Abstract
The success of the cutaneous immune system reflects its ability to rapidly and efficiently recruit leukocytes to areas of trauma and infection. Skin-homing memory T cells expressing cutaneous lymphocyte-associated Ag tether on the walls of postcapillary venules in inflamed skin via interaction with endothelial E-selectin and roll in response to the shear stress imparted by flowing blood. Rolling cells sample the vascular surface for chemoattractant compounds (e.g., thymus- and activation-regulated chemokine/CCL17 interacting with CCR4 on the leukocyte surface) and, if successfully stimulated, progress to firm arrest and transmigration mediated by LFA-1 and vascular ICAM-1. Although it is established that this sequence of events draws T cells into inflamed skin, the mechanisms directing trafficking of T cells to noninflamed skin are less well characterized. We hypothesized that basal expression and colocalization of E-selectin, chemokine (e.g., CCL17), and ICAM-1 in dermal vessels could serve to recruit T cells to noninflamed human skin. Immunohistochemical staining for E-selectin and CD31 demonstrated E-selectin expression in a restricted subset of dermal vessels in noninflamed human skin from three different sites. Confocal multicolor immunofluorescence imaging revealed a nonuniform distribution of E-selectin in dermal vessels as well as colocalization of E-selectin with CCL17 and ICAM-1. Coexpression of these molecules on blood vessels in noninflamed skin provides the basis for a model of cutaneous immunosurveillance system active in the absence of pathologic inflammation.
The homing of leukocytes from the blood to tissue is a critical component of mammalian immune responses. Because of its exposure to the outside environment, the skin must recruit leukocytes effectively to ward off potential pathogens. The importance of acquired immune responses in skin is highlighted by their absence in immunosuppressed patients, who are at higher risk for cutaneous infections and malignancies. The concept of a constitutive acquired immune surveillance system in noninflamed skin has been proposed, but is largely uncharacterized (1).
We hypothesized that noninflamed skin uses the same mechanisms as inflamed skin to recruit T cells. Trafficking of T cells to inflamed skin is known to require the interactions of molecules located on the endothelial cells of dermal blood vessels with their counterreceptors on T cells and can be described as a series of distinct steps: tethering of T cells to the blood vessel wall, rolling of tethered T cells in response to the shear force of blood flow, T cell activation and up-regulation of integrin avidity, firm adhesion via integrins to the vascular wall, extravasation into the underlying tissue, and chemotaxis toward the focus of inflammation. Tethering of T cells to vascular endothelia in shear flow is mediated primarily by E-selectin (CD62E), which is induced on endothelial cells by inflammation, binding to cutaneous lymphocyte-associated Ag (CLA),3 a carbohydrate epitope expressed on skin-homing memory T cells (2, 3). T cells tethered to the vascular wall via selectins roll due to the shear stress imparted by flowing blood. As they roll, T cells sample the endothelial surface for activating molecules such as chemokines, which are produced by resident skin cells and transported to the luminal surface of local endothelial cells (reviewed in Ref.4). Skin-homing T cells express the chemokine receptor CCR4, whose ligands include the chemokines thymus- and activation-regulated chemokine (TARC), recently renamed CCL17 (5), and macrophage-derived chemokine (CCL22) (6). Specific interactions between CCL17 and CCR4 lead to the up-regulation of the avidity of integrins on T cells, such as LFA-1 (CD11a/CD18, αLβ2), for their endothelial ligands, including ICAM-1 (CD54), which results in the firm arrest of T cells on the endothelium (7). Bound T cells then extravasate through the endothelial layer to enter the surrounding tissue, where they encounter further chemotactic stimuli directing them toward inflammatory sites. It is important to recognize that recruitment occurs only when a T cell successfully completes each of the steps preceding transmigration: tethering, activation, and firm adhesion. Lack of the appropriate counterreceptor or ligand at any of these stages will block transition to the next step, resulting in release of the T cell back into the circulation.
