Search

Changes in surface glycan determinants, specifically sialic acid loss, determine platelet life span. The gradual loss of stored platelet quality is a complex process that fundamentally involves carbohydrate structures. Here, we applied lipophilic extraction and glycan release protocols to sequentially profile N- and O-linked glycans in freshly isolated and seven-day room temperature stored platelet concentrates. Analytical methods including matrix assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS), tandem mass spectrometry (MSn), and liquid chromatography (LC) were used to obtain structural details of selected glycans and terminal epitopes. The fresh platelet repertoire of surface structures revealed diverse N-glycans, including high mannose structures, complex glycans with polylactosamine repeats, and glycans presenting blood group epitopes. The O-glycan repertoire was largely comprised of sialylated and fucosylated core-1 and core-2 structures. For both N- and O-linked glycans, we observed a loss in sialylated epitopes with a reciprocal increase in neutral structures as well as increased neuraminidase activity following platelet storage at room temperature. The data indicate that loss of sialylated glycans is associated with diminished platelet quality and untimely removal of platelets following storage.

Our understanding of cell biological processes involved in aging has advance greatly over the past decades. Platelets are small cells that circulate for 4-5 days in mice and 7-10 days in humans. And even though, platelets are anucleated cells, a growing body of evidence shows that platelet clearance is a well-regulated mechanism. We have recently demonstrated that platelets lose sialic acid as they circulate and age in blood and are rapidly cleared by the hepatic Ashwell-Morell receptor (AMR). And others have shown, in a series of studies using genetically modified mice or pharmacological inhibitors that platelets undergo apoptosis by triggering the intrinsic mitochondrial apoptotic machinery. Here, we investigate if desialylation and apoptosis are related events. First, using a newly developed state-of-the-art technique called dynamic BH3 profiling (DBP), we investigated the mitochondria readiness to undergo apoptosis on platelets derived from WT and AMR deficient (Asgr2-/-) mice. In our assay, digitonin-permeabilized platelets were exposed to activators signaling peptides (such as Bim, Bid and PUMA), and as cells undergo apoptosis due to peptide treatment, they released Cytochrome C. Our data showed that desialylated platelets derived from Asgr2-/-mice have high background levels of Cytochrome C release when compared to WT platelets in the presence of all activator peptides, indicating that desialylated platelets are highly primed to apoptosis. We also tested the level of dependence on pro-survival protein, by using sensitizer peptides (Bad, Hrk and MS1). We observed that desialylated platelets (Asgr2-/-platelets), and to a certain degree, WT platelets, are extremely sensitive to BCL-xL inhibition, as indicated by the extremely high response to Bad and Hrk peptides even at lower concentrations (0.1 and 1uM). Surprisingly, WT and Asgr2-/-platelets show very little response to the MS1 peptide, indicating that they are not dependent on MCL1 for survival, as otherwise suggested. Flow cytometry analysis revealed desialylated platelets from Asgr2-/-mice have a ~2-fold increase in Phosphatidylserine (PS) surface exposure when compared to WT platelets. In addition, western blot analysis showed increased expression of cleaved caspase 3 in Asgr2-/-platelets, but no changes in Bcl-xL protein expression between WT and Asgr2-/-platelets. Next, WT and Asgr2-/-mice received a single dose of the BH3 mimetic, ABT-737, which binds and inhibits pro-survivor proteins (Bcl-2, Bcl-xL and Bcl-w) inducing apoptosis in vivo. Approximately 2 hours after the injection of ABT-737, we observed a big drop on platelets counts in both WT (~42%) and Asgr2-/-(~59%) mice. Importantly, platelets from Asgr2-/-mouse were cleared more efficiently (~20%) from the circulation when compared to those in WT mice, consistent with the ~20% increment in platelet number observed in this mouse model and supporting the notion that the platelets that circulate longer in the Asgr2-/-mice are more sensitive to apoptotic events. To investigate if apoptosis could be triggering platelet desialylation, WT mice were treated with ABT-737 and 1hour later (time point before platelet count drop), platelets were collected and analyzed by flow cytometry. Interestingly, analysis of galactose exposure by RCA-I lectin showed no differences in desialylation between ABT-737 and PBS control groups. On the other hand, Phosphatidylserine (PS) exposure was significantly elevated on ABT-737 group, indicating that platelets were undergoing apoptosis without changing their sialylated status. To confirm our in vivodata, freshly isolated washed WT platelets were treated with ABT-737 to induce apoptosis or Neuraminidase (NA) to desialylated platelets. NA treatment induced platelet desialylation (increased RCA-I binding) in WT platelets, as expected, and interestingly triggered apoptosis, judge by increased PS exposure in both ABT-737 and NA treated groups. However, ABT-737 treatment wasn't able to induce desialylation as levels of RCA-I binding to platelets was the same when compared to PBS control platelets. Taken together, our data shows that desialylated platelets in circulation are prone to apoptosis. In addition, our findings strongly support the hypothesis that desialylation of platelet surface glycoproteins trigger the intrinsic apoptotic pathway in platelets in vivo. Disclosures No relevant conflicts of interest to declare.

