CD112 Regulates Angiogenesis and T Cell Entry into the Spleen

Abstract

Junctional adhesion proteins play important roles in controlling angiogenesis, vascular permeability and leukocyte trafficking. CD112 (nectin-2) belongs to the immunoglobulin superfamily and was shown to engage in homophilic and heterophilic interactions with a variety of binding partners expressed on endothelial cells and on leukocytes. Recent in vitro studies suggested that CD112 regulates human endothelial cell migration and proliferation as well as transendothelial migration of leukocytes. However, so far, the role of CD112 in endothelial cell biology and in leukocyte trafficking has not been elucidated in vivo. We found CD112 to be expressed by lymphatic and blood endothelial cells in different murine tissues. In CD112-deficient mice, the blood vessel coverage in the retina and spleen was significantly enhanced. In functional in vitro studies, a blockade of CD112 modulated endothelial cell migration and significantly enhanced endothelial tube formation. An antibody-based blockade of CD112 also significantly reduced T cell transmigration across endothelial monolayers in vitro. Moreover, T cell homing to the spleen was significantly reduced in CD112-deficient mice. Overall, our results identify CD112 as a regulator of angiogenic processes in vivo and demonstrate a novel role for CD112 in T cell entry into the spleen.

Keywords: T cell homing; angiogenesis; blood vessels; nectin-2; spleen.

Figure 1

CD112 is expressed by endothelial cells in vivo. (A,B) Flow cytometry analysis of mouse ear skin single-cell suspensions. (A) Endothelial cells were identified as CD31⁺CD45⁻ cells (left) and further divided into blood endothelial cells (BECs) and lymphatic endothelial cells (LECs) based on podoplanin (podo) expression (right). (B) Representative flow cytometry plot showing in vivo expression of CD112 by LECs (green; CD31⁺podo⁺) and BECs (red; CD31⁺podo⁻). (C) Summary of median fluorescent intensity (MFI) values of CD112 expression on murine LECs and BECs of three experiments. Data points from the same experiment are connected by a line. (D) Low magnification confocal images of ear skin whole-mount immunofluorescence staining visualizing CD112 expression (white) by lymphatic vessels (LVs) and blood vessels (BVs), indicated by white arrow heads. Scale bar: 50 μm. (E) High magnification confocal images revealed a button-like expression pattern for CD112 (white) in lymphatic capillaries (LYVE-1⁺/VE-cadherin⁺ upper panel), respectively, and a zipper-like expression pattern in lymphatic collectors (LYVE-1⁻/VE-cadherin⁺ lower panel). Scale bar: 20 μm.

Figure 2

CD112 regulates endothelial cell function in vitro. (A,B) Flow cytometry analysis showing CD112 expression by human (A) and conditionally immortalized murine LECs (imLECs) (B). Representative flow cytometry plots of CD112 expression comparing steady-state (blue line) and inflamed conditions (red line: TNFα/IFNγ treated; grey lines: isotype controls) are shown on the left. Summary of MFI values of CD112 expression of three to four experiments are shown on the right. Data points of the same experiment are connected by a line. (C) Western blot analysis configure 112. protein (~65 kDa) expression in in vitro cultured human LECs in steady-state (SS) and upon TNFα/IFNγ-mediated inflammation. β-actin was used as an internal control. Blots were cut at the 50 kDa ladder mark. Representative data from one out of three experiments are shown. (D) Summary of densitometry signal intensities of CD112 protein in steady state (SS) and upon TNFα/IFNγ-mediated inflammation normalized to β-actin signal of three Western blot experiments. Data points from the same experiment are connected by a line. (E) Quantitative real-time PCR analysis of CD112 mRNA levels in in vitro cultured human LECs in steady-state (SS) and upon TNFα/IFNγ-mediated inflammation. Induction of ICAM-1 mRNA levels by TNFα/IFNγ was analysed as a positive control. Delta cycle threshold (CT)values: Difference between the CT values measured for CD112 and for the housekeeping gene RPLP0. Pooled data (experimental means) from three independent experiments (biological replicates) are shown. Each experiment was performed in triplicate (technical replicates). (F) Representative immunofluorescence images of CD112 expression (white) on in vitro cultured imLEC monolayers colocalizing with the junctional molecule VE-cadherin (green) at the intercellular junctions. Scale bar: 20 μm. (G,H) Human LEC migration upon CD112 blockade was investigated in an in vitro scratch assay. (G) Representative images of an LEC scratch assay in VEGF-A-induced wound closure. Scale bars: 200 µm. (H) Pooled quantitative analysis from three independent experiments are shown. (I) Representative image (left) of an in vitro tube formation assay. Confluent human LEC monolayers were incubated overnight in collagen gel solution supplemented with blocking antibody for CD112 or isotype control. Total tube length was manually analysed using a self-made macro in FIJI (ImageJ). Quantitative analysis (right) of total tube length upon CD112 blockade. One out of three similar independent experiment is shown. (J) In vitro permeability assay, in which 70 kDa FITC dextran was applied onto confluent pretreated LEC monolayers grown on a Boyden transwell membrane. After 30 min, diffused FITC dextran was quantified in the lower chamber. Pooled data from three similar and independent experiments are shown.

