NKG2C zygosity influences CD94/NKG2C receptor function and the NK-cell compartment redistribution in response to human cytomegalovirus

Authors


Abstract

Human cytomegalovirus (HCMV) infection promotes a persistent expansion of a functionally competent NK-cell subset expressing the activating CD94/NKG2C receptor. Factors underlying the wide variability of this effect observed in HCMV-seropositive healthy individuals and exacerbated in immunocompromized patients are uncertain. A deletion of the NKG2C gene has been reported, and an apparent relation of NKG2C genotype with circulating NKG2C+ NK-cell numbers was observed in HCMV+ children. We have assessed the influence of NKG2C gene dose on the NK-cell repertoire in a cohort of young healthy adults (N = 130, median age 19 years). Our results revealed a relation of NKG2C copy number with surface receptor levels and with NKG2C+ NK-cell numbers in HCMV+ subjects, independently of HLA-E dimorphism. Functional studies showed quantitative differences in signaling (i.e. iCa2+ influx), degranulation, and IL-15-dependent proliferation, in response to NKG2C engagement, between NK cells from NKG2C+/+ and hemizygous subjects. These observations provide a mechanistic interpretation on the way the NKG2C genotype influences steady-state NKG2C+ NK-cell numbers, further supporting an active involvement of the receptor in the HCMV-induced reconfiguration of the NK-cell compartment. The putative implications of NKG2C zygosity over viral control and other clinical variables deserve attention.

Introduction

Human cytomegalovirus (HCMV) is a herpes virus that infects with high prevalence all human populations. Upon primary infection, HCMV persists in latency eventually undergoing subclinical reactivations. HCMV infection constitutes a common complication in immunocompromised individuals, and is the leading infectious cause of congenital neurological disorders [1, 2].

We originally reported that HCMV infection in healthy individuals promotes a persistent increase of NK cells displaying the CD94/NKG2C activating receptor [3]. The expansion of NKG2C+ NK cells can be detected in HCMV-seropositive children and in patients with congenital infection [4, 5]. HCMV-induced NKG2C+ NK cells lack NKG2A, display low levels of activating NKp30 and NKp46, express inhibitory leukocyte immunoglobulin-like receptor subfamily B, member 1 (LILRB1) and killer-cell immunoglobulin-like receptors (KIRs) specific for self HLA class I molecules, and are functionally competent [3, 6, 7]. The observation that NKG2C+ lymphocytosis during an acute HCMV infection in an SCID patient coincided with a reduction of viremia suggests that this NK-cell subset may, at least partially, contain the viral infection in the absence of T cells [8]. Similar kinetics of NKG2C+ NK-cell expansion have been recently reported upon HCMV reactivation in solid organ and hematopoietic cell transplant recipients [7, 9-11], supporting their putative contribution in antiviral defense as a compensatory mechanism in scenarios of T-cell suppression.

The response of the NKG2C+ NK-cell population to HCMV is reminiscent of that shown by murine Ly49H+ NK cells, which expand in response to murine cytomegalovirus (MCMV) infection [12-14] and persist in the circulation conferring a more efficient defense against reinfection [15]. The term “memory NK cell” has been coined to define this pattern of response [16], and it has been speculated that NKG2Cbright NK cells might play a similar role in humans [17].

Remarkably, HCMV-induced NKG2C+ NK cells have also been shown to expand in patients coinfected with other viruses such as hantavirus, chikungunya virus, hepatitis B virus (HBV), HCV, or HIV-1 [18-21], and to display enhanced function against antibody-targeted infected cells [22]. Whether HCMV-induced NKG2C+ NK cells play any role in the response against these pathogens remains unanswered.

CD94/NKG2C is a heterodimeric C-type lectin-like receptor whose expression and signaling are dependent on the ITAM-containing DAP12 adaptor [23-25]. CD94/NKG2C recognition of the HLA-E class Ib molecule triggers NK-cell mediated cytokine production, cytotoxicity, and proliferation [23, 26]. Interaction with HCMV-infected fibroblasts promoted an expansion of NKG2C+ NK cells in vitro [27], suggesting the involvement of a cognate interaction of the receptor with a still undefined ligand on infected cells.

The magnitude of the NKG2C+ NK-cell expansion in HCMV+ individuals is quite variable, representing >50% of the peripheral NK-cell pool in some subjects whereas in others their proportions are similar to those found in seronegative donors. Although the effect appears particularly marked in immunocompromised individuals, supporting an inverse correlation with the T-cell mediated control of HCMV infection [7-11, 20, 21, 28], other factors such as host genetic traits might hypothetically underlie the interindividual variability. In this regard, a homozygous deletion of the NKG2C gene was reported in ∼4% of Dutch and Japanese populations [29], and we detected a similar frequency in a Spanish cohort [30].

