Epidermal γδ T cells originate from yolk sac hematopoiesis and clonally self-renew in the adult

Abstract

The murine epidermis harbors two immune cell lineages, Langerhans cells (LCs) and γδ T cells known as dendritic epidermal T cells (DETCs). LCs develop from both early yolk sac (YS) progenitors and fetal liver monocytes before locally self-renewing in the adult. For DETCs, the mechanisms of homeostatic maintenance and their hematopoietic origin are largely unknown. Here, we exploited multicolor fate mapping systems to reveal that DETCs slowly turn over at steady state. Like for LCs, homeostatic maintenance of DETCs is achieved by clonal expansion of tissue-resident cells assembled in proliferative units. The same mechanism, albeit accelerated, facilitates DETC replenishment upon injury. Hematopoietic lineage tracing uncovered that DETCs are established independently of definitive hematopoietic stem cells and instead originate from YS hematopoiesis, again reminiscent of LCs. DETCs thus resemble LCs concerning their maintenance, replenishment mechanisms, and hematopoietic development, suggesting that the epidermal microenvironment exerts a lineage-independent influence on the initial seeding and homeostatic maintenance of its resident immune cells.

Graphical abstract

Figure 1.

Adult DETCs assemble in clonal units. (A) Cx3cr1^gfp/wt mice were lethally irradiated and reconstituted with BM from Cx3cr1^gfp/wt Ubc^tdt/tdt mice. 3 mo later, the contribution of tdTomato⁺ donor BM to DETCs (TCR Vγ5⁺ CD3⁺) and LCs (F4/80⁺ MHCII⁺) was assessed by flow cytometry. To assess total reconstitution levels by donor BM, circulating blood leukocytes (blood CD45⁺) were analyzed as controls. Data are pooled from three independent experiments with one or two individual mice per experiment. (B) CD45.1 WT and CD45.2 Cx3cr1^gfp/wt mice were surgically joined. After 2 mo, the origin of DETCs and LCs was determined by flow cytometry. This origin is expressed as the relative contribution of CD45.1⁺ or CD45.2⁺ nonhost cells to the individual parabionts of the respective other genotype. Lymph node B cells served as controls for the establishment of a shared circulation. Three parabiotic pairs were analyzed. Data in A and B are displayed as mean (columns) ± SEM (error bars), with data points corresponding to individual mice. (C–E) Multicolor fate mapping of DETCs in Ubow^+/+ mice. (C) Left: In the absence of Cre, all cells express tdTomato. Upon Cre activity, cells lose tdTomato expression and stochastically acquire YFP or CFP. This choice is definitive and transmittable to the progeny of the Cre-expressing cells. Right: Flow cytometric analysis of DETCs and LCs harvested from the epidermis of Cx3cr1^Cre/wt Ubow^+/+ mice. Cells were pregated as CD45⁺ and TCR Vγ5⁺ (DETC) or MHCII⁺ (LC). (D and E) Confocal imaging analysis of DETC clusters in Cx3cr1^Cre/wt Ubow^+/+ mice. Epidermal sheets were stained for TCR Vγ5 (step 1). Fluorescent signals corresponding to TCR Vγ5⁺ DETCs were digitally isolated using Imaris Image Analysis software (step 2). This enables analysis of fluorescent reporter recombination in DETCs only (step 3). The resulting confocal images were then digitally rendered into voronoi tessellated pictures (step 4) that are amenable to computational simulation. The number of CFP^+/+ DETCs per cluster was determined in the experimental data (“original”) and compared with Monte Carlo simulations in which DETCs were randomly distributed (step 5; 1 of 10,000 simulations displayed). For further explanations, see Materials and methods section. ***, P < 0.0001, calculated using a nonparametric test (Mann–Whitney). Data in C–E are representative of five different experiments with two mice per experiment. Bars, median. Scale bar, 100 µm. For details on digital rendering and analysis of cluster formation, see Materials and methods section.

Figure 2.

