Role of IL-15 Signaling in the Pathogenesis of Simian Immunodeficiency Virus Infection in Rhesus Macaques Free

This article is featured in In This Issue, p.2747

The homeostatic cytokine IL-15 is produced by numerous cell types in diverse tissues, in particular monocytes/macrophages and dendritic cells, and its expression can be upregulated by inflammatory signaling, including ligation of TLR with their ligands or induction of type I IFN (1–4). IL-15 signals through the common γ-chain receptor cytokine family and provides survival and differentiation signals involved with the development and maintenance of NK cell and memory T cell (T_M) homeostasis and function (4–9). Because of its ability to enhance the production, homeostasis, and functional (effector) activity of various lymphoid effector populations, IL-15 is currently under investigation as a therapeutic option in a number of disease conditions, such as viral infections (6, 10), immune reconstitution (11), and cancer (7, 8). IL-15 is upregulated during HIV and SIV infection and its activity may support antiviral immunity by 1) enhancing the activation of innate immune effectors, such as NK cells, early postinfection (pi); 2) promoting the generation, expansion, and maintenance of HIV/SIV-specific CD8⁺ T cell responses (12); and 3) promoting the expansion and differentiation of viral infection-depleted CD4⁺, CCR5⁺ transitional memory T cells (T_TrM) and effector memory T cells (T_EM) from their CCR5⁻ naive T cell (T_N) and/or central memory T cell (T_CM) precursors, providing for reconstitution of the depleted populations (13, 14).

Exogenous IL-15 administration in both nonhuman primates and humans results in selective induction of proliferation among CD4⁺ and CD8⁺ T_M with effector differentiation (T_EM and T_TrM) (15–18). This suggests that IL-15 can stimulate the expansion of Ag-specific effector T cell populations and supports the use of IL-15 as an adjuvant to boost immune response against HIV/SIV. Indeed, administration of the IL-15 superagonist ALT-803 to SIV-infected rhesus macaques (RM) was shown to enhance SIV-specific T cell responses, reduce plasma viral loads (pvl), and even direct SIV-specific T cells into B cell follicles (19, 20). Similarly, there was an increase in SIV-specific CD8⁺ T cells with granzyme B expression and a decrease in viral RNA in lymph nodes (LN) of simian/HIV-infected RM after treatment with heterodimeric IL-15, a stable noncovalent complex of IL-15 and the IL-15R α-chain (21). IL-15 can also induce the proliferative expansion and migration of both CD4⁺ and CD8⁺ T_EM/T_TrM from the periphery into extra lymphoid effector sites in RM (17). Collectively, these studies suggest that modulation of IL-15 signaling may be an effective therapeutic approach to enhance control of virus replication and/or to support immune reconstitution pi.

However, despite the potential beneficial effect of IL-15–mediated expansion of antiviral CD8⁺ effectors and CD4⁺ T_M in facilitating control of HIV/SIV infection and CD4⁺ T cell reconstitution, respectively, several lines of evidence suggest IL-15 activity may also potentiate HIV/SIV disease progression. Plasma isolated during acute HIV infection revealed that the level of certain plasma cytokines predicted 66% of variation in pvl, with IL-15 levels significantly associated with higher pvl set points (22). In untreated HIV-infected patients with pvl above 100,000 copies/ml, IL-15 levels also strongly correlate with increasing HIV-1 viremia (23). IL-15 administration to RM during acute SIV infection was associated with enhanced activation and proliferation of CD4⁺ T_M, higher postpeak virus replication set points, and accelerated disease progression despite higher SIV-specific CD8⁺ T cell responses (24). Although the precise mechanisms by which IL-15 administration appears to enhance rates of virus replication are unclear, IL-15–mediated upregulation of the HIV/SIV receptor CD4 (25), and/or coreceptor CCR5 (26) might increase target cell susceptibility and/or number. It has also been suggested that overexposure of developing virus-specific T cells to IL-15 might compromise the generation and/or long-term stability of effector CD8⁺ T cell responses, which could account for the increased pvl observed in animals treated with IL-15 during acute SIV infection (27). Finally, elevated IL-15 levels during HIV/SIV infection could have an overall negative impact on CD4⁺ T cell homeostasis. Indeed, we have shown that IL-15 supports the homeostatic maintenance of CD4⁺ T_TrM and T_EM in vivo and can induce CD4⁺ T_CM to simultaneously proliferate and differentiate into T_EM in vitro (26). As such, long-term exposure of CD4⁺ T_CM to elevated levels of IL-15 during HIV/SIV infection might disturb the balance between CD4⁺ T_CM self-renewal and CD4⁺ T_CM differentiation into T_EM, promoting population failure by inducing excessive differentiation and ultimately contributing to the collapse of CD4⁺ T_CM homeostasis and overt immunodeficiency (13).

IL-15 signaling is also a primary regulator of NK cell homeostasis and function, and in fact, this cytokine is essential for peripheral maintenance of NK cells (4, 26, 28). Therefore, to the extent that NK cells contribute to the control of HIV/SIV replication, IL-15 would be required for this activity. However, the contribution of NK cells in the control of HIV/SIV replication remains controversial. Recent data suggests that expression of HLA-A may impair control of HIV infection through the inhibition of NKG2A-mediated NK cell licensing (29), suggesting NK cells may play a role in viral control. Previous work evaluating the role of NK cells in the SIV-RM model involved the use of a depleting anti-CD16 mAb administered during primary SIV infection (30). Although NK cell depletion was incomplete, the partial depletion in peripheral blood CD16⁺ NK cells achieved was associated with no demonstrable effect on pvl or kinetics of CD4⁺ T cell depletion (31). In contrast, control of SIV replication in LN B cell follicles by NK cells was associated with high levels of local expression of IL-15 in African green monkeys and distinguishes SIV infection of African green monkeys from pathogenic SIV infection of RM (32), suggesting that NK cell activity plays an important role in maintaining control of viral replication in LN of this natural host species.