We recently reported in a small cohort of HCMV+ children that the NKG2C genotype appeared related with the absolute numbers of NKG2C+ and total NK cells, as well as with surface receptor levels [5]. In the present study, we analyzed in healthy young adults (N = 130) the putative influence of NKG2C copy number on the steady-state NK-cell subset distribution as well as on CD94/NKG2C receptor function. The NKG2C+/+ genotype was associated with higher frequencies and absolute numbers of NKG2C+ NK cells in HCMV+ individuals. Moreover, NKG2C copy number was related with CD94/NKG2C surface expression levels and function. These results provide clues to understand the mechanisms underlying the influence of the NKG2C gene dose on the adaptation of the NK-cell compartment to HCMV infection, further supporting an active involvement of the CD94/NKG2C receptor in this process.

Results

Two NKG2C+ NK-cell subsets are defined by the pattern of NK-cell receptor (NKR) coexpression in HCMV+ subjects

HCMV infection promotes a persistent expansion of an NK-cell subpopulation displaying high surface levels of the NKG2C activating NKR (NKG2Cbright) [3, 6]. To study the putative factors underlying the interindividual variability in the proportions of NKG2C+ cells found among HCMV+ healthy subjects, a phenotypic, serologic, and genotypic analysis was carried out in blood samples from young healthy donors (N = 130, median age 19 years). The gating strategy in flow cytometry analyses is displayed as Supporting Information. The HCMV seroprevalence in the studied cohort was 51% and no differences in either age and gender distribution, or in the proportions and absolute numbers of NK and T cells were observed when individuals were stratified according to HCMV serostatus (Supporting Information Table 1). Analysis of NKR distribution confirmed the association between HCMV seropositivity and increased proportions and numbers of NKG2C+ NK and T cells (Supporting Information Table 1). Remarkably, the increase of NKG2C+ NK-cell numbers was concurrent with a significant reduction of the NKG2CNKG2A NK-cell pool, whereas no significant difference in the average of NKG2A+ NK cells was observed.

In a number of HCMV+ individuals, NKG2Cbright NK cells were undetectable, showing instead small proportions of NKG2C+ cells with low surface levels of the receptor (NKG2Cdim). We analyzed the phenotype of NKG2Cbright and NKG2Cdim NK cells from two subgroups of HCMV+ donors, displaying high (MFI >95) or low (MFI<95) NKG2C surface levels, and from a HCMV control group, comparing the coexpression of NKG2A, NKG2D, NKp30, NKp46, CD161, CD16, LILRB1, CD57, and KIR. The cutoff value (95) was defined as the mean NKG2C MFI + 2SD in ex vivo analyzed whole blood samples from HCMV donors (n = 57). In line with previous reports [3, 6], NKG2Cbright NK cells were found in HCMV+ donors displaying elevated proportions of NKG2C+ NK cells. As compared to the NKG2C subset, these cells lacked NKG2A and displayed lower natural cytotoxicity receptors (NCR) and CD161, but higher LILRB1, KIR, and CD57 expression levels. By contrast, NKG2Cdim cells found in other HCMV+ subjects presented a NKR coexpression profile comparable to the NKG2C subset and to NKG2Cdim NK cells from HCMV individuals (Fig. 1 and Supporting Information Fig. 1). Of note, NKG2D and CD16 expression was similar in both NKG2Cbright and NKG2Cdim NK-cell subsets, which coexisted in some HCMV+ donors (Supporting Information Fig. 1, HCMV+ NKG2Cbright individual). Follow-up over a 5-year period of several HCMV+ donors (n = 6) revealed no substantial changes in the frequencies and phenotype of NKG2C+ NK cells, thus supporting the stability of the HCMV imprint on the NK-cell receptor repertoire in healthy donors (Fig. 2).

Figure 1.

Phenotypic characterization of NKG2Cdim and NKG2Cbright NK-cell subsets. Coexpression of NKG2A, LILRB1/ILT2, NKG2D, NKp46, NKp30, CD161, CD16, KIRmix and CD57 in NKG2C+ NK cells analyzed by multicolor flow cytometry in PBMCs from two groups of HCMV+ individuals with low (<95, group 1) and high (>95, group 2) NKG2C surface expression levels (MFI) and HCMV donors (n = 6, each group). Box-and-whisker plots displaying the median (first to third quartiles, whiskers: min to max values) of the proportions of CD56dimCD3NKG2C+ (+) and CD56dimCD3NKG2C NK cells (−) expressing each NKR. Intragroup comparison between NKG2C+ and NKG2C NK cells using paired Student's t-test. Comparison of NKG2C+ NK-cell markers between groups by Mann–Whitney U test. Statistically significant differences are shown (*p < 0.05; **p < 0.01; ***p < 0.001).

