IL-15 Treatment during Acute Simian Immunodeficiency Virus (SIV) Infection Increases Viral Set Point and Accelerates Disease Progression despite the Induction of Stronger SIV-Specific CD8⁺ T Cell Responses¹

Increasing function and effectiveness of CD8⁺ T cells is one of the major goals in the fight against viral infections such as HIV infection. Although studies indicated the important role of virus-specific CD8⁺ T cells in HIV and SIV infection (1, 2, 3, 4, 5), these CTLs are unable to control viral replication and progression of HIV infection. In recent years, intrinsic defects were suggested for HIV- (6, 7) and SIV-specific CD8⁺ T cells (8) which result in reduced survival and skewed differentiation of effector memory virus-specific CD8⁺ T cells. Improving the survival and function of these virus-specific CD8⁺ T cells by cytokine therapy may therefore increase viral control. IL-15 is a proinflammatory cytokine produced by a variety of cells including macrophages/monocytes, dendritic cells, and bone marrow stromal cells (9, 10, 11, 12) and binds to the IL-15R, composed of the specific IL-15Rα chain expressed on lymphoid, myeloid, nonhemopoietic cells, including macrophages and monocytes, NK cells, and CD8⁺ T cells (13, 14), the IL-2Rβ chain, and the common γ-chain. Transpresentation of IL-15 has been shown to be a major signaling mechanism (15, 16, 17) although this may not be true for all cell types (18). Several studies have shown the ability of IL-15 to enhance the activation, function, and proliferation of NK cells and CD8⁺ T cells (18, 19, 20, 21, 22, 23, 24, 25, 26), and to function as a chemoattractant for the homing of lymphocytes to peripheral lymph nodes (LN)³ and inflamed tissue either directly (11, 27, 28) or indirectly by inducing IL-8 and MCP-1 by monocytes (29). IL-15 overexpression suppresses progression to murine acquired immunodeficiency syndrome in mice infected with LP-BM5 murine leukemia virus (30) and can act as a plasmid adjuvant (31, 32, 33). In vitro IL-15 enhances survival, inhibits CD95-induced apoptosis, and augments effector function of HIV-specific CD8⁺ T cells (34, 35). Furthermore, IL-15 stimulates expansion of HIV-specific CD8⁺ T cells (36, 37) and restores NK cytotoxicity and deficient IL-12 production by PBMC from HIV-infected individuals (19).

The above effects of IL-15 have suggested its use as an immune augmenter in HIV infection. In previous studies, we have shown that high dose (100 μg/kg) IL-15 treatment of chronically infected cynomolgus macaques results in a 3-fold expansion of CD8⁺ T cells and NK cells. Others have shown that low-dose (10 μg/kg) IL-15 treatment of chronically infected animals may restore the CD4⁺ T cell compartment but only when viral replication is controlled by antiretrovirals (38). These studies led us to investigate the effect of IL-15 treatment during acute SIV infection in Mamu-A*01⁺ rhesus macaques. Our data show that IL-15 treatment increases viral set point by 3 logs and leads to accelerated disease progression to simian AIDS in two of six animals with one developing SIV meningoencephalitis. IL-15 induced 2- to 3-fold increases in SIV-specific CD8⁺ T cell and NK cells during the acute infection without effect on viremia. Although increased viral set point was accompanied by reduced blood activated SIV-specific CD8⁺ T cells and NK cells, LN numbers of SIV-specific CD8⁺ T cells were increased whereas LN NK cell numbers were not affected. Expression of Ki-67 in CD4⁺ T cells at day 7 postinfection was positively correlated with viral load at day 84. These findings suggest that IL-15 may be altering the course of the disease by enhancing the proliferation of CD4⁺ T cells early during infection and promoting their migration into tissue hence increasing the target pool of SIV infection. Thus, the level of CD4⁺ T cell activation during acute infection may determine the viral set point months later. Finally, our findings also indicate that IL-15 treatment of acutely SIV-infected nonhuman primates may prove useful as an animal model to elucidate what controls viral set point.

The 12 Mamu-A*01⁺ male Indian rhesus macaques (Macaca mulatta) included in this study were housed at the Bioqual animal facility according to standards and guidelines as set forth in Animal Welfare Act and The Guide for the Care and Use of Laboratory Animals and according to animal care standards deemed acceptable by the Association for the Assessment and Accreditation of Laboratory Animal Care International. All experiments were performed following institutional animal care and use committee approval. The animals were inoculated i.v. with 100 ID₅₀ SIVmac251 and either treated with 100 μg/kg bodyweight rhesus macaque IL-15 or with saline (control) s.c. twice weekly for 4 wk. SIV RNA levels were measured using a quantitative TaqMan RNA RT-PCR assay (sensitivity of 200 SIV RNA copies/ml blood; Applied Biosystems) (39). Recombinant rhesus macaque IL-15 was provided through the National Center for Research Resources (NCRR) funded Resource for Nonhuman Primate Immune Reagents (40).

