Myeloid-Derived Suppressor Cells Inhibit T Follicular Helper Cell Immune Response in Japanese Encephalitis Virus Infection

Japanese encephalitis virus (JEV) is a member of the genus flavivirus and is symptomatically, genetically, and ecologically similar to Zika virus (ZIKV), West Nile virus, and St. Louis encephalitis virus (1). JEV can cause the worst mosquito-borne disease and epidemic viral encephalitis (Japanese encephalitis) and has recently been reported in previously nonaffected areas such as Australia (2) and Pakistan (3). It is estimated that 30,000–50,000 cases of Japanese encephalitis occur, resulting in 10,000–15,000 deaths each year (1). Its expanding geographical range and ability to cause devastating epidemics in nonimmune populations call for urgent efforts to develop efficient antiviral therapies (4) or alternate strategies to reduce JEV immunopathogenesis.

Many studies have focused on the importance of crossing the blood–brain barrier (BBB) in pathogenesis and immune responses of the CNS during JEV infection (5). JEV induces activation of endothelial cells to increase adhesion and transport of infected leukocytes across the BBB (6). These findings imply that JEV must escape immune surveillance at the periphery until the essential alterations occur at the BBB. Therefore, from the periphery to the CNS, evasion of immune clearance may be a prerequisite for JEV. However, little is known regarding the mechanism by which JEV escapes the immune response at the periphery. ZIKV outbreaks have recently occurred in French Polynesia (7, 8), Brazil (9), and the Central and South American countries (10, 11), which have revealed that the lack of knowledge of flaviviruses impedes clinical therapy and drug development. As flaviviruses, JEV and ZIKV have similar characteristics (12), and thus the mechanism of JEV immune evasion may also be applicable in studies of ZIKV.

Infection with some flaviviruses can suppress the initial immune response and further prevent detection and clearance by the host or enable the virus to subvert adequate development of adaptive immunity. For example, infection with yellow fever virus causes depletion of lymphocytic elements and necrosis of the B cell germinal center, with striking changes in spleen and lymph node architecture (13). Infection with dengue virus is often associated with leukopenia and bone marrow suppression (14, 15). West Nile virus appears to evade NK cell cytotoxicity by increasing surface expression of MHC class I molecules (16, 17). Furthermore, the early stage of JEV pathogenesis is characterized by a general state of immunosuppression (18). Thus, it is imperative to investigate the detailed mechanism of immunosuppression for resolving the viral escape from immune surveillance at the periphery.

Myeloid-derived suppressor cells (MDSCs), derived from myeloid cells, with strong immunosuppressive function, have been recognized for >20 y since their initial description in patients with cancers (19–21). In mice, MDSCs are broadly characterized as CD11b⁺Gr-1⁺ cells, and there are functionally distinct subsets in this cell population (22, 23). Early studies indicate that MDSCs are not only induced in cancer patients or tumor-bearing animals but are also presented during inflammation and infections (24–26). Recent studies have evaluated the roles that MDSCs play in the pathogenesis of viral infectious diseases. Infection with influenza A virus in mice or human patients results in the expansion of MDSCs, which suppress influenza A virus–specific immune responses (27). Adenoviral infection in mice induces rapid accumulation of granulocytic MDSCs in the early stage, and depletion of the MDSC population in vivo leads to an accelerated clearance of adenoviruses in the liver (28). Expansion of monocytic MDSCs dampens T cell function in HIV-1–seropositive patients (29, 30). All of these studies suggest that MDSCs may play a significant role in immunosuppression during viral infection. However, it remains unknown whether MDSCs mediate immunosuppression during JEV infection at the periphery.

