Research ArticleT CELLS

Human CD4+CD103+ cutaneous resident memory T cells are found in the circulation of healthy individuals

See allHide authors and affiliations

Science Immunology  05 Jul 2019:
Vol. 4, Issue 37, eaav8995
DOI: 10.1126/sciimmunol.aav8995

Human resident memory T cells hit the road

Skin and other barrier tissues are home to long-lived tissue-resident memory T cells (TRM) that function as sentries capable of rapidly responding to previously encountered antigens. Klicznik et al. investigated the relationship between human skin CD4+ TRM and human blood CD4+ memory T cells expressing skin-homing markers by comparing immunophenotypes, gene expression, and T cell receptor sequences. Shared phenotype, function, and clonotypes between blood and skin CLA+CD103+CD4+ T cells indicated that blood CD4+CLA+CD103+ T cells were previously skin resident. Analysis of immunodeficient mice bearing human skin xenografts revealed that human skin CD4+ TRM can exit the skin, reenter the circulation, and home to secondary human skin sites. These findings establish that there is basal recirculation of human CD4+ skin TRM. See the related Focus by Carbone and Gebhardt.

Abstract

Tissue-resident memory T cells (TRM) persist locally in nonlymphoid tissues where they provide frontline defense against recurring insults. TRM at barrier surfaces express the markers CD103 and/or CD69, which function to retain them in epithelial tissues. In humans, neither the long-term migratory behavior of TRM nor their ability to reenter the circulation and potentially migrate to distant tissue sites has been investigated. Using tissue explant cultures, we found that CD4+CD69+CD103+ TRM in human skin can down-regulate CD69 and exit the tissue. In addition, we identified a skin-tropic CD4+CD69CD103+ population in human lymph and blood that is transcriptionally, functionally, and clonally related to the CD4+CD69+CD103+ TRM population in the skin. Using a skin xenograft model, we confirmed that a fraction of the human cutaneous CD4+CD103+ TRM population can reenter circulation and migrate to secondary human skin sites where they reassume a TRM phenotype. Thus, our data challenge current concepts regarding the strict tissue compartmentalization of CD4+ T cell memory in humans.

INTRODUCTION

T cell memory is compartmentalized into circulating and tissue-resident cell populations. Whereas circulating memory T cells continually patrol the body via the blood and lymphatics, tissue-resident memory T cell (TRM) populations establish residence in nonlymphoid organs, where they can provide potent recall responses (1). TRM populations at barrier surfaces such as the intestines, lungs, and skin are best defined by expression of the markers CD103 and/or CD69, which together function to restrict their recirculation and maintain tissue residence (2, 3). However, despite extensive studies, there is no single-cell definition for TRM. Instead, the term TRM is used to describe a cell population within a tissue that is in substantial disequilibrium with cells in the circulation as measured by depletion, tissue transplantation, or parabiosis studies (2, 4, 5).

TRM were first identified in the context of CD8+ T cell responses to infection (5, 6). Although cutaneous CD8+ TRM have been well studied in the mouse, the behavior of CD4+ memory T cells in mouse skin has been more controversial. Initial studies indicated that CD4+ T cells in the skin exhibit a more dynamic pattern of migration and recirculation than cutaneous CD8+ T cells, resulting in their equilibration with the circulating T cell pool (7, 8). However, skin inflammation or infection increased recruitment and retention of murine CD4+ T cells in the skin (8, 9) and, in some cases, led to the formation of sessile cutaneous CD69+CD103+CD4+ T cells with superior effector functions (10, 11). Within the skin, TRM are most abundant at the site of initial infection (11, 12). However, long-term maintenance of this biased distribution may pose a disadvantage for a large barrier organ like the skin where pathogen reencounter at a secondary tissue site is possible.

As in experimental animals, human CD4+ TRM are generated in response to cutaneous microbes such as Candida albicans (11), but aberrantly activated or malignant TRM are implicated in skin diseases, including psoriasis and mycosis fungoides (13). However, in studying cutaneous CD4+ TRM, reliance on animal models can be problematic due to fundamental structural differences in the skin in humans versus mice and a lack of direct correspondence between cutaneous T cell populations in these species. For instance, whereas nearly all CD4+ T cells in murine skin are found in the dermis, the human epidermis is much thicker than in mice, and memory CD4+ T cells are found throughout human skin, in both the dermal and the epidermal compartments (2). In human skin, most CD4+ T cells express CD69, and a fraction of these are also CD103+. Moreover, studies following depletion of circulating T cells with anti-CD52 (alemtuzumab) demonstrated that both CD103 and CD103+ CD4+CD69+ T cell populations can persist locally in the skin in the absence of continual replacement by circulating cells (2), thereby defining them functionally as TRM populations.

Most human skin–resident memory T cells express the cutaneous lymphocyte antigen (CLA), a glycan moiety that promotes skin homing of immune cells by acting as a ligand for E-selectin (14). CLA is also expressed by skin-tropic memory CD4+ T cells in human blood (15), but the clonal and functional relationship between the CD4+CLA+ T cells in blood and TRM populations in the skin is not well defined. To elucidate the relationship between CD4+CLA+ T cells in the blood and skin, we characterized circulating and cutaneous T cell populations, taking advantage of new technological tools including cytometry by time of flight (CyTOF), transcriptional profiling of rare cell populations by RNA sequencing (RNA-seq), and a human skin xenograft mouse model. In tissue explant and skin xenograft mouse systems, we showed that CD4+CLA+CD69+CD103+ TRM in human skin can down-regulate CD69 and exit the tissue. Furthermore, we identified a distinct population of CD4+CLA+CD69CD103+ T cells in human lymph and blood that shared phenotypic, functional, and transcriptomic signatures with the CD4+CLA+CD103+ TRM population in the skin. T cell receptor (TCR) repertoire analysis confirmed the common clonal origins of the circulating and skin-resident CD4+CLA+CD103+ T cells. In skin-humanized mice, we further showed that after exit from the skin, CD4+CLA+CD103+ T cells could migrate into distant human skin sites and regain CD69 expression upon reentering the tissue. Thus, basal recirculation of the CD4+CLA+CD103+ TRM population can be detected in the steady state, and this can promote the spread of skin TRM throughout this large barrier tissue.

