Migration-induced cell shattering due to DOCK8 deficiency causes a type 2-biased helper T cell response

Abstract

Mutations that impact immune cell migration and result in immune deficiency illustrate the importance of cell movement in host defense. In humans, loss-of-function mutations in DOCK8, a guanine exchange factor involved in hematopoietic cell migration, lead to immunodeficiency and, paradoxically, allergic disease. Here, we demonstrate that, like humans, Dock8^-/- mice have a profound type 2 CD4⁺ helper T (T_H2) cell bias upon pulmonary infection with Cryptococcus neoformans and other non-T_H2 stimuli. We found that recruited Dock8^-/-CX3CR1⁺ mononuclear phagocytes are exquisitely sensitive to migration-induced cell shattering, releasing interleukin (IL)-1β that drives granulocyte-macrophage colony-stimulating factor (GM-CSF) production by CD4⁺ T cells. Blocking IL-1β, GM-CSF or caspase activation eliminated the type-2 skew in mice lacking Dock8. Notably, treatment of infected wild-type mice with apoptotic cells significantly increased GM-CSF production and T_H2 cell differentiation. This reveals an important role for cell death in driving type 2 signals during infection, which may have implications for understanding the etiology of type 2 CD4⁺ T cell responses in allergic disease.

Figure 1.. Dock8 deficiency leads to a type-2 immunity bias upon infection.

a, Peripheral blood mononuclear cells from either control or DOCK8-deficient humans restimulated with PMA/ionomycin; data are representative of 4 DOCK8-deficient and 13 healthy humans. b-d, Lung cells from wild type (WT) or Dock8 deficient (KO) mice 20 days post C. neoformans infection were restimulated ex vivo and production of IL-4, IL-5, IL-13, and TNFα assessed by flow cytometry. A representative flow cytometry plot (gated on CD4⁺ FoxP3⁻ CD44^hi T cells) from a WT and KO mouse where numbers indicate percent cells in each quadrant (b). Data are summarized from 4 independent experiments (c). Frequency of activated CD4+ T cells making any of the Th2 cytokines IL-4 or IL-5 or IL-13, determined as in (b,c) from 4 independent experiments (d). e, Representative histograms of T helper lineage transcription factor GATA-3 expression measured by flow cytometry in WT naïve (CD44^lo, shaded) or WT versus KO activated (CD44^hi) CD4⁺ T cells in the lung on day 20 post C. neoformans infection (left) and summary of relative fluorescent intensity of GATA-3 (RFI) from 3 independent experiments (right). f-h, Total C. neoformans colony forming units (CFU) in the lung (f), and number of total immune cells (g) or eosinophils (h) in the lung 20 days p.i. with C. neoformans compiled from 4-5 independent experiments. i, Representative hematoxylin- and eosin-stained lungs of C. neoformans-infected lung from WT and KO mice 20 days p.i. Scale bars are 1mm. Insets are magnified 14x. In all graphs error bars represent median ± min/max; each data point is from an individual mouse.

Figure 2.. Dock8-deficiency results in a type-2 biased CD4⁺ T cell response to non-Th2 stimuli and is associated with greater GMCSF production.

a, Heatmap of cytokine combinations (columns) produced by lung CD44^hi CD4₊ T cells at 20 days p.i. with C. neoformans upon ex vivo restimulation. Only those combinations differing between WT and KO mice (rows) by more than 0.3 standard deviations of the mean separating WT from KO in an unbiased clustering analysis (dendrogram on right side) are shown. Data are from 5 mice per group. b, Frequency of IFNγ, TNFα, IL-17, or GM-CSF producing CD4⁺ T cells determined as in (a) summarized from 4 independent experiments. c, Frequency of type-2 cytokine (IL-4, IL-5 or IL-13), IFNγ, IL-17, TNFα, or GM-CSF producing activated CD4⁺ T cells determined in lungs of mice that were infected intranasally with influenza A virus 7 days prior. d, Frequency of type-2 cytokines, IFNγ, IL-17, TNFα, or GM-CSF producing activated CD4⁺ T cells from mice that were immunized subcutaneously with Complete Freund’s adjuvant (CFA) and OVA peptide, measured in the iliac lymph node 7 days later. Data are representative of 4 independent experiments. In all graphs error bars represent median ± min/max; each data point is from an individual mouse.

