The β-Chemokine Receptor D6 Is Expressed by Lymphatic Endothelium and a Subset of Vascular Tumors

Antibodies

The production of mouse monoclonal anti-D6 antibody is detailed below. Anti-podoplanin antibody was prepared as described recently. 4 Control murine IgGs and rabbit Ig were purchased from Sigma (Poole, UK) and DAKO (Glostrup, Denmark), respectively. Fluorescein isothiocyanate-coupled goat anti-mouse IgG was from Jackson ImmunoResearch Laboratories (West Grove, PA).

In Situ Binding Assay

MIP-1αP was labeled with ¹²⁵I using Iodogen (Pierce, Rockford, IL) as previously described. 33 In situ ligand-binding analysis with this labeled protein on skin pieces was performed as detailed before. 22 In brief, intact viable skin was removed in the process of elective reductive surgery on informed consent. Immediately on removal the skin was cubed into 1 to 2 mm 3 pieces by a sharp scalpel blade. Five to six cubes were incubated with ∼10 pg and 100 pg ¹²⁵I-MIP-1αP (specific radioactivity, ∼10 μCi/μg) in 0.2 ml of Hanks’ balanced salt solution supplemented with 0.1% bovine serum albumin and 20 mmol/L HEPES for 1 hour at room temperature on a platform shaker (200 rpm). The pieces were washed three times for 10 minutes in 2 ml of the buffer on the platform shaker and then fixed overnight in 4% buffered paraformaldehyde. Afterward the pieces were washed, dehydrated in increasing concentrations of ethanol, embedded in paraplast (Monoject Scientific, Athy, Ireland), and 5-μm-thin sections cut. These were deparaffinized, rehydrated, air-dried, and coated with K2 nuclear track emulsion (Ilford, Mobberley, UK). Slides were exposed at 4°C in the dark for 4 to 12 weeks and then developed with D-19 developer (Kodak-Pathe, Paris, France), fixed (Unifix, Kodak-Pathe), and counterstained with hemalaun. Slides were studied and photographs taken under a BX 60 Olympus microscope equipped with a DP10 digital camera (Olympus Austria, Vienna, Austria).

Generation of Anti-Human D6 Monoclonal Antibodies

Monoclonal antibodies to human D6 were raised using techniques described elsewhere. 34 Briefly, human D6 was expressed in the murine cell line L1.2 and this transfectant was used to immunize C57BL/6 mice. Fusion of spleen and myeloma SP2/0 was performed following standard procedures and the hybridoma supernatants were screened by flow cytometry for positively staining the D6 transfectant. Specificity of the antibodies was tested by staining wild-type L1.2 or Chinese hamster ovary (CHO) cells and L1.2 or CHO cells transfected with other human chemokine receptors. CHO cell transfectants were tested using immunohistochemistry as outlined below. To assess L1.2 transfectants, ∼10 6 cells were incubated with 50 μl of hybridoma supernatant for 30 minutes at 4°C, washed, then similarly incubated with goat anti-mouse IgG (fluorescein isothiocyanate-coupled), and subsequently assessed by fluorescence-activated cell sorting (FACS). Supernatants from several clones that specifically stained the D6 transfectant were selected, further subcloned by limiting dilution, expanded, and tissue culture supernatant collected. Specific D6 immunoreactivity was again confirmed and each antibody titrated to give a maximal differential between D6 transfectants and untransfected controls. One antibody (4A5) was purified using a Protein-G column. Purified antibody and antibody supernatants were stored at −20°C until use.

Immunohistochemistry

Normal human skin was removed during elective plastic surgery on informed consent. The skin samples were snap-frozen in liquid nitrogen-chilled isopentane. 5-μm-thin serial frozen sections were cut onto glass slides coated with 3-(triethoxysilyl)-propylamin (TESPA; Sigma Chemical Co., St. Louis, MO). The sections were fixed for 10 minutes in acetone at room temperature and thereafter were kept at −20°C for up to 8 weeks until the staining procedure.

