Polymorphonuclear Myeloid-Derived Cells That Contribute to the Immune Paralysis Are Generated in the Early Phase of Sepsis via PD-1/PD-L1 Pathway

ABSTRACT

Immune paralysis is a protracted state of immune suppression following the early/acute inflammatory phase of sepsis. CD11b⁺ Gr-1⁺ cells induced during sepsis are heterogeneous myeloid-derived cells (MDCs). This study investigated the contribution of MDCs to immune paralysis. Treatment of mice with zymosan (ZM) induced a marked increase in the total number of splenocytes with an increase in the proportion of Gr-1^hi cells and a decrease in the proportion of T cells on day 7; levels of these cells eventually return to levels similar to those of control mice on day 21. T-cell activation and gamma interferon (IFN-γ) expression by CD8⁺ T cells were clearly impaired in ZM-treated mice on day 21 (d21-ZM mice). Gr-1^hi cells showed a CD11b⁺ Ly6G^hi polymorphonuclear phenotype. Injection of lipopolysaccharide (LPS) into d21-ZM mice impaired interleukin 6 (IL-6) production in serum, accompanied by accumulation of CD11b⁺ Gr-1^hi cells in the peripheral blood. Transfer of Gr-1^hi cells from d21-ZM mice into intact mice impaired IL-6 production, but similar transfer of Gr-1^hi cells from PD-1/PD-L1-deficient d21-ZM mice showed no such suppressive effect. Conversely, either depletion of Gr-1^hi cells by treatment with anti-Gr-1 monoclonal antibody (MAb) or neutralization of the PD-1/PD-L1 pathway by anti-PD-1 and anti-PD-L1 MAbs during the induction phase of sepsis ameliorated ZM-induced immune suppression. Our results suggest that the PD-1/PD-L1-mediated generation of Gr-1^hi cells in the early phase of sepsis is required for the late phase of immune paralysis.

INTRODUCTION

Sepsis is defined as life-threatening organ dysfunction caused by a dysregulated host immune response against microbial infection, and it shows a biphasic progression (1). The early phase of sepsis is an acute hyperinflammatory state characterized by excessive production of proinflammatory mediators by innate immune cells. In contrast, the later phase of sepsis shows protracted immune suppression, termed immune paralysis, which promotes chronic infection. The mortality rate in late sepsis is higher than that in the early phase because of hyporesponsiveness with persistent primary or secondary infection caused by innate and adaptive immune dysfunction (2). Induction of immune paralysis during sepsis is caused by increased numbers of immune-suppressive cells and enhanced expression of immune-suppressive factors (1).

Myeloid-derived suppressor cells (MDSCs) are heterogeneous innate immune cells involved in regulation of the immune response in diverse pathological conditions (3). In mice, MSDCs are defined as CD11b⁺ Gr-1⁺ cells and consist of two subpopulations: CD11b⁺ Ly6G⁺ Ly6C^lo polymorphonuclear MDSCs (PMN-MDSCs) and CD11b⁺ Ly6G-Ly6C^hi monocytic MDSCs (M-MDSCs). Both cell populations play important roles in the pathogenesis of sepsis (4). In the early phase of sepsis, MDSCs secrete more proinflammatory cytokines, such as interleukin 6 (IL-6) or tumor necrosis factor alpha (TNF-α), for the activation of innate immune cells to exclude pathogens, which cause the malfunction of several tissues associated with an increase in the mortality rate (5). In the late phase, however, MDSCs express anti-inflammatory cytokines, suppress gamma interferon (IFN-γ) production by CD8⁺ T cells, and inhibit lipopolysaccharide (LPS)-induced proinflammatory cytokine release, thus contributing to immune paralysis (5–7). The MDSC subsets in the early phase of sepsis have been elucidated (8), but those in the late phase are unclear.

