M(IL-4) Tissue Macrophages Support Efficient Interferon-Gamma Production in Antigen-Specific CD8+ T Cells with Reduced Proliferative Capacity

CD8+ cytotoxic T cell (CTL) responses are necessary for the lysis of virally infected cells and control of infection. CTLs are activated when their TCRs bind a major histocompatibility complex (MHC)-I/peptide complex on the surface of antigen presenting cells such as macrophages (MΦ). It is now apparent that MΦ display remarkable plasticity in response to environmental signals to polarize into classically activated M(LPS + IFN-γ) or alternatively activated M(IL-4). However, little is known about how MΦ activation status influences their antigen presentation function to CD8+ T cell in models of virus infection. Consequently, we tested how polarization of spleen-derived (Sp)-MΦ impacts direct presentation of viral antigens to influence effector and proliferative CD8+ T-cell responses. We show that M(IL-4) Sp-MΦ retain MHC-I surface expression and the ability to stimulate IFN-γ production by CTL following peptide stimulation and lymphocytic choriomeningitis virus infection to levels similar to M0 and M(LPS + IFN-γ) MΦ. However, memory CD8+ T cells cultured in the presence of M(IL-4) MΦ underwent significantly reduced proliferation and produced similar IFN-γ levels as coculturing with M0 or M(LPS + IFN-γ) cells. Thus, these results show a novel ability of polarized MΦ to regulate CD8+ T-cell proliferation and effector functions during virus infection.

inTrODUcTiOn Tissue macrophages (MΦ) comprise an important member of the mononuclear phagocyte system where they regulate inflammation, cancer, and autoimmunity (1). They are involved in innate and adaptive immune responses to invading pathogens (2), and adapt their phenotype and function in accordance with their environment through a process termed MΦ polarization (3)(4)(5). It is now understood that both tissue MΦ and bone marrow (BM)-MΦ can develop into pro-inflammatory (M1) or anti-inflammatory (M2) (6)(7)(8).
During lymphocytic choriomeningitis virus infection (LCMV), MΦ support viral replication, process, and present viral antigens to activate CD8 + T cells (21)(22)(23)(24)(25)(26). Activated CD8 + T cells proliferate, gradually acquire cytotoxic T lymphocyte (CTL) effector function and home to the site of infection to secrete IFN-γ and lyse virally infected upon recognition of viral epitopes on MHC-I (27). It is known that M(IL-4) peritoneal MΦ and BM-MΦ inhibit OT-II proliferation in a signal transducer and activator of transcription (STAT)-6-dependent fashion (28). Moreover, in a murine norovirus infection model, helminth-induced M2 cells inhibit CD8 + T-cell proliferation (29). Yet, how polarized MΦ engage CD8 + T cells to control proliferation and functions during RNA virus infection remains unexplored. Here, we report on a novel finding supporting a dichotomized regulatory role of M(IL-4) tissue MΦ where they can inhibit CD8 + T-cell proliferation without affecting their IFN-γ production after peptide-specific antigen presentation.
Professional antigen presenting cells (pAPC) such as MΦ are needed for the activation of adaptive immune cells (30), as they are involved in antigen presentation via both MHC-I and MHC-II as well as their expression of costimulatory molecules (31). Nevertheless, LCMV has evolved mechanisms to interrupt APC activation and costimulatory molecule expression (32). Therefore, in order to assess the ability of polarized Sp-MΦ to engage CD8 + T-cell receptors, we characterized surface expression of activated Sp-MΦ markers following 24 h of LCMV infection ( Figure 1B). With regard to CD80 expression, M0 and M(LPS + IFN-γ) cells increased surface levels following viral infection, while M(IL-4) cells expression of CD80 remained largely unchanged (column 1). Interestingly, M0 cells slightly decreased CD86 expression following LCMV infection compared with M(LPS + IFN-γ) and M(IL-4) cells where no change was detected (column 2). M0 cells exhibited slight MHC-I reduction but not M(LPS + IFN-γ) or M(IL-4) Sp-MΦ (column 3). In addition, we also assessed expression of the inhibitory molecule PD-L1 (column 4). We observed that M(LPS + IFN-γ) cells expressed the greatest levels of PD-L1, while M0 and M(IL-4) had similar expression levels, which confirmed data in BM-MΦ published by another group (33). LCMV infection increased expression of PD-L1 in M0 and M(IL-4), while reduced expression in M(LPS + IFN-γ) Sp-MΦ. These data demonstrate that polarized cells are not negatively affected by LCMV infection when considering CD80/86 or MHC-I expression, while LCMV increases inhibitory molecule PD-L1 expression in M2 and M0 cells, but not M(LPS + IFN-γ).
To characterize further the functional profile of polarized cells, we investigated the release of pro-and anti-inflammatory cytokines in uninfected and LCMV-infected (24 h) Sp-MΦ. As expected, for the secretion of the cytokines TNF-α and IL-6 ( Figure 1C), M0 and M(IL-4) cells were poor, while M(LPS + IFN-γ) stimulation produced substantial levels agreeing with what has been described previously (34). Interestingly, 24 h post-LCMV infection, M0 and M(IL-4) cells both significantly increased production of TNF-α and IL-6. Moreover, M(LPS + IFN-γ) cells had reduced production of TNF-α after infection but were still producing significantly higher amounts than M0 and M(IL-4). No changes in IL-6 secretion were observed with M(LPS + IFN-γ) after the infection.
Lymphocytic choriomeningitis virus infection significantly decreased production of IL-12p40, in M0 and M(LPS + IFN-γ) cells while the opposite is true for M(IL-4), where production levels increased. Collectively, these data point to LCMV-promoting M(IL-4) cells to acquire a mixed M(LPS + IFN-γ)/M(IL-4) phenotype considering the ability to produce pro-inflammatory cytokines post-infection. For the anti-inflammatory cytokine IL-10, LCMV infection increased secretion in all subsets; however, M(LPS + IFN-γ) and M(IL-4) produced substantially less amounts than M0 infected cells ( Figure 1C).

