Commentary Lymphocyte trafficking through blood and lymphatic vessels: more than just selectins, chemokines and integrins

Abbreviations:

Lymphoid organs are the professional sites in the body that allow lymphocytes to encounter antigen, become activated and proliferate. Naive T cells home very efficiently to lymph nodes, with about 15,000 lymphocytes extravasating each second in a typical lymph node of the sheep 1. The entry of lymphocytes from the blood into lymphoid tissue occurs at specialized sites of the vascular tree, the so-called high endothelial venules (HEV). These microvessels are post-capillary venules that are distinguishable from other blood vessels by their plump, cuboidal "high" endothelial cells. Such endothelial cells specifically present constitutively expressed adhesion mechanisms and chemoattractive stimulators that guide naive lymphocytes in a cascade of molecular interactions to enter lymphoid tissue 2, 3.

Antigen-experienced T cells, and other leukocytes such as neutrophils and monocytes that actively execute immune defense, exit from blood vessels at any sites of injury or infection. Efficientand well-controlled extravasation occurs preferentially in post-capillary venules (except for spleen, lung and liver) and principally similar molecular mechanisms as in lymph node HEV mediate the initiation of leukocyte–endothelial interactions, and leukocyte migration and diapedesis. In contrast to HEV, where these mechanisms are constitutively active, the endothelium at sites of injury up-regulates such mechanisms upon exposure to inflammatory stimulators. Depending on the type of tissue and the type of inflammatory stimulus, slightly variable subsets of cell adhesion molecules can beinvolved in this process 2, 3.

Despite such variations, the mechanism usually requires a well-known cascade of molecular interactions starting with selectin-dependent rolling of leukocytes on the endothelial cell surface that initiates the interaction with the blood vessel wall. The three selectins are carbohydrate-binding adhesion molecules, which are either constitutively expressed on most leukocytes (L-selectin) orinducible on endothelium (E- and P-selectin). Their ligands are sialomucins that present oligosaccharides related to sialyl-Lewis^X 4. The binding parameters of selectin–ligand interactions allow binding with high on-rates and high tensile strength, rendering the selectins ideal tools for capturing leukocytes from the flowing blood to the blood vessel wall. Selectin-mediated cell contacts enable leukocytes to sense chemokines presented on the endothelial cell surface, leading to activation of leukocyte integrins and active migration of leukocytes on the endothelial cell surface. The most important integrins in this process are the β2-integrin LFA-1 (αLβ2 or CD11a–CD18) and the two α4 integrins α4β1 (VLA-4) and α4β7. Their ligands on endothelium are members of the immunoglobulin supergene family such as ICAM-1, ICAM-2, VCAM-1 and mucosal addressin cell adhesion molecule-1 (MAdCAM-1) 2, 3. Finally, the leukocytes move through the endothelial cell layer at sites of endothelial junctions by a mechanism that is not yet understood in detail 5.

Whereas the entry of lymphocytes from blood into lymph nodes has been well analyzed, exit of lymphocytes from lymph nodes is much less well studied. It requires migration of lymphocytes through the endothelium of lymphatic sinusoids in the lymph nodes to reach the efferent lymphatic vessels. An alternative route for lymphocytes to enter lymph nodes is via the afferent lymph although far fewer lymphocytes enter lymph nodes via this route than via HEV. In addition to lymphocytes, DC and macrophages enter lymph nodes through afferent lymphatics whereas only lymphocytes are able to leave lymph nodes again via the efferent lymph. These trafficking pathways illustrate the need for mechanisms that allow leukocytes to traverse lymphatic endothelium within lymph nodes. Outside lymphoid tissues, leukocytes such as DC traverse lymphatic endothelium in order to enter afferent lymphatics on their journey to the lymph nodes.

Lymphatic vessels have only very little or no basal lamina, simplifying the diapedesis process. Junctional structures and typical endothelial cell contact proteins between lymphatic endothelial cells seem to be less abundant than between microvascular blood endothelial cells. However, VE-cadherin, catenins and CD31 as well as adherens, gap and even tight junctions were recently documented between primary isolated lymphatic endothelial cells and between cultured lymphatic endothelial cells of human skin 6. Thus, it is possible that lymphatic endothelial cells in lymphatic vessels represent a more stringent barrier function than previously anticipated. For two reasons it would seem likely that the molecular mechanisms of lymphocyte migration through the barrier of lymphatic endothelium differs from that through blood endothelium. First, lymphocytes that exit from lymph nodes need to traverse through the barrier of the lymphatic sinusoids from the basal/abluminal side to the apical/luminal side of the endothelial cells, in contrast to the situation of extravasating lymphocytes in blood vessels. Second, although DC and those few lymphocytes that enter lymph nodes via the afferent lymph vessels traverse the lymphatic endothelium of the subcapsular sinus from the luminal to the abluminal site, they do this in the absence of flow and shear.

