|Home | About | Journals | Submit | Contact Us | Français|
The field of platelet biology has rapidly expanded beyond the classical role of platelets in preventing blood loss and orchestrating clot formation. Despite the lack of transcriptional ability of these anuclear cell fragments, platelet function is now thought to encompass such diverse contexts as tissue repair, immune activation, primary tumor formation, and metastasis. Recent studies from multiple groups have turned the spotlight on an exciting new role for platelets in the formation of lymphatic vessels during embryonic development. Genetic experiments demonstrate that Podoplanin, a transmembrane protein expressed on lymphatic endothelial cells, engages the platelet CLEC-2 receptor when exposed to blood, leading to SYK-SLP-76-dependent platelet activation. When components of this pathway are disrupted, aberrant vascular connections form, resulting in blood-lymphatic mixing. Furthermore, platelet-null embryos manifest identical blood-lymphatic mixing. The identification of platelets as the critical cell type mediating blood-lymphatic vascular separation raises new questions in our understanding of lymphatic development and platelet biology.
Hematopoiesis and vascular development have shared origins as progenitors of both lineages initially arise in extraembryonic blood islands. In the embryo, endothelial precursors surround clusters of primitive hematopoietic precursors prior to vascular lumen formation.1 Definitive hematopoietic stem cells subsequently arise from a subset of competent endothelial cells of the dorsal aorta in the aorta-gonad-mesonephros (AGM) region of the developing embryo.2, 3 There has been a long and heated debate over the contribution of bone marrow-derived cells (BMDCs) to endothelium during development that has been complicated by the close spatiotemporal association and common molecular markers between the two lineages. Fatemapping studies conducted in blood vessels using Vav-Cre transgenic mice 4 and in lymphatic vessels using Runx1-MER-Cre-MER inducible knockin mice 5 showed that most, if not all, endothelial cells are not of hematopoietic origin. However, the distinct blood-lymphatic mixing vascular phenotype in mice lacking the hematopoietic proteins SYK or SLP-76 suggested there could be a small but critical requirement for bone marrow-derived endothelium. This review will highlight definitive evidence that SYK/SLP-76 knockout vascular abnormalities do not arise from a failure of BMDCs to incorporate into the vasculature but rather from the requirement of hematopoietic cells, specifically platelets, to regulate blood-lymphatic separation. We will present contributions by a number of groups that describe the discovery of a novel embryonic role for platelets in regulating vascular development.
Lymphatic vessels form an extensive vascular network that facilitates immune trafficking and surveillance, maintains tissue fluid homeostasis, and absorbs dietary lipids in the intestine through mesenteric lacteals.6 These specialized vessels are composed of lymphatic endothelial cells (LECs) that originate in the cardinal vein as venous endothelial cells. Starting at E9.75, polarized induction of PROX1, a homeobox transcription factor, is necessary and sufficient to activate the lymphatic gene profile.5, 7 These LECs then migrate away from the cardinal vein largely in response to VEGFR3/Neuropilin-2 signaling induced by VEGF-C in the mesoderm.8 The LECs assemble into vessels to form a de novo parallel but separate collecting system from the blood vessels, sprouting centrifugally from the initial sacs and migrating deeper into tissues to form the inner lymphatic plexus.5 The resulting collecting vessels form luminal valves to prevent backflow and are covered by mural cells, while the blind-ended capillaries lack mural cells and remain largely uninvested by pericytes. Lymphatic capillary endothelium is characterized by loose, intercellular junctions that facilitate the homeostatic function of these vessels.