In this study, we sought to better define the role of IL-15–dependent immune regulation in acute and chronic SIV infection by specifically inhibiting IL-15 activity in vivo using a rhesusized anti–IL-15 mAb that we have previously shown selectively abrogates IL-15 signaling in RM (26). If IL-15–dependent NK cell and/or CD8⁺ T cell effector responses were essential for SIV control, we would expect anti–IL-15 mAb-treated RM to exhibit enhanced viral replication and rapid disease progress relative to controls. In contrast, if CD4⁺ T_M regeneration was a dominant effect of IL-15, reduced IL-15 might be expected to initially reduce viral loads due to accelerated CD4⁺ target cell depletion but at the same time hasten disease pathogenesis by facilitating collapse of CD4⁺ T_M populations (13). Instead, we found that despite the profound loss of NK cells and significant perturbations in T_EM homeostasis, the magnitude and kinetics of SIV replication, CD4⁺ T cell depletion, and overall SIV disease progression were not significantly different between anti–IL-15–treated RM versus controls. However, prolonged anti–IL-15 mAb treatment during primary SIV infection did accelerate reactivation of RM rhadinovirus (RRV), a simian γ-herpesvirus closely related to human herpesvirus type 8/Kaposi’s sarcoma–associated herpesvirus (33, 34). We also observed cases of non-Hodgkin lymphoma in anti–IL-15–treated RM that were predominantly associated with lymphocryptovirus (LCV), the simian homolog of EBV (35, 36). Collectively, these data suggest that although IL-15 signaling is not a primary regulator of SIV pathogenesis, it may play an important role in maintaining immune control of oncogenic γ-herpesviruses in SIV-infected immunodeficient RM.

A total of 55 purpose-bred male and female RM (Macaca mulatta) of Indian genetic background and free of Macacine alphaherpesvirus 1, D type simian retrovirus, and primate T-lymphotropic virus 1 were used in this study. For chronic SIV infection studies, a subgroup of 18 SIVmac239-infected RM (between 196 and 565 d after SIV infection) was administered with the rhesus recombinant anti–IL-15 mAb, clone M111 (n = 9), or a rhesus recombinant IgG control mAb (n = 9) i.v. once every 2 wk at 20 mg/kg on day 0 and 10 mg/kg on days 14, 28, 42, 56, and 70. A subgroup of 14 SIV naive RM were used as uninfected controls. For primary SIV infection studies, 23 RM were intrarectally inoculated with 3000 50% tissue culture-infective dose of SIVmac239. Of these, a subgroup of seven RM was administered anti–IL-15 mAb i.v. once every 2 wk at 20 mg/kg on day −42 and at 10 mg/kg on days −28 and −14 prior to SIV infection (group A). Another subgroup of 16 RM was administered with anti–IL-15 mAb (n = 8) (group B) or IgG control mAb (n = 8) (group C) i.v. once every 2 wk at 20 mg/kg on day −42 and at 10 mg/kg on days −28 and −14 prior to SIV infection, and 10 mg/kg on days 0, 14, 28, 42, and 56 after SIV challenge. BrdU (Sigma-Aldrich) was prepared and administered i.v. in three separate doses of 30 mg/kg body weight over a 24-h period on day 40–41 pi as previously described (26). All RM were housed at the Oregon National Primate Research Center in accordance with standards of the Center’s Institutional Animal Care and Use Committee and the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Whole blood and mononuclear cells isolated from LN, bronchoalveolar lavage (BAL; or lung airspace), bone marrow, and small intestinal mucosa were obtained and stained for flow cytometric analysis as previously described (13, 17, 27). Polychromatic (8–12-parameter) flow cytometric analysis was performed on an LSR II instrument (BD Biosciences) using Pacific Blue, AmCyan, FITC, PE, PE–Texas Red, PE-Cy7, PerCP-Cy5.5, allophycocyanin, allophycocyanin-Cy7, and Alexa Fluor 700 as the available fluorescent parameters. List mode multiparameter data files were analyzed using FlowJo software (Tree Star). Delineation of T_N and T_M subsets and for setting + versus − markers for CCR5 and Ki-67 expression have been previously described in detail (13, 17, 27). Briefly, for each subset to be quantified, the percentages of the subset within the overall small lymphocyte and/or small T cell (CD3⁺ small lymphocyte) populations were determined. For quantification of peripheral blood subsets, absolute small lymphocyte counts were obtained using an AcT5diff cell counter (Beckman Coulter) and, from these values, absolute counts for the relevant subset were calculated based on the subset percentages within the light scatter–defined small lymphocyte population on the flow cytometer. Baseline values were determined as the average of values at days −14, −7, and 0 relative to first treatment. Results are presented as percentage of baseline, with baseline shown as 100%, fold change from baseline, with baseline shown as 1, or changes in proliferative fraction, indicated as the difference in the %Ki-67⁺ (Δ%Ki-67⁺) measured as the designated time points from baseline (0% = no change).

SIV-specific CD4⁺ and CD8⁺ T cell responses were measured in blood and tissues by flow cytometric intracellular cytokine analysis, as previously described (37, 38). Briefly, mixes of sequential (11 aa overlapping) 15-mer peptides (AnaSpec) spanning the SIVmac239 Gag, Env, Pol, Nef, Rev, Tat, Vif, Vpr, and Vpx proteins were used as Ags in conjunction with anti-CD28 (CD28.2, purified 500 ng/test: Custom Bulk 7014-0289-M050; eBioscience) and anti-CD49d stimulatory mAb (9F10, purified 500 ng/test: Custom Bulk 7014-0499-M050; eBioscience). Cells were incubated at 37°C with peptide mixes and Abs for 1 h, followed by an additional 8 h incubation in the presence of brefeldin A (5 μg ml⁻¹; Sigma-Aldrich). Stimulation in the absence of peptides served as background control. After incubation, stimulated cells were stored at 4°C until staining with combinations of fluorochrome-conjugated mAbs including: anti-CD3 (SP34-2: Pacific Blue; Custom Bulk 624034 and PerCP-Cy5.5; BD Biosciences; Custom Bulk 624060; BD Biosciences), anti-CD4 (L200: FITC; Custom Bulk 624044 and AmCyan; BD Biosciences; Custom Bulk 658025; BD Biosciences), anti-CD8α (SK1: allophycocyanin-Cy7; Custom Bulk 7047-0087-M002; eBioscience), anti–TNF-α (MAB11: allophycocyanin; Custom Bulk 624076 and FITC; BD Biosciences; Custom Bulk 624046; BD Biosciences and PE; Custom Bulk 624049; BD Biosciences), anti–IFN-γ (B27: allophycocyanin; Custom Bulk 624078 and FITC; BD Biosciences; 554700; BD Biosciences) and anti-CD69 (FN50: PE; Custom Bulk CUST01282 and PE-TexasRed; eBioscience; Custom Bulk 624005; BD Biosciences). Data were collected on an LSR II flow cytometer (BD Biosciences). Analysis was performed using FlowJo software (Tree Star). In all analyses, gating on the lymphocyte population was followed by the separation of the CD3⁺ T cell subset and progressive gating on CD4⁺ and CD8⁺ T cell subsets. Ag-responding cells in both CD4⁺ and CD8⁺ T cell populations were determined by their intracellular expression of CD69 and either or both of the cytokines IFN-γ and TNF-α. After subtracting background, the raw response frequencies were memory corrected, as previously described (39, 40). Neutralizing Abs against tissue culture-adapted SIVmac251 were measured in luciferase reporter gene assays using TZM-bl cells as previously described (41).

