Human Cytomegalovirus Infection Is Associated with Increased Proportions of NK Cells That Express the CD94/NKG2C Receptor in Aviremic HIV-1–Positive Patients

Killer immunoglobulin-like receptors (KIR), CD94/NKG2 killer lectin-like receptors, and CD85j (also termed ILT2, LIR1, and LILRB1) are expressed by NK and T cell subsets and specifically recognize HLA class I molecules [1]. In particular, NKG2A and NKG2C are C-type lectins encoded at the NK gene complex in human chromosome 12 and, for their surface expression, require a covalent assembly with CD94 [1, 2]. The CD94/NKG2A heterodimer constitutes an inhibitory receptor that recruits the SHP-1 tyrosine phosphatase via the immunoreceptor tyrosine-based motif–bearing NKG2A subunit, whereas CD94/NKG2C is an activating receptor coupled to a tyrosine kinase–activation pathway via the DAP12 adapter [3, 4]. In humans, both of these NK receptors (NKRs) specifically recognize HLA-E, which presents peptides derived from the signal sequences of other HLA class I molecules [5, 6]

Several changes in the NKR repertoire in T and NK cells have been reported in HIV-1–positive patients. Increased proportions of a functionally defective CD56⁻/CD16⁺ NK cell population, associated with a higher expression of inhibitory NKRs and a significant down-regulation of the natural cytotoxicity receptors (NCRs), have been reported in HIV-1–positive patients with detectable viral loads [7 –9]. Moreover, increased expression of inhibitory NKR, particularly CD94/NKG2A, has been found in HIV-specific cytolytic T lymphocytes [10]

Recently, a predominance of the NK cell subset expressing the CD94/NKG2C heterodimer has been reported in HIV-positive individuals, compared with healthy control subjects; this finding suggests that the viral infection is associated with a switch from NKG2A to NKG2C expression [11]. Yet, the mechanisms underlying the impact that HIV infection has on NKR expression are uncertain

Elsewhere, we have reported that increased proportions of CD94/NKG2C⁺ NK cell lymphocytes and ILT2⁺ T cell lymphocytes are detectable in peripheral-blood lymphocytes (PBLs) from human cytomegalovirus (HCMV)–seropositive donors and that these increased proportions are independent of other herpesviruses (e.g., Epstein-Barr and herpes simplex); this finding indicates that the viral infection might shape the NK cell–receptor repertoire [12]. Moreover, we have obtained evidence that an in vitro interaction between PBLs from HCMV-positive donors and HCMV-infected fibroblasts promotes the expansion of the NKG2C⁺ NK and ILT2⁺ T cell subsets [13]

Both HCMV and HIV-1 selectively down-regulate the expression of HLA class I molecules to evade immune recognition by cytolytic T cells but preserve HLA-E, the endogenous ligand for CD94/NKG2 receptors [14, 15]. Because of the potential influence that HCMV has on the expression of CD94/NKG2C, we analyzed the expression of CD94/NKG2A and CD94/NKG2C receptors in HIV-1–positive patients. In the present study, we provide evidence supporting the conclusion that the increased expression of CD94/NKG2C in blood from aviremic HIV-1–positive patients reflects the influence of the underlying HCMV infection rather than a direct effect of HIV-1

Subjects, materials, and methods Blood samples from a cohort of 45 aviremic HIV-1–positive patients (age, 23–41 years; mean±SD, 35.0±3.8 years) attending the Hospital Universitari Germans Trias i Pujol were selected for the analysis. The patients had been receiving successful highly active antiretroviral therapy (HAART) for ⩾1 year and had undetectable HIV-1 RNA loads (<50 copies/mL). The mean ± SD time of suppression of HIV-1 viremia to undetectable levels was 3.5±2.1 years. Control blood samples were taken from an age-matched cohort of healthy volunteer donors (age, 22–54 years; mean±SD, 36.9±10.2 years): 10 donors were from Hospital Universitari Germans Trias i Pujol, and 21 were from Universitat Pompeu Fabra

Standard clinical diagnostic tests (Abbott Laboratories) and a commercially available ELISA kit (Bioelisa CMV Colour; Biokit) were used to determine whether circulating antibodies against HCMV were present in plasma samples from control donors and HIV-positive patients. Plasma levels of HIV-1 RNA were quantified using an Amplicor HIV-1 Monitor 1.5 (Roche Diagnostic Systems) with a sensitivity of 50 copies/mL