A careful kinetic study of murine intestinal lymphatic development reveals that mature blood vessels are established days before lymphatics invade the mesentery and the intestinal tube. By E13.5–14.5, although local blood vessel development is complete, there are few or no lymphatics in the wall of the small intestine and immature lymphatics have just begun to invade the mesentery. By E16.5, there has been a remarkable increase in branching and further migration into the intestinal tube, but it isn’t until E19.5 that they have formed a mature lymphatic network capable of transporting the large amounts of fat required for nutritional absorption.9 The completed vascular and lymphatic endothelial networks connect only at the thoracic duct in the mouse, where the collected lymphatic fluid is returned to the blood circulation through the subclavian vein.5, 10, 11
Mice lacking SYK,12, 13 SLP-76,14 or PLCγ-2 15 develop blood-lymphatic mixing in distinct vascular beds at specific timepoints during embryonic development. This is the result of primary misconnections between lymphatics and blood vessels.14 Disruptions in normal development are first detectable at E11.5 as the nascent lymphatic sacs are emerging from the cardinal vein along the antero-posterior axis of the embryo. Small connections between lymphatic and venous endothelial cells persist along the cardinal vein in the mutant mice, resulting in downstream filling of the superficial dermal lymphatics with blood and general edema by mid-gestation. Lymphatic development continues largely unimpeded, but severe vascular abnormalities also become apparent in the intestine by E16.5. Mesenteric lymphatics are invading the vascularized intestinal tube and the mutants develop a gross perturbation of what are normally separate, arborized vessels. The mesenteric lymphatic vessels themselves appear structurally normal, but once these vessels reach the intestinal wall they develop myriad connections with the blood circulation. Real-time injection of fluorescent dextran into the circulation reveals that the intestinal lymphatics become perfused with dextran and are functionally indistinguishable from their venous counterparts. The vast majority of these animals die perinatally.14 Notably, some vascular beds in these animals appear completely unaffected, including those found in the lungs where there is a high density of both lymphatic and blood vessels.
The observation that SLP-76 knockout mice develop a primary vascular defect was particularly surprising because expression of this protein is confined to the hematopoietic compartment where it nucleates signaling complexes downstream of immunoreceptor tyrosine-based activation motif (ITAM) activation. Expression of SLP-76 has not been detected in endothelial cells, even at the cardinal vein during the emergence of LECs.14 Reconstitution of the bone marrow and circulating blood cells in lethally irradiated wildtype mice with Slp-76−/− hematopoietic cells results in the development of intestinal blood-lymphatic mixing, highlighting the requirement for SLP-76 in BMDCs.14 Furthermore, transgenic rescue of SLP-76 in a limited profile of hematopoietic cells but not in endothelial cells is sufficient for normal lymphatic development.16 Runx1−/− mice, which fail to undergo definitive hematopoiesis, demonstrate similar blood-filled lymphatic sacs at E11.5 also suggesting a hematopoietic requirement for normal lymphatic development.5 Careful analysis of hematopoietic lineage tracing using a Runx1-MER-Cre-MER inducible knock-in mouse 17 to examine the emerging lymphatic endothelial sacs failed to reveal any direct contribution of BMDCs to these vessels.5 Finally, although hematopoietic-specific deletion of SLP-76 using Vav-Cre confers the vascular phenotypes seen in the straight knockout, lineage tracing in these mice also failed to demonstrate any hematopoietic origin of LECs, even in the highly affected gut where the requirement is greatest.18
How does a hematopoietic signaling pathway regulate lymphatic vessel growth and development? A major insight into this question came with the discovery that Podoplanin (PDPN) null mice exhibit blood-lymphatic mixing and vascular abnormalities that closely phenocopy those of mice lacking SYK, SLP-76 and PLCγ2.18–20 PDPN is a heavily O-glycosylated Type I transmembrane protein whose expression is largely restricted to glomerular podocytes, lung alveolar type I cells, and LECs, but is not seen in hematopoietic cells or blood endothelial cells. Endothelial loss of the T-synthase enzyme required for PDPN synthesis is sufficient for blood-lymphatic mixing like that observed in mice lacking the immune receptor signaling pathway in blood cells.19 Radiation chimeras confirm that the requirement for PDPN is not in blood cells,18 consistent with an essential role in LECs. These studies and the finding that loss of PDPN magnified the penetrance of the vascular mixing phenotype in Vav-Cre; Slp-76fl/fl animals placed PDPN in the same pathway as SYK, SLP-76 and PLCγ2, but in a different cell type.
