| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Published online before print September 4, 2008
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Department of Pathology;* the Department of Surgery, Division of Cardiac-Thoracic Surgery; the Core Unit for Medical Statistics and Informatics; the Department of Internal Medicine, Division of Rheumatology; the Department of Dermatology,¶ Research Unit of Biology and Pathobiology of Human Skin; the Department of Surgery,|| Division of Plastic and Reconstructive Surgery; and the Department of Dermatology, Division of General Dermatology, Vienna Medical University, Vienna, Austria; and the Department of Reconstructive Surgery,** University Hospital Zurich, Zurich, Switzerland
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Blood vascular endothelial cells (BECs) and LECs were isolated ex vivo from dermatome skin sheets from cosmetic breast surgery specimens of healthy females, in a two step protocol using mechanical and enzymatic dissociation, followed by fluorescence-activated cell sorting (FACStar Plus; BD Biosciences, Franklin Lakes, NJ) using CD45 antibodies for gating and CD31 and podoplanin antibodies for separation as reported previously.2,3 LECpodo-low and LECpodo-high were purified by fluorescence-activated cell sorting separation of the total LEC fraction by increasing the resolution of the podoplanin channel. The collected cells were lysed for RNA isolation or immobilized on slides by cytospin (Cytospin 3; Dako Cytomation, Glostrup, Denmark) or taken into tissue culture. The study complied with the Declaration of Helsinki and was as approved by the local ethics committee (permit no. 449/2001).
Antibodies
We used the IgG fraction of a rabbit polyclonal anti-podoplanin and of the corresponding preimmune serum.3,4 The other reagents were FITC-conjugated monoclonal anti-CD31 (catalog no. 555445; BD Biosciences Pharmingen, Franklin Lakes, NJ), RPE-Cy5.1-conjugated mouse monoclonal anti-CD45 (catalog no. PM IM2653; Beckman Coulter, Inc., Fullerton, CA), mouse clone 2C3 anti-Duffy blood group antigen receptor for chemokines (DARC)/Fy (kindly provided by Dr. Ives Colin; INSERM, Institute National de Transfusion Sanguine, Paris5,6 ), mouse monoclonal anti-CCL27 (catalog no. MAB3761; R&D Systems, Minneapolis, MN), neutralizing rat anti-CCL27 antibody (clone 68623; catalog no. MAB725; R&D Systems), rat IgG-Isotype control (catalog no. ab37361; Abcam, Cambridge, UK), mouse monoclonal anti-CCL21 (catalog no. AF366; R&D Systems), goat polyclonal anti-CCR10 (catalog no. NB 100-707; Novus Biologicals, Littleton, CO), rat monoclonal anti-human CLA antibody (catalog no. 55946; BD Biosciences Pharmingen), mouse monoclonal anti-CD31 (catalog no. M0823; Dako Cytomation), and phycoerythin-conjugated mouse monoclonal anti-CD3 (catalog no. 345765; Becton Dickinson) antibodies. Secondary agents were phycoerythin-conjugated donkey polyclonal anti-rabbit IgG (WG, catalog no. 711-116-152; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA), Alexa Fluor 594-conjugated goat anti-mouse and Alexa Fluor 488- and 350-conjugated goat anti-rabbit (1 µg/ml; nos. catalog nos. A-11020, A-11034 and A-11046; Molecular Probes, Inc., Eugene, OR), and 10 nm gold-labeled goat anti-rabbit (catalog no. RPN421; Amersham Biosciences, Uppsala, Sweden) antibodies.