The following Abs were used in this study: anti-CD3 (SP34-2: Alexa700; Custom Bulk 624042 and Pacific Blue; BD Biosciences; Custom Bulk 624034; BD Biosciences), anti-CD4 (L200: AmCyan; Custom Bulk 658025; BD Biosciences), anti-CD8α (DK25: Pac Blue; PB98401-1; DAKO), anti-CD8α (SK1: PerCP-Cy5.5, AmCyan allophycocyanin-Cy7; Custom Bulk 7047-0087-M002; eBioscience), anti-CD95 (DX2: PE; Custom Bulk CUST00525; Life Technologies), anti-CD28 (CD28.2: PE-Texas Red; Custom Bulk 624005; BD Biosciences), anti-CCR5 (3A9: allophycocyanin; Custom Bulk 624046; BD Biosciences), anti-Ki-67 (B56: FITC; Custom Bulk 624046; BD Biosciences), anti-CD16 (3G8: Pac Blue; Custom Bulk 624034; BD Biosciences), anti-HLA-DR (L243: PE-Texas Red; Custom Bulk 624004; BD Biosciences), anti-CD20 (L27: allophycocyanin-Cy7; Custom Bulk 655118; BD Biosciences), anti-CD56 PerCP-Cy5.5 (MEM-188; Invitrogen), anti-CD20 allophycocyanin-Cy7 (L27; BD Biosciences), anti-NKG2A PE (Z199; Beckman Coulter), anti-CD14 FITC (M5E2; BD Biosciences, R&D Systems), and anti-BrdU FITC, allophycocyanin (B44; BD Biosciences). Anti-CCR7 Biotin (150503) was purchased as purified Ig from R&D Systems, conjugated to biotin using a Pierce Chemical biotinylation kit, and visualized with anti-streptavidin (Pac Blue, S11222; Life Technologies). Rhesus recombinant anti–IL-15, clone M111, and IgG isotype control mAb were provided through the National Institutes of Health’s Nonhuman Primate Reagent Resource Program.

Plasma SIV RNA was assessed using a real-time RT-PCR assay (threshold sensitivity = 30 SIV gag RNA copies/ml of plasma; interassay coefficient of variation ≤ 25%) as previously described (13, 27). RRV DNA was assessed in whole blood essentially as described (42). Specifically, DNA was purified as previously described (43), and ∼100 ng of purified total DNA was subsequently analyzed by quantitative real-time PCR on an ABI StepOnePlus (Applied Biosystems). DNA primers used for the analysis were designed to amplify a segment of RRV ORF3, which encodes the vMIP. The sequence of the TaqMan primers are as follows: vMIP-1, 5′-CCTATGGGCTCCATGAGC-3′; and vMIP-2, 5′-ATCGTCAATCAGGCTGCG-3′. The probe sequence is 5′-TCATCTGCCGCCACCCGGTTTA-3′. Rhesus CMV (RhCMV) DNA in BAL was assessed by quantitative real-time PCR amplification of a segment of the RhCMV UL121 gene, essentially as previously described (44). Briefly, nucleic acid was purified from samples using a Magnapure Compact instrument (Roche Diagnostics) according to the manufacturer’s instructions. Total nucleic acid in the samples was quantified by absorbance at 260 nm using a Nanodrop ND-1000 spectrophotometer. Duplicate 10 μl samples of purified nucleic acid were analyzed using real-time RT-PCR (Primers included forward: 5′-GGGCATCCTCAGGATCACAG-3′, reverse: 5′-CGACACCAAGAGGGTATGGG-3′ and fluorescently labeled probe 6FAM-5′-ACTCCGAAGACCACAAGGACCCACG-3′-TAMRA), and the resulting copies per sample RhCMV UL121 gene were divided by the micrograms of total nucleic acid in the sample. LCV DNA were assessed by real-time RT-PCR. Nucleic acid was purified from whole blood, resuspended in nuclease-free water, and quantified using a Nanodrop ND-1000 spectrophotometer. One hundred nanogram of total DNA were analyzed in duplicate using primers and TaqMan probe specific against the LCV IR1 repeat region and TaqMan PCR conditions as previously described (45). PCR amplicons encompassing the IR1 region were quantified by spectrophotometry and diluted from 10⁶ to 10¹ copies to generate a standard curve for deriving viral copy numbers.

