Bone-Marrow-Resident NK Cells Prime Monocytes for Regulatory Function during Infection

Abstract

Tissue-infiltrating Ly6C(hi) monocytes play diverse roles in immunity, ranging from pathogen killing to immune regulation. How and where this diversity of function is imposed remains poorly understood. Here we show that during acute gastrointestinal infection, priming of monocytes for regulatory function preceded systemic inflammation and was initiated prior to bone marrow egress. Notably, natural killer (NK) cell-derived IFN-γ promoted a regulatory program in monocyte progenitors during development. Early bone marrow NK cell activation was controlled by systemic interleukin-12 (IL-12) produced by Batf3-dependent dendritic cells (DCs) in the mucosal-associated lymphoid tissue (MALT). This work challenges the paradigm that monocyte function is dominantly imposed by local signals after tissue recruitment, and instead proposes a sequential model of differentiation in which monocytes are pre-emptively educated during development in the bone marrow to promote their tissue-specific function.

Figure 1. Monocytes are educated prior to tissue entry during infection

CX3CR1-GFP mice were infected per-orally with T. gondii. (A) Flow-cytometric analysis of Ly6C^hi monocytes (Mo) present in the small intestinal lamina propria (SILP) of naïve animals or at day 8 post-infection (p.i.). (B) Phenotype of blood Ly6C^hi monocytes at defined time-points p.i. as measured by flow cytometry. (C) Frequency of blood Ly6C^hi monocytes within the total blood monocyte compartment. (D) Ex vivo PGE₂ production by Ly6C^hi monocytes sorted from the blood of naïve or day 8 infected mice and cultured in the presence or absence of E. coli lysate. (E) IL-10 and TNF-α production by blood Ly6C^hi monocytes from naïve and day 8 T. gondii infected animals cultured in the presence or absence of E. coli lysate. (F) Accumulation of monocytes in the SILP. (G) Serum levels of IFN-γ and TNF-α in infected mice. Dots represent individual animals. Error bars represent standard deviation. Data are representative of at least two independent experiments, n = 3-6 per group. Statistical comparisons were performed using one-way ANOVA or unpaired student's t test adjusted for multiple comparisons. **: p<0.01, ***: p<0.001. See also Figure S1.

Figure 2. IFN-γ remodels the blood monocyte compartment

(A-B) CX3CR1-GFP mice were infected intravenously with P. yoelii infected red blood cells, per-orally with Y. pseudotuberculosis, or per-orally with T. gondii. (A) Phenotype of blood Ly6C^hi monocytes (Mo) following infection with P. yoelli (day 4 p.i.), Y. pseudotuberculosis (day 5 p.i.) and T. gondii (day 8 p.i.). (B) Frequency of blood Ly6C^hi monocytes within the total monocyte compartment at time-points described in (A). (C-D) CX3CR1-GFP mice were administered IFN-γ or PBS once per day for three consecutive days. (C) Phenotype of blood Ly6C^hi monocytes assessed by flow cytometry. (D) Frequency of Ly6C^hi blood monocytes within the total monocyte compartment. (E-F) CX3CR1-GFP mice were infected with T. gondii and treated with anti-IFN-γ Ab or isotype control (IgG). (E) Phenotype of blood Ly6C^hi monocytes. (F) Frequency of Ly6C^hi monocytes within the total blood monocyte compartment. (G-H) Chimeric mice comprised of equal numbers of WT CD45.1⁺ and Ifngr1^−/− (CD45.2⁺) leukocytes were infected or not with T. gondii. (G) Mean fluorescent intensity (MFI) of MHCII and Sca-1 expression by WT and Ifngr1^−/− blood Ly6C^hi monocytes measured by flow cytometry at 8 days p.i. Values of WT and Ifngr1^−/− cells from the same host are joined by a line. (H) Proportion of blood Ly6C^hi monocytes in total blood monocytes within WT and Ifngr1^−/− compartments in naïve and infected mice. Error bars represent standard deviation. Data are representative of at least two independent experiments, n = 3-5 per group. Statistical comparisons were performed using unpaired student's t test, or paired t test adjusted for multiple comparisons. **: p<0.01, ***: p<0.001.