DETCs clonally proliferate at steady state. (A) Homeostatic turnover of DETCs and LCs. WT mice were provided with BrdU in their drinking water for the indicated duration. BrdU incorporation was quantified by flow cytometry at the indicated time points. Pooled data from two independent experiments (20 mice per experiment, 5 mice per time point) are shown as mean (symbols) ± SEM (error bars; B–D). Adult Cx3cr1^CreERT2/wt Ubow^+/+ mice were administered a single dose of tamoxifen. (B) Cre-mediated recombination of fluorescent reporters by DETCs and LCs was determined by flow cytometry 1 wk after tamoxifen treatment. (C and D) Confocal imaging analysis of reporter recombination in TCR Vγ5⁺ DETCs of Cx3cr1^CreERT2/wt Ubow^+/+. Epidermis was harvested 1 wk or 1, 2, or 6 mo after tamoxifen treatment. Top: Raw confocal images showing TCR Vγ5 staining and CFP/YFP fluorescence in all epidermal cells. Bottom: Fluorescent reporter signals corresponding to DETCs only (i.e., for digitally isolated TCR Vγ5⁺ DETCs (see steps 1–3 in Fig. 1). Arrowheads and dashed lines indicate mono-colored foci of DETCs. (D) Quantification of C. Numbers of CFP^+/+ DETCs per DETC foci were determined experimentally (“Original”) and compared with computer-simulated random distribution (“Randomized”) as in Fig. 1. Data in B–D are representative of two independent experiments (12 mice per experiment, 3 mice per time point). Bars in D, median. n.s., not significant; *, P < 0.05; ****, P < 0.0001, calculated using a nonparametric test (Mann–Whitney).

Figure 3.

Accelerated local self-renewal facilitates DETC replenishment following skin wounding. (A–D) Sources of DETCs following TS. The left ears of WT mice (A), chimeras previously reconstituted with Ubc^gfp/gfp BM (B), and Cx3cr1^CreERT2/wt Ubow^+/+ mice (C and D) were tape stripped (TS). Contralateral ears served as controls (no TS). (A) 10 and 30 d after TS, mice were injected with EdU, and the percentages of EdU⁺ DETCs and LCs were assessed 14 h later by flow cytometry. Data are pooled from three independent experiments with three mice each and displayed as mean (symbols) ± SEM (error bars). (B) Contribution of GFP⁺ donor BM-derived cells to DETCs and LCs 30 d after TS. Data are derived from two independent experiments and shown as mean (columns) ± SEM (error bars). (C and D) Cx3cr1^CreERT2/wt Ubow^+/+ mice received 1 mg tamoxifen and were tape stripped 1 wk later. The formation of mono-colored DETC clusters in tape stripped and control ears was analyzed 1 mo later by confocal microscopy (C) and quantified as in Figs. 1 and 2 (D). Bars, mean. Error bars, SEM. Arrowheads and dashed lines in C indicate mono-colored DETCs. Scale bar, 100 µm (C). **, P < 0.01; ****, P < 0.0001, calculated with the nonparametric Mann–Whitney test. (E) Skin from nonfluorescent WT donors was grafted on the back of Cx3cr1^gfp/wt Ubc^tdt/tdt mice. After 4 mo, the frequency of GFP⁺ TCR Vγ5⁺ host DETCs within the grafted skin was determined at various distances from the wound edge. Bars, mean + SEM. Representative confocal images (right) and quantification (left, bottom) of three independent experiments are shown. Scale bars, 70 µm (left); 40 µm (right).

Figure 4.

DETCs develop from HSC-independent YS hematopoiesis. (A) WT mice were shield-irradiated as newborns and transplanted with CD11b-depleted BM from Ubc^tdt/tdt donors, and the contribution of the grafted BM to DETCs and LCs was analyzed by flow cytometry when mice had reached adulthood. Donor contribution to brain microglia and circulating leukocytes was analyzed as references for efficacy of shielding and engraftment, respectively. (B–D) Cdh5-CreERT2^+/wt Rosa^tdt/wt (B), Runx1^CreERT2/wt Rosa^yfp/wt (C), and Csf1r^CreERT2/wt Rosa^yfp/wt (D) mice were administered a single dose of 4OHT in utero at the indicated time points. Fluorescent reporter recombination was assessed in DETCs and LCs of adult mice (B and C) or at the indicated ages (D). All data are pooled from minimally two independent experiments (litters) and depicted as mean (columns) + SEM (error bars), with single data points corresponding to individual mice (B and C) or mean (symbols) ± SEM (error bars; D). Microglia and circulating leukocytes served as controls for fate mapping of the output of YS (microglia) and adult-type definitive (blood CD45⁺) hematopoiesis, respectively. n.s., not significant; *, P < 0.05; ***, P < 0.001; ****, P < 0.0001, derived from a nonparametric test (Mann–Whitney).