Blood of animals was processed within 6 h after blood draw. Freshly density centrifugation isolated PBMC (Percoll, 1.075 g/ml; Amersham Biosciences) were stained ex vivo or used for apoptosis, cytotoxicity, and cytokine production assays. Mamu-A*01-β2-microglobulin tetramers were loaded with SIV Gag 181–189 CM9 (CTPYDINQM) or SIV Tat SL8 28–35 (STPESANL) peptide (41). The following rhesus macaque cross-reacting anti-human Abs were used: anti-CD3 (clone SP34), -CD4 (L200), -CD8 (3B4), -CD20 (2H7), -CD25 (M-A251), -CD69 (FN50), -CD45RA (5H9), -CD62L (SK11), -HLA-DR (L243/G46-6), -Ki-67 (B56), -IL-2 (MQ1–17H12), -IFN-γ (4S.B3), -TNF-α (Mab11), -IL-10 (JES3-9D7), -CD195 (CCR5, 3A9), and annexin V (BD Biosciences; eBioscience; Invitrogen Life Technologies; Southern Biotechnology Associates). PBMC were stained with up to 10 different Abs in HBSS (Cellgro), 3% heat-inactivated horse serum (Invitrogen Life Technologies), 0.02% NaN₃ for 30 min on ice, washed two times, and fixed with 1% paraformaldehyde. For intracellular staining, PBMC were first surface stained, fixed, and permeabilized (Cytofix-Cytoperm; BD Biosciences), incubated with the Abs and isotype controls and fixed with 1% paraformaldehyde. Lymphocytes were isolated from inguinal LN by incubation of minced LN tissue in digestion medium (0.15 mg/ml DNase (Roche Applied Science), 3 mg/ml collagenase A (Roche Applied Science)) for 2 h at 37°C followed by red blood lysis (150 mM NH₄Cl, 10 mM KHCO₃, 0.1 mM Na₂EDTA (pH 7.2)) for 5 min at room temperature. Samples were collected on a FACSCalibur and FACSAria (BD Biosciences) and analyzed using FlowJo software (Tree Star).

Freshly isolated PBMC were stimulated for 6 h with SIV Gag 181–189 CM9 (CTPYDINQM) peptide or SIV Tat SL8 28–35 (STPESANL) peptide at a concentration of 10 μg/ml in the presence of 1 μg/ml anti-CD28 and anti-CD49e Ab (eBioscience) with brefeldin A (10 μg/ml; BD Biosciences) added for the last 4.5 h. Surface and intracellular staining was performed directly afterward. Background (cells cultivated without peptide) was subtracted from the peptide-stimulated samples and absolute cell number of cytokine-producing cells was calculated.

For spontaneous apoptosis, freshly isolated PBMC from SIV-infected untreated and IL-15-treated animals were cultured in RPMI 1640/10% heat-inactivated FBS/2 mM l-glutamine/100 U/ml penicillin/100 μg/ml streptomycin sulfate (Cellgro). For CD95-induced apoptosis, PBMC were incubated in the presence of 10 ng/ml soluble FasL (Axxora) and 1 μg/ml enhancer (Axxora). Cells were harvested after 14-h incubation and stained for flow cytometry analysis.

To measure cytotoxicity of NK cells, freshly isolated PBMC were incubated at E:T ratios as indicated with 5 × 10⁴ Na₂⁵¹CrO₄ (PerkinElmer) radiolabeled K562 target cells in 96-well U-bottom plates for 5 h at 37°C. Thirty microliters of supernatants were transferred to 96-well LumaPlates (Packard Bioscience) and counted in a TopCount microplate scintillation counter (Packard Bioscience). Specific cytotoxicity was determined using the formula: ((cpm experimental release − cpm spontaneous release)/(cpm maximal release − cpm spontaneous release)) × 100.

Total SIV-specific Abs (IgM and IgG) were measured in plasma of untreated and IL-15-treated acutely SIV-infected rhesus macaques using the Genetic Systems HIV-2 EIA (Bio-Rad) according to the manufacturer’s instructions.

For each animal, detailed gross and microscopic neuropathological evaluation was performed. Brain specimens were fixed in formalin. Multiple representative sections from the cerebral neocortex and corresponding portions of hemispheric white matter of the frontal, parietal temporal and occipital lobes, hippocampus, basal ganglia, diencephalic structures (thalamus, hypothalamus, subthalamus), brain stem, and cerebellum were analyzed (20 paraffin/animal). Histological sections were stained with H&E and Luxol fast blue/cresyl violet and evaluated. The extent and distribution of inflammatory lesions was analyzed and rated after Boche et al. (42) as follows: (±) rare, scant/sparse; (+) frequent, scant/sparse; (++) frequent and marked.