PD-1 is a co-inhibitory receptor expressed on the surface of activated T cells. PD-L1, a ligand of PD-1, is widely expressed on dendritic cells (DCs), macrophages, and nonhematopoietic cells in patients with sepsis (9). In several studies using different mouse models of sepsis, PD-1 expression was upregulated on T cells and monocytes (10, 11) and PD-L1 expression was upregulated on monocytes/macrophages (12, 13). PD-1/PD-L1 interactions suppress T-cell activation and cytokine production by downregulating T-cell receptor (TCR) signaling, resulting in immune suppression or tolerance (14). Anti-PD-1 antibody (Ab) treatment during the late phase of sepsis reactivates antigen-presenting cells and T cells, resulting in attenuation of secondary infection (15, 16). In contrast, anti-PD-L1 Ab treatment in the early phase of sepsis increases the survival rate by reversing monocyte dysfunction and inhibiting T-cell apoptosis (13). PD-1- or PD-L1-deficient mice also show resistance to sepsis-induced lethality and have a less intense inflammatory response due to the prevention of sepsis-induced macrophage dysfunction (11, 12). Higher PD-L1 expression on neutrophils in the early phase of sepsis is correlated with increased risk of death from sepsis as a result of elevated inflammatory cytokine expression (12). Taken together, these observations indicate that activation of the PD-1/PD-L1 pathway in the early phase of sepsis causes malfunction of innate immune cells. The PD-1/PD-L1 pathway in the early phase of sepsis may contribute to immune suppression by MDSCs during immune paralysis because MDSCs transition from a proinflammatory to an anti-inflammatory phenotype during the course of sepsis (5). In antitumor immunity, PD-1 regulates the metabolism-driven lineage fate commitment of myeloid progenitors and differentiation of effector myeloid cells (17). However, the relationship between the PD-1/PD-L1 pathway and the differentiation of MDSCs in sepsis is unclear.

The present study investigated whether myeloid-derived cells generated in the early phase of sepsis affect immune suppression in immune paralysis. Cecal ligation and puncture (CLP) is commonly used as an animal model of sepsis because it mimics the characteristics of human sepsis (18). CLP requires invasive procedures on the abdomen, which are accompanied by a risk of mortality during the progression of CLP-induced sepsis. To induce mild sepsis in a minimally invasive manner, we used zymosan (ZM), a glucan found on the surface of fungi, as a model of fungal sepsis.

RESULTS

Zymosan injection-induced immune paralysis model.

To induce sepsis, ZM was injected intraperitoneally (i.p.) into mice daily from day −4 to 0. The body weight of mice was reduced by 15% on day −3 but increased gradually to the same level as phosphate-buffered saline (PBS)-injected control mice by day 21 (Fig. 1A). The percentage of T cells decreased to a nadir on day 7 and was restored by day 21 (see Fig. S1 in the supplemental material). Therefore, the immunological status of ZM-treated mice was similar to that of control mice on day 21, because all immunological parameters were restored.

FIG 1.

Zymosan-induced immune paralysis model. Zymosan (ZM) or PBS (control [Cont]) was injected i.p. into WT mice from day −4 to day 0. (A) Body weight changes in PBS-treated control and ZM-treated mice. Body weight was monitored from day −5 to day 21. (B and C) Functions of splenic T cells in vitro. Splenic CD4⁺ T or CD8⁺ T cells were isolated from Cont or ZM mice on day 21. Violet-labeled cells were stimulated with immobilized anti-CD3ε MAb for 72 h. Next, cells were stained with APC–Cy7–anti-CD3ε and either PE–Cy7–anti-CD4, PerCP–Cy5.5–anti-CD8α, or APC–anti IFN-γ MAbs or the appropriate fluorochrome-conjugated control Ab and analyzed by flow cytometry. An electric gate was placed on CD3⁺ CD4⁺ T cells or CD3⁺ CD8⁺ T cells, and dilution of violet-labeled cells and the percentage of IFN-γ⁺ cells were assessed. The percentages of separated CD4⁺ T and CD8⁺ T cells (B) and IFN-γ⁺ among CD4⁺ T and CD8⁺ T cells (C) are shown. Representative FACS profiles are shown in the lower portions of panels B and C. (D) LPS was injected i.p. into control or ZM mice on day 21. Sera were collected 6 h after injection. Serum concentrations of IL-6 were measured by ELISA. Bars show means ± standard deviations (SD) for four mice. *, P < 0.05.