M(il-4) sp-MΦ Present siinFeKl Peptide
Bound to Mhc-i at lower levels compared with M(lPs + iFn-γ)  (c) Cells were pulsed 10 −7 or 10 −9 M SIINFEKL for 2 h at 37°C before coincubation with the T-cell B3Z hybridoma for 18 h (1:1 ratio). The detection assay was carried out as described in Section "Materials and Methods" and OD was measured at 415 nm. Graphs depicting mean ± SD from three experimental replicates. MFI, mean fluorescent intensity; MHC, major histocompatibility complex.  Figure 2A demonstrate that each population of Sp-MΦ are able to display p:MHC on their surface. Measuring the fold change in mean fluorescent intensity (MFI) over unpulsed controls revealed M(LPS + IFN-γ) were best at binding and presenting the peptide and that Sp-M(IL-4) cells were the least efficient ( Figure 2B). This suggests that the polarized all MΦ subsets should be able to present H2-K b restricted epitopes to CD8 + T cells to varying degrees.
Based on the above observations (Figures 2A,B), we reasoned that this would translate to differential abilities to activate CD8 + T cells by the MΦ population. To test this, we employed the CD8 + T-cell hybridoma system for which inducible Lac-Z is under the NFAT enhancer that binds the IL-2 promoter (36,37). In this system, SIINFEKL-K b -specific TCR ligation results in binding to the IL-2 promotor and expression of downstream Lac-Z that can be detected by colorimetric changes (38). Using two different concentrations of the SIINFEKL peptide (10 −7 and 10 −9 M), we observed that M(LPS + IFN-γ) MΦ populations elicited B3Z activation better than M0 and M(IL-4) ( Figure 2C). Yet, M(IL-4) were still very proficient in activating B3Z T cells indicating that M(IL-4) cells retain sufficient antigen presentation capabilities that was close to M0 cells.