The frozen serial skin sections were rehydrated and incubated with anti-D6 mouse monoclonal or anti-podoplanin rabbit polyclonal antibodies. The bound primary antibodies were detected by sequential incubations with alkaline phosphatase-conjugated goat anti-rabbit and rabbit anti-mouse antibodies (for podoplanin and D6, respectively) and an alkaline phosphatase-anti-alkaline phosphatase staining kit (DAKO), according to the manufacturer’s instructions. To control the specificity of antibody binding, mouse and rabbit IgG and Ig control antibodies (DAKO) were used at equimolar concentrations. Each immunostaining protocol was performed on skin from at least three different donors.

All of the other immunohistochemistry was done using 5-μm-thin paraffin sections cut from blocks obtained from the archives of the Department of Pathology, University of Glasgow. The sections were deparaffinized in Histoclear (Fisher Scientific, Loughborough, UK), rehydrated through decreasing concentrations of ethanol and then pure water, and incubated for 10 minutes in 3% hydrogen peroxide. To unmask the antigens, slides were either microwaved for 10 minutes in 1 mmol/L ethylenediaminetetraacetic acid (pH 8) (for D6) or for 15 minutes in 10 mmol/L citrate buffer (pH 6) (for podoplanin). After blocking with serum, avidin, and biotin, the appropriate primary antibodies were added and the sections left for 45 to 60 minutes at room temperature. The bound antibodies were detected using the Vector avidin-biotin-peroxidase detection system and Nova-Red substrate (Vector Laboratories, Peterborough, UK), according to the manufacturers instructions. Sections were fixed with neutral buffered formalin, counterstained with hematoxylin (Surgipath Europe Ltd., Peterborough, UK), dehydrated through increasing concentrations of ethanol, cleared in Histoclear, and mounted in Histomount (Fisher Scientific). The staining was evaluated by at least two independent examiners and photographed with an Olympus BX60 microscope.

In Situ Hybridization

Adjacent 7-μm frozen sections of formaldehyde-fixed human skin were either stained with anti-podoplanin antibodies as above, or used for in situ hybridization as follows. Sections were rehydrated, immersed in 100 mmol/L glycine-phosphate-buffered saline, permeabilized in 0.3% Triton X-100, and treated with Proteinase K (20 μg/ml) for 15 minutes at 37°C in 10 mmol/L Tris-HCl, 1 mmol/L EDTA, pH 7.4 buffer. After fixing in 4% paraformaldehyde at 4°C, sections were acetylated in TE buffer containing 0.25% acetic anhydride. Prehybridization and hybridization were done in buffers supplied by Novagen (Madison, WI). Sense and antisense D6 probes were generated from linearized human D6 cDNA plasmid using T7 or SP6 polymerases in combination with digoxigenin RNA-labeling mix (Roche Molecular Biochemicals, Mannheim, Germany). Transcripts were subjected to alkaline hydrolysis and 100- to 200-bp fragments purified. Probes were added to the hybridization buffer to 100 ng/ml, denatured, hybridized for 16 hours at 42°C, and the sections then washed in several changes of 1× standard saline citrate at 37°C, treated for 30 minutes with RNase A (10 μg/ml), and washed extensively at 42°C with 0.1× standard saline citrate. Pre-incubation for 1 hour in blocking solution (150 mmol/L NaCl, 100 mmol/L Tris, pH 7.5, 1% rabbit serum, 1% goat serum, 1% sheep IgG, 0.15 Triton X-100), was followed by 2 hours in blocking solution containing alkaline phosphatase-conjugated sheep anti-DIG. After washing extensively, sections were equilibrated in AP buffer (100 mmol/L Tris, pH 9.5, 100 mmol/L NaCl, 50 mmol/L MgCl₂, 1 mmol/L levamisole) and BCIP/NBT solution applied in AP buffer. Color development was done for 16 hours at room temperature in the dark.