RESULTS

Cutaneous CD4+CLA+CD103+ TRM can down-regulate CD69 and exit the skin

Confirming previous analyses (2), we found that the vast majority of both CD8+ and CD4+ T cells in human skin expressed CD69, and a subset of CD69+ cells also expressed CD103 and thus had the phenotype of cutaneous TRM populations resistant to alemtuzumab-mediated depletion of circulating cells (Fig. 1, A and B; see fig. S1 for representative T cell gating strategies from skin and blood). Consistent with their localization in the skin, most of these CD69+CD103+ T cells also expressed CLA (Fig. 1C) (14). To directly assess the ability of different skin T cell populations to exit the tissue, we performed tissue explant cultures using human skin obtained from surgical samples. Despite high expression of CD69 in the tissue, we found that a fraction of cutaneous CD4+CLA+ T cells exited the tissue in these explant cultures and could be detected in the culture medium. This tissue exit was associated with down-regulation of CD69 by a fraction of both CD103+ and CD103 cells (Fig. 1, D and E). The input skin contained virtually no CD69CD103+CD4+ T cells (Fig. 1D), and CD103 expression was not induced on blood-derived CD4+ T cells cultured in parallel, indicating that CD69 was down-regulated by CD103+ TRM in these cultures. In addition, the CD103+ T cells found in the medium did not express the chemokine receptor CCR7 (Fig. 1D)and thus are distinct from CD69CD103+/lo T cells that undergo CCR7-dependent migration from the skin to the draining lymph nodes in mice (16) and from the CCR7+CD62L migratory memory T (TMM) cells described in human skin (2). By contrast, compared with CD4+ T cells, fewer CD8+ T cells exited the tissue in these skin explant cultures (Fig. 1D), consistent with more prolonged tissue residency of the skin CD8+ TRM population in some mouse models (7, 8).

Fig. 1 CD4+CLA+CD103+T cells down-regulate CD69 and exit the skin.

(A) Representative flow cytometric analysis of CD69 and CD103 expression by live gated CD8+ and CD4+ T cells from human skin. (B) Graphical summary of the proportions of CD69- and CD103-defined T cell populations among CD8+ and CD4+ skin T cells. (C) Representative flow cytometric analysis of CLA expression by live gated CD103+CD69+ TRM in human skin. (D) Human skin was adhered to tissue culture plates and cultured for 7 days submerged in medium. The ratio of CD4+ and CD8+ T cells and the expression of CLA and CD103 by T cells in the indicated samples were analyzed by flow cytometry. Representative data (n = 4). APC, allophycocyanin; PE, phycoerythrin. (E) Graphical summary of the proportion of CD69 cells among CD103+ or CD103 live gated CD45RACD4+CLA+ T cells from the indicated samples. Open symbols represent data from an individual with mammary carcinoma but no skin condition. Significance was determined by one-way repeated-measures ANOVA with Tukey’s posttest for pairwise comparisons. (F) Three 8-mm punch biopsies of healthy human skin per animal (n = 3) were placed on the back of NSG mice, and grafts and spleens were analyzed by flow cytometry 50 days later. Representative flow cytometric analysis of CLA and CD103 expression by live gated human CD45+CD3+CD4+CD25CD45RA T cells. (G) Graphical summary showing CD103 expression by live gated human CD45+CD3+CD4+CD25CD45RACLA+ T cells from skin grafts and spleens of skin-grafted NSG mice.

To determine whether CD4+ TRM could exit the skin in vivo, we used a xenografting model in which human skin was transplanted onto immunodeficient NOD (nonobese diabetic), scid (severe combined immunodeficient), common-γ chain-deficient (NSG) mice (17). T cells that had exited the skin were analyzed in the spleen. Similar to our explant studies, T cells exited the skin in all animals examined, including CLA+CD103+ T cells in two of the three recipient animals (Fig. 1, F and G). Expression of CD103 in the periphery is not induced by the xenogeneic system we used, as we did not observe induction of CD103 expression by CD4+ T cells in NSG mice upon transfer of total peripheral blood mononuclear cells (PBMCs; fig. S2).

Identification of CD4+CLA+CD103+ T cells in circulation of healthy humans

Among the CD4+CLA+CD45RA T cells in human blood, we noted a small population of CD103+CD69 cells (Fig. 1D) and reasoned that these may represent CD4+CLA+CD103+ T cells that had exited the skin and reentered the circulation. To more comprehensively examine circulating CLA+CD103+ T cells and determine whether they constitute a phenotypically distinct T cell population, we performed mass cytometry analysis of CLA+ T cells in the blood of five healthy individuals using markers associated with cell lineage, functional differentiation, and migration of CD4+ T cells. Cluster analysis following t-distributed stochastic neighbor embedding (t-SNE) revealed 10 clusters of CD3+CD45RACLA+ memory T cells present in all individuals examined (Fig. 2A and fig. S3), including five clusters of CD4+ T cells (Fig. 2B). Most CD4+CD103+ cells clustered together as a phenotypically discrete population (cluster 10). In addition to being positive for CD103 and its dimerization partner β7 integrin, cells in this cluster expressed chemokine receptors strongly indicative of skin tropism such as CCR4, CCR6, and CCR10, but were largely negative for CCR7 (Fig. 2, B and C) (18, 19). In addition, cells in cluster 10 were low for expression of markers of regulatory T cells (Foxp3 and CD25), T helper 1 (TH1) cells (CXCR3), TH17 cells (CD161), or natural killer (NK) T cells (CD56).

Fig. 2 CD4+CLA+CD103+ T cells constitute a unique cell population in human blood.

(A) Left: Mass cytometry analysis of CD45RA and CLA expression by live gated CD3+CD45+ PBMCs showing the gate used to define CD3+CLA+ T cells for subsequent clustering analysis. Right: t-SNE analysis and clustering of CD3+CLA+ T cells from blood of five healthy donors based on the expression of CD4, CD8, CCR7, CD103, β7 integrin, CXCR3, CCR6, CCR4, CCR10, Foxp3, CD27, CD25, CD161, and CD56. (B) Heat map showing relative expression of the indicated markers in each of the CD3+CLA+ cell clusters. (C) t-SNE analysis of CD3+CLA+ T cells overlaid with relative expression of the indicated markers.

Using conventional flow cytometry, we directly compared the abundance and phenotype of circulating and skin-resident CD4+CLA+CD103+ T cells. In both blood and skin, CD103 and CCR7 delineated three populations of CLA+ T cells (Fig. 3, A and B), but the distribution of these populations differed markedly between sites. Whereas CD4+CLA+CD103+ memory T cells were common in the skin (26 ± 9% of CD4+CLA+ T cells), they were a rare population in the blood, representing, on average, <2% of circulating CD4+CLA+CD45RA memory T cells and <0.2% of total CD4+ T cells (Fig. 3C and fig. S4A). On the basis of previous estimates of the total number of T cells in human skin and blood (14), we calculated that the number of CD4+CLA+CD103+ cells in the skin is approximately 250-fold higher than in the blood, highlighting the disequilibrium of this population between the skin and circulation (fig. S4B). Within the skin, we found CD4+CLA+CD103+ cells in both the dermis and epidermis (fig. S4C).

Fig. 3 Shared phenotype of CD4+CLA+CD103+ T cells from human blood and skin.