Although there is abundant evidence for skin-specific T cell homing in inflamed tissue, there is little known about the mechanisms regulating T cell trafficking in noninflamed skin. The ready identification of mature CLA+, CCR4+ memory T cells in normal skin (8, 9, 10) suggests that a method must exist for T cell recruitment to noninflamed skin. Moreover, this process may depend on the baseline expression of the vascular components involved in attracting CLA+, CCR4+ T cells to inflamed skin, specifically E-selectin, CCL17, and ICAM-1. As indicated by the cascade of events outlined above, colocalization of these factors in dermal vessels would also be required to support T cell extravasation into normal skin.
There is strong evidence for the constitutive presence of ICAM-1 in vascular endothelial cells in noninflamed skin (11, 12, 13). However, there is conflicting data as to whether E-selectin and CCL17 are constitutively expressed. Intravital microscopy experiments have shown spontaneous leukocyte rolling in the postcapillary venules of noninflamed rodent skin (14, 15) and have pinpointed the involvement of selectins in this process (16, 17). However, the issue has remained controversial in humans, with studies supporting both the presence (13, 18, 19) and absence (20, 21, 22) of E-selectin in noninflamed skin. Studies investigating CCL17 expression in normal human skin have been less abundant, but have also produced inconsistent results, reporting either the presence (7, 8, 23) or absence (24, 25) of CCL17. In addition, while these various reports have documented the individual expression of E-selectin, CCL17, and ICAM-1 in noninflamed skin, none have confirmed coexpression and have been sufficient to establish a model for a cutaneous immunosurveillance system.
In this study, we have sought to establish evidence for the basal expression and colocalization of E-selectin, CCL17, and ICAM-1 in noninflamed human skin. As indicated by the multistep recruitment model described above, these factors would fulfill the minimum criteria necessary to support the trafficking of CLA+ CCR4+ T cells to these sites. In this study, we report highly reproducible staining of E-selectin in noninflamed skin and demonstrate that it is restricted to a subset of CD31+ vessels in the mid to upper dermis. Confocal microscopy of 30-μm sections of normal human skin shows that E-selectin is not uniformly distributed on dermal blood vessel endothelium, but is expressed in an irregular, patchy distribution that may have significance for the immune surveillance model. We also document colocalization of E-selectin and CCL17; E-selectin and ICAM-1; and E-selectin, CCL17, and ICAM-1 in dermal vessels via two- and three-color immunofluorescence staining. These observations indicate that E-selectin, CCL17, and ICAM-1 are coordinately expressed by endothelia in noninflamed human skin and may form the basis for the constitutive recruitment of memory T cells to skin.
Materials and Methods
Abs and reagents
Mouse anti-human E-selectin mAb (IgG1, clone 68-5H11), FITC-conjugated mouse anti-human CD31 mAb (IgG1, clone WM-59), mouse IgG1 (clone MOPC-31C), and FITC-conjugated mouse IgG1 (clone MOPC-21) were obtained from BD PharMingen (San Diego, CA). Mouse anti-human CD31 mAb (IgG1, clone JC/70A) and HRP-conjugated streptavidin were purchased from DAKO (Carpinteria, CA). Rabbit anti-human TARC/CCL17 polyclonal Ab was obtained from PeproTech (Rocky Hill, NJ). Rabbit IgG was purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). FITC-conjugated mouse anti-human ICAM-1 mAb (IgG1, clone 84H10) was obtained from Immunotech (Marseille, France). Alexa Fluor 568-conjugated goat anti-mouse IgG, Alexa Fluor 647-conjugated goat anti-mouse IgG, and Alexa Fluor 568-conjugated goat anti-rabbit IgG were purchased from Molecular Probes (Eugene, OR). FITC-conjugated goat anti-rabbit IgG was obtained from Zymed Laboratories (San Francisco, CA). Biotinylated goat anti-mouse IgG was obtained from Vector Laboratories (Burlingame, CA).
Preparation of fresh human skin tissue
In compliance with local institutional review board policies, normal skin samples were obtained as discarded human tissue from skin-removal surgeries of 12 individual subjects without known inflammatory disorders (4 face-lift surgeries, 4 abdominoplasties, and 4 breast reduction surgeries). Skin samples were immediately trimmed to isolate tissue distal to the surgical incision sites, placed on ice, embedded in OCT (Miles, Elkart, IN), frozen in cold isopentane or in liquid nitrogen, and stored at −80°C. Six- or 30-μm adjacent sections were cut in a freezing cryotome (Leica, Bannockburn, IL) and placed onto charged microscope slides (Fisher, Hampton, NH).