The human body produces and removes 1011 platelets daily to maintain a normal steady state platelet count. Platelet production must be regulated to avoid spontaneous bleeding or arterial occlusion and organ damage. Multifaceted and complex mechanisms control platelet production and removal in physiological and pathological conditions. This review will focus on different mechanisms of platelet senescence and clearance with specific emphasis on the role of posttranslational modifications. It will also briefly address platelet transfusion and the role of glycans in the clearance of stored platelets.

Platelets are fundamentally important in normal hemostasis and pathological thrombosis (i.e. cardiovascular diseases, stroke, etc.). Platelets mediate the initial first-step in hemostasis through surface glycoproteins like the GPIb-IX-V complex and integrin αIIbβ3 (GPIIbIIIa). Although the functions of platelet surface glycoproteins are well known, the roles of posttranslational modifications on those surface glycoproteins are poorly understood. We have recently shown that sialic acid is a key regulator of platelet survival. As platelets circulate and age in blood, they lose sialic acid and are rapidly cleared by the hepatocytes where they stimulate liver TPO production and consequently regulate thrombopoiesis. Here, we investigated the importance of glycosylation to platelet function by measuring the impact of sialic acid content on platelet responses to thrombin activation. Freshly isolated wild-type washed platelets were treated with a2-3, -6, -8 sialidase (neuraminidase, NA) to remove sialic acid from the platelet surface glycoproteins or with a competitive NA inhibitor, 2-deoxy-2,3-dehydro-N-acetylneuraminic acid (DANA) to prevent sialic acid loss by the action of sialidases. After treatment with both NA and DANA, platelets were activated with Thrombin (THR, 0.1U/mL). As controls, aliquots of freshly isolated wild-type washed platelets were left untreated (Rest), only treated with neuraminidase (NA) or activated with Thrombin (THR). First, we measured b-galactose exposure using RCA-I lectin to test the efficacy of treatments. As expected, NA treated platelets showed significantly higher RCA-I binding when compared to Rest, THR and DANA treated platelets. Noteworthy, RCA-I binding to THR activated platelets was higher than Rest or DANA + THR platelets. We next investigated the effect of NA and DANA treatments of platelet degranulation. Thrombin activated platelets showed high level of P-selectin surface exposure when compared to Rest platelets. NA treatment alone caused low P-selectin exposure (~18% positive platelets) and NA + THR treated platelets showed high levels of P-selectin similar to THR only treatment. Interestingly, DANA +THR platelets showed a lower percentage of P-selectin positive platelets when compared to THR activation only (~50% compared to ~85%). In platelets, thrombin signaling is mediated by PARs, G-protein-coupled receptors that trigger several intracellular pathways, including phosphorylation of several proteins. We next investigated if glycan remodeling of surface glycoproteins could alter the intracellular signaling triggered by Thrombin. Our data shows that NA + THR platelets have increased phosphorylated Akt when compared to THR alone and pretreatment with DANA dampens the phosphorylation signal triggered by THR activation. These data suggest that the glycosylation status of surface glycoproteins on platelets regulates thrombin-induced activation. Neuraminidases are lysosome-resident enzymes, they act primarily intracellularly but can also be recruited to the cell surface. Studies have shown that Neu1, one of the neuraminidase isoforms, regulates lysosome exocytosis by desialylation of LAMP1. Flow cytometry analysis of LAMP1 surface expression showed that THR activation induced LAMP1 surface exposure when compared to Rest. NA treatment did not affect LAMP1 surface exposure caused by THR, but DANA treatment completely blocked LAMP1 translocation to the surface, suggesting that Neuraminidase is a regulator of lysosomal exocytosis in platelets. Taken together, our data shows that sialic acid is a potential regulator of platelet function. More studies are needed to identify platelet glycoproteins affected by sialic acid changes. Nonetheless, these data illustrate that glycan remodeling is ideally suited for therapeutic manipulation to prevent undesired platelet activation. Disclosures No relevant conflicts of interest to declare.

