Altered distribution of natural killer cell subsets identified by CD56, CD27 and CD70 in primary and chronic human immunodeficiency virus-1 infection

Abstract

Human natural killer (NK) (CD3⁻ CD56⁺) cells can be divided into two functionally distinct subsets, CD3⁻ CD56^dim and CD3⁻ CD56^bright. We analysed the distribution of NK cell subsets in primary and chronic human immunodeficiency virus-1 (HIV-1) infection, to determine if HIV infection stage may influence the subset distribution. In primary infection, contrary to chronic infection, the CD3⁻ CD56^dim subset was expanded compared to healthy controls. We also studied the effect of antiretroviral therapy administered early in infection and found that NK cell subset distribution was partially restored after 6 months of antiretroviral therapy in primary infection, but not normalized. Recently, NK cells have been divided into CD27⁻ and CD27⁺ subsets with different migratory and functional capacity and CD27-mediated NK cell activation has been described in mice. We therefore investigated whether CD27 and/or CD70 (CD27 ligand) expression on NK cells, and thus the distribution of these novel NK subsets, was altered in HIV-1-infected patients. We found up-regulated expression of both CD27 and CD70 on NK cells of patients, resulting in higher proportions of CD27^high and CD70^high NK cells, and this phenomenon was more pronounced in chronic infection. Experiments conducted in vitro suggest that the high interleukin-7 levels found during HIV-1 infection may participate in up-regulation of CD70 on NK cell subsets. Imbalance of NK cell subsets and up-regulated expression of CD27 and CD70 initiated early in HIV-1 infection may indicate NK cell activation and intrinsic defects initiated by HIV-1 to disarm the innate immune response to the virus.

Introduction

Natural killer (NK) cells are a key component of the innate arm of the immune system, which is important as a first-line of defence against viral infections. In addition to being of potential use in fighting human immunodeficiency virus-1 (HIV-1) infection, NK cells are targets of the infection themselves^1–3 and are phenotypically and functionally impaired from the early phase of infection.^4,5 HIV-1 infection induces profound immune activation, which may be a root cause of NK cell dysfunction,⁶ but the mechanisms underlying these disturbances have not yet been clarified. Aberrant NK cell activation is reflected in altered expression of NK cell receptors and these phenotypic changes are usually accompanied by important functional defects.^3,6,7

Based on the surface expression of CD56, human NK cells are divided into two subsets: CD56^dim and CD56^bright.⁸ The majority (∼90%) of peripheral blood NK cells belong to the dim subset and are mainly cytolytic while CD56^bright subset (10% of NK cells) primarily produces cytokines.⁹ These subsets have also been shown to migrate in response to different sets of chemokines, and CD56^bright NK cells have been found at sites of inflammation and in lymphoid tissues where they may affect adaptive immune responses.

It was recently shown that CD27, a tumour necrosis factor receptor (TNFR) family member, defines specific subsets of NK cells with distinct responsiveness and migratory capacity.¹⁰ In a murine system, Hayakawa and collaborators showed that CD27 expression divides mature NK cells into CD27^high and CD27^low populations; with the CD27^high NK cells possessing a lower activation threshold and producing larger amounts of cytokines than CD27^low NK cells.^10,11 CD27, one of the non-DD-containing members of the TNFR family, is constitutively expressed by a majority of lymphocytes and has been defined as a costimulatory molecule for T and B cells.¹² In order to contribute to immune responses, TNFRs require the expression of their corresponding ligands, which are themselves members of a family of type II membrane-bound or secreted proteins, known as the TNF family of ligands. The CD27 ligand (CD27L), CD70, is mainly expressed on B and T lymphocytes following antigen stimulation, though it can also be expressed by NK cells.¹² CD27–CD70 interactions have been shown to play a role in several important biological lymphocyte functions¹³ and we previously showed that CD70 is up-regulated on B and T cells of HIV-1-infected patients.^14,15 It has also been suggested that CD27-mediated activation of murine NK cells may be involved in NK cell-mediated innate immunity against virus-infected cells.¹⁶ Given the apparent importance of proper NK cell subset distribution, functional distinctiveness of identified NK cell subsets in both mouse and man, and the malfunction of NK cells in HIV infection, we investigated whether the distribution of CD56⁺ NK cells is altered in HIV infection. We also studied whether the expression of CD27 and CD70 on NK cells, and thus the distribution of CD27⁺ NK-cell subsets, may be modulated during HIV-1 infection. In order to determine what role disease progression may play in altering NK-cell subset distribution, we studied patients with both primary (PHI) and chronic (CHI) HIV infection.