H&E analysis
A standard H&E staining technique was used. The slides were subsequently analyzed by a dermatopathologist and showed small perivascular lymphocytic infiltrates with no lymphocytic accumulation in this tissue, which is consistent with normal noninflamed skin (Fig. 1⇓).
H&E staining of abdominal, breast, and facial skin samples recovered from elective surgeries. Minimal perivascular infiltrates are present in two sections of abdominal skin (A and B) and two sections of breast skin (C and D) removed from separate donors. Perivascular infiltrates seen in samples of face skin (E and F) are somewhat more prominent than those in abdomen and breast skin. There were no inflammatory cells identified in the tissue parenchyma of any of the samples tested. These findings were reviewed by a dermatopathologist and judged within normal limits and characteristic of noninflamed skin at each of these sites.
Immunohistochemical analysis
Sections were stained by the ABC-immunoperoxidase method, according to the manufacturer’s instructions (Vector Laboratories). Briefly, air-dried 6-μm sections were fixed in acetone for 10 min at 4°C and then immersed in a 0.03% hydrogen peroxide in PBS solution for 5 min. After blocking with 10% normal goat serum in PBS for 10 min, the sections were incubated with an avidin-biotin blocking solution (DAKO), followed by an overnight incubation at 4°C with the primary Ab (mouse anti-human E-selectin mAb, mouse anti-human CD31 mAb, or mouse IgG1) in a humid chamber. The slides were then washed in three changes of PBS over 30 min and incubated with biotinylated goat anti-mouse Ab for 30 min at room temperature. The slides were again washed and incubated with HRP-conjugated streptavidin for 30 min at room temperature. The tissue sections were then washed and stained using the NovaRed Substrate Kit (Vector), counterstained in methyl green, dehydrated in ethanol and xylene, and coverslipped with PolyMount mounting medium (Polysciences, Warrington, PA).
Two-color immunofluorescence analysis
A standard two-stage double-immunofluorescence labeling technique was used. Briefly, air-dried 30-μm sections were fixed in acetone for 15 min at 4°C, and treated with a blocking serum solution of 10% normal goat serum in PBS for 10 min. The sections were incubated with primary Ab (mouse anti-human E-selectin mAb or mouse IgG1) overnight at 4°C. Washed sections were then incubated with Alexa Fluor 568-conjugated goat anti-mouse IgG for 30 min at room temperature. For double staining of E-selectin and CD31 or ICAM-1, the slides were washed and incubated with a directly conjugated primary Ab (FITC-conjugated mouse anti-human CD31 mAb or mouse IgG1) for 30 min at room temperature, or FITC-conjugated mouse anti-human ICAM-1 mAb or mouse IgG1 for 2 h at room temperature. For double staining of E-selectin and CCL17, overnight primary Ab incubation (mouse anti-human E-selectin mAb or mouse IgG1) at 4°C and washing were followed by incubation with a second primary Ab (rabbit anti-human TARC/CCL17 polyclonal Ab or rabbit IgG) for 1 h at room temperature. The washed sections were then incubated with both Alexa Fluor 568-conjugated goat anti-mouse IgG and FITC-conjugated goat anti-rabbit IgG for 30 min at room temperature. After washing, the tissue sections were coverslipped with VectaMount mounting medium (Vector Laboratories). Tests for nonspecific binding between secondary Abs and subsequent primary Abs and for cross-reactive binding between primary and secondary Abs were negative and precluded any need for additional blocking steps (data not shown).