M(il-4) MΦ effectively stimulate epitope-specific cD8 + T cells to synthesize iFn-γ
Recently, it was shown that splenic marginal zone MΦ are responsible for activating an LCMV-specific CD8 + T cells when left unprimed by DC (39). Therefore, we extended our testing into the LCMV system, which includes H2-K b and H2-D b binding epitopes in H-2 b mice (40). Polarized Sp-MΦ were pulsed with either GP33-41 or NP396-404 then cocultured with splenocytes isolated from LCMV-infected mice (2 × 10 5 pfu i.p.), 8 days postinfection. The ability of virus-specific CD8 + T cells to produce IFN-γ after in vitro re-stimulation with the ICS assay was used to assess the polarized Sp-MΦ antigen presentation functions (Figure 3).
We found that the percentage of IFN-γ antigen-specific CD8 + T cells remained constant irrespective of MΦ activation status (Figures 3A,B). Moreover, in agreement with immunodominance hierarchy reported for this model (41), there was a greater percentage of IFN-γ-secreting GP33-41-specific CD8 + T cells than NP396-404 T cells (Figures 3A,B). This finding was identical irrespective of the activation MΦ state even at low-peptide concentrations of 10 −9 M. We further tested IFN-γ induction in LCMV-memory-specific CD8 + T cells using the LCMV GP33-41 peptide. After ex vivo stimulation by GP33-41 pulsed MΦ, we found that M(IL-4) cells were able to elicit similar activation of the memory T cells compared with M0 (data not shown). Therefore, we were able to demonstrate that M(IL-4)-MΦ polarization does not negatively influence their ability to stimulate activation and release of IFN-γ from antigen-specific effector or memory CD8 + T cells after virus infection.

evaluating the ability of activated MΦ to Present Viral antigens after infection
To assess whether the functional abilities with regard to antigen presentation were retained by the MΦ during viral infection, we infected Sp-MΦ with LCMV (MOI 5), and assessed their ability to present LCMV antigens to epitope-specific CTL. It was clear from the data ( Figure 4A) that LCMV was to infect the polarized MΦ in vitro and initiate its replication cycle as evident by LCMV-NP detection 24 h post-infection ( Figure 4A). This protein was not detected immediately during the first hour  presenting cells when interacting with CD8 + T cells, we discovered that they were proficient in presenting either peptides or processing viral antigens in different model systems. M(IL-4) MΦ inhibit antigen-specific T-cell proliferation (28,29). These publications utilized culture systems with M(IL-4) BM-MΦ and naïve P14 CD8 + T cells (specific for GP33-41) or anti-TCR and anti-CD28 antibodies to artificially stimulate naïve CD8 + T-cell proliferation in an antigen-independent fashion proliferation (28,29). However, how M(IL-4) cells direct memory recall responses is unknown. Therefore, we asked how CD8 T cells from LCMV memory, non-transgenic mice would respond to antigen presentation by polarized Sp-MΦ.
To assess this, we cultured peptide-pulsed either polarized or unstimulated Sp-MΦ with carboxyfluorescein succinimidyl ester (CFSE) labeled splenocytes from LCMV immune WT mice in the absence or presence of IL-2 (Figures 5A,B). Without exogenous IL-2 added to cultures, all the three MΦ populations failed to induce recall proliferation in vitro (data not shown), confirming previously published data by other groups (44). We then increased IL-2 culture concentrations to determine the minimum threshold of exogenous IL-2 required for T-cell expansion. At 5 U/mL, M0 and M(LPS + IFN-γ) MΦ induced 40 and 60% CD8 + T-cell proliferation, respectively (Figures 5A,B).
We were able to restore the deficit in the stimulation ability of M(IL-4) cells by increasing the IL-2 levels in the culture system to 20 U/mL. This finding was not restricted to a single specificity of epitope-specific T cells, as both GP33-41-and NP396-404-specific T cells proliferated to similar levels when comparing M(IL-4) with either M(LPS + IFN-γ) or M0 (Figures 5C,D). Collectively, the data from cD8 + T cells expanded for 6 Days by M(il-4) stimulators exhibit attenuated iFn-γ secretion compared with M0 or M(lPs + iFn-γ) cells From our observations above, we noted that although CD8 + T-cell proliferation was impaired compared with M0 and M(IL-4), approximately 20-30% of CD8 + T cell were still proliferating after peptide antigen presentation with M(IL-4) MΦ ( Figure 5B). As CD8 + T cells proliferate, they progressively acquire CTL functions (45). To test whether the epitope-specific CD8 + T cells were fully functional after the expansion period, we restimulated the expanded T cells using a common (BMA) MΦ cell line as antigen presenting cells so that the only variable factor in the assay would be the expansion difference of the Sp-MΦ (23,42,46).
Consequently, we cocultured the cells in vitro for 6 days using 5 U/mL IL-2 because this was the concentration at which we observed minimal M(IL-4)-induced CTL expansion (Figure 5A). We then tested for IFN-γ production by ICS as described earlier, but this time using the BMA cell line as APC to present the LCMV peptide to T cells (Figure 6A).
M0 and M(LPS + IFN-γ)-expanded CTL were ~60% IFN-γ positive, whereas M(IL-4)-expanded CTL displayed lower level of were pulsed with LCMV GP33-41 and cultured with CFSE-labeled splenocytes from LCMV immune mice (10:1 ratio) for 6 days in the presence of 5 U/mL IL-2 (a) or 10 U/mL IL-2 (B). Expanded cells were restimulated with GP33-41 pulsed BMA or unpulsed controls for detection of IFN-γ production by ICS in a restimulation assay to detect their epitope-specific activation. Graphs show mean ± SD from three replicates where *P < 0.05 and ***P < 0.0005. CFSE, carboxyfluorescein succinimidyl ester; LCMV, lymphocytic choriomeningitis virus infection.  Figure 5B, after 4 days of coculturing, the CD8 + T cells were tested for CFSE dilution as described before. CFSE, carboxyfluorescein succinimidyl ester; LCMV, lymphocytic choriomeningitis virus infection.  (47). Evidence suggests that for efficient CD8 + T-cell responses to occur, the APC must be helped by CD4 + T cells (48). In particular, direct engagement between CD40 and CD40L activates APC to enable IL-2 production for memory CD8 + T-cell proliferation (49). Given that IL-2 rescued CD8 + T-cell proliferation in M(IL-4) MΦ cultures in our experimental model, we reasoned that activated CD4 + T from an in vivo LCMV infection would compensate for the M(IL-4) ability to support CD8 + T cells in our model when the exogenous IL-2 levels are limiting in culture.
To test this hypothesis, we sorted CD4 + T cells from LCMVinfected mice on day 4 because it has been shown that CD69 expression levels are near their peak after virus infection in vivo (50). We cocultured the sorted CD4 + T cells with the MΦ plus CD8 + T cells during the CFSE proliferation assays in our model, when exogenous IL-2 levels are limiting. Although we observed that addition of CD4 + T cells enhanced the ability of M0 cells to increase CD8 + T-cell proliferation (from 38 to 50%), there was no positive influence in the M(IL-4) cultures (Figure 7) possibly because they may require additional help. Interestingly, no further proliferation (60%) was noted with M(LPS + IFN-γ) cells when we included the helper cells, indicating that a possible maximum proliferation had already been reached at such IL-2 levels without the CD4 + T cells present in the culture.