Figure 3

CD112 deficiency increases blood vessel coverage in the retina and spleen in mice. (A–E) Whole mounts from postnatal 6 (P6) retinas were stained with isolectin B4 (iB4) to visualize the blood vessels of the superficial vascular plexus, followed by quantification of vascular parameters. (A) Representative confocal images of iB4 stained P6 retinal whole mounts of CD112^−/− and Wild type (WT) pups. 4× objective 1.0 zoom. Image-based morphometric analysis showing (B) the iB4+ area (left) and percent covered by blood vessels (right) and (C) the superficial vascular plexus length (SVPL). Scale bar: 500 µm. Data from one out of 3 similar experiments (n = 4–6 mice/group). (D,E) Upon vessel maturation, vessels undergo pruning, which can be quantified by analysing the blood vessel density. (D) Representative images of P6 retinas of WT and CD112^−/− stained with iB4. 20× objective, 1.0 zoom, scale bar: 80 µm. Analysis of (E) the branch length formed in the peri-arterial space in CD112^−/− mice. Data from one experiment are shown (n= 4–6 mice/group). (F–M) Immunofluorescence and flow cytometry analysis of spleen single-cell suspension. (F–I) Sections sizes of 40 μm of optimum cutting temperature (OCT)-frozen spleens from CD112^−/− and WT mice were stained for MadCAM-1 (F) and MECA-32 (G). Representative pictures are shown. Scale bar: 200 µm. (H,I) Quantification of the MadCAM-1⁺ (H) and MECA-32⁺ (I) area in spleen of CD112^−/− mice; n = 5 mice/group. (J,K) Quantification of the number of (J) podo⁻CD31⁺ and (K) CD31⁺MadCAM-1⁺ BECs present in the spleen of WT and CD112^−/− mice. Pooled data from three experiments are shown. (L) Gating strategy of spleen single-cell suspensions. BECs were gated on CD45⁻podo⁻CD31⁺ (left), MadCAM-1⁺ CD31⁺ (right). (M) Representative histogram of several independent experiments showing CD112 expression by BECs in the spleen. Blue: CD31⁺, red: CD31⁺MadCAM-1⁺, grey: isotype controls.