(A) Representative flow cytometric analysis of CD45RA and CLA expression by live gated CD4+ T cells from blood and skin of healthy donors. (B) Representative flow cytometric analysis of CCR7 and CD103 expression by live gated CD4+CD45RACLA+ memory T cells from blood and skin of healthy donors. (C) Graphical summary of the proportions of CCR7- and CD103-defined T cell populations among CD4+CD45RACLA+ T cells from blood and skin. (D and E) Representative flow cytometric analysis and graphical summary of expression of the indicated markers by CD4+ T cell populations in the blood and skin as indicated. Significance was determined by one-way repeated-measures ANOVA with Tukey’s posttest for pairwise comparisons.

CD4+CLA+CD103+ T cells in the skin shared the CCR4+CCR6+CXCR3 chemokine receptor profile with circulating CD4+CLA+CD103+ memory T cells. In addition, CD4+CLA+CD103+ T cells in the blood and all populations in the skin were largely negative for CD49d (Fig. 3D). Like CD103, CD49d (also known as α4 integrin) can pair with β7 integrin and, in this combination, promotes T cell localization to the intestinal mucosa (20). CD4+CLA+CD103+ T cells in the skin and blood were also CD27, indicating that they are terminally differentiated (21), whereas most circulating CD4+CLA+CD103 memory T cells were CD27+. In addition, substantial fractions of circulating CD4+CLA+CD103+ memory T cells expressed the markers CD101 and CD9, which were also expressed by most of the CD4+CLA+CD103+ T cells in the skin (Fig. 3E). CD101 was recently identified as a marker of CD8+CD103+ TRM populations in various human tissues that was also expressed by CD4+ lung TRM (3, 22), whereas CD9 is expressed by immune cells and keratinocytes and modulates transforming growth factor–β (TGF-β) signaling, integrin function, cell migration, and wound healing (23, 24). However, consistent with down-regulation of CD69 by T cells that migrated out of the skin in our tissue explant cultures, CD4+CLA+CD103+ cells in the blood were almost all CD69 (Fig. 3E).

CD4+CLA+CD103+ T cells from blood and skin share a transcriptional and functional profile

To assess the transcriptional signature of circulating and skin-resident CD4+CLA+CD103+ T cells, we performed RNA-seq on sorted CLA+ memory CD4+ T cell populations from blood and skin (see fig. S5 for sort scheme). Analysis of circulating CD4+CLA+CD103+ T cells identified a unique signature of 83 genes that were differentially expressed [false discovery rate (FDR) < 0.05 and fold change > 2] compared with both CD4+CLA+CD103CCR7 and CD4+CLA+CD103CCR7+ memory populations in the blood (Fig. 4A). This CD103+ gene signature derived from blood was also significantly enriched in CD4+CLA+CD103+ versus CD4+CLA+CD103CCR7 T cells in the skin (Fig. 4B), consistent with the notion that circulating CD4+CLA+CD103+ cells represent skin T cells that down-regulated CD69 to exit the tissue. Hierarchical clustering based on this gene signature grouped CD4+CLA+CD103+ cells from skin and blood into a single branch (Fig. 4C).

Fig. 4 CD4+CLA+CD103+ T cells from human blood and skin share a transcriptional profile.

Whole-transcriptome profiling by RNA-seq was performed on sorted CLA+ T cell subsets from blood or skin. (A) Venn diagram showing the number of significantly differentially expressed genes [FDR < 0.05 and log2 fold change (FC) > 1] between CLA+CD103+ T cells and either CLA+CD103CCR7+ or CLA+CD103CCR7 T cells as indicated. The overlapping 83 genes were designated as the CD103+ gene signature. (B) Barcode plot showing the distribution of the CD103+ signature genes (red, up-regulated in CD103+; blue, down-regulated in CD103+) relative to gene expression changes comparing CD103+ and CD103CCR7 T cells from the blood or skin as indicated. Significance was determined by rotation gene set testing for linear models. (C) Heat map and hierarchical clustering of RNA-seq samples from the indicated blood and skin cell populations based on the CD103+ gene signature. (D) Venn diagram showing functional annotation of key genes up- or down-regulated by CLA+CD103+ T cells in blood or skin identified in our phenotypic, functional, and transcriptional analyses. Category names were assigned on the basis of described functions of the indicated genes in the published literature. Underlined gene names indicate proteins whose expression pattern was validated by flow cytometry in Figs. 3 and 5.

CD103 expression by cutaneous TRM is induced upon their migration into the skin by TGF-β (25), which is produced and activated by epidermal keratinocytes and dermal fibroblasts (2, 26, 27). Consistent with this, several TGF-β–related genes were up-regulated in both circulating and skin-resident CD4+CLA+CD103+ T cells (Fig. 4D and fig. S6). Moreover, along with ITGAE (the gene encoding CD103), CD27, and CD101, we identified additional TRM-associated genes that were differentially expressed by circulating CD4+CLA+CD103+ T cells, including CCR8 (28), CXCR6 (3), EOMES (29), and PPARG (30). We also identified overlapping functional modules of genes differentially expressed by CD4+CLA+CD103+ T cells that control cellular migration and adhesion and that modulate host defense and tissue inflammation (Fig. 4D and fig. S6).

To complement this transcriptomic study, we assessed the function of memory T cells from skin and blood after ex vivo stimulation and intracellular cytokine staining. Among effector cytokines, production of interleukin-22 (IL-22) and IL-13 was significantly enriched in CD4+CLA+CD103+ T cells from both tissues, but these cells were largely negative for interferon-γ (IFN-γ), IL-17A, or IL-4 (Fig. 5). Although granulocyte-macrophage colony-stimulating factor (GM-CSF) production was also highly enriched in CD4+CLA+CD103+ T cells in the blood, production in the skin was low in all T cell populations (Fig. 5, D and E). This cytokine phenotype is consistent with that of TH22 cells (31, 32) and distinguished CD4+CLA+CD103+ T cells from CD4+CLA+CD103 and CD4+CLACD103 T cells in both skin and blood. Coproduction of IL-22 and IL-13 (fig. S7) is indicative of a role for CD4+CLA+CD103+ T cells in promoting normal tissue homeostasis and repair in the skin (33, 34). Consistent with this, CD4+CLA+CD103+ T cells in the blood and skin differentially expressed a set of genes implicated in tissue repair responses, such as CD9 (24), MUC16 (35, 36), and LGALS3 (37), as well as the receptor components for the damage-associated epithelial cell products IL-25 (IL17RA and IL17RB) and prostaglandin E2 (PTGER3) (Fig. 4F and fig. S6) (33, 38, 39). CD4+CLA+CD103+ T cells in both the skin and blood also expressed IL26, a cytokine that has direct antimicrobial activity against Staphylococcus aureus and other extracellular bacteria (40, 41). Thus, the CD4+CLA+CD103+ T cell population is functionally well suited to promote both tissue repair and host protective responses in the skin.