Tissues (rectum pinch biopsy samples, inguinal LN) were fixed with formalin, embedded in paraffin, and were assayed for SIV viral RNA expression by in situ hybridization as previously described (43). Briefly, the tissue sections were hybridized overnight at 50°C with either sense or antisense SIVmac239 digoxigenin-UTP labeled riboprobe, blocked with 3% normal sheep and horse serum in 0.1 M Tris (pH 7.4), and incubated with sheep anti-digoxigenin-alkaline phosphatase (Roche Molecular Biochemicals) and NBT/5-bromo-4-chloro-3-indolyl phosphate (Vector Laboratories).

Statistical analysis was performed using a two-way ANOVA when data were normally distributed or could be transformed to attain that distribution. Otherwise, a series of nonparametric tests (Mann-Whitney U) between the groups at each time was used.

To examine whether IL-15 treatment induced clinical side effects, animals were tested weekly for several parameters. Glucose, blood urea nitrogen, creatinine, total protein, albumin, globulin, total bilirubin, alkaline phosphatase, sodium, potassium, triglyceride lipid, and amylase concentration showed no differences when untreated and IL-15-treated animals were compared. The concentration of phosphorus was significantly reduced in the blood with IL-15 treatment at weeks 4 and 6, with no differences at later time points (data not shown), whereas the concentration of calcium was comparable between untreated and IL-15 treatment. We also observed a transient decrease in the amount of alanine aminotransferase at weeks 3 and 4 in the IL-15-treated animals. Additionally, hematological measurements (white blood cell numbers, RBC numbers, platelets, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, platelet numbers) were also performed weekly, with no differences observed between untreated and IL-15-treated animals. IL-15 treatment was therefore well-tolerated by the animals and did not lead to toxic side effects, which is in agreement with the results of our previous study in which we treated chronically SIV-infected cynomolgus macaques (40).

Viral load was followed in all 12 SIV-infected animals up to week 20. IL-15-treated animals had a slightly increased plasma viral load initially with 5.7 ± 3.6 × 10⁶ SIV RNA molecules/ml at day 7 and 20 ± 5.9 × 10⁶ SIV RNA molecules/ml at day 14 compared with untreated animals (2.6 ± 1.4 × 10⁶ SIV RNA molecules/ml at day 7 and 10 ± 1.5 × 10⁶ SIV RNA molecules/ml at day 14) (Fig. 1). After week 3, viral load in the untreated animals slowly decreased whereas in IL-15-treated animals, viral load remained high. This resulted in a 1 log higher viral load in IL-15-treated animals at week 6 (4.4 ± 3.1 × 10⁶ SIV RNA molecules/ml for IL-15-treated animals and 0.18 ± 3.1 × 10⁶ SIV RNA molecules/ml for untreated animals, p = 0.042) and a 3 log higher viral load at week 20 (42 ± 28 × 10⁶ SIV RNA molecules/ml for IL-15-treated animals and 2.4 ± 0.9 × 10⁴ SIV RNA molecules/ml for untreated animals, p = 0.017) (Fig. 1).

IL-15-treated animals at week 2 postinfection had a 2.3-fold increase in Gag- and a 3-fold increase in Tat-specific CD8⁺ T cells (p = 0.044) compared with untreated animals (Fig. 2,A). However, at all later time points virus-specific CD8⁺ T cell numbers were comparable between IL-15-treated and untreated animals. In IL-15-treated animals, we observed 1.6-fold more CD8⁺ T cells at week 2 (Fig. 2,B). After cessation of IL-15, the CD8⁺ T cell numbers were comparable between IL-15-treated and untreated animals throughout the study. No differences were found in the absolute CD4⁺ T cell and CD20⁺ B cell numbers (Fig. 2 B).

To determine whether IL-15 treatment altered activation of cells, we analyzed expression of the activation markers CD69, CD25, and HLA-DR. IL-15 treatment reduced the frequency of HLA-DR⁺ Gag-specific CD8⁺ T cells in the peripheral blood compared with untreated animals beginning at week 6 (untreated animals: 23 ± 3.4%, IL-15-treated animals: 8.2 ± 1.3%; p < 0.05) (Fig. 3,A). A similar decrease was observed for total CD8⁺ T cells (Fig. 3 A). IL-15 treatment did not alter the expression of CD69, CD25, and HLA-DR on CD4⁺ T cells and CD20⁺ B cells (data not shown).

An increase in cell numbers due to IL-15 treatment was also found for CD3⁻CD8⁺ NK cells with a 2.3-fold increase at week 1 (p = 0.049; Fig. 2,B). However, although IL-15 was given until week 4, the absolute NK cell numbers were not different for IL-15-treated and untreated animals from weeks 2 to 4. After week 4, IL-15-treated animals had reduced NK cell numbers compared with untreated animals with up to significantly 67% reduction at week 16 (p = 0.02; Fig. 2,B). However, as mentioned below, week 30 LN from IL-15-treated animals showed no decrease in CD3⁻CD8⁺ NK cells suggesting that CD3⁻CD8⁺ NK cells are specifically reduced from circulation (Fig. 3 D).