In Situ Binding of ¹²⁵I-MIP-1αP in Viable Human Skin

Our previous in situ binding experiments 22 had demonstrated the presence on lymphatic endothelium of a chemokine receptor to which RANTES, MCP-1, and MCP-3 (but not MIP-1αS and IL-8) bind and fully cross-compete for each others binding. To our knowledge D6 is the only chemokine receptor that can bind RANTES and MCP-1 but not IL-8 or MIP-1αS. This prompted us to investigate the possible D6 expression on lymphatic ECs.

To further assess the specificity of lymphatic chemokine receptor, we performed in situ ligand-binding experiments on human skin using ¹²⁵I-MIP-1αP, a MIP-1α isoform, which unlike the previously studied MIP-1αS, 22 binds to D6 with high affinity. 30 As shown in Figure 1 ▶ , MIP-1αP bound selectively to the EC lining of lymphatics and some resident dermal cells but weakly, if at all, to vascular ECs. Thus, our current results, together with our previous findings, 22,30 suggest the presence of chemokine receptors with ligand specificity identical to D6 on the surface of dermal lymphatic ECs. The almost complete cross-competition of RANTES binding by MCP-1 and vice versa 22 excludes the significant involvement of specific chemokine receptors, eg, CCR-2 and CCR-5 in chemokine in situ binding to lymphatic ECs.

Generation and Analysis of Monoclonal Anti-Human D6 Antibodies

To further examine tissue distribution and confirm lymphatic EC expression of D6, a panel of monoclonal antibodies was generated against cells transfected with the human D6 gene (see Materials and Methods). Supernatants from hybridomas were screened by immunohistochemistry and FACS analysis for binding to D6 expressed on heterologous cell lines. Several supernatants contained D6 binding antibodies (Figure 2) ▶ . One clone (4A5) appeared to have the highest titer and was further purified on a protein G column. None of the antibody supernatants tested was able to block ¹²⁵I-MIP-1α binding to D6 transfectants, or detect human D6 in D6 transfectant cell extracts using Western blotting (not shown), suggesting that they recognize conformational epitopes outside the ligand-binding domain. Clones were tested for their ability to bind to a panel of cell lines stably transfected with individual human chemokine receptors [CCR1–5 (Figure 2) ▶ , CXCR1–3 (not shown)] and no cross-reaction was observed. Furthermore, using FACS analysis, the 4A5 monoclonal did not bind to any cell types from peripheral blood, including permeabilized cells or cells fixed before antibody staining. Human D6 transfectants and the Jurkat T cell line that expresses the D6 gene, both bound the 4A5 antibody in FACS analysis (data not shown). Moreover, we have not observed staining by 4A5, or other clones, of cells within blood vessels in any of the tissue sections that we have screened (see below). These results effectively exclude cross-reaction with a wide array of chemokine receptors known to be expressed on peripheral blood cells. Also, importantly, these data show that D6 expression on, or within, peripheral blood cells is below the level of detection with our monoclonal antibodies. These results are in accordance with gene expression analysis that has shown that D6 mRNA is of low abundance in peripheral blood samples. 28 However, it remains a possibility that hemopoietic cells express this protein if they are subjected to the appropriate stimulation or activation.

Monoclonal Anti-D6 Antibodies Stain Lymphatic ECs in Human Skin

Next, we used the mouse monoclonal anti-D6 antibodies in immunohistochemical analysis of frozen and paraffin-embedded sections of human skin. All of the anti-D6 antibody clones tested stained ECs lining thin-walled channels within the dermis that had the morphological appearance of afferent lymphatics (Figure 3; A, B, and D ▶ ). Blood vessels corresponding to different segments of the circulatory tree were immunonegative in all sections tested (Figure 3B ▶ and not shown). In some skin sections, both frozen and paraffin-embedded, a few individual cells scattered within the lower and more rarely upper dermis showed D6 immunoreactivity (Figure 3B ▶ , insets). These cells were not associated with either blood or lymphatic vessels and their identity is currently under investigation. In these experiments as in all of the subsequent analyses, immunohistochemistry using anti-D6 antibodies was performed alongside controls on adjacent sections using no primary antibody and equimolar isotype-matched irrelevant control antibody that showed no specific staining (not shown). The stainings presented are representative of that seen in several sections from skin from at least five different individuals.