To confirm the immune paralysis in ZM-treated mice on day 21 (d21-ZM mice), T-cell proliferation and IFN-γ expression stimulated with anti-CD3ε MAb were assessed. CD4⁺ T and CD8⁺ T cells showed impaired cell division in ZM-treated mice (Fig. 1B). The percentage of IFN-γ⁺ cells among CD8⁺ T cells, but not CD4⁺ T cells, was significantly lower in ZM-treated mice (Fig. 1C). To mimic secondary bacterial infection in the late phase of sepsis, mice were injected with LPS and the serum level of IL-6 was measured. The IL-6 concentration was lower in ZM-treated mice than in controls following injection of LPS (Fig. 1D). These observations indicate that immune responses are impaired in d21-ZM mice.

Gr-1^hi cell populations expand after LPS injection and suppress LPS-induced IL-6 production.

The number of Gr-1^hi cells in the spleen increased due to sepsis until day 7 and then decreased to the level of the control group by day 21 (Fig. S1B). Gr-1^hi cells in d21-ZM mice showed higher forward scatter (FSC) and expressed CD11b⁺ (Fig. 2A); the major population of CD11b-gated cells consisted of Ly6G⁺ Ly6C^lo, a polymorphonuclear myeloid-derived cell (PMN-MDC) phenotype (Fig. 2A), and F4/80^dim/− (Fig. S2). In peripheral blood, the proportion of CD11b⁺ Gr-1^hi cells in d21-ZM mice was comparable to that in control mice. These cells showed more rapid expansion in the peripheral blood of d21-ZM mice than control mice at 6 h after injection of LPS (Fig. 2B). At 24 h after LPS injection, the proportion of CD11b⁺ Gr-1^hi cells in control mice was similar to that in the ZM-treated group (Fig. 2C). These results indicate that the number of CD11b⁺ Gr-1^hi cells in the blood increased in response to LPS in ZM-treated mice. Next, we examined whether CD11b⁺ Gr-1^hi cells induced by ZM are required for suppression of LPS-induced IL-6 production (Fig. 1D). CD11b⁺ cells were isolated using magnetic beads from d21-ZM or control mice and adoptively transferred to intact recipient mice. LPS was injected into recipient mice 1 day after cell transfer, and only mice with cell transfer from d21-ZM mice showed suppression of IL-6 production (Fig. 2D), suggesting that CD11b⁺ Gr-1^hi cells generated in d21-ZM mice suppress LPS-induced IL-6 production.

FIG 2.

ZM-induced Gr-1^hi cell populations expand and suppress IL-6 after LPS injection. (A) Splenocytes were isolated from ZM mice on day 21. Cells were stained with V500–anti-CD11b and either V450–anti-Gr-1, PE–Cy7–anti-Ly6C, or FITC–anti–Ly6G MAbs or the appropriate fluorochrome-conjugated control Ab. The stained cells were analyzed by flow cytometry. An electric gate was placed on FSC^hi Gr-1^hi cells, and the expression of CD11b was assessed (middle); Ly6C and Ly6G expression was assessed on FSC^hi CD11b⁺ cells (right). Representative profiles are shown. (B and C) Peripheral blood cells (PBLs) from control (Cont) or ZM mice on day 21 before or 6 h after LPS injection (B) or on day 21 at the indicated time points after LPS injection (C) were collected, and ACK lysis buffer was added to lyse red blood cells. PBLs were stained with APC-Cy7–anti-CD11b and V450–anti-Gr-1 MAbs or the appropriate fluorochrome-conjugated control Ab. The stained cells were analyzed by flow cytometry. The expression profiles of CD11b and Gr-1 were assessed. Representative profiles of the CD11b⁺ Gr-1^hi fractions or mean values ± SD for four mice are shown. (D) CD11b⁺ cells were isolated from control or ZM mice on day 21 and adoptively transferred into intact WT mice. One day after transfer, LPS was injected i.p. into nontransfer, control CD11b⁺-transfer, or ZM CD11b⁺-transfer recipient mice. Sera were collected at 6 h after injection. Serum concentrations of IL-6 were measured by ELISA. Bars show means ± SD for four mice. *, P < 0.05.

Immune-suppressive Gr-1^hi cells are generated in the early phase of ZM-induced sepsis.