Figure 2.

Stability of the NKG2C phenotype over time. Dot plots showing the NKG2C/NKG2A coexpression pattern in NK cells from three different individuals in two determinations sampled along a 5-year interval. Insets indicate the percentage of cells in each quadrant.

Influence of NKG2C copy number on the NK-cell compartment reconfiguration in response to HCMV

NKG2C homozygosity was recently associated with higher numbers of NKG2C+ and total NK cells in a small cohort of HCMV+ children, including newborns with congenital infection [5]. Thus, we assessed whether NKG2C copy number variation would impact the steady-state NK-cell subset distribution in young healthy adults. Analysis of the NKG2C deletion revealed genotype frequencies similar to those previously described [29, 30]. NKG2C+/+, NKG2C+/del, and NKG2Cdel/del individuals, respectively, represented 66, 26, and 8% of the studied population, and no differences in the NKG2C-genotype distribution between HCMV+ and HCMV donors were observed (Supporting Information Table 1).

NKG2C zygosity appeared unrelated with NK- or T-cell numbers, regardless of the HCMV serostatus (Fig. 3A and Supporting Information Table 2), as well as with NKG2C+, NKG2A+, and NKG2CNKG2A NK-cell numbers in the HCMV population (Fig. 3). By contrast, NKG2C+/+ HCMV+ subjects showed significantly increased proportions and absolute numbers of NKG2C+ NK cells, as compared to hemizygous individuals (Fig. 3B). Only in homozygous donors, the increase of NKG2C+ cells was encompassed by reduced numbers of the NKG2CNKG2A population (Fig. 3B and D), without any significant relation of NKG2C zygosity with the average numbers of NKG2A+ NK cells (Fig. 3C). On the other hand, no differences in the proportions of NKG2C+ or NKG2A+ T cells were detected between individuals stratified according to NKG2C copy number (Supporting Information Table 2).

Figure 3.

Influence of NKG2C genotype and HCMV infection on NKG2C+, NKG2A+, and NKG2CNKG2A NK-cell subset distribution. The proportion of CD56+CD3 cells and the expression of NKG2C and NKG2A were analyzed in blood samples by flow cytometry (left panels). Absolute cell numbers were calculated based on the numbers of total lymphocytes per microliter (right panels). HCMV serology and the NKG2C genotype were determined as specified in Methods. The proportions and absolute numbers of (A) NK, (B) NKG2C+, (C) NKG2A+, and (D) NKG2CNKG2A NK cells are displayed. Each symbol represents the result obtained in a single test from an individual donor (n = 130). Significant differences are shown (*p < 0.05; **p < 0.01; ***p < 0.001; Mann–Whitney U test); obvious significant differences with NKG2Cdel/del individuals in panel B are not displayed.

The presence of differentiated NKG2Cbright NK cells was observed in NKG2C+/+ and NKG2C+/del HCMV+ donors (Fig. 4A and B). To more precisely address the influence of NKG2C copy number on the NKG2C profile, HCMV+ donors were arbitrary categorized as NKG2Cbright or NKG2Cdim according to the criteria defined in methods. NKG2Cbright profiles were observed in ∼50% of both NKG2C+/+ and NKG2C+/del HCMV+ donors (Fig. 4C) suggesting that the differentiation of this NK-cell subset occurred comparably in HCMV+ subjects regardless of their NKG2C genotype.

Figure 4.

NKG2C genotype does not determine NKG2Cbright NK-cell differentiation. (A) Presence of differentiated NKG2Cbright cells in NKG2C+/+ and NKG2C+/del donors according to the coexpression of NKG2A, LILRB1, NKp30, KIR. Representative donors out of three analyzed for each NKG2C genotype are shown. (B) Dot plots of NKG2C and NKG2A expression in three examples of NKG2C+/+ and NKG2C+/del HCMV+ donors classified as NKG2Cbright or NKG2Cdim according to the criteria specified in Materials and methods. (C) Bar graph presents the frequency of HCMV+ donors displaying NKG2Cdim or NKG2Cbright NK cells categorized by their NKG2C genotype; number of donors included in each category are indicated.

