Experiments were performed using sex- and age-matched inbred C57BL/6 mice (Harlan Laboratories, Germany) and wild type and transgenic DEREG mice (25). At day 0 of experiments, mice were between 7 and 10 weeks old. For the experiments, the FV complex containing B-tropic Friend murine leukemia helper virus and polycythemia-inducing spleen focus-forming virus was used. The FV stock was prepared as a 15% spleen cell homogenate from susceptible BALB/c mice infected 14 days previously with 3,000 SFFU of FV. Mice were injected intravenously with 0.1 ml phosphate-buffered saline (PBS) containing 20,000 SFFU of FV, if not indicated differently. The virus stock did not contain lactate dehydrogenase-elevating virus. Mice were sacrificed at 3 dpi by cervical dislocation.

Detection of infectious centers was done by 10-fold dilutions of single-cell suspensions onto Mus dunnis cells. Co-cultures were incubated for 72 h and fixed with ethanol. Mus dunni cells were stained with the F-MuLV envelope-specific monoclonal antibody 720, and developed with a peroxidase-conjugated goat anti-mouse antibody. In a final step, cells were incubated with aminoethylcarbazol for the detection of foci.

Multi-parameter flow cytometry was done with the following antibodies: CD3 (17A2), CD4 (RM4-5), CD11b (M1/70), CD11c (N418), CD49b (DX5), CD69 (H1.2F3), CD80 (16-10A1), CD86 (GL1), F4/80 (BM8), FasL (MFL3), Gr-1 (RB6-8C5), GzmB (GB11), ICOS (7E.17G9), IL-10 (JES6-5H4), KI-67 (SolA15), KLRG-1 (2F1), NK1.1 (PK136), PD-L1 (10F.9G2), Ter119 (Ter119), TGF-β1 (TW7-16B4), TNFα (MP6-T22) and Foxp3 (FjK-16S). For the identification of FV-infected cells a FV protein gp70 (Ab720) Alexa Fluor 647-conjugated antibody was used (26). To exclude dead cells, cells were stained with Zombie UV (Fixable Viability Kit, BioLegend) dye. For gating on lineage-negative (lin⁻) cells, dead cells, T cells and NK cells were excluded from the analysis. Splenocytes were restimulated with ionomycin (500 ng/ml), phorbol myristate acetate (PMA; 25 ng/ml), monensin (1X), and brefeldin A (2 μg/ml) diluted in Iscove's modified Dulbecco's medium (IMDM) buffer at 37°C for 3 h. For intracellular stainings, cells were fixed with Fixation/Permeabilization Solution Kit (BD Biosciences) whereas cells were fixed with Foxp3 Transcription Factor Fixation/Permeabilization kit (Thermofisher) for intranuclear stainings. Data were acquired at LSR II flow cytometer (BD).

NK cells were isolated from spleens with the MojoSort Mouse NK cell Isolation Kit (BioLegend) according to the manufacturer's protocol. YAC-1 cells or FBL-3 cells were stained with carboxyfluorescein succinimidyl ester (CFSE, 2.5 μM). Cells were co-incubated in an ET ratio of 25:1. The co-incubation was performed in 96-well U-bottom plates at 37°C in a humidified 5% CO₂ atmosphere. After 18 h cells were washed and stained with fixable viability dye. Cells were measured immediately at LSR II.

Total RNA was isolated using the DNA/RNA Shield (Zymo research) and the innuPREP RNA mini kit (Analytik Jena). cDNA was synthesized with innoScipt reverse transcriptase (Analytik Jena). Real time-PCR analysis of IL-15 and IL-18 was performed using innuMIX quantitative PCR (qPCR) MasterMix SyGreen (Analytik Jena). Oligonucleotide sequences were ordered at Biomers as follows: for β-actin, 5′-AAATCGTGCGTGACATCAAA-3′ and 5′-CAAGAAGGAAGGCTGGAAAA-3′; IL-15, 5′-CATTTTGGGCTGTGTCAGTG-3′ and 5′-TCTTCAAAGGCTTCATCTGCAA-3′. For the detection of mouse IL-18 Mm-Il18-1-SG QuantiTect primer assay was purchased from Qiagen. The quantitative mRNA levels were determined by using Rotor-Gene Q series software (Qiagen) and were normalized to the β-actin mRNA expression levels.

Mice were injected intraperitoneally with the NK1.1-specific monoclonal antibody PK136 1 day prior FV infection and 1 day after infection to deplete NK cells. More than 90% of NK cells (CD3⁻ CD49b⁺ NK1.1⁺) were depleted in the spleen. To deplete regulatory T cells in transgenic DEREG mice, mice were injected intraperitoneally with DT (0.5 μg, Calbiochem) diluted in PBS at −1 and 1 dpi.