Immature myeloid cells, which express Gr-1, are generated in the bone marrow in the early phase of sepsis and are converted into immune-suppressive cells under septic conditions (3). We investigated whether the elimination of Gr-1⁺ cells in the early phase of sepsis would reduce the immune-suppressive activity of Gr-1^hi cells in d21-ZM mice. Compared to controls, FSC^hi Gr-1^hi cells expanded into the peripheral blood after ZM injection but were depleted by anti-Gr-1 MAb (Fig. 3A). The inhibitory effects in Gr-1^hi cells from d21-ZM mice on T-cell proliferation (Fig. 3B), IFN-γ production by CD8⁺ T cells (Fig. 3C), and LPS-induced IL-6 production (Fig. 3D) were alleviated by anti-Gr-1 MAb. These data demonstrate that immune-suppressive Gr-1^hi cells induced by ZM injection are generated immediately after the induction of sepsis.

FIG 3.

Depletion of Gr-1⁺ cells during induction of sepsis breaks immune paralysis. Before PBS or ZM injection, anti-Gr-1 MAb (200 μg/mouse) was injected intraperitoneally. (A) PBLs from control (Cont), ZM, or anti-Gr-1 MAb-injected ZM mice were isolated on day −3 as described for Fig. 2B and stained with V450–anti-Gr-1 MAb or an appropriate fluorochrome-conjugated control Ab. The stained cells were analyzed by flow cytometry. Representative data are shown. (B and C) On day 21, splenic CD4⁺ T or CD8⁺ T cells were isolated from ZM mice with or without anti-Gr-1 MAb injection. Violet-labeled cells were stimulated with immobilized anti-CD3ε MAb for 72 h as for Fig. 1. After incubation, T cells were stained with APC–Cy7–anti-CD3ε and either PE–Cy7–anti-CD4, PerCP–Cy5.5–anti-CD8α, or APC–anti-IFN-γ MAbs or the appropriate fluorochrome-conjugated control Ab and analyzed by flow cytometry. An electric gate was placed on CD3⁺ CD4⁺ for CD4⁺ T cells or CD3⁺ CD8⁺ for CD8⁺ T cells, and dilution of violet-labeled cells and the percentage of IFN-γ⁺ cells were assessed. Percentages of separated CD4⁺ T and CD8⁺ T cells (B) and IFN-γ⁺ among CD8⁺ T cells (C) are shown. Representative FACS profiles are shown in the lower portion of panel B. Bars show means ± SD for four mice. (D) On day 21, sera from ZM mice with or without anti-Gr-1 MAb injection were collected at 6 h after LPS injection. Serum concentrations of IL-6 were measured by ELISA. Data are representative of two independent experiments with similar results. *, P < 0.05.

Gr-1^hi cells in d21-ZM mice lacking PD-1/PD-L1 signaling in the early phase of sepsis lose the effect on immune paralysis.

When we injected ZM into PD-1/PD-L1 double-knockout mice (KO mice), the inhibition ratio of LPS-induced IL-6 production was lower than that in wild-type (WT) mice (Fig. 4A). To examine the suppression of IL-6 production by PD-1- and PD-L1-deficient Gr-1^hi cells in d21-ZM mice, Gr-1^hi cells isolated from d21-ZM KO mice were adoptively transferred into intact WT mice. LPS-induced IL-6 production was not inhibited 1 day after cell transfer (Fig. 4B), suggesting that Gr-1^hi cells from d21-ZM mice lacking PD-1/PD-L1 signaling do not suppress IL-6 production. We detected transient enhancement of PD-L1 expression in CD11b⁺ Gr-1^hi cells in the bone marrow (BM) and spleen at 6 h after the first injection of ZM (Fig. 4C). To examine the importance of the PD-1/PD-L1 pathway for generating ZM-induced Gr-1^hi cells in the early phase of sepsis, we used blocking antibodies against PD-1 and PD-L1. This restored T-cell proliferation (Fig. 4D), IFN-γ production by CD8⁺ T cells (Fig. 4E), and LPS-induced IL-6 production (Fig. 4F). Taken together, these results suggest that activation of the PD-1/PD-L1 pathway during sepsis induction is required to generate immune-suppressive Gr-1^hi cells in d21-ZM mice.