NK-cell subset redistribution in response to HCMV is not influenced by the HLA-E allelic dimorphism

CD94/NKG2C specifically interacts with HLA-E, which exhibits a characteristic allelic dimorphism (Arg or Gly at position 107) that correlates with different surface expression levels of the class Ib molecule [31]. Therefore, we assessed whether this variable might modulate the influence of NKG2C copy number on the expansion of NKG2C+ NK cells in response to HCMV. HLA-EG107 and HLA-ER107 were represented at similar frequencies (Supporting Information Fig. 2A); the distribution of HLA-E genotypes did not differ significantly from Hardy–Weinberg equilibrium in the whole HCMV+ cohort, neither after stratification for the NKG2C genotype (N = 58, Supporting Information Fig. 2A and B). In agreement with previous results [3], no relationship between the HLA-E dimorphism and the numbers of NKG2C+, NKG2A+, or NKG2CNKG2A NK cells in HCMV+ individuals was perceived (Supporting Information Fig. 2C). Importantly, linear regression analysis did not support any influence of HLA-E allotypes on the association of NKG2C zygosity with the NKG2C+ NK-cell expansion.

Influence of NKG2C copy number on the CD94/NKG2C receptor expression and function

As previously reported in children [5], NKG2C surface expression levels (MFI) were significantly higher in the group of NKG2C+/+ donors, regardless of their HCMV serostatus (Fig. 5). Remarkably, the frequency of NKG2C+ NK cells showed a significant positive correlation with the receptor surface density (MFI) only in HCMV+ individuals, being stronger in NKG2C+/+ (ρ = 0.67, p < 0.0001) than in NKG2C+/del (ρ = 0.48, p = 0.04) donors (Fig. 5B and C).

Figure 5.

Relationship between NKG2C surface levels and the proportion of NKG2C+ NK cells in HCMV+ individuals with NKG2C+/+ and NKG2C+/del genotypes (A) Geometric MFI of NKG2C in NK cells according to NKG2C copy number. Statistically significant comparisons are displayed (*p < 0.05; **p < 0.01; ***p < 0.001; Mann–Whitney U test). (B and C) Correlation between the frequencies of NKG2C+ NK cells and the NKG2C MFI in NKG2C+/+ (filled dots) and NKG2C+/del (empty dots) individuals stratified by their HCMV serostatus. Data were obtained from CD56+CD3NKG2C+ cells stained in whole blood samples and analyzed in a FACSCalibur.

To address whether differences in NKG2C copy number might influence receptor signaling, pairs of NKG2C+/+ and NKG2C+/del individuals displaying similar proportions of NKG2C+ NK cells were selected, and the iCa2+ increase triggered upon receptor engagement in purified NK cells was assessed. Consistent with the higher NKG2C MFI displayed by NKG2C+/+ compared to NKG2C+/del individuals included in each assay, the increase of iCa2+ levels was higher in NKG2C+/+ samples; results were reproducible in four donor pairs with NKG2Cbright (HCMV+, n = 2) and NKG2Cdim (HCMV, n = 2) profiles (Fig. 6). Of note, the rise in iCa2+ was greater in NKG2Cbright than in NKG2Cdim samples (Fig. 6).

Figure 6.

Influence of NKG2C zygosity on Ca2+ influx induced by receptor cross-linking. Indo-1-AM-loaded NK cells from donors with similar percentages of NKG2C+ cells and differing in NKG2C copy number were stained with anti-NKG2C-PE mAb. Basal and receptor-induced iCa2+ mobilization was recorded prior and upon NKG2C engagement with a GaM F(ab)′2 in an LSR II flow cytometer. (A and B) NKG2C surface expression and calcium flux triggered by NKG2C in NK cells from (a) HCMV and (b) HCMV+ individuals with NKG2Cbright phenotype. NKG2C+/+ (black lines) and NKG2C+/del (gray lines) genotypes. Dotted lines correspond to the iCa2+ response of NKG2C cells in the samples. (A, B) Data are shown as mean ± SEM of three independent determinations from each donor and are representative of four experiments. (C) Comparison of the NKG2C-triggered iCa2+ flux in each NKG2C+/+ and NKG2C+/del donor pair. Data are shown as mean ± SEM of three independent determinations from each donor. The percentages of NKG2C+ NK cells, the NKG2C MFI, and the NKG2C+ NK-cell phenotype in each sample are indicated. Statistically significant differences (Student's t-test) are displayed. (D) Comparison of the mean NKG2C-triggered iCa2+ flux in NKG2C+/+ and NKG2C+/del individuals (paired Student's t-test) and between NKG2Cdim and NKG2Cbright NK cells (Student's t-test; *p < 0.05; **p < 0.01; ***p < 0.001).

