In Vivo Activation of Human NK Cells by Treatment with an Interleukin-15 Superagonist Potently Inhibits Acute In Vivo HIV-1 Infection in Humanized Mice

ABSTRACT

Natural killer (NK) cells with anti-HIV-1 activity may inhibit HIV-1 replication and dissemination during acute HIV-1 infection. We hypothesized that the capacity of NK cells to suppress acute in vivo HIV-1 infection would be augmented by activating them via treatment with an interleukin-15 (IL-15) superagonist, IL-15 bound to soluble IL-15Rα, an approach that potentiates human NK cell-mediated killing of tumor cells. In vitro stimulation of human NK cells with a recombinant IL-15 superagonist significantly induced their expression of the cytotoxic effector molecules granzyme B and perforin; their degranulation upon exposure to K562 cells, as indicated by cell surface expression of CD107a; and their capacity to lyse K562 cells and HIV-1-infected T cells. The impact of IL-15 superagonist-induced activation of human NK cells on acute in vivo HIV-1 infection was investigated by using hu-spl-PBMC-NSG mice, NOD-SCID-IL2rγ^−/− (NSG) mice intrasplenically injected with human peripheral blood mononuclear cells (PBMCs) which develop productive in vivo infection after intrasplenic inoculation with HIV-1. IL-15 superagonist treatment potently inhibited acute HIV-1 infection in hu-spl-PBMC-NSG mice even when delayed until 3 days after intrasplenic HIV-1 inoculation. Removal of NK cells from human PBMCs prior to intrasplenic injection into NSG mice completely abrogated IL-15 superagonist-mediated suppression of in vivo HIV-1 infection. Thus, the in vivo activation of NK cells, integral mediators of the innate immune response, by treatment with an IL-15 superagonist increases their anti-HIV activity and enables them to potently suppress acute in vivo HIV-1 infection. These results indicate that in vivo activation of NK cells may represent a new immunotherapeutic approach to suppress acute HIV-1 infection.

IMPORTANCE Epidemiological studies have indicated that NK cells contribute to the control of HIV-1 infection, and in vitro studies have demonstrated that NK cells can selectively kill HIV-1-infected cells. We demonstrated that in vivo activation of NK cells by treatment with an IL-15 superagonist that potently stimulates the antitumor activity of NK cells markedly inhibited acute HIV-1 infection in humanized mice, even when activation of NK cells by IL-15 superagonist treatment is delayed until 3 days after HIV-1 inoculation. NK cell depletion from PBMCs prior to their intrasplenic injection abrogated the suppression of in vivo HIV-1 infection observed in humanized mice treated with the IL-15 superagonist, demonstrating that activated human NK cells were mediating IL-15 superagonist-induced inhibition of acute HIV-1 infection. Thus, in vivo immunostimulation of NK cells, a promising therapeutic approach for cancer therapy, may represent a new treatment modality for HIV-1-infected individuals, particularly in the earliest stages of infection.

INTRODUCTION

The crucial role of the human immunodeficiency virus (HIV)-specific T cell and antibody response mounted by the adaptive immune system to control HIV-1 infection is well established (1). However, during acute infection, viremia is not controlled because it takes several weeks after the initiation of infection for the adaptive immune response to activate and clonally expand sufficient numbers of HIV-1-specific T cells and B cells to suppress HIV-1 infection (2). This delay in the mobilization of the adaptive immune response permits HIV-1 to rapidly replicate and disseminate during the acute phase of infection, leading to the production of high plasma viral loads, which are associated with an adverse disease course (3, 4). Early control of HIV-1 replication can have a beneficial impact on the subsequent disease course, as evidenced by the ability of some individuals whose viremia was suppressed by combination antiretroviral therapy (cART) during acute infection to achieve long-term infection control despite lacking protective HLA-B alleles (5–7). Prior to the development of an effective HIV-1-specific adaptive immune response, natural killer (NK) cells, crucial innate immune effector cells which are large granular cytotoxic lymphocytes, are rapidly activated and expanded and may contribute to controlling the initial phase of HIV-1 replication (8, 9). Infection induces changes in the cellular expression of ligands recognized by NK cell receptors, which enables NK cells to specifically identify and kill virus-infected cells to control and/or abort viral infections prior to the initiation of antigen-specific responses (10). One mechanism by which HIV-1-infected cells become susceptible to killing by NK cells is through a reduction in their surface expression of major histocompatibility complex (MHC) class I molecules mediated by HIV-1 Nef as a means of evading killing by HIV-1-specific CD8⁺ cytotoxic T cells (11). Further support for the role of NK cells in controlling HIV-1 replication and improving clinical outcomes is provided by several genetic population studies linking slower HIV-1 disease progression with the expression of specific allotypes of killer cell immunoglobulin-like receptors (KIRs) and their respective HLA class I ligands (10). However, the correlates of protective HIV-1-specific immunity conferred by NK cells are not well characterized, and there is no direct evidence that in vivo stimulation of NK cell activity can enhance their capacity to inhibit acute in vivo HIV-1 infection.