DiscUssiOn
It is now understood that polarized MΦ can modulate the outcome of infection and play a role in exacerbating or combating multiple diseases (4, 10). We previously reported that M-CSF induces in vitro differentiation of Sp-MΦ into cells that resemble splenic red pulp MΦ (23,51). Further characterization revealed that Sp-MΦ are effective antigen presenting cells capable of direct and cross presentation (23). Additionally, mature Sp-MΦ remain plastic and can be induced to M(LPS + IFN-γ) and M(IL-4) phenotypes, similar to BM-MΦ (8).
In our system, we observed a significant reduction of antigenspecific memory CD8 + T-cell proliferation when stimulated with M(IL-4) cells compared with M0 or M(LPS + IFN-γ) when employing IL-2 at limiting concentrations. Interestingly, supplementing the coculture medium with additional IL-2 restored CD8 + T-cell proliferation and subsequent IFN-γ secretion upon restimulation. The most likely explanation for this finding is likely due to the ability of IL-2 to tune TCR sensitivity and regulate cell-cycle progression. With regard to TCR sensitivity, exogenous IL-2 restores basally depressed CD3ζ expression in patients with chronic myeloid leukemia (59). Moreover, it was recently demonstrated that IL-2 reduces the threshold of activation in CD8 + T cells promoting responsiveness to low antigen levels (60). Therefore, by adding IL-2 we ostensibly improved the TCR sensitivity to antigen. In terms of cell cycle, IL-2 enhances expression of cyclin D3 and CDK4, and activates CDK2 promoting cell-cycle progression into S phase (61,62). Therefore, it is plausible that additional IL-2 overrides the M(IL-4) cell-induced CD8 + T-cell impairments by increasing CD3ζ expression and promoting entrance into cell cycle. In agreement with this notion, delivery of IL-2 complexed to anti-IL-2 monoclonal antibody breaks established CD8 + T-cell tolerance in an FBL model of murine leukemia (63). As such, future research into the biological mechanisms of our reported novel finding that IL-2 can overcome M(IL-4) stimulatory deficits CD8 + T cell is of immense interest to the immunotherapy field.
Another novel finding that is reported here is the ability of the polarized MΦ to elicit IFN-γ production by antigen-specific CTL following peptide stimulation or LCMV infection. Given the impairment of CD8 + T-cell TCR signaling described above following M(IL-4) coculture, we anticipated M(IL-4) cells to poorly induce IFN-γ secretion from CTL (57,59). A limitation of these studies was that they did not address the consequence of shortterm (<24 h) M(IL-4) MΦ-T-cell interactions on CD3ζ expression (57,59). However, the duration of our cytokine detection assays (approximately 6 h) allows us to separate the short-term and long-term impacts of M(IL-4) cells on CD8 + T-cell effector function and memory CD8 + T-cell proliferative response.
In our system, the observed ability of M(IL-4) cells to stimulate CTL IFN-γ release is likely attributed to the lower threshold of CTL activation for cytokine production compared with proliferation (64). It was previously demonstrated that low concentrations of TCR ligands result in IFN-γ, but not IL-2 production nor proliferation in CD4 + T cells (64). However, as ligand concentration increases, so does the diversity of cytokine response and proliferation levels (64,65). This implies that additional characteristic of polyfunctional CD8 + T cells requires a successive increase in signaling threshold. Conceptually, this might serve as a safeguard to control against unwanted CD8 + T-cell effector activity and cytokine production.
In summary, our results uncover a previously unknown ability of M(IL-4) cells to induce IFN-γ secretion by CTL and yet negatively regulate memory CD8 + T-cell proliferation compared with other forms of MΦ employed in this study. Thus, it is plausible that M(IL-4) cells possess novel undefined characteristics that need to be uncovered regarding how they may regulate immunity during infections.