Bone marrow (BM) macrophages maintain both survival and retention of hematopoietic stem cells and regulate erythropoiesis. The role of macrophage lectins and glycans in thrombopoiesis remains unclear. We report a novel role for bone marrow macrophage galectin-3 in maintaining platelet counts, by phagocytosing megakaryocytes (MKs) expressing the Thomsen-Friedenreich (TF) antigen, which is often exposed under pathological conditions, such as cancer and malignancies. The TF antigen is a disaccharide presented in cryptic form on O-glycans and covered by a sialic acid moiety. The sialyltransferase ST3Gal1 transfers sialic acid onto the TF antigen. To investigate the role of O-glycans in thrombopoiesis, we generated mice with increased TF antigen in MKs by generating St3gal1loxP/PF4+ mice specifically lacking ST3Gal1 in the MK lineage. As expected, St3gal1loxP/PF4+ circulating platelets and BM MKs had increased TF antigen expression, compared to controls, as evidenced by peanut agglutinin (PNA) binding. Other blood cell lineages had no increase in TF antigen expression. St3gal1loxP/PF4+ mice developed mild thrombocytopenia, but surprisingly had virtually normal platelet clearance. BM MK colony forming units and in vitro proplatelet production were normal in St3gal1loxP/PF4+ mice, suggesting that extrinsic factors in the St3gal1loxP/PF4+BM environment affected platelet production. St3gal1loxP/PF4+ BM smears revealed increased hemophagocytosis, indicative of an increase in phagocytic macrophages. In vivo macrophage ablation by injection of clodronate-encapsulated liposomes significantly reduced the numbers of activated macrophages, thereby normalizing blood platelet counts and size. Flow cytometric phenotypic analysis of BM-derived macrophages showed an increased population of activated macrophages in St3gal1loxP/PF4+ mice, compared to controls, specifically macrophages with increased galectin-3 expression, a ligand for the TF antigen. Immunofluorescence staining of BM sections using a specific antibody towards the TF antigen showed that MK progenitors and pro-platelet-like structures expressed TF antigen in control BMs, which is significantly increased in St3gal1loxP/PF4+ mice and co-localized with galectin-3 expressing macrophages, supporting the notion that MK O-glycans and macrophage galectin-3 play a role in thrombopoiesis under steady state and pathological conditions. Consistent with this notion, galectin-3 deficient mice have slightly, but significantly increased blood platelet counts. We conclude that galactin-3 plays a minor role in normal thrombopoiesis. Activation of galectin-3 expressing macrophages by the MK TF antigen leads to MK phagocytosis, inhibition of platelet formation and thrombocytopenia. Disclosures No relevant conflicts of interest to declare.

The human body produces and removes 1011 platelets daily to maintain a normal steady-state platelet count. However, the regulatory mechanisms remain elusive. We have shown that platelets lacking sialic acid (desialylated platelets) are removed by the hepatic Ashwell-Morell receptor (AMR or asialoglycoprotein receptor type 2), thereby regulating platelet survival and hepatic TPO levels. Platelet counts and lifetime were increased in Asgr2-/- mice (AMR-null mice), compared to wild type (WT) mice. Platelet volume and immature platelet fraction (IPF) are decreased in AMR-null mice, consistent with the notion that platelets in AMR-null mice (AMR-null platelets) circulate longer and are older. By contrast, deficiency of the sialyltransferase St3gal4 gene induces a marked thrombocytopenia (St3gal4-null platelets), due to rapid platelet clearance by the hepatic AMR. Consistent with the rapid platelet clearance, platelet volume and IPF were increased in St3gal4-/- mice, reflecting high platelet turnover and younger platelets. While both AMR-null and St3gal4-null platelets are desialylated, they differ substantially in their time, i.e. age, in circulation. Here we investigated the effects of in vivo and in vitro aging on platelet function. Freshly isolated St3gal4-null platelets showed significantly increased integrin activation when stimulated with convulxin and thrombin, while AMR-null platelets showed a significantly lower response compared to WT platelets. Secretion of α-granule was significantly increased in St3gal4-null platelets. By contrast no significant difference was measured between WT and AMR-null platelets. Despite increased platelet counts, the tail-bleeding time was significantly prolonged in AMR-null mice, compared to WT mice, suggesting that increased circulatory time (age) negatively affects platelet function in vivo. We next performed in vitro storage for up to 72 hours at room temperature to stress platelet aging. Stored St3gal4-deficient platelets had increased integrin activation, α-granule secretion in response to convulxin and thrombin and showed increased phosphatidyl serine exposure as detected by Annexin V binding in response to calcium ionophore, compared to stored control platelets. By contrast, stored AMR-null platelets had significantly impaired integrin activation, α-granule secretion and Annexin V binding compared to controls. To further evaluate the propensity to undergo apoptosis, we tested caspase-3 activation and mitochondrial membrane potential. Surprisingly, we found that both St3gal4-null (young) and AMR-null (old) platelets showed a significantly lower caspase-3 activation in response to calcium ionophore and ABT-737 compared to WT platelets. Furthermore, the mitochondrial membrane potential was lower in both St3gal4-deficient and AMR-null platelets compared to WT platelets, indicating functionally impaired mitochondria. Taken together, our data indicate that younger St3gal4-null platelets have an increased baseline function while by contrast older AMR-null platelets have decreased function in vitro and in vivo. Interestingly, younger and older (longer circulating platelets) had reduced propensity to undergo apoptosis and impaired mitochondrial function. Overall, younger platelets represent a highly favorable profile during storage. Identification of donors with a larger fraction of younger platelets could result in safer and more efficacious platelet transfusions. Disclosures No relevant conflicts of interest to declare.