Three-color immunofluorescence analysis
Briefly, air-dried 30-μm sections were fixed in acetone for 15 min at 4°C, and treated with a blocking serum solution of 10% normal goat serum in PBS for 10 min. The sections were incubated with primary Ab (mouse anti-human E-selectin mAb or mouse IgG1) overnight at 4°C. Washed sections were then incubated with a second primary Ab (rabbit anti-human TARC/CCL17 polyclonal Ab or rabbit IgG) for 1 h at room temperature. Slides were washed and incubated with Alexa Fluor 647-conjugated goat anti-mouse IgG and Alexa Fluor 568-conjugated goat anti-rabbit IgG for 30 min at room temperature. Washed slides were incubated with a directly conjugated primary Ab (FITC-conjugated mouse anti-human ICAM-1 mAb or mouse IgG1) for 1 h at room temperature. After washing, the tissue sections were coverslipped with VectaMount mounting medium (Vector Laboratories). Tests for nonspecific and cross-reactive binding of both primary and secondary Abs were negative, and thereby precluded any need for additional blocking steps (data not shown).
Microscopic evaluation
Immunoperoxidase- and H&E-stained samples were viewed under an Eclipse E600 (Nikon, Tokyo, Japan) biological research microscope at ×40 and ×100 magnification. Immunofluorescence slides were examined using a Bio-Rad 1024-MP laser confocal microscope (Bio-Rad, Hercules, CA). For confocal microscopy, a series of optical sections spanning the z-axis was obtained from each slide by scanning the sample from the base to the apex at 2-μm intervals. A single composite image was then compiled for each slide by overlaying the optical sections. Sample images were compared with their respective controls.
Results
E-selectin is expressed in a subset of mid to upper dermal vessels in noninflamed skin
Expression of E-selectin was assessed in noninflamed skin obtained from healthy individuals undergoing elective surgery. Absence of inflammation in the skin samples used for analysis was determined by clinical observation and confirmed by histological analysis of H&E sections. Each sample showed a small number of perivascular lymphocytes, considered within the range of normal. Of note, all samples lacked tissue-infiltrating leukocytes or other evidence of subclinical inflammation (Fig. 1⇑). E-selectin was readily identified on blood vessels in sections of noninflamed facial skin, abdominal skin, and breast skin by avidin-biotin immunoperoxidase staining (Fig. 2⇓, A, D, and G). The general distribution of E-selectin-positive vessels in areas of the mid to upper dermis was similar in skin from all three regions of the body. Serial sections stained with an Ab to CD31, an established endothelial cell marker (Fig. 2⇓, B, E, and H), revealed the broader distribution of vessels in these three skin types, indicating that E-selectin is expressed on a restricted subset of endothelial cells in noninflamed skin. In contrast to the homogeneous CD31 staining seen on blood and lymphatic vessels, E-selectin immunoreactivity on individual vessels in noninflamed skin was not uniformly contiguous. Not only would some vessels appear E-selectin positive in one section and negative in an adjacent serial section (data not shown), but even within one vessel, E-selectin expression could be seen to vary from one area of the lumen to another (Fig. 2⇓G).
E-selectin and CD31 expression in serial cryostat sections of noninflamed human abdominal, breast, and face skin via immunoperoxidase staining. A small subset of mid to upper dermal vessels was immunostained with anti-human E-selectin mAb (red-brown) in noninflamed abdominal (A), breast (D), and face (G) skin. Noninflamed face skin (G) demonstrated focal E-selectin expression (open arrow) within an upper dermal vessel. All dermal vessels were identified in serial sections of the same skin tissue blocks by staining with anti-human CD31 mAb (red-brown) in noninflamed abdominal (B), breast (E), and face (H) skin. The specificity of E-selectin and CD31 immunostaining was demonstrated by the absence of signal using an isotype-matched control mAb in noninflamed abdominal (C), breast (F), and face (I) skin. Original magnification: A–F, ×100, G–I, ×40. (Note: The epidermis is oriented at the top in all immunoperoxidase and immunofluorescence pictures.)