After infection, NK cells are rapidly activated by interleukin-15 (IL-15), a multifunctional cytokine produced by activated dendritic cells and macrophages which enables NK cells to generate protective responses capable of clearing viral infections (12–14). In contrast to IL-2 and other cytokines, which are secreted and circulate until they bind directly to their cognate receptors on target cells, IL-15 secreted by dendritic cells and macrophages binds to the IL-15-specific receptor alpha chain (IL-15Rα/CD215) embedded on their cell surfaces to form a membrane-bound IL-15:IL-15Rα complex. This complex is then presented in trans to bind to the IL-15Rβ (IL-2Rβ/CD122) and IL-15Rγ (IL-2Rγ/CD132) chains expressed by adjacent NK cells and CD8⁺ T cells, which activate them (13, 15). Alternatively, IL-15:IL-15Rα complexes are cleaved from dendritic cell and macrophage membranes and are released into the serum, where they circulate as stable heterodimers that display greater in vivo bioactivity and a longer half-life (4 h versus 30 min) than those of secreted IL-15 monomers (16), to potently stimulate NK cells and/or CD8⁺ T cells to eliminate virus-infected cells (17). Consequently, the functional activity of recombinant IL-15 can be greatly augmented by complexing it with soluble recombinant IL-15Rα and thereby converting IL-15 from an agonist to a superagonist (6). To replicate the increased functional activity of the IL-15:IL-15Rα complex, we constructed a fusion protein, IL-15RαSu/Fc, which consists of two linked IgG1 Fc domains, each fused to the 65-amino-acid (aa) Sushi (Su) domain of IL-15Rα (18). The Sushi domain is then bound to IL-15N72D, an IL-15 molecule with the asparagine at position 72 mutated to aspartic acid, which increases its affinity for the human IL-15Rβ chain and its biological activity by ∼5-fold compared to native human IL-15 (19). As a consequence of these structural modifications, the IL-15 superagonist, identified as ALT-803 for phase I studies evaluating its efficacy as a candidate cancer therapeutic, exhibits a ∼25-fold-higher biological activity and a >35-fold-longer serum half-life (∼25 h versus ∼40 min) than those of soluble IL-15 and potently stimulates in vivo NK cell and CD8⁺ T cell proliferation in mice (20). In the current study, we investigated whether the capacity of NK cells to inhibit acute HIV-1 infection could be stimulated by the in vivo administration of the IL-15 superagonist to humanized mice days after inoculation with HIV-1. For these experiments, we used a humanized mouse model, hu-spl-PBMC-NSG mice, generated by intrasplenically injecting NOD-SCID-IL2rγ^−/− (NSG) mice with activated human peripheral blood mononuclear cells (PBMCs), which develop rapid and robust HIV-1 infection after intrasplenic inoculation with HIV-1.

MATERIALS AND METHODS

Generation of HIV-1 infectious molecular clones.

Virus stocks were generated by transient transfection of 293T cells with a plasmid, NL-LucR.T2A-JR-CSF.ecto, which encodes an infectious HIV-1 molecular clone whose Env protein ectodomain sequence is derived from JR-CSF Env and which expresses the Renilla reniformis luciferase (LucR) gene (HIV-LucR), as previously described (21). LucR has a short cellular half-life of ∼3 h (22), so its expression reflects active replication. HIV-LucR, which was engineered to express the LucR reporter gene and the heterologous JR-CSF env gene in cis with all of the HIV-1 open reading frames, continues to express LucR over multiple cycles of replication after inoculation, permitting highly sensitive and specific detection of active HIV-1 replication for several weeks after inoculation (21). The infectious titer of HIV-LucR (5 × 10⁷ to 10 × 10⁷ infectious units [IU]/ml) was determined by limiting-dilution infection of TZM-bl cells, as described previously (21).

Generation of the IL-15 superagonist fusion complex.

The expression vector pMSGV-IL-15RαSu/Fc expressing the IL-15RαSu/Fc fusion gene was constructed by overlap PCR amplification of DNA templates encoding the Su domain of human IL-15Rα (aa 1 to 65 of human IL-15Rα) and the human IgG1 Fc fragment and was ligated into the pMSGV-1 expression vector, as described previously (23). To bind the two IL-15 binding sites in the IL-15RαSu/Fc fusion protein, we used IL-15N72D, human IL-15 mutated to replace the native asparagine at position 72 with an aspartic acid. This IL-15 mutein displays ∼5-fold-higher levels of biological activity than that of native IL-15 due to its increased affinity for and binding to the human IL-15Rβ chain (19). The complex of IL-15RαSu/Fc bound to IL-15N72D (IL-15 superagonist) was generated by cotransfecting expression vectors encoding IL-15RαSu/Fc and IL-15N72D into CHO cells and purifying the soluble IL-15 superagonist fusion protein complex by using a two-step affinity and ion-exchange chromatography-based process, as described previously (20).

Measurement of the in vivo capacity of the IL-15 superagonist to inhibit in vivo HIV-1 infection.

NSG mice (Jackson Laboratories, Bar Harbor, ME) were bred and maintained under isolation and biocontainment conditions, as described previously (24). The in vivo anti-HIV-1 activity of the IL-15 superagonist was determined by using humanized mice generated through a modification of an in vivo adoptive transfer system that we developed (25). PBMCs isolated from HIV-1-naive donors were cultured in complete media, RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum, penicillin (100 U/ml), streptomycin (10 μg/ml), glutamine (2 mM), and HEPES (10 mM); activated with phytohemagglutinin (PHA) (4 μg/ml) and IL-2 (100 U/ml) for 24 h; and washed and resuspended (∼8 × 10⁷ cells/ml) in sterile cold phosphate-buffered saline (PBS). NSG mice were intrasplenically injected with activated human PBMCs (∼8 × 10⁶ cells) to generate hu-spl-PBMC-NSG mice. Some of the hu-spl-PBMC-NSG mice were challenged in vivo with HIV-1 by intrasplenic injection of 100 μl of HIV-LucR (1 × 10⁸ IU/ml) in parallel with intrasplenic injection of the activated PBMCs. One or three days after intrasplenic HIV-1 challenge, the mice were treated with one intravenous dose of the IL-15 superagonist (0.2 mg/kg of body weight), which provided the mice with a maximum concentration of drug in serum (C_max) of ∼25 nM. For some experiments, PBMCs were depleted of CD8⁺ T cells or NK cells (>95% depletion) prior to their injection into mice by immunomagnetic sorting using anti-human CD8 or anti-human CD56 microbeads (Miltenyi Biotec, Cambridge, MA). The level of HIV-1 infection that developed in transferred human PBMCs in the spleens of hu-spl-PBMC-NSG mice was quantified by measuring the LucR activity in the mouse splenic lysates by using the Renilla luciferase assay system (Promega, Madison, WI), as described previously (26). To visualize and quantify in vivo HIV-1 infection using bioluminescent imaging, mice were imaged with the IVIS Spectrum imager (Caliper LifeSciences, Hopkinton, MA) after intravenous injection (5 μg/mouse) of the bioluminescence substrate RediJect Coelenterazine h (Caliper Life Sciences), and the images were analyzed by using the Wizard bioluminescent selection tool for automatic wavelength and exposure detection. The bioluminescent and grayscale images were overlaid by using the LivingImage 4.0 software package to create a pseudocolor image that represents bioluminescence intensity. The bioluminescence intensity in the mouse spleens was quantified by using the LivingImage 4.0 software package and reported as photon counts/second.