FIG 4.

PD-1/PD-L1 pathway blockade during sepsis induction breaks immune paralysis. (A and B) Serum levels of IL-6 after LPS injection. On day 21, sera from WT or PD-1/PD-L1-deficient (KO) mice treated with PBS (control [Cont]) or ZM were collected at 6 h after LPS injection (A). CD11b⁺ Gr-1^hi cells were isolated from control KO mice or ZM KO mice on day 21 and were transferred into intact WT mice. One day after transfer, sera from recipient mice were collected at 6 h after LPS injection (B). Serum concentrations of IL-6 were measured by ELISA. Bars show means ± SD for four mice. (C) Bone marrow cells and splenocytes were isolated from control or ZM mice at 6 h after first PBS or ZM injection, respectively. Cells were stained with APC–Cy7–anti-CD11b, V450–anti-Gr-1, and PE–anti-PD-L1 MAbs or the appropriate fluorochrome-conjugated control Ab and analyzed by flow cytometry. An electric gate was placed on CD11b⁺ Gr-1^hi cells. Data are representative of two independent experiments with similar results. (D and E) On day 21, splenic CD4⁺ T or CD8⁺ T cells were isolated from ZM mice with or without anti-PD-1 and anti-PD-L1 MAb injection. Violet-labeled cells were stimulated with immobilized anti-CD3ε MAb for 72 h as for Fig. 1. After incubation, T cells were stained with APC–Cy7–anti-CD3ε MAb and either PE-Cy7–anti-CD4, PerCP–Cy5.5–anti-CD8α, or APC–anti-IFN-γ MAb or the appropriate fluorochrome-conjugated control Ab and analyzed by flow cytometry. An electric gate was placed on CD3⁺ CD4⁺ for CD4⁺ T cells or CD3⁺ CD8⁺ for CD8⁺ T cells, and the dilution of violet-labeled cells and the percentage of IFN-γ⁺ cells were assessed. The percentages of separated CD4⁺ T and CD8⁺ T cells (D), and IFN-γ⁺ among CD8⁺ T cells (E) are shown. Bars show means ± SD for four mice. (F) On day 21, sera from ZM mice with or without anti-PD-1 and anti-PD-L1 MAb injection were collected at 6 h after LPS injection. Serum concentrations of IL-6 were measured by ELISA. Bars show means ± SD from two independent experiments with similar results. *, P < 0.05.

DISCUSSION

We found that bone marrow-derived cells related to immune paralysis are generated in the early phase of sepsis, during which time the PD-1/PD-L1 pathway is required to generate ZM-induced immune-suppressive Gr-1^hi cells, resulting in immune paralysis.

In our model, involving serial injection of low-dose ZM, Gr-1^hi cells accumulated in the spleen after 10 days but decreased and reached the basal level by day 21. T cells isolated from d21-ZM mice showed impaired in proliferation and IFN-γ production. The number of T cells was transiently reduced but was restored by day 21 (Fig. S1). After sepsis-induced apoptosis of T cells, homeostatic proliferation of the T cells occurs in a lymphopenic environment together with changes in the composition of the T-cell compartment (19). During proliferation, expression of inhibitory receptors, such as 2B4 and PD-1, is increased on surviving cells, leading to increased mortality and dysfunction of T cells (20, 21). By interacting with the receptor, immune-suppressive Gr-1^hi cells or their precursor cells generated in the early phase of sepsis modulate the function of proliferating T cells, resulting in suppression of T-cell function in immune paralysis.