Glycosylation defects have been associated with low platelet counts. Six genes encoding sialyltransferases (ST), ST3gal1 to 6, that synthesize an α2,3 sialic acid (SA) linkage have been identified in the mammalian genome, and deletion of St3gal1 and St3gal4 genes has been associated with macrothrombocytopenia in mice. Despite the similarity in transferring SA in a α2,3-linkage to terminal galactose residues, St3gal1 and St3gal4 sialylate distinct glycans: St3gal1 is associated with core 1 O-glycan Galβ1,3GalNAcα1-Ser/Thr expression, also known as tumor-associated or Thomsen-Friedenreich antigen (T-antigen), whereas St3gal4 sialylates lactosaminyl Galβ1,4GlcNAc N-glycans. It has been previously shown that St3gal4-null platelets are cleared by the hepatic Ashwell-Morell receptor, causing severe thrombocytopenia in these mice. Herein, we generated St3gal1loxP/PF4+ mice specifically lacking ST3Gal1 in the megakaryocyte (MK) lineage to investigate the detailed mechanisms of macrothrombocytopenia associated with St3gal1 deficiency. Both St3gal1loxP/PF4+ circulating platelets and bone marrow (BM) MKs had increased T-antigen expression, compared to control, as evidenced by peanut agglutinin (PNA) binding. As expected, other blood cell lineages had no increase in T-antigen expression. Blood platelet counts were reduced by ~50% and platelets were enlarged in St3gal1loxP/PF4+ mice, compared to control, despite a virtually indistinguishable platelet clearance. BM MK numbers were normal despite the observed thrombocytopenia, BM MK colony forming units (CFUs) were reduced and in vitro proplatelet production was normal in St3gal1loxP/PF4+ mice, suggesting that extrinsic factors in the St3gal1loxP/PF4+ BM environment affected platelet production. We hypothesize that recognition of the T-antigen epitope on MKs mediate phagocytosis by macrophages. Macrophages in St3gal1loxP/PF4+ mice had increased expression of CD68 (macrosialin), indicative of an activated macrophage state. Flow cytometric analysis of BM derived macrophages of St3gal1loxP/PF4+ mice showed an increased population of resolving M2-type macrophages, which are normally involved in apoptotic cell clearance. Additionally, St3gal1loxP/PF4+ BM smears revealed increased hemophagocytosis, as evidenced by May-Grunwald/Giemsa, indicative of an unspecific increase in phagocytic macrophages. Macrophage ablation by in vivo injection of clodronate-encapsulated liposomes significantly reduced the numbers of activated macrophages in St3gal1loxP/PF4+ mice, thereby normalizing blood platelet counts and size. Taken together data show the contrasting effects of different SA loss on platelet homeostasis: Platelets lacking α2,3-linked SA on N-glycans have increased platelet clearance, whereas a lack of α2,3-linked on O-glycans do not affect platelet half-life, but cause defective thrombopoiesis in MKs. Disclosures No relevant conflicts of interest to declare.