animals, cell lines, and Virus
C57BL/6 (H-2 b ) mice (6-8 weeks) were purchased from (Jackson Laboratories) and were kept under specific pathogen-free conditions. Animal experiments were carried out in accordance with the guidelines of the Canadian Council of Animal Use and with approval from Queen's University Animal Care Committee. The BMA cell line BMA3.1A7 (a gift from Dr. K. Rock, University of Massachusetts Medical School, Worcester, MA, USA) is an adherent murine MΦ cell line generated from the BM of female C57BL/6 mice by overexpressing myc and raf oncogenes and has recently been characterized by our group to be a good model to study MΦ polarization (66,67). For viral infections, mice were injected at indicated plaque forming units (PFUs) or LCMV-WE intraperitoneal in 200 µL of sterile PBS.

Macrophage generation and activation
Bone marrow-MΦ and Sp-MΦ were generated as previously described (8,23,51 assessment of arginase and inducible nitric Oxide activity M(LPS + IFN-γ) polarization was determined by assessing NO production using Griess reagent as published previously (8,24). MΦ (200,000 cells/well) were seeded in a round-bottom 96-well plate and activated in phenol-red free RPMI, where nitrite concentration was determined by measuring OD at 540 nm using the Varioskan microplate reader. Sodium nitrite was used for the standard graph (0-100 µM) to calculate nitrite concentrations in the test samples and was purchased from Fisher Scientific (Whitby, ON, Canada). M(IL-4) polarization was assessed by arginase activity as described (8,68). Activated MΦ were resuspended in lysis buffer (0.1% Triton-X, 25-mM Tris-Cl, pH 8.0) purchased from Sigma (Oakville, ON, Canada) and incubated for 30 min at 4°C followed by centrifugation at 10,000× g for 20 min at 4°C. Protein content in supernatant was determined using Bradford reagent (Bioshop) and sample concentrations were equalized to 100 µg/mL. For the reaction, 100 µL of sample was added to 1.5-mL eppendorfs followed by 10 μL of 10-mM MnCl2 an incubation at a 55°C for 10 min. 100 µL 0.5-M l-arginine solution (pH 9.7) was then added before incubation for 1 h at 37°C. To stop the reaction, 800 µL of acid solution (7:3:1; H2O:H2PO4, 85%: H2SO4, 95%) and 40 µL of α-isonitrosopropiophenone (ISPF) (9% w/v in absolute ethanol) were added. Samples were heated at 100°C for 30 min and urea concentration was determined with the help of a urea standard graph (0-25 M) by measuring OD at 550 nm using a Varioskan spectrophotometric microplate reader.