At euthanasia and concurrent with blood sampling, tissues, including tumor tissues, were harvested and fixed with 4% paraformaldehyde, diluting 32% solution (Electron Microscopy Sciences) with PBS, overnight at room temperature. The fixative was replaced with 80% ethanol, and tissues were paraffin embedded. For immunohistochemistry, 4-μm-thick sections were stained. All stained slides were scanned at high magnification (×400) using a whole-slide scanning microscope (Aperio AT2; Leica Biosystems), yielding high-resolution data from the entire tissue section. The scanned slides were analyzed, and figures were produced with HALO Image Analysis Software version 2.3 (Indica Labs). For LCV detection in tissues, immunohistochemistry was performed with the following Abs: CD20 (clone L26, DAKO, used in 1:400), CD3 (clone SP34-2, BD, 1:400), and Epstein–Barr nuclear Ag 2 (EBNA2) (PE2, 1:1600; Abcam) using the Bond RX platform (Leica Biosystems) according to the manufacturer’s protocol. Briefly, tissue sections were baked, deparaffinized, and rehydrated. Epitope retrieval (ER) for CD20 and EBNA2 was performed using Leica ER1 solution (pH 6) whereas ER for CD3 was performed using Leica ER2 solution (pH 9) heated to 100°C for 20 min. Endogenous peroxidase activity was quenched with hydrogen peroxide prior to addition of primary Ab. The Bond Polymer Red Refine Detection Kit (Leica Biosystems) was used to detect chromogen for EBNA2 whereas ImmPress Excel Amplified HRP Polymer Staining Kit Anti-Mouse IgG (Vector Laboratories) were used for CD3 and CD20. For RRV detection in tissues, RNAscope in situ hybridization was performed using the Bond RX platform (Leica Biosystems) and the RNAscope 2.5L reagent kit (Advanced Cell Diagnostics [ACD]) according to the manufacturer’s protocol. Briefly, freshly cut or previously frozen 4-μm-thick sections were stained. Following heat-induced ER (ACD heat-induced ER 15 min with ER2 at 88°C) and proteinase digestion (ACD 15 min Protease), the slides were incubated for 2 h at 40°C with 2.5LS Probe-V-RRV (ACD reference 448028), which targets R6, PAN, K8, ORF25, ORF73, and ORF74 (46). Amplification steps were performed according to the ACD protocol. The chromogen was detected with the Bond Polymer Red Refine Detection kit (Leica Biosystems).

Mixed model, repeated-measures ANOVA was used for longitudinal outcome measures and to compare the T cell and NK cell dynamics outcome changes between treatment groups (anti–IL-15 Ab [three times]/anti–IL-15 Ab [eight times]/IgG control Ab) over the time pi. Because in a typical experiment using repeated measures, two measurements taken at adjacent times are more highly correlated than two measurements taken several timepoints apart, optimal covariance structure chosen by Bayesian information criteria was used to account for within subject correlation. One-way ANOVA was used for the cross-sectional evaluation including T cell and NK cell dynamics in chronic SIV-infected RM prior to IL-15 blockade and at the time of SIV infection. False discovery rate correction method for multiple comparison adjustment was applied. Statistical significance was determined at the significant level of 0.05.

We have previously established that a 20 mg/kg loading dose, followed by biweekly administration of 10 mg/kg of rhesusized anti–IL-15 mAb, was sufficient to establish and maintain essentially complete IL-15 signaling inhibition in vivo in SIV-uninfected RM (26). The most dramatic consequence of this inhibition is profound, sustained depletion of NK cells in blood and tissues. In addition, both CD4⁺ and CD8⁺ T_EM were substantially (∼70%) depleted in blood early after the onset of treatment (with little to no depletion of the T_N, T_CM, or T_TrM populations), but whereas essentially complete NK cell depletion was maintained throughout the IL-15 inhibition period, T_EM depletion was countered by the onset of massive T_TrM and T_EM proliferation, which largely restored circulating T_EM numbers, even with continued IL-15 inhibition. Tissue T_EM, however, remained significantly reduced, and T_EM maintained a very high proliferative rate throughout anti–IL-15 treatment, which appeared to be due to development of increased sensitivity of these cells to alternative γ-chain cytokines, particularly IL-7 (26).

In this study, we randomized 18 RM with asymptomatic, chronic-phase SIVmac239 infection (≥196 d after challenge; Supplemental Table I) to receive either the same anti–IL-15 regimen described above (n = 9) or a matched regimen using an IgG isotype control mAb (n = 9), with anti–IL-15 or control IgG treatment administered every 2 wk for 10 wk (Fig. 1A), resulting in at least 12 wk of potent in vivo IL-15 inhibition. Prior to anti–IL-15 or control IgG treatment, both groups of SIV⁺ monkeys manifested profound depletion of circulating CD4⁺ T_M, with sparing of CD4⁺ T_N and essentially normal numbers of CD8⁺ T cells and NK cells relative to SIV⁻ controls (the latter excepting a modest expansion of the CD16⁻CD56⁻ NK cell and the T_TrM CD8⁺ T cell subsets in the SIV⁺ RM; Fig. 1B–D). As shown in Fig. 1D, depletion of CD4⁺ T_M lineage cells was most marked for effector-differentiated T_TrM and T_EM (which express the CCR5 coreceptor), and all CD4⁺ T_M subsets manifested elevated frequencies of Ki-67 expression, indicating hyperproliferation and consistent with the quasistable high turnover state that characterizes progressive SIV infection (13, 14). CD8⁺ T_CM and T_TrM were also hyperproliferative in the SIV⁺ RM at baseline (Fig. 1C), consistent with persistent generalized immune activation and an ongoing adaptive CD8⁺ T cell response to SIV infection.

As observed in SIV⁻ RM (26), the overall NK cell population was profoundly depleted (mean 95% reduction) in the blood of these chronically SIV-infected RM in the first 14 d of anti–IL-15 treatment, with the dominant CD16⁺ CD56⁻ “cytotoxic” subset essentially eliminated (98% reduced) and the “minor” CD16⁻ CD56⁻ and “regulatory” CD16⁻ CD56⁺ subsets reduced by 93 and 82%, respectively (Fig. 2A). The profound depletion of the CD16⁺ CD56⁻ NK cells was maintained throughout the treatment period, whereas the level of depletion became more variable for the CD16⁻ CD56⁻ and CD16⁻ CD56⁺ subsets after day 42. The response of circulating CD8 lineage T cells in SIV⁺ RM to anti–IL-15 treatment was also essentially identical to that observed in SIV⁻ RM (26), with marked initial depletion of the dominant CD8⁺ T_EM subset (75%), the rapid onset of hyperproliferation within both CD8⁺ T_TrM and T_EM subsets, and gradual reconstitution of circulating CD8⁺ T_EM to baseline levels while still on continuing anti–IL-15 treatment (Fig. 2B). The effect of anti–IL-15 treatment on the already markedly depleted, circulating CD4⁺ T_EM subset in SIV⁺ RM was difficult to resolve, as the observed ∼73% depletion on day 14 after treatment did not reach statistical significance owing to variability in this tiny, residual population (Fig. 2C). However, we did observe significant hyperproliferation of the remaining CD4⁺ T_EM (Fig. 2C), suggesting IL-15 blockade had similar effects as observed in SIV⁻ RM [at the very least, the induction of the same increased sensitivity of these cells to alternative γ-chain cytokines (26)].