RNA Analysis
Isolation of total RNA, generation of cRNA, and DNA chip hybridzation were performed as described previously.4
DNA chip hybridzation was performed on GeneChip U133A (Affymetrix, Inc., Santa Clara, CA). Raw and processed data were deposited at ArrayExpress (accession no. E-TABM-227). DNA chip raw data were normalized to a subset of 10 housekeeping genes using MAS 5.0 software and exported as spreadsheets. Considering the sample size of n = 2, only postnormalized probe set values tagged as "present" were accepted. Ratios of values of LECpodo-low and LECpodo-high sample pairs deriving from each patient were calculated. Both ratios were averaged, and a threshold of differential induction of 
fivefold was applied. Finally, bibliographic search on PubMed database for co-occurrence of the gene name and the keywords "skin" and "chemokine" defined a selected list of eight genes induced in LECpodo-low versus LECpodo-high. The two housekeeping genes in nonquantitative RT-PCR were breakpoint cluster region and β 2-microglobulin, as described previously.3
In addition, early passage cultures of LECpodo-low and LECpodo-high were subjected to quantitative real-time PCR (Applied Biosystems, Foster City, CA) for relative podoplanin mRNA quantification using 
Ct.
Morphological Analysis
For electron microscopy, human skin was fixed in 4% PFA and 0.1% glutaraldehyde and embedded in Lowicryl-K4M, and sections were immunolabeled with 20 µg/ml anti-podoplanin antibody, followed by 10 nm goat anti-rabbit gold conjugate (final concentration, 1:10). The number of gold particles bound to 
500 µm of LEC plasma membranes were counted and expressed as number of gold particles per 1 µm. Lymphatic capillaries were identified by their expression of podoplanin, the absence of continuous basement membranes, and lack of erythrocytes in their lumen.
For reconstruction of lymphatic vessels in three dimensions, double immunofluorescence was performed on 100 µm cryostat sections of tissue samples that were fixed in 4% PFA for 1.5 hours at 22°C, immersed into 30% sucrose at 4°C overnight, and frozen in liquid nitrogen.7 The sections were immunolabeled with antibodies against podoplanin (2.8 µg/ml) and CCL27 (2.5 µg/ml). After extensive washing, they were postfixed in 4% paraformaldehyde for 10 minutes at 22°C and permeabilized with 0.3% Triton X-100 for 60 minutes. Primary antibodies were incubated for 3 days at 22°C, and secondary antibodies were incubated for 6 hours. Sections were analyzed by confocal laser scanning microscopy using a Zeiss LSM 510 (Carl Zeiss, Jena, Germany). Three-dimensional visualization was realized in a program in JAVA, based on the Visualization Toolkit 5.0 (Kitware, Clifton Park, NY; http://www.vtk.org). Negative controls omitting the first antibodies consistently gave negative results, as did incubation with nonimmune sera (data not shown).
Immunofluorescence labeling was performed on 2-µm paraffin sections of formaldehyde-fixed human tissues. After antigen retrieval by microwaving twice for 5 minutes in 0.01 mol/L citrate buffer, pH 6, primary antibodies were combined with antibodies conjugated to fluorescent dyes for double or triple staining.8 Images were captured with the Axiophot epifluorescence system (Zeiss). We combined immunolabeling by antibodies toward podoplanin with those specific for DARC (final concentration 20 µg/ml), CCL21 (15 µg/ml), CCR10 (4 µg/ml), and CCL27. For quantification of the distribution of CCR10+ in the perivascular stroma, the perilymphatic area was divided into concentric rings of incremental 10-µm radiuses (R1–3) measured from the vessel’s outer surface, and CCR10+ cells were counted in each segment and normalized to 1000 µm2. Also, the number of CCR10+ cells was determined in the vascular lumen.
Cell Culture
Culturing of human LEC subpopulations was performed on 10 µg/ml fibronectin-coated dishes (catalog no. 354008; BD Biosciences) in EGM-2MV complete growth medium (catalog no. CC-3202; Cambrex, Inc., East Rutherford, NJ). Cells were detached by 1x trypsin-EDTA (catalog no. 25300-05; Invitrogen, Carlsbad, CA), lysed in SDS sample buffer, and analyzed by reduced gel electrophoresis and Western blotting. For immunofluorescence, endothelial cells were fixed for 15 minutes at 22°C in 4% paraformaldehyde. Nuclei were stained with 10 µg/ml 4',6'-diamidino-2-phenylindole hydrochloride (SERVA GmbH, Heidelberg, Germany). The concentration of CCL27 in tissue culture supernatants was determined by ELISA (R&D Systems).