Z199 (anti-NKG2A) monoclonal antibody (MAb) has been described in detail elsewhere [12] and was provided by A. Moretta (University of Genova). Anti-NKG2C MAb (MAb1381) and anti–NKG2C-PE MAb were from R&D Systems, Inc.; Z199 (anti-NKG2A) MAb was conjugated to fluorescein isothiocyanate (FITC) (Sigma). Indirect immunofluorescence analysis was performed with phycoerythrin (PE)-tagged F(ab′)₂ rabbit anti–mouse Ig antibodies (Dakopatts). Anti–CD3-PerCP, anti–CD56-PE, and anti–CD56-APC were from BD Biosciences Pharmingen

Heparinized peripheral-blood samples were obtained by venous puncture. Immunophenotypic analysis was performed using samples of whole fresh blood—except in the case of 21 control donors, whose PBLs, isolated by centrifugation on Ficoll-Hypaque (Axis-Shield PoC AS), were analyzed. Indirect immunofluorescence staining was performed according to the protocols detailed below, and samples were subsequently analyzed by use of flow cytometry (FACS) (FACScan; Becton Dickinson). Whole-blood samples were incubated with anti–NKG2A-FITC; subsequently, samples were incubated with anti–NKG2C-PE (R&D Systems Inc.), anti–CD56-APC, and anti–CD3PerCP (BD Biosciences Pharmingen). After being washed, erythrocytes were lysed by FACS lysis buffer (Becton Dickinson), and the cells were fixed and resuspended in 1% paraformaldehyde in PBS. As described in more detail elsewhere [12], PBLs were incubated with individual anti-NKG2C or anti-NKG2A unconjugated MAbs and then were washed and were labeled with FITC-tagged F(ab′)₂ rabbit anti–mouse Ig antibody (Dakopatts); subsequently, samples were incubated with anti–CD56-PE and anti-CD3PerCP (BD Biosciences Pharmingen). For both the whole-blood samples and the PBLs, appropriate anti-isotypic MAbs were used to assess nonspecific staining. As reported elsewhere [12], NKG2C⁺ cells were defined as cells brightly stained for NKG2C (MAb1381⁺). With regard to the detection of NKG2A⁺ and NKG2C⁺ NK cells, there were no significant differences between the two groups of samples. NK cells were conventionally defined as CD3⁻CD56⁺ lymphocytes; it is noteworthy that a CD56⁻CD16⁺ NK cell subset found to be increased in HIV-1–positive subjects who were viremic was negligible in those who were aviremic [7]

To assess the relationship between a categorical variable with 2 levels and either normally or nonnormally distributed quantitative variables, Student’s t test or the Mann-Whitney U test, respectively, was applied. Multivariate linear regression was used to assess the relationship between CD94/NKG2C expression and HIV status, with HCMV status considered to be a potential confounding variable. Analyses were performed with an SPSS (version 11.0; SPSS) statistical package. Results were considered to be significant at a P value (2-tailed) of .05

Results and discussion We studied the expression of CD94/NKG2 receptors in NK cells from HIV-1–positive patients (n = 45) with undetectable HIV-1 RNA load who had been receiving successful HAART for ⩾1 year. Because the proportions of NKG2C⁺ (mean±SD, 12.9% ± 20.7% vs. 16.8% ± 22.4%; P=.3) and NKG2A⁺ (mean±SD, 45.1% ± 15.5% vs. 41.7% ± 15.6%; P=.6) cells were not significantly different between 2 cohorts of independently analyzed control donors, the data were pooled to compare them with those obtained from HIV-1–positive patients

In HIV-1–positive patients, the mean ± SD proportions of NKG2C⁺ and NKG2A⁺ NK cells were 25.9%±23.0% (range, 0.1%–72.0%) and 42.7%±19.1% (range, 8.5%–82.8%), respectively; in control donors, they were 16.1%±20.7% (range, 2.8%–83.0%) and 43.1%±15.0% (range, 9.2%–75.0%), respectively. Figure 1 shows the distribution of NKG2C⁺ and NKG2A⁺ NK cells in 2 representative HIV-positive patients (A and B) and 2 control donors (C and D), illustrating the wide variability in the expression of both markers. The proportions of CD94/NKG2C⁺ NK cells in control donors were higher than those reported by Mela et al. [11]; yet, in agreement with their results, a significant increase of NKG2C⁺ NK cells was also observed in HIV-1–positive patients compared with control donors (mean±SD for HIV-1–positive subjects vs. control donors, 25.9%±23.0% vs. 16.1%±20.7%; P=.04) (table 1). By contrast, no relationship was seen between HIV-1 infection and the proportions of NKG2A⁺ NK cells (mean±SD for HIV-1–positive subjects vs. control donors, 42.7%±19.1% vs. 43.1%±15.0%; P=.7), which confirms a previous report indicating that the proportions of NKG2A⁺ NK cells are similar in both healthy and aviremic HIV-1–positive subjects [7]