Next, we assessed the influence of the NKG2C genotype on NK-cell degranulation, comparing NK cells from donor pairs presenting similar percentages of NKG2C+ NK cells, but differing in NKG2C copy number. NKG2Cdim NK cells from HCMV NKG2C+/+ and NKG2C+/del individuals tested in parallel, responded similarly to higher anti-NKG2C mAb concentrations, whereas degranulation of NKG2C+/+ samples was greater at low mAb doses (15–8 ng, Supporting Information Fig. 3A). Similar results were obtained in three independent donor pairs, confirming a consistently lower response to receptor engagement of NK cells from NKG2C+/del subjects (Supporting Information Fig. 3B). By contrast, degranulation of NKG2Cbright NK cells from HCMV+ donors was comparable regardless of NKG2C copy number (Supporting Information Fig. 3C and D). In all cases, the NKG2C MFI was higher in NKG2C+/+ than in NKG2C+/del individuals assayed in parallel.

NKG2C zygosity influences NK-cell proliferation in response to CD94/NKG2C receptor engagement

We previously showed that NKG2C engagement triggers NK-cell proliferation [32]. Moreover, Ly49H signaling through DAP12 has previously been shown to enhance murine NK-cell sensitivity to low IL-15 concentrations [33]. Based on these observations, we assessed whether NKG2C copy number might influence receptor-triggered NKG2C+ NK-cell proliferation. The basal proliferation of NKG2Cbright NK cells to low dose IL-15 stimuli was initially assessed. NKG2Cbright NK cells consistently showed a lower proliferation in response to IL-15 in comparison to NKG2Cdim and NKG2C NK cells (Supporting Information Fig. 4A and B). Engagement of NKG2C by a specific mAb in combination with IL-15 promoted a specific and vigorous proliferation of NKG2Cbright NK cells, whereas anti-NKG2D did not significantly alter the basal response to the cytokine (Supporting Information Fig. 4C). To assess the influence of NKG2C copy number on IL-15-dependent NKG2Cbright NK-cell proliferation we tested, in parallel, PBMCs from HCMV+ donor pairs with similar proportions of NKG2C+ NK cells but different NKG2C genotypes. NKG2C engagement promoted a proliferative response of NKG2Cbright NK cells, stronger in NKG2C+/+ than in NKG2C+/del donors. The proportions of dividing NKG2Cbright NK cells and, particularly, their proliferation index (average number of divisions of the responding cells) appeared higher in NKG2C+/+ individuals (Supporting Information Fig. 5A and B). To verify the apparent influence of NKG2C copy number on the proliferation of NKG2Cbright NK cells, CFSE-labeled, purified NK cells from HCMV+ donor pairs, presenting similar proportions of NKG2Cbright NK cells but a different genotype (Fig. 7B), were cultured with the HLA class I-deficient 721.221 cell line or with the HLA-E-transfected 721.221 cell line (.221-AEH) in the presence of IL-15. Interaction with .221-AEH cells promoted NKG2Cbright NK-cell division and, in line with the results obtained with the specific mAb, samples from NKG2C+/+ individuals showed moderately increased proportions of dividing cells as compared to NKG2C+/del donors, but a higher proliferation index (Fig. 7A, C, and D). To compare the data in three donor pairs independently tested, the proliferation index of NKG2Cbright NK cells at 10:1 NK cell:AEH ratio was normalized to their basal proliferation with IL-15 (Fig. 7C and D). In all cases, CD94/NKG2C-receptor engagement promoted a greater number of cell divisions in NKG2Cbright NK cells from NKG2C+/+ compared to NKG2C+/del individuals, indicating that NKG2C copy number does influence IL-15-dependent proliferation. Remarkably, no significant differences in the proliferation of NKG2Cbright NK cells from NKG2C+/+ and NKG2C+/del donors to IL-15 nor to .221+IL15 was observed (Supporting Information Fig. 6).

Figure 7.

Influence of NKG2C copy number on the CD94/NKG2C receptor-induced proliferation in response to IL-15. (A) CFSE-labeling profiles of NKG2C+ NK cells from the NKG2C+/+ and the NKG2C+/del donors displaying similar percentage of NKG2Cbright NK cells cocultured in parallel with .221-AEH at the indicated NK:target ratios in the presence of IL-15 for 6 days. Insets indicate the proportions of proliferating NKG2Cbright NK cells (CFSE low) and their proliferation index. NKG2C MFI is indicated. Results correspond to one representative experiment of three performed with different donors. (B) Dot plots displaying the expression of NKG2C and NKG2A in NK cells from the selected donor pairs assayed. (C) Bar graph showing the proliferation index (PI) at 10:1 E:T of NKG2Cbright NK cells from three NKG2C+/+ and NKG2C+/del donor pairs, normalized to the basal proliferation to IL-15. Data are shown as mean + SD of duplicate measurements. The percentage of NKG2C and the NKG2C MFI at the end of the assay in nontreated cells are indicated. (D) Comparison of the mean normalized proliferation index between the three pairs of NKG2C+/+ and NKG2C+/del donors. Statistically significant differences by paired Student's t-test.