Common cytokine receptor γ-chain (γ_c) cytokines such as interleukin (IL)-2, IL-7, and IL-15 have been shown to play important roles in NK cell generation, differentiation and maturation, and have also been evaluated as immune therapy agents during HIV-1 infection either alone or in combination with antiretroviral therapy.^17–19 Because we have previously shown that serum IL-7 levels are significantly increased in HIV-1 infected patients²⁰ we investigated whether IL-7 could affect the expression of CD27 and CD70 on NK cells in vitro and thus potentially play a role in NK cell dysfunction in HIV infection.

Materials and methods

Study population

Sixty patients with primary HIV-1 infection (PHI) [CD4⁺ T-cell count: 645 cells/μl (range 204–1412); HIV-1 RNA: 4·60 (range 2·69–5·46)] followed at the San Raffaele Institute, Milan were studied. We have previously described the PHI cohort,¹⁴ which consisted of subjects fulfilling at least one critical criterion (signs/symptoms of acute retroviral syndrome (ARS) at visit or during the previous 60 days; HIV exposure in previous 3 months and negative HIV test in previous 6 months) and one laboratory criterion (detectable plasma HIV-RNA; detectable HIV-p24; gp120, gp160 ± p24 bands on Western blot; weakly positive enzyme-linked immunosorbent assay with increasing reactivity over time), according to international guidelines.

Eighteen patients with CHI¹⁴[CD4⁺ T cell count: 591 (68–850); HIV-1 RNA: 2·30 (1·70–4·50)] (infected for 5 years or longer) and on antiretroviral therapy (ART) for at least 1 year were also analysed. Twenty-seven of the 60 PHI patients were sampled at diagnosis and 6 months after ART [CD4⁺ T-cell count: [906 (608–1100); HIV-1 RNA: 2·23 (1·90–2·40)]; all the CHI subjects were on ART. Seventeen age-matched healthy blood donors served as controls. The ethical committees of the Karolinska Institutet and the San Raffaele Institute approved the study.

Cell culture and IL-7 stimulation

PBMC from eight patients and four controls were cultured for 8 days in RPMI medium, supplemented with 10% heat inactivated fetal bovine serum (FBS), penicillin streptomycin (PeSt) 1000 ng/ml and 2 mm l-glutamine, in the presence or absence of 25 ng/ml IL-7. Every second day of culture, cells were washed and old medium discarded and replaced with fresh medium and IL-7 at the same concentration.

Flow cytometry

Peripheral blood mononuclear cells (PBMC) were stained with CD56-phycoerythrin (PE), CD3-Cychrome (Cyc), CD27-fluoroscein isothiocyanate (FITC) or CD70-FITC monoclonal antibodies (BD Pharmingen, San Diego, CA). NK cells were defined as CD56⁺ CD3^– lymphocytes and expression of CD27 and CD70 on NK cells was analysed using three-colour (freshly isolated PBMC) or four-colour (cultured cells) analysis on the gated CD56⁺ CD3^– cells.

Statistical analysis

Data analysis was performed using Sigma Stat for Windows software (SPSS Inc, Chicago, IL). Differences between groups were analysed by parametric (t-test, one-way anova followed by multiple comparisons by Dunn's method) or non-parametric tests (Mann–Whitney or Wilcoxon test, one-way anova on ranks) as appropriate. Differences between baseline and follow-up values in PHI were tested by paired t-tests or Wilcoxon signed rank tests. Correlations between variables were tested using linear regression and Spearman rank tests.

