Type I and Type II Interferon Coordinately Regulate Suppressive Dendritic Cell Fate and Function during Viral Persistence

Abstract

Persistent viral infections are simultaneously associated with chronic inflammation and highly potent immunosuppressive programs mediated by IL-10 and PDL1 that attenuate antiviral T cell responses. Inhibiting these suppressive signals enhances T cell function to control persistent infection; yet, the underlying signals and mechanisms that program immunosuppressive cell fates and functions are not well understood. Herein, we use lymphocytic choriomeningitis virus infection (LCMV) to demonstrate that the induction and functional programming of immunosuppressive dendritic cells (DCs) during viral persistence are separable mechanisms programmed by factors primarily considered pro-inflammatory. IFNγ first induces the de novo development of naive monocytes into DCs with immunosuppressive potential. Type I interferon (IFN-I) then directly targets these newly generated DCs to program their potent T cell immunosuppressive functions while simultaneously inhibiting conventional DCs with T cell stimulating capacity. These mechanisms of monocyte conversion are constant throughout persistent infection, establishing a system to continuously interpret and shape the immunologic environment. MyD88 signaling was required for the differentiation of suppressive DCs, whereas inhibition of stimulatory DCs was dependent on MAVS signaling, demonstrating a bifurcation in the pathogen recognition pathways that promote distinct elements of IFN-I mediated immunosuppression. Further, a similar suppressive DC origin and differentiation was also observed in Mycobacterium tuberculosis infection, HIV infection and cancer. Ultimately, targeting the underlying mechanisms that induce immunosuppression could simultaneously prevent multiple suppressive signals to further restore T cell function and control persistent infections.

Fig 1. In vivo localization and identification of immunoregulatory DCs during viral persistence.

A. Sections from IL-10 reporter mice infected with LCMV-Cl13 for 9 or 36 days were stained for B cells, CD4+ T cells, and CD90.1 (IL-10) and visualized at 10x magnification. B. CD39 and CD95 expression on splenic CD11b+ DCs from naive IL-10 reporter mice or IL-10 reporter mice infected with LCMV-Cl13 for 9 or 30 days and their corresponding expression of CD90.1 (IL-10) and PDL1 within the iregDC (red) and stimDC (blue) populations. Bar graphs indicate the geometric mean fluorescence intensity (MFI) of CD90.1 (IL-10) and PDL1 expression by iregDCs (red) and stimDCs (blue). DC are characterized as being viability dye-, CD45+, Thy1.2-, NK1.1-, Ly6G-, CD11c++ (high), CD11b+. C. The number of iregDCs (red) and stimDCs (blue) based on CD39 and CD95 expression at the indicated time point after LCMV-Cl13 infection. D. Histograms of the indicated protein on iregDCs (red) and stimDCs (blue) at day 9 and day 30 following LCMV-Cl13 infection. E. iregDCs and stimDCs were sorted from splenocytes at day 9 of LCMV-Cl13 infection and cultured with LCMV specific CD4+ T cells (SMARTA) for 3 days with IL-10R blocking antibody, PDL1 blocking antibody, or media alone. Bar graph represents the proportion of proliferated SMARTA cells after the culture. Data in 1E show a single experiment using iregDC and stimDC sorted from a pool of 8 mice in order to obtain adequate numbers of each population. Data are representative of 2 or more independent experiments each consisting of 3–4 mice per group. *, p<0.05.

Fig 2. iregDCs have a distinct expression profile compared to stimDCs.

StimDCs and iregDCs were obtained from day 9 LCMV-Cl13 infected mouse splenocytes, RNA was harvested and subjected to RNA-seq. A. Graph scatter shows the log-transformed expression values for genes expressed at 2 RPKM or more in at least one sample. Dots are color-coded based on the ratio of RPKM values to indicate differentially expressed (DE) genes between the two cell types. The number of genes in each DE group are indicated in the top left (higher in stimDCs) and bottom right (higher in iregDCs) corners. Selected DE genes from this dataset are shown in the rest of the panels to highlight important gene groups as indicated on top of each panel. B. Gene Ontology (GO) term over representation analysis was performed on a set of genes that were at least 5-fold differentially expressed between the iregDC and stimDC samples. Plots show the statistical significance of GO terms that are enriched in genes expressed at higher levels in iregDCs and stimDCs. The number of genes that were classified under a particular GO term are indicated in brackets. GO terms with an adjusted (Benjamini-Hochberg method) p-value of < = 0.05 were considered to be enriched in the DE gene set. C. Bar graphs show the fold change in the indicated gene. Up on the y-axis indicates increase in iregDC (red). Down on the y-axis indicates increase in stimDC (green). The number in parenthesis indicates the RPKM value of the highest sample (to indicate relative expression level).