Flow cytometric analysis for the evaluation of cellular phenotype and perforin and granzyme B expression.

Splenocytes isolated from the mice were stained with allophycocyanin (APC)-labeled anti-human CD45 monoclonal antibody (MAb), fluorescein isothiocyanate (FITC)-labeled anti-human CD3 MAb, phycoerythrin (PE)-Cy7-labeled anti-human CD4 MAb, Pacific Blue-labeled anti-human CD8 MAb, PE-labeled anti-human CD56/CD16 MAb, and APC-labeled anti-human CD16 MAb (all from BioLegend, San Diego, CA) and were analyzed by using an LSRII instrument (BD Biosciences, San Jose, CA) and FlowJo software (Treestar, Ashland, OR), as described previously (24). The effect of IL-15 superagonist treatment on perforin and granzyme B expression in NK cells was determined by treating human PBMCs from seronegative donors with the indicated concentrations of the IL-15 superagonist for 2 days. PBMCs were washed, stained with APC-Cy7-labeled anti-human CD3 MAb and PE-labeled anti-human CD56 MAb, incubated in fixation and permeabilization buffers, intracellularly stained with FITC-conjugated anti-human granzyme B MAb and peridinin chlorophyll protein (PerCP)-Cy5.5-labeled perforin MAb (BioLegend), and washed. The mean fluorescence intensity (MFI) (geometric mean) of granzyme B and perforin expression by gated CD3⁻ CD56⁺ NK cells was determined by analysis on a FACSverse instrument with FACSuite software (BD Biosciences).

Measurement of NK cell cytotoxic function directed against K562 cells and HIV-1-infected ACH2 cells.

The capacity of IL-15 superagonist treatment to stimulate NK cell degranulation was examined by quantifying the expression of CD107a by flow cytometry after incubation of PBMCs harvested from HIV-negative donors with K562 cells, as described previously (27). In parallel, we determined the effect of IL-15 superagonist stimulation on the functional cytolytic activity of NK cells by quantifying their capacity to lyse K562 target cells using a modification of a previously reported technique (28). Briefly, K562 target cells were labeled with CellTrace Violet (Invitrogen-Life Technologies, Grand Island, NY). Human PBMCs either untreated or treated with the IL-15 superagonist (10 nM) were mixed with CellTrace Violet-labeled K562 target cells at the indicated effector-to-target cell (E:T) ratios in complete medium and incubated at 37°C with 5% CO₂ for 3 days. K562 target cell viability was assessed by analysis of propidium iodide-positive staining of the CellTrace Violet-labeled K562 target cells on a BD FACSverse instrument, and cytotoxicity was calculated as the percentage of dead target cells in samples cocultured with effector cells.

To determine the capacity of the IL-15 superagonist to stimulate NK cell-mediated lysis of HIV-1-infected cells, PBMCs were harvested from HIV-negative donors and cultured (5 × 10⁵ cells/well) with the indicated concentrations of the IL-15 superagonist for 2 days. ACH2 cells, T cells latently infected with a single integrated HIV-1 provirus which display minimal HIV-1 replication unless stimulated with phorbol myristate acetate (PMA) and/or tumor necrosis factor alpha (TNF-α), were obtained through the NIH AIDS Reagent Program, Division of AIDS, NIAID, NIH (29). ACH2 cells were used as target cells and either were not activated or were stimulated with PMA (2 ng/ml) and TNF-α (2 ng/ml) for 24 h to stimulate HIV-1 production, as described previously (30). After the effector cells were cultured for 48 h with the indicated doses of the IL-15 superagonist, the effector cells were cocultured with unactivated or activated ACH2 cells at an E:T ratio of 25:1 for 24 h, as described previously (31). Lysis of target cells was then measured by using the fluorescence-based CytoTox-One homogeneous membrane integrity assay (Promega) according to the manufacturer's instructions. Background fluorescence was determined by adding unstimulated target ACH2 cells (2 × 10⁴ cells/well) to effector PBMCs (5 × 10⁵ cells/well) that were either untreated or treated with the IL-15 superagonist at an E:T ratio of 25:1. The percentage of specific cell lysis was determined by subtracting background fluorescence from experimental fluorescence and dividing this value by the maximum lactate dehydrogenase (LDH) release from target cells, determined by lysing (2 × 10⁴ cells/well) stimulated ACH2 cells and then multiplying the resulting number by 100.

Statistical analysis of data.

GraphPad Prism statistical software was used for statistical analysis with an independent two-tailed Student's t test. Differences were considered statistically different for P values of <0.05.

