KIR Polymorphism Modulates the Size of the Adaptive NK Cell Pool in Human C Ytomegalovirus−Infected Individuals

KIR Polymorphism Modulates the Size of the Adaptive NK Cell Pool in Human C ytomegalovirus−Infected Individuals

This information is current as Angela R. Manser, Nadine Scherenschlich, Christine Thöns, of September 26, 2021. Hartmut Hengel, Jörg Timm and Markus Uhrberg J Immunol published online 13 September 2019 http://www.jimmunol.org/content/early/2019/09/12/jimmun ol.1900423 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2019/09/12/jimmunol.190042 Material 3.DCSupplemental http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 26, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2019 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published September 13, 2019, doi:10.4049/jimmunol.1900423 The Journal of Immunology

KIR Polymorphism Modulates the Size of the Adaptive NK Cell Pool in Human Cytomegalovirus–Infected Individuals

Angela R. Manser,* Nadine Scherenschlich,* Christine Tho¨ns,† Hartmut Hengel,‡,x Jo¨rg Timm,† and Markus Uhrberg*

Acute infection with human CMV (HCMV) induces the development of adaptive NKG2C+ NK cells. In some cases, large expansions of this subset, characterized by coexpression of HLA-C–speciﬁc KIR, are stably maintained during the life-long latent phase of infection. The factors that control these unusual expansions in vivo are currently unknown. In this study, the role of KIR polymorphism and expression in this process was analyzed. It is shown that strong NKG2C+ NK cell expansions are dominated by single KIR clones, whereas moderate expansions are frequently polyclonal (p < 0.0001). Importantly, the choice of KIR was not arbitrary but biased toward usage of HLA-C–speciﬁc KIR encoded by the centromeric part of group A (cenA) haplotypes.

Consideration of KIR allelic variation and gene copy number revealed that the cenA effect was predominantly due to the Downloaded from HLA-C2–speciﬁc KIR2DL1 receptor; presence of KIR2DL1 on NKG2C+ NK cells led to signiﬁcantly larger clonal expansions than the cenB-encoded KIR2DL2 (p = 0.002). Expansion of NKG2C+KIR2DL1+ NK cells was always accompanied by the cognate ligand HLA-C2. Moreover, in these donors the frequency of NKG2C+ NK cells correlated with the concentration of anti-HCMV IgG (r = 0.62, p = 0.008), suggesting direct relevance of NKG2C+KIR2DL1+ NK cells for virus control. Altogether, the study suggests that the homeostasis of NKG2C+ NK cells in HCMV infection is at least partly controlled by coexpression of cognate

inhibitory KIR. In particular, the strong interaction of KIR2DL1 and HLA-C2 ligands seems to promote large and stable http://www.jimmunol.org/ expansion of adaptive NK cells in HCMV infection. The Journal of Immunology, 2019, 203: 000–000.

hallmark of NK cells is the expression of HLA class haplotypes represent combinations of conserved centromeric and I–specific inhibitory receptors. They constitute an im- telomeric gene blocks separated by a unique region constituting a A portant detection system for virally infected and ma- hotspot of recombination (4–6). The most common haplotype is lignantly transformed cells, which tend to downregulate HLA referred to as A haplotype and encodes inhibitory KIR for all three class I to escape adaptive immune responses (1) and at the same major HLA class I–encoded epitopes: on the centromeric part time mediate NK cell tolerance toward healthy cells. Two receptor KIR2DL1 and KIR2DL3, which are specific for the HLA-C2 and -C1 families are key to this system: the Ig-like killer cell Ig-like re- epitopes, respectively, and on the telomeric part the HLA-Bw4– by guest on September 26, 2021 ceptor (KIR) genes on human chromosome 19 and the lectin-like specific KIR3DL1. All non-A haplotypes are referred to as group B CD94 and NKG2 genes on chromosome 12. To detect even sub- haplotypes and are on the centromeric side characterized by lack tle changes in expression of single HLA class I or combinations or attenuation of the C2-specific KIR2DL1 receptor. Instead, thereof on aberrant cells, the members of these two receptor group B haplotypes encode the KIR2DL2 receptor that recognizes families are expressed in a clonally distributed fashion, leading to C1 as well as C2, albeit the latter with less avidity than KIR2DL1. the formation of complex NK cell repertoires (2). Importantly, KIR Another major difference between group A and B haplotypes genes are highly polymorphic, and besides allelic polymorphisms pertains to the possession of stimulatory KIR genes; group A hap- also exhibit extensive haplotypic diversity (3). The common KIR lotypes possess only one stimulatory KIR gene, KIR2DS4,which is often a nonproductive allele, whereas group B haplotypes harbor up to five different stimulatory KIR genes (2, 3). *Institute for Transplantation Diagnostics and Cell Therapeutics, Heinrich Heine University, University Hospital of Du¨sseldorf, 40225 Du¨sseldorf, Germany; The CD94/NKG2A heterodimer represents the only inhibitory †Institute of Virology, Heinrich Heine University, University Hospital of Du¨sseldorf, receptor within the CD94/NKG2 family of lectin-like receptors and 40225 Du¨sseldorf, Germany; ‡Institute of Virology, University Medical Center, x is also the evolutionary most conserved HLA class I–specific in- Albert-Ludwigs-University Freiburg, 79104 Freiburg, Germany; and Faculty of Medicine, Albert-Ludwigs-University Freiburg, 79104 Freiburg, Germany hibitory receptor in primates. It interacts with the similarly con- ORCIDs: 0000-0002-3482-816X (H.H.); 0000-0001-7799-3045 (J.T.); 0000-0001- served HLA-E, which is an ubiquitously expressed nonclassical 9553-1987 (M.U.). HLA class I molecule that requires loading with leader peptides, Received for publication April 12, 2019. Accepted for publication August 14, 2019. predominantly from other HLA class I molecules, for expression This work was supported by the Deutsche Forschungsgemeinschaft (Research Grant on the cell surface (7–9). NKG2A plays a key role for tolerance UH91/7 to M.U. and HE2526/9-1 to H.H.) and by Infect-ERA Grant TANKACY induction of NK cells because it is the first class I–specific in- (BMBF 031L0090 to H.H.). hibitory receptor on the cell surface of developing NK cells and Address correspondence and reprint requests to Dr. Markus Uhrberg, Institute for mediates inhibition by all healthy body cells because of its Transplantation Diagnostics and Cell Therapeutics, Heinrich Heine University, Uni- versity Hospital of Du¨sseldorf, Building 14.80, Moorenstrasse 5, 40225 Du¨sseldorf, ubiquitously expressed ligand HLA-E. NKG2A expression thus Germany. E-mail address: [email protected] secures tolerance and enables full functionality of NK cells at a The online version of this article contains supplemental material. stage before the initiation of inhibitory KIR expression. The stim- Abbreviations used in this article: HCMV, human CMV; KIR, killer cell Ig-like ulatory NKG2C receptor, which like NKG2A is specific for HLA-E, receptor; MFI, mean fluorescence intensity; PCR-SSP, PCR with sequence-specific is expressed at a later developmental stage and preferentially found primer. on terminally differentiated NK cells (10). In mature NK cells, all Copyright Ó 2019 by The American Association of Immunologists, Inc. 0022-1767/19/$37.50 possible combinations of inhibitory and stimulatory NKG2 and

