Dermis resident macrophages orchestrate localized ILC2 eosinophil circuitries to promote non-healing cutaneous leishmaniasis

Abstract

Tissue-resident macrophages are critical for tissue homeostasis and repair. We previously showed that dermis-resident macrophages produce CCL24 which mediates their interaction with IL-4⁺ eosinophils, required to maintain their M2-like properties in the T_H1 environment of the Leishmania major infected skin. Here, we show that thymic stromal lymphopoietin (TSLP) and IL-5⁺ type 2 innate lymphoid cells are also required to maintain dermis-resident macrophages and promote infection. Single cell RNA sequencing reveals the dermis-resident macrophages as the sole source of TSLP and CCL24. Generation of Ccl24-cre mice permits specific labeling of dermis-resident macrophages and interstitial macrophages from other organs. Selective ablation of TSLP in dermis-resident macrophages reduces the numbers of IL-5⁺ type 2 innate lymphoid cells, eosinophils and dermis-resident macrophages, and ameliorates infection. Our findings demonstrate that dermis-resident macrophages are self-maintained as a replicative niche for L. major by orchestrating localized type 2 circuitries with type 2 innate lymphoid cells and eosinophils.

Fig. 1. IL5⁺ ILC2s are required to maintain M2-like dermal TRMs which play a critical role in non-healing L. major infection.

a Representative flow cytometric analysis of ear isolates prepared from naive R5: ROSA26-LSL-tdTomato mice. CD45.2⁺CD11b^-CD90.2⁺NK1.1^--gated cells were analyzed for the expression of TCRβ and TCRγδ, and subdivided into DETCs (green; TCRγδ^highTCRβ^-), γδT cells (blue; TCRγδ^intTCRβ^-), ILC2s (Red; TCRγδ^-TCRβ^-), and αβT cells (orange; TCRγδ^-TCRβ⁺). Both CD2/CD3 and tdTomato expression were shown by overlaying color-coded populations in CD2-CD3 axes and histogram respectively. b tdTomato expression in ILC2s and (c) the absolute numbers of indicated cells recovered from ears of R5 (n = 8) vs. R5: ROSA26iDTR (n = 7) mice were measured at 12 days p.i. with 2 × 10⁵ LmSd followed by DT treatment. d, e The same parameters are shown as above in ear isolates prepared from R5 (n = 4) vs. R5: Gata3^f/f (n = 4) mice at day 12 p.i. with 2 × 10⁵ LmSd, or (f) WT (n = 8) vs. RT: Gata3^f/f (n = 8) vs. R5: Gata3^f/f (n = 8) mice adoptively transferred with in 0.5 × 10⁶ in vitro expanded, splenic ILC2s from WT mice, intravenously administered on day 0 p.i. with 2 × 10⁵ LmSd. g Lesion development and pathology score over the course of infection with 10³ LmSd metacyclic promastigotes in the ear dermis of Gata3^f/f, Il4^−/−, and R5: Gata3^f/f animals (n = 8). h Parasite burdens were quantified at 12 weeks p.i. Values represent mean ± standard deviation (n = 8). *P < 0.05, **P < 0.01, and ***P < 0.01 by two-sided non-parametric Mann-Whitney test (b–e) and one-way ANOVA with Dunn’s posttest compared with Gata3^f/f (f, g). Data are representative of two independent experiments (a–g). Source data are provided as a Source Data file.

Fig. 2. TSLP receptor signaling promotes non-healing L. major infection.

a Lesion development and pathology score over the course of infection with 10³ LmSd metacyclic promastigotes, and (b) Parasite burdens were quantified at 12 weeks p.i. in the ear dermis of WT (n = 10), Il25^−/− (n = 10), Il33^−/− (n = 6), and Tslpr^−/− (n = 10) mice. c The absolute numbers of indicated cells and (d) intracellular staining for IL5/IL13 expressions in ILC2s from ear isolates of WT and Tslpr^−/− mice were compared at 12 days p.i. with 2 × 10⁵ LmSd (n = 10). Values represent mean ± standard deviation. *P < 0.05, **P < 0.01, and ***P < 0.001 by one-way ANOVA with Dunn’s posttest compared with Tslpr^−/− (a) and two-sided non-parametric Mann-Whitney test (b–d). Data are representative of two independent experiments (a–d). Source data are provided as a Source Data file.