Fig. 5 CD4+CLA+CD103+T cells from human blood and skin share a functional profile.

(A) Representative flow cytometric analysis of CD103 and CLA expression by live gated CD4+CD45RA T cells from blood and skin of healthy donors. (B to D) Representative flow cytometric analysis of indicated CLA/CD103 subpopulations of blood and skin CD4+CD45RA T cells producing IL-13, IL-4, IL-22, IL-17A, IFN-γ, and GM-CSF as indicated upon ex vivo stimulation with PMA/ionomycin and intracellular cytokine staining. (E) Graphical summary of the proportions of CLACD103, CLA+CD103, and CLA+CD103+ live gated CD4+CD45RA T cells producing cytokines as indicated. Open symbols represent data from an individual with mammary carcinoma. Significance was determined by one-way repeated-measures ANOVA with Tukey’s posttest for pairwise comparisons.

CLA+CD103+ T cells from human blood and skin are clonally related

Their shared phenotype, transcriptional signature, and functions suggest that the CD4+CLA+CD103+ T cells in blood and skin are closely related and may represent circulating and skin-resident fractions of the same TRM population. To directly determine whether CD4+CLA+CD103+ T cells in the skin and blood have a shared clonal origin, we performed TCRβ sequencing on CD4+CLA+ memory CD4+ T cells from paired blood and skin samples from four individual donors that were sorted as in fig. S5. Analysis of unique CDR3 clonotypes showed that sequences from skin CD4+CLA+CD103+ T cells were found at high frequency in the circulating CD4+CLA+CD103+ T cells and also showed some overlap with CD4+CLA+CD103 skin T cells. By contrast, little clonal overlap was observed with circulating CLA+CD103CCR7+ and CLA+CD103CCR7 T cells (Fig. 6A and fig. S8A). Quantitative analysis of TCR repertoire overlap using the Morisita index (42), which accounts for both the species presence and abundance, confirmed that the repertoire of skin CD4+CLA+CD103+ T cells is most similar to that of the circulating CD4+CLA+CD103+ T cell population (Fig. 6A, right). Reciprocally, circulating CD4+CLA+CD103+ cells from the blood showed extensive TCR repertoire overlap with skin CLA+CD103+ T cells but little clonal similarity with the other circulating CD4+CLA+ T cell populations (Fig. 6B and fig. S8A). Together, these analyses demonstrate the shared clonal origin of the CD4+CLA+CD103+ T cell populations in the blood and skin, thereby providing strong evidence that CLA+CD103+ T cells in the blood represent the circulating counterpart of the cutaneous CLA+CD103+ TRM population. CD4+CLA+CD103+ T cells in both blood and skin showed diverse Vβ usage (fig. S8B), and there was virtually no overlap in the TCRβ sequences in any of the populations examined between the four different donors. Thus, CD4+CLA+CD103+ T cells in the blood and skin do not appear to be a clonally restricted or invariant T cell population such as NK T cells or mucosa-associated invariant T cells (43).

Fig. 6 CD4+CLA+CD103+ T cells in skin and blood are clonally related.

TCRβ sequencing was performed on sorted CLA+ memory CD4+ T cell subsets (as in Fig. 4) from blood and skin sorted on the basis of the expression of CD103 and CCR7 as indicated. (A) Left: Circle plot of unique productive TCRβ sequences from each of the indicated populations of CLA+ T cells from one representative donor (donor 1). Connections highlight sequences from skin CLA+CD103+ T cells found in each of the other populations. For visualization, sequences were downsampled (weighted for relative abundance) for populations containing >1000 unique sequences. Right: Graphical summary of the Morisita index as a measure of TCR repertoire similarity over all productive rearrangements between CLA+CD103+ T cells in the blood and each of the indicated populations across all four donors examined. (B) Circle plot (left) and graphical summary of the Morisita index (right) as in (A) using CLA+CD103+ T cells from blood as the reference population. Significance was determined by one-way ANOVA with Dunnett’s multiple comparisons test.

CD4+CLA+CD103+ T cells recirculate from tissues via the lymphatics

To further determine whether CD4+CLA+CD103+ T cells recirculate from the tissue under physiological conditions, we analyzed human thoracic duct lymph (TDL) collected from patients with chylothorax. As in the blood and skin, we identified a CD4+CLA+CD103+ T cell population that lacked the expression of CCR7 (Fig. 7, A to C). Moreover, a majority of these cells expressed CCR4 and CCR6 but were low for CD27 and negative for CD69 expression (Fig. 7D). CCR7 is required for trafficking to the lymph node via the blood, and it is therefore unlikely that these CCR7 T cells are recirculating by exiting the blood to the lymph node and then returning to the circulation via the TDL (44). Rather, this further supports the idea that circulating CD4+CLA+CD103+ T cells in blood and TDL have exited directly from the skin.

Fig. 7 CD4+CLA+CD103+T cells are present in human lymph.

(A) Representative flow cytometric analysis of CD45RA and CLA expression by live gated CD4+ T cells from blood and TDL. (B) Representative flow cytometric analysis of CCR7 and CD103 expression by live gated CD4+CD45RACLA+ memory T cells from blood and TDL. (C) Graphical summary of the proportions of CCR7- and CD103-defined T cell populations among CD4+CD45RACLA+ T cells from blood and TDL. (D) Representative flow cytometric analysis and graphical summary of expression of the indicated markers by CD4+ T cell populations in the blood and TDL as indicated. Significance was determined by one-way repeated-measures ANOVA with Tukey’s posttest for pairwise comparisons.

Circulating CD4+CLA+CD103+ TRM can reseed distant skin sites

Exit of cutaneous CD4+CLA+CD103+ T cells and their recirculation may allow them to migrate to distant tissue sites, thereby promoting the efficient distribution of functionally specialized T cells throughout the skin. To directly test this hypothesis in vivo, we used a skin-xenografting mouse model designed to track tissue exit of human cutaneous T cells and their subsequent migration to secondary human skin sites. In this system, cultured human keratinocytes and fibroblasts are placed in a grafting chamber that is surgically implanted on NSG mice. The cells undergo spontaneous cell sorting to form epidermal and dermal layers, generating engineered skin (ES) tissue with histological features of human skin and the organotypic expression of structural proteins such as human type VII collagen at the epidermal-dermal junction (Fig. 8A) (45). Thus, the ES closely resembles human skin but lacks resident immune cells, and therefore, T cell migration into the ES can be definitively monitored.

Fig. 8 CD4+CLA+CD103+TRMcan exit the skin and reseed distant skin sites in a xenograft model.