Despite these differences between leukocyte migration through blood and lymphatic endothelia, Irjala et al. 7 surprisingly discovered a novel glycoprotein that is able to support lymphocyte adhesion to lymphatics as well as to HEV. It was identified by a classical approach using mAb obtained after immunizing mice with efferent lymphatics isolated from the stroma of human lymph nodes. Two mAb against this antigen stained afferent and efferent lymphatic endothelium as well as HEV. For this reason, the antigen was named common lymphatic endothelial and vascular endothelial receptor-1 (CLEVER-1). Interestingly, the antigen was present in afferent lymphatic vessels in all non-lymphoid tissues analyzed (except cerebellum) whereas it was absent from blood vessel endothelium of normal, non-inflamed tissues. It was also absent from blood leukocytes. In inflamed tissues, e.g. inflamed synovium and inflamed skin, CLEVER-1 was found on HEV-like blood vessels and its expression correlated with the extent of infiltrating lymphocytes. This expression pattern is highly suggestive for a role of CLEVER-1 in lymphocyte trafficking.

To test this possibility Irjala et al. 7 analyzed whether the mAb against CLEVER-1 would interfere with the binding of lymphocytes to lymph node HEV and lymphatic vessels. They used the classical "frozen-section assay" for this analysis 8 which is based on the incubation of lymphocytes with frozen sections of lymph nodes under conditions of mild rotation at 7°C. Under these conditions lymphocytes bind specifically only to areas of HEV on the tissue sections. This rather unusual "cell adhesion" assay has proved to be of fundamental importance for the identification of the classical adhesion molecules that mediate lymphocyte homing such as L-selectin 9 and the two vascular addressins peripheral lymph node addressin (PNAd)10 and MAdCAM-1 11, which were found on HEV. PNAd turned out to be a carbohydrate motif that acts as ligand for L-selectin on various glycoprotein carriers on HEV of peripheral and mesenteric lymph nodes, and this epitope is the classical marker for HEV.

Using this assay, Irjala et al. 7 found that the two mAb against CLEVER-1 inhibit the binding of lymphocytes to human lymph node HEV by 50%. Binding to HEV-like blood vessels of inflamed synovium was also inhibited by the anti-CLEVER-1 mAb and this was shown not only for lymphocytes but also for granulocytes and monocytes. Under conditions of shear (mild rotation) no adhesion of lymphocytes can be observed to lymphatics. Performing the assay under static conditions (still at 7°C) allowed detection of specific lymphocyte binding only to HEV and to lymphatic sinusoids. The two anti-CLEVER-1 mAb could inhibit this binding of lymphocytes to lymphatic endothelium, demonstrating that one and the same glycoprotein can mediate lymphocyte binding to blood and to lymphatic endothelium.

A function for CLEVER-1 in lymphocyte homing is also suggested by in vivo experiments in rabbits. The antibody 3-372 crossreacts with the rabbit homolog of CLEVER-1 and the epitope was found to be accessible on the luminal surface of HEV following i.v. injection of the antibody and staining with a secondary antibody on sections of lymph nodes that had been isolated 5 min after the injection. To test the effect of the antibody on lymphocyte recruitment to rabbit lymph nodes in vivo, animals were immunized in footpads with keyhole limpet hemocyanin and mAb 3-372 wasadministered i.v. at the same time and at days 1 and 2. At day 6 the draining lymph nodes were weighed and histologically analyzed. The weight of the lymph node was indeed reduced by the anti-CLEVER-1 mAb when compared with an irrelevant control antibody. Similar to the result of the frozen-section assay, the inhibitory effect was only partial and the immunization-dependent size-increase of the lymph nodes was still twofold when compared with non-immunized lymph nodes. Nevertheless, the combined results of tissue distribution, frozen-section assays and immunization experiments represent the first evidence for a novel adhesive mechanism that assists lymphocyte binding to HEV as well as lymphocyte interactions with the lymphatic sinusoids in lymph nodes.