Fig. 3. TSLP is expressed by dermal TRMs during L. major infection.

a UMAP plot representing of 17,355 cells combined from four samples; naïve WT (4592 cells) and eoCre: Il4/13^f/f (5913 cells), and infected WT (3728 cells) and eoCre: Il4/13^f/f (3552 cells) mice, 12 days post-challenge with 2 × 10⁵ LmSd in the ear dermis. b, c Selected gene expression UMAP plots. The identities of expressing cells were labeled and color-coded as the magnitude of expressions in scale. d Heat map to visualize Tslp, Il25, and Il33 expression of indicated dermal myeloid cells, generated from Expression (GEO) database. e Comparison of Tslp transcription from various immune cells published by Immgen. The relative expression of Tslp gene is shown as a heatmap among the indicated populations. Each square represents either a specified subset or population of the specified cell isolated from different organs. f Dot plots of expressed transcripts within dermal TRMs from infected WT (n = 86) and eoCre: Il4/13^f/f (n = 69) animals and (g) ILC2s from infected WT (n = 73) and eoCre: Il4/13^f/f (n = 34) animals. h The absolute numbers and % IL4/IL13⁺ ILC2s from ear isolates of WT and eoCre: Il4/13^f/f mice at 12 days p.i. with 2 × 10⁵ LmSd (n = 6). Values represent mean ± standard deviation. *P < 0.05 and **P < 0.01 by two-sided non-parametric Mann-Whitney test (f–h). Data are representative of two independent experiments (h). See also Supplementary Fig. 1, 2, and 3. Source data are provided as a Source Data file.

Fig. 4. Characterization of MHCII⁺ and MHCII^- subpopulation of dermal TRMs in scRNA-seq.

a cMAP analysis of dermal TRMs showing their enrichment for either MHCII⁺ (gray) or MHCII^- (red) dermal macrophage transcriptomes (Ly6C^lowCD64^highMerTK⁺) which were previously published. Cells having transcriptional similarity to neither subset are marked as zero cMAP score (clear). b UMAP of dermal TRMs overlayed by MHCII⁺ (gray) and MHCII^- (red) subsets color-coded by cMAP analysis in (a). c Selected gene expression UMAP plots of dermal TRMs. d 3-dimensional scatter plot showing statistical correlation between Mrc1, Tslp, and Ccl24 gene expressions in MHCII⁺ (gray) or MHCII^- (red) dermal TRMs. e The heatmap of DEGs (absolute logFC > 0.4 and p_val_adj (FDR corrected < 0.05)). DEGs between two groups of cells were identified using Wilcoxon Rank Sum Test. f IPA analysis is performed using the DEGs from the comparison between MHCII⁺ (gray) and MHCII^- (red) dermal TRMs. The p value of overlap is calculated using the right-tailed Fisher’s Exact Test. See also Supplementary Fig. 4.

