral immunity. Curr Opin Virol. 2016;16:15-23. polymorphism: a new area for clinical applications? Tissue 4. Trowsdale J, Moffett A. NK receptor interactions with MHC Antigens. 2013;82(6):363-373. class I molecules in pregnancy. Semin Immunol. 2008;20(6): 25. Moesta AK, Parham P. Diverse functionality among human 317-320. NK cell receptors for the C1 epitope of HLA-C: KIR2DS2, 5. Hiby SE, Apps R, Sharkey AM, et al. Maternal activating KIRs KIR2DL2, and KIR2DL3. Front Immunol. 2012;3:336. protect against human reproductive failure mediated by fetal 26. Chewning JH, Gudme CN, Hsu KC, Selvakumar A, Dupont B. HLA-C2. J Clin Invest. 2010;120(11):4102-4110. KIR2DS1-positive NK cells mediate Alloresponse against the C2 6. Moffett A, Colucci F. Co-evolution of NK receptors and HLA HLA-KIR ligand group in vitro. J Immunol. 2007;179(2):854-868. ligands in humans is driven by reproduction. Immunol Rev. 27. Carlomagno S, Falco M, Bono M, et al. KIR3DS1-mediated rec- 2015;267(1):283-297. ognition of HLA-*B51: modulation of KIR3DS1 responsiveness 7. Huhn O, Zhao X, Esposito L, Moffett A, Colucci F, by self HLA-B Allotypes and effect on NK cell licensing. Front Sharkey AM. How do uterine natural killer and innate lym- Immunol. 2017;8:581. phoid cells contribute to successful pregnancy? Front Immunol. 28. Wilson MJ, Torkar M, Haude A, et al. Plasticity in the organi- 2021;12:607669. zation and sequences of human KIR/ILT gene families. Proc 8. Moretta A. Existence of both inhibitory (p58) and activatory Natl Acad Sci U S A. 2000;97(9):4778-4783. (p50) receptors for HLA-C molecules in human natural killer 29. Trowsdale J, Barten R, Haude A, Stewart CA, Beck S, cells. J Exp Med. 1995;182(3):875-884. Wilson MJ. The genomic context of natural killer receptor 9. Moretta A, Bottino C, Vitale M, et al. Receptors for HLA class-I extended gene families. Immunol Rev. 2001;181:20-38. molecules in human natural killer cells. Annu Rev Immunol. 30. Hsu KC, Liu XR, Selvakumar A, Mickelson E, O'Reilly RJ, 1996;14:619-648. Dupont B. Killer Ig-like receptor haplotype analysis by gene FALCO ET AL. 117 content: evidence for genomic diversity with a minimum of six 46. Gagne K, Willem C, Legrand N, et al. Both the nature of basic framework haplotypes each with multiple subsets. KIR3DL1 alleles and the KIR3DL1/S1 allele combination affect J Immunol. 2002;169(9):5118-5129. the KIR3DL1 NK-cell repertoire in the French population. Eur 31. Jiang W, Johnson C, Jayaraman J, et al. Copy number variation J Immunol. 2013;43(4):1085-1098. leads to considerable diversity for B but not a haplotypes of the 47. Della Chiesa M, Sivori S, Carlomagno S, Moretta L, Moretta A. human KIR genes encoding NK cell receptors. Genome Res. Activating KIRs and NKG2C in viral infections: toward NK cell 2012;22(10):1845-1854. memory? Front Immunol. 2015;6:573. 32. Pyo CW, Guethlein LA, Vu Q, et al. Different patterns of evolu- 48. Dunphy SE, Guinan KJ, Chorcora CN, et al. 2DL1, 2DL2 and tion in the centromeric and telomeric regions of group A and B 2DL3 all contribute to KIR phenotype variability on human haplotypes of the human killer cell Ig-like receptor locus. PLoS NK cells. Genes Immun. 2015;16(5):301-310. One. 2010;5(12):e15115. 49. Horowitz A, Djaoud Z, Nemat-Gorgani N, et al. Class I HLA 33. Pyo CW, Wang R, Vu Q, et al. Recombinant structures expand haplotypes form two schools that educate NK cells in different and contract inter and intragenic diversification at the KIR ways. Science Immunol. 2016;1(3):eaag1672. locus. BMC Genomics. 2013;14:89. 50. Saunders PM, Pymm P, Pietra G, et al. Killer cell 34. Uhrberg M, Valiante NM, Shum BP, et al. Human diversity in immunoglobulin-like receptor 3DL1 polymorphism defines dis- killer cell inhibitory receptor genes. Immunity. 