After stimulation by LPS, the IL-6 level in serum was reduced in ZM-treated mice, and this reduction was accompanied by accumulation of Gr-1^hi cells in the peripheral blood in d21-ZM mice. Transfer of Gr-1^hi cells from d21-ZM mice into intact WT mice also suppressed IL-6 production. In sepsis, inflammatory cytokine production (IL-6, IL-1β, and TNF-α) is important to activate the innate immunity to exclude microorganisms from the host. In immune paralysis, however, the innate responses of myeloid-derived cells, such as inflammatory-cytokine production, to secondary infection are impaired (22). IL-6 production by LPS-stimulated macrophages was reduced by coculture with sepsis-induced MDSCs (7), suggesting that Gr-1^hi cells in d21-ZM mice gain the ability to suppress IL-6 production or consume IL-6 secreted by macrophages during the course of sepsis. Under septic conditions, populations of Gr-1⁺ immature myeloid cells are expanded and converted to immune-suppressive MDSCs by epigenetic modification, egress from the bone marrow (BM) into the peripheral blood, and accumulate in lymphoid organs (3, 4). Our results indicated that the depletion of Gr-1^hi myeloid cells in the early phase of sepsis inhibits the expansion and accumulation of immune-suppressive Gr-1^hi cells in lymphoid organs in ZM-treated mice, alleviating immune paralysis.

We demonstrated that the activation of the PD-1/PD-L1 pathway in the early phase of sepsis is important for differentiation of Gr-1^hi cells in d21-ZM mice. In our model, PD-L1 expression in myeloid cells in the BM or spleen was transiently upregulated just after ZM injection. STAT3 is a key transcription factor for expansion and accumulation of MDSCs (23) and is activated by IL-6 produced in response to sepsis. The bacillus Calmette-Guérin tuberculosis vaccine induces PD-L1 expression on DCs or macrophages via the IL-6–STAT3 pathway (24). As we detected IL-6 production at 6 h after the first ZM injection (data not shown), we speculated that after enhancement of PD-L1 expression on the precursors of immune-suppressive Gr-1^hi cells by ZM-induced IL-6 via STAT3, the interaction of PD-L1 with PD-1 on other cells promoted the differentiation of Gr-1^hi cells in d21-ZM mice. A recent study showed that PD-L1 expression in PMN-MDSC from BM and spleens of CLP mice was higher in the early phase of sepsis than in sham controls (8). In addition, PMN-MDSCs isolated from CLP mice at 1 day postsurgery, which have elevated PD-L1 expression, have the ability to inhibit T-cell proliferation in vitro. These results indicate that immune-suppressive myeloid cells are induced in the early phase of sepsis via the PD-1/PD-L1 pathway.

We cannot rule out a role in immune paralysis of PD-1 on Gr-1^hi cells in d21-ZM mice. The interaction between PD-L1 on antigen-presenting cells (APCs) and PD-1 on T cells suppressed T-cell functions, and blocking antibodies against PD-1 or PD-L1 in the late phase of sepsis improved the function of T cells or APCs, respectively, in fungal sepsis (16). As well as activated T cells, PD-1 is expressed on other immune cells, such as myeloid cells (25). PD-1 expression was also induced on CD11b⁺ Gr-1^hi cells within 6 h after the first ZM injection (Fig. S3). PD-1 expression in macrophages suppresses the innate response to sepsis and inhibits phagocytosis of Mycobacterium tuberculosis in active tuberculosis (11, 26). PD-L1 on T cells engaged PD-1-expressing macrophages to induce tolerogenic macrophages (27). Such back-signaling via the PD-1/PD-L1 pathway may regulate the differentiation of ZM-induced immune-suppressive Gr-1^hi cells in immune paralysis.

In summary, the activation of the PD-1/PD-L1 pathway in the early phase of sepsis generates immune-suppressive myeloid cells. CD11b⁺ Gr-1^hi cells show therapeutic promise for immune paralysis.

MATERIALS AND METHODS

Mice.

Seven- to 8-week-old female wild-type (WT) BALB/c mice were purchased from Japan SLC (Hamamatsu, Japan). PD-1/PD-L1 double-knockout (KO) mice in the BALB/c background were generated as described previously (28). All mice were maintained under specific-pathogen-free (SPF) conditions at the Animal Facility of Tokyo Medical and Dental University. All animal procedures were reviewed and approved by the Institutional Animal Care and Use Committee of Tokyo Medical and Dental University (A2019-206C, G2019-014C5).

Zymosan-induced immune paralysis.