Results

Percentages of total NK cells are normal in HIV-1 infection but the distribution of CD56^dim and CD56^bright NK cell subsets is altered

Figure 1(a) shows a representative dot plot of the PBMC stainings and the gating of total NK cells and CD56^dim and CD56^bright NK cell subsets. Total NK cell percentages are expressed as percentage of total lymphocytes, while NK cell subsets are expressed as percentage of total NK cells.

Figure 1.

NK cell percentages and subset distribution in HIV-1 infection. (a) Representative dot plot showing gating of total, CD56^bright and CD56^dim NK cells.(b) Percentage of NK (CD56⁺ CD3^–) cells in PBMC from healthy controls, patients with primary (PHI) and chronic (CHI) HIV-1 infection. (c) Percentages of CD56^bright and (d) CD56^dim NK cells were determined in healthy controls and in patients with primary (PHI) and chronic (CHI) HIV-1 infection. The boxes depict the 25th and 75th percentiles, with a line at the median and whiskers depicting the 5th and 95th percentiles.

NK cell percentages in primary (PHI) and chronic (CHI) HIV-1 infection were comparable to controls (Fig. 1b) and we did not find any correlation between HIV-1 RNA or CD4 counts and NK cell (CD56⁺ CD3^–) percentage.

In PHI, the CD56^bright subset was contracted compared to controls (P = 0·005). In CHI however, the CD56^bright subset was significantly expanded compared to controls (P = 0·02) (Fig. 1c). Conversely, the CD56^dim subset was expanded in PHI compared to controls (P = 0·005), contrary to CHI in which the CD56^dim subset was diminished compared to both controls (P = 0·02) and PHI (P < 0·001) (Fig. 1d).

We followed 27 PHI patients during 6 months of ART in whom we have previously shown that therapy significantly reduced viraemia.^14,21 Following treatment, the percentages of CD56^bright NK cells increased [from 2% (1·3 to 4·7) to 3·8% (2·2–6·5)], while the percentages of CD56^dim NK cells decreased [from 98% (95·4 to 98·7) to 96·2% (93·6–97·8)] (P = 0·04, for both subsets) though the values still differed significantly compared to healthy controls [CD56^bright: 8·3% (4·2–95·8), CD56^dim: 91·7% (8·5–95·8)] (P = 0·02 for both subsets).

Expression of both CD27 and CD70 is up-regulated on NK cells during HIV-1 infection

Compared to controls, a significantly higher percentage of NK cells expressed CD27 in CHI (P = 0·01) (Fig. 2a), while the percentage of NK cells expressing CD27 in PHI was comparable to that of controls. In general, the MFI of CD27 was also higher on the NK cells from the patients compared to controls (results not shown). There was thus a higher percentage of CD27^high NK cells in HIV-1 infected patients compared to controls.

Figure 2.

Expression of CD27 and CD70 on NK cells in HIV-1 infection. Percentages of NK (CD56⁺ CD3^–) cells in PBMC expressing (a) CD27 and (b) CD70. Percentages of CD56^bright NK cells expressing (c) CD27 and (d) CD70. Percentages of CD56^dim NK cells expressing (e) CD27 and (f) CD70. The boxes depict the 25th and 75th percentiles, with a line at the median and whiskers depicting the 5th and 95th percentiles.

The proportion of NK cells expressing CD70 was also increased in both PHI and CHI compared to controls (P < 0·001) (Fig. 2b). A significantly higher percentage of the CD56^bright subset expressed CD27 compared to the CD56^dim subset in all the groups (Fig. 2c,e). In controls and PHI, similar percentages of CD56^bright and CD56^dim NK cells expressed CD70, while in the CHI group a significantly higher proportion of the CD56^bright NK cells expressed CD70 compared to the CD56^dim subset (Fig. 2d,f).