We next analyzed the expression of CCR5 which is up-regulated by IL-15 (44). Although during the first 4 wk, no differences were found, at later time points the percentage of CCR5⁺CD4⁺ T cells in the IL-15-treated animals decreased to <1% whereas in the untreated animals >5% CCR5⁺CD4⁺ T cells were found between weeks 6 and 12 (p = 0.049; Fig. 3 B). CCR5 expression on Gag-specific CD8⁺ T cells, total CD8⁺ T cells, NK cells, and monocytes was comparable between untreated and IL-15-treated animals (data not shown).

Using CD45RA and CD62L, we also analyzed memory cell distribution of CD4⁺ T cells which did not differ when untreated and treated animals were compared during the initial IL-15 treatment period. Thereafter, however, a 50% reduction of CD45RA⁻CD62L⁻ CD4⁺ T cells were found in IL-15-treated animals (week 6, 12 ± 2% for treated and 23 ± 3% for untreated animals (p = 0.016); week 20: 10 ± 1% for treated and 25 ± 6% for untreated animals (p = 0.049)). In contrast, CD45RA⁻CD62L⁺ cells were increased in IL-15-treated animals (week 6: 63 ± 8%, week 20 56 ± 4%) compared with untreated animals (week 6: 51 ± 9%, week 20: 38 ± 3%) (Fig. 3 C).

Memory cell distribution and spontaneous and CD95-induced apoptosis was also examined in SIV-specific CD8⁺ T cells as reduced survival and skewed memory phenotype was described as intrinsic defects for HIV-specific CD8⁺ T cells. No difference was found in the distribution of memory subpopulations of Gag- and Tat-specific CD8⁺ T cells and total CD8⁺ T cells when IL-15-treated and untreated animals were compared (data not shown). Spontaneous and CD95-induced apoptosis was comparable in SIV-specific CD8⁺ T cells, total CD8⁺ T cells, and CD4⁺ T cells when PBMC from untreated and IL-15-treated animals were compared (data not shown).

Biopsies of inguinal LN were collected 30 wk after SIV infection and IL-15 treatment. Viral loads were 25,387 ± 17,994 SIV RNA molecules/ml blood in the IL-15-treated animals (n = 3) and 8,440 ± 4,210 SIV RNA molecules/ml blood in the untreated animals (n = 3). The LN of IL-15-treated animals were significantly smaller (169 ± 49 mg) than the ones from untreated animals (367 ± 51 mg; p < 0.05). Accordingly, total cell, lymphocyte and CD4⁺ T cell numbers per LN were significantly reduced in IL-15-treated animals. Despite this overall reduction, the percentage (data not shown) and absolute cell numbers of Gag-specific CD8⁺ T cells were increased in the IL-15-treated animals compared with untreated animals (p = 0.049; Fig. 3,D). No significant differences in total CD8⁺ T cell, CD3⁻CD8⁺ NK cell, or CD20⁺ B cell numbers were observed between LN from untreated and IL-15-treated animals (Fig. 3 D).

We next examined whether viral load at days 14 (peak) and 84 (viral set point) correlated with activation and proliferation of CD4⁺ T cells and Gag-specific CD8⁺ T cells at days 7, 14, and 84.

We observed a significant increase in the percentage of Ki-67⁺ CD4⁺ T cells at day 7 postinfection in the IL-15-treated animals compared with the untreated animals (Fig. 4,A). When the correlation was examined between the percentage of Ki-67⁺CD4⁺ T cells on day 7 postinfection and viral load, a positive correlation was observed for day 14 viral load (R² = 0.38, p = 0.03) and day 84 viral load (R² = 0.78, p = 0.002) (Fig. 4 B). However, we did not see any significant correlation when percentage of days 14 and 84 Ki-67⁺CD4⁺ T cells were analyzed, which is in agreement with our observations that there was no difference between the percentage of Ki-67⁺CD4⁺ T cells in untreated and IL-15-treated animals at these later time points. A positive correlation was also seen when the absolute cell numbers of Ki-67⁺CD4⁺ T cells at day 7 and viral replication at day 14 (R²: 0.72, p = 0.03) and day 84 (R²: 0.68, p = 0.04) were compared. No correlation was found between percentage and absolute cell numbers of CD25⁺CD4⁺ T cells and viral load. Whereas no correlation was found between percentage of HLA-DR⁺CD4⁺ T cells and viral load, absolute HLA-DR⁺CD4⁺ T cells on day 7 correlated positively with viral load at day 14 (R²: 0.82, p = 0.01) and day 84 (R²: 0.73, p = 0.03). When we examined the correlation between absolute SIV-specific CD8⁺ T cell numbers and viral replication, we found no significant correlation. The same was true for absolute cell numbers of CD25⁺, HLA-DR⁺, and Ki-67⁺ SIV-specific CD8⁺ T cells and viral load. However, the percentage of day 84 CD25⁺ and Ki-67⁺ SIV-specific CD8⁺ T cells correlated positively with viral load (data not shown), indicating that increased viral replication maintains activation of virus-specific CD8⁺ T cells.