Figure 3.. Selective loss of Dock8 in CX3CR1⁺ mononuclear phagocytes recapitulates the type-2 immunity bias.

a, Immune cell population counts found in the lung at day 20 p.i. with C. neoformans in WT (x-axis) compared to KO (y-axis) mice, gated as shown in Supplementary Fig. 3b. Dotted line is drawn where WT and KO cell numbers would be equal. Each point represents the mean ± S.D. of 15 (WT) and 17 (KO) mice determined in 3 independent experiments. b, Lung cells from WT, KO, or Dock8^fl/fl mice crossed to Lck, CD11c, LysM or eosinophil-specific cre lines were harvested 20 days post C. neoformans infection, restimulated ex vivo and production of IL-4, IL-5, or IL-13 assessed by flow cytometry. c,d, Congenically labelled (CD45.1+) T cells from WT mice were transferred into DOCK8 KO (CD45.2+) recipients that were infected with C. neoformans 7 days later (c) and percent of Th2 cytokine- or GM-CSF- producing CD4⁺ T cells was assessed in 2 independent experiments for WT and KO mice, as well as WT T cells in KO recipients (d). e, CX3CR1 expression on lung cell populations 20 days post-infection with C. neoformans shown as a representative histogram from 3 independent experiments. f-h, WT, Dock8 KO, and CX3CR1-Dock8^−/− mice were infected with C. neoformans, with a tamoxifen treatment schedule as shown (f). Number of activated CD4⁺ T cells (g) and frequency of type-2 cytokine and GMCSF producing lung CD4⁺ T cells (h) measured 8 days post-infection. Control group includes: WT mice (black circles), corn-oil treated Dock8^fl/fl.CX3CR1-ERcre+ (white circles), and tamoxifen-treated Dock8^fl/fl.CX3CR1-ERcre− (gold circles). Data are from 3 independent experiments. In all graphs except (a), error bars represent median ± min/max; each data point is from an individual mouse.

Figure 4.. Mononuclear phagocytes lacking Dock8 are highly sensitive to migration-induced shattering, caspase activation and death.

a, Percent of immune cells undergoing apoptosis (apop, AnnexinV+) or are dead (AnnexinV−, viability stain+) in the lung on day 8 or 20 post C. neoformans infection. Data from 2-3 independent experiments. b-d, Frequency of cleaved caspase-3+ cells analyzed by flow cytometry 8 or 20 days post-infection with C. neoformans in DCs, granulocytes, lymphocytes and non-DC MNPs (gated as in Supplementary Fig. 3b). Representative histograms from day 20 (b) and data summarized from 2 independent experiments for DCs and non-DC MNPs in WT and KO mice (c), and data from CX3CR1-Dock8^−/− mice or controls treated as in Fig. 3f (d). e-h, WT and KO LifeAct-GFP+ BMDCs migrating in a 1.7mg/ml collagen matrix ex vivo. Still images taken at 0 and 3 hours from two representative movies, Supplementary Video 2 (e). Relative viability measured 45 min from start of migration assay as compared to cells kept in media for the same duration in 2 independent experiments (f). BMDC cell area, perimeter, major and minor axes quantified after 3 hours of migration, summarized from 5 movies and 3 mice per group; n=2106 WT, n=4479 KO cells (g). Example confocal images of individual WT and KO BMDCs (h). i-k, Dock8^−/− mice were treated with a pan-caspase inhibitor (Q-VD-OPh) on day 1, 3, 5, and 7 post C. neoformans infection (i), and on day 8 post infection number of activated (CD44^hi) CD4⁺ T cells (j) and frequency of activated CD4⁺ T cells making Th2 cytokines (IL-4, IL-5 or IL-13) or GM-CSF following ex vivo restimulation (k) was assessed in the lung. Data from 2 independent experiments. In all graphs error bars represent median ± min/max except in (f) where mean ± SD is shown; each data point is from an individual mouse except in (g) where each data point is from an individual cell.