Figure 3. IFN-γ primes BM monocytes for regulatory function during infection

(A) Ex vivo PGE₂ production by Ly6C^hi monocytes sorted from BM of naïve or T. gondii infected mice at day 5 p.i. and cultured with or without E. coli lysate, Soluble Toxoplsama antigen (STAg), or various TLR ligands. (B) Ex vivo PGE₂ production by BM Ly6C^hi monocytes sorted from naïve or infected mice at day 24 p.i. and stimulated with E. coli lysate. (C) WT mice were infected as in (A) and treated with anti-IFN-γ Ab or isotype control. Ly6C^hi monocytes were sorted from BM of naïve and infected mice and assessed ex vivo for PGE₂ production in response to E. coli lysate. (D) Ex vivo PGE₂ production by Ly6C^hi monocytes sorted from BM of CX3CR1-GFP mice treated with IFN-γ or PBS once per day for three consecutive days. (E) Ex vivo PGE₂ production by BM Ly6C^hi monocytes sorted from naïve WT mice, cultured for 6 hours with or without IFN-γ, and subsequently stimulated with E. coli lysate. (F-G) WT mice were infected as in (A) and BM Ly6C^hi monocytes were sorted from naïve or day 5 infected mice. Sorted monocytes were cultured for 6 hours in the presence or absence of LPS, and gene expression assessed using the NanoString platform. (F) Principle component analysis comparing gene expression of untreated and LPS stimulated monocytes from naïve and infected mice. Plot represents clustering of samples in a 2-dimensional matrix of principle components 1 and 2. (G) Gene expression of untreated naïve monocytes from (C) was compared to each of the other three groups. Numbers in each section of the Venn diagram represent the number of genes altered in expression between that group and untreated naïve controls, with overlapping regions representing genes changed in more than one group. (H) Heat maps representing the relative expression of the 30 genes changed only in monocytes from infected mice upon LPS stimulation. Error bars represent standard deviation. Data are representative of two or more independent experiments, n = 3-5 replicates per group (A-E). Columns represent biological replicates of 2 pooled samples (H). Statistical comparisons were performed using unpaired student's t test, adjusted for multiple comparisons (A-E) or by Welch's t test adjusted for multiple comparisons (F-H). *: p<0.05, **: p<0.01, ***: p<0.001.

Figure 4. IFN-γ controls gene expression of monocyte progenitors early during infection

(A) CX3CR1-GFP mice were infected with T. gondii. At defined time-points p.i., the phenotype of monocyte progenitors (cMoP) present in the BM was assessed by flow cytometry. (B) cMoP were sorted from the BM of naïve or day 5 infected mice, and gene expression was assessed by NanoString. Heat maps represent the relative expression of 37 genes by cMoP selected by pathway analysis as being affected by IFN-γ signalling from naïve and infected mice. (C) CX3CR1-GFP mice were administered 5 μg IFN-γ or PBS by i.p. injection once per day for three consecutive days. Mice were sacrificed eighteen hours after the final injection, and MHCII and Sca-1 expression by cMoP was assessed. (D) CX3CR1-GFP mice were infected and administered anti-IFN-γ Ab or isotype control. Expression of MHCII and Sca-1 was assessed at day 5 p.i. Error bars represent standard deviation. Data are representative of two or more independent experiments, n = 3-5 mice per group (A, C, D). Columns represent individual samples pooled from 3 mice (B). Statistical comparisons were performed using one-way ANOVA (A, D), Welch's t test adjusted for multiple comparisons (B), or unpaired students t test (C). ***: p<0.001. See also Figures S2, S3, and S4.

Figure 5. BM NK cells educate cMoP and Ly6C^hi monocytes via IFN-γ

(A) WT mice were infected with T. gondii. At defined time-points, lymphocyte IFN-γ production was assessed by flow cytometry. Plots are gated on TCR-β⁻NK1.1⁺DX5⁺ NK cells. (B) Absolute numbers of lymphocytes producing IFN-γ in the BM. (C-D) Confocal images of BM of IFN-γ YFP mice at day 5 p.i. (C) Yellow arrows indicate NK cells producing IFN-γ. (D) Ly6G⁻Ly6B2⁺ monocytes in contact with (yellow arrows) or in close proximity to (orange arrows) NK cells producing IFN-γ. (E) CX3CR1-GFP mice were infected as in (A) and injected intravenously with NK cell depleting antibody (αNK1.1) or isotype control. MHCII and Sca-1 expression by cMoP were assessed at day 5 p.i. Error bars represent standard deviation. Data are representative of two or more independent experiments, n = 3-5 mice per group. Statistical comparisons were performed using oneway ANOVA. **: p<0.01, ***: p<0.001. See also Figure S5.

Figure 6. IL-12 produced by Batf3-dependent DC in the MALT stimulates NK cells to secrete IFN-γ

(A) Serum IL-12p70 levels post-infection. (B-C) WT mice were infected with T. gondii and treated with anti-IL-12p70 Ab or isotype control. (B) Absolute numbers of IFN-γ producing lymphocytes at day 5 p.i. (C) MHCII and Sca-1 expression by cMoP at day 5 p.i. (D-E) WT mice were administered IL-12 or PBS once per day for two days. (D) Absolute numbers of IFN-γ producing lymphocytes in the BM. (E) MHCII and Sca-1 expression by cMoP. (F) Batf3^−/− mice and WT controls were infected and CD8α⁺ and CD11b⁺ resident DC in the mesLN were assessed at day 5 p.i. Dot plots are gated on TCR-β⁻CD11c⁺MHCII^int resident DC. Numbers represent mean frequency of cells gated in this manner +/− standard deviation. (G) Serum IL-12p70 in Batf3^−/− and WT mice, naïve and day 5 p.i.. (H) Absolute numbers of IFN-γ producing lymphocytes in Batf3^−/− mice and WT controls at day 5 p.i. (I) MHCII expression by cMoP in Batf3^−/− mice and WT controls at day 5 p.i. (J) Ten percent of the total cells from each tissue were cultured for 16 hours and cell supernatants were assessed for IL-12p70. Error bars represent one standard deviation. Data are representative of two or more independent experiments (A-I) or the pooled results of 2 independent experiments (J), n = 3-5 per group. Statistical comparisons were performed using one-way ANOVA or unpaired student's t test, corrected for multiple comparisons. *: p < 0.05, **: p<0.01, ***: p<0.001. See also Figure S6.