In addition to the identification and functional analysis of CLEVER-1, the antigen was cloned. It is a large glycoprotein (270–300 kDa in SDS-PAGE) that separates as three or four isoforms by gel electrophoresis. The full-length cDNA comprises 7879 bp with a coding sequence for 2570 amino acids. The same cDNA was reported in the last year to code for a protein called stabilin-1 expressedby stimulated monocytes and macrophages 12 and a protein called FEEL-1 13, 14 that is expressed on the leukemia cell line KG-1. FEEL-1 was reported to act as an endocytotic receptor ("scavenger receptor") that binds advanced-glycation end-products as well as bacteria and that may have angiogenesis-modulating activity. CLEVER-1 is a typical type I transmembrane protein with only a single transmembrane domain close to its C terminus. The large protein is composed of several fasciclin domains and EGF-like repeats, and it contains two RGD motifs plus a domain that resembles the proteoglycan link protein domain, a type of domain that is also present in CD44. Such a domain composition is in line with the adhesive properties of CLEVER-1.

The fact that CLEVER-1 seems to mediate interactions between lymphocytes and lymphatic endothelium suggests a function as an adhesion molecule in the absence of shear, arguing that CLEVER-1 onHEV may rather support adhesion at a step after the L-selectin-mediated capturing. It is interesting that CLEVER-1 was not expressed on the cell surface of transfected cells, although cell surface expression on HEV was found in situ in rabbits. Whether it is also found in intracellular compartments of endothelial cells is not reported.

Another protein that was originally identified by the same group on HEV-like vessels of inflamed synovium is the non-classical adhesion molecule vascular adhesion protein-1 (VAP-1) 15–17. This protein controls lymphocyte migration directly and acts as a cell surface amine oxidase 18. VAP-1 is inducible on blood vessels in inflamed tissue and was found to be expressed in vesicles inside endothelial cells, distinct from Weibel-Palade bodies. It would be interesting to analyze whether CLEVER-1 and VAP-1 are found in similar intracellular locations and what would regulate their transport to the cell surface. Another non-classical adhesion molecule, that was found by the same group (actually by the same experimental approach as was used for CLEVER-1), is the mannose receptor that was described to function as an L-selectin ligand for lymphocytes on lymphatic endothelium 19. In contrast to CLEVER-1, it was not expressed on HEV. Interestingly, CLEVER-1 and the mannose receptor share functional features of scavenger receptors.

Efferent lymphatics provide the routes for lymphocyte exit from lymph nodes. Afferent lymphatics take up DC and some lymphocytes from peripheral tissue and allow them to migrate into lymph nodes. The spread of tumor cells may also occur via afferent lymphatics, guiding metastasis to lymph nodes. It is still debated to what extent lymphatic endothelium forms a closed lining of cells or whether gaps are present. A continuous basal lamina seems not to exist for these vessels. Electron microscopic studies have provided evidence for migration of DC through the intercellular space of directly adjacent lymphatic endothelial cells as well as for the migration through larger gaps 20. In vitro cultures of primary isolated lymphatic endothelial cells form junctions and dense cell monolayers with clearly detectable, uninterrupted VE-cadherin expression at all cell contacts, arguing for a better barrier function of this type of endothelium than usually anticipated 6. It will be interesting to study in the future whether the mannose receptor and CLEVER-1, which are both expressed on afferent lymphatics, are involved in the entry of lymphocytes and DC into afferent lymphatics.

Likewise, it is not known how DC and those few lymphocytes that enter lymph nodes via the afferent lymphatic vessels actually move into the lymphoid tissue. Upon arriving in the subcapsular sinus, DC adhere to the sinus floor, traverse through the sinusoidal endothelium (from the apical/luminal to the basal/abluminal side) and migrate further into the lymphoid tissue, probably following and adhering to the reticular network of fibroblastic reticular cells that surround collagen fibers 21. Recently the architecture of the cortex of lymph nodes was analyzed in a thought-provoking paper 22. The cortex is the site where naive T cells coming from the HEV encounter antigen-presenting cells that enter the cortex from the afferent lymph. Kaldjianet al. 22 suggest that the fibroblastic reticular cells, which wrap around extracellular matrix (ECM)-surrounded collagen fibers in the cortex, also form the inner layer of the cortex, which they described as a "labyrinthine cavity". Where this cavity abuts the sinus there is a triple boundary made of a thin layer of collagenous ECM sandwiched between lymphatic endothelial sinus-lining cells and the reticular cells that cover the inner surface of the cavity. From this picture it would follow that antigen-presenting cells entering the cortex from the afferent lymph would needto traverse through this triple boundary. A possible involvement of CLEVER-1 in this process will be interesting to study.