Next, we examined whether the increased cell numbers were due to proliferation of these cells by analyzing the expression of Ki-67. No difference in Ki-67 expression of SIV-specific CD8⁺ T cells was observed in IL-15-treated animals (Fig. 4,A). IL-15-treated animals had a significant increase in frequency of Ki-67⁺ total CD8⁺ T cells (59 ± 5.3%) compared with the untreated group (13 ± 2.1%) at week 1 (p = 0.004), but lower Ki-67⁺ CD8⁺ T cells at weeks 6 and 8 (Fig. 4,A). Similarly, the IL-15-treated group had significantly more Ki-67⁺ NK cells at week 1 (88 ± 2%) compared with the untreated group (26 ± 6%; p = 0.004) and significantly less Ki-67⁺ NK cells in IL-15-treated animals at weeks 6 and 8 (Fig. 4,A). Although absolute CD4⁺ T cell numbers did not differ between IL-15-treated and untreated animals (Fig. 2,B), the IL-15-treated group had a 2-fold increase of Ki-67⁺ CD4⁺ T cells at week 1 (21 ± 2.5%) compared with the untreated animals (10 ± 1.4%; p = 0.015; Fig. 4 A). At weeks 6 and 8, increased Ki-67⁺ CD4⁺ T cells were observed in the untreated but not in the IL-15-treated animals.

The central question of this study was whether IL-15 could enhance the innate and adaptive immune response against SIV infection. We therefore examined cytotoxicity of NK cells and peptide-mediated cytokine secretion of CD8⁺ T cells in untreated and IL-15-treated animals. IL-15 treatment increased the absolute numbers of TNF-α and IFN-γ-producing CD8⁺ T cells stimulated with the Gag peptide at week 2 (Fig. 5,A). However, because the absolute Gag-specific CD8⁺ T cell number was higher in the IL-15-treated animals, this did not result in an increase in the percentage Gag-peptide responding cells through IL-15 treatment. The same was seen when CD8⁺ T cells were stimulated with the Tat peptide although only an increase in IFN-γ-producing CD8⁺ T cells but not in TNF-α-producing CD8⁺ T cells was observed at week 2 (Fig. 5,A and data not shown). No difference was found in IL-2 and IL-10 production by SIV-specific CD8⁺ T cells between untreated and IL-15-treated animals (data not shown). Finally, NK cell cytotoxicity did not differ between three IL-15-treated and three untreated animals analyzed at week 0 preinfection and weeks 1–4 and 6 postinfection (Fig. 5 B).

Inguinal LN and rectal biopsies were collected from three IL-15-treated and three untreated animals and analyzed for SIV infection by in situ hybridization. No statistical difference in the number of SIV-infected cells was found in rectal biopsies when untreated and IL-15-treated animals were compared (week 1: 1.51 ± 1.39 cells/mm² and 1.44 ± 2.04 cells/mm²; week 8: 0.09 ± 0.16 cells/mm² and 0.08 ± 0.13 cells/mm² for untreated and IL-15-treated animals, respectively). In LN biopsies significantly less SIV-infected cells were found at week 2 (peak viremia) in the IL-15-treated (1.6 ± 0.8 SIV-infected cells/mm² tissue) compared with the untreated animals (7.8 ± 1.5 SIV-infected cells/mm² tissue; p = 0.021; Fig. 5 C). Although no statistical differences were observed at later time points it appeared that IL-15 treatment increased the number of SIV-infected cells in LN. When infected macrophages and infected lymphocytes were estimated based on morphology, no difference in numbers of SIV-infected macrophages and lymphocytes were found between untreated and IL-15-treated groups.

To examine whether IL-15 treatment altered B cell function, we analyzed anti-SIV Abs in the plasma of untreated or IL-15-treated animals at week 3 (early infection) and at week 16 (viral set point). Significant less anti-SIV Abs were found at both time points in the IL-15-treated animal group compared with the untreated one (week 3: 0.05 ± 0.03 arbitrary units (AU) vs 0.22 ± 0.06 AU, p = 0.046 and week 16: 3.3 ± 2 AU vs 16 ± 5.2 AU, p = 0.025, for IL-15-treated and untreated animals, respectively) (Fig. 5,D). The decrease in Ab production was, however, not accompanied by a change in B cell numbers either in peripheral blood (Fig. 2,B) or in LN (Fig. 3 D). To further examine whether the reduced anti-SIV Ab concentration could result in increased viral load, correlations between Ab and viral replication were analyzed at weeks 3 and 16. No correlation was observed for week 3 Ab concentration and viral replication, however, an inverse correlation was found for week 16 Ab concentration and viral replication (R²: 0.34, p = 0.045). If IL-15-induced changes in Ab production are associated with CD4 activation/proliferation, a correlation between Ki-67⁺CD4⁺ T cells and amount of Ab should be observed. Indeed, we found an inverse correlation between the percentage of Ki-67⁺CD4⁺ T cells at week 1 and the amount of Abs at week 3 (R²: 0.68, p = 0.001) and 16 (R²: 0.45, p = 0.017).