The immune-suppressive activity of MDSCs is highly pleiotropic and involved in a variety of mechanisms (31). Suppression of effector T cell activity by MDSCs is one of the major mechanisms (29, 30). However, some studies have reported that MDSCs can mediate the function and differentiation of a subset of T cells. MDSCs suppress Th2 responses during primary Heligmosomoides polygyrus bakeri infection (32). MDSCs convert ex vivo naive CD4⁺ T cells into Foxp3⁺-expressing regulatory T cells (Tregs) (33, 34). Mouse MDSCs promote Th17 cell differentiation in experimental autoimmune encephalomyelitis (35). There is evidence that MDSCs regulate the immune response of effector B cells directly, accompanied by serum Ab reduction in murine models of systemic lupus erythematosus (36), and Tfh cells provide help to B cells by expressing CD40L and IL-21 (37), which are factors that promote B cell proliferation, isotype switching, germinal center formation, and the differentiation of memory B cells and long-lived plasma cells (38). Thus, these findings imply that it is possible for MDSCs to suppress Tfh function and Ab production.

In the present study, we investigated the immunosuppressive function of MDSCs during JEV infection. Our findings showed that infection with JEV virulent P3 strain expanded MDSCs, which was companied by an increase in Tregs and production of IL-10. However, only MDSCs facilitated the progression of infection, instead of Tregs and IL-10. P3-induced MDSCs suppressed the immune responses of CD4⁺ T cells, especially the subset of Tfh cells, and the population of splenic B cells (CD19⁺) and blood plasma cells (CD19⁺CD138⁺) decreased as well as a low level of total IgM and JEV-specific neutralizing Abs. Furthermore, the Tfh cells, B cells, plasma cells, and Ab responses recovered after depletion of MDSCs in vivo. The above findings indicated the potential function of MDSCs in mediating immune suppression by inhibiting Tfh cell responses. Additionally, humoral immunity was impaired, and the progression of infection was facilitated.

Adult C57BL/6 mice (female, 6–8 wk old) were obtained from the animal housing facility of the Chinese Academy of Sciences (Changsha, China) and maintained according to the Committee for Protection, Supervision, and Control of Experiments on Animals guidelines, Huazhong Agricultural University. Foxp3^gfp mice were provided as a courtesy of Dr. M. Zhang from Tsinghua University. JEV-P3 is a virulent strain, and JEV-SA14-14-2 is a vaccine strain. The two strains were previously preserved in our laboratory.

Mice were i.v. injected with 10⁵ PFU of JEV (strain P3 or SA14-14-2) in 100 μl of DMEM. Control animals received the same volume of DMEM via the same route. The time before infection was referred to as day 0. Infection with the wild strain of P3 continued for 10–15 d, and mice presented clinical symptoms at 5 d postinfection (dpi). Mice infected with SA14-14-2 presented no clinical symptoms. JEV-infected mice (three or more mice) were sacrificed every other day up to 9 d for either flow cytometry analysis or RNA extraction.

Single-cell suspensions from the spleen and lymph nodes were stained with different combinations of the following mAbs conjugated with FITC, PE, AF-647, PE-Cy7, and allophycocyanin.

For cell surface maker staining, single-cell suspensions were incubated with the following Abs: FITC anti-CD4 (GK1.5; MACS), FITC anti–Gr-1 (RB6-8C5; BioLegend), PE anti-CD11c (N418; BioLegend), PE anti-CD11b (M1/70; BioLegend), PE anti-CCR4 (anti-CD194, 2G12; BioLegend), PE anti-CXCR3 (anti-CD183, CXCR3-173; BioLegend), AF-647 anti-BTLA (anti-CD272, 8F4; BioLegend), allophycocyanin anti-CTLA4 (anti-CD152, UC10-4B9; BioLegend), allophycocyanin anti-CD25 (PC61; BioLegend), and allophycocyanin anti-CCR6 (anti-CD196, 29-2L17; BioLegend) in 0.2% BSA (BioSharp) with PBS (pH 7.4) for 25 min at 4°C. Cells were washed with PBS (400 × g, 5 min, 4°C) and fixed with 1% paraformaldehyde. Cells were determined using a FACSCalibur (BD Biosciences) system, and the data were analyzed using CellQuest Pro software (BD Biosciences).