Dynamins are large and highly conserved GTPases involved in endocytosis and vesicle trafficking. Mutations K562E/del in the ubiquitous dynamin 2 (DNM2) have been associated with thrombocytopenia in humans. To determine the role of DNM2 in megakaryopoiesis we generated Dnm2fl/fl Pf4-Cre mice specifically lacking DNM2 in the megakaryocyte (MK) lineage. Dnm2fl/fl Pf4-Cre mice were viable, but had severe macrothrombocytopenia with moderately accelerated platelet clearance and prolonged bleeding due to poorly functional platelets. Dnm2-null bone marrow MKs had altered demarcation membrane system, appearing at times as a compact, narrow twisting membrane system of clathrin-coated vesicles. Fetal liver cell derived Dnm2-null MKs formed proplatelets poorly in vitro, showing that DNM2 plays a major role in MK membrane formation and thrombopoiesis. Both endogenous DNM2 and overexpressed DNM2 WT, but not DNM2 K562E/del mutants localized with the early endosome in bone marrow MKs. The endocytic pathway was disrupted in Dnm2-null MKs, as evidenced by severely reduced early endosome EEA1 and APPL1 staining and weak LysoTracker internalization. Endocytosis of the thrombopoietin (TPO) receptor Mpl was impaired in Dnm2-null platelets, causing constitutive phosphorylation of the tyrosine kinase JAK2 and elevated circulating TPO levels. MK-specific DNM2 deletion severely disrupted bone marrow homeostasis, as reflected by massive MK hyperplasia and myelofibrosis, and consequent extramedullary hematopoiesis and splenomegaly. However, additional Mpl genetic deletion failed to rescue the severe splenomegaly of Dnm2fl/fl Pf4-Cre mice, and Mpl-/- Dnm2fl/fl Pf4-Cre mice instead died at 4-5 weeks of age. Taken together, our data demonstrates that unrestrained MK growth and proliferation results in rapid myelofibrosis independently of Mpl expression and other bone marrow cell types, and establishes a previously unrecognized role for DNM2-dependent endocytosis in megakaryopoiesis, thrombopoiesis and bone marrow homeostasis. Disclosures No relevant conflicts of interest to declare.

Congenital amegakaryocytic thrombocytopenia (CAMT) is a disorder caused by c-mpl mutations that block or severely reduce thrombopoietin (Tpo) signaling, leading to amegakaryocytic thrombocytopenia and ultimately aplastic anemia. In neonates and infants, CAMT can have different clinical manifestations than in adults, and cases have been described of newborn infants who presented with severe thrombocytopenia but normal numbers of immature-appearing megakaryocytes (MKs) in the bone marrow. To elucidate the potential mechanisms underlying the developmental differences in the manifestations of Tpo deficiency, we evaluated MKs and platelets in newborn c-mpl-/- mice on postnatal day 1 (P1). Since the liver is the main site of megakaryopoiesis at that developmental stage (Liu et al., Blood 2014), liver MKs in P1 mice were immunohistochemically stained for VWF, quantified, measured, and ultrastructurally evaluated using transmission electron microscopy (TEM), as previously described. Histological analysis showed that MKs in newborn c-mpl-/- mice were present at a concentration of 30% and were significantly smaller compared to WT pups, a finding previously reported in adult mice. Unlike in adult c-mpl-/- mice, however, which have ultrastructurally normal MKs, TEM analysis of neonatal c-mpl-/- liver MKs revealed a significant shift toward immature cells and maturational abnormalities in 68% of MKs, including a disorganized demarcation system, reduced granules, and an abnormally wide peripheral zone. Ultrastructural analysis of platelets from c-mpl-/- P5 mice also revealed a larger open canalicular system and more glycosomes compared to age-matched WT platelets. These findings were not associated with increased levels of baseline P-selectin expression, suggesting that they were not the result of activation during processing. Next, we serially evaluated platelet counts in WT and c-mpl-/- mice during the transition from fetal/neonatal to adult megakaryopoiesis. At P2, c-mpl-/- platelet counts were 170±54x103/µl, representing 26% of those in P2 WT mice (648±137x103/µl). Over the first two weeks of life, WT platelet counts doubled and reached nearly adult levels, as previously described. In contrast, c-mpl-/- platelet counts failed to increase, and remained stably low at P14. Since we previously showed that the rise in platelet counts during this period was associated with a one-day prolongation of the neonatal platelet lifespan (Liu et al., Blood 2014), we evaluated platelet survival in c-mpl-/- compared to WT mice, using in vivo biotinylation. These studies suggested that the platelet lifespan was significantly shorter in c-mpl-/- compared to WT newborn (P2) mice. In contrast, we found no differences in platelet lifespan between c-mpl-/- and WT adult mice. Taken together, our findings suggest that, in fetal and neonatal life, Tpo signaling is required for normal MK proliferation, maturation, platelet formation, and platelet survival. This is in contrast to adult life, when Tpo deficiency leads only to reduced MK and platelet counts. We hypothesize that these differences reflect the lack of Tpo-independent pathways of MK maturation in the environment of the fetal/neonatal liver, and have implications in the developmental stage-specific clinical manifestations of CAMT. Disclosures No relevant conflicts of interest to declare.