The increase in viral load indicated already that IL-15 treatment during acute infection may increase the normally slow progression of SIV infection in Mamu-A*01⁺ rhesus macaques (45, 46). We therefore followed the animals up to week 30. Two of the 6 IL-15-treated animals had to be euthanized at weeks 25 and 26, respectively, due to simian AIDS with a plasma viral load of 103 × 10⁶ SIV RNA molecules/ml and 53 × 10⁶ SIV RNA molecules/ml, respectively. The other four IL-15-treated animals are alive at week 30 with plasma viral load of 53 ± 51 × 10⁴ SIV RNA molecules/ml whereas the six untreated animals had a viral load of 7.4 ± 6.6 × 10⁴ SIV RNA molecules/ml blood.

One of the two animals euthanized due to simian AIDS showed clinical signs of CNS involvement such as signs of visual impairment. We therefore examined the brain to analyze whether IL-15 treatment not only accelerates progression to AIDS but also leads to early involvement of the brain. Gross examination of the brain of the two euthanized animals was essentially unremarkable. One of two animals showed a mild, focal meningoencephalitis characterized by rare, discrete, randomly distributed and scant inflammatory infiltrates—consisting predominantly of monocytes/macrophages, to a lesser extent lymphocytes—involving the cerebellar and cerebral leptomeninges (Fig. 6, A–C). Neither substantial meningitis nor leptomeningeal thickening were observed. Scant perivascular mononuclear inflammatory cell infiltrates were detected only in the cortex (data not shown). In addition, patchy microglial nodules were detected focally in the superficial/subpial cortical gray matter together with mild to moderate subpial gliosis (Table I, Fig. 6 D). No perivascular syncytial/multinucleated giant cells, microglial nodules, endothelial hypertrophy were identified. The cerebral white matter was devoid of myelin pallor and/or vacuolar changes. Collectively, these changes are consistent with early, minimal CNS SIV-associated changes in the context of “early encephalitis.”

Several in vitro and in vivo studies in recent years have indicated that cytokine therapy may be a promising adjunctive therapeutical treatment for HIV infection. Candidate cytokines that have been tested in animals are IL-7 (47, 48), IL-12 (49), IL-2 (50, 51), and IL-15 (38, 40, 51). The effect of such cytokine treatment on acute infection, however, is unexplored. By providing a cytokine adjuvant during acute infection, one can envision changing the whole dynamics of host virus interaction leading to either better control or accelerating disease and thus allowing us to understand better the mechanisms that control viral loads. One of the most promising cytokine is IL-15, which was shown in in vitro studies to increase survival and effector function of HIV- (34) and SIV-specific CD8⁺ T cells (8). This study was therefore performed to examine whether in vivo IL-15 treatment during acute SIV infection of rhesus macaques could enhance the immune control of viral replication. However, we found that IL-15 treatment during acute infection increased the viral set point in animals and accelerated disease progression in some animals.

The increased viral set point of animals treated with IL-15 could be due to a direct effect of IL-15 on viral replication. In vitro IL-15 has either a marginal effect (19, 52, 53) or no effect (54, 55) on replication of HIV in naturally infected PBMC. However, in PHA-stimulated freshly HIV-infected PBMC IL-15 can induce various amounts of HIV replication that differ from study to study (44, 53, 54, 55, 56). This suggests that the effect of IL-15 on viral replication may depend on the activation status of the cells. In agreement with this, we have observed an increase in viral load in our present study treating acutely SIV-infected monkeys when intense immune activation is occurring. In contrast, earlier studies showed that IL-15 treatment of chronically SIV-infected animals did not increase viral set point (38, 40). An important difference in acute infection may be the presence of high amounts of IL-2, as IL-2 and IL-15 synergize to increase viral replication (54). Although no IL-2 mRNA was found in peripheral blood during acute HIV infection, LN biopsies have high amounts of IL-2 mRNA (57). Thus, during acute infection endogenous IL-2 and exogenous IL-15 could synergize leading to increased viral replication. Although this could be happening in our current study of IL-15 treatment, the fact that no difference in peak viral loads was seen when animals were still being treated with IL-15 would argue against this as a mechanism of increased viral set point. Most importantly, viral set points were higher at time points long after IL-15 treatment had ceased, indicating that some other mechanism is at action.