To neutralize IL-10, mice were injected with 50 μg LEAF Purified anti-mouse IL-10 antibody (JES5-2A5, BioLegend) at day 1, 2, and with 100 μg at day 1. For the neutralization of TGF-β, mice were injected i. p. with 200 μg of InVivoMAb anti-mouse TGF-β (1D11.16.8, BioXCell) every other day starting 1 day prior infection.

Statistical analyses were computed with Graph Pad Prism version 6. Statistical differences between two different groups were determined by the Mann–Whitney test (non-parametric) or unpaired t-test (parametric test), depending on the Gaussian distribution. Differences between more than two groups were analyzed by Kruskal–Wallis test (non-parametric) or Ordinary one-way ANOVA (parametric test). If indicated, Rout method was used to remove outliers.

Since application of diphtheria toxin (DT) into transgenic DEREG mice results in a depletion of up to 90% of Tregs (25), we used these mice to analyze the regulation of NK cells by Tregs. To address the question whether Tregs suppress NK cell functions during initial FV infection (3 dpi), we depleted Tregs and subsequently determined NK cell numbers and their phenotype. All infection experiments were performed with a medium dose of FV inoculum, which we previously demonstrated to induce only very weak NK cell responses during acute infection (19). We did not detect any changes in the absolute numbers of CD3⁻ CD49b⁺ NK1.1⁺ cells in the spleen after FV infection (3 dpi) nor increased NK cell numbers after Treg depletion (Figure 1A). However, Treg depletion increased the percentage of proliferating (KI-67), activated and matured (KLRG-1), effector phenotype (CD86, GzmB, and FasL) and cytokine producing (TNFα) NK cells during acute infection (Figure 1B). To test the cytotoxic potential of NK cells against target cells in vitro, we co-incubated isolated NK cells with YAC-1 cells and determined their killing capacity (Figure 1C). We did not detect any increased killing of target cells after FV infection in comparison to naïve mice whereas the ablation of Tregs resulted in a significant increase of NK cell-mediated target cell elimination. Similarly, we analyzed the killing of the FV-induced tumor cell line FBL-3 (Figure 1D), which expresses FV antigens but does not produce infectious virus. Again, we detected a significantly augmented elimination of FBL-3 target cells when using NK cells from infected, Treg-ablated mice compared to the only FV group. Thus, the absence of suppressive Tregs improved NK cell proliferation, activation and cytotoxicity during acute FV infection.

Activation of NK cells depends on diverse signals such as cytokines and cell-to-cell contacts with myeloid cells (2, 11, 12). Several bidirectional interactions between NK cells and DCs or between NK cells and macrophages were reported (10, 19, 27). Beside NK cells, also Tregs are able to interact with DCs and macrophages. Tregs can block the maturation and functions of DCs and macrophages, e. g. through the ligation of CTLA-4 with CD80 and/or CD86, via binding of LAG3 to MHC II molecules or by the Fas/FasL pathway (28, 29). To address the question whether Tregs suppress the activation of macrophages and DCs, hence indirectly regulate NK cell functions, we analyzed the APC activation by detection of the CD80 co-stimulatory molecule after acute FV infection. FV infection alone did not result in increased percentages of activated APCs whereas the depletion of Tregs during FV infection resulted in an increased percentage of CD80⁺ DCs (Figure 2A) and macrophages (Figure 2B). In a previous study, we demonstrated that during the acute phase of FV infection activated DCs and macrophages were sources of the NK cell-stimulatory cytokines IL-15 und IL-18 (19). Because macrophages and DCs are able to produce large amounts of IL-15 and IL-18, we analyzed IL-15 and IL-18 mRNA levels in splenocytes from FV-infected and Treg depleted animals during initial infection. Interestingly, we did not detect a significant increase in IL-18 mRNA levels but we found a substantially augmented mRNA level for IL-15 in Treg-ablated mice (Figure 2C). Taken together, Tregs suppress the activation of DCs and macrophages that correlates with reduced mRNA levels for IL-15 during the first days of FV infection.

To analyze the influence of Treg depletion on the viral replication and spread, we measured viral loads in spleens of FV-infected mice (3 dpi) with an infectious center assay (Figure 3A). Depletion of Tregs (FV + DT) resulted in a more than 3-fold increase of viral loads in comparison to the only FV-infected group (FV). NK cell depletion in animals infected with 20,000 SFFU of FV, had no influence on viral loads (19). The question was, why viral loads increased at 3 dpi after Treg depletion and one possible explanation was that Tregs also control the activation of the target cells for FV. As target cells for FV infection, erythroid precursor cells, monocytes, granulocytes, T cells and B cells were described, and FV is known to efficiently infect activated and dividing cells (30, 31). We therefore analyzed cell subsets, which are potential targets for early FV infection. To this end, we used a viral envelope protein gp70-specific antibody (Ab720) that allows the detection of FV-infected cells (26). Absolute numbers of infected erythroid precursor cells (Ter119⁺) and granulocytes (Gr-1⁺) as well as macrophages (F4/80⁺) and DCs (CD11c⁺) were analyzed by multi-parameter flow cytometry and calculated for whole spleens. Initial FV infection resulted in moderate numbers of gp70⁺ cells whereas depletion of Tregs led to a significant increase in gp70⁺ DCs, erythroblasts, granulocytes and macrophages (Figure 3B). Thus, reduced suppression of target cell activation due to Treg depletion resulted in enhanced infection levels during acute FV infection.