The anti–IL-15–associated proliferative burst of effector-differentiated CD4⁺ and CD8⁺ T_M was also observed in BAL (Fig. 2D). Importantly, a total of ∼12 wk of IL-15 inhibition had no effect on the absolute counts in blood of any CD4⁺ T cell subset or on CD4/CD8 T cell ratios in BAL (Fig. 2B–D), suggesting that IL-15 signaling was not required to maintain CD4⁺ T cell homeostasis over the 14 wk of observation. In addition, IL-15 blockade had no effect on pvl during or immediately following treatment (Fig. 2E), indicating that IL-15 signaling to effector-differentiated CD8⁺ T cells was dispensable for maintaining the antiviral effector responses associated with viral load stability (47, 48) and that NK cells have little to no role in determining these viral set points in chronic SIV infection.

The above described data did not reveal a nonredundant role for IL-15 signaling or NK cells in chronic SIV infection, once viral replication set points and CD4⁺ T cell homeostasis processes are established. However, it remained possible that IL-15 signaling or NK cells would play an important role in the initial establishment of the plateau-phase quasistable steady-state. To evaluate this possibility, we sought to determine the impact of IL-15 blockade prior to and during primary SIV infection by implementing a 3-arm study in which two groups of RM received anti–IL-15 at a 20 mg/kg initial dose and 10 mg/kg at 2 wk intervals thereafter, either given only at 6, 4, and 2 wk prior to infection (group A; short duration treatment group) or from 6 wk prior to SIV infection to 8 wk pi (group B; long duration treatment group) (Fig. 3A). A third cohort of RM (group C) was administered an isotype-matched control mAb on the same schedule as the long duration treatment to serve as a control group for both treatment groups. The short duration (treatment-prior-to-infection) arm was designed to assess the effect of pre-establishing the “IL-15–inhibited state”—including NK cell depletion and CD4⁺ and CD8⁺ T_TrM and T_EM hyperproliferation—on viral and lymphocyte dynamics during and after primary infection, while allowing the IL-15 inhibition to wane prior to development of the plateau-phase quasi-steady-state at approximately day 56 pi (see NK cell and T_EM kinetic analysis below). The long duration treatment arm was designed to additionally evaluate the effect of maintaining IL-15 inhibition throughout primary SIV infection—including NK cell depletion, and interference with overall CD4⁺ and CD8⁺ T cell homeostasis and with the adaptive T cell response to SIV infection—on SIV viral replication setpoints, CD4⁺ depletion rates, and overall disease progression.

As expected, the 6 wk of anti–IL-15 treatment of groups A and B RM prior to SIV challenge resulted in the same changes in NK and T_EM dynamics described above (Supplemental Fig. 1), leading to the following differences between the treated (groups A and B) versus control (group C) RM at the time of SIV: 1) profound NK cell depletion in blood and tissues, 2) hyperproliferation of CD4⁺ and CD8⁺ T_EM and CD4⁺ T_TrM in blood, and 3) hyperproliferation of CD4⁺ T_M in tissues with high representation of T_EM and T_TrM (lung airspace and bone marrow) (Fig. 3B–E). Of note, CD4⁺ and CD8⁺ T_EM counts in blood were not significantly different between treated and control groups (Supplemental Fig. 1), as the 6-wk interval between initiation of IL-15 inhibition and challenge was sufficient for the homeostatic regeneration of these populations.

After SIV challenge, cytotoxic (CD16⁺ CD56⁻) and minor (CD16⁻ CD56⁻) NK cells remained profoundly depleted in blood throughout acute primary infection in both anti–IL-15–treated RM groups. Rebound of these subsets to control group levels occurred between day 56 and 91 pi in group A RM and between day 112 and 140 pi in group B RM (Fig. 4A), consistent with full decay of anti–IL-15 activity occurring by 8–10 wk after the last dose. The small regulatory (CD16⁻ CD56⁺) NK cell subset fluctuated after challenge in group A RM but appeared to largely return to control levels by day 35 pi in this group, whereas in the group B RM, depletion of this subset was maintained through approximately day 56 pi, before slowly reconstituting to control levels by day 130 pi.

CD8⁺ T_CM numbers in blood did not appreciably change in any of the three treatment groups after challenge, and all three groups showed a similar postchallenge increase in proliferation (Fig. 4B). In the group C controls, CD8⁺ T_EM counts in blood transiently increased pi (peaking at day 21 pi) associated with a transient burst of proliferation (Fig. 4B). In contrast, CD8⁺ T_EM in group A and B RM were hyperproliferative prior to SIV infection, and this hyperproliferative state did not change pi until waning of the IL-15 blockade (e.g., in concert with NK cell reconstitution), at which time proliferative rates returned to control levels. Group A and B RM did not show any pi increase in CD8⁺ T_EM absolute counts, and indeed, these counts increased with waning of the IL-15 blockade and the associated hyperproliferation. These data indicate that CD8⁺ T_EM response to infection was restricted in the IL-15–deficient environment, perhaps due to a maximally proliferative, high turnover state. CD8⁺ T_TrM manifested an intermediate phenotype between the T_CM and T_EM, showing statistically significant hyperproliferation but only a modest restriction in absolute counts that did not achieve statistical significance.

To directly determine the relative rate of turnover of these CD8⁺ T cell subsets, proliferating cells were BrdU labeled at day 40–41 pi (BrdU given over 24 h), and the %BrdU⁺ cells in each subset were subsequently followed (with BrdU decay reflecting overall cell turnover including cell death and proliferative dilution). As shown in Fig. 4B, decay of BrdU⁺ cells among CD8⁺ T_CM was identical in groups A–C, indicating IL-15 blockade had no effect on the overall turnover of this population. Among CD8⁺ T_TrM, BrdU decay rates appeared slightly faster in the anti–IL-15–treated groups, but this difference did not achieve statistical significance. However, for CD8⁺ T_EM, BrdU decay in the first 2–3 wk postlabeling was significantly faster in both treated groups relative to controls, confirming rapid turnover of this subset in anti–IL-15–treated RM and that this high turnover state persists for ∼10 wk after the last anti–IL-15 dose.