Study approval.

All the studies were performed under protocols approved by the Institute for Animal Studies and the Institutional Review Board at the Albert Einstein College of Medicine in compliance with the human and animal experimentation guidelines of the U.S. Department of Health and Human Services.

RESULTS

The in vitro cytolytic activity of human NK cells is induced by IL-15 superagonist treatment.

The IL-15 superagonist IL-15N72D:IL-15RαSu/Fc is a homodimer of two engineered IgG1 Fc domains fused with the IL-15Rα binding domain, each of which is bound to a mutated IL-15 molecule, IL-15N72D, which activates IL-15Rβγ chain-expressing NK cells by mimicking the trans-presentation of IL-15 by the soluble IL-15:IL-15Rα complex (Fig. 1A). Based on its capacity to potently activate the cytotoxic activity of NK cells and CD8⁺ T cells, this IL-15 superagonist, also identified as ALT-803, has shown promise as an immunostimulatory agent to treat cancer (32). We investigated whether the activation of NK cells by treatment with this IL-15 superagonist would increase their cytotoxic activity to enable them to eliminate HIV-1-infected cells and inhibit the in vivo acquisition of HIV-1 infection. The predominant killing pathway utilized by NK cells to kill target cells that they identify as malignant or infected is by exocytosis of preformed granules containing lytic proteins, mainly perforin and granzyme B, into a cytotoxic synapse formed with the target cells (33). Because increasing cellular levels of perforin and granzyme B increase the cytotoxic capacity of NK cells (34), we examined whether IL-15R superagonist treatment increased the cytolytic capacity of NK cells by inducing perforin and granzyme B production. Human PBMCs from six seronegative donors were treated with the IL-15 superagonist or untreated for 2 days and then analyzed by flow cytometry to quantify the intracellular expression of perforin and granzyme B by NK cells. Compared to the NK cells in untreated PBMCs, the NK cells in human PBMCs treated with the IL-15R agonist for 2 days displayed significantly higher levels of perforin (Fig. 1B) and significantly higher levels of granzyme B (Fig. 1C) in a dose-responsive fashion. CD107a is a marker of NK cell degranulation, and its upregulation on cell surfaces is induced by interaction with MHC-negative target cells and correlates with NK cell lysis of target cells and cytokine secretion (27). To evaluate the stimulating capacity of the IL-15 superagonist to increase the cytolytic activity of NK cells, we incubated human PBMCs isolated from seronegative donors with K562 cells, a highly sensitive in vitro target for NK cells, and examined the effect of treatment with the IL-15 superagonist on the NK cell surface expression of CD107a. In the absence of K562 cells, incubation of PBMCs with the IL-15 superagonist (1,000 pM) did not significantly increase cell surface CD107a expression by NK cells (∼5% CD107a positive) (data not shown). When the PBMCs were incubated with the target K562 cells, cell surface CD107a expression increased by 5-fold (∼26%), and treatment of the PBMCs with the IL-15 superagonist significantly increased their cell surface CD107a expression by an additional 2.3-fold (∼62%) in a dose-responsive manner (Fig. 1D). Increased cell surface expression of CD107a induced by treatment with the IL-15 superagonist was associated with an increase in the in vitro capacity of the NK cells in the PBMCs to lyse K562 cells (Fig. 1E), the standard NK-sensitive target cells for human NK cell assays (28). In addition, NK cell stimulation by the IL-15 superagonist was also indicated by its induction of the early activation marker CD69 (data not shown). Taken together, these data demonstrate that the IL-15 superagonist enhances the in vitro cytotoxic capability of human NK cells, possibly via upregulating their expression of the effector molecules granzyme B and perforin and increasing their activation state.

FIG 1.

Treatment of human PBMCs with the IL-15 superagonist increases their cytotoxic activity. (A) Schematic structure of the IL-15 superagonist. (B and C) Intracellular perforin expression (B) and intracellular granzyme B expression (C) in CD3⁻ CD16⁺ CD56⁺ cells were determined by flow cytometry after human PBMCs from six seronegative donors were incubated for 2 days with the indicated concentrations of the IL-15 superagonist. Data represent the mean ± standard error of MFI values obtained from the six donors. (D) After human PBMCs were stimulated with the indicated concentrations of the IL-15 superagonist for 16 h, they were coincubated with K562 target cells (5:1 E:T ratio) for 4 h, and CD107a expression by CD3⁻ CD56⁺ NK cells was determined by flow cytometry. The results are expressed as mean percentages of CD107a-positive CD3⁻ CD56⁺ NK cells from PBMCs isolated from 4 normal donors ± standard errors. (E) Human PBMCs were cultured with K562 cells at the indicated E:T ratios with or without the IL-15 superagonist (10 nM). After 3 days, the fraction of dead K562 cells was quantified by flow cytometry. (F) Human PBMCs were cultured with activated ACH2 cells with or without the indicated doses of the IL-15 superagonist. After 24 h, the percent cell lysis of target cells was measured by using the CytoTox-One homogeneous membrane integrity assay. Data represent results from at least three replicates per group with the mean ± standard error for the group. *, P < 0.05; **, P ≤ 0.01; ***, P ≤ 0.001 (by Student's t test).

To determine whether IL-15 superagonist stimulation increased the capacity of NK cells to kill HIV-1-infected cells, we used ACH2 cells, a latently infected cell line which can be induced to produce HIV-1 after PMA and/or TNF-α treatment (29) and which has previously been used as target cells to evaluate HIV-1-specific NK cell-mediated antibody-dependent cellular cytotoxicity (ADCC) activity (31). Unstimulated PMBCs and IL-15 superagonist-stimulated PBMCs were incubated with unactivated ACH2 cells or PMA/TNF-α-stimulated ACH2 cells, and lysis of ACH2 cells was quantified by measuring the amount of LDH released into the culture supernatant. Human PBMCs treated with the IL-15 superagonist (100 nM) displayed ∼8-fold increased lysis of stimulated ACH2 cells (P < 0.05) compared to unstimulated PBMCs (Fig. 1F).