Because we detected increased viral loads after Treg depletion but, counter-intuitive, also an increased cytotoxic activity of NK cells, we determined whether or not NK cell activation had an effect on the levels of virus replication. Additional depletion of NK cells led to a mean viral burden of 1,474 FV-infected cells per million cells in comparison to 248 infected cells in the FV + DT group (Figure 3A). In addition, NK cell depletion in Treg-ablated mice substantially increased numbers of gp70⁺ macrophages (Figure 3B). Thus, activated NK cells reveal an anti-retroviral function and efficiently reduce numbers of infected cells in the absence of Tregs.

On one hand, Tregs are essential for maintaining self-tolerance and homeostasis but on the other hand they can contribute to viral spread and chronic infections. However, whether these suppressive cells control the initial immune response to acute viral pathogens needs to be further elucidated. To study the phenotype of Tregs during acute infection, we first determined the absolute numbers of CD4⁺ Foxp3⁺ Tregs in the spleen at 3 dp FV infection (Figure 4A). We detected almost two-times higher numbers of Tregs in FV-infected spleens compared to naïve organs. Furthermore, we analyzed the activation of Tregs by the early activation marker CD69, the inducible co-stimulator ICOS (CD278) and the Programmed cell death ligand 1 (PD-L1, CD274) that are upregulated on activated T cells (32, 33). Compared to the naïve group, FV infection substantially increased the percentage of CD69⁺, ICOS⁺ and PD-L1⁺ Tregs at 3 dpi, indicating early Treg activation and expansion (Figure 4B). Treg are able to suppress immune cells through several mechanisms such as metabolic disruption, modulation of DCs or release of inhibitory cytokine (29). Particularly the inhibitory cytokines IL-10 and TGF-β have been the focus of considerable attention as mediators of Treg suppression. Therefore, we analyzed these molecules in CD4⁺ Foxp3⁺ Tregs during acute FV infection via multi-parameter flow cytometry. Although we did not detect significant changes in the percentages of IL-10 expressing Tregs in FV infection compared to naïve mice, we showed a significant higher Mean fluorescence intensity (MFI) of IL-10 in Tregs from the group of FV-infected animals (Figure 4C). TGF-β is first synthesized as precursor TGF-β protein and then non-covalently linked to the latency-associate peptide (LAP) (34). Involving also other proteins, this protein complex increases the secretion efficiency of TGF-β and targets the complex to the extracellular matrix (35, 36). Here, we used an anti-mouse LAP antibody to detect TGF-β1⁺ Tregs in the spleen of naïve and FV-infected mice (Figure 4D). The analysis of TGF-β revealed an increased percentage of TGF-β1⁺ Tregs but no changes in the MFI for TGF-β in Tregs were detected. Thus, we demonstrated that IL-10⁺ Tregs produced more IL-10 and additional Tregs initiated the production of TGF-β after a few days of FV infection.

Our observation that Tregs produced immunosuppressive cytokines and NK cells were inhibited by Tregs raised the question, whether this inhibition might be mediated via the cytokines IL-10 and/or TGF-β. Therefore, we neutralized IL-10 or TGF-β by repeated injections of neutralizing antibodies as it is shown in Figure 5A. In a previous study, we have demonstrated that activated macrophages and DCs efficiently augment the antiviral activity of NK cells (19). Thus, we first addressed the question whether IL-10 and TGF-β suppressed the activation of these APCs. Figures 5B,C show that we did not detect an IL-10- or TGF-β-mediated suppression of DC or macrophage activation, measured by CD80 expression. In addition, IL-10 or TGF-β neutralization did not significantly affect levels of mRNA for IL-15 or IL-18 in splenocytes (Figure 5D). In a next step, we addressed the question whether the Treg-produced cytokines suppressed the NK cell activity. Indeed, we detected a significantly increased percentage of proliferating NK cells in FV-infected mice after IL-10- and TGF-β-neutralization. Beside the proliferation (KI-67), we analyzed activated (KLRG-1) NK cells, determined the percentage of cytokine-producing (TNFα) NK cells and measured the cytotoxicity-associated marker FasL (Figure 5E). Interestingly, only IL-10 neutralization resulted in increased percentages of KLRG-1⁺, FasL⁺ and TNFα⁺ NK cells, but not TGF-β neutralization. Collectively, the data indicate that IL-10 directly suppressed NK cell proliferation and functions during initial FV infection, but did not influence the activation of macrophages and DCs.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.