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1900423 2 KIR POLYMORPHISM MODULATES ADAPTIVE NK CELL EXPANSIONS

KIR receptors are realized, leading to complex NK cell rep- Materials and Methods ertoires. In a given individual, these repertoires are stably Blood samples maintained over time, whereas they are highly variable be- PBMC were isolated from randomly selected buffy coats (n = 270; kindly tween individuals. provided by the DRK Blutspendedienst Baden-Wu¨rttemberg/Hessen, The exogenous factors that are involved in shaping NK cell Frankfurt a. M.) or whole blood donations from healthy volunteers (n =6) repertoires are only partly understood. In this regard, it became using density gradient centrifugation (Biocoll separating solution; Biochrom). recently clear that infection with human CMV (HCMV) can lead to Prior to PBMC isolation, serum was collected and stored at 220˚C. Cells dramatic changes in the composition of NK cell repertoires (11). were cryopreserved in RPMI 1640 (Lonza) containing 50% FBS (Biochrom) and 10% DMSO (Santa Cruz Biotechnology) and stored in liquid nitrogen. HCMV (human herpesvirus 5) is a member of the b-herpesvirus Samples were excluded from analyses (8/276) if ,1000 NK cells could be family that has coevolved with its host species to establish measured (because of poor sample quality after thawing, 5/276) or if no a commensal relationship that leads to life-long infection in HCMV serology could be determined (because of missing serum samples, most cases without clinical symptoms (12). However, there are 3/276). This study was approved by the ethics committee of the Medical Faculty of Heinrich-Heine-University (no. 3880). conditions that are associated with life-threatening complications (i.e., HCMV reactivation in immune-compromised patients and KIR, HLA, NKG2C, and CD16 genotyping, HCMV serology congenital infection of fetuses as well as infection of preterm DNA was extracted from whole blood or buffy coat using QIAamp DNA infants [13–15]). Thus, the clinical relevance of understanding Blood Mini Kit (Qiagen). Generic KIR genotyping was performed as how NK cells interact and contribute to control HCMV in the previously described (27), including subtyping for KIR2DL1, KIR2DL2, latent phase of infection is high. A large part of the HCMV ge- and KIR2DL3 according to Shilling et al. (28) and KIR3DL1 subtyping nome encodes for gene products that interfere with detection by according to Boudreau et al. (29). KIR2DL1 copy number variation Downloaded from was performed as described previously (30). HLA-E alleles (*0101 or the innate and adaptive immune system (12). The fact that a large *0103) were typed by PCR with sequence-specific primers (PCR- majority of the adult population in most parts of the world is la- SSP) as previously described (31) as well as the HLA-C–encoded tently infected with HCMV is impressive testament to the highly KIR ligands C1 and C2 (32). For typing of the NKG2C deletion successful immune escape strategy of this virus (16). variant, PCR-SSP was performed as previously described (33). The Pioneering work by the group of Lo´pez-Botet (17) had shown FCGR3A polymorphism at position 158V/F was determined using PCR-SSP as previously described (34). HLA class I allele typing was that HCMV infection is associated with an increased frequency of performed by Luminex technology (One D). HCMV IgG serology was http://www.jimmunol.org/ + + NKG2C NK cells. NKG2C NK cells are referred to as adaptive tested using a chemiluminescent immunoassay (CLIA, Liaison CMV NK cells because of several unique features such as longevity and IgG II; DiaSorin). apoptosis resistance, exquisite sensitivity to Ab-mediated activa- Flow cytometrical Abs and staining procedure tion, and unresponsiveness to IL-12 stimulation, which correlate with major transcriptional changes and epigenetic remodeling For phenotypical characterization, the following mAbs (clone names indicated in parentheses) were used: KIR3DL1 (DX9), anti-IgG1– (18–20). The role of KIR expression was first addressed in the allophycocyanin-Vio770 (X-56), NKG2A-Biotin (REA110), anti-Biotin– setting of HCMV reactivation following hematopoietic stem cell VioGreen (Bio3-18E7), KIR2DL1/S1–VioBlue (11PB6), KIR3DL1/2–PE transplantation; the expansion of NKG2C+ NK cells was accom- (5.133), KIR2DS4-PE-Vio770 (JJC11.6), all from Miltenyi Biotec; panied by coexpression of KIR with a strong bias toward the KIR2DL1-FITC or -allophycocyanin (143211), KIR2DL3-FITC (180701), by guest on September 26, 2021 HLA-C–specific KIR2DL2/3 (21). Subsequently, in healthy NKG2C-PE or -allophycocyanin (134591), all from R&D Systems; + KIR2DL2/3/S2–PC5.5 (GL183), NKG2A-PC7 or -allophycocyanin (Z199), all HCMV individuals, clonal expansions of the NKG2C subset from Beckman Coulter; KIR3DL1-FITC, –Alexa Fluor 700, –Brilliant Violet were identified as the major reason for perturbations of the KIR 421, or –PE/Cy7 (DX9), CD56–PE/Dazzle594 (HCD56), CD16–PE/Cy5 repertoire (22). Similar to the observations in acute infection, (3G8), CD57–Pacific Blue (HCD57), CD3–Pacific Blue, –Brilliant Violet these clonal expansions were largely restricted to HLA-C–specific 510, –allophycocyanin/Cy7 or –PE/Cy7 (UCHT1), all from BioLegend. 3 inhibitory KIR, whereas other inhibitory KIR such as the HLA-B– PBMC were thawed and washed two times with 1 PBS containing 0.5% w/v BSA and 2 mM EDTA and resuspended in 100 mlof13 PBS/ specific KIR3DL1 and the HLA-A–specific KIR3DL2 were rarely BSA/EDTA. To enable appropriate analysis of the clonal distribution of + found on NKG2C NK cells (22, 23). Importantly, the size of the KIR, flow cytometric analysis of KIR repertoires was adjusted based on NKG2C+ subset varies vastly in the population; whereas in the each donors’ KIR genotype. majority of the HCMV-infected population only small subsets of For optimal staining results (good separation of KIR2DL1 and KIR2DS1 + as well as KIR2DL2/KIR2DS2 and KIR2DL3) and to avoid cross-reactivity, NKG2C -adaptive NK cells are maintained, in some individuals Abs were stained in a sequential order: first step: KIR3DL1, 10 min 4˚C; the NK cell repertoires are dominated by large clonal expansions second step: anti-IgG1–allophycocyanin–Vio770, 10 min 4˚C; third step: of NKG2C+ NK cells exceeding 50% of the NK cell compartment. NKG2A-Biotin, 10 min 4˚C; fourth step: anti-Biotin–VioGreen, 10 min Remarkably, the expansion of NKG2C NK cells could be recapit- 4˚C; fifth step: KIR-specific Abs for KIR2DL1 and KIR2DL3, 10 min 4˚C; ulated in vitro by cocultures using HCMV-infected cells, allowing to sixth step: all other Abs, 20 min 4˚C. Except for step 5, cells were washed with 13 PBS/BSA/EDTA. demonstrate the influence by several HCMV-encoded immune response genes (24–26). In contrast, the contribution of host-intrinsic Flow cytometric devices, software, statistics, and definition of genetic variation is currently mostly unexplored. Specifically, it is NKG2C-expanded donors unclear why HCMV-specific NK cell expansions are consistently Flow cytometry was performed on a Beckman Coulter CyAn ADP or associated with HLA-C–specific KIR. CytoFLEX, data were analyzed using Kaluza Flow Cytometry Analysis In the current study, we have performed a population survey v1.2 (Beckman Coulter). Samples positive for both KIR2DL2/S2 and to analyze how KIR polymorphism and clonal expression KIR2DL3*005 were excluded from KIR repertoire analyses, as these KIR influences expansion of the NKG2C+-adaptive NK cell subset. cannot be distinguished with currently commercially available Abs as described before (35). For the comparison of groups, a parametric (one- The present data suggest that the selected KIR is a major deter- way ANOVA, t test) or nonparametric (Kruskal–Wallis, Mann–Whitney minant regulating homeostasis of HCMV-induced adaptive NK cells. U,Spearmancorrelation,x2) test was used depending on normal distribution KIR2DL1, the high-avidity receptor for the HLA class I–encoded C2 (Shapiro–Wilk normality test). Significance levels were set as *p , 0.05, , , epitope is shown to be associated with large expansions of NKG2C+ **p 0.01, ***p 0.001 or adjusted for pairwise comparisons by the Bonferroni method (for Fig. 3B, 3C). Statistical analyses were performed NK cells. The correlation of this subset with anti-HCMV plasma using GraphPad Prism 5.03 (GraphPad Software). According to Tukey, concentration raises novel questions concerning the role of adaptive samples exceeding the 75th percentile plus the 1.5-fold interquartile range NKG2C+ NK cells in the control of HCMV infection. of NKG2C expression on HCMV2 NK cells (i.e., .9.25% for NKG2C The Journal of Immunology 3 frequency and .9.85 for NKG2C mean fluorescence intensity [MFI]) were defined as NKG2C-expanded donors. Results Strong expansions of NKG2C+ NK cells are dominated by single KIR clones To evaluate the role of KIR in establishing and maintaining HCMV- specific expansions of NKG2C+ NK cells, a cohort of 262 healthy blood donors (age 18–59 y) was analyzed. Donors were tested for anti-HCMV–specific IgG Abs (HCMV infection rate: 53%) and analyzed for NKG2C expression by flow cytometry (Fig. 1). A strong positive association with HCMV infection was found in terms of frequency and strength of expression as previously described (p , 0.0001, Table I). Nonetheless, the large majority of HCMV-infected donors exhibited NKG2C frequencies similar to noninfected donors; 96% of the donors in the noninfected, but also 63% in the infected cohort exhibited moderate frequencies of NKG2C expression ranging from 0 to 10% of NK cells. A clear separation between the HCMV-seropositive and -negative cohort Downloaded from was achieved only for NKG2C frequencies around 10% of NK cells. To differentiate more objectively between donors with HCMV-mediated expansion of NKG2C+ NK cells and those without clear evidence for shaping by HCMV, a suitable algorithm by Tukey was applied (36). The resulting thresholds for frequency