E-selectin colocalizes with CD31 in a subset of mid to upper dermal vessels in noninflamed human skin
To further explore the expression of E-selectin on endothelial surfaces in noninflamed skin, two-color immunofluorescence staining for E-selectin and CD31 was performed. Confocal microscopy was used to examine the immunostained samples, allowing for the imaging of thicker sections and evaluation of more extensive vascular structures. The figures presented are images of 30-μm skin sections scanned at 2-μm intervals and displayed as overlays of the individual optical sections into single composite images. Examination of E-selectin- and CD31-stained abdominal, breast, and face skin showed overlap of immunoreactivity, indicating colocalization of E-selectin and CD31 in vessels in the mid to upper dermis (Fig. 3⇓, A, C, E, and G). As shown in Fig. 2⇑, E-selectin was seen in only a subset of the CD31-positive vessels present in skin. Moreover, focal variation in the distribution of E-selectin on CD31-stained vessels, as highlighted in Fig. 3⇓C, shows that E-selectin expression can be highly variable within a vessel and suggests control at the level of individual endothelial cells. This unusual expression pattern was consistent with the observations made with immunoperoxidase staining of 6-μm sections shown in Fig. 2⇑ and may help to explain some of the inconsistency regarding E-selectin expression reported in previous studies.
Two-color immunofluorescence staining for E-selectin and CD31 in noninflamed human abdominal, breast, and face skin. Coexpression of E-selectin (red) and CD31 (green) was present in a subset of mid to upper dermal vessels in noninflamed abdominal (A and C), breast (E), and face (G) skin. E-selectin expression (open arrows, C) could be seen in distinct areas within a single upper dermal vessel in noninflamed abdominal skin. Staining with isotype-matched control mAbs showed minimal background staining in noninflamed abdominal (B and D), breast (F), and face (H) skin. (Note: All confocal immunofluorescence images shown in this figure and Figs. 4⇓ and 5⇓ are composite overlays integrating a series of optical sections taken at 2-μm intervals which span the z-axis.)
E-selectin colocalizes with CCL17 in a subset of mid to upper dermal vessels in noninflamed human skin
To investigate whether E-selectin expression coincided with CCL17 expression in dermal vessels in noninflamed skin, two-color immunofluorescence staining for E-selectin and CCL17 was performed on 30-μm sections of noninflamed skin and analyzed by confocal microscopy, as above. Staining for CCL17 and E-selectin on abdominal skin revealed that a number of mid to upper dermal vessels contained both CCL17 and E-selectin (Fig. 4⇓A). Identical two-color immunofluorescence studies of face and breast skin confirmed colocalization of CCL17 and E-selectin in a similar distribution of vessels in these tissues as well (data not shown).
Two-color immunofluorescence staining for E-selectin and TARC/CCL17, and E-selectin and ICAM-1 in noninflamed human skin. A, E-selectin (red) and CCL17 (green) co-localize in a select group of mid to upper dermal vessels, as shown in noninflamed abdominal skin. B, Minimal staining was observed using isotype-matched control Abs for E-selectin and CCL17. C, E-selectin (red) and ICAM-1 (green) colocalize in a subgroup of mid to upper dermal vessels, as shown in noninflamed face skin. D, Minimal staining was observed using isotype-matched control Abs for E-selectin and ICAM-1.
E-selectin colocalizes with ICAM-1 in a subset of mid to upper dermal vessels in noninflamed human skin
To determine whether E-selectin is coexpressed with ICAM-1, two-color immunofluorescence staining was performed on 30-μm sections of noninflamed skin and analyzed using confocal microscopy, as above. Immunostaining for E-selectin and ICAM-1 on face skin demonstrated that ICAM-1 was more broadly expressed than E-selectin. Areas of overlapping expression were once again seen only on a restricted subset of vessels in the mid to upper dermis (Fig. 4⇑C). Similar patterns of E-selectin and ICAM-1 coexpression were observed using two-color immunofluorescence staining of breast and abdominal skin (data not shown).
E-selection, CCL17, and ICAM-1 colocalize in blood vessels of noninflamed skin
Three-color immunofluorescence staining was performed to analyze the relative expression of E-selectin, CCL17, and ICAM-1 in blood vessels of noninflamed skin. Confocal microscopy revealed a subset of vessels in the mid to upper dermis of abdominal skin with overlapping E-selectin, CCL17, and ICAM-1 expression (Fig. 5⇓, A–C, G, and H). Similar to the results described above, ICAM-1 showed a broader distribution of vessel expression than either CCL17 or E-selectin. This is reflected in the general appearance of segments and branches of vessels with ICAM-1 expression alone, and areas with all three molecules present. Similar results were obtained with face skin in which three-color immunofluorescence staining (Fig. 5⇓, D–F, I, and J) showed coexpression of E-selectin, CCL17, and ICAM-1 in a subset of dermal vessels.