Figure 5.. Increased IL-1β is a necessary but not sufficient signal in the greater GM-CSF and type-2 cytokine production by CD4⁺ T cells in Dock8-deficiency.

a,b, BMDC were kept in dissociated culture (non-migrating) or added to 1.7mg/ml collagen matrices. Caspase and MLKL activation was examined after 40 min by western blot. Pro-caspase 3 gel is shown at high and low exposure (hi exp, lo exp). Two independent replicates are shown per group (a). IL-1β released was quantified after 3 hours by ELISA in 2 independent experiments. Each data point represents a BMDC batch generated from an individual mouse (b). c-d, Cytokine production was measured in mice infected with C. neoformans (day 20) either in BAL fluid by cytokine array (c) or in whole lung by ELISA (d). Data in (c) is from 4 mice per group and shown as fold change relative to WT group mean. Data in (d) is from 2 independent experiments. e-g, WT, KO and Dock8/IL-1R double KO mice (KO x IL1R KO) were infected with C. neoformans and on day 20 post infection number of activated (CD44^hi) CD4⁺ T cells (e), frequency of activated CD4⁺ T cells making Th2 cytokines (IL-4, IL-5 or IL-13) or GM-CSF following ex vivo restimulation (f), and eosinophil numbers (g) were evaluated in the lung. Data are from 2 independent experiments. h-k, Congenically labelled (CD45.2+) T cells from IL1R KO mice were transferred into CD45.1+ TCRβ KO or Dock8/TCRβ double KO recipients that were infected with C. neoformans (h) and on day 20 p.i. number of activated (CD44^hi) CD4⁺ T cells (i), frequency of activated CD4⁺ T cells making Th2 cytokines (IL-4, IL-5 or IL-13) or GM-w ex vivo restimulation (j), and eosinophil numbers in the lung was assessed in both recipient groups compared to Dock8^−/− mice (k). Data are from 3 independent experiments. l, Dock8^−/− mice were treated with a pan-caspase inhibitor (Q-VD-OPh) with or without IL-1β on day 1, 3, 5, and 7 post C. neoformans infection, and on day 8 post infection the frequency of activated CD4⁺ T cells making Th2 cytokines (IL-4, IL-5 or IL-13) following ex vivo restimulation was assessed in the lung. Data from 2 independent experiments. For all graphs each data point is from an individual mouse. In (b,d) lines represent means; in all other graphs error bars represent median ± min/max.

Figure 6.. GM-CSF production by CD4⁺ T cells plays a role in the type-2 bias, amplifying IL-1β production.

a, Lung cells from WT and KO mice 20 days post C. neoformans infection were restimulated ex vivo, and production of GM-CSF by CD4⁺ T cells assessed by flow cytometry. Representative flow cytometry plots (gated on CD4⁺FoxP3⁻ CD44^hi T cells, top) or gated on any Th2 cytokine-producing CD4⁺ T cells (bottom), with numbers indicating percent cells in each quadrant, are shown. b,c, GMCSF-production by CD4⁺ T cells was determined as in (a) upon infection with C. neoformans (lung, day 20 p.i.), influenza A (lung, day 7 p.i.) or following immunization with CFA + OVA (iliac lymph node, day 7 p.i.). and the frequency of GMCSF-producing Th2 cells is summarized from 2-4 independent experiments. c-f, KO mice were treated with αGMCSF every 2 days from day 8 post C. neoformans infection and compared to untreated KO and WT mice (c). On day 20 post infection number of activated (CD44^hi) CD4⁺ T cells (d), frequency of activated CD4⁺ T cells making Th2 cytokines (IL-4, IL-5 or IL-13) or GM-CSF following ex vivo restimulation (e), eosinophil numbers (f) were evaluated in the lung. Data are from 3 independent experiments. g, IL-1β production was measured in WT, KO, or KO treated with αGMCSF infected with C. neoformans (day 20) either in whole lung by ELISA. h, Monocytes were cultured in media with or without GM-CSF in presence or absence of LPS and protein expression of pro-IL1β measured by flow cytometry. A representative histogram is shown. For all graphs error bars represent median ± min/max, except in (g) where group means are shown; each data point is from an individual mouse.

Figure 7.. Cell death provides a necessary and sufficient type-2-inducing signal.

a-e, WT mice were infected with C. neoformans and given intrapharyngeal IL-1β, 1x10⁷ irradiated thymocytes i.v. on days 4, 5, 6, and 7 p.i., or injected with PBS alone, as shown (a). The number of activated (CD44^hi) CD4⁺ T cells (b), frequency of activated CD4⁺ T cells making Th2 cytokines (IL-4, IL-5 or IL-13), GM-CSF, TNFα or IL-17 following ex vivo restimulation (c), and eosinophil numbers (d) were evaluated in the lung. Levels of IL-1β were measured in total lung homogenate. Data are from 2-5 independent experiments. For all graphs error bars represent median ± min/max, except in (e) group means are shown; each data point is from an individual mouse.