Lymphatic vessel transit seeds precursors to cytotoxic resident memory T cells in skin draining lymph nodes

Abstract

Resident memory T cells (T_RM) provide rapid, localized protection in peripheral tissues to pathogens and cancer. While T_RM are also found in lymph nodes (LN), how they develop during primary infection and their functional significance remains largely unknown. Here, we track the anatomical distribution of anti-viral CD8⁺ T cells as they simultaneously seed skin and LN T_RM using a model of skin infection with restricted antigen distribution. We find exquisite localization of LN T_RM to the draining LN of infected skin. LN T_RM formation depends on lymphatic transport and specifically egress of effector CD8⁺ T cells that appear poised for residence as early as 12 days post infection. Effector CD8⁺ T cell transit through skin is necessary and sufficient to populate LN T_RM in draining LNs, a process reinforced by antigen encounter in skin. Importantly, we demonstrate that LN T_RM are sufficient to provide protection against pathogenic rechallenge. These data support a model whereby a subset of tissue infiltrating CD8⁺ T cells egress during viral clearance, and establish regional protection in the draining lymphatic basin as a mechanism to prevent pathogen spread.

Keywords: infection; lymph nodes; lymphatic vessels; migration; resident memory.

Figure 1.. Lymph node resident memory T cells form in draining lymph nodes following skin infection

(A-G) P14 TCR-tg T cells (CD90.1⁺) were transferred into naïve mice (CD90.1⁻) and the following day, ear skin was infected by scarification with vaccinia virus expressing GP33_33–41 (VV-GP33). (A) Representative flow plots (left, gated on live, CD8α⁺, CD45 IV⁻, lymphocytes) and quantification (B) of P14 T cells in lymph nodes (LN) 28 days post infection (d.p.i.). Draining LN (dLN, ipsilateral parotid); non-draining LN (ndLN, contralateral parotid). (C) Representative flow plots (left, gated on CD44⁺CD90.1⁺ cells in (A)) and quantification (D) of CD69⁺CD62L⁻ P14 T cells in LNs 28 d.p.i. (E) Representative histograms at 28 d.p.i. comparing P14 T cells. Top is CD69⁻CD62L⁺ P14 T cells in the dLN; middle and bottom are CD69⁺CD62L⁻ in dLN (middle) or skin (bottom). (F) Percent CD69⁺CD62L⁻ P14 T cells in ipsilateral and contralateral parotid (Parot), brachial (Brac), and inguinal (Inguin) LNs as well as mesenteric (Mes) LNs and spleen 28 d.p.i. (G) Number of CD69⁺CD103⁺ P14 T cells in skin and draining parotid LN over time. (A-E) Data are combined from 4 experiments for a total of n=13 mice. (F) Data are representative of two experiments with 4 mice per experiment. (G) Data are combined from at least two experiments per time point with 3–4 mice per experiment. Day 102 time point represents one experiment at day 101 and another at day 103 post infection. Bars represent average + SEM. Each point represents an individual mouse. Statistical significance determined using paired student’s t test. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001

Figure 2.. Dermal lymphatic vessels are necessary for LN T_RM formation after skin infection