Increased viral set points in IL-15-treated animals could be due to increased CD4⁺ T cell targets for infection. IL-15 can directly affect CD4⁺ T cells and macrophages, therefore increasing the target cell pool for SIV infection. The first studies describing the effect of IL-15 and IL-15Rα knockout mice concluded that CD4⁺ T cells are not affected in these mice (27, 58). This is in accordance with our study of IL-15-treated chronically SIV-infected cynomolgus macaques where we have not observed any changes in activation, cell numbers, and memory phenotype of blood CD4⁺ T (40). However, we and others have observed that under certain circumstances, IL-15 can indeed act as costimulator albeit weak for memory CD4⁺ T cells (38, 59, 60, 61). Furthermore, mitogen-activated human CD4⁺ T cells express the IL-15Rα chain (62) and IL-15-stimulated effector memory CD4⁺ T cells can migrate into tissue (38), suggesting that IL-15-responsive CD4⁺ T cells may migrate to sites of viral infection and replication. In the present study, we did observe increased Ki-67 expression in CD4⁺ T cells at week 1 in IL-15-treated animals, however, this was not accompanied by increased peripheral blood CD4⁺ T cell numbers. This may indicate CD4⁺ T cells migration into tissue. We have not observed increased infected lymphocytes in inguinal LNs or rectal biopsies in the acute infection. It is possible that CD4⁺ T cells migrated into GALT, which was shown before to be one of the major site of viral replication (63). In week 30 LN, the absolute CD4⁺ T cell number is decreased, suggesting that CD4⁺ T cells in our animal model may be migrating preferentially into GALT and not LN. We found that CD45RA⁻CD62L⁻ CD4⁺ T cells are diminished in the blood of IL-15-treated animals which suggests that these cells are either migrating into GALT or they are being consumed at sites of viral replication, because memory CD4⁺ T cells are the prime target for SIV infection. IL-15 treatment also reduced blood CCR5⁺ CD4⁺ T cells, something previously reported in chronic infection (63). Because IL-15 up-regulates CCR5 on human CD4⁺ T cells (44, 64), the loss of blood CCR5⁺ CD4⁺ T cells suggests migration of these cells from peripheral blood into tissue and increased likelihood of infection with the CCR5-tropic SIVmac251 virus. This, however, remains to be determined. It is unlikely that monocytes, another target cell population for SIV infection, is the source of increased viral replication as the numbers and CCR5 expression of monocytes were not affected in IL-15-treated animals. That indeed activation and proliferation of CD4⁺ T cells early during acute SIV infection can alter the viral replication months later was further supported by the positive correlation we have found between percentage and absolute cell numbers of Ki-67⁺CD4⁺ T cells at day 7 and viral load at days 14 and 84. Absolute numbers of HLA-DR⁺CD4⁺ T cells was also positively correlated with early (day 14) and late (day 84) viral replication. These findings suggest that viral replication during viral set point is already at least partially determined at day 7 by the magnitude of CD4⁺ T cell activation and proliferation.

Besides altering replication and target cell populations of SIV, changing the immune response against SIV could be another mechanism contributing to the disease acceleration and high viral loads induced by IL-15 treatment. The failure of NK cells and SIV-specific CD8⁺ T cells to control virus could result in the increased viral set point found in IL-15-treated animals. Although we only observed increases in blood SIV-specific CD8⁺ T cell numbers at week 2 in IL-15-treated animals, increased numbers of Gag-specific CD8⁺ T cells were found in LNs even at week 30. Because numbers of blood HLA-DR⁺ Gag-specific CD8⁺ T cells were reduced in IL-15-treated monkeys, this may suggest migration of activated cells into tissues. The fact that viral replication is higher in IL-15-treated animals despite having more SIV-specific CD8⁺ T cells in the LNs suggests that these cells may be either functionally defective or die there. If activated virus-specific CD8⁺ T cells accumulate at sites of viral replication, bystander killing or functional defects could be enhanced especially if viral replication is high. Further studies are necessary to determine whether SIV-specific CD8⁺ T cells in LNs and possibly other sites are fully functional or exhausted effector CD8⁺ T cells.

IL-15 treatment did reduce significantly the numbers of blood NK cells without increasing them in LNs. This may suggest an impairment of NK cells to control viral replication. In contrast, it could suggest that the homeostasis of NK cells has been affected with possibly higher turnover of these cells as a result of higher viral replication in the IL-15-treated animals. It should be stressed that for both reduced activated SIV-specific CD8⁺ T cells and NK cells and the accompanied increased viral loads, two alternative explanations can be given. Either loss of activated SIV-specific CD8⁺ T cells and NK cells is resulting in reduced control of viral replication and hence increased viral set points in IL-15-treated animals, or alternatively, increased viral replication due to nonimmunological mechanisms such as increased seeding of target cells results in exhaustion of activated SIV-specific CD8⁺ T cells and NK cells and thus the reduced numbers.

We also observed a decrease in anti-SIV Abs in IL-15-treated compared with untreated animals. However, no concurrent decrease in B cell numbers in peripheral blood and LN was observed. More studies are necessary to examine more carefully the effect of IL-15 on B cell function. The decreased anti-SIV Abs could indeed influence the outcome of SIV infection, as it was shown before that during progression to HIV and SIV, Abs specific for HIV and SIV decrease or are absent (65, 66). We observed an inverse correlation between the amount of anti-SIV Abs and viral replication at week 16. However, no correlation of anti-SIV Abs at week 3 and viral replication at week 3 was observed, indicating that although anti-SIV Abs are already significantly decreased early in SIV infection, this does not associate with viral set point.