We next evaluated whether IL-15 blockade affected adaptive immune responses to SIV infection and the determination of viral load setpoints. As shown in Fig. 5A, the magnitude of the SIV-specific CD8⁺ T cell response in blood appeared reduced in the anti–IL-15–treated groups, but this reduction did not achieve statistical significance. However, the development of SIV-specific CD8⁺ T cell responses in lung airspace, representing an extralymphoid effector site, was significantly reduced in the anti–IL-15–treated groups during acute primary infection relative to the control group (Fig. 5A). SIV-specific CD4⁺ T cells were similar in magnitude in blood across groups A–C, but again, in lung airspace, these responses were reduced in both anti–IL-15–treated groups (Fig. 5B). CD4⁺ T cell help to B cells did not appear to be affected by anti–IL-15 treatment, as the kinetics and magnitude of the SIVenv-specific Ab response was essentially identical across all the RM groups (Fig. 5C). Taken together, these results suggest that although anti–IL-15 treatment did not block development of SIV-specific CD8⁺ and CD4⁺ T cell responses, it did compromise effector cell production and delivery to extralymphoid sites, in addition to almost complete ablation of NK cell populations. These effects, however, did not change SIV replication dynamics as there were no significant differences in the timing or magnitude of peak or plateau-phase pvl in groups A–C (Fig. 5D).

CD4⁺ T_M depletion is a hallmark of SIV pathogenesis, and the onset of overt immunodeficiency in SIV infection is closely linked to homeostatic collapse of CD4⁺ T_M populations (14). We therefore next asked whether IL-15 blockade affects CD4⁺ T_M depletion dynamics in group A–C RM. As expected, CD4⁺ T_M dynamics in blood after SIV challenge were dominated by massive depletion and postdepletion hyperproliferation involving all subsets (Fig. 6A). CD4⁺ T_CM numbers were not appreciably different between the three treatment groups after challenge, and all three groups showed a similar postchallenge increase in proliferation. In addition, decay of BrdU⁺ cells among CD4⁺ T_CM was identical in groups A–C, confirming that IL-15 blockade has no detectable effect on the turnover of this population, similar to their CD8⁺ T_CM counterparts (Fig. 4A). CD4⁺ T_EM and T_TrM were hyperproliferative prior to infection, and this hyperproliferation decayed to control levels by day 14 pi in both subsets, consistent with the massive depletion of CD4⁺ T cells pi. Following the postdepletion proliferative burst (day 21 pi), proliferation and decay rates of BrdU⁺ cells among CD4⁺ T_TrM were similar between all three treatment groups. In contrast, the CD4⁺ T_EM proliferative burst was delayed in both groups of anti–IL-15–treated RM, before gradually increasing by day 42 pi and remaining elevated for up to 24 wk. CD4⁺ T_EM BrdU decay rates were also slightly faster in anti–IL-15–treated RM relative to controls, although this difference was only significant at later time points. Despite the hyperproliferation observed among CCR5⁺ CD4⁺ T_TrM and T_EM populations in anti–IL-15–treated RM at time of SIV infection (Supplemental Fig. 1), the rate and extent of CD4 depletion was not appreciably different between the three treatment groups (Fig. 6A). Similarly, CD4⁺ T_M dynamics (depletion, proliferation, and turnover) in the extralymphoid effector sites of lung airspace and small intestine were not significantly different between the three treatment groups (Fig. 6B). Collectively, these results suggest that IL-15 signaling activity does not play a major role in determining CD4⁺ T_M population stability following primary SIV infection.

Based on the fact that CD4⁺ T_M homeostasis was not significantly altered in anti–IL-15–treated RM in comparison with controls, it was unsurprising to observe that IL-15 inhibition during primary SIV infection had no discernable effect on SIV disease progression, as time to end stage disease was not appreciably different between the three treatment groups (Fig. 7A). At necropsy, pathological findings revealed a wide range of lymphoproliferative disorders, including a large fraction of anti–IL-15–treated RM with AIDS-defining B cell lymphomas (Supplemental Table II), indicating the possible involvement of oncogenic γ-herpesviruses. Indeed, although most captive-bred RM that are not specifically bred and maintained in specific pathogen-free colonies are latently infected with γ-herpesviruses such as RRV, incidences of lytic replication and oncogenesis are extremely low and mainly manifest in the context of immunodeficiency, such as following SIV infection (43). In this study, we performed a longitudinal assessment of both RRV and LCV DNA in the blood to determine the impact of IL-15 signaling inhibition and the associated loss of NK cells on the control of RRV and LCV replication following primary SIV infection. As shown in Fig. 7B and 7C, there was a significant decrease in time to RRV reactivation in anti–IL-15–treated RM relative to controls, as measured by the detection of RRV DNA in the blood. RRV DNA levels increased as early as 6 wk pi in group B and ∼12 wk pi in group A, in comparison with 24 wk pi in the control group (Fig. 7C). In addition, RRV reactivation was observed in six of eight RM in group B in comparison with just three RM in both groups A and C. Of note, three RM in group A and two RM in group B with RRV viremia rapidly progressed to AIDS <6 mo pi (Fig. 7A). Thus, if RM with rapid disease progression are excluded, then three of four (75%) RM in group A and six of six (100%) RM in group B were shown to have RRV reactivation.

In contrast to the dramatic effects of IL-15 inhibition on RRV replication, the impact of anti–IL-15 treatment on LCV replication was negligible. LCV DNA was detectable in the blood prior to SIV infection, and although levels increased following primary SIV infection, they were not significantly different between the three treatment groups (Fig. 7D). Similarly, anti–IL-15 treatment had minimal effects on RhCMV, a β-herpesvirus. As shown in Fig. 7E, RhCMV DNA in the lung airspace increased transiently in one RM in group A and three RM in group B before returning to baseline levels in most RM by day 154 pi. Collectively, these data suggest that IL-15 signaling activity has differential effects on the immune control of herpesviruses during progressive SIV infection, as a greater proportion of IL-15–inhibited monkeys showed significant increases in RRV but not LCV or RhCMV replication. However, despite the profound effect IL-15 inhibition had on RRV replication, immunohistochemical staining of malignant lymphoma tissues taken at necropsy from RM with confirmed cases of B cell lymphomas revealed high frequencies of EBNA2-positive B cells, indicating predominantly LCV-associated tumorigenesis (Fig. 8A, 8B, Supplemental Table III). RRV was not detectable in any of the malignant tissues, despite the fact that six of the nine RM that had diagnosed B cell lymphoma at necropsy had RRV RNA detectable in the spleen (Fig. 8C, Supplemental Table III).

