Podoplanin maintains high endothelial venule integrity by interacting with platelet CLEC-2

Abstract

Circulating lymphocytes continuously enter lymph nodes for immune surveillance through specialized blood vessels named high endothelial venules, a process that increases markedly during immune responses. How high endothelial venules (HEVs) permit lymphocyte transmigration while maintaining vascular integrity is unknown. Here we report a role for the transmembrane O-glycoprotein podoplanin (PDPN, also known as gp38 and T1α) in maintaining HEV barrier function. Mice with postnatal deletion of Pdpn lost HEV integrity and exhibited spontaneous bleeding in mucosal lymph nodes, and bleeding in the draining peripheral lymph nodes after immunization. Blocking lymphocyte homing rescued bleeding, indicating that PDPN is required to protect the barrier function of HEVs during lymphocyte trafficking. Further analyses demonstrated that PDPN expressed on fibroblastic reticular cells, which surround HEVs, functions as an activating ligand for platelet C-type lectin-like receptor 2 (CLEC-2, also known as CLEC1B). Mice lacking fibroblastic reticular cell PDPN or platelet CLEC-2 exhibited significantly reduced levels of VE-cadherin (also known as CDH5), which is essential for overall vascular integrity, on HEVs. Infusion of wild-type platelets restored HEV integrity in Clec-2-deficient mice. Activation of CLEC-2 induced release of sphingosine-1-phosphate from platelets, which promoted expression of VE-cadherin on HEVs ex vivo. Furthermore, draining peripheral lymph nodes of immunized mice lacking sphingosine-1-phosphate had impaired HEV integrity similar to Pdpn- and Clec-2-deficient mice. These data demonstrate that local sphingosine-1-phosphate release after PDPN-CLEC-2-mediated platelet activation is critical for HEV integrity during immune responses.

Figure 1. Loss of FRC PDPN or platelet CLEC-2 leads to spontaneous mucosal LN bleeding

a, Gross morphology of MLNs. Insets contain montages of confocal images of MLN cryosections showing PDPN expression in WT and Pdpn^f/f;CagCre mice. b, Images of H&E–stained MLN sections. Arrows indicate extravasated RBCs outside HEVs (dashed line) of Pdpn^f/f;CagCre mice. c, Confocal images of MLNs from WT and Pdpn^f/f;CagCre mice reveal extravasated RBCs (arrows) outside HEVs, some of which are picked up by lymphatic vessels (LVs). Ter119 indicates RBCs. CD31 marks endothelial cells. Lyve-1 marks LVs. d, Immunostaining of MLN cryosections using HEV-specific marker PNAd. Inset shows that no bleeding occurred around CD31⁺/PNAd⁻ non-HEV vessels in MLNs. e, Gross morphology and confocal images of MLN cryosections from WT and Pdpn^f/f;PdgfrbCre stained for Ter119, Lyve-1, and CD31. f, Gross morphology and confocal images of MLNs from P15 WT mice treated with isotype control rat IgG1κ or the CLEC-2 depleting antibody, INU1. g, Gross morphology and confocal images of MLN cryosections from WT and Clec-2^f/f;Pf4Cre mice stained for Ter119, Lyve-1, and CD31. Data are representatives of ≥ 12 mice/group. Scale bars, 2 mm (gross images), 50 μm (b, and confocal images). Asterisk indicates bleeding in the LN. Arrows indicate extravasated RBCs. Tissues were from 1-month-old mice unless otherwise specified.

Figure 2. FRC PDPN and platelet CLEC-2 protect LN vascular integrity during immune responses

a, Confocal images of MLN cryosections stained with antibodies to PNAd, platelets, and PDPN. Arrows indicate platelets on the abluminal side of HEVs. Arrowheads indicate platelets that are not associated with HEVs. Bar graph on the right represents percentage of HEVs with platelets on their abluminal side (mean ± s.d., 250 HEVs/mouse (n = 3)). b, Confocal images of PNAd⁺HEVs, PDPN, and transfused fluorescently labelled WT platelets (plts) in MLNs. Arrows indicate plts on the abluminal side of HEVs. c, Comparison of Evans blue permeability between MLNs and popliteal LNs (PLNs) of 1-month-old WT mice 5 min after intravenous injections (mean ± s.d., n = 15/group). d, Gross morphology (insets) and confocal images of PLNs after OVA/CFA challenge. e, Gross morphology (insets) and confocal images of CLNs from 3-week-old mice with (Pdpn^f/f;PdgfrbCre) or without (Pdpn^f/f;PdgfrbCre/Rag1^−/−) lymphocytes. f, Gross morphology (insets) and confocal images of PLNs from 1-month-old mice one week after OVA/CFA challenge and injections with a mAb (Mel-14) that blocks L-selectin-dependent lymphocyte homing or with an isotype control rat IgG. Data represent at least 8 LNs per group from at least three experiments. Scale bars, 2 mm (gross), 50 μm (confocal). Asterisk indicates bleeding, and arrows mark extravasated RBCs around HEVs (d-f). ***, P < 0.001.