In vivo HIV-1 infection is potently suppressed by treatment of hu-spl-PBMC-NSG mice with the IL-15 superagonist.

We next examined whether the increased in vitro NK cell cytotoxic activity induced by the IL-15 superagonist correlated with an increased capacity of IL-15 superagonist-activated NK cells to inhibit acute in vivo HIV-1 infection. For these studies, we used a humanized mouse model that supports in vivo HIV-1 infection, hu-spl-PBMC-NSG mice, NSG mice intrasplenically injected with human PBMCs, which we previously used to evaluate HIV-1-specific CD8⁺ T cell activity (25). The spleens of these mice continue to be populated with human PBMCs, including NK cells, and support HIV-1 infection for at least 1 month after intrasplenic injection. This model also permits us to identify effector cells responsible for anti-HIV-1 activity by removing specific cell populations from human PBMCs prior to intrasplenically injecting them into mouse spleens. As the infectious inoculum, we used HIV-LucR, a replication-competent molecular HIV clone expressing the HIV-1_JR-CSF Env protein and a Renilla luciferase reporter gene, which enables HIV-1 infection to be quantified by directly measuring the luciferase activity of infected cells (21). HIV-LucR has been well established as an infectious inoculum that can be used to evaluate the capacity of antibodies, NK cells, and CD8 T cells to inhibit HIV-1 infection (35–42). One day after hu-spl-PBMC-NSG mice were infected with HIV-1 by intrasplenic injection with HIV-LucR, one group of mice was intravenously injected with a dose of the IL-15 superagonist. Five days later, HIV-1 infection was quantified by determining the level of expression of the luciferase reporter gene in the mouse spleens. Splenic lysates of the IL-15 superagonist-treated mice displayed a >99% reduction in luciferase activity (average of 804 relative light units [RLU]) compared to untreated mice (average of 827,829 RLU), indicating that the IL-15 superagonist treatment delivered 1 day after direct intrasplenic challenge with HIV-1 significantly suppressed (P < 0.001) acute HIV-1 infection (Fig. 2A). The use of infectious HIV-1 expressing a LucR reporter gene enabled us to directly visualize in vivo infection by intravital bioluminescence imaging and observe suppression of acute HIV-1 infection by the activation of NK cells by the IL-15 superagonist. hu-spl-PBMC-NSG mice were infected with HIV-1 by intrasplenic injection with HIV-LucR. One day later, the mice were treated with either the IL-15 superagonist or PBS, and 4 days later, the level of in vivo HIV-1 infection was visualized and quantified by bioluminescence imaging of LucR activity using the IVIS Spectrum intravital imaging system. Direct visualization demonstrated that the level of bioluminescence in the spleens of IL-15 superagonist-treated hu-spl-PBMC-NSG mice was markedly reduced compared to that in untreated mice (Fig. 2B). Quantification of the bioluminescence signal by measurement of total photon flux demonstrated that IL-15 superagonist treatment significantly reduced (P < 0.01) bioluminescence due to HIV-LucR infection (Fig. 2C).

FIG 2.

Treatment of hu-spl-PBMC-NSG mice with the IL-15 superagonist inhibits acute in vivo HIV-1 infection. NSG mice were intrasplenically injected with activated human PBMCs (∼8 × 10⁶ cells) and with HIV-LucR (1 × 10⁷ IU) and 1 day later were either untreated or intravenously treated with the IL-15 superagonist (0.2 mg/kg). (A) Five days after HIV-LucR inoculation, the level of HIV-1 infection in the treated mice was evaluated by measuring LucR activity in splenic lysates. (B) Four days after HIV-LucR inoculation, in vivo infection in the mice was evaluated by intravenously injecting them with a luciferase bioluminescent substrate and then scanning the mice with the IVIS Spectrum imager to measure the bioluminescence intensity in the spleens. Pseudocolor images of the bioluminescence intensity measured as photon counts/second overlaid with grayscale images of a control uninoculated mouse and IL-15 superagonist- or PBS-treated mice are shown. (C) Quantification of the total flux of the bioluminescent signals detected by the IVIS Spectrum imager, reported as photons/second, in IL-15 superagonist- or PBS-treated mouse spleens. **, P ≤ 0.01; ****, P ≤ 0.001 by Student's t test. The experimental protocol is shown at the bottom.

To examine the effect of acute HIV-1 infection and treatment with the IL-15 superagonist on the human T cell and NK cell populations in hu-spl-PBMC-NSG mouse spleens, one group of hu-spl-PBMC-NSG mice was inoculated with HIV-LucR, and 1 day later, some of the HIV-LucR-inoculated mice and HIV-LucR-uninoculated mice were intravenously injected with a dose of the IL-15 superagonist. Six days later, the human CD3⁻ CD56⁺ NK cell and CD3⁺ T cell populations (Fig. 3A) and CD4⁺ and CD8⁺ T cell populations (Fig. 3B) in the mouse spleens were evaluated by flow cytometry. These results demonstrated that 1 week after intrasplenic injection, the hu-spl-PBMC-NSG mouse spleens were well populated with human NK cells, CD4⁺ T cells, and CD8⁺ T cells, which were not markedly altered as a consequence of HIV-LucR inoculation or IL-15 superagonist treatment.

FIG 3.