Transmigation Assays
BECs, LECpodo-low, and LECpodo-high were seeded onto the bottom of transwell dishes with polycarbonate membrane inserts with 5-µm pores (catalog no. 3421; Costar, Cambridge, MA).9 CD3+ CCR10+ human peripheral blood T lymphocytes were purified from pooled human blood samples by fluorescence-activated cell sorting, labeled with CellTracker (2 µg/ml CellTracker green CMFDA; catalog no. C2925; Invitrogen), and resuspended in 100 µl RPMI with 10% fetal calf serum; and 1 x 105 cells were added into the top chamber of each insert. After 4 hours, the trans-migrated cells in the lower chamber were collected by cytospinning onto microscope slides and counted in a fluorescence microscope. Fixed cells were additionally immunostained with 10 µg/ml anti-CLA antibody. In some experiments, 10 µg/ml neutralizing monoclonal rat anti-CCL27 antibody was added into the bottom chamber. As control, isotype matched irrelevant rat IgG was used in the same concentration in parallel experiments.
Statistical Analysis
All quantitative experiments were performed at least in triplicate (RT-PCR, numbers of immunogold particles per micrometer of LEC cell membrane representing podoplanin, capillary densities in histological sections, localization of CCR10+ T lymphocytes in relation to lymphatic vessels in situ, and CCR10+ T-lymphocyte transmigration assays), and the results were subjected to the two-sided Student’s t-test.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
The fluorescence-activated cell sorting pattern of isolated dermal microvascular CD45–CD31+Podoplanin+ LECs and CD45–CD31+Podoplanin– BECs has previously suggested the presence of two overlapping LEC subpopulations that varied in their surface expression of podoplanin.2
Here, we have isolated these two LEC cohorts, and designated them as LECpodo-low and LECpodo-high (Figure 1A)
. The distinctly different levels of podoplanin expression of LECpodo-low and LECpodo-high were confirmed by indirect immunofluorescence in ex vivo cytospin preparations (Figure 1B)
, and their lymphatic phenotype was supported by nonquantitative RT-PCR for the established LEC markers LYVE-1, Prox1, and the pan-endothelial marker von Willebrand factor (Figure 1, C and F)
. In cell culture, small amounts of podoplanin protein were detected by Western blotting in LECpodo-low that, however, remained abundant in LECpodo-high (Figure 1D)
. This was also confirmed at the level of mRNA by quantitative RT-PCR (Figure 1E)
. In later passages, the expression of podoplanin in LECpodo-low increased toward the levels of LECpodo-high, whereas the expression of the LECpodo-low typic chemokine CCL27 (Figure 1G)
was lost after the second passage (data not shown).
|
To identify additional markers for the precise localization of LECpodo-low and to gain insight into the LECpodo-low’s distinct phenotype in normal and particularly in inflamed tissues, a DNA chip analysis was performed in a shotgun approach. We purified RNA from freshly isolated LECpodo-low, LECpodo-high, total LECs, and total BECs2,3
that were isolated from the dermis of two patients without skin disease. In Affymetrix arrays, we identified 107 genes that were amplified by at least fivefold in LECpodo-low. Of these, eight were associated with inflammation (Table 1)
and included members of the chemokine/chemokine receptor family, such as Ccl27, Cxcl12, Cxcl14, and Darc. In addition, LECpodo-low overexpressed E-selectin, components of the major histocompatibility complex type II (MHC II), and placental growth factor. By contrast, LECpodo-high significantly overexpressed Ccl21. The chemokine CCL27 was expressed in cultured LECpodo-low, but not in LECpodo-high (Figure 1G)
, and released into cell culture media in a concentration <1 ng/ml.