Figure 3. Interactions of FRC PDPN and platelet CLEC-2 are critical for HEV junctional integrity

a, Confocal images of intravenously injected FITC-dextran (2,000 kDa) in P15 MLNs. Arrows indicate vascular leak of FITC-dextran. Insets show non-HEVs blood vessels. b, Confocal images of intravenously injected/fluorescently labelled RBCs (red) in MLNs from 2-month-old mice. Arrow and insets show labelled RBCs outside of HEVs (CD31⁺). c, Transmission electron micrographs of HEVs in MLNs. Arrow indicates gaps and RBCs between high endothelial cells. d, Confocal images of VE-cad in P8 MLNs. Insets show VE-cad staining in non-HEV blood vessels. e, Confocal analysis of intravenously injected FITC-dextran in MLNs from WT or Clec2^−/− BM chimaeras with or without previous intravenous injections of WT platelets (plts). Arrows indicate vascular leak of FITC-dextran. Inset depicts non-HEV blood vessels. Graphs on the right quantify number of leaking vessels (mean ± s.d., 300 vessels per group (n = 3)). f, Immunofluorescence analysis of anti-VE-cad staining of HEVs and non-HEV blood vessels in WT BM chimaeras (WT chimaera), and Clec2^−/− BM chimaeras (Clec2^−/− chimaeras) without or with previous intravenous injections with WT plts. # indicates non-specific staining as also observed in negative controls. Graphs on the right quantify VE-cad staining (mean ± s.d., 300 vessels per group (n = 3)). Tissues were from 1-month-old (a-c) or 12-week post BM transplantation (e, f) mice unless otherwise specified. Dashed lines mark HEVs. Data represent at least three individual experiments. Scale bars, 2 μm (TEM), 50 μm (confocal images (inset a and b, 25 μm)). ***, P < 0.001.

Figure 4. S1P from platelets after PDPN-CLEC-2-dependent activation contributes to HEV barrier function

a, ELISAs of S1P concentrations in supernatants of WT and Clec-2^−/− platelets after incubation with CLEC-2 activating antibody, INU1 (INU1-treated) or isotype control rat IgG (sham-treated). S1P in platelet lysates (plt lysates) was used as the positive control (mean ± s.d., n = 4 mice per group representing two individual experiments). b, Representative images of VE-cad staining of HEVs from WT LN slices incubated for 1.5 hrs with DMEM and normal FBS (media), DMEM and lipid-depleted FBS (lipid-depleted), lipid-depleted and WT platelets (+WT plts), lipid-depleted and Clec-2^−/− platelets (+Clec-2^−/− plts), lipid-depleted and S1Pless platelets (+S1Pless plts), or lipid-depleted and WT plts plus S1PR1 antagonist W146 (+WT plts/W146). Pdpn^f/f;CagCre LN slices incubated with lipid-depleted and WT platelets (+WT plts) were controls. 100 HEVs analyzed per condition. Dashed lines mark HEVs. # marks non-specific staining as also observed in negative controls. Graphs represent ratios of VE-cad intensities on HEVs relative to that of WT LN slices cultured with media (mean ± s.d., n = 20 HEVs/group). c, Gross morphology (insets) and confocal images of draining popliteal LNs (PLN) after immunisation. Asterisk indicates bleeding. LV, lymphatic vessel (Lyve-1⁺). Arrow marks bleeding (Ter119⁺) around an HEV (CD31⁺). d, Confocal images of PLN HEVs from S1Pless mice transfused with Clec-2^−/− or WT platelets (plts) for 4 days and littermate controls transfused with WT plts after intravenous FITC-dextran injection. Arrows indicate vascular leak of FITC-dextran. Graphs on right quantify leaking HEVs (mean ± s.d., 50 HEVs/group (n = 3)). e, Model depicting how PDPN maintains HEV integrity during lymphocyte trafficking. FRC PDPN engages CLEC-2 on extravasated platelets in the perivenular space of HEVs and induces local release of S1P, which promotes VE-cad expression on the WT HEV (left). In contrast, loss of the interaction results in impaired HEV integrity and subsequent bleeding (right). b–d, Data is from three individual experiments. Scale bars, 50 μm. ***, P < 0.001.