Discussion

HCMV infection promotes a persistent expansion of NK cells that express the activating receptor CD94/NKG2C yet, it is uncertain which factors determine the wide interindividual variability of this effect in healthy HCMV+ subjects [3, 4, 6]. In this regard, a deletion of the NKG2C gene has been identified at similar frequencies (∼4% NKG2Cdel/del) in different populations [29, 30]. We recently reported that HCMV+ NKG2C+/+ children displayed greater numbers of NKG2C+ NK cells as compared to hemizygous subjects [5]. In the present study, stronger evidence supporting a relationship of NKG2C copy number with steady-state NKG2C+ NK-cell numbers in healthy HCMV+ adults is provided. Moreover, functional studies revealed an influence of NKG2C zygosity on NK-cell activation in response to CD94/NKG2C receptor engagement, thus establishing a putative causal link between the genotype and the NK-cell compartment redistribution.

A detailed analysis in a cohort of young healthy blood donors revealed that HCMV+ NKG2C+/+ subjects display significantly greater numbers of NKG2C+ NK cells than hemizygous individuals. The NKG2C genotype appeared unrelated with NKG2A+ and total NK cells, indicating that the association with broad changes in the NK-cell compartment, previously observed in children [5], did not correspond to steady-state conditions and presumably involved other variables (e.g. closeness of primary infection). HCMV imprint on the NK-cell repertoire showed a remarkable stability in a longitudinal 5-year follow-up of some donors, nonetheless, studies addressing whether differences on the numbers of HCMV-induced NKG2C+ NK cells in NKG2C+/+ and NKG2C+/del populations are maintained in the elderly are warranted.

As previously reported [3], increased proportions of NKG2C+ T lymphocytes were also detected in HCMV+ individuals, but no relation between the NKG2C genotype and the numbers of CD3+ NKG2C+ cells was noticed, consistent with a different regulation of NKR expression in the T-cell lineage. We previously reported that in vitro interaction with HCMV-infected fibroblasts promoted an expansion of NKG2C+ NK cells from HCMV+ donors suggesting an engagement of the CD94/NKG2C receptor [27]. The relationship between the NKG2C gene dose, the steady-state NKG2C+ NK-cell numbers, and the receptor function, further supports an active role of CD94/NKG2C in the HCMV-driven expansion of this NK-cell subset. However, the nature of the viral stimulus promoting the reconfiguration of the NKR repertoire and the putative involvement of HLA-E remain uncertain issues. Of note, the influence of NKG2C copy number on the NK-cell compartment was independent of the HLA-E allelic dimorphism.

Our results highlight that, among HCMV+ blood donors, two NKG2C+ NK-cell subsets can be discriminated based on their phenotype. NKG2Cbright NK cells bear low levels of NKp46, NKp30, and CD161, and display inhibitory NKR, including LILRB1 and KIR specific for self-HLA-C1 or/and HLA-C2 allotypes with an oligoclonal distribution pattern [3, 6, 7]. These cells were only detected in half of the HCMV+ subjects whereas the other half presented instead the NKG2Cdim subset, which partially coexpresses NKG2A and phenotypically resembles NKG2C+ cells from HCMV donors. Of notice, the NKG2Cbright subset was detected with similar frequency in NKG2C+/+ and NKG2C+/del subjects, indicating that other factors (i.e. time of primary infection, viral/hosts genetics, efficacy of the T-cell response) might determine the differentiation of this particular subpopulation. In this regard, a putative influence of NKG2C and CD94 polymorphisms deserves attention [34].

CD56dim NKG2A+ KIR+ NK cells have been proposed to differentiate into polyclonal CD56dim NKG2A KIR+ cells [35-37]. This schematic view of human NK-cell maturation should be revised defining the origin of the differentiated CD56dim NKG2C+ NKG2A KIR+ cells expanding in response to HCMV. Studies on the molecular mechanisms regulating CD94/NKG2C receptor expression and stability in NK cells are warranted to complement our current understanding of human NK cell maturation.