(A) In vitro expanded human keratinocytes and fibroblasts were grafted onto the backs of NSG mice using a grafting chamber. After 99 days of healing and differentiation, the ES or adjacent murine skin was excised, frozen in O.C.T., and stained either with hematoxylin and eosin (H&E) (left) or with anti-human type VII collagen before immunofluorescence analysis (right). Human skin from a healthy donor was used as control. (B) Experimental schematic for the generation of ES followed by xenografting human skin onto NSG mice. (C) Representative photograph of ES and skin grafts on day 144. (D) Representative flow cytometric analysis and (E) graphical summary of CLA+CD103+ T cells by live gated human CD45+CD3+CD4+CD45RA T cells from skin grafts, spleen, and ES (3 to 5 weeks after skin grafting). Open and filled symbols denote samples derived from two different skin donors. Each symbol represents data from one recipient animal. (F) Representative flow cytometric analysis and graphical summary of expression of CD27, CD9, and CD69 by live gated CD45+CD4+CD45RACD103+CLA+ T cells in the skin grafts, spleen, and ES 5 weeks after skin grafting (day 145 relative to ES generation). Significance was determined by one-way ANOVA with Tukey’s posttest for pairwise comparisons. (G) Experimental schematic for the generation of ES followed by adoptive transfer of 2.5 × 106 PBMCs (autologous to the ES)/mouse into NSG mice. i.v., intravenous. (H) Representative flow cytometric analysis and (I) graphical summary of CLA+CD103+ T cells by live gated human CD45+CD3+CD4+CD45RA T cells from spleen and ES 25 days after PBMC transfer. Each symbol represents data from one recipient animal. Significance was determined by paired t test.

After healing of the ES (>110 days), mice received skin grafts from healthy donors, and tissues were analyzed 3 to 5 weeks later (Fig. 8, B and C). As in Fig. 1 (F and G), CD4+CLA+CD103+ T cells had exited the skin grafts and were found in the spleens of all recipient animals. In addition, in five of seven recipient mice, CD4+CLA+CD103+ T cells also migrated to the ES (Fig. 8, D and E), whereas no human cells were found in adjacent murine skin (fig. S9). Similar to what we observed in human blood and skin, CD9 and CD69 were downmodulated on CD4+CLA+CD103+ T cells found in the spleen but were reexpressed by cells entering the ES and CD27 expression remained low in all tissue sites (Fig. 8F). Last, we used the ES system to interrogate the in vivo migration behavior of circulating CD4+CLA+CD103+ T cells from the blood. Upon transfer of PBMCs into NSG mice carrying ES grafts (Fig. 8G), we found that T cells migrated to the ES and that CD4+CLA+CD103+ T cells were significantly enriched in the ES versus the spleen (Fig. 8, H and I). Together, these data demonstrate that CD4+CLA+CD103+ T cells that exit from the skin upon grafting or that are found in the blood of healthy individuals have the ability to migrate to and populate secondary skin sites.

DISCUSSION

TRM populations mediate optimal protective responses to site-specific challenges in nonlymphoid tissues and are most readily identified by their expression of CD69 and/or CD103 (2, 4). Using tissue explant cultures and a skin-xenografting mouse model, we made the unexpected discovery that CD4+CLA+CD69+CD103+ T cells in human skin can down-regulate CD69 and exit the tissue. This is consistent with recent observations in murine systems showing that secondary stimulation mobilized CD8+ T cells from nonlymphoid tissues (including the skin) that subsequently established residence within draining secondary lymphoid organs (46). However, this study did not establish from which skin-resident population these cells were derived or whether these mobilized cells could further recirculate to other peripheral nonlymphoid tissue sites. We also identified CD4+CLA+CD103+ T cells as a distinct population of circulating T cells that are clonally related to the CD4+CLA+CD103+ TRM population in the skin. The phenotypic, functional, and transcriptional profile of circulating CD4+CLA+CD103+ T cells is consistent with their origin and function within the skin, a TGF-β–rich barrier site exposed to microbial threats and frequent tissue damage. Last, we show that upon exiting the skin, CD4+CLA+CD103+ T cells can migrate via the circulation to secondary skin sites where they reacquire markers of tissue residency such as CD69. On the basis these features, we propose that blood and skin CD4+CLA+CD103+ T cells represent components of the same TRM population that undergoes basal recirculation but is maintained in substantial disequilibrium between these tissues.

Our data challenge current concepts regarding the strict tissue compartmentalization of T cell memory in humans and instead support a model in which cells of the CD4+CLA+CD103+ TRM population can transiently forgo their cutaneous location before reassuming residency at distant skin sites. Whether tissue exit of cutaneous CD4+CLA+CD103+ T cells is a stochastic process or actively triggered mobilization remains to be determined. In the context of our studies in explant cultures and in skin-humanized mice, tissue damage unavoidably associated with surgical skin acquisition is one potential trigger that may have affected TRM mobilization. However, we detected CD4+CLA+CD103+ cells in the blood and lymph of all donors analyzed, which indicates that a small fraction of the CD4+CLA+CD103+ TRM population recirculates even in the absence of clinical skin infection, inflammation, or tissue damage.

Our xenografting studies allowed us to further follow the fate of tissue-derived T cells, and we found that CD4+CLA+CD103+ T cells from either skin grafts or human blood can migrate to and preferentially seed secondary human skin sites. This tissue seeding occurred in the absence of tissue damage or local inflammation since the recipient ES tissue was fully healed (>110 days) and thus lacked expression of damage-associated molecules such as IL-1α, IL-1β, IL-18, and TNF-α that might increase recruitment of circulating cells (47). However, it remains possible that recruitment to secondary skin sites is increased by damage-associated stress. This might offer an intriguing explanation to the hitherto unexplained Koebner phenomenon, in which lesions in TRM-mediated diseases such as psoriasis and mycosis fungoides can spread to otherwise healthy (noninfected) skin sites upon triggers such as mechanical trauma, burns, friction, or ultraviolet irradiation (4850).

Human skin–resident αβ T cells can promote tissue repair in skin organ culture models (51). Although the precise role of the CD4+CLA+CD103+ T cell population in skin immunity and homeostasis remains to be established in vivo, their transcriptional and functional profile is indicative of a function in wound healing and tissue repair responses. The signature cytokines produced by CD4+CLA+CD103+ T cells, IL-22 and IL-13, both have important tissue repair functions in the skin. IL-22 acts directly on keratinocytes to promote their survival, proliferation, migration, and antimicrobial functions (34), whereas IL-13 activates cutaneous fibroblasts and promotes M2 macrophage differentiation and wound healing (52). IL-22 can also induce production of antimicrobial peptides by keratinocytes (53). Thus, the CD4+CLA+CD103+ population has many of the hallmarks of other lymphocyte populations implicated in antimicrobial and tissue repair responses, such as cutaneous IL-22–producing γδ T cells, and IL-13– or IL-22–producing innate lymphoid cells (34, 54, 55). In these contexts, note that we and others found CD4+CLA+CD103+ T cells in the epidermis and the dermis of the skin (2), and thus, they are ideally positioned to modulate the responses of keratinocytes, fibroblasts, and skin macrophages and promote both tissue repair and host protective responses.