(A-D) CD90.1⁺ P14 TCR-tg T cells were transferred to WT or K14-VEGFR3-Ig (K14-V3) mice and the following day ear skin was infected with vaccinia virus expressing GP33_33–41 (VV-GP33) via scarification. (A) Number of CD69⁺CD62L⁻ P14 T cells in the skin 49–50 days post infection. (B) Representative flow plots gated on live, CD44⁺CD8α⁺CD90.1⁺ lymphocytes and (C) quantification of percent and (D) total number of CD69⁺CD62L⁻ in draining lymph node (dLN). Data are combined from 2 experiments with 3–5 mice per group per experiment. (E) Experimental design for (F-H) where CD90.1⁺ P14 T cells were transferred to WT or K14-V3 mice and infected with LCMV the following day. After 30 days, mouse ears were challenged with VV-GP33 and rested at least 30 days before sacrifice. Data are representative of three experiments with 3–5 mice per group per experiment. (F) Number of CD45 IV⁻CD90.1⁺CD69⁺CD62L⁻ P14 T cells in skin. (G) Percent and (H) number of CD45 IV⁻CD90.1⁺CD69⁺CD62L⁻ P14 T cells in dLN. (I) Experimental design for (J-M) where photoconvertible Kaede-tg mice were infected with VV-GP33. Infected ears were photoconverted 14–15 days later and sacrificed 24 hrs later. (J) Representative flow plot of tetramer staining for H2-K^b-B8R-specific T cells. Gated on live, CD8α⁺ lymphocytes. (K) Representative flow plots and (L) quantification of percent and (M) number of Kaede-Red fluorescence in tetramer⁺ cells. Data are representative of two experiments with 5 mice per experiment. Statistical significance determined using unpaired student’s t test (A,C,D,F,G,H) or paired student’s t test (L,M). Bars represent average + SEM. Each point represents an individual mouse. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001

Fig. 3. Egressing T cells share transcriptional similarities with LN T_RM

Kaede-tg ears were infected with VV-GP33 by scarification and photoconverted (KaedeGreen → KaedeRed) 11 days later. 24 hours after photoconversion, KaedeGreen⁺ and KaedeRed⁺ cells were sorted by FACS (Live, CD45 IV⁻ CD8⁺CD44⁺) from draining lymph nodes pooled from 29 mice and analyzed by scRNA-seq. Prior to sorting, viral specific cells were stained with 2 APC conjugated tetramers (H2-K^b-B8R_20–27 and H2-D^b-GP_33–41). APC was then tagged with an oligonucleotide to identify a subset of viral-specific cells by scRNA-Seq. (A) UMAP projection of cells colored according to cluster. (B) CITE-seq detection of tetramers visualized by UMAP projection. (C) UMAP projection showing distribution of KaedeGreen⁺ or KaedeRed⁺ and (D) quantification of their proportions within each cluster. (E) Expression of select genes across clusters normalized from −1 to 1. Size of dot corresponds to percentage of cells in cluster expressing the transcript. (F) UMAP projection of select transcripts or (G) signature score for LN resident memory T cells (Molodtsov et al. 2021).

Figure 4.. T cell transit through skin is required for the formation of LN T_RM

(A) Experimental design for (B-F) where P14 TCR-tg T cells (CD90.1⁺) were transferred into naïve mice (CD90.1⁻) and the following day ear skin was infected by scarification with vaccinia virus expressing GP_33–41 (VV-GP33). At 13 days post infection (d.p.i.), mice were treated with IgG or αCD90.1 to deplete P14 T cells in blood and secondary lymphoid organs and sacrificed 3 or 15 days later. (B) Frequency and (C) number of all P14 T cells in untreated (14 d.p.i.) or depleted mice (16 d.p.i., 3 days post depletion) in draining lymph node (dLN). (D) Prevalence of P14 T cells in LNs of depleted mice 15 days post depletion (contralateral, non-draining LN, ndLN). (E) Percent of P14 T cells with a CD69⁺CD62L⁻ residence phenotype and (F) representative surface phenotype of P14 T cells in dLNs of depleted (orange) or IgG treated (black). Data are combined from two experiments (B,C,E,F) or representative of two experiments (D) with 3–4 mice per group per experiment. (G) Experimental design for (H-L) where P14 T cells were transferred into naïve mice and the following day ear skin was infected with VV-GP33. At variable times, the infected ear was resected and mice were all sacrificed 28 d.p.i. Data are combined from two experiments with 3–5 mice per group per experiment. (H) Total number of P14 T cells per spleen 28 d.p.i. (I) Representative flow plots and (J) quantification of CD69⁺CD62L⁻ P14 T cells in dLNs. (K) Representative flow plots and (L) quantification of CD69⁺CD103⁺ P14 T cells in dLNs. For M and N, P14 T cells were transferred into naïve mice and the following day skin from both ears were infected with VV-GP33. One ear was resected at 5 d.p.i. and dLNs from the resected and unresected ear were compared 28 d.p.i.. Quantification of CD69⁺CD62L⁻ (M) or CD69⁺CD103⁺ (N) P14 T cells in dLNs. Data are combined from two experiments (M,N) with 3–4 mice per group per experiment. Bars represent average + SEM. Statistical significance was determined using unpaired student’s t test (B,C,E), paired student’s t test (D,M,N) or one way ANOVA (H,J,L). * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001