In accordance with observations on the response of mature CD56dim CD57+ NK cells to cytokines [36, 37], NKG2Cbright NK cells displayed a lower basal response to IL-15 than NKG2C and NKG2Cdim subsets. The lower basal sensitivity of HCMV-induced NKG2Cbright NK cells to other cytokines such as IL-12 and IL-18 has already been described [18]. Remarkably, NKG2C engagement promoted vigorous proliferation of NKG2Cbright NK cells to IL-15 thus supporting that their clonal expansion is dependent on receptor signaling, similarly to T lymphocytes. CD94/NKG2C signals through the DAP12 adaptor, which was shown to be strictly required for sustaining Ly49H+ NK-cell expansion upon MCMV infection [33]. Oligoclonal expansions of NKG2Cneg NK cells, which displayed DAP12-coupled activating KIR, have been also reported in some HCMV+ individuals [6]. Whether the ability to promote CD56dim NK cell expansion and differentiation is differentially regulated by DAP12 as compared to other ITAM-bearing adaptors is an intriguing possibility.

We show a clear influence of NKG2C copy number on the receptor expression and early signaling (iCa2+ mobilization) in both NKG2Cbright and NKG2Cdim NK cells. In NKG2Cdim NK cells, NKG2C zygosity conditioned as well the degranulation triggered upon CD94/NKG2C engagement. Consistent with their higher NKG2C surface expression levels (MFI), the response of NKG2C+/+ NK cells was greater than that of NKG2C+/del samples upon stimulation with low concentrations of an agonistic anti-NKG2C mAb. These results predicted that NKG2Cdim NK cells from NKG2C+/+ individuals would be more efficiently activated under a limiting expression of the NKG2C ligand. In contrast, NKG2Cbright NK cells from NKG2C+/+ and NKG2C+/del individuals showed a comparable degranulation but differed in their capacity to undergo prolonged proliferation in response to IL-15 upon NKG2C-engagement. Indeed, NKG2Cbright NK cells from NKG2C+/+ individuals underwent more mitotic cycles than NKG2C+/del samples upon receptor engagement with either an mAb or HLA-E. It is likely that the greater NKG2C-induced iCa2+ mobilization in NKG2Cbright NK cells overcame the degranulation threshold in homozygous and hemizygous samples but still modulated downstream signaling pathways regulating proliferation and/or survival. An alternative possibility is that upon cell division, a higher biosynthetic rate of NKG2C in NKG2C+/+ NK cells may restore the steady-state surface levels of the receptor earlier than in NKG2C+/del cells, thus enhancing their responsiveness to further receptor engagement.

Functional consequences of gene copy number variation have been reported for some immunoreceptors [38, 39]. Interestingly, a recent study proposed that the number of copies of KIR3DS1 and KIR3DL1, in the presence of their ligands, might regulate the expansion of KIR3DS1+ NK cells favoring the control of HIV-1 infection [40]. The pattern of response of NKG2C+ NK cells to HCMV is reminiscent of that reported for the murine Ly49H+ NK-cell subset, which specifically recognizes the m157 MCMV glycoprotein [12, 13]. After expansion in response to MCMV infection, Ly49H+ NK cells persist in the circulation conferring protection against reinfection, being accordingly termed “memory” NK cells [15, 16]. Despite that formal evidence supporting a role of NKG2C+ NK cells in defense against HCMV infection is still missing, it is conceivable that they might control viral replication more efficiently in NKG2C+/+ individuals than in hemizygous subjects, particularly in the context of a defective T-cell response (e.g. immunosuppression). Interestingly, a subset of FcR gamma chain negative NK cells detected in HCMV+ subjects, including NKG2C+NKG2A NK cells, has been reported to mediate an enhanced antibody-dependent cell-mediated cytotoxicity against HCMV-infected cells [22].

In summary, our results support that the NK-cell compartment reconfiguration in HCMV-infected young healthy individuals is tuned by the effect of NKG2C copy number on the receptor expression and signaling, which quantitatively modulates the proliferation and/or survival of NKG2Cbright NK cells, increasing their steady-state numbers in NKG2C+/+ HCMV+ individuals. Our data further support an active involvement of the receptor in this process and contributes to partially explain the variability of the NK-cell compartment redistribution in HCMV+ subjects. This might be of special relevance in the context of the putative role played by HCMV-induced NKG2C+ cells in viral control and other clinical variables in immunosupressed patients [7, 8, 20, 21], particularly in hematopoietic transplant recipients [9-11].

Materials and methods

Ethics statement

PBMCs and NK cells used in this study were obtained from peripheral blood of volunteer donors. Written informed consent was obtained from every donor, and the study protocol was approved by the local ethics committee (CEIC, Parc de Salut Mar n°2010/3766/I).