|
. Within single lymphatic vessels, there was consistently no intermixing of the two LEC subpopulations. Immunoelectron microscopy also confirmed the at least fourfold different podoplanin content on the plasma membrane surface of LECpodo-low and LECpodo-high (Figure 2B)
.
|
. The latter exclusively expressed CCL2112
and large amounts of podoplanin, in contrast to LECpodo-low vessels that were immunolabeled for CCL27 (Figure 2D)
. This indicated that LECpodo-low were restricted to lymphatic precollectors that directly originate from canonical initial LECpodo-high lymphatic capillaries, and that the LEC subpopulations expressed the same molecular signatures discovered also in situ as seen before in isolated cell. A similar compartmentalization of the lymphatic microvasculature with distinct segregation of LECpodo-high- and LECpodo-low-containing segments was found in 6 samples of normal skin, 15 cases of chronic dermal inflammatory diseases, and in sporadic cases of renal transplant rejection (Figure 3A)
and pulmonary disease (Figure 3B)
.
|
Lymphatic microvessel densities were determined in all samples of normal and inflammatory skin, using podoplanin and CCL27 and/or DARC as immunohistochemical markers. We observed that the overall lymphatic vessel density increased threefold in chronic skin inflammation over normal controls and that the relative proportion between LECpodo-high lymphatic capillaries and LECpodo-low precollectors remained constant. Blood capillaries that were identified by expression of DARC+ endothelial cells and erythrocytes were 50 times more abundant than LECpodo-low and seven times more abundant than LECpodo-high microvessels in normal skin and in all disease entities studied (Figure 4A)
.
|
It is established that in chronic inflammatory skin diseases, the CCL27 receptor CCR10 on CLA+ pathogenic T lymphocytes mediates homing to the dermis.13
In normal human skin, few sporadic CCR10+ T lymphocytes were found in the dermis, without significant relation to microvessels (data not shown). By contrast, CCR10+ inflammatory infiltrates in human skin samples of psoriasis (n = 7), eczema (n = 7), and chronic atrophic acrodermatitis (n = 1) that constitute almost all lymphocytes in these tissues were centered around CCL27-expressing LECpodo-low, as visualized by triple immunofluorescence (Figure 4B)
. Quantification of the density of the CCR10+ T lymphocytes supported this view, with almost threefold higher densities around LECpodo-low-containing precollectors than LECpodo-high initial capillaries. CCR10+ T lymphocytes were found within the precollector’s lumen but not in that of initial capillaries (Figure 4C)
.
Chemotactic Attraction of CCR10+ T Cells by LECpodo-low in Vitro
LECpodo-low, but not LECpodo-high expressed (Figure 1G)
and secreted small amounts of CCL27 (<1 ng/ml) into the medium. This raised the possibility to test whether or not CCL27+ LECpodo-low could attract CCR10+ T lymphocytes also in vitro, as suggested by the immunohistological in vivo results described above. Isolated LECpodo-low, LECpodo-high, or BECs not later than in passage 2 were grown to confluency in the lower compartments of transmigration chambers, and CCR10+ CD4+ CLA+ T lymphocytes that were purified from human peripheral blood were added to the upper chambers. Staining with anti-CLA antibody showed that these T lymphocytes were statistically significantly (P < 0.05) more attracted by LECpodo-low than by LECpodo-high or BECs. Addition of anti-CCL27 antibody with established specific inhibitory activity13
into the lower Transwell chamber reduced the chemoattractive capacity of LECpodo-low to baseline levels (Figure 4D)
.