IL-7 differentially alters expression of CD27 and CD70 on NK cells in vitro

We found increased percentages of NK cells expressing CD27 and CD70 in patients, and have previously found increased serum levels of IL-7 in HIV-1-infected patients.²⁰ Given that cytokines such as TNF-α and IL-4 differentially regulated expression of both CD27 and CD70 molecules²² we wondered if the increased levels of IL-7 could be influencing expression of CD27 and CD70 on NK cells of patients. We analysed the effects of extrinsic IL-7 on expression of CD27 and CD70 on NK cells following 8 days of culture of PBMC in an attempt to determine whether IL-7 may alter expression of CD27 and CD70 on NK cells. We observed a slight decrease in percentage of NK cells expressing CD27 after culture of PBMC with IL-7, but this change did not reach statistical significance (results not shown). However, the percentage of cells expressing CD70 was significantly (P = 0·002) increased compared to baseline (day 0) values in six out of eight HIV-1 infected patients and in three out of four controls. In spite of the fact that there was a significant increase of CD70⁺ NK cells in cultures kept with medium alone for 8 days, the difference between IL-7⁺ and IL-7^– cultures at day 8 was statistically significant in nine (six patients and three controls) out of 12 subjects studied (Fig. 3). This increase in CD70 expression was observed in both the CD56^bright and CD56^dim NK cell subpopulations (data not shown). In the remaining three subjects (two patients and one control) studied, the levels of CD70 expression after 8 days of culture was higher in the IL-7^– wells than in the than in the IL-7⁺ wells with values above 6% of CD70⁺ NK cells in the non-treated cultures.

Figure 3.

CD70 expression on NK cells following short-term in vitro culture of PBMC with or without IL-7. PBMC from patients and controls were cultured for 8 days in RPMI medium alone (untreated) or in the presence of 25 ng/ml IL-7 (IL-7-treated). PBMC were collected on day 8 and stained for CD70 expression. Bars represent the mean ± standard deviation.

Discussion

The work of several investigators (reviewed in ⁶) revealed that, during HIV-1 infection, there is a decrease in percentages/numbers, and cytolytic activity of NK cells. However, contrary to a recently published report²³ we found that NK cell percentages in primary (PHI) and chronic (CHI) HIV-1 infection were comparable to controls. We also showed that the distribution of CD56^bright/dim CD27⁺ and CD70⁺ NK cell subsets, is altered both in primary and chronic HIV-1 infection, with an overall increase in the percentage of CD27^high and CD70^high NK cells in the patients we studied. In addition, 6 months of ART treatment did not normalize the distribution of CD56^bright/dim NK cell subsets.

Increased expression of activation markers by NK cells is observed in individuals with CHI, often accompanied by impaired cytolytic activity^7,24 and perturbation of NK-cell subset distribution is thought to contribute to this dysfunction.^25,26 The effect of HIV-1 infection and the role of high viral loads on NK cell numbers and functions are still conflicting^25,27 although perturbed proportions of NK cells have been suggested to reflect the level of ongoing viral replication.²⁶ Given the loss of CD4⁺ T cells²⁸ and B cells¹⁴ in PHI, the alterations in blood NK cell subsets might be a reflection of total lymphocyte dynamics. There was however, no correlation between CD4⁺ T-cell counts and NK-cell subsets or total NK cell percentage in PHI patients, indicating that altered frequency of NK cell subsets may reflect an intrinsic damage to the NK cell compartment occurring early in HIV-1 infection.

Reports on the effects of ART on NK cell disturbances are also conflicting;^23,25 control/clearance of viraemia has been shown to be associated with a decline in NK cell activity.²⁵ Other reports show decreased NK cell activity²⁶ and imbalance of NK cell subsets in treatment-naïve viraemic patients.²⁹ The partial restoration of the NK cell subset distribution we observed in the PHI patients, indicated that early ART might improve, but not correct the imbalance in NK-cell subsets. However, the reverse pattern of subset distribution in CHI (expanded CD56^bright and contracted CD56^dim) suggests that NK-cell subset imbalance may indeed reflect disease progression.