Changes in glycans expression have been associated with defects in blood platelet counts. However, the role of posttranslational modifications on platelet production is poorly understood. Six genes encoding sialyltransferases (ST)3Gal-I to -VI that form a2-3 sialic acid linkage have been identified in the mammalian genome, and deletion of St3gal1 and St3gal4 genes has been associated with macrothrombocytopenia in mice. We and others have shown previously that St3gal4-null platelets are cleared by the hepatic Ashwell-Morell receptor. Loss of ST3Gal-I activity has been associated with core 1 O-glycan Galβ1-3GalNAcα1-Ser/Thr expression, also known as tumor-associated or Thomsen-Friedenreich antigen (T antigen). We here investigated the detailed mechanisms of macrothrombocytopenia associated with St3gal1 deficiency by generating St3gal1loxP/PF4+ mice that lack ST3Gal-I specifically in the megakaryocyte (MK) lineage. Blood platelet counts were reduced by ~50% in St3gal1loxP/PF4+ mice, compared to control mice. Other blood cell counts were normal in St3gal1loxP/PF4+ mice. The clearance rate of St3gal1-null platelets was increased by ~15%, as determined by in vivo platelet biotinylation. Bone marrow MK numbers were normal in St3gal1loxP/PF4+ mice, compared to control mice, indicating that mechanisms other than clearance regulate circulating platelet counts in St3gal1loxP/PF4+ mice. Both St3gal1loxP/PF4+ platelets and bone marrow MKs had increased T antigen expression, as evidenced by flow cytometry using peanut agglutinin (PNA) binding. St3gal1loxP/PF4+ mice had increased bone marrow macrophage numbers, as evidenced by immunohistochemistry and flow cytometry using the macrophage marker F4/80. Macrophages in St3gal1loxP/PF4+ mice had increased expression of CD68 (macrosialin), as determined by immunohistochemistry and flow cytometry, indicative of an activated macrophage state. Consistently, St3gal1loxP/PF4+ bone marrow smears stained with May-Grunwald/Giemsa revealed increased hemophagocytosis. Macrophage ablation by in vivo injection of clodronate-encapsulated liposomes normalized blood platelet counts and size, and significantly reduced the numbers of activated macrophages in St3gal1loxP/PF4+ mice. Together, our data indicates that platelet production in the bone marrow is reliant on correct glycosylation on MK surface proteins and that the intimate interaction between MKs and macrophages play an important role in regulating platelet production and bone marrow homeostasis. Disclosures No relevant conflicts of interest to declare.

Background:Immune thrombocytopenia (ITP) is a common bleeding disorder caused primarily by autoantibodies against platelet GPIIbIIIa (70-80%) and/or GPIb-complex (20-40%). Current theory suggests antibody-mediated platelet destruction occurs in the spleen, via macrophages through Fc-FcγR interactions. However, evidence from us and others demonstrated that anti-GPIbα, but not anti-GPIIbIIIa, can induce thrombocytopenia via an Fc-independent pathway, which is resistant to intravenous IgG (IVIG) therapy in murine ITP models (Blood 2006) and subsequent IVIG studies in human ITP patients, including our recent large patient cohort study (JTH 2014). This suggests that binding of anti-GPIbα antibodies may induce platelet clearance through a presently unidentified mechanism different than that of anti-GPIIbIIIa. Methods: We developed unique mouse anti-mouse monoclonal antibodies (mAbs) in GPIIIa-/- or GPIba-/- mice, which also recognize GPIbα and GPIIbIIIa of different species including human. Flow cytometry, immunofluorescence, and western blotting were used to evaluate whether these mAbs induced platelet activation, neuraminidase-1 translocation and desialylation of the heavily glycosylated GPIbα in the presence of sialidase inhibitor N-acetyl-2,3-dehydro-2-deoxy neuraminic acid (DANA). These experiments were repeated with human platelets and human ITP patient plasma. We further investigated the effects of anti-GPIbα antibodies on platelet activation, desialylation and clearance in vivo; BALB/c mice were injected with anti-GPIbα or anti-GPIIbIIIa mAbs and following, platelet activation and desialylation were measured by flow cytometry. Hepatocytic Ashwell-Morell receptor (AMR) mediated anti-GPIbα platelet clearance in the liver was examined using immunohistochemistry or blocking the AMR with asialofetuin in both wild-type and macrophage depleted mice. Therapeutic administration of DANA in a murine ITP model assessed the significance of Fc-independent anti-GPIbα mediated platelet clearance in ITP. Results and Discussion: We found that anti-GPIbα, but not anti-GPIIbIIIa antibodies, induced significant P-selectin expression, JON/A binding, neuraminidase-1 translocation and desialylation in murine platelets. Interestingly, certain human platelets were activated (P-selectin expression) and desialylated in the presence of both anti-GPIbα and anti-GPIIbIIIa mAbs or ITP patient plasma. However, we demonstrate that the anti-GPIIbIIIa antibody mediated platelet effects are dependent on the FcγRIIa present exclusively on human platelets as FcγRII blocker IV.3 completely attenuated the response. In contrast, IV.3 had little effect on anti-GPIbα mediated platelet activation or desialylation. Anti-GPIbα Fab fragments and platelet signal pathway inhibitors