For intracellular staining, cells were fixed and permeabilized using mouse Treg staining kit no. 1 (eBioscience) for 10 min at 25°C and stained with PE anti–Bcl-6 (BCL-DWN; eBioscience) and PE anti-Foxp3 (3G3; MACS). For analysis of cytokine production, cells were restimulated in vitro with PMA (50 ng/ml; Sigma Systems) and ionomycin (1 μg/ml; Sigma Systems) for 5 h at 37°C in the presence of brefeldin A (1×; BioLegend). Generally, cells were first stained with surface makers and then fixed and permeabilized using a fixation/permeabilization kit (BD Biosciences) for 20 min at 25°C and stained with PE anti–IL-10 (JES5-16E3; BD), PE anti–IFN-γ (XMG1.2; BD Biosciences), allophycocyanin anti–IL-4 (11B11; BD Biosciences), and PE anti–IL-21 (FFA21; eBioscience) or PE-conjugated rat IgG2b (BD Biosciences), PE-conjugated rat IgG1 (BD Biosciences), and allophycocyanin-conjugated rat IgG2a (BD Biosciences) for 40–60 min at 4°C. Cells were finally determined via FACS.

The spleen of each mouse was collected and homogenized in DMEM. Total RNA from spleen was extracted with TRIzol (Invitrogen, Grand Island, NY). cDNA was synthesized using a ReverTra Ace quantitative PCR reverse transcription kit (Toyobo, Osaka, Japan) according to the manufacturer’s instructions. Quantitative real-time PCR was performed using SYBR Green (Invitrogen) on a StepOnePlus system with StepOne software v2.2.2 (Applied Biosystems, Foster City, CA). The expression of mRNA was normalized to β-actin. The following primers were used: β-actin (5′-CACTGCCGCATCCTCTTCCTCCC-3′ and 5′-CAATAGTGATGACCTGGCCGT-3′), IL-10 (5′-TGTCCAGCTGGTCCTTTGTT-3′ and 5′-ACTGCACCCACTTCCCAGT-3′), IL-21 (5′-CTGAACTTCTATCAGCTCCACAAGATG-3′ and 5′-TTGGGTGTCCTTTTCTCATACGAATCAC-3′), Bcl-6 (5′-GGCTACGTCCGAGAGTTCAC-3′ and 5′-CTTGTGCGCTCTTAGGGGT-3′), T-bet (5′-GCCAGGGAACCGCTTATATG-3′ and 5′-GACGATCATCTGGGTCACATTGT-3′), and GATA3 (5′-CTCCTTTTTGCTCTCCTTTTC-3′ and 5′-AAGAGATGAGGACTGGAGTG-3′).

Naive CD4⁺ T cells and Tfh cells were purified from the spleen of uninfected mice using MoFlo XDP (Beckman Coulter) and labeled with 2.5 μM cytosolic CFSE (10 min, 25°C). The cells were seeded into 96-well plates (2 × 10⁵ cells per well) in RPMI 1640 medium stimulated with plate-coated anti-CD3 (5 μg/ml) and anti-CD28 Abs (2 μg/ml). MDSCs (CD11b⁺Gr-1⁺) were sorted from P3-infected mice and cultured with naive CD4⁺ T cells at different ratios in the presence of either IL-10 neutralizing Ab (10 μg/ml; R&D Systems) or anti-IgG1 Ab (10 μg/ml; R&D Systems) for 4 d at 37°C with 5% CO₂. MDSCs (CD11b⁺Gr-1⁺) sorted from P3-infected mice were also cultured with Tfh cells for 4 d at 37°C with 5% CO₂. The proliferation of CD4⁺ T cells and Tfh cells was measured by the loss of CFSE fluorescence using FACS (39).