Effect of in vivo IL-15 superagonist treatment and/or HIV-1 infection on the populations of human NK cells and T cells in hu-spl-PBMC-NSG mouse spleens. NSG mice were intrasplenically injected with activated human PBMCs (∼8 × 10⁶ cells). In parallel, some of these mice were also inoculated by intrasplenic injection with HIV-LucR (1 × 10⁷ IU). One day later, some of the uninoculated and HIV-LucR-inoculated mice were intravenously treated with the IL-15 superagonist (0.2 mg/kg). Six days later, the spleens were harvested, and after gating on the live human CD45⁺ population, the fraction of splenocytes that were NK cells (CD3⁻ CD56⁺ CD16⁺) or T cells (CD3⁺) (A) and CD8⁺ T cells or CD4⁺ T cells (B) was determined by flow cytometry. Mean values ± standard error of the mean for each group of hu-spl-PBMC-NSG mice, which were inoculated with HIV-LucR and/or treated with the IL-15 superagonist, as indicated, are shown.

We next investigated whether the activation of NK cells by the IL-15 superagonist could inhibit the acquisition of HIV-1 infection even when the IL-15 superagonist was administered 3 days after inoculation with HIV-1. hu-spl-PBMC-NSG mice were intrasplenically challenged with HIV-LucR, and 3 days later, some mice were treated with a dose of the IL-15 superagonist. Five days after treatment, the level of HIV-1 infection in hu-spl-PBMC-NSG mouse spleens was quantified and was significantly reduced (P < 0.005) in the IL-15 superagonist-treated mice by almost 95% compared to untreated mice (Fig. 4A). We extended these findings by examining whether the potent suppressive effect mediated by IL-15 superagonist treatment was sustained, by quantifying the level of HIV-1 infection in mouse spleens at 2 weeks and at 3 weeks after treatment. When hu-spl-PBMC-NSG mice were treated 3 days after intrasplenic injection with HIV-LucR with an intravenous injection of the IL-15 superagonist, the development of HIV-1 infection in humanized mouse spleens was significantly inhibited (P < 0.05) by >90% when evaluated at days 14 and 21 posttreatment compared to untreated mice (Fig. 4B). The continued reduction in the level of HIV-1 infection by >97% at 21 days after inoculation suggested that IL-15 superagonist treatment even when initiated 3 days after HIV-1 inoculation effectively eliminated cells initially infected with HIV-1 and thereby potently suppressed the establishment of primary HIV-1 infection. However, when IL-15 superagonist treatment is delayed until 5 days after HIV-1 inoculation, the effect of sustained suppression of primary HIV-1 infection was no longer significant (data not shown).

FIG 4.

IL-15 superagonist treatment of hu-spl-PBMC-NSG mice administered 3 days after inoculation inhibits acute in vivo HIV-1 infection. Three days after NSG mice were intrasplenically injected with activated human PBMCs (∼8 × 10⁶ cells) and HIV-LucR (1 × 10⁷ IU), the mice were either untreated or treated with one intravenous dose of the IL-15 superagonist (0.2 mg/kg). (A) Five days after HIV-LucR inoculation, the levels of HIV-1 infection in IL-15 superagonist-treated mice and untreated mice were evaluated by measuring LucR activity in splenic lysates. (B) HIV-1 infection was quantified by measuring LucR activity in the splenic lysates in another experimental group 14 days after inoculation and in another experimental group 21 days after HIV-LucR inoculation. A dot plot shows the LucR values from each mouse with the mean ± standard error for each group. *, P < 0.05; ***, P < 0.005 by Student's t test. The experimental protocol is shown at the bottom.

NK cells mediate IL-15 superagonist-induced inhibition of in vivo HIV-1 infection.

To confirm that NK cells activated by the IL-15 superagonist were responsible for inhibiting in vivo HIV-1 infection, we depleted by immunomagnetic sorting either the population of CD8⁺ T cells or the population of NK cells from PBMCs prior to intrasplenic injection (Fig. 5A). NSG mice were intrasplenically injected with either activated unfractionated PBMCs, CD8⁺ T cell-depleted PBMCs, or NK cell-depleted PBMCs and in parallel infected with HIV-1 by intrasplenic injection with HIV-LucR. One day later, some groups of mice were either untreated or treated with one intravenous dose of the IL-15 superagonist, and 5 days after treatment, the level of acute HIV-1 infection was quantified by determining the level of expression of the luciferase reporter gene in mouse splenocytes. While the depletion of CD8⁺ T cells from PBMCs had only minimal effects on reducing the IL-15 superagonist-mediated inhibition of HIV-1 infection, NK cell depletion of human PBMCs completely abrogated (P < 0.0001) IL-15 superagonist-mediated inhibition of in vivo HIV-1 infection (Fig. 5B).

FIG 5.

NK cells mediate IL-15 superagonist inhibition of HIV-1 infection in hu-spl-PBMC-NSG mice. Activated PBMCs that were unfractionated or were depleted of CD8⁺ T cells or NK cells by immunomagnetic sorting were intrasplenically injected into NSG mice (∼8 × 10⁶ cells) in parallel with HIV-LucR (1 × 10⁷ IU). One day later, the mice were treated with an intravenous dose of the IL-15 superagonist or PBS. (A) The effectiveness of the depletion of CD8⁺ T cells or CD56⁺ NK cells from the PBMCs prior to injection was assessed by flow cytometry after staining with antibodies to human CD3, CD8, and CD56. (B) Six days after intrasplenic injection, HIV-1 infection in the spleens of mice treated as indicated was quantified by measuring LucR activity in the splenic lysates. The dot plot shows the LucR values for each mouse with the mean ± standard error for each group. ****, P < 0.0001 by Student's t test.