1997;7(6): tinct hierarchies of HLA class I recognition. J Exp Med. 2016; 753-763. 213(5):791-807. 35. Norman PJ, Abi-Rached L, Gendzekhadze K, et al. Meiotic 51. Pende D, Marcenaro S, Falco M, et al. Anti-leukemia recombination generates rich diversity in NK cell receptor activity of alloreactive NK cells in KIR ligand- genes, alleles, and haplotypes. Genome Res. 2009;19(5):757-769. mismatched haploidentical HSCT for pediatric patients: 36. Vendelbosch S, de Boer M, van Leeuwen K, et al. Novel evaluation of the functional role of activating KIR and insights in the genomic organization and hotspots of recombi- redefinition of inhibitory KIR specificity. Blood. 2009; nation in the human KIR locus through analysis of intergenic 113(13):3119-3129. regions. Genes Immun. 2015;16(2):103-111. 52. Falco M, Romeo E, Marcenaro S, et al. Combined genotypic 37. Cisneros EMM, Gomez-Lozano N, Muntasell A, L opez- and phenotypic killer cell Ig-like receptor analyses reveal Botet M, Vilches C. Haplotype-based analysis of KIR-gene pro- KIR2DL3 alleles displaying unexpected monoclonal antibody files in a south European population-distribution of standard reactivity: identification of the amino acid residues critical for and variant haplotypes, and identification of novel recombi- staining. J Immunol. 2010;185(1):433-441. nant structures. Front Immunol. 2020;11:440. 53. Czaja K, Borer AS, Schmied L, Terszowski G, Stern M, 38. Solloch UV, Schefzyk D, Schäfer G, et al. Estimation of Ger- Gonzalez A. A comprehensive analysis of the binding of anti-KIR man KIR allele group haplotype frequencies. Front Immunol. antibodies to activating KIRs. Genes Immun. 2014;15(1):33-37. 2020;11:429. 54. Aramburu JBM, Ramírez A, Silva A, et al. A novel functional 39. Valiante NM, Uhrberg M, Shilling HG, et al. Functionally and cell surface dimer (Kp43) expressed by natural killer cells and structurally distinct NK cell receptor repertoires in the periph- gamma/delta TCR+ T lymphocytes. II. Modulation of natural eral blood of two human donors. Immunity. 1997;7(6):739-751. killer cytotoxicity by anti-Kp43 monoclonal antibody. 40. Gardiner CM, Guethlein LA, Shilling HG, et al. Different NK J Immunol. 1990;15(144):3238-3247. cell surface phenotypes defined by the DX9 antibody are due to 55. Pérez-Villar JJCM, Navarro F, Melero I, et al. Biochemical and KIR3DL1 gene polymorphism. J Immunol. 2001;166(5):2992- serologic evidence for the existence of functionally distinct 3001. forms of the CD94 NK cell receptor. J Immunol. 1996;157(12): 41. Yu J, Heller G, Chewning J, Kim S, Yokoyama WM, Hsu KC. 5367-5374. Hierarchy of the human natural killer cell response is deter- 56. Beziat V, Traherne J, Malmberg JA, et al. Tracing dynamic mined by class and quantity of inhibitory receptors for self- expansion of human NK-cell subsets by high-resolution analy- HLA-B and HLA-C ligands. J Immunol. 2007;179(9):5977-5989. sis of KIR repertoires and cellular differentiation. Eur J 42. Fauriat C, Ivarsson MA, Ljunggren HG, Malmberg KJ, Immunol. 2014;44(7):2192-2196. Michaelsson J. Education of human natural killer cells by acti- 57. Biassoni R, Cantoni C, Falco M, et al. The HLA-C-specific vating killer cell immunoglobulin-like receptors. Blood. 2010; "activatory" or "inhibitory" natural killer cell receptors display 115(6):1166-1174. highly homologous extracellular domains but differ in their 43. Beziat V, Traherne JA, Liu LL, et al. Influence of KIR gene transmembrane and intracytoplasmic portions. J Exp Med. copy number on natural killer cell education. Blood. 2013; 1996;183(2):645-650. 121(23):4703-4707. 58. Della Chiesa M, Romeo E, Falco M, et al. Evidence that the 44. Charoudeh HN, Terszowski G, Czaja K, Gonzalez A, KIR2DS5 gene codes for a surface receptor triggering natural Schmitter K, Stern M. Modulation of the natural killer cell KIR killer cell function. Eur J Immunol. 