Three-color immunofluorescence staining for E-selectin, TARC/CCL17, and ICAM-1 in noninflamed human skin. A subset of mid to upper dermal vessels revealed expression of E-selectin (blue, A and D), CCL17 (red, B and E), and ICAM-1 (green, C and F) in noninflamed abdominal (A–C) and face skin (D–F). Overlap of E-selectin (blue), CCL17 (red), and ICAM-1 (green) staining confirms coexpression of these three molecules in dermal vessels of both abdominal (G) and face (I) skin. E and J, Isotype-matched control mAbs produced minimal signals and confirmed staining specificity in noninflamed abdominal (H) and face (J) skin.
Discussion
In this study, we show that E-selectin, CCL17, and ICAM-1 are present in noninflamed skin from multiple individuals and demonstrate for the first time the colocalization of all three molecules in vessels of the mid to upper dermis. These results suggest that dermal vessels express the factors necessary to recruit T cells in the absence of clinical and histological evidence of inflammation. We used both avidin-biotin immunoperoxidase and immunofluorescence staining methods to examine the expression of E-selectin in noninflamed skin from three different sites, including both typically sun-exposed (face) and non-sun-exposed (breast and abdomen) areas. Similar levels of E-selectin expression were observed in all three skin sites, and E-selectin appeared in only a subset of vessels labeled by CD31. To analyze E-selectin expression along extended areas of endothelium, we compiled images obtained from 30-μm sections of normal human skin via confocal microscopy. Using this method, we observed that E-selectin expression is not uniformly distributed in the vessels of noninflamed skin, but appears concentrated in focal areas of vessel endothelium in the mid to upper dermis.
The presence of E-selectin in the dermal vessels of noninflamed skin from multiple sites implies that this finding is a common feature of skin and may play a significant role in cutaneous immune surveillance. Because of the constitutive expression of E-selectin in dermal blood vessels, skin-homing memory T cells will slow their passage through skin via their tethering and rolling action. This would both reduce the response time necessary to recruit T cells to sites of skin injury and foreign invasion, and enhance their ability to respond to low-level inflammatory signals. In this manner, E-selectin expression can sustain a baseline cutaneousdefense shield by effectively delaying individual skin-homing memory T cells in dermal vessels and maintaining a steady state population of memory T cells poised to respond to signs of distress in the skin. The patchy distribution of E-selectin in dermal vessels is striking and suggests recruitment occurs only in restricted areas.
The demonstrated coexpression of E-selectin with CCL17 and ICAM-1 in these vessels supports this hypothetical early defense system by providing the substrates necessary for integrin activation and firm adhesion of rolling T cells. T cells rolling at low velocity on E-selectin could respond to small levels of chemokines and firmly adhere to the vascular endothelium in preparation for extravasation into the skin. This model is consistent with the small number of lymphocytes observed in the perivascular areas of dermal blood vessels in noninflamed skin on H&E sections (Fig. 1⇑) and is further supported by reports that T cells recovered from noninflamed skin show increased levels of CLA and CCR4 compared with circulating T cells (47% in skin vs 10% in blood for CLA; 63% in skin vs 20% in blood for CCR4) (8), suggesting a bias toward recruitment of CLA+ CCR4+ T cells into noninflamed skin.