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 1

Comparative analysis of CD94/NKG2C and CD94/NKG2A expression in blood from healthy control donors and from human immunodeficiency virus type 1 (HIV-1)–positive patients. Fresh blood was stained with anti-NKG2C, anti-NKG2A, anti-CD56, and anti-CD3 monoclonal antibodies (see the “Subjects, materials, and methods” section). Samples were analyzed by use of flow cytometry, and the proportions of NKG2A⁺ and NKG2C⁺ cells in gated CD56⁺CD3⁻ populations were calculated. The staining patterns in samples from 2 representative HIV-1–positive patients (A and B) and 2 healthy control donors (C and D) are displayed. Samples were derived from human cytomegalovirus (HCMV)-negative (A and C) and HCMV-positive (B and D) individuals. FITC, fluorescein isothiocyanate; PE, phycoerythrin

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

Proportions of NK cells expressing CD94/NKG2C and CD94/NKG2A, by human immunodeficiency virus (HIV) and human cytomegalovirus (HCMV) infection status

To ascertain whether the differences in NKG2C expression were directly related to the HIV-1 infection (and not to the influence of HCMV, as has elsewhere been reported for healthy donors [12]), both HIV-positive patients and HIV-negative control donors were subclassified on the basis of the presence or absence of HCMV-specific circulating antibodies, as follows: HIV-negative/HCMV-positive (n=17 [55% of HIV-negative control donors]), HIV-positive/HCMV-positive (n=38 [85% of HIV-positive patients]), HIV-negative/HCMV-negative (n = 14 [45% of HIV-negative control donors]), and HIV-positive/HCMV-negative (n=7 [15% of HIV-positive patients]). A strong association between the proportions of NKG2C⁺ cells and HCMV seropositivity was observed, as NKG2C⁺ NK cells were significantly increased in HCMV-positive individuals, in both the HIV-positive and the HIV-negative groups (P=.01 and P=.003, respectively) (table 1 and figure 1). The proportions of NKG2C⁺ and NKG2A⁺ cells in HCMV-positive and HCMV-negative individuals appeared to be similar in both the HIV-positive and the HIV-negative groups (table 1). The relationship between CD94/NKG2C expression and HIV infection, when HCMV serologic status is considered to be a potential confounding variable, was analyzed by use of a multivariate linear regression model; a significant correlation between HIV status and the proportion of NKG2C⁺ NK cells was observed (P=.04), but it vanished when the HCMV variable was jointly considered in the model (P=.5)

When considered together, these data support the conclusion that HIV infection does not directly influence the distribution of the NKG2C⁺ cell subset in aviremic patients, and that the differences between HIV-positive and HIV-negative individuals can be attributed to the higher proportion of HCMV-positive individuals among the HIV-positive population. Mavilio et al. [8] detected in viremic HIV-positive patients a subset of CD56⁻CD16⁺ cells that expressed inhibitory NKRs (i.e., KIR and ILT2) but displayed low surface levels of NCR and depressed cytolytic activity. Moreover, these cells were reported to be CD94⁺/NKG2A⁻, leaving open the possibility that they may express NKG2C; this possibility was not directly analyzed. It is noteworthy that this phenotype appears reminiscent of that originally reported in NKG2C⁺ NK cells from HCMV-positive healthy donors [12] and recently confirmed, by Mela et al. [11], in HIV-1–positive patients. The results of the present study emphasize the importance of considering HCMV infection when one is interpreting the impact that immunodeficiencies and immunosuppression have on NK cells. Little information is available on the relationship between HIV and HCMV viremia and on the latter’s possible impact on NKG2C⁺ cells. Further studies are necessary to sort out the relative influence that HIV and HCMV exert on the NKR repertoire in HIV-1–positive patients with detectable levels of HIV-1 RNA

• Received December 20, 2005.
• Accepted February 23, 2006.

• © 2006 by the Infectious Diseases Society of America