Two milligrams of zymosan A (Sigma-Aldrich, St. Louis, MO) in 0.5 ml Dulbecco’s phosphate-buffered saline (PBS) (Sigma-Aldrich, St. Louis, MO) was injected intraperitoneally (i.p.) into mice daily from day −4 to day 0. Mice in the control group received the same volume of PBS alone. To stimulate IL-6 production, 21 days after the last injection, 20 μg of lipopolysaccharide (LPS; Sigma-Aldrich) in PBS was injected i.p. to mimic a secondary Gram-negative bacterial infection.

Measurement of serum concentration of IL-6.

Serum samples were collected before and 6 h after LPS injection, and the levels of IL-6 were measured using an enzyme-linked immunosorbent assay (ELISA) (eBioscience, San Diego, CA).

Antibodies and flow cytometry.

Monoclonal antibodies (MAbs) against CD3ε (145-2C11), CD4 (RM4-5), CD8 (53-6.7), CD11b (M1/70), CD11c (N418), CD49b (DX5), CD45R (RA3-6B2), major histocompatibility complex (MHC) class II (M5/114.15.2), F4/80 (BM8), Gr-1 (RB6-8C5), Ly6G (1A8-Ly6g), Ly6C (HK1.4), interleukin 6 (IL-6) (MP5-20F3), and IFN-γ (XMG1.2) were used. All fluorescein isothiocyanate (FITC)-, phycoerythrin (PE)-, allophycocyanin-, peridinin chlorophyll protein (PerCP)-carbocyanine 5.5 (Cy5.5)-, PE-Cy7-, allophycocyanin-eFluor780-, V450-, and V500-conjugated MAbs, biotinylated MAbs, and allophycocyanin-eFluor780-streptavidin were obtained from eBioscience, BD-Pharmingen (San Diego, CA), or BioLegend (San Diego, CA). 2.4G2 hybridoma culture supernatant was used to block nonspecific binding via FcγR. Multicolor staining for intracellular cytokines and cell surface antigens was performed as described previously (29). Stained cells were analyzed with a FACSVerse system (BD Biosciences) with FACSuite software (BD Biosciences). Data were analyzed using FlowJo software (Tree Star, Ashland, OR).

Measurement of T-cell proliferation and cytokine production.

CD4⁺ and CD8⁺ T cells were isolated from PBS- and ZM-treated mice by negative selection, respectively. Isolated T cells (1 × 10⁶ cells/well) were labeled using a CellTrace Violet cell proliferation kit (Invitrogen, Carlsbad, CA), stimulated with an immobilized anti-CD3ε (0.3 μg/ml) MAb, and cultured for 72 h. To detect IFN-γ-producing cells, cells were further incubated with 5 μg/ml brefeldin A (Sigma-Aldrich) for 6 h, and cell division and IFN-γ expression were analyzed by flow cytometry.

Isolation of peripheral blood leukocytes.

Peripheral blood was collected in heparinized hematocrit tubes (Drummond Scientific Company, Broomall, PA). Peripheral blood leukocytes (PBLs) were purified after lysing red blood cells with ammonium-chloride-potassium (ACK) lysis buffer.

Adoptive transfer of CD11b⁺ cells.

After preparation of a single-cell suspension of splenocytes, CD11b⁺ cells were isolated using CD11b microbeads and magnetic separation on a magnetic-activated cell sorting (MACS) system (Miltenyi Biotec Inc., San Diego, CA) according to the manufacturer’s protocol. The CD11b⁺ Gr-1^hi fraction was confirmed by flow cytometry to have a purity of >95%. The sorted cells were injected intravenously (i.v.) into intact mice to evaluate the effects on LPS-induced IL-6 production.

MAb treatment in vivo.

To deplete Gr-1⁺ cells, 200 μg anti-Gr-1 MAb (RB6-8C5; rat IgG2b) was injected i.p. on days −4 and 0. To block the PD-1/PD-L1 pathway; 200 μg anti-PD-1 MAb (RPM1-14; rat IgG2a) and 200 μg anti-PD-L1 MAb (MIH-5; rat IgG2a) were injected i.p. on day −5. PBS was used as the vehicle control.

Statistical analyses.

Statistical analyses were performed using Prism 6 software (GraphPad Software, San Diego, CA) with the Mann-Whitney U test or two-way analysis of variance (ANOVA). In all analyses, a P value of <0.05 was taken to indicate statistical significance.