Although IL-15 treatment induced a 2- to 3-fold increase in SIV-specific CD8⁺ T cells during acute infection, it was rather surprising that this had no effect on peak viremia. IL-15 could be enhancing the proliferation or survival of these cells. IL-15 inhibits CD95-induced apoptosis of HIV-specific CD8⁺ T cells (34, 60) and in vivo IL-15 induces massive expansion of CD8⁺ T cells (40). Given that SIV-specific CD8⁺ T cells are already >90% Ki-67⁺ as we show, it is unlikely that IL-15 is augmenting their proliferation; most likely, we are inhibiting the death of these cells. The fact that IL-15 increased SIV-specific CD8⁺ T cells and NK cells by 2- to 3-fold without any impact on peak viremia indicates how potent an immune response induced by an effective vaccine may need to be to impact on viral replication during acute infection.

IL-15 not only increased the viral set point but also led to accelerated disease progression to simian AIDS in two of the six treated animals. In one of the two animals, histology revealed that the brain was affected in this animal. Neuropathologically, SIV-infected macaques are known to reveal four distinct lesion profiles: minimal changes, early encephalitis, leukoencephalopathy, and encephalitis (42). Detailed morphological evaluation of one animal that died of AIDS in our study revealed evidence of minimal changes and incipient early encephalitis in the form of rare mononuclear inflammatory cell infiltrates in the cerebral and cerebellar, in conjunction with rare microglial nodules in the subpial portions of the cerebral cortex. The overall paucity of meningeal and parenchymal infiltrates coupled by the lack of multinucleated giant cells, substantial cortical and subcortical presence of microglial nodules and/or myelin pallor in the cerebral hemispheric white matter argue against the diagnosis of a full-blown SIV encephalitis or leukoencephalopathy. Our findings suggest that this animal suffered from early minimal CNS SIV-associated disease and incipient early encephalitis. It should be noted that previous studies to date have failed to establish a relationship between neuropathological findings, numbers of SIV-replicating cells, T cell infiltration, and repertoires of proinflammatory cytokines (42). However, a recent study suggests that IL-15 mRNA levels in the brain are increased in SIV infection and that this may indeed lead to maintenance of SIV-specific CD8⁺ T cells in the CNS (67). Our results could therefore be interpreted in different ways. Either IL-15 treatment increased the numbers of activated CD8⁺ T cells that could enter the brain and may therefore induce cytotoxic damage and inflammatory pathology, or more SIV-infected cells could enter the brain in the IL-15-treated animals and promote brain pathology. Furthermore, as mentioned above, we found much lower anti-SIV Abs in IL-15-treated animals compared with untreated animals. In a study using SIV/δ-induced immunodeficiency disease CNS involvement was observed in animals which did not mount a good Ab response to the virus (68). This could indicate that the significant lower amount of anti-SIV Abs found in the IL-15-treated animals at weeks 3 and 16 may be involved in the CNS damage we have observed. A strong anti-SIV response was also found in the plasma and cerebrospinal fluid of slow progressors but not of fast progressors when rhesus macaques were infected with SIVmac251 (69). Detailed studies will be necessary to reveal the mechanism responsible for this early involvement of the CNS in IL-15-treated animals and whether Ab levels play a part in this.

In summary, we have shown here that IL-15 treatment for 4 wk during acute SIV infection of rhesus macaques increased the viral set point in all animals and accelerated disease progression to simian AIDS in two of six animals with one of the animals developing SIV meningoencephalitis. Although 2- to 3-fold more SIV-specific CD8⁺ T cells and NK cells were observed in IL-15-treated animals this had no impact on peak viremia although numbers of infected cell in LN biopsies were reduced. Whereas activated SIV-specific CD8⁺ T cells and NK cells were reduced in the blood of IL-15-treated animals at viral set point, SIV-specific CD8⁺ T cells were increased in LN while NK cell numbers did not differ with IL-15 treatment. Early activation of CD4⁺ T cells followed by reduction of blood CCR5⁺ and CD45RA⁻CD62L⁻CD4⁺ T cells with IL-15 treatment suggest an increased use of these cells as targets of infection that leads to increased viral set point and infected cells in LN. The positive correlation between day 7 Ki-67⁺CD4⁺ and HLA-DR⁺CD4⁺ T cells and viral replication at day 84 further supports a direct association between early CD4⁺ T cell activation and viral set point. Finally, this study suggests that IL-15 treatment of acutely SIV-infected nonhuman primates may serve as a useful model to understand what controls viral set point in HIV infection.

We thank Louise Bertrand (Department of Neurobiology and Anatomy, Drexel University College of Medicine) for assisting with figure preparations. We thank the NCRR-funded Resource for Nonhuman Primate Immune Reagents for the recombinant rhesus macaque IL-15.

The authors have no financial conflict of interest.