Figure 5.. LN T_RM establishment is dependent upon antigen recognition in skin

(A) Experimental design for (B-E) where CD45.1⁺ OT-1 TCR-tg T cells were transferred to naïve mice. The next day mice were infected on the right ear skin with vaccinia virus (VV) and left ear with VV expressing SIINFEKL (VV-OVA) via scarification and sacrificed 27–28 days later. (B) Representative flow plots and (C) quantification of frequency of CD45.1⁺ OT-1 T cells 27–28 d.p.i. in draining lymph nodes (dLN). Flow plots are gated on live, CD8α⁺ lymphocytes. (D) Representative flow plots gated on CD45.1+ OT-1 T cells in dLN and (E) quantification of number of CD69+CD62L- OT-T T cells in dLNs. Data are combined from two experiments with 4 mice per experiment. (F) Quantification of GFP expression in CD45.1⁺ OT-1 Nur77-GFP T cells transferred to mice one day prior to infection with VV-OVA. Skin, dLN, and non-draining LN (ndLN). (G) Experimental design for (H and I) where CD45.1⁺ OT-1 Nur77-GFP T cells were transferred into naïve mice and the ear was infected the following day with VV-OVA. 80 d.p.i mice either received a secondary VV-OVA challenge (Challenged) or did not (Unchallenged). (H) Representative flow plots and (I) quantification of GFP expression in CD45.1⁺ OT-I T cells. Data are combined from 2 experiments. Each point represents an individual mouse. Statistical significance determined using paired student’s t test (C,E) or one way ANOVA (H). * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001

Figure 6.. LN T_RM are poised for cytotoxicity and localized protection

(A) UMAP representation of single-cell RNA-seq data from CD45 IV⁻ P14 TCR-tg T cells isolated from draining lymph nodes (dLN) 45 d.p.i. with VV expressing GP33_33–41 (VV-GP33). (B) Cells scored against a published gene signature for LN resident memory T cells (T_RM). (C) Single-cell regulatory network inference and clustering (SCENIC) analysis inferring gene regulatory networks in each cluster. (D) Differentially expressed genes between resident (cluster 2) and circulating memory (clusters 0,1,3,5) cells. (E) Expression of Gzmb, Gzmk, and Prf1 in T_CIRC (clusters0,1,3,5) and T_RM (cluster 2). (F) Expression of GranzymeB by OT-1 TCR-tg T cells at least 60 days after ear skin infection with VV expressing SIINFEKL (VV-OVA) as measured by flow cytometry. Data are representative of two experiments with 5 mice per experiment. (G) Experimental design for (H and I) where naïve mice were infected on one ear with VV-GP33 and treated with 100μg of αCD8α 11 days after infection. At least 30 days later mice were challenged with lymphocytic choriomeningitis virus (LCMV) via intraperitoneal injection. (H) Representative image of plaques and (I) quantification of plaque forming units (PFU) measured 3 days following challenge dLN (ipsilateral) or contralateral, non-draining LN (ndLN). Data are combined from 3 experiments with 3–4 mice per experiment. Each point represents an individual mouse. Statistical significance determined using pairwise Wilcoxon rank test (D,E) or paired student’s t test (F,I). ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001