($9.25%) and mean expression (MFI $ 9.85) of NKG2C led to http://www.jimmunol.org/ the definition of a subgroup of 55 expanded donors within an extended cohort of 268 donors (Fig. 1). Three of the 55 samples with expanded NKG2C+ NK cells were from HCMV2 donors. The negative status of these samples was confirmed by repeated serological testing. Additionally, a lack of T cell reactivity was observed in an HCMV peptide–specific proliferation assay (data not shown). For these three HCMV2 donors, an acute HCMV infection preceding the increase of virus-specific IgG titer could thus not be excluded as well as HCMV-independent viral (all by guest on September 26, 2021 threedonorswereEBV+) or nonviral origins of their NKG2C expansion. Several genetic polymorphisms that have the potential to in- FIGURE 1. Defining HCMV-driven NK cell expansions on the basis of + NKG2C expression. Scatterplot of the distribution (frequency and MFI) of fluence the expansion of NKG2C NK cells were defined in the dim 2 cohort (Table I). Between subgroups with and without expanded NKG2C on CD56 NK cells. Samples were divided in HCMV (upper panel; n = 123) and HCMV+ donors (lower panel; n = 145). HCMV2 NKG2C+ NK cells, the frequencies of HLA-C–encoded ligands samples were used to define thresholds of NKG2C expansion according to C1 and C2, the KIR haplotype groups A and B, and the dimor- Tukey; values that were higher than the 75th percentile + 1.5-fold inter- phism of the NKG2C ligand HLA-E at position 107 (R/G) were all quartile range (IQR) (9.25% for NKG2C frequency, 9.85 for NKG2C-MFI; not significantly different (Table I). Because adaptive NK cells are dashed lines) were defined as expansions (light gray filled circles). Non- highly sensitive to IgG-mediated activation leading to Ab-dependent expanded donors were marked by dark gray filled circles (HCMV2 donors; cellular cytotoxicity (19, 37), the functionally important dimorphism n = 120) or open circles (HCMV+ donors; n = 93). at position 158 (V/F) in CD16 was analyzed, too. Again, no significant difference was detected between subgroups. Finally, the (clone size: .50% of NKG2C subset), whereas donors with in- cohort was typed for a deletion variant of NKG2C (38). By defi- termediate and low NKG2C frequency were associated with nition, homozygous deletion of NKG2C was restricted to the progressively more complex KIR repertoires composed of dif- nonexpanded cohort (n = 7). Irrespective of these homozygous ferent receptor combinations (Fig. 2C). This observation, which samples, there was a significantly decreased frequency of the wassimilarlyobservedindonorswithC1andC2KIRligands del NKG2C variant in expanded cases (Table I). This in turn (data not shown), indicates that large NKG2C expansions are + translated into reduced frequencies of NKG2C NK cells in do- often based on only few “founder” clones, whereas less prom- del nors carrying the NKG2C variant (Supplemental Fig. 1), which inent NKG2C subsets remain polyclonal in their KIR repertoire. is consistent with previous studies (33, 39). Next, the KIR repertoire of NKG2C+ NK cells was analyzed by Cognate interaction between KIR2DL1 and HLA-C2 promotes flow cytometry employing panels of KIR-specific Abs that were expansion of the NKG2C compartment adjusted according to the respective KIR genotypes. As illustrated The above analyses suggest that the continued expansion of in Fig. 2, the combinatorial expression mode of KIR enabled to NKG2C+ NK cells is a highly selective process, recruiting only a define clonal expression patterns within the NKG2C+ NK cell very small part of the KIR repertoire. We next asked if the KIR subset (Fig. 2A, 2B). Notably, increase in NKG2C frequency was that are coexpressed on expanded NKG2C+ NK cells were ran- + strongly correlated with decrease in clonal diversity (rs = 0.6702; domly chosen. KIR repertoire analysis in HCMV donors with p , 0.0001). Donors with high NKG2C frequency (.45% of NKG2C expansion (35.9% of HCMV+ donors as defined in Fig. 1) NK cells) were consistently dominated by single KIR clones revealed coexpression of mainly HLA-C–specific KIR. Only in 4

Table I. Characteristics of the initial cohort and the extended HCMV+ cohort: genetic association of NKG2C dimorphism with presence of NKG2C expansion