D6-Immunoreactive Dermal Structures Express Podoplanin Immunoreactivity

To confirm the identity of the D6-immunoreactive ECs as lymphatic, we stained the adjacent serial sections of human skin with antibodies against podoplanin, a known lymphatic EC marker. 4 As shown in Figure 3, C and D ▶ , all of the vessels in skin that expressed D6 contained podoplanin immunoreactivity. However, podoplanin expression was considerably broader than D6 with many podoplanin immunoreactive vessels being D6-negative (not shown). This suggests that only a subset of lymphatics in skin expresses D6. The exact fingerprint of D6 chemokine specificity observed in the in situ binding to lymphatics, and the D6 immunoreactivity of lymphatic ECs, together strongly imply that D6 is specifically expressed on lymphatic endothelium.

Lymphatic ECs Express D6 mRNA

Next, sections of skin were examined for expression of D6 mRNA using in situ hybridization. Although sense D6 probes produced no specific hybridization (Figure 4A) ▶ , antisense D6 probes hybridized to ECs lining vessels in the dermis, both in regions near the epidermis (Figure 4B) ▶ and also deeper within the dermis (Figure 4B ▶ , inset). These vessels had the morphological appearance of lymphatics. To confirm this, serial sections were either hybridized with antisense D6 probes or subjected to immunohistochemistry using anti-podoplanin antibodies (Figure 4 ▶ ; C, D, E, and F). ECs lining the superficial (Figure 4C) ▶ and the deep dermal (Figure 4E) ▶ vessels that hybridized effectively to the antisense D6 probe also showed podoplanin immunoreactivity (Figure 4, D and F) ▶ . ECs lining blood vessels, such as the one seen in Figure 4, E and F ▶ , did not hybridize to the antisense D6 probe and showed no podoplanin immunoreactivity. These observations provide further evidence that lymphatic ECs express D6, and strongly suggest that the anti-D6 antibodies are not cross-reacting with other antigens on these cells.

D6 Is Expressed by the Lymphatics in the Gut

Next we embarked on a search for D6-immunoreactive lymphatics in a panel of normal human tissues. As shown in Figure 5 ▶ , anti-D6 antibodies stained abundant lymphatics in the wall of large and small intestine and appendix. Small lymphatics located in the villi of small and large intestine and in the lamina propria mucosae of colon (Figure 5; A, B, and C ▶ ) and large collective lymphatics located in the muscular layer (Figure 5D) ▶ displayed D6 immunoreactivity. However, not all of the lymphatics identified as such by their podoplanin immunoreactivity were D6-positive (not shown). In appendix, a network of D6-positive lymphatics was observed in the lymphoid tissue of the lamina propria (Figure 5E) ▶ and in the lamina muscularis externa (Figure 5F) ▶ . As observed in skin sections (Figure 3) ▶ , blood vessels of different caliber representing different segments of circulatory tree were consistently D6-negative (eg, Figure 5D ▶ ). Isotype control antibody showed no specific staining in any of the tissues examined (not shown). However, in the gut one anti-D6 clone out of four tested, weakly immunostained additional structures, eg, the enterocytes lining the lumen (not shown). The lack of the epithelial immunoreactivity by all of the other anti-D6 monoclonal antibody clones tested (all stained the afferent lymphatics) strongly suggests that enterocyte staining may be because of cross-reactivity with an unrelated epitope. All of the observations were on normal tissue from at least three different individuals; D6-immunoreactive lymphatics were also seen in a section from an acute appendicitis specimen (Figure 5G) ▶ . Anti-D6 antibodies failed to reveal the presence of D6-immunoreactive ECs in paraffin sections of heart, kidney, liver, skeletal muscle, brain, cerebellum, pancreas, prostate, and thyroid, although in some tissues (eg, lung, liver, and placenta) D6 immunoreactivity was seen in cells that were not lymphatic ECs. The identity of these cells is being investigated.