Subjects and sample collection

Blood samples from a cohort of 130 healthy adults were fractionated to obtain DNA, CMV serology, a basic hemogram, and the analysis of NKR surface expression in NK and T cells by flow cytometry. Standard clinical diagnostic tests were used to analyze serum samples for circulating immunoglobulin G and M antibodies against CMV (bioMérieux Clinical Diagnostics, Marcy l'Etoile, France).

NKG2C and HLA-E genotyping

DNA was isolated from total blood using the Puregene, BloodCore Kit B (Qiagen). NKG2C zygosity was assessed by PCR as in Moraru et al. [41]. Two HLA-E allotypes, HLA-E*0101 (HLA-E107R) and HLA-E*0103 (HLA-E107G), were determined as described [3]. Briefly, HLA-E exons 2 and 3 were PCR-amplified, and codon 107 was analyzed by DNA sequencing using BigDye Terminator v.3.1 Cycle sequencing kit and a 3100 ABI automatic sequencer (Applied Biosystems, Foster City, CA).

Antibodies and flow cytometry analysis

Expression of NKG2A and NKG2C was analyzed by four-color flow cytometry in fresh peripheral blood samples with anti-CD3-PerCP, anti-CD56-APC, anti-NKG2C-PE, and anti-NKG2A-FITC as previously described [42]. Samples were analyzed in a FACSCalibur flow cytometer and data were analyzed with CellQuest software (BD Biosciences). Absolute numbers of NK- and T-cell subsets were calculated based on the hemogram counts. Donor profiles were classified as NKG2Cbright when their NKG2C+ NK cells fitted both of the following criteria: (a) NKG2C MFI ≥95 (mean + 2SD of NKG2C MFI in HCMV donors, excluding NKG2Cdel/del individuals), (b) >60% NKG2C+ NK cells lacking NKG2A (mean + 1SD of the percentage of NKG2C+ NK cells lacking NKG2A in HCMV donors). Otherwise, donor profiles were classified as NKG2Cdim. Since the NKG2C MFI was not sufficiently discriminative (Supporting Information Fig. 7), and based on the results displayed in Figure 1, the lack of NKG2A coexpression was considered for supporting the definition of NKG2Cbright NK cells.

The coexpression of NKG2A, LILRB1, NKG2D, NKp30, NKp46, CD161, CD16, CD57, and KIR (2DL1/S1, 2DL2/S2/L3, 3DL2 and 3DL1) in NKG2C+ NK cells was analyzed by multicolor flow cytometry in CD56dimCD3 PBMCs from HCMV seronegative (n = 6) and two groups of HCMV-seropositive donors displaying low (<95) or high (>95) NKG2C MFI levels (n = 6, each group) as detailed in Supporting Information methods.

Functional assays

NK-cell intracytoplasmic calcium (iCa2+) flux, degranulation (CD107a), and proliferation (CFSE dilution) in response to NKG2C engagement were measured, using standard flow cytometry protocols, in isolated NK cells or PBMCs from NKG2C homozygous and hemizygous donor pairs presenting similar percentages of NKG2C+ NK cells. Experiments were performed with NK cells from HCMV individuals (NKG2Cdim) or from HCMV+ individuals displaying an NKG2Cbright NK-cell phenotype. Details on stimulation conditions for iCa2+ flux measurements, redirected degranulation assays, and proliferation assessment are provided in Supporting Information methods.

Statistical analysis

Continuous variables between two groups were compared using the Mann–Whitney U test; Kruskal–Wallis test was employed to compare medians between three or more groups. p-Values were corrected for multiple comparisons using the Benjamini–Hochberg–Yekutieli procedure and p-values <0.05 were considered statistically significant. Chi-squared test or Fisher's exact test were used as appropriate to compare proportions. Spearman's rank correlation coefficient was used to evaluate the association between continuous variables. Multivariate analysis was carried out using linear regression analysis.

Acknowledgments

This work was supported by grants from Plan Nacional de I+D (SAF2010-22153-C03), EU SUDOE program (SOE2/P1/E341), Red HERACLES (Instituto de Salut Carlos III), and Fundació La Marató TV3 (121531). AM is supported by Asociación Española Contra el Cáncer (AECC). We are grateful to Esther Menoyo for collaborating in obtaining blood samples, Dr. Oscar Fornas for advice in flow cytometry, and to volunteer blood donors for their participation.

Conflict of interest

The authors declare no commercial or financial conflict of interest.

Abbreviation
HCMV

human cytomegalovirus

Ancillary