demonstrate that anti-GPIbα mediated platelet activation and desialylation are consequences of GPIbα cross linking and are reinforced by a positive feedback loop. In vivo, we found significant increases in P-selectin and desialylation in anti-GPIbα injected mice, independent of IgG subclass. A significant role for the hepatic AMR in the clearance of deglycosylated platelets was observed; particularly in macrophage depleted mice whereby, although anti-GPIIbIIIa mediated platelet clearance was completely attenuated, anti-GPIbα mediated platelets clearance still occurred, but was completely rescued with asialofetuin. Immunohistochemistry revealed significant co-localization of anti-GPIbα opsonized platelets with AMR. These suggest the AMR is the dominant Fc-independent anti-GPIbα mediated platelet clearance pathway in the absence of macrophages. Remarkably, sialidase inhibitor DANA ameliorated anti-GPIbα mediated thrombocytopenia in mice. Thus, we demonstrate for the first time that anti-GPIbα antibodies induce platelet activation leading to GPIbα desialyation and platelet clearance via a novel Fc-independent pathway: the hepatic AMR. Our data also suggested that some anti-GPIIbIIIa autoantibodies in human patients may also induce platelet activation and desialylation via the platelet FcR signaling pathway. These findings may lead to novel therapeutic regimens including designating desialylation as a potential diagnostic biomarker and therapeutic target in the treatment of both anti-GPIIbIIIa and anti-GPIbα mediated and/or refractory ITP. Disclosures No relevant conflicts of interest to declare.

The human body produces and removes 1011 platelets daily to maintain a normal steady-state platelet count, and the level of production can be greatly increased under conditions of platelet destruction. Thrombopoietin (TPO) is the primary regulator of platelet production, supporting the survival, proliferation and differentiation of platelet precursors, bone marrow megakaryocytes. Hepatocytes are a major source of production and secretion of circulating TPO. However, mechanisms regulating circulating TPO levels have been debated for decades. Here, we provide experimental evidence that platelets lacking sialic acid (desialylated platelets) are removed by the hepatic Ashwell-Morell receptor (AMR or asialoglycoprotein receptor), thereby regulating platelet survival and hepatic TPO levels. These conclusions are based on the following evidence: 1) Mice lacking the AMR Asgr2 subunit had increased platelet survival, compared to wild type (WT) mice. Platelets from Asgr2-null mice showed increased loss of sialic acid, as evidenced by flow cytometry using the galactose specific lectins RCAI and ECL, showing that removal of desialylated platelets by the AMR regulates in vivo platelet survival. 2) Livers isolated from Asgr2-null mice had TPO mRNA levels decreased by 40%, compared to WT mice. In contrast, liver TPO mRNA levels were increased by 30% in St3gal4-null mice lacking the sialyltransferase ST3GalIV, where desialylated platelet clearance is increased and specifically mediated by the AMR. Both plasma TPO levels and platelet TPO contents were similarly altered in both mutant mice. Thus, desialylated platelet uptake by the AMR regulated liver TPO levels. 3) Desialylated platelets isolated from St3gal4-null or Asgr2-null mice infused into WT mice increased hepatic TPO mRNA levels as early as 12h post-infusion. Plasma TPO concentrations and bone marrow megakaryocyte numbers increased in parallel with TPO mRNA levels, peaking by day 2 post-infusion, followed by new platelet release at day 10 post-infusion. In contrast, desialylated platelets infused into Asgr2-null mice had no effect on TPO mRNA synthesis, TPO plasma levels and bone marrow megakaryocyte numbers. 4) Incubation of human hepatoma cell line, HepG2 cells, with human desialylated platelets by sialidase treatment resulted in TPO mRNA expression increase by 2.2 and 2.9-fold after 4 and 6h, respectively, followed by significant increase in TPO secretion. 5) The signaling pathways activated by uptake of desialylated platelets by the AMR to induce TPO mRNA transcription were investigated in vivo and in vitro. Major polypeptides of 60-70 and 125 kDa were highly tyrosine phosphorylated in WT liver cells, as evidenced by SDS-PAGE. Using a specific antibody directed against JAK2, we identified the 125-kDa phosphoprotein as the tyrosine kinase JAK2 in mouse liver cells and human HepG2 cells. Analysis of liver samples revealed a marked reduction in JAK2 phosphorylation in Asgr2-null mice and significant increase in St3gal4-null mice. 6) The JAK1/2 inhibitor AZD1480 significantly decreased phosphorylation of JAK2, phosphorylation and translocation to the nucleus of the acute phase response transcription factor STAT3, TPO mRNA expression and TPO secretion in HepG2 cells incubated with desialylated platelets. In vivo treatment of WT mice with AZD1480 blocked TPO mRNA increase promoted by injection of endogenously desialylated platelets. Therefore we conclude that platelets desialylate as they circulate, thereby becoming the primary AMR ligand and providing a novel physiological feedback mechanism to regulate plasma TPO levels and platelet production in vivo and in vitro. Disclosures No relevant conflicts of interest to declare.