The CD56^bright subset has a high level of expression of CCR7, unlike the CD56^dim subset on which CCR7 is absent and the CD56^bright subset is thus the predominant NK cell population in lymph nodes.³⁰ One mechanism for the imbalance in distribution of NK cells could thus be skewed homing of different NK cell subsets (at different stages of infection) to lymphoid tissues. The contraction of the CD56^bright subset observed in PHI may be a reflection of intact/increased migration to the lymphoid tissues while the expansion of the CD56^bright subset in CHI may reflect an impaired homing to lymphoid tissue. NK cells are also thought to be rapidly recruited to sites of inflammation/infection and since this process may be dependent on chemokine receptor expression, differential expression of chemokines such as CCR7 on different subsets may be an important underlying mechanism for altered NK-cell subset distribution. CD56^bright NK cells in man have also been shown to express CXCR3, a chemokine that plays an important role in NK cell trafficking, and in this regard are thought to be similar to CD27^high mouse NK-cell subset.¹¹ Detailed studies of lymphoid tissues and NK cell migration and distribution during HIV-1 infection may further elucidate the mechanisms leading to NK-cell subset imbalance in HIV-1.

According to a recent report by Hayakawa et al. NK cells can also be divided into two functionally distinct subsets namely CD27^high and CD27^low, and it was suggested that the CD27^high subset possessed a greater ability to produce IFN-ã in response to cytokine stimulation.¹⁰ Our results suggest a tendency for the balance of NK-cell subsets to shift towards the CD27^high and CD70^high subsets in HIV-1 patients. One may speculate whether this can compensate for impaired cytokine production commonly observed in HIV infection. In a recent mouse model study, blockade of the CD27–CD70 interaction resulted in control of chronic lymphocytic choriomeningitis virus (LCMV) infection and improved neutralizing antibody responses.³¹ Further, it was shown that stimulation of murine NK cells with immobilized anti-CD27 monoclonal antibody or CD70 transfectants could induce proliferation and interferon-γ production.¹⁶ These observations suggest an important role for CD27-mediated activation of NK cells and CD27–CD70 interactions in innate immunity against CD70-expressing virus-infected cells and in the pathogenesis of chronic viral infections. Given that CD27 expression promotes long-term survival of functional effector memory CTLs in HIV-infected patients³² NK cell survival may also be augmented by CD27 expression. Altered expression of CD27 and CD70 on NK cells suggests that interactions between NK cells and B and/or T cells mediated by this receptor-ligand pair may be altered in HIV-1 infection. NK cells with up-regulated expression of CD27 and CD70 in HIV-1 infection may contribute to the persistent delivery of costimulatory signals via CD27–CD70 interactions which may play a role in exhaustion of the T cell pool and induction of lethal immunodeficiency.³³

The expression pattern of inhibitory receptors has also been shown to be distinct on different NK cell subsets, with CD56^bright NK cells expressing little or no killer immunoglobulin-like (KIR) receptors.³⁴ In the mouse system it was also observed that the killer lectin-like receptor G1 (KLRG1) is almost exclusively expressed on CD27^low NK cells.¹⁰ Because NK cells lacking these inhibitory receptors are more efficient at eliciting proper NK cell effector functions, an expansion of the CD27^high CD56^bright subsets would be largely to the benefit of the patients.

Our results indicate that IL-7 may also participate in the deregulation of expression of CD27 and CD70 on NK cells. Studies in an adaptive transfer model suggest that IL-7 may play a role in NK cell homeostasis.³⁵ In light of our observations in HIV-1 patients, the nature of such possible homeostatic regulation merits further investigation. The regulatory effect of IL-7 on NK cells may be indirect, possibly involving IL-7R expressing cells (such as T cells) as in our hands the NK cells stained negative for IL-7Rα (results not shown). A recent paper has however, shown that the CD56^bright NK cell subset expresses IL-7Rα³⁶ and expansion of the bright subset in CHI may restore/improve the ability of this subset to respond directly to IL-7, in an attempt to correct the generalized lymphopenia that characterizes CHI. It may be important to take into account such potential regulatory/modulatory effects of IL-7 on NK cells when evaluating the benefits of including cytokines such as IL-7 in immune therapy for HIV-1.