PDPN, also known as aggrus, was identified as a tumor cell surface protein capable of initiating platelet aggregation.21, 22 The mechanism for this response was recently identified as PDPN-mediated activation of a novel platelet receptor, C-type lectin-like receptor 2 (CLEC-2), that is highly expressed on platelets.23 The intracellular tail of CLEC-2 contains a single YxxL motif that initiates downstream signaling through SYK and SLP-76 upon ligand engagement of a CLEC-2 dimer,24 providing a molecular explanation for how PDPN and these hematopoietic signaling proteins may be linked. Genetic deletion of CLEC-2 in mice reveals a striking phenocopy of the blood-lymphatic mixing defects seen in mice deficient for SLP-76 or PDPN.18, 25 Radiation chimeras reconstituted with Clec-2−/− fetal liver cells develop the intestinal mixing phenotype, confirming a hematopoietic requirement for CLEC-2.18 In situ and flow cytometry studies indicate that CLEC-2 is restricted to platelet-generating megakaryocytes and platelets in the developing embryo,18 although in older animals it is also reported to be present on peripheral blood neutrophils.26 These studies suggested that PDPN on LECs activates CLEC-2 receptors and downstream SYK-SLP-76 signaling in platelets to initiate and maintain blood-lymphatic vascular separation. This unexpected mechanism is confirmed by the finding that platelet-specific deletion of SLP-76 is sufficient to confer the blood-lymphatic mixing phenotype.18 Importantly, separate lines of investigation have provided genetic confirmation of the role of platelets in this vascular regulatory mechanism.18, 27 Platelet-null embryos generated either by deletion of the transcription factor MEIS1 or through platelet-specific expression of diptheria toxin exhibit the blood-lymphatic mixing phenotype.27 These studies demonstrate that platelets are responsible for mediating blood and lymphatic separation through activation of the CLEC-2 receptor by PDPN ligand presented on the surface of LECs.
It remains unknown how abnormal blood-lymphatic vascular connections arise in the absence of CLEC-2-PDPN signaling. Platelet-LEC interactions can be seen in the cardinal vein during the specification of PDPN-expressing LECs from venous endothelial cells,18, 20 but similar interactions have not been detected in the intestine where large numbers of blood-lymphatic connections are formed.18 Since both PDPN and CLEC-2 are transmembrane proteins, the mechanism requires direct contact between LECs and platelets. The use of platelets as a means of sensing blood vessel contact by LECs makes sense, as platelets are one of the few blood cell types that cannot extravasate and enter lymphatic vessels, even following trauma. Precisely how platelet activation by LECs negatively regulates the formation of LEC connections to blood vessels remains speculative. The requirements for SYK and SLP-76 indicate a need for platelet activation downstream of the PDPN-CLEC-2 interaction. The fact that mice lacking platelet integrins required for platelet aggregation do not form blood-lymphatic connections suggests that the formation of a platelet plug, the hallmark of platelet-mediated hemostasis, may not be the key step downstream of platelet activation. Instead, it is tempting to speculate that platelet degranulation may release regulators of LEC growth that inhibit the formation of blood-lymphatic connections. Platelet alpha granules are known to contain numerous angiogenic regulators, supporting the hypothesis that degranulation is the mechanism by which platelet activation controls lymphatic growth. Further genetic studies will be required to test the effects of such agents on LEC growth and lymphatic development.
The studies described above are also significant because they describe a clear embryonic role for platelets that is not connected to hemostasis. Platelet activation has been associated with many non-hemostatic roles in mature animals, including inflammation,28, 29 wound healing,30, 31 primary tumor growth,32 and tumor metastasis,33 but few embryonic roles have been defined. The lack of an embryonic phenotype in NF-E2-deficient mice lacking most but not all platelets 34, 35 has suggested that platelets do not have such roles, but this is clearly not the case and it seems likely that future studies will identify other roles for platelets in regulating development of the cardiovascular and other systems.
Finally, it remains unclear whether PDPN and/or CLEC-2 play roles outside of lymphatic vascular development. The number of CLEC-2 receptors on the surface of human or mouse platelets has not yet been determined. However, comparison of hematopoietic and megakaryocyte gene analysis libraries, as well as in situ hybridization studies of mouse megakaryocytes and binding of anti-CLEC-2 antibodies and PDPN-Fc fusion proteins suggests that CLEC-2 may be one of the most highly expressed activating receptors on the platelet surface.18, 23 In vitro studies of Clec-2−/− platelets have suggested that CLEC-2 deficiency does not affect collagen-induced aggregate formation.36 Furthermore, tyrosines in the YxxL motif of platelet CLEC-2 receptors are not phosphorylated upon flow over collagen,36 suggesting that CLEC-2 is not involved in platelet activation by classical hemostatic stimuli. The existence of a PDPN-independent role for CLEC-2 remains unclear and is an open question in the field. Future studies addressing the hemostatic role(s) of CLEC-2 and PDPN are likely to yield new insights into platelet biology.