Platelets have the shortest shelf-life of all major blood components and are the most difficult to store complicating platelet transfusion practices. Transfused fresh radiolabeled autologous platelets differ significantly in recovery and survival among healthy subjects, however the cause of the inter-individual differences remains unclear. We demonstrated that the loss of sialic acid from the surfaces of cold-stored and transfused platelets promotes their clearance by Ashwell Morell receptors. The loss of platelet surface sialic acid correlates with increases in surface sialidase activity during platelet storage. Here we investigated whether fresh platelets from individual donors exhibit differences in surface sialidase expression and glycan exposure and sialic acid content changes with storage. Methods Platelets were isolated by standard methods from the venous blood of healthy volunteers or from standard platelet concentrates (PCs) and analyzed by flow cytometry for surface β-galactose using FITC-conjugated E. cristagalli lectin (ECL). Platelet surface sialidase expression was measured by flow cytometry using antibodies to sialidases Neu1 and Neu3. Sialidase activities were assayed using standard methods, Platelet uptake by hepatocytes was measured by using the human hepatoma cell line HepG2. To further elucidate these issues in a structural biology context we performed baseline study of the N- and O-linked glycans and glycosphingolipids (GSLs) in platelets, and any structural changes observed during storage, by employing HPLC, LC-MS/(MS), and sequential mass spectrometry (MSn) approaches. Results We found that terminal galactose on freshly-isolated platelet glycoproteins varies considerably among healthy subjects: Seven of ten individuals had low levels of exposed galactose (15.3 ± 4.1, MFI) and three subjects exhibited significantly higher levels of terminal galactose as detected by flow cytometry using lectins. Reduced sialic acid content correlated with increased surface sialidase activity and expression. Platelets with high terminal galactose were ingested with a higher rate by HepG2 cells, i.e via Ashwell Morell receptors. Importantly, individuals with low sialic acid levels correlate with low platelet counts at steady state. Structural analysis revlealed that fresh platelet N-glycan pools include a significant amount of high-mannose (Man5-Man9) and asialo complex glycans, however, are dominated by a diverse range of complex sialylated structures with two to four antennae, up to four NeuAcs, and include antennary fucosylation, and five or more lactosamine extensions. The O-linked fractions are comprised of core-1 and core-2 glycans having zero, one, or two NeuAc residues. A significant decrease in sialylation during conventional platelet storage at room temperature was confirmed at the level of individual O-glycan structures. Quantitative analysis of the more structurally complex N-glycan pools is ongoing. Conclusion Our results show that fresh platelets from healthy individuals vary in surface sialidase activity and sialic acid content and exhibit a high complexity in glycan structures. Collectively we propose that individual platelet counts may be dependent on surface sialic acid content and that the surface sialic acid could represent a factor that affects the recovery and survival of transfused platelets. Disclosures: No relevant conflicts of interest to declare.

It becomes increasingly apparent that, besides the intrinsic apoptotic machinery, surface glycan modifications regulate platelet survival. Platelets with reduced α2,3-linked sialic acid during sepsis due to S. pneumoniae infection, after cold storage, or in mice lacking the sialyltransferase ST3GalIV are cleared by the hepatic Ashwell-Morell receptor (AMR, ASGPR1/2). Platelet survival in Asgr2-/- mice was increased by ∼35% when compared to that of WT mice, which results in a ∼50% increase in circulating platelet counts, despite a loss of surface sialic acid. We reasoned that