Although the coordinate expression of all three elements of the T cell recruitment cascade would allow these endothelial sites to serve as gateways for T cell entry into skin, the mechanisms restricting expression to a limited subset of vessels in the mid to upper dermis are unclear. Up-regulation of the expression of all three components, E-selectin, CCL17, and ICAM-1, in response to inflammation is linked to transcriptional up-regulation via the NF-κB pathway (26, 27). Although UV light is known to initiate inflammatory cell recruitment by activating NF-κB regulatory cascades, and can induce E-selectin, CCL17, and ICAM-1 expression in skin (26, 27, 28, 29), the presence of these three components in the multistep leukocyte recruitment cascade in multiple skin samples from typically non-sun-exposed areas, such as the breast and abdomen, suggests that UV exposure is not the only regulatory trigger for T cell recruitment to resting skin. Whether low-level NF-κB activation drives localized constitutive expression of E-selectin, or whether another control mechanism exists, remains undetermined. Although the studies reported in this work do not address the possibility of low-level NF-κB activation directly, the lack of tissue-infiltrating leukocytes on H&E-stained sections, and the similarity of expression in skin from multiple sites and in multiple donors argue that the results presented in this work represent a normal feature of human skin. In fact, the small perivascular collections of lymphocytes seen in these samples may be taken to represent the expected results of immunosurveillance in the skin.
Although we have demonstrated colocalization of E-selectin and CCL17, other chemokines could also play a role in cutaneous immunosurveillance. Liver- and activation-regulated chemokine)/macrophage-inflammatory protein-3α/CCL20 has been proposed as a candidate molecule, as its receptor CCR6 is also highly expressed on skin-homing CLA+ memory T cells (30). In fact, CCL20 has been demonstrated to induce the firm arrest of memory T cells in vitro (31) and to be constitutively present in dermal vessels (32). In preliminary studies, we have observed CCL20 staining in dermal blood vessels of noninflamed skin andcoexpression of CCL20 with E-selectin, similar to our findings with CCL17 (data not shown). Whether CCL20 and CCL17 occur in the same, or distinct, areas of dermal blood vessels has not yet been determined. Other candidate chemokines, including cutaneous T cell-attracting chemokine/CCL27, stromal-cell derived factor-1/CXCL12, and CCL22, were not detected in dermal vessels in noninflamed skin (data not shown).
In summary, we have demonstrated the colocalization of E-selectin, CCL17, and ICAM-1 in a subset of dermal vessels of noninflamed human skin from both sun-exposed and non-sun-exposed areas in multiple individuals. Based on these observations, and supporting evidence from other studies detailed above, we propose that skin-homing T cells are recruited to noninflamed skin in a manner similar to that in inflammatory sites via the coordinate actions of E-selectin, CCL17 (or other chemokines), and ICAM-1, and that these components are constitutively expressed and colocalized on dermal microvessels in noninflamed skin. Because E-selectin provides the critical initial step in recruiting circulating T cells, both the number of T cells marginated to skin, and the capacity of these T cells to respond quickly to early, or subclinical, inflammatory signals are increased in areas with constitutive E-selectin expression. This enhanced monitoring mechanism provides a foundation for a strong cutaneous defense mechanism and most likely reflects the unique role of the skin as a primary interface with the environment.
Acknowledgments
We thank Dr. Thomas Cochran and the staff of the Boston Center for Plastic Surgery, and Dr. Dennis Orgill and Dr. Carl Schanbacher of Brigham and Women’s Hospital for providing skin tissue from various skin-removal surgeries; Dr. Phillip Allen of Brigham and Women’s Hospital for assistance with the laser confocal microscope; Dr. Scott Granter of Brigham and Women’s Hospital for analyzing H&E-stained slides; and Denise Long-Woodward for advice with histological studies.
Footnotes
-
↵1 This work was supported by the National Institutes of Health/National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant AI-41707. B.F.C. is supported by a Howard Hughes Medical Institute Medical Student Research Training Fellowship.
-
↵2 Address correspondence and reprint requests to Dr. Robert C. Fuhlbrigge, Harvard Institutes of Medicine, 77 Avenue Louis Pasteur, Room 668, Boston, MA 02115. E-mail address: rfuhlbrigge{at}rics.bwh.harvard.edu
-
↵3 Abbreviations used in this paper: CLA, cutaneous lymphocyte-associated Ag; TARC, thymus- and activation-regulated chemokine.
- Received July 3, 2003.
- Accepted November 13, 2003.
- Copyright © 2004 by The American Association of Immunologists
References
In this issue