HIV/SIV pathogenesis is driven by viral replication, but the intensity of this replication and its deleterious effects on the host immune system is thought to be determined by dynamic interplay between 1) activation, destruction and homeostasis (or lack thereof) of CD4⁺ T cells (the major viral target population); 2) inflammation and immune activation; and 3) immune responses, potentially both innate and adaptive, that counter both HIV/SIV infection and infection with the opportunistic pathogens that ultimately result in AIDS (14, 49, 50). The central role of IL-15 in the development, function, and homeostasis of CD4⁺ and CD8⁺ memory/effector T cells and NK cells would suggest that this cytokine orchestrates key elements of these outcome-determining processes and, therefore, that its experimental modulation in SIV-infected RM offers an opportunity both to elucidate the relative important of these processes to SIV infection and disease and to determine whether IL-15–targeted therapeutics would be useful in this setting.

Previous work correlating IL-15 expression levels with SIV infection outcome or administering relatively high dose IL-15 or IL-15 superagonists to SIV-infected monkeys has suggested that IL-15 can, paradoxically, both promote SIV infection and pathogenesis by enhancing viral replication and inflammation and counter SIV infection and pathogenesis by enhancing antiviral immunity (19, 21, 24, 32). Similarly, in humans, IL-15 has been shown to enhance CD4⁺ T cell susceptibility to HIV infection and IL-15 levels have been correlated with increasing HIV viremia and inflammation (23, 51), but at the same time, IL-15 has been shown to enhance T/NK cell function (7, 52–54). Correlation analysis and the study of IL-15 biology via pharmacologic IL-15 administration, although informative, do not provide a complete picture of IL-15 biology, as the former does not distinguish cause from effect, and the latter is complicated by the use of supraphysiologic doses, which both accentuate responses and, after initial stimulation, engage counterregulatory mechanisms that have the potential to dysregulate subsequent responses (19, 27, 55). To circumvent these limitations, we developed a method for long-term inhibition in vivo of IL-15 in RM, using a rhesusized, neutralizing anti–IL-15 mAb, an approach that abrogates function of physiologic IL-15 expression. Using this approach, we previously demonstrated that in SIV-uninfected RM, this IL-15 inhibition profoundly and systemically depletes NK cell populations and disrupts CD4⁺ and CD8⁺ T_M homeostasis by initially (partially) depleting T_EM populations but then eliciting a compensatory homeostatic response (likely mediated by increased sensitivity to other common γ-chain cytokines, in particular IL-7) that stabilized T_EM numbers at the cost of sustained T_TrM and T_EM hyperproliferation (26).

These changes would be expected to substantially, if not completely, abolish any putative viral suppression mediated by NK cells and, possibly, impair viral suppression by CD8⁺ effector T cells, which would potentially increase viral replication rates. Because proliferating CD4⁺ T_M are both highly infectible and highly virus productive (56), the increased frequency of actively proliferating CD4⁺ T_M induced and maintained by IL-15 inhibition might also amplify viral replication. In addition, because CD4⁺ T_M hyperproliferation is required to maintain normal or reduced CD4⁺ T_EM populations in the absence of SIV infection (26), it is likely that the homeostatic response to SIV-mediated depletion would be less effective in IL-15–inhibited than in control RM, possibly accelerating the homeostatic failure of CD4⁺ T_EM that contributes to the onset of AIDS (14, 49). Strikingly, despite the fact that these changes were documented for at least 14 wk in chronically SIV-infected RM subjected to IL-15 inhibition, pvl, CD4⁺ T cell homeostasis in blood, and clinical course were not different from control-treated RM, observations strongly suggesting that in the quasistable steady-state of chronic, asymptomatic SIV infection, IL-15 signaling is dispensable over a 3-mo time frame, and NK cells play a minimal role, if any, in determining plateau-phase viral replication rates.

Immune perturbations during primary SIV infection, prior to establishment of steady-state, tend to have a much bigger impact on SIV pathogenesis than when applied in chronic SIV infection. For example, anti-CD8α mAb-mediated CD8α⁺ cell depletion (which depletes both CD8⁺ T cells and NK cells) in primary SIV infection invariably results in sustained excess plasma viremia and rapid disease progression, whereas in chronic SIV infection, similar treatment typically results in only a transient increase in plasma viremia (27, 48, 57). However, when IL-15–inhibited monkeys with the aforementioned characteristic changes of this treatment were infected with SIV, there was again no difference in either peak or plateau-phase plasma viremia or in both blood and tissue CD4⁺ T_M depletion relative to controls, even when IL-15 inhibition was maintained into plateau-phase. Adaptive CD4⁺ and CD8⁺ T cell responses to SIV were not significantly impaired in blood of IL-15–inhibited RM, although peak, but not plateau-phase, effector frequencies of SIV-specific T cells were significantly, but only partially, reduced in pulmonary airspace, consistent with the described role of IL-15 in promoting memory/effector T cell delivery to extralymphoid tissues (17). However, the essentially identical 1.5 log drop in plasma viremia from peak to plateau-phase in both anti–IL-15–treated and control RM groups strongly suggests that antiviral immunity, to the extent that it is effective in SIV-infected RM, was not impaired with IL-15 inhibition.

This outcome is substantially different from the outcome of nearly universal hyperviremia and rapid progression with CD8⁺ cell depletion in the same timeframe (27). Both anti-CD8α and anti–IL-15 treatments result in profound NK cell depletion and in CD4⁺ T_M hyperproliferation, but whereas virus-specific CD8⁺ T cells are only marginally affected by IL-15 inhibition, this population is substantially (if only transiently) depleted by anti-CD8α treatment. These differences imply that CD8⁺ T cells play the major role in determining acute phase viral dynamics, in particular the drop in postpeak viral loads that protects against rapid progression, with no measurable contribution by either CD4⁺ T_M hyperproliferation or by NK cell–mediated viral suppression. With regard to CD4⁺ T_M hyperproliferation, we have previously demonstrated that CD4⁺ T_M hyperproliferation associated with the anti-CD8α depletion is predominantly IL-15 mediated and that transient anti–IL-15 treatment can block this hyperproliferation without ameliorating the loss of viral control and accelerated disease progression associated with the anti-CD8α treatment (27).