Fig 3. Two-pronged mechanism of IFN-I mediated immunosuppression: IFN-I programs iregDC suppressive functions and limits stimDC generation.

A. CD39 and CD95 expression by splenic CD11b+ DCs from WT and IFNαR-/- mice at day 9 after LCMV-Cl13 infection. Bar graphs indicate the number of iregDC (CD39+, CD95+), stimDC, and total splenocytes. B. Histogram shows PDL1 expression on iregDCs from WT (grey) and IFNαR-/- (white) mice. Bar graphs show quantification of PDL1 expression (MFI) and IL-10 RNA by iregDC from WT and IFNαR-/- mice at day 9 of LCMV-Cl13 infection. To obtain sufficient cell numbers for the IL-10 RNA analysis, iregDC and stimDC were sorted from a pool of 8 mice. Data for IL-10 RNA analysis show one of two experiments. C. StimDC and iregDC were FACSorted from WT or IFNR-/- mice 9 days after LCMV-Cl13 infection and co-cultured with CFSE labeled naïve LCMV-specific CD4 SMARTA T cells. No peptide was added to the cultures. Flow plots show CFSE dilution by the virus-specific CD4 T cells 5 days after co-culture. Numbers in the flow plot indicate the percent of virus-specific CD4 T cells that have diluted CFSE. The bar graph indicates the percent of virus-specific CD4 T cells that have undergone 3 or more divisions (gated on divided cells), calculated as in [21]. D. Flow plots show reconstitution of CD11b+ DCs from WT CD45.1+) and IFNαR-/- CD45.2+) lineage cells in bone marrow chimera pre-infection (blood– 9 weeks post reconstitution) and post-infection (spleen—day 9 of LCMV-Cl13 infection). Bar graph shows quantification of the percent of lineage derived CD11b+ DCs. E. Flow plots show iregDC and stimDC gated on splenic CD11b+ DCs at day 9 after LCMV-Cl13 infection of the WT: IFNαR-/- bone marrow chimera mice. Flow plots are gated on the WT and IFNαR-/- lineage from the same mouse. Histogram shows PDL1 expression on iregDCs from WT (grey) and IFNαR-/- (white) lineage cells in the same chimeric mice. Bar graph indicates the total number of iregDC and stimDC and PDL1 MFI from each lineage in the chimeric mice. F. Bar graph shows IL-10 expression in the supernatant of sorted CD11b+ DC after stimulation with media alone or IFNβ. Supernatants were harvested 24 hours after culture. Data are representative of 2 or more independent experiments each consisting of 3–4 mice per group. *, p<0.05. For the WT vs IFNR-/- DC co-culture with T cells (Fig 3C), data are shown from 1 of 2 experiments each using samples pooled from 10–15 mice per DC group to obtain adequate numbers of each DC population.

Fig 4. Suppressive iregDC are of monocyte origin.

A. Flow plots show CCR2 and FcγRI expression on splenic CD11b+ DCs on day 9 after LCMV-Cl13 infection. CCR2+, FcγRI+ (red) and CCR2-, FcγRI- (blue) populations are subsequently plotted on iregDC markers CD95 and CD39. B. Expression of FcεR1α, FcγRII/III, and Ly6C on iregDCs (red) and stimDCs (blue) gated on CD95 and CD39 from CD11b+ DCs on day 9 after LCMV-Cl13 infection. C. Dendogram representing the gene expression profile relation between naïve monocytes, naïve DCs, iregDCs and stimDCs at day 9 of LCMV-Cl13 infection. Bar graph of the expression of key monocyte-associated (red) and conventional DC-associated (green) genes. Numbers in parenthesis indicate the RPKM value for the highest sample. D. Left flow plots depict CCR2 and FcγRI expression on splenic CD11b+ DCs at day 25 after LCMV-Cl13 infection. Right flow plots depict CD39 and CD95 expression gated on CCR2+, FcγRI+ (red) and CCR2-, FcγRI- (blue) populations. E. Monocytes were isolated from naïve WT CD45.1+ mice and transferred into WT CD45.2+ mice that had been infected for 3 days with LCMV-Cl13. Left flow plots show DC and iregDC marker expression on isolated monocytes prior to transfer (gated on CCR2+, Ly6C+, FcγRI+ cells). Right flow plots depict transferred monocytes on day 6 post-transfer (day 9 after infection; gated on CD45.1+, CCR2 +, FcγRI+, CD11b+ cells). Histogram represents the expression of PDL1 of the transferred monocytes based on iregDC (red) and stimDC (blue) gates. F. Naïve CD45.1+ monocytes were transferred into WT CD45.2+ mice that had been infected with LCMV-Cl13 for 19 days. Six days later (day 25 after infection), DC and iregDC differentiation were analyzed in the spleen. Flow plots are gated on the transferred CD45.1+, CCR2+, FcγRI+ cells and then into the CD11c high (DC) and CD11c- populations. The histogram shows the expression of PDL1 on the CD11c high (shaded grey) and CD11c- (solid black line) monocyte derived subsets 6 days after transfer. G. Naïve WT or IFNαR-/- monocytes were transferred into LCMV-Cl13 infected WT mice as in E. Flow plots show CD39 and CD95 expression on transferred monocytes in the spleen on day 9 after LCMV-Cl13 infection (gated on CCR2+, FcγRI+, CD11c++, CD11b+ DC). Bar graph indicates PDL1 expression on iregDCs derived from the transferred monocytes. Data are representative of 2 or more independent experiments each consisting of 3–4 mice per group. *, p<0.05. RNA-seq is from 1 experiment using samples each pooled from 10 mice per group.