For Treg depletion, mice were i.p. injected with purified anti-mouse CD25 (50 mg/mouse, PC61; eBioscience) or IgG1 isotype control (50 mg per mouse, eBRG1; eBioscience) 1 d prior to P3 infection. Peripheral blood was collected and analyzed by a FACSCalibur for percentage of CD4⁺CD25^hi cells and CD4⁺Foxp3⁺ cells from day 1 to day 16. CD25 was detected using an anti-CD25 mAb (3C7; MACS), which recognizes an epitope distinct from that of PC61.5 mAb.

Mice were i.p. injected with 100 μg per mouse of neutralizing Ab (R&D Systems) specific for IL-10 or rat IgG1 isotype control Ab (R&D Systems) 1 d prior to P3 infection.

The pellets of all trans retinoic acid and placebo samples were obtained from Innovative Research of America (Sarasota, FL). Retinoic acid pellets (21-d release, 5 mg) or placebo pellets were implanted by trocar injection 3 d before viral infection. The frequency of MDSCs was determined in splenocytes by flow cytometry after the administration of retinoic acid or placebo pellets.

Total serum IgM was measured with mouse IgM ELISA quantitation kits (4A Biotech). The level of IgM was determined at 450 nm absorbance using a standard enzyme instrument.

At the indicated times, serum was obtained from P3-infected or control mice. A baby hamster kidney fibroblast cell line (BHK-21) was used for the plaque reduction neutralization test (PRNT). Serum were diluted to 1:100, 1:200, 1:400, and 1:800 with DMEM containing 2% FCS, mixed with the virus, and incubated at 4°C overnight. BHK-21 cell monolayers were grown in 12-well plates and inoculated with the prepared serum. After incubation for 2 h, the incubation medium was aspirated, and the cells were washed twice with PBS. An overlay consisting of DMEM containing 1.5% carboxymethylcellulose (Wako) and 2% FCS (carboxymethylcellulose-DMEM) was added to the cells, and the plates were incubated at 37°C in a CO₂ incubator. After incubation for 5 d, the plaques were counted, and PRNT endpoint titers were expressed as the reciprocal of the serum dilution. The PRNT titer was calculated based on a 50% reduction in plaque count (40).

All data were analyzed with GraphPad Prism v5.0 (GraphPad Software, La Jolla, CA). Data are expressed as the mean ± SEM, and the significance of differences between groups was evaluated by using a parametric ANOVA test with a Tukey posttest or a t test. A p value <0.05 was considered significant.

How JEV escapes from the peripheral immune response remains unclear. It is well known that MDSCs comprise a heterogeneous population of immature myeloid cells that share a common property of suppressing T cell responses (29). These cells are generally identified in mice by coexpression of CD11b and Gr-1 (Ly6G/Ly6C) surface markers (22, 23). In the present study, the frequency of CD11b⁺Gr-1⁺ cells and Tregs was measured via flow cytometry to determine whether both types of immune cells expanded during the early stage of JEV infection. The result showed that P3 caused a significant expansion of CD11b⁺Gr-1⁺ cells in the spleen (Fig. 1A). CD11b⁺Gr-1⁺ cells increased from 1 to 5 dpi but began to decrease from 7 to 9 dpi (Fig. 1B). The percentage of CD11b⁺Gr-1⁺ cells in the P3 group was twice that as the control group, and the SA14-14-2 group was the same as the control group at 3 and 5 dpi (Fig. 1C).

Tregs are powerful inhibitory T cells that negatively regulate immune responses to maintain immune homeostasis (41). In some viral infections, Tregs can suppress effector T cell responses and promote pathogen persistence (42–44). This study demonstrated that the frequency of Tregs in the spleen significantly increased after P3 infection (Fig. 1D) from 3 to 7 dpi (Fig. 1E). There was a marked increase in growth according to the frequency of Tregs at 5 and 7 dpi during P3 infection (Fig. 1F). After expansion of CD11b⁺Gr-1⁺ cells, the frequency of Tregs increased, which dramatically increased in P3-infected mice that exhibited disease symptoms but not in P3-infected and disease-resistant mice (Fig. 1G).