In conclusion, whereas in several organs known to contain lymphatics (eg, heart and liver) no D6-immunoreactive vessels could be detected, skin and gut, both tissues in intimate contact with external environment, contained D6-immunoreactive lymphatics.

D6 Immunoreactivity in Secondary Lymphoid Organs

Because of the presence of D6-immunoreactive channels within the lymphoid tissue of the appendix, we also examined sections from other secondary lymphoid organs (ie, tonsils, spleen, and lymph nodes), for the presence of D6-immunoreactive structures. The parafollicular areas of the tonsils contained abundant D6-immunoreactive sinus-like channels (Figure 6, A and B) ▶ ; HEVs were D6-negative (Figure 6A) ▶ . The white pulp of spleen contained no D6-immunoreactive structures, whereas in one out of four different spleens tested, D6-immunoreactive vessels were observed in the red pulp (Figure 6C) ▶ .

In the lymph nodes, afferent lymphatics entering the nodes were D6-positive (not shown) suggesting, together with the findings in skin and gut, that afferent lymphatics may express D6 throughout their entire length. Also, cells lining the subcapsular and medullary sinuses displayed strong D6 immunoreactivity (Figure 6, D and F) ▶ . Again, no blood vessel ECs were D6-immunoreactive in any secondary lymphoid tissue, including cells lining HEVs. Using the anti-podoplanin antibodies on adjacent sections from lymph nodes and tonsils revealed that D6-immunoreactive vessels also expressed podoplanin-immunoreactivity (Figure 6, E and G ▶ , and not shown), but additionally, there were lymphatic sinuses detected that reacted with the anti-podoplanin antibodies only (not shown).

D6 Is Expressed by a Subset of Vascular Tumors

There has been considerable debate concerning the origin and differentiation of the aberrant ECs seen in malignant vascular tumors, ie, whether they are blood vessel EC-like or lymphatic EC-like. Recently, the lymphatic EC markers podoplanin and VEGFR-3 have been shown in some vascular tumors. 1,2,4,35 As D6-immunoreactive ECs are exclusively lymphatic, we reasoned that D6 immunoreactivity might be expressed in the vascular tumors that show lymphatic origin or differentiation. As an initial assessment, we screened a panel of 15 vascular tumors predominantly from the skin for the expression of D6 immunoreactivity, and by using adjacent sections, for the expression of podoplanin immunoreactivity. As shown in Table 1 ▶ , there was considerable variation in the presence of these two markers. Examples of vascular tumors stained with the D6 antibodies are shown in Figure 7 ▶ . In general, those tumors that expressed podoplanin also displayed, on a variable proportion of their cells, D6 immunoreactivity. The same 11 tumor samples showed some immunoreactivity with both antibodies, although in only three of these cases D6 immunoreactivity was present in almost all of the cells of the tumor (Figure 7 ▶ ; A, B, and D): podoplanin immunoreactivity was expressed on 100% of the tumor cells in these three cases plus a further four samples (Figure 7, E and F ▶ , and Table 1 ▶ ). In fact, podoplanin immunoreactivity was consistently seen to be expressed on a higher percentage of the tumor cells than D6. Also, the intensity of the staining with the D6 antibodies varied between samples: six samples showed intense immunoreactivity, three stained only weakly, and two tumors contained populations of cells that expressed different levels of D6 immunoreactivity (Table 1) ▶ . These results further support the observations made by others, 1,2,4,35 showing that vascular tumors often express markers of lymphatic ECs. However, there is considerable heterogeneity between vascular tumor samples with respect to the distribution and levels of expression of these markers.