2 HCMV HCMV+ HCMV+ Not Expanded HCMV+ Expanded

n =123 % n = 139 % p Value n =93 % n =52 % p Value HLA-C C1/C1 53 43.1 38 27.3 0.0111a 30 32.3 11 21.2 0.2821a C1/C2 60 48.8 78 56.1 0.0111a 47 50.5 33 63.5 0.2821a C2/C2 10 8.1 23 16.5 0.0111a 16 17.2 8 15.4 0.2821a KIR A/A 31 25.2 45 32.4 0.1247a 36 38.7 13 25.0 0.2451a A/B 73 59.3 65 46.8 0.1247a 40 43.0 27 51.9 0.2451a B/B 19 15.4 29 20.9 0.1247a 17 18.3 12 23.1 0.2451a I OYOPIMMDLTSAATV KCL EXPANSIONS CELL NK ADAPTIVE MODULATES POLYMORPHISM KIR HLA-E R/R 38 30.9 44 31.7 0.7313a 29 31.2 18 34.6 0.8120a R/G 61 49.6 73 52.5 0.7313a 48 51.6 27 51.9 0.8120a G/G 24 19.5 22 15.8 0.7313a 16 17.2 7 13.5 0.8120a CD16 polymorphism 158 V/V 13 10.6 21 15.1 0.3207a,b 17 18.3 7 13.5 0.2892a,b V/F 59 48.0 69 49.6 0.3207a,b 41 44.1 30 57.7 0.2892a,b F/F 51 41.5 46 33.1 0.3207a,b 33 35.5 14 26.9 0.2892a,b n/a 0 0.0 3 2.2 2 2.2 1 1.9 NKG2C +/+ 74 60.2 99 71.2 0.032a,b 59 63.4 45 86.5 0.0247c,d +/del 46 37.4 32 23.0 0.032a,b 26 28.0 7 13.5 0.0247c,d del/del 3 2.4 7 5.0 0.032a,b 7 7.5 0 0.0 n/a 0 0.0 1 0.7 1 1.1 0 0.0 % NKG2C on CD56dim (6SD) 4.6 6 8.1 13.4 6 16.6 <0.0001e 4.6 6 3.5 32.1 6 17.7 <0.0001e MFI NKG2C on CD56dim (6SD) 7.0 6 1.6 10.5 6 4.5 <0.0001e 8.2 6 2.6 15.3 6 5.1 <0.0001e Anti-HCMV IgG (U/ml) (6SD) 5.3 6 0.9 90.0 6 40.8 <0.0001e 88.5 6 40.6 93.4 6 39.8 0.3441e % CD56dim (6SD) 13.1 6 8.4% 12.7 6 8.2% 0.61e 12.5 6 8.2% 13.7 6 8.1% 0.2543e The p values ,0.05 are presented in bold type. ap values were calculated using x2 test. bSamples with unavailable typing were excluded from analysis. cp values were calculated using Fisher exact test. dSamples that were typed homozygous for the deletion variant and samples with unavailable typing were excluded from analysis. ep values were calculated using Mann–Whitney U test.

n/a, not available.

Downloaded from from Downloaded http://www.jimmunol.org/ by guest on September 26, 2021 26, September on guest by The Journal of Immunology 5

level as well as copy number variation analysis was performed. As shown in Fig. 3B, NKG2C expansions were significantly affected by the centromeric part of KIR haplotypes, which harbor the HLA-C–specific inhibitory KIR genes (p = 0.012). The mean NKG2C frequencies were 40.7% in donors with two copies of cenA (cenA/A donors) and 21.9% in donors with zero copies (cenB/B donors). Separate analysis of the HLA-C–specific KIR genes revealed that the cenA effect was mainly driven by the C2-specific KIR2DL1 gene (Fig. 3C). NKG2C expansions using KIR2DL1 were significantly stronger than those using the group B–specific KIR2DL2 (size of expansion 42.3% of all NK cells for KIR2DL1-expanded donors versus 22.0% for KIR2DL2-expanded donors, p = 0.0023). Furthermore, strong NKG2C expansions (.45% of all NK cells) were always coexpressing KIR from cenA haplotypes and in most cases recruited KIR2DL1 (Fig. 3D). This effect was independent from HLA-C gene-dose. Moreover, the analysis of outliers in KIR expression revealed that KIR2DL1 expression was highly sensi-

tive to HCMV-mediated perturbations; all but one KIR2DL1 Downloaded from outliers were due to NKG2C expansions (p = 0.0018; Fig. 3E). In contrast, the majority of KIR2DL3 outliers were found in donors lacking NKG2C expansions, and in the case of KIR2DL2 only one outlier was found (Fig. 3E). Notably, all donors with KIR2DL1 outliers also had the respective ligand HLA-C2.

Anti-HCMV plasma concentrations correlate with expansion of http://www.jimmunol.org/ NKG2C+KIR2DL1+ NK cells As shown above, signiﬁcant expansions of NKG2C+ NK cells are detectable in only one third of HCMV+ donors. It is so far unknown if this expanded population of adaptive NK cells is directly involved in the control of virus reactivation during latent infection. To shed more light on this, plasma concentrations of HCMV-speciﬁc IgG Ab were compared. In the present cohort,

Ab concentrations did not show a significant difference between by guest on September 26, 2021 expanded and nonexpanded donors (Table I). However, within the expanded donor group, a nonsignificant positive correlation was noted between virus-specific Ab concentrations and frequency of the NKG2C subset (r = 0.236, p = 0.0923; data not shown). Further stratification on the basis of the coexpressed KIR revealed that this effect was mediated by the KIR2DL1-expressing expansions; a significant correlation of the Ab concentration was seen in donors coexpressing KIR2DL1 but not in those expressing KIR2DL2 or KIR2DL3 (Fig. 4). A strong connection between KIR2DL1 and size of the NKG2C subset is also suggested by linear regression analysis (r =0.62,p =0.008). FIGURE 2. KIR clonal diversity decreases with increasing size of NKG2C subset. (A) Representative flow cytometrical staining and gating Discussion strategy of a NKG2C-expanded donor coexpressing KIR2DL1. After The presence of NKG2C+ NK cells is tightly correlated with dim + initial gating on lymphocytes CD56 NKG2C NK cells were deter- HCMV infection. The acute phase is in most cases characterized mined for further analyses. (B) Repertoire analyses of three represen- + + by a rapid and extensive expansion of NKG2C NK cells. Sub- tative HCMV donors with a large (79.2%; black bars), intermediate sequent phases of infection are usually accompanied by contrac- (40.1%; white bars), or small (17.7%; gray bars) NKG2C expansion tion of the NKG2C+ NK cell subset often back to baseline levels. illustrating how clonal diversity decreases with the size of the NKG2C However, in some donors the NKG2C subset remains prominent subset. (C) A strong correlation (n =106;p , 0.0001, rs = 0.6702) could be observed between the size of the NKG2C subpopulation and the or even increases in frequency up to the point where the naive clonal diversity of KIR (defined as the largest clone within NKG2C+ polyclonal NK cell repertoire is largely replaced by adaptive + CD56dim NK cells). For reliable analyses of KIR expression patterns, NKG2C NK cells. These expanded NKG2C subsets are stable samples with very low NKG2C frequency (,250 NKG2C+ NK cells per over at least several years and are characterized by coexpression sample; n = 38) or undefined KIR clonality (n = 1) were excluded from of an inhibitory KIR for self–HLA class I. The present study the analysis. suggests that the large interindividual variation in the adaptive response to HCMV infection is substantially influenced by the two cases the Bw4-specific KIR3DL1 receptor was dominantly underlying genetic diversity of the KIR family. Large expansions expressed in NKG2C expansions (Fig. 3A). To increase resolution of NKG2C+ NK cells were preferentially associated with cenA of our analysis and enable assignment of clonally expanded KIR KIR haplotypes. The genetic effect was largely due to the cenA- to known haplotype groups, additional KIR typing on the allelic located KIR2DL1 that is specific for the C2 epitope of HLA-C. 6 KIR POLYMORPHISM MODULATES ADAPTIVE NK CELL EXPANSIONS Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 3. Large NKG2C expansions preferentially express cenA-encoded KIR2DL1 alleles. (A) Distribution of KIR on expanded NKG2C+ NK cells in HCMV+ donors (KIR2DL1: n = 17; KIR2DL2: n = 10; KIR2DL3: n = 16; KIR3DL1: n = 2; KIR-Mix: n = 7 [at least two equally expanded KIR]). (B) Tukey box plots illustrating the frequency of NKG2C+ NK cells in expanded donors divided into either three (left columns, p = 0.0117, Kruskal–Wallis test) or two (right columns, p = 0.0059, Mann–Whitney U test) groups based on the presence of centromeric group A and B KIR motifs (B/x = A/B and B/B donors). (C) Frequency of NKG2C+ NK cells in expanded donors stratified by expression of HLA-C–specific inhibitory KIR into either three (left columns, p = 0.0052, one-way ANOVA; KIR2DL1 versus KIR2DL2 p = 0.0023, unpaired t test) or two (right columns, p = 0.0103, Mann–Whitney U test) groups. Donors were color-coded based on their HLA-C genotype (white symbols: C1/C1, gray symbols: C1/C2, black symbols: C2/C2). (D) Frequency and MFI of NKG2C expression in expanded donors according to coexpression of KIR. An area of high-frequency expansions (.45% NKG2C) expressing exclusively cenA-encoded KIR is indicated. (E) Frequencies of KIR2DL1 (n = 247 donors), KIR2DL2 (n = 138 donors), and KIR2DL3 (n = 224 donors) in the cohort (n = 268). Donors were subdivided in no HCMV-driven NKG2C expansion (white circles) and HCMV-driven NKG2C expansions (black circles). According to Tukey (36), upper fences defining outliers are marked by a horizontal dashed line, respectively. Significance was calculated by x2 test (x2 test, p = 0.0179). For (B and C) significance level was corrected for pairwise comparisons (*p , 0.0167, **p , 0.0033).