Figure 4

CD112 regulates CD4⁺ T cell transmigration through BECs in vitro. (A,B) In vitro transmigration assay with human T cells and primary human LECs. Quantification of transmigration efficiency upon either ICAM-1 antibody blockade (clone BBIG-I1, mouse IgG1) (A) or CD112 antibody blockade (clone R2.525, mouse IgG1) (B). Pooled data from three similar experiments are shown (n = 3 replicates per group and per experiment, i.e., 9 per pooled group). (C) Flow cytometry analysis showing CD112 protein expression by cultured blood vascular MS-1 cells. Representative histograms from three independent experiments are shown (black: CD112, grey: isotype control). (D) Representative histograms demonstrating the cross-reactivity of mouse anti-human CD112 antibody (R2.525, R&D systems) with mouse CD112 (black doted: antibody binding to CD112; grey: isotype controls). (E–H) In vitro transmigration assay of freshly isolated murine CD4⁺ T cells across blood vascular MS-1 monolayers in presence of anti-ICAM-1 (clone YN1, rat IgG2b) (E,F) or anti-CD112 (clone R2.525, mouse IgG1) (G,H). The number of transmigrated T cells was quantified by flow cytometry after 4 h. (E,G) show absolute numbers of transmigrated cells and (F,H) show values normalized to the isotype control. Pooled data from three similar experiments are shown (n = 3 replicates per group and per experiment, i.e., 9 per pooled group). (I) Representative histogram showing CD112 expression in WT (red) and CD112^−/− (blue) primary LN LECs compared to isotype controls (grey). Data from one out of three independent experiment are shown. (J) In vitro transmigration of freshly isolated murine CD4⁺ T cells across primary WT or CD112^−/− LN LECs. Pooled data from four similar experiments are shown (n = 4 replicates per group and per experiment, i.e., 12 per pooled group).

Figure 5

CD4⁺ T cells express CD226. (A–F) Analysis of CD112 binding partner expression on T cells. CD226 (A) and T cell immunoreceptor Ig tyrosine (TIGIT) (B) expressions on freshly isolated splenic CD4⁺ T cells (red) and in vitro generated T_H1 cells (blue). (C) Characterisation of naïve, effector/memory and central memory T cells based on CD44 and CD62L expressions of splenic CD4⁺ T cells. (D) CD226 (left) and TIGIT (right) expression on naïve (red), effector/memory (green) and central memory (black) T cells. Representative graphs from one out of three similar experiments are shown. (E,F) Percentages of CD226⁺ (E) or TIGIT⁺ (F) T cells. T_eff/mem: effector/memory; Table 3. 5. CD112 Deficiency Reduces Homing of T Cells to the Spleen.

Figure 6

Absence of CD112 reduces homing of adoptively transferred T cells to spleen. (A–I) T cell homing experiment into secondary lymphoid organs (SLOs). (A) CD2-dsRED⁺ leukocytes were injected intravenously into either WT or CD112-deficient mice. After 2.5 h, SLOs were harvested and the number of homed T cells was quantified with flow cytometry. (B) Representative gating strategy of adoptively transferred CD2-dsRED⁺ T cells in murine spleen. Single-cell suspensions were gated on CD45⁺CD2-dsRED⁺ and further subdivided into CD4⁺/CD8⁺ cells. (C) Numbers of adoptively transferred CD2-dsRED⁺ T cells in different SLOs (PPs: Payer’s patches, MLN: mesentery LN, PLN: peripheral LN). Representative data from one out of three experiments are shown. n = 5–7 mice/group, ns = not-significant. (D–F) Quantification of (D,E) total CD2-dsRED⁺, (F,G) CD2-dsRED⁺CD4⁺ and (H,I) CD2-dsRED⁺CD8⁺ T cells in the spleen of WT and CD112^−/− mice from. (D,F,H) show representative data from one out of three independent experiments (n = 5–7 mice/group/experiment) performed. (E,G,I) show a summary of the mean values obtained in the three individual experiments. Data points (WT—CD112^−/−) from the same experiment are connected by a line. (J–L) Analysis of CD112 expression in LN HEVs. (G) Gating strategy of LN single-cell suspensions. CD45⁻ stromal cells were further divided by CD31 and podo expression. (K) HEV BECs were identified as podo⁻CD31⁺PNAd⁺ cells. Representative plots from three independent experiments are shown. (L) Representative histogram showing CD112 expression in CD31⁺ LN BECs (blue) and CD31⁺PNAd⁺ HEV BECs (red). (M,N) Comparison of CD112 and ICAM-1 expression in CD31⁺PNAd⁺ LN BECs and splenic CD31⁺ BECs. (M) Representative histograms and (N) quantification of the median fluorescent intensities (MFIs) from three independent experiments.