Fig. 5. TSLP promotes ILC2 infiltration and interaction with dermal TRMs during L. major infection.

a Representative flow cytometric analysis of isolates from epidermal and dermal layers of a naive ear skin to determine the tissue distribution of DETC, γδT, ILC2, and αβT cells. b Image obtained from an IVM of the ear of R5: ROSA26-LSL-tdTomato reporter mouse intravenously injected with efluor450 anti-CD31 Abs to label blood vessels, and its 3-dimensional surface rendering (right panel). Hair follicles are demarcated by dotted lines. The closest distance between dermal ILC2 and blood vessel was measured and plotted in the right panel. c Image obtained from an IVM of the ear of R5: ROSA26-LSL-tdTomato reporter mouse infected intradermally with 1 × 10³ LmSd-GFP parasites and treated with either TSLP-neutralizing Abs or control IgG for 9 days. R5: ROSA26-LSL-tdTomato mice were injected with Cy5-Manocept^TM to label MR^high dermal TRMs and efluor450 anti-CD31 Abs for blood vessels. Yellow-colored “R5 track” shows the paths followed by tdTomato⁺ ILC2s. Both the number and % surface of contacts per timepoint between ILC2s and dermal TRMs were analyzed and plotted in right panels. ****P < 0.0001 by two-sided nonparametric Mann-Whitney test (c). Data are representative of more than four independent mouse experiments (a–c). Source data are provided as a Source Data file.

Fig. 6. Ccl24-cre transgenic mice target MHCII^-MR^high dermal TRMs and TRMs in multiple tissues.

a Representative flow cytometric analysis of ear isolates from ROSA26-LSL-tdTomato mice vs. Ccl24-cre: ROSA26-LSL-tdTomato mice infected with LmSd-GFP for 8 days. b Live CD45.2⁺ singlets were gated into eosinophils, PMNs, inflammatory monocytes, moDCs, and dermal TRMs. The tdTomato⁺ cells in red were overlayed onto gatings. c Representative confocal image of ear dermis of infected Ccl24-cre: ROSA26-LSL-tdTomato mice which were intravenously injected with efluor450 anti-CD31 Abs and Cy5-Manocept^TM. d Quantification of percent tdTomato positive of dermal TRMs in (b) (n = 6). e Flow cytometric analysis of tdTomato⁺ cells in Ccl24-cre: ROSA26-LSL-tdTomato mice which were overlayed onto gatings for indicated organs. pgWAT; perigonadal white adipose tissue. The complete gating strategies are shown in Supplementary Fig. 6. Percent tdTomato positive of TRMs are labeled in red. f–j Immunohistochemistry analysis of Alexa488-Laminin Abs stained tissue sections or whole mount from Ccl24-cre: ROSA26-LSL-tdTomato mice which were intravenously injected with efluor450 anti-CD31 Abs and Cy5-Manocept^TM. G glomerulus. Tubules; Renal tubules which were visualized with filtered Cy5-Manocept^TM. Data are representative of more than three independent mouse experiments (a–c). See also Supplementary Fig. 5 and 6. Source data are provided as a Source Data file.

Fig. 7. TSLP from dermal TRMs promotes non-healing L. major infection.

a Quantification of Tslp expression from the sorted MR^high dermal TRMs of WT, Ccl24-cre: Tslp^f/+, Ccl24-cre: Tslp^f/f animals at 12 days p.i. with 2 × 10⁵ parasites (n = 4). b The absolute numbers of indicated cells and (c) % IL-5⁺ or IL-13⁺ ILC2s recovered from ears of WT, Ccl24-cre: Tslp^f/+, Ccl24-cre: Tslp^f/f animals were measured at 12 days p.i. with 2 × 10⁵ LmSd (n = 6). d Lesion development and pathology scores over the course of infection with 10³ LmSd metacyclic promastigotes in the ear dermis of WT (n = 10), Ccl24-cre: Tslp^f/+ (n = 6), Ccl24-cre: Tslp^f/f animals (n = 18). e Parasite burdens were quantified at 9 weeks p.i. in the ear dermis of WT (n = 10), Ccl24-cre: Tslp^f/+ (n = 6), Ccl24-cre: Tslp^f/f animals (n = 8). Values represent mean ± standard deviation. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by one-way ANOVA with Dunn’s post-test compared to Ccl24-cre: Tslp^f/f animals (a–e). Data are representative of two independent experiments (a–e). Source data are provided as a Source Data file.