Compared with the cenB-located KIR2DL2 that interacts with C1 following allogeneic hematopoietic stem cell transplantation and C2 ligands, NKG2C expansions coexpressing KIR2DL1 were reported a strong bias toward expression of KIR2DL2/3 on ex- significantly larger (mean: 42.3% for KIR2DL1 versus 22.0% for panded NKG2C+ NK cells up to 1 y posttransplantation (21, 40). KIR2DL2). In general, HCMV infection had a much larger impact The prevalence of KIR2DL2/3 in these studies probably mirrors on KIR2DL1 expression than on KIR2DL2 as illustrated by the the sequential expression mode of KIR during NK cell develop- fact that the expansion of NKG2C+ NK cells created a multitude ment from hematopoietic progenitors that is characterized by early of high-frequency outliers for KIR2DL1, whereas KIR2DL2 ex- expression of KIR2DL2/3 and late expression of KIR2DL1 (41, 42). pression was only marginally affected. KIR2DL3 exhibited in- Thus, when HCMV reactivation occurs during NK cell reconsti- termediate characteristics with some outliers attributed to NKG2C tution, NKG2C+ NK cells are rapidly generated from NK cell expansions and others that were independent of HCMV. progenitors, leading to the reported bias toward KIR2DL2/3 ex- The present finding that most of the large NKG2C expansions in pression as first receptor. However, it is unclear if during sub- latently infected people express the KIR2DL1 receptor was rather clinical infection, HCMV infection in C2+ donors similarly starts unanticipated. Previous studies analyzing reactivation of HCMV with a wave of hematopoietic progenitors going through this The Journal of Immunology 7

that binding of KIR2DL1 to C2 ligands has a higher avidity compared with interactions between KIR2DL2 or KIR2DL3 with C1 alleles (43, 44). Notably, in the current study only high-avidity, cenA-encoded KIR2DL1 alleles were found on strongly expanded NKG2C+ NK cell clones, whereas the lower avidity, cenB- encoded KIR2DL1*004 allele was only found in two cases, as predominant cognate KIR and, interestingly, in these cases the NKG2C expression levels were relatively low compared with the other expanded cases (mean MFI: 10.1). Nonetheless, no signiﬁcant differences in KIR2DL1 allele usage could be detected. Mecha- nistically, the strong binding of KIR2DL1 to C2 could provide stronger survival signals compared with the other receptor/ligand combinations. It was previously shown that cross-linking of KIR provides crucial survival signals to NK and T cells and that cognate KIR/KIR ligand interaction on licensed NK cells leads to anti-apoptotic signaling (45–47). The presence of KIR2DL1 on NKG2C+ NK cells could thus lead to preferential survival via stronger cognate interaction with C2 resulting in the replacement

of the KIR2DL3 receptor. Obviously, the suggested mechanism Downloaded from requires presence of HLA-C ligands on the target cells. In this regard, it is well described that HLA class I is generally downregulated on HCMV-infected cells (48). However, whereas this is clearly shown for HLA-A and -B, this is less clear for HLA-C. It was previously shown that the complex mechanisms of HCMV-

mediated HLA-I regulation differentially affect HLA-A and -B http://www.jimmunol.org/ versus HLA-C expression and that at least certain HLA-C allotypes are spared from downregulation (49–51). The situation is reminiscent of HIV-mediated HLA-I regulation, which affects HLA-A and -B much more than HLA-C, which is downregulated to a certain degree but still able to mediate NK cell inhibition (52). Given the exquisite specificity and binding strength of KIR2DL1 for HLA-C2 allotypes, we hypothesize that the levels of HLA-C expression on the infected target cell are sufficient to enable interaction with KIR and that the residual HLA-C expression might by guest on September 26, 2021 even favor triggering of KIR2DL1+NKG2C+ NK cells compared with the less strong HLA-C binders KIR2DL2 and KIR2DL3. Differential regulation of HLA class I genes could also explain the near exclusion of non–HLA-C–specific KIR such as KIR3DL1 and KIR3DL2 from expanded NKG2C+ NK cells because of low expression of the corresponding HLA-Bw4– and HLA-A–carrying allotypes, which might still enable proper inhibition but NS expansion of the respective NK cells. Perhaps the most surprising finding in this study was that the expansion of KIR2DL1+NKG2C+ NK cells was significantly correlated with the concentration of HCMV-specific IgG Ab. The correlation was restricted to adaptive NK cells expressing KIR2DL1 and not seen for NKG2C+ NK cells expressing KIR2DL2 or KIR2DL3, although the latter exhibited a tendency into the same direction as KIR2DL1. On the one hand, the observed correlation could be due to an improved virus-specific response along the T cell/B cell axis mediated by the KIR2DL1+NKG2C+ subset. FIGURE 4. Expansion of KIR2DL1+-adaptive NK cells correlates with How adaptive NK cells would support this kind of heightened anti-HCMV plasma concentration. HCMV-specific IgG Ab concentration serological response is, however, unclear; although NK cells are was plotted against NKG2C frequency for donors coexpressing either known to interact with dendritic cells in secondary lymph nodes KIR2DL1 (n = 17; upper panel), KIR2DL2 (n = 13; lower left), or and to support Th1-mediated antiviral responses, Th2-triggered KIR2DL3 (n = 20; lower right). A significant correlation by linear re- B cell responses are usually downregulated by NK cell–mediated gression analysis was found for KIR2DL1 (p = 0.0083, r = 0.618) but not IFN-g production (53, 54). In future studies, it would be interesting KIR2DL2 (p = 0.8744, r = 0.049) or KIR2DL3 (p = 0.0881, r = 0.391). to analyze the quality and quantity of HCMV-specific T cell responses in relation to the size of the NKG2C+ NK cell sequential acquisition process eventually leading to expression of compartment. On the other hand, the present observation of KIR2DL1+. heightened anti-HCMV–specific IgG in the presence of clonal The preference for KIR2DL1 in NKG2C-expanded donors might KIR2DL1+NKG2C+ NK cells could also reflect a less stringent be related to differences in binding strength between the three control of HCMV in the latent phase of infection, for example inhibitory HLA-C receptors KIR2DL1-3. It was previously shown on the level of local, subclinical reactivation. In this context, the 8 KIR POLYMORPHISM MODULATES ADAPTIVE NK CELL EXPANSIONS general absence of stimulatory KIR such as KIR2DS2 and genetic phenomenon is the increased incidence of unsuccessful KIR2DS3 on the cenA haplotype could also contribute to this reproduction in mothers lacking cenB haplotypes and carrying effect. Actually, the role of NKG2C+ NK cells in resistance or fetuses with C2 alleles from the father (60). Obviously, repro- control of HCMV infection is currently unclear. So far, evidence ductive success has a direct impact on selection of protective gene concerning NK cell–mediated control of CMV infection by variants. In contrast, HCMV is the leading cause of congenital adaptive NK cells largely stems from experiments in mice, birth defects (61). Vertical transmission of HCMVoccurs in 0.5–2% showing that Ly49H-expressing NK cells recognizing the murine of pregnancies, leading to severe complications in ∼10% of cases. CMV glycoprotein m157 confer resistance to murine CMV in- It is so far unknown if congenital infection with HCMV is fection in a memory-like manner (55). A similar protective effect associated with a combination of cenA KIR haplotypes and C2 by NKG2C+ NK cells via direct recognition of HCMV products or ligands in the fetus or with NKG2C expansions carrying indirect recognition of HCMV peptides presented by HLA-E was KIR2DL1, and this merits further investigation. In any case, so far not described. Interestingly, in a recent population study in HCMV constitutes an ubiquitous potential thread in pregnancy The Gambia, which is characterized by a near 100% HCMV in- as well as various states of immunosuppression and has thus the fection rate within the first 3 y of life, a deletion variant of NKG2C potential to act as a selective force influencing the frequencies (NKG2Cdel) was found at high frequency; 10.5% of cases had a and the epistatic interaction of KIR and HLA class I alleles. In homozygous deletion of NKG2C, and close to 50% were deletion summary, as shown in the current study, cenAKIRhaplotypes allele carriers (33) compared to only 3.8% homozygous carriers in encoding KIR2DL1 alleles strongly binding C2 ligands pro- the present cohort with a much lower incidence of HCMV in- mote large expansions of NKG2C+ NK cells, suggesting an fection. For a resistance locus, the correlation would have been important role for KIR polymorphism in the regulation of Downloaded from expected to be opposite (i.e., high frequency of productive adaptive NK cell responses. NKG2C alleles in regions with high virus incidence). Notably, heterozygosity for NKG2Cdel is significantly associated with an Acknowledgments overall lower frequency of NKG2C+ cells and importantly also a We thank all healthy volunteers for donating blood and Lutz Walter (German lower frequency of cases with NKG2C expansions as shown in Primate Center, Go¨ttingen, Germany) for input and critical discussions.