Human NK cells activated by in vivo IL-15 superagonist treatment eliminate HIV-1-infected cells in hu-spl-PBMC-NSG mouse spleens.

To determine whether treatment with the IL-15 superagonist during acute HIV-1 infection induced in vivo activation of human NK cells in hu-spl-PBMC-NSG mouse spleens, we examined its effect on the expression of CD69, a functional marker for NK cell activation that correlates with the expression of the degranulation marker CD107a (43, 44). hu-spl-PBMC-NSG mice were infected with HIV-1 by intrasplenic injection with HIV-LucR. One day later, one group of mice was treated with the IL-15 superagonist. Two days after IL-15 superagonist treatment, flow cytometric analysis demonstrated that CD3⁻/CD56⁺/CD16⁺ human NK cells in the spleens of humanized mice displayed an almost 3-fold increase (P < 0.001) in CD69 expression compared to human NK cells in spleens of untreated hu-spl-PBMC-NSG mice (Fig. 6). We next investigated whether in vivo IL-15 superagonist activation of NK cells increased their capacity to kill HIV-1-infected cells in hu-spl-PBMC-NSG mice with established in vivo infection. hu-spl-PBMC-NSG mice were infected by intrasplenic injection of HIV-LucR. Five days after inoculation, which provided sufficient time for in vivo infection to develop, one group of mice was treated with an intravenous injection of the IL-15 superagonist. One day after IL-15 superagonist treatment, the level of HIV-1 infection in mouse spleens, as determined by measurement of LucR activity, was reduced by >90%, compared to LucR levels in the spleens of untreated mice (Fig. 7). Taken together, these results indicated that in vivo treatment with an IL-15 superagonist activated human NK cells, which increased their in vivo cytolytic activity and enabled them to kill HIV-1-infected cells and suppress acute HIV-1 infection.

FIG 6.

In vivo activation of human NK cells by the IL-15 superagonist in spleens of HIV-1-infected hu-spl-PBMC-NSG mice. NSG mice were intrasplenically injected with human PBMCs (∼8 × 10⁶ cells) and in parallel inoculated with HIV-LucR (1 × 10⁷ IU) and then either untreated or treated with one intravenous dose of the IL-15 superagonist (0.2 mg/kg) 1 day after infection. Two days later, the harvested splenocytes were analyzed by flow cytometry for the expression of human CD3, CD56, CD16, and CD69. After gating on the human CD3⁻/CD56⁺/CD16⁺ NK cell population, the fraction of NK cells expressing CD69 was determined. The data points for each mouse are shown with the mean ± standard error for each group. ***, P < 0.001 by Student's t test. The experimental protocol is shown at the bottom.

FIG 7.

In vivo activation of human NK cells by the IL-15 superagonist stimulates elimination of HIV-1-infected cells in the spleens of HIV-1-infected hu-spl-PBMC-NSG mice. Five days after NSG mice were intrasplenically injected with human PBMCs (∼8 × 10⁶ cells) and in parallel inoculated with HIV-LucR (1 × 10⁷ IU), they were either untreated or treated with one intravenous dose of the IL-15 superagonist (0.2 mg/kg). One day later, HIV-1 infection in the mouse spleens was quantified by measuring LucR activity in the splenic lysates. The dot plot shows the LucR values for each mouse with the mean ± standard error for each group. **, P < 0.01 by Student's t test. The experimental protocol is shown at the bottom.

DISCUSSION

To our knowledge, this study is the first reported demonstration that in vivo activation of NK cells can inhibit acute HIV-1 infection. NK cells were activated by in vivo treatment with a superagonist of IL-15, a multifunctional cytokine which plays a crucial role in NK cell and CD8⁺ T cell development, activation, proliferation, differentiation, and functional activity (12, 15, 45). PBMCs treated in vitro with the IL-15 superagonist displayed increased NK cell activity, as indicated by their increased expression of perforin and granzyme B, as well as cell surface CD107a upon exposure to K562 cells and by their augmented capacity to lyse K562 cells, an NK cell-sensitive target, and activated ACH2 cells, an HIV-1-producing T cell line. Activation of human NK cells by in vivo treatment with the IL-15 superagonist after intrasplenic injection of human PBMCs into NSG mice was indicated by their increased expression of CD69, a marker associated with NK cell activation (43, 44). Acute in vivo HIV-1 infection was potently inhibited by IL-15 superagonist treatment of NSG mice intrasplenically injected with unfractionated human PBMCs or with CD8⁺ T cell-depleted PBMCs. In contrast, when NK cells were depleted from human PBMCs prior to their intrasplenic injection into NSG mice, acute HIV-1 infection was not inhibited by treatment with the IL-15 superagonist. This demonstrated that human NK cells were the crucial effector cells activated by in vivo treatment with the IL-15 superagonist that mediated inhibition of acute in vivo HIV-1 infection. Taken together, our data indicate that in vivo activation of human NK cells by treatment with the IL-15 superagonist, even when delayed until 3 days after HIV-1 inoculation, suppressed the establishment of productive infection. It is likely that this is due to NK cell-mediated elimination of a large fraction of the cells initially infected with HIV-1 during acute infection, which markedly reduced the population of HIV-1-infected cells and the subsequent level of productive HIV-1 infection. This was supported by our demonstration that 1 day after treatment with the IL-15 superagonist, the level of HIV-1 infection in the spleens of hu-spl-PBMC-NSG mice with established HIV-1 infection was reduced by >90%. While treatment with the IL-15 superagonist initiated 5 days after inoculation markedly reduced the number of HIV-1-infected cells measured 1 day later (i.e., measured at day 6 after inoculation, as shown in Fig. 7), sustained suppression of infection was not observed in these mice as it was in mice treated with the IL-15 superagonist within 3 days after inoculation. This decline in sustainable inhibition of HIV-1 infection may be due to the additional 2-day deferral in treatment, which provides HIV-1 with more time to spread sufficiently to a critical number of infected cells that escape NK cell-mediated killing and disseminate infection. Thus, administration of the IL-15 superagonist at day 3 or day 5 after inoculation can potently inhibit HIV infection, but sustained suppression of HIV infection occurs only when the IL-15 superagonist is administered within 3 days of exposure. NK cells contribute to the control of HIV-1 replication in HIV-1-infected individuals, as indicated by the capacity of HLA-B Bw4-801-activated NK cells that express the activating killer immunoglobulin-like NK cell receptor (KIR3DS1) to potently inhibit HIV-1 replication (46) and the delayed progression of HIV-1 disease observed for individuals expressing HLA-B Bw4-801, the ligand for KIR3DL1 receptors, which are highly expressed on NK cells (47). NK cells constitutively express inhibitory and activating receptors capable of identifying and rapidly lysing virus-infected cells (48). Consequently, even during initial exposure to HIV-1 and prior to the development of HIV-specific adaptive immune responses, NK cells can recognize and eliminate HIV-1-infected cells whose HLA class I MHC molecules are downregulated by HIV-1 Nef (48). The capacity of some NK cells to recognize and kill HIV-1-infected cells may be limited by the selective Nef-mediated reduction of only HLA-A and HLA-B molecules, with continued expression of HLA-C and HLA-E molecules (49). It is possible that IL-15 superagonist treatment may activate other NK cells that are capable of effectively killing HIV-infected CD4⁺ T cells despite their continued expression of HLA-C and HLA-E, because this subpopulation of NK cells does not express inhibitory receptors specific for HLA-C and HLA-E molecules (11). Activated NK cells may also identify and kill HIV-1-infected cells by alternative mechanisms, including the detection of stress molecules whose expression is induced by viral infection (50, 51). In addition, activated NK cells may also inhibit the spread of acute HIV-1 infection by secreting CC chemokines, such as RANTES, macrophage inhibitory protein-1-alpha (MIP1-α), and MIP1-β, which block HIV-1 binding to CCR5 (52).