2008;38(8):2284-2289. repertoire by cytomegalovirus infection. Eur J Immunol. 2013; 59. Winter CC, Gumperz JE, Parham P, Long EO, Wagtmann N. 43(2):480-487. Direct binding and functional transfer of NK cell inhibitory 45. Djaoud Z, David G, Bressollette C, et al. Amplified NKG2C+ receptors reveal novel patterns of HLA-C allotype recognition. NK cells in cytomegalovirus (CMV) infection preferentially J Immunol. 1998;161(2):571-577. express killer cell Ig-like receptor 2DL: functional impact in 60. Hilton HG, Norman PJ, Nemat-Gorgani N, et al. Loss and gain controlling CMV-infected dendritic cells. J Immunol. 2013; of natural killer cell receptor function in an African hunter- 191(5):2708-2716. gatherer population. PLoS Genet. 2015;11(8):e1005439. 118 FALCO ET AL. 61. Nemat-Gorgani N, Hilton HG, Henn BM, et al. Different hematopoietic transplantation for acute myeloid leukemia: chal- selected mechanisms attenuated the inhibitory interaction of lenging its predictive value. Blood. 2007;110(1):433-440. KIR2DL1 with C2(+) HLA-C in two indigenous human 69. Locatelli F, Pende D, Falco M, Della Chiesa M, Moretta A, populations in southern Africa. J Immunol. 2018;200(8):2640- Moretta L. NK cells mediate a crucial graft-versus-leukemia 2655. effect in Haploidentical-HSCT to cure high-risk acute leuke- 62. Beziat V, Hilton HG, Norman PJ, Traherne JA. Deciphering mia. Trends Immunol. 2018;39(7):577-590. the killer-cell immunoglobulin-like receptor system at super- 70. Meazza R, Falco M, Loiacono F, et al. Phenotypic and func- resolution for natural killer and T-cell biology. Immunology. tional characterization of NK cells in αβT-cell and B-cell 2017;150(3):248-264. depleted Haplo-HSCT to cure pediatric patients with acute leu- 63. VandenBussche CJ, Mulrooney TJ, Frazier WR, kemia. Cancer. 2020;12(8):2187. Dakshanamurthy S, Hurley CK. Dramatically reduced surface 71. Hiby SE, Walker JJ, O'Shaughnessy KM, et al. Combinations of expression of NK cell receptor KIR2DS3 is attributed to multi- maternal KIR and fetal HLA-C genes influence the risk of pre- ple residues throughout the molecule. Genes Immun. 2009; eclampsia and reproductive success. J Exp Med. 2004;200(8): 10(2):162-173. 957-965. 64. Fan QR, Long EO, Wiley DC. Crystal structure of the human 72. Hilton HG, Guethlein LA, Goyos A, et al. Polymorphic HLA-C natural killer cell inhibitory receptor KIR2DL1-HLA-Cw4 com- receptors balance the functional characteristics of KIR haplo- plex. Nat Immunol. 2001;2(5):452-460. types. J Immunol. 2015;195(7):3160-3170. 65. Biassoni R, Pessino A, Malaspina A, et al. Role of amino acid position 70 in the binding affinity of p50.1 and p58.1 receptors SU PP O R TI N G I N F O RMA TI O N for HLA-Cw4 molecules. Eur J Immunol. 1997;27(12):3095-3099. Additional supporting information may be found in the 66. Meazza R, Falco M, Canevali P, et al. Characterization of KIR+ NK cell subsets with a monoclonal antibody selectively online version of the article at the publisher's website. recognizing KIR2DL1 and blocking the specific interaction with HLA-C. HLA. 2022;100(2):119-132. How to cite this article: Falco M, Meazza R, 67. Ruggeri L, Capanni M, Mancusi A, Martelli MF, Velardi A. The impact of donor natural killer cell alloreactivity on alloge- Alicata C, et al. Epitope characterization of a neic hematopoietic transplantation. Transpl Immunol. 2005; monoclonal antibody that selectively recognizes 14(3–4):203-206. KIR2DL1 allotypes. HLA. 2022;100(2):107‐118. 68. Ruggeri L, Mancusi A, Capanni M, et al. Donor natural killer doi:10.1111/tan.14630 cell allorecognition of missing self in haploidentical