The immunosuppressive molecule of IL-10 is also capable of negatively regulating T cell activation (45, 46). The level of IL-10 largely increased after P3 infection (Fig. 1H). Furthermore, P3-induced CD11b⁺Gr-1⁺ cells produced more IL-10 compared with the control and SA14-14-2 groups (Fig. 1I, 1J). Briefly, all three factors, CD11b⁺Gr-1⁺ cells, Tregs, and IL-10, were induced during P3 infection.

To determine whether CD11b⁺Gr-1⁺ cells, Tregs, or IL-10 was the key factor contributing to immune evasion of P3, depletion of CD11b⁺Gr-1⁺ cells or Tregs and neutralization of IL-10 were performed upon P3 infection. Retinoic acid can specifically reduce the number of CD11b⁺Gr-1⁺ cells with no significant decrease in T cells, NK cells, macrophages, or B cells (47). Thus, retinoic acid was applied to deplete CD11b⁺Gr-1⁺ cells, and the specificity of the depletion was tested. The results illustrated that depletion of CD11b⁺Gr-1⁺ cells did not affect the frequency of dendritic cells, macrophages, or lymphocytes in the spleen (Supplemental Fig. 1A–D) but reduced the weight of mice (Supplemental Fig. 1E). Based on the body weights of mice and Kaplan–Meier curves, the survival rate of mice significantly increased in the retinoic acid group (Fig. 2B), even with a reduction in body weight (Fig. 2A).

To further study the pathogenic role of Tregs and IL-10, Treg depletion and IL-10 neutralization were performed respectively in vivo. PC61.5 mAb was applied to deplete Tregs, and the percentage of CD4⁺CD25^hi cells and CD4⁺Foxp3⁺ cells in the blood decreased from day 1 to day 16 (Supplemental Fig. 2A, 2B). However, depletion of Tregs did not affect the body weight or mortality of the mice after P3 infection (Fig. 2C, 2D). Similarly, neutralization of IL-10 did not influence the mortality (Fig. 2F) even though the body weight of P3-infected mice reduced from day 8 to day 12 in the isotype group (Fig. 2E), suggesting that P3-induced CD11b⁺Gr-1⁺ cells might contribute to disease progression after P3 infection.

To further study the function of CD11b⁺Gr-1⁺ cells, the response of CD4⁺ T cells was tested. The percentage of CD3⁺CD4⁺ T cells decreased from 3 dpi and fell to the lowest level at 5 dpi, but recovered at 7 dpi (Fig. 3A, 3B). The percentage of CD3⁺CD4⁺ T cells was significantly decreased at 3 and 5 dpi compared with the SA14-14-2 and control groups (Fig. 3C). Furthermore, the substantial increase in CD11b⁺Gr-1⁺ cells during P3 infection coincided with the suppression of the CD4⁺ T cell immune response (Fig. 3D), raising the possibility that CD11b⁺Gr-1⁺ cells suppressed the CD4⁺ T cell immune response during P3 infection.

To determine whether P3-induced CD11b⁺Gr-1⁺ cells mediated immunosuppression of CD4⁺ T cells, CD11b⁺Gr-1⁺ cells were sorted from splenocytes of P3-infected mice and cocultured with CD4⁺ T cells stimulated with anti-CD3/CD28. Significant suppression of CD4⁺ T cell proliferation was observed in cocultures with CD11b⁺Gr-1⁺ cells compared with only CD4⁺ T cells (Fig. 3E, 3F). P3-induced CD11b⁺Gr-1⁺ cells shared the same makers and a common property of suppressing T cell responses with MDSCs, so we called them MDSCs instead of CD11b⁺Gr-1⁺ cells. Additionally, more IL-10 was induced in the P3-induced MDSCs (Fig. 1I), indicating that MDSC immunosuppressive activity might depend on IL-10. In coculture with P3-induced MDSCs, CD4⁺ T cell proliferation was suppressed significantly, whereas blocking Abs of IL-10 did not alleviate the suppression of cell proliferation (Supplemental Fig. 2C, 2D). These data indicated that P3-induced MDSCs limited CD4⁺ T cell immune responses but IL-10 was dispensable for the suppression of CD4⁺ T cell proliferation in vitro.