With regard to NK cell–mediated viral suppression, NK cells have been proposed to contribute to antiviral effector activity via direct recognition of infected cells by germline-encoded receptors recognizing MHC-related, stress, or viral ligands or indirect recognition mediated by antiviral Abs binding to the CD16 Fc receptor, resulting in killing of infected cells or viral suppression via elaboration of soluble mediators (58–61). Because many of the NK cell receptor–ligand interactions are highly polymorphic, the best evidence for participation of NK cells in HIV/SIV control is the epidemiologic association of these genetic polymorphisms with viral load set point and/or disease course (29, 62, 63). For example, across multiple HIV⁺ cohorts, the combination of the activating killer cell–Ig (KIR) receptor 3DS1 with HLA Bw4-80I correlated with lower HIV setpoints and improved outcome (63), whereas HLA polymorphisms resulting in higher HLA-E expression (which triggers the inhibitory NKG2A receptor on NK cells) had a deleterious effect on HIV control and disease outcome (29). However, although these correlations are highly significant, the differences in viral loads between groups are quite small, a fraction of a log, far too small to resolve without large population-based cohorts. Thus, if the antiviral activity of NK cells is limited to these effect sizes, our study would not be powered to resolve these differences. In contrast, these effects might be the tip of the iceberg, with NK cells playing a broader role in viral control across genetic polymorphisms, perhaps by Ab-dependent cell-mediated cytotoxicity, or perhaps by editing adaptive responses (59, 60). However, our data contradict this possibility, suggesting that in primary SIV infection of nonvaccinated RM, NK cells have a negligible role in determining viral load dynamics by any mechanism. This result confirms an earlier study which, using CD16 mAb treatment, showed that transient >88% depletion of CD16⁺ NK cells from blood during primary SIVmac251 infection did not significantly alter pvl dynamics (31). Mathematical modeling of population dynamics following SIV infection also agrees with this conclusion, suggesting that NK cells have little impact on viral load or CD4⁺ T cell destruction (64).

We have previously proposed that collapse of a quasistable homeostasis of CD4⁺ T_M populations, in particular the production and delivery of effector-differentiated CD4⁺ T_M in tissues (T_EM and T_TrM), precedes and underlies the development of the overt immunodeficiency in SIV-infected monkeys that results in the onset of AIDS (14, 49). Despite the fact that anti–IL-15 treatment partially depletes T_EM and induces a hyperproliferative state in the residual T_EM and T_TrM populations, we saw no significant difference in CD4⁺ T_M depletion dynamics in monkeys with short- or long-term IL-15 inhibition in primary infection versus controls, nor was there a significant overall difference in time to AIDS between these groups. These data suggest that IL-15 is dispensable for the homeostatic response to CD4⁺ T cell depletion that stabilizes CD4⁺ T_M numbers above the overt immunodeficiency threshold in the chronic (pre-AIDS) phase of infection, with the caveat that longer treatment intervals and/or larger cohort sizes might have revealed effects that were not discernible in this study.

Herpes family virus reactivation accounts for a significant burden of opportunistic infection in RM AIDS, in particular RhCMV, Rhesus LCV, and RRV (46, 65). Analysis of these viruses in our study RM indicated that LCV replication was similar in IL-15–inhibited versus control monkeys, and CMV reactivation, although slightly higher in IL-15–inhibited RM (4/15) versus controls (1/8), was not a dominant infection in either group. However, the onset of RRV reactivation was significantly faster in IL-15–inhibited RM versus controls, and this reactivation was associated with B cell hyperplasia in treated RM. Interestingly, LCV-associated lymphomas developed in 39% of RM in this study, including 50% of RM given the longer-term anti–IL-15 treatment (versus 37.5% of control RM). Although definitive conclusions cannot be made in this small study, it is tempting to speculate that the loss of IL-15–supported effector populations, in particular NK cells [which are known to be especially important in control of herpesviruses (66, 67)], combined with CD4⁺ T_M depletion to accelerate the loss of immune control of RRV, possibly also enhancing LCV tumorigenesis by inducing B cell hyperplasia and expanding lytic LCV infection (46, 68). The possibility that continuation of IL-15 inhibition deeper into chronic phase SIV infection might have further accelerated immunodeficiency will have to be resolved in future studies.

In summary, our results show that inhibition of IL-15–mediated signaling for up to 10 wk in primary SIV infection and 12 wk in chronic SIV infection does not materially change viral or CD4⁺ depletion dynamics in RM. This lack of effect is surprising given the nearly complete NK cell ablation with this treatment, suggesting that NK cells play no role, or at most, a minimal role in SIV control, and that loss of adaptive CD8⁺ T cell responses almost certainly account for the abrogation of viral control and rapid progression associated with CD8α⁺ cell depletion in primary infection. Whether this lack of NK cell contribution to SIV control is related to inefficient recognition of SIV-infected cells via the array of germline-encoded receptors in RM that govern NK cell triggering or to a mismatch between the location, expansion, and migration of NK cells relative to viral infection trajectory (69) remains to be determined, but these results suggest that unless key parameters are different in HIV-infected humans or that NK cell–targeted vaccines can be developed that dramatically change these parameters, CD8⁺ T cells remain the most potent cellular immune mechanism for HIV/SIV control. In addition, our data suggest that SIV replication and memory CD4⁺ T cell depletion dynamics are relatively insensitive to T_M hyperproliferation and loss of a single common γ-chain cytokine, respectively.

We thank the National Institutes of Health's Nonhuman Primate Reagent Resource, from whom we obtained the rhesus recombinant anti–IL-15 and rhesus IgG control mAb, the Quantitative Molecular Diagnostics Core of the AIDS and Cancer Virus Program of the Frederick National Laboratory for viral load analyses, and D. Montefiori for neutralizing Ab assays. The authors also thank J. Turner, S. Planer, T. Swanson, M. Fischer, and A. Legasse for expert animal husbandry. We also thank J. Estes, D. Siess, M. Reyes, J. Clock, A. Kiddle, N. Hamilton, C. Pexton, C. Xu, W. Brantley, A. Maxwell, M. Lidell, M. Marenco, K. Sheffield, S. Hagen, A. Sylwester, H. Park, B. Varco-Merth, A. Townsend, and L. Boshears for technical or administrative assistance.

The authors have no financial conflicts of interest.