These results extend data from previous studies which first suggested a role for IL-15 in treating HIV-1 infection based on its capacity to potently activate CD8⁺ T cells and NK cells (53), enhance the in vitro function and survival of HIV-specific CD8⁺ T cells (54), delay the recurrence of HIV-1 viremia after antiretroviral therapy interruption (55), and prevent HIV-1 transmission during breastfeeding (56). While IL-15-induced stimulation of NK cells and CD8⁺ T cells may have the beneficial effect of inhibiting HIV-1 replication during acute infection, as shown in our study, during chronic infection, IL-15-mediated activation of CD4⁺ T cells may potentially have the opposite and deleterious effect of increasing their susceptibility to HIV-1 infection. This was indicated in studies of acutely simian immunodeficiency virus (SIV)-infected macaques treated with IL-15 whose memory CD4⁺ T cells displayed increased susceptibility to SIV infection (57) and who subsequently displayed increased viral set points and accelerated progression of disease (58). It is possible that our observation that the IL-15 superagonist potently inhibited HIV-1 infection in hu-spl-PBMC-NSG mice is due to its ∼25-fold-higher biological activity and >35-fold-longer serum half-life than those of soluble IL-15 (20). The greater potency of the IL-15 superagonist than that of IL-15 may enable it to stimulate a more potent anti-HIV-1 immune response than that of the recombinant IL-15 used to treat the macaques and may outweigh the potential deleterious effect of IL-15-mediated activation of CD4⁺ T cells, particularly during acute infection. It is also possible that IL-15, particularly when delivered as a complex with the IL-15Rα chain, has different effects on macaque NK cells and T cells than on human NK cells and T cells.

Multiple studies have demonstrated that immunostimulation of CD8⁺ T cells and NK cells by IL-15 results in the in vivo clearance of tumors and has led to the initiation of phase I trials for their use to treat cancer (59). While studies in mice have indicated that CD8⁺ T cells were the crucial effector cells mediating IL-15-induced clearance of tumors (20, 60), our results demonstrate that activated NK cells were the crucial effector cells activated by the IL-15 superagonist to inhibit acute HIV-1 infection. The functional activity of NK cells is impaired during chronic HIV-1 infection (9), and it is possible that this impairment may be reversed by treatment with the IL-15 superagonist. Further studies in humanized mice infected with HIV-1 and treated with the IL-15 superagonist should enable us to further delineate the effects of NK cell activation mediated by IL-15 superagonist treatment on inhibiting in vivo HIV-1 infection. Treatment of simian-human immunodeficiency virus (SHIV)-infected monkeys with broadly neutralizing HIV-specific antibodies improved the function of their HIV-specific CD8⁺ T cells and lowered their viral load set points, suggesting that neutralizing antibody therapy could increase antiviral immune responses to enable better control of SHIV infection (61). It is possible that NK cell-mediated antibody-dependent cellular cytotoxicity (ADCC) may contribute to the reported reduction in rebound viremia in humanized mice treated with broadly neutralizing HIV-specific antibodies (62). Thus, an intriguing possibility is that increasing NK cell activity by treatment with the IL-15 superagonist may further facilitate the anti-HIV-1 activity of neutralizing antibody therapy by augmenting ADCC activity. Further studies in humanized mice infected with HIV-1 and treated with the IL-15 superagonist should enable us to further delineate the effects of IL-15 superagonist-mediated NK cell activation on directly inhibiting in vivo HIV-1 infection and further suppressing HIV-1 infection through ADCC when combined with broadly neutralizing HIV-specific antibody therapy.