Next, Th1 cells, Th2 cells, Tregs, and Tfh cells were analyzed to obtain direct evidence of which subset of CD4⁺ T cells was inhibited during JEV-P3 infection. First, splenic Th1 (CD4⁺IFN-r⁺) and Th2 (CD4⁺IL-4⁺) cell populations were not affected by P3 infection on 3 and 5 dpi (Fig. 2A, 2B). Furthermore, the expression levels of differentiation-related transcription factor T-bet and GATA3 in the spleen were not reduced during P3 infection (Supplemental Fig. 3A, 3B). Second, the population of Tregs (CD4⁺CD25⁺Foxp3⁺) was increased at 5 dpi in the spleens of mice infected with P3 (Fig. 1D), but the percentage of Tfh (CD4⁺CXCR5⁺PD-1^hi) cells in the spleen was dramatically reduced at 3 and 5 dpi (Fig. 4C). The expression of differentiation-related transcription factor Bcl-6 in CD4⁺ T cells and the cytokine IL-21 in Tfh cells was lower in mice infected with P3 compared with that in SA-14-14-2–infected mice or control mice (Fig. 4D, 4E). Additionally, there was a marked reduction in the levels of IL-21 and Bcl-6 in the spleens of P3-infected mice (Fig. 4F, 4G), and the expression of the chemokine receptor CXCR5 in CD4⁺ T cells and the ligand CXCL13 in the spleen further illustrated the reduction in the Tfh cell population. The percentage of CXCR5 in CD4⁺ T cells was not reduced at 3 and 5 dpi (Supplemental Fig. 3C), but the expression of CXCL13 was significantly decreased at 3 and 5 dpi (Supplemental Fig. 3D). Thus, we found that P3 infection is sufficient to cause immune suppression of Tfh cells in vivo.

It is well known that Tfh cells can drive follicular B cell development to plasma cells and that they play essential roles in affinity maturation of the Ab response (38). Tfh cell differentiation is initiated at the early stage of acute viral infection (46). To determine whether the inhibition of Tfh cells in vivo affects the immune response of B lymphocytes and plasma cells, B lymphocytes (CD19⁺) and plasma cells (CD19⁺CD138⁺) were measured with flow cytometry in peripheral blood and spleen after P3 infection. The percentage of B lymphocytes in peripheral blood decreased dramatically (Fig. 5A). The percentage of B lymphocytes in peripheral blood infection (Fig. 5B) and plasma cells in the blood (Fig. 5D) decreased from 3 to 9 dpi of P3. The percentage of B lymphocytes in peripheral blood (Fig. 5C) and the percentage of plasma cells in the blood (Fig. 5E) and spleen (Fig. 5G) were lower at 5, 7, and 9 dpi of P3 compared with the SA14-14-2 and control groups. However, the percentage of B lymphocytes in spleen decreased at 5 and 7 dpi of P3 (Fig. 5F). These data suggested that the immune response of B lymphocytes and plasma cells was suppressed following immune suppression of Tfh cells in vivo.