this study (Table I) and elsewhere (33, 39), illustrating that the http://www.jimmunol.org/ NKG2Cdel variant efficiently regulates the size of the NKG2C+- Disclosures adaptive NK cell compartment in the population. Finally, the The authors have no financial conflicts of interest. analysis of individuals completely lacking NKG2C did not show any evidence for increased susceptibility to infection or lack of virus control (56). Altogether, it remains unclear what constitutes References del 1. Lanier, L. L. 2005. NK cell recognition. Annu. Rev. Immunol. 23: 225–274. the driving force for selection of the NKG2C variant and more 2. Manser, A. R., S. Weinhold, and M. Uhrberg. 2015. Human KIR repertoires: + generally if NKG2C NK cells play a protective role in HCMV shaped by genetic diversity and evolution. Immunol. Rev. 267: 178–196. infection. 3. Uhrberg, M., N. M. Valiante, B. P. Shum, H. G. Shilling, K. Lienert-Weidenbach, B. Corliss, D. Tyan, L. L. Lanier, and P. Parham. 1997. Human diversity in killer by guest on September 26, 2021 A possible interpretation of the present data is that clonal cell inhibitory receptor genes. Immunity 7: 753–763. NKG2C subsets that expand beyond a certain size, in particular 4. Yawata, M., N. Yawata, L. Abi-Rached, and P. Parham. 2002. Variation within those expressing high-avidity KIR2DL1 alleles, impair the overall the human killer cell immunoglobulin-like receptor (KIR) gene family. Crit. Rev. Immunol. 22: 463–482. control of HCMV. From an immunological standpoint, oligoclonal 5. Hsu, K. C., X. R. Liu, A. Selvakumar, E. Mickelson, R. J. O’Reilly, and repertoires are generally hallmarks of an aging immune system and B. Dupont. 2002. Killer Ig-like receptor haplotype analysis by gene content: evidence for genomic diversity with a minimum of six basic framework haplo- are associated with impaired response to vaccination and newly types, each with multiple subsets. J. Immunol. 169: 5118–5129. encountered pathogenic challenges (57). NKG2C expansions 6. Uhrberg, M. 2005. The KIR gene family: life in the fast lane of evolution. Eur. J. comprising 40–80% of the whole NK cell repertoire, as described Immunol. 35: 10–15. 7.Braud,V.M.,D.S.Allan,C.A.O’Callaghan,K.So¨derstro¨m, A. D’Andrea, in this study, are extreme cases of a continuum of HCMV-induced G. S. Ogg, S. Lazetic, N. T. Young, J. I. Bell, J. H. Phillips, et al. 1998. HLA-E responses that are heavily biased toward a single specificity. The binds to natural killer cell receptors CD94/NKG2A, B and C. Nature 391: highly clonal NK cell compartment might be less fit to counteract 795–799. 8. Borrego, F., M. Ulbrecht, E. H. Weiss, J. E. Coligan, and A. G. Brooks. 1998. HCMV-induced changes occurring in infected host cells over Recognition of human histocompatibility leukocyte antigen (HLA)-E complexed time, for example regarding HLA class I regulation. Notably, in a with HLA class I signal sequence-derived peptides by CD94/NKG2 confers + protection from natural killer cell-mediated lysis. J. Exp. Med. 187: 813–818. previous case report, a strong clonal expansion of KIR2DL1 NK 9. Lee, N., M. Llano, M. Carretero, A. Ishitani, F. Navarro, M. Lo´pez-Botet, and cells was associated with recurrent episodes of HCMV infection D. E. Geraghty. 1998. HLA-E is a major ligand for the natural killer inhibitory (58). In contrast to such strong clonal expansions, we show that receptor CD94/NKG2A. Proc. Natl. Acad. Sci. USA 95: 5199–5204. 10. Bjo¨rkstro¨m, N. K., P. Riese, F. Heuts, S. Andersson, C. Fauriat, M. A. Ivarsson, the large majority of infected individuals contain much smaller A. T. Bjo¨rklund, M. Flodstro¨m-Tullberg, J. Michae¨lsson, M. E. Rottenberg, et al. subsets of NKG2C+-adaptive NK cells that are characterized by a 2010. Expression patterns of NKG2A, KIR, and CD57 define a process of more polyclonal KIR repertoire and might thus be more robust CD56dim NK-cell differentiation uncoupled from NK-cell education. Blood 116: 3853–3864. toward newly developing viral immune escape variants. The 11. Lo´pez-Botet, M., A. Muntasell, and C. Vilches. 2014. The CD94/NKG2C+ NK- dominance of clonal NKG2C+ subsets present in certain donors cell subset on the edge of innate and adaptive immunity to human cytomegalovirus infection. Semin. Immunol. 26: 145–151. might also impair other components of the NK cell–mediated 12. Picarda, G., and C. A. Benedict. 2018. Cytomegalovirus: shape-shifting the response against HCMV. In this regard, NK cell responses to immune system. J. Immunol. 200: 3881–3889. HCMV infection are not restricted to NKG2C+ NK cells but are 13. Hamprecht, K., J. Maschmann, M. Vochem, K. Dietz, C. P. Speer, and G. Jahn. 2 2001. Epidemiology of transmission of cytomegalovirus from mother to preterm also found in the NKG2C compartment, which is obviously infant by breastfeeding. Lancet 357: 513–518. strongly reduced in the expanded donors (22, 56). It is also cur- 14. Pass, R. F., K. B. Fowler, S. B. Boppana, W. J. Britt, and S. Stagno. 2006. rently unclear if NKG2C+ NK cells have a regulatory impact on Congenital cytomegalovirus infection following first trimester maternal infection: symptoms at birth and outcome. J. Clin. Virol. 35: 216–220. HCMV-specific T cells controlling virus-specific Ab production. 15. Fishman, J. A. 2007. Infection in solid-organ transplant recipients. N. Engl. J. Finally, it is also worth mentioning that there is a worldwide Med. 357: 2601–2614. 16. Manicklal, S., V. C. Emery, T. Lazzarotto, S. B. Boppana, and R. K. Gupta. 2013. inverse correlation between cenA KIR haplotypes and HLA-C The “silent” global burden of congenital cytomegalovirus. Clin. Microbiol. Rev. alleles carrying the C2 epitope (59). One possible driver for this 26: 86–102. The Journal of Immunology 9