antigen Presentation assays
In order to compare direct antigen presentation by activated MΦ, we utilized LCMV-specific T cells, either in vitro generated LCMV CTL or the B3Z hybridoma T cells. For ex vivo stimulation of LCMV-specific CD8 + T cells by polarized MΦ, splenocytes isolated from an LCMV-infected mouse on day 8 and were restimulated with GP33-41 or NP396-404 (10 −7 M) peptide-pulsed MΦ (ratio of 10 splenocytes: 1 APC) for 2 h. ICS for IFN-γ production was then performed as described below. LCMV-specific CTL were cultured with activated MΦ (1:1) that were pulsed with decreased peptide molarity (GP33-41/NP396-404: 10 −7 to 10 −9 M) or infected with LCMV-WE (MOI 3 or 5) for B3Z assay B3Z CD8 + T-cell hybridoma cell line specific for OVA residues 257-264 (SIINFEKL) presented on murine MHC-I (H2-K b ) was also used (36,69). Cells were cultured in IMDM medium containing 500 µg/mL G418 to ensure positive selection of reporter cells until time of experiment. For antigen presentation assays, APC were pulsed with SIINFEKL peptide (10 −7 or 10 −9 M) at 37°C for 2 h before extensive washing in warm PBS. Thereafter, the APC were cocultured with B3Z in a 96-well round-bottom plate (Thermo Scientific) at a ratio of 1:1 for 18 h in IMDM medium (5% FCS) at 37°C. For Laz-Z detection, Z Buffer (150 µL), containing 0.125% NP-40, 9-mM MgCl2, 100 mM of β-mercaptoethanol, and 5-mM ONPG, was added to the B3Z:APC cell pellets and incubated at 37°C for 4 h to allow for colorimetric change. Thereafter, 100 µL of buffer was transferred to a flat-bottom 96-well plate for absorbance measurement at 410 nm using a Varioskan plate reader.

Detection of lcMV-nP
For LCMV-NP detection, infected cells were fixed and permeabilized with PBS containing 4% paraformaldehyde and 0.5% saponin for 20 min at room temperature. Cells were washed in PBS with 0.25% saponin and incubated for 1 h with rat anti-LCMV-NP Ab (clone VL4) supernatants (a gift from Dr. M. Groettrup, University of Constance, Germany) to detect NP expression (70). After washing twice, FITC-conjugated goat anti-rat IgG Ab (Invitrogen) was incubated with the cells overnight at 4°C. In separate experiments, propidium idodide (1 µg/mL; Sigma Aldrich, Oakville, ON, Canada) was added to uninfected and LCMV-infected samples for assessment of cell death. Data were acquired with the Epics XL MCL (Beckman) and analyzed with the FlowJo software.
In Vitro cD8 + Proliferation assay Splenocytes were harvested from LCMV-WE immune (30 days post-infection) and lymphocytes were purified by ficoll-gradient centrifugation with lymphocytes separation medium (Fisher, Whitby, ON, Canada). Purified cells were labeled with CFSE (0.4 µM) for 15 min at 37°C then washed twice in warm PBS. Lymphocytes were cultured in a 96-well flat-bottomed plate with GP33-41 or NP396-41 peptide pulsed polarized MΦ for 3-6 days (at decreasing concentrations of recombinant IL-2) before staining for CD3 and CD8α expression and assessed for CFSE staining by FCM. Percent proliferation was determined by calculating the % of CD8 + CFSE Low cells compared with unstimulated CFSE labeled controls (CFSE Hi ). For the CD4 + T-cell coculture experiment: CD4 + T cells were negatively selected for by gating out B220 + , CD11b + , F480 + , and CD8 + cells from the spleen using a BD FACS Aria III sorter (Beckman-Dickinson). CD4 + T cells were added to MΦ + CFSE labeled splenocyte at the indicated ratio and incubated for 4 days.

statistical analysis
Statistical analysis was performed using Prism 7.0 (La Jolla, CA, USA) with unpaired (Student's) t-test.