We further sought to determine whether immune suppression of Tfh cells in the early stage of P3 infection was mediated by P3-induced MDSCs. To test the effect on the Tfh subset by P3-induced MDSCs in the early stage of P3 infection, first, MDSCs were sorted from splenocytes of P3-infected mice and cocultured with Tfh cells stimulated with anti-CD3/CD28 Abs. Significant suppression of Tfh cells was observed in cocultures with MDSCs compared with only Tfh cells (Fig. 6A). Second, we depleted MDSCs with retinoic acid in vivo. The result showed that the percentages of Tfh cells (Fig. 6B) and Bcl-6 in CD4⁺ T cells (Fig. 6C) in the spleen were greatly decreased at 3 and 5 dpi of P3 but recovered to normal levels after depletion of MDSCs in vivo. Additionally, the effect on the percentages of Th1, Th2, and Tregs by P3-induced MDSCs was determined in the early stage of P3 infection. No differences were observed between placebo and retinoic acid treatment in the populations of Th1, Th2, and Tregs at 3 and 5 dpi (Supplemental Fig. 4). These data suggested that depletion of MDSCs in vivo halts suppression of the Tfh immune response at the early stage of P3 infection.

We next questioned whether immune suppression of B lymphocytes and plasma cells could be relieved following recovery of the Tfh cell immune response. We compared the percentage of B lymphocytes in the spleen in the retinoic acid treatment group with the placebo group, and the results showed that the percentage of B lymphocytes recovered after depletion of MDSCs (Fig. 6D). The population of plasma cells in the blood was also restored after MDSC depletion (Fig. 6E). Additionally, the levels of IgM production in the serum were determined after P3 infection by ELISA, which revealed a higher level of IgM production after depletion of MDSCs at 3 and 5 dpi (Fig. 6F, 6G). The titer of viral neutralizing Abs in the serum of MDSC-depleted mice was 1:800, whereas the titer of placebo-treated mice was only 1:400 (Fig. 6H). Therefore, these results clearly indicated that JEV-P3–induced MDSCs were sufficient for mediating immune suppression of Tfh cells in vivo, leading to a low level of IgM and neutralizing Abs.

Various mechanisms contribute to immunosuppression, which facilitates viral immune evasion from the host defenses. This study identified a new immune evasion pathway by which virus infection–induced MDSCs suppress CD4⁺ T cells, especially Tfh function and subsequent humoral immune responses in a JEV-infected mouse model. These findings demonstrate an immunosuppressive pathway during JEV infection and highlight the importance of innate MDSCs in inhibiting adaptive immune responses during viral infection.

Diminished frequency and function of CD4⁺ T cells attributed to the persistence of viral infections in both humans and mice (48, 49). In our study, it was shown that the expansion and accumulation of MDSCs were correlated with T cell suppression during JEV-P3 infection. It has been reported that infection-induced MDSCs can suppress T cell responses to mediate immunosuppression and facilitate immune evasion during Staphylococcus aureus infection (50). In both hepatitis C virus– and HIV-infected patients, virus-induced MDSCs suppress T cell responses, which is associated with the persistence of hepatitis C virus infection (25, 29). These studies have confirmed that MDSCs suppress T cell activity to modulate immune responses and, in some cases, suggest that the function of MDSCs relies on production of cytokines such as IL-10 (51, 52). However, the immunosuppressive function of MDSCs was independent of IL-10 during JEV-P3 infection. MDSCs can also exert immunosuppressive function through the production of NO, reactive oxygen species, and H₂O₂ (53–55) or via cell-to-cell contact (50, 56). Based on these observations, it remains unclear whether P3-induced MDSCs suppress T cells via cell surface molecules or the release of other short-lived soluble mediators in addition to IL-10.