17. Guma´, M., A. Angulo, C. Vilches, N. Go´mez-Lozano, N. Malats, and M. Lo´pez- influences CD94/NKG2C receptor function and the NK-cell compartment Botet. 2004. Imprint of human cytomegalovirus infection on the NK cell re- redistribution in response to human cytomegalovirus. Eur. J. Immunol. 43: ceptor repertoire. Blood 104: 3664–3671. 3268–3278. 18. Schlums, H., F. Cichocki, B. Tesi, J. Theorell, V. Beziat, T. D. Holmes, H. Han, 40. Davis, Z. B., S. A. Cooley, F. Cichocki, M. Felices, R. Wangen, X. Luo, S. C. Chiang, B. Foley, K. Mattsson, et al. 2015. Cytomegalovirus infection T. E. DeFor, Y. T. Bryceson, D. J. Diamond, C. Brunstein, et al. 2015. Adaptive drives adaptive epigenetic diversification of NK cells with altered signaling and natural killer cell and killer cell immunoglobulin-like receptor-expressing T cell effector function. Immunity 42: 443–456. responses are induced by cytomegalovirus and are associated with protection 19. Lee,J.,T.Zhang,I.Hwang,A.Kim,L.Nitschke,M.Kim,J.M.Scott,Y.Kamimura, against cytomegalovirus reactivation after allogeneic donor hematopoietic cell L. L. Lanier, and S. Kim. 2015. Epigenetic modification and antibody-dependent transplantation. Biol. Blood Marrow Transplant. 21: 1653–1662. expansion of memory-like NK cells in human cytomegalovirus-infected individuals. 41. Fischer, J. C., H. Ottinger, S. Ferencik, M. Sribar, M. Punzel, D. W. Beelen, Immunity 42: 431–442. M. A. Schwan, H. Grosse-Wilde, P. Wernet, and M. Uhrberg. 2007. Relevance of 20. Luetke-Eversloh, M., Q. Hammer, P. Durek, K. Nordstro¨m, G. Gasparoni, C1 and C2 epitopes for hemopoietic stem cell transplantation: role for sequential M. Pink, A. Hamann, J. Walter, H. D. Chang, J. Dong, and C. Romagnani. 2014. acquisition of HLA-C-specific inhibitory killer Ig-like receptor. J. Immunol. 178: Human cytomegalovirus drives epigenetic imprinting of the IFNG locus in 3918–3923. NKG2Chi natural killer cells. PLoS Pathog. 10: e1004441. 42. Zhao, X., S. Weinhold, J. Brands, M. Hejazi, O¨ . Degistirici, G. Ko¨gler, 21. Foley, B., S. Cooley, M. R. Verneris, M. Pitt, J. Curtsinger, X. Luo, S. Lopez- R. Meisel, and M. Uhrberg. 2018. NK cell development in a human stem cell Verge`s, L. L. Lanier, D. Weisdorf, and J. S. Miller. 2012. Cytomegalovirus niche: KIR expression occurs independently of the presence of HLA class I reactivation after allogeneic transplantation promotes a lasting increase in edu- ligands. Blood Adv. 2: 2452–2461. cated NKG2C+ natural killer cells with potent function. Blood 119: 2665–2674. 43. Hilton, H. G., L. A. Guethlein, A. Goyos, N. Nemat-Gorgani, D. A. Bushnell, 22. Be´ziat, V., L. L. Liu, J. A. Malmberg, M. A. Ivarsson, E. Sohlberg, P. J. Norman, and P. Parham. 2015. Polymorphic HLA-C receptors balance the A. T. Bjo¨rklund, C. Retie`re, E. Sverremark-Ekstro¨m, J. Traherne, P. Ljungman, functional characteristics of KIR haplotypes. J. Immunol. 195: 3160–3170. et al. 2013. NK cell responses to cytomegalovirus infection lead to stable im- 44. Winter, C. C., J. E. Gumperz, P. Parham, E. O. Long, and N. Wagtmann. 1998. prints in the human KIR repertoire and involve activating KIRs. Blood 121: Direct binding and functional transfer of NK cell inhibitory receptors reveal 2678–2688. novel patterns of HLA-C allotype recognition. J. Immunol. 161: 571–577. 23. Lopez-Verge`s, S., J. M. Milush, B. S. Schwartz, M. J. Pando, J. Jarjoura, 45. Marti, F., C. W. Xu, A. Selvakumar, R. Brent, B. Dupont, and P. D. King. V. A. York, J. P. Houchins, S. Miller, S. M. Kang, P. J. Norris, et al. 2011. Ex- 1998. LCK-phosphorylated human killer cell-inhibitory receptors recruit Downloaded from pansion of a unique CD57+NKG2Chi natural killer cell subset during acute human and activate phosphatidylinositol 3-kinase. Proc. Natl. Acad. Sci. USA 95: cytomegalovirus infection. Proc. Natl. Acad. Sci. USA 108: 14725–14732. 11810–11815. 24. Hammer, Q., T. Ru¨ckert, E. M. Borst, J. Dunst, A. Haubner, P. Durek, 46. Young, N. T., and M. Uhrberg. 2002. KIR expression shapes cytotoxic reper- F. Heinrich, G. Gasparoni, M. Babic, A. Tomic, et al. 2018. Peptide-specific toires: a developmental program of survival. Trends Immunol. 23: 71–75. recognition of human cytomegalovirus strains controls adaptive natural killer 47.Felices,M.,T.R.Lenvik,D.E.Ankarlo,B.Foley,J.Curtsinger,X.Luo, cells. Nat. Immunol. 19: 453–463. B. R. Blazar, S. K. Anderson, and J. S. Miller. 2014. Functional NK cell 25. Guma´, M., M. Budt, A. Saéz, T. Brckalo, H. Hengel, A. Angulo, and M. Lo´pez- repertoires are maintained through IL-2Ra and Fas ligand. J. Immunol. 192:

Botet. 2006. Expansion of CD94/NKG2C+ NK cells in response to human 3889–3897. http://www.jimmunol.org/ cytomegalovirus-infected ﬁbroblasts. Blood 107: 3624–3631. 48. Halenius, A., C. Gerke, and H. Hengel. 2015. Classical and non-classical MHC I 26. Ro¨lle, A., J. Pollmann, E. M. Ewen, V. T. Le, A. Halenius, H. Hengel, and molecule manipulation by human cytomegalovirus: so many targets—but how A. Cerwenka. 2014. IL-12-producing monocytes and HLA-E control HCMV- many arrows in the quiver? Cell. Mol. Immunol. 12: 139–153. driven NKG2C+ NK cell expansion. J. Clin. Invest. 124: 5305–5316. 49. Ameres, S., J. Mautner, F. Schlott, M. Neuenhahn, D. H. Busch, B. Plachter, and 27. Uhrberg, M., P. Parham, and P. Wernet. 2002. Deﬁnition of gene content for nine A. Moosmann. 2013. Presentation of an immunodominant immediate-early common group B haplotypes of the Caucasoid population: KIR haplotypes CD8+ T cell epitope resists human cytomegalovirus immunoevasion. PLoS contain between seven and eleven KIR genes. Immunogenetics 54: 221–229. Pathog. 9: e1003383. 28. Shilling, H. G., L. A. Guethlein, N. W. Cheng, C. M. Gardiner, R. Rodriguez, 50. Furman, M. H., H. L. Ploegh, and D. J. Schust. 2000. Can viruses help us to D. Tyan, and P. Parham. 2002. Allelic polymorphism synergizes with variable gene understand and classify the MHC class I molecules at the maternal-fetal inter- content to individualize human KIR genotype. J. Immunol. 168: 2307–2315. face? Hum. Immunol. 61: 1169–1176. 29. Boudreau, J. E., J. B. Le Luduec, and K. C. Hsu. 2014. Development of a novel 51. Schust, D. J., D. Tortorella, J. Seebach, C. Phan, and H. L. Ploegh. 1998. Tro-

multiplex PCR assay to detect functional subtypes of KIR3DL1 alleles. PLoS phoblast class I major histocompatibility complex (MHC) products are resistant by guest on September 26, 2021 One 9: e99543. to rapid degradation imposed by the human cytomegalovirus (HCMV) gene 30. Babor, F., A. R. Manser, J. C. Fischer, N. Scherenschlich, J. Enczmann, products US2 and US11. J. Exp. Med. 188: 497–503. O. Chazara, A. Moffett, A. Borkhardt, R. Meisel, and M. Uhrberg. 2014. KIR 52. Ko¨rner, C., C. R. Simoneau, P. Schommers, M. Granoff, M. Ziegler, ligand C2 is associated with increased susceptibility to childhood ALL and A. Ho¨lzemer, S. Lunemann, J. Chukwukelu, B. Corleis, V. Naranbhai, et al. confers an elevated risk for late relapse. Blood 124: 2248–2251. 2017. HIV-1-mediated downmodulation of HLA-C impacts target cell recogni- 31. Grimsley, C., A. Kawasaki, C. Gassner, N. Sageshima, Y. Nose, K. Hatake, tion and antiviral activity of NK cells. Cell Host Microbe 22: 111–119.e4. D. E. Geraghty, and A. Ishitani. 2002. Definitive high resolution typing of HLA- 53. Martıń-Fontecha, A., L. L. Thomsen, S. Brett, C. Gerard, M. Lipp, A. Lanzavecchia, E allelic polymorphisms: identifying potential errors in existing allele data. and F. Sallusto. 2004. Induced recruitment of NK cells to lymph nodes provides Tissue Antigens 60: 206–212. IFN-gamma for T(H)1 priming. Nat. Immunol. 5: 1260–1265. 32. Frohn, C., P. Schlenke, B. Ebel, C. Dannenberg, G. Bein, and H. Kirchner. 1998. 54. Morandi, B., G. Bougras, W. A. Muller, G. Ferlazzo, and C. Mu¨nz. 2006. NK DNA typing for natural killer cell inhibiting HLA-Cw groups NK1 and NK2 cells of human secondary lymphoid tissues enhance T cell polarization via by PCR-SSP. J. Immunol. Methods 218: 155–160. IFN-gamma secretion. Eur. J. Immunol. 36: 2394–2400. 33. Goodier, M. R., M. J. White, A. Darboe, C. M. Nielsen, A. Goncalves, 55. Sun, J. C., S. Lopez-Verges, C. C. Kim, J. L. DeRisi, and L. L. Lanier. 2011. NK C. Bottomley, S. E. Moore, and E. M. Riley. 2014. Rapid NK cell differentiation cells and immune “memory”. J. Immunol. 186: 1891–1897. in a population with near-universal human cytomegalovirus infection is attenu- 56. Liu, L. L., J. Landskron, E. H. Ask, M. Enqvist, E. Sohlberg, J. A. Traherne, ated by NKG2C deletions. Blood 124: 2213–2222. Q. Hammer, J. P. Goodridge, S. Larsson, J. Jayaraman, et al. 2016. Critical role 34. Liu, B., L. A. Maier, N. Hamzeh, K. MacPhail, M. M. Mroz, H. Liu, and L. Li. of CD2 co-stimulation in adaptive natural killer cell responses revealed in 2019. Polymorphism of FCGR3A gene in chronic beryllium disease. Genes NKG2C-deficient humans. Cell Rep. 15: 1088–1099. Immun. 20: 493–499. 57. Simon, A. K., G. A. Hollander, and A. McMichael. 2015. Evolution of 35. Be´ziat, V., J. Traherne, J. A. Malmberg, M. A. Ivarsson, N. K. Bjo¨rkstro¨m, the immune system in humans from infancy to old age. Proc. Biol. Sci. 282: C. Retie`re, H. G. Ljunggren, J. Michae¨lsson, J. Trowsdale, and K. J. Malmberg. 20143085. 2014. Tracing dynamic expansion of human NK-cell subsets by high-resolution 58. Gazit, R., B. Z. Garty, Y. Monselise, V. Hoffer, Y. Finkelstein, G. Markel, analysis of KIR repertoires and cellular differentiation. Eur. J. Immunol. 44: G. Katz, J. Hanna, H. Achdout, R. Gruda, et al. 2004. Expression of KIR2DL1 2192–2196. on the entire NK cell population: a possible novel immunodeficiency syndrome. 36. Tukey, J. W. 1977. Exploratory Data Analysis. Pearson, London. Blood 103: 1965–1966. 37. Zhang, T., J. M. Scott, I. Hwang, and S. Kim. 2013. Cutting edge: antibody- 59. Parham, P., P. J. Norman, L. Abi-Rached, and L. A. Guethlein. 2012. Human- dependent memory-like NK cells distinguished by FcRg deficiency. J. Immunol. specific evolution of killer cell immunoglobulin-like receptor recognition of 190: 1402–1406. major histocompatibility complex class I molecules. Philos. Trans. R. Soc. Lond. 38. Hikami, K., N. Tsuchiya, T. Yabe, and K. Tokunaga. 2003. Variations of human B Biol. Sci. 367: 800–811. killer cell lectin-like receptors: common occurrence of NKG2-C deletion in the 60. Moffett, A., and F. Colucci. 2015. Co-evolution of NK receptors and HLA li- general population. Genes Immun. 4: 160–167. gands in humans is driven by reproduction. Immunol. Rev. 267: 283–297. 39. Muntasell, A., M. Lo´pez-Montan˜e´s, A. Vera, G. Heredia, N. Romo, J. Penãfiel, 61. Pereira, L., E. Maidji, S. McDonagh, and T. Tabata. 2005. Insights into viral M. Moraru, J. Vila, C. Vilches, and M. Lo´pez-Botet. 2013. NKG2C zygosity transmission at the uterine-placental interface. Trends Microbiol. 13: 164–174. Supplemental Figure S1

100 *** * ***

60 NK cells

+ 40

% NKG2C % 20

0 +/+ +/del +/+ +/del +/+ +/del HCMV- HCMV+ total cohort

Supplemental Figure S1: Increased frequency of NKG2C+ NK cells in donors homozygous for the functional NKG2C allele. NKG2C frequency on NK cells in donors with two functional NKG2C alleles (+/+; n = 74 for HCMV- donors and n = 104 for HCMV+ donors) compared to heterozygous donors (+/del; n = 46 for HCMV- donors and n = 33 for HCMV+ donors) and the total cohort (+/+ donors: n = 178 and +/del donors: n = 79). Significant differences were found in all three groups by Mann-Whitney U test: HCMV- p = 0.0002, HCMV+ p = 0.0135, and total cohort p < 0.0001.