Fig 5. iregDC extrinsic MyD88 signaling, but not direct LCMV infection or IFNβ signaling is required for iregDC differentiation.

A. WT mice were CD8 depleted and then infected with LCMV-Cl13-GFP. Flow plots show GFP expression (i.e., LCMV infection) in splenic iregDC and stimDC (gated on CD39 and CD95) on day 9 after infection. In parallel, naïve splenic DC (N), and iregDC and stimDC were FACSorted based on GFP expression on day 9 after infection. The level of GFP+ stimDC was too small to obtain adequate numbers for RNA isolation. Bar graphs show relative expression of IL-10 RNA, LCMV-GP RNA and the MFI of PDL1 protein expression from the sorted populations. To obtain sufficient cell numbers for the RNA analyses, infected and uninfected iregDC and stimDC were sorted from a pool of 8 mice. Data show one of two experiments. B. Flow plots show the percent and bar graphs the number of iregDC and stimDC in the indicated mouse strain from CD11b+ DCs at day 9 after LCMV-Cl13 infection. C. Bar graphs indicate plasma IL-10 concentration and PDL1 expression in WT, IFNαR-/-, and MyD88-/- mice at day 9 after LCMV-Cl13 infection. D. Naïve WT and MyD88-/- monocytes were transferred into WT mice and analyzed on day 9 after LCMV-Cl13 infection as described in Fig 4E. E. Mixed chimera mice were generated using bone marrow from WT and MyD88-/- bone marrow. Flow plots depict WT and MyD88-/- lineage splenic CD11b+ DCs at day 9 after LCMV-Cl13 of WT. Histogram shows PDL1 expression on iregDCs from WT (white) and MyD88-/- (grey) lineage cells. Bar graphs indicate PDL1 expression and the number of iregDC and stimDC from each lineage in the chimera mice. Data are representative of 2 or more independent experiments each consisting of 3–5 mice per group. *, p<0.05.

Fig 6. Monocytes require direct IFNγR signaling for DC differentiation throughout persistent virus infection.

A. Plasma IFNγ concentration in WT and MyD88-/- (KO) mice at day 9 after LCMV-Cl13 infection. B. Flow plots represent the percent of moDC, iregDC and stimDC (gated on total CD11b+ CD11c+ DC) on day 9 in LCMV-Cl13 infected WT and IFNγR-/- mice. Bar graphs indicate the number of moDC, iregDC, stimDC, total CD11b+ DC, total splenocytes and plasma IL-10 concentration in WT and IFNγR-/- mice at day 9 after LCMV-Cl13 infection. C. WT and IFNγR-/- naïve monocytes were transferred into LCMV-Cl13 infected WT mice and gated and analyzed as in Fig 4E. Flow plots show CD11c++ DC, and iregDC and stimDC development in WT (left flow plots) and IFNγR-/- (right flow plots) transferred monocytes. Bar graphs indicate the total number of transferred cells at day 9 after infection and the number of CD11b+ DC, iregDC, and stimDCs derived from the transferred monocytes. Data are representative of 2 or more independent experiments each consisting of 3–5 mice per group. *, p<0.05.

Fig 7. Monocyte-derived iregDCs develop in Mtb infection, HIV infection and cancer.

A. Flow plots show CD39 and PDL1 expression on splenic CD11b+ DCs from IL-10 reporter mice either uninfected or infected with virulent Mtb for 36 days. Histograms indicate expression of the indicated protein on iregDC (red) and stimDC (blue). Bar graphs indicate the number of iregDC and stimDC in uninfected and Mtb infected mice. B. Flow plots show CD39, PDL1 and monocyte-associated proteins on CD11c++, HLA-DR+, CD14+ iregDC (red) and stimDC (blue) from the peripheral blood of naïve humanized mice and HIV infected humanized mice. C. Flow plots show CD39, PDL1 and monocyte-associated proteins on CD11b+, CD11c++ iregDC (red) and stimDC (blue) on B16 melanoma infiltrating leukocytes. Data are representative of 2 or more independent experiments each consisting of 4–7 mice per group. *, p<0.05.