MDSCs can secrete a variety of inflammatory and anti-inflammatory cytokines, endowing their involvement in the polarization of naive CD4⁺ T cells in different subsets (57). MDSCs contribute to the differentiation of Tregs not only in cancer patients and tumor-bearing mice but also in virus-infected patients (33, 34, 58, 59). MDSCs also promote Th17 cell differentiation in experimental autoimmune encephalomyelitis (35). Although no reports have confirmed that MDSCs regulate Tfh cell immune responses directly, MDSCs can reduce serum Abs of dsDNA and decrease the population of effector B cells in the germinal center and blood in murine models of systemic lupus erythematosus (36). The Tfh cell is one of the most important CD4⁺ T cell subsets, which guide B cell differentiation and Ab class switching as well as affinity maturation (60–62), suggesting that MDSCs may be involved in the functional regulation of the Tfh subset and humoral immunity. Tfh cell commitment is an early event that can be initiated as early as 2 d after acute viral infection (63), and we found that the Tfh population decreased as MDSCs expanded (Figs. 1A, 4C). Cytokine IL-21 promotes Tfh cell differentiation, inducing B cells to undergo class switch recombination and to produce large amounts of IgG (37, 64). Bcl-6 is also a master regulator of Tfh cell differentiation and necessary for Tfh cells to provide help to B cells in vivo (65). Our results showed that the expression level of IL-21 significantly decreased, and the level of transcription factor Bcl-6 also decreased after JEV-P3 infection; B cell and plasma cell immune responses were both suppressed after JEV-P3 infection as the Tfh cells decreased. Therefore, MDSCs may suppress the differentiation of Tfh cells during the early stage of JEV-P3 infection, leading to humoral immune suppression.

Although all three immunosuppressive factors, MDSCs, Tregs, and IL-10, were induced during P3 infection, only depletion of MDSCs reduced the mortality of mice after P3 infection. There was a positive correlation between Tregs and disease progression. The Treg population increased from 3 dpi, which occurred later than the expansion of MDSCs and the suppression of Tfh cells. In our previously published study, we found that JEV invades the CNS as early as 3 or 4 dpi (66). Therefore, Tregs may be a result of disease caused by JEV. JEV-mediated immunosuppression and viral immune evasion are independent of Tregs, similar to S. aureus infection (51).

Furthermore, IL-10 is important for MDSC immunosuppressive function (53) and plays an important role in the expansion of Tregs (67, 68). In our study, neutralization of IL-10 did not affect the mortality or body weight of mice after P3 infection (Fig. 2E, 2F). Although there was a higher level of IL-10 induction in the P3-induced MDSCs (Fig. 1I), IL-10 was dispensable in the suppression of CD4⁺ T cells mediated by MDSCs. Additionally, JEV-mediated immunosuppression and viral immune evasion are independent of Tregs, and therefore IL-10 may be not involved in the immunosuppression and persistence of viral infection during JEV infection.

In this study, JEV-P3 infection leading to acute encephalopathy depended on suppression of the adaptive immune responses mediated by innate cells. Meanwhile, JEV-SA14-14-2, a successful JEV live vaccine strain (69), did not induce immunosuppressive factors MDSCs, Tregs, and IL-10. However, more expression of T-bet and GATA3 in mice infected with JEV-SA14-14-2 as well as more Bcl-6 expression and a high level of CXCL13 were determined. T-bet and GATA3 regulate T cell development, proliferation, and maintenance (70, 71). Moreover, Bcl-6 is necessary for Tfh cell function (65). Besides, CXCR5 is the canonical Tfh cell marker. Tfh cells migrate in response to CXCL13 (72–74). These observations suggest the potential mechanism of the immune escape of P3 and the high immunopotency of SA14-14-2.

The large amount of MDSCs during chronic inflammation and chronic infection appears to be a common feedback mechanism in both mice and humans. Thus, many different avenues to combat MDSC number, function, and generation are being actively pursued (56). In our study, MDSCs were implicated in the suppression of Tfh subset function and the subsequent humoral immune responses in the JEV-infected mouse model; thus, we think that counteracting the effects of MDSCs may represent an alternate way to confront the difficulties in treating not only chronic but also acute infections. In summary, this study highlighted the polyphasic nature of innate responses to viral infections and revealed the significance of innate cell inhibition in the adaptive immune responses during viral infection.

The authors have no financial conflicts of interest.