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
The major novel finding in this report is that initial lymphatic capillaries and precollectors are endowed with two distinctly different subpopulations of endothelial cells. LEC heterogeneity was initially discovered by virtue of the differential expression of the LEC-specific membrane sialomucin podoplanin, and was extended further to several other significant differences in gene expression, including characteristic gene expression of different sets of chemokines and their receptors, suggesting potentially different functions in normal tissues and in inflammatory processes. For example, LECpodo-low in precollectors highly up-regulated CCL27, whereas CCL21 was characteristically expressed by LECpodo-high in initial capillaries. Previous work has raised the possibility that podoplanin is involved in the generation of chemotactic gradients around initial capillaries, because CCL21 that attracts CCR7+ dendritic cells and T lymphocytes14 avidly bound to podoplanin and complexes were shed from the surfaces of LEC.15 Recently, experimental evidence was provided that the interaction of CCL21/CCR7 is involved in clearing sporadic CCR7+ cells from normal skin.16
The discovery that LECpodo-low in lymphatic precollectors produce CCL27 raised the possibility that this vascular segment could be involved in trafficking of pathogenic CCR10+ T cells.13 These lymphocytes are the principal pathogenic effector cells in inflammatory skin diseases and in graft-versus-host reactions,17 constituting >95% of the inflammatory infiltrate. Their import route was recently mapped out in detail: CCL27 is produced by inflammatory chemokine-activated epidermal keratinocytes and diffuses toward blood microvessels.13 There, it is presented at the vascular lumen to attract circulating CCR10+ T lymphocytes and to initiate their emigration into the stroma. The indispensable role of CCL27 in this process was demonstrated by abrogation of experimental murine dermatitis by neutralizing CCL27 antibodies.13,18 These results on the CCL27-driven mechanisms of lymphocyte entrance into the site of inflammation then raise the question of what the role of CCL27 secreting LECpodo-low in lymphatic precollectors could be. We provide arguments that this segment of the lymphatic microvasculature could serve as exit route for CCR10+CLA+ T cells toward the regional lymph node. First, this view is based on a quantitative immunohistological analysis of 15 cases of human inflammatory skin disease that shows a significant selective recruitment of CCR10+CLA+ T lymphocytes around precollectors and within their lumen, in contrast to initial capillaries. Second, in vitro transmigration assays using cultured LECpodo-low indicate that CCL27 secreted by these cells into the medium drives the transmigration of isolated human CCR10+CLA+ T cells. The efficiency of precollectors as potential exit gate for lymphocytes could be further enhanced by the selective presentation of endothelial adhesion proteins on LECpodo-low, such as E-selectin, that specifically interacts with CLA on the surface of pathogenic CCR10+ T lymphocytes.19 Intriguingly, also cutaneous T-cell lymphoma expresses CCR10, and it remains to be determined whether any interaction with LECpodo-low occurs and is related to their dissemination.20
Expression of CCR10 was observed not only in T lymphocytes but also in LECpodo-low in inflamed tissues (Figure 4)
and was detected in cultured LECs after exposure to TNF
in vitro,21
suggesting an autocrine regulatory system that requires further elucidation. Also, the functional significance of other gene products expressed preferentially by LECpodo-low, such as CXCL12 and CXCL14, remains to be investigated.
Synoptically, our findings establish that distinct subpopulations of LECs are defined by their expression of podoplanin and distinct patterns of chemokines. We further show that these LEC subtypes constitute different segments of the initial lymphatic vasculature. These results dovetail into the recent findings that the CD31-containing interendothelial cell junctions of initial lymphatic capillaries differ from those in precollectors.22 Finally, our results raise the possibility that the LECpodo-low- containing precollector segment of the lymphatic vasculature is involved in the trafficking of pathogenic CCR10+ T lymphocytes in inflamed tissues.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by the Austrian Science Foundation (16472-B08), the Vienna Science and Technology Fund (LS139), the EU-FP6 framework program Lymphangio-genomics (LSHG-CT-2004-503573), and the Austrian Ministry for Education (Trafo Program).
Accepted for publication June 24, 2008.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