Mobilization of cutaneous CD4+CLA+CD103+ T cells to the circulation would support the distribution of immunity in a large barrier organ such as the skin, as well as provide a reservoir of specialized circulating T cells that could be rapidly recruited to infected or damaged skin to promote host defense and tissue repair. Our identification of CD4+CLA+CD103+ T cells as a unique population of circulating T cells in healthy individuals greatly facilitates the isolation and study of cutaneous TRM from a broadly available human tissue, the blood. This observation may yield new insights into the biology and function of the human skin TRM population and also provides an opportunity for therapeutic manipulation of skin TRM in the contexts of cutaneous autoimmunity, infection, and tissue repair.

MATERIALS AND METHODS

Study design

The objectives of this research were to characterize the phenotype, function, and migratory behavior of CD4+ T cell populations that express CLA and to define the relationship between CLA+ T cells in the blood and CLA+ TRM in the skin. This was accomplished using blood and skin samples from healthy donors by flow cytometric analysis of cellular phenotype and function, transcriptomic analysis by RNA-seq, TCR clonotype analysis by TCR sequencing, and experimental studies of cellular behavior in explant culture models and skin xenograft studies using immunodeficient mice. Blood samples from healthy donors were obtained by standard phlebotomy. Normal human skin was obtained from patients undergoing elective surgery (panniculectomy and elective breast reduction), in which skin was discarded as a routine procedure. In one case, skin and blood were obtained from a treatment-naïve individual undergoing surgery for mammary carcinoma, and data from this individual are specifically marked in the figures. Samples of individuals of both sexes were included in the study. Ages ranged from 17 to 70. All samples were obtained upon written informed consent at the University Hospital Salzburg, Austria, the University of Pennsylvania and the Children’s Hospital of Philadelphia (Philadelphia, PA), or the Virginia Mason Medical Center in Seattle, WA, USA. All studies were approved by the Salzburg state Ethics Commission (decision: according to Salzburg state hospital law no approval required) (Salzburg, Austria), the Institutional Review Board of the University of Pennsylvania and the Children’s Hospital of Philadelphia (Philadelphia, PA), or the Institutional Review Board of the Benaroya Research Institute (Seattle, WA). NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice were obtained from The Jackson Laboratory and bred and maintained in a specific pathogen–free facility in accordance with the guidelines of the Central Animal Facility of the University of Salzburg. All animal studies were approved by the Austrian Federal Ministry of Science, Research and Economy. No statistical method was used to predetermine sample size. Sample sizes were selected on the basis of the availability of human blood and skin specimens and were large enough to achieve a greater than 80% probability of identifying an effect of >20% in measured variables. Two to three independent experiments (biological replicates) were conducted to validate each finding.

Skin explant cultures

Skin was washed in phosphate-buffered saline (PBS) with 1% penicillin/streptomycin and 0.1% Primocin (InvivoGen, ant-pm-1) for 5 min. Small skin pieces of 1 to 2 mm were generated using forceps and sharp scissors. Pieces were placed in a 60-mm dish and allowed to adhere for 10 min. Crawl-out medium consisting of 50% EpiLife (Gibco, MEPICF500) and 50% RPMI-complete [RPMIc: RPMI 1640 (Gibco, 31870074) with 5% human serum (Sigma-Aldrich, H5667 or H4522), 1% penicillin/streptomycin (Sigma-Aldrich, P0781), 1% l-glutamine (Gibco, A2916801), 1% non-essential amino acid solution (NEAA) (Gibco, 11140035), 1% sodium pyruvate (Sigma-Aldrich, S8636), and 0.1% β-mercaptoethanol (Gibco, 31350-010)] was added to explant cultures. Seven days later, cells in culture medium were analyzed by flow cytometry.

Cytometry by time of flight

Human frozen PBMCs were thawed and rested for 12 to 15 hours. The samples were washed with Ca- and Mg-free PBS (Sigma, D8537) and stained with 50 μM cisplatin (Enzo Life Sciences, ALX-400-040-M250) in PBS for 1 min to exclude dead cells. The cells were washed and resuspended with Human TruStain FcX (BioLegend, 422302) for 5 min before adding the primary surface staining cocktail for 20 min, washing, and staining with the secondary surface cocktail for 20 min. Intracellular staining was performed after fixation and permeabilization using the Maxpar Nuclear Antigen Staining Buffer Set (Fluidigm, 201063) for 60 min, after which cells were incubated overnight at 4°C with Maxpar Fix and Perm Solution containing 125 nM Cell-ID Intercalator-Ir (Fluidigm, 201192A) for DNA staining. Cells were washed with MilliQ H2O and resuspended in MilliQ H2O spiked with 1/20th Maxpar EQ Four Element Calibration Beads (Fluidigm, 201078) to a density of <5 × 105 cells/ml. Data were acquired on a CyTOF 1.5 (Fluidigm) instrument. For analysis, FCS files on gated CD3+CLA+ cells were generated using FlowJo software (Tree Star Inc.). t-SNE analysis was performed in R using the cytofkit R package, and clustering was performed with FlowSOM.

T cell isolation from human skin for flow cytometry and RNA-seq

For isolation of cells from full-thickness skin, tissue was minced and digested using collagenase type 4 and deoxyribonuclease (DNase), as previously described (56). Briefly, subcutaneous fat was removed before skin was minced with dissection scissors and surgical scalpel. About 1 cm2 of skin was digested overnight in 5% CO2 at 37°C with 3 ml of digestion mix containing collagenase type 4 (0.8 mg/ml; Worthington, LS004186) and DNase (0.02 mg/ml; Sigma-Aldrich, DN25) in RPMIc. Samples were filtered, washed with RPMI and PBS, and stained for flow cytometry/cell sorting or stimulated for intracellular cytokine staining. For separation of human epidermis and dermis, the skin was washed in PBS with 1% penicillin/streptomycin and 0.1% Primocin for 5 min before removal of fat tissue. Skin pieces of 2 to 3 mm in diameter were incubated overnight with 0.25% Dispase II (Sigma-Aldrich, D4693) in RPMIc, and dermis and epidermis were separated with forceps. The separated tissue layers were then further minced and digested with collagenase type 4 and DNase to generate single-cell suspensions for flow cytometric analysis.

PBMC isolation for flow cytometry and RNA-seq

Human PBMCs were isolated using Ficoll-Hypaque (GE Healthcare, GE17-1440-02) gradient separation. For RNA-seq, CD4+ T cells were enriched using CD4 microbeads (Miltenyi, 130-045-101), and 4 × 106 cells/ml were resuspended in RPMIc with IL-2 (50 U/ml; ImmunoTools, 11340023) in 24-well plate wells and stimulated for 20 hours with ImmunoCult CD3/CD28 T cell activator (25 μl/ml; STEMCELL, 10971) before staining and cell sorting.

Flow cytometry

Cryopreserved PBMCs or TDL cells were rested overnight at 37°C and 5% CO2 in complete medium [RPMI supplemented with 10% fetal bovine serum (FBS), 2 mM l-glutamine, penicillin (100 U/ml), and streptomycin (100 mg/ml)], washed, and stained for flow cytometry. For sorting, cells were stained in fluorescence-activated cell sorting (FACS) buffer (PBS + 1% FBS + 1 mM EDTA) for surface markers. For detection of intracellular cytokine production, skin single-cell suspensions and PBMCs were stimulated with phorbol 12-myristate 13-acetate (PMA) (50 ng/ml; Sigma-Aldrich, P8139) and ionomycin (1 μg/ml; Sigma-Aldrich, I06434) with brefeldin A (10 μg/ml; Sigma-Aldrich, B6542) for 3.5 hours. For permeabilization and fixation, a Cytofix/Cytoperm kit was used (Becton Dickinson, RUO 554714). Data were acquired on LSR Fortessa, LSRII, FACSymphony (all from BD Biosciences), or Cytoflex LS (Beckman Coulter) flow cytometers and analyzed using FlowJo software (FlowJo LLC).

Antibodies and reagents

A detailed list of the antibodies and reagents used can be found in table S1.

Analysis of TDL samples

TDL was collected and processed from patients with chylopericardium and chylothorax as previously published (57, 58). Cells were washed in PBS and stained for CXCR3, CCR4, CCR6, and CCR7 at 37°C, 5% CO2 for 20 min. All subsequent incubations were performed in the dark at room temperature. Cells were stained for viability exclusion for 10 min. Antibody cocktail mix was added and incubated for 20 min. Cells were fixed with PBS with 1% paraformaldehyde and stored in the dark at 4°C overnight until acquisition. Data were collected on a BD FACSymphony flow cytometer (BD Biosciences) and analyzed using FlowJo software (FlowJo LLC).

RNA sequencing

Five hundred cells per population were sorted into lysis buffer using a BD FACSARIA III or FACSFUSION instruments (BD Biosciences), and complementary DNA (cDNA) was prepared using the SMART-Seq v4 Ultra Low Input RNA Kit for Sequencing (Takara). Library construction was performed using the Nextera XT DNA sample preparation kit (Illumina) using half the recommended volumes and reagents. Dual-index, single-read sequencing of pooled libraries was run on a HiSeq2500 sequencer (Illumina) with 58-base reads and a target depth of 5 million reads per sample. Base calling and demultiplexing were performed automatically on BaseSpace (Illumina) to generate FASTQ files.

RNA-seq analysis

The FASTQ files were processed to remove reads of zero length (fastq_trimmer v.1.0.0), remove adapter sequences (fastqmcf tool v.1.1.2), and perform quality trimming from both ends until a minimum base quality ≥ 30 (FASTQ quality trimmer tool v.1.0.0). Reads were aligned to the human reference genome (build hg38) with TopHat (v.1.4.0), and read counts per Ensembl gene ID were quantified with htseq-count (v.0.4.1). Quality metrics for the FASTQ and BAM/SAM files were generated with FastQC (v.0.11.3) and Picard (v.1.128). Processing of FASTQ and BAM/SAM files was executed on the Galaxy workflow platform of Globus Genomics. Statistical analysis of gene expression was assessed in the R environment (v.3.4.4). Samples with a total number of fastq reads below 106, mapped reads below 70%, or median coefficient of variation (CV) of coverage > 1 were excluded from further analysis. Mapping Ensembl Gene IDs to Human Gene Nomenclature Committee (HGNC) gene symbols was achieved through biomaRt (GRCh38.p10). Genes were filtered for protein coding genes and those with an expression of counts per million (CPM) > 2 in at least 10% of the libraries. A linear model for gene expression was fit to the filtered 12,293 genes using limma (v3.34.9) (59), considering donor effects through a random factor. For visualizations, the random effect of the model was approximated by removing the donor effect from the expression data with limma::removeBatchEffect. Genes found to be significantly different (adjusted P < 0.05 and fold change > 2) between CD103+ and CD103CCR7+ cells as well as between CD103+ and CD103CCR7 cells in the blood were defined as the CD103+ gene signature. Enrichment of the CD103+ gene signature in the ranked list of CD103+ cells versus CD103CCR7 cells in the skin was visualized with limma::barcodeplot, and significance was determined by rotation gene set testing with limma::roast.

TCRβ sequencing and analysis

A minimum of 2000 T cells from the indicated populations was sorted into RPMIc, and genomic DNA was prepared using the QIAamp DNA Micro Kit (Qiagen). Amplification and sequencing were performed using the immunoSEQ Assay (Adaptive Biotechnologies, Seattle, WA), which combines multiplex polymerase chain reaction with high-throughput sequencing and a bioinformatics pipeline for TCRβ CDR3 analysis (60). Data analysis was performed using Adaptive Biotechnologies ImmunoSeq Analyzer 3.0 software and R version 3.5.1. Data for the analysis in R were exported through the export function in the Rearrangement details view. Circle plots for individual donors were created by downsampling populations with more than 1000 unique rearrangements (weighted based on relative abundance of each individual clonotype) and matching TCR chains with the R package TCRtools (https://github.com/mjdufort/TCRtools). Links between the blood or skin CD103+ reference population and all other populations are displayed in the circle plot. For V gene usage, we removed unknown and ambiguous mappings and computed the percentage of clones using each V gene among each sample. The plot includes V genes that have a usage ≥5% in at least one sample.

Generation of ES

Human keratinocytes and fibroblasts were isolated from normal human skin and immortalized using human papilloma virus type oncogenes E6/E7 as previously described (61). These were cultured in EpiLife (Gibco, MEPICF500) and Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, 11960-044) containing 2% l-glutamine, 1% penicillin/streptomycin, and 10% FBS. For transplantation, 80% confluent cells were trypsinized (TrypLE Express, Gibco, 12604021), washed with PBS, and counted. ES tissue was generated in vivo in mice by placing 1 × 106 to 2 × 106 keratinocytes mixed 1:1 with autologous fibroblasts in 400 μl of MEM (Gibco, 11380037) containing 1% FBS, 1% l-glutamine, and 1% NEAA in grafting chambers as previously described (45).

Transplantation of human skin or PBMC transfer

Punch biopsies (8 mm) of human skin were trimmed to an average thickness of 0.5 to 1 mm. Transplants were soaked in PBS + 1% penicillin/streptomycin + 0.1% Primocin (Invitrogen, ant-pm-1) for 5 min and kept on ice in a sterile container with PBS-soaked gauze until transplantation. NSG mice were anesthetized, and full-thickness wound bed was prepared using surgical scissors. Three grafts per mouse were placed on the back and fixed using surgical skin glue (Histoacryl, B. Braun). Transplants were covered with paraffin gauze dressing and bandaged with self-adhesive wound dressing. Bandages were removed after 7 days. In other experiments, PBMCs were thawed and rested in medium overnight before transfer into NSG recipient mice (2.5 × 106 cells per animal). After transfer of human cells (PBMCs or skin grafting), mouse neutrophils were depleted with anti–Gr-1 antibody (100 μg per animal intraperitoneally every 5 to 7 days; InVivoMab clone RB6-8C5) (62, 63).

Histological staining of skin sections

Normal human skin, ES grafts, and adjacent murine skin were excised and frozen in Tissue-Tek O.C.T. Compound (Sakura, TTEK). Cryosections (7 μm) were stained with Hemalum solution acid (Carl Rorth, T865.1) and Eosin Y aqueous solution (Sigma, 201192A). Human type VII collagen was stained by immunofluorescence using anti–human type VII collagen antibody [anti–NC-1 domain of type VII collagen (LH7.2) provided by A. Nyström, University of Freiburg, Germany] and goat anti-rabbit immunoglobulin G (IgG) Alexa Fluor 488 (Thermo Fisher, A11008) secondary antibody, and nuclear 4′,6-diamidino-2-phenylindole (DAPI) staining (ProLong Gold Antifade Mountant with DAPI, Invitrogen, P36931).

Tissue preparation from mice

Mice were euthanized using CO2 asphyxiation followed by cervical dislocation. Single-cell suspensions were generated from spleen, ES, and human skin grafts, and leukocytes were analyzed by flow cytometry. For T cell isolation from murine skin for flow cytometry, about 3 cm2 of shaved dorsal mouse skin was harvested and single-cell suspensions were prepared as in (64) and stained for flow cytometry.

Statistical analysis

Statistical significance of data was calculated with Prism 6.0 software (GraphPad) by one-way analysis of variance (ANOVA) with Tukey’s or Dunnett’s multiple comparisons test or by paired t test as indicated. Error bars indicate means ± SD.

SUPPLEMENTARY MATERIALS

immunology.sciencemag.org/cgi/content/full/4/37/eaav8995/DC1

Fig. S1. Representative flow cytometry gating strategies used to identify T cell subsets in human blood and skin.

Fig. S2. CD103 expression is not induced on human CD4+ T cells in NSG mice.

Fig. S3. CyTOF analysis of CLA+ T cells in PBMCs.

Fig. S4. Frequencies of CLA+ T cell subsets in blood and skin.

Fig. S5. Experimental schematic of cell isolation and sort gates for RNA-seq.

Fig. S6. Gene expression in the different populations of CD4+CLA+ T cells from blood and skin as determined by RNA-seq.

Fig. S7. CLA+CD103+ T cells from human blood and skin coproduce IL-22 and IL-13.

Fig. S8. Analysis of TCRβ repertoire overlap and Vβ gene usage in CLA+ T cells from blood and skin.

Fig. S9. Human skin–derived immune cells do not infiltrate murine skin.

Table S1. Detailed list of antibodies and reagents.

Table S2. RNA-seq pairwise comparisons.

Table S3. Raw data file.

REFERENCES AND NOTES

Acknowledgments: We thank A. Sir and R. Reitsamer (Breast Center of the University Hospital Salzburg, Paracelsus Medical University Salzburg, Austria) for providing skin and blood samples. We also thank A. Hochreiter and A. Raninger from the Flow Cytometry Core facility at the Cell Therapy Institute, PMU Salzburg. We thank S. Hainzl (EB House Austria, Department of Dermatology, University Hospital of the Paracelsus Medical University Salzburg, Austria) for the immortalization of primary human keratinocytes and fibroblasts. We further thank A. Benedetti for technical assistance and for performing the skin histology. At the Benaroya Research Institute in Seattle, we thank K. Benoscek and F. Roan for help in obtaining skin and blood samples, T. Son-Nguyen for frozen PBMCs, A. Weidemann for help with CyTOF, V. Gersuk for help with RNA-seq, and S. Presnell, M. Dufort, and H. DeBerg for helpful discussions on RNA-seq analysis. We also thank M. Itkin (University of Pennsylvania) and Y. Dori (Children’s Hospital of Philadelphia) for obtaining TDL samples. Funding: This work was supported by the Focus Program “ACBN” of the University of Salzburg, Austria; a grant from the Dystrophic Epidermolysis Bullosa Research Association (DEBRA) International and DEBRA Austria to I.K.G.; NIH grant R01AI127726 awarded to I.K.G. and D.J.C.; and NIH grants R01AI076066 and R01AI118694 awarded to M.R.B.; M.M.K. was part of the PhD program Immunity in Cancer and Allergy, funded by the Austrian Science Fund (FWF; grant W 1213), and received a DOC Fellowship from the Austrian Academy of Sciences. Author contributions: M.M.K., P.A.M., B.H., S.R.V., T.D., D.J.C., and I.K.G. designed the experiments and participated in the statistical analysis of the data; M.M.K., M.D.R., and I.K.G. developed xenografting methods; M.M.K., P.A.M., B.H., S.R.V., S.J.M., and T.D. performed experiments; B.H. performed computational analysis; L.K.-C. and E.G. performed analysis of TDL samples; M.R.B. planned and supervised the analysis of TDL samples; S.A.L. helped develop the CyTOF panels; G.B. helped set up the sort panels; D.J.C. and I.K.G. wrote the manuscript; M.M.K., P.A.M., B.H., S.R.V., S.J.M., T.D., and M.D.R. reviewed and edited the paper; D.J.C. and I.K.G. supervised the project. Competing interests: M.M.K. and I.K.G. are inventors on a pending patent application covering use of the skin-humanized mouse model (European file number EP18168258; U.S. file number US16389821). The other authors declare that they have no competing interests. Data and materials availability: The complete RNA-seq data are available from the Gene Expression Omnibus under accession number GSE131770. Pairwise comparisons between T cell populations are included in table S2. The CyTOF data are available at flowrepository.org under accession number FR-FCM-Z24B. HPV16 E6/E7–immortalized human keratinocyte and skin fibroblast cell lines are available to interested investigators through a material transfer agreement with the University of Salzburg. All other data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.
View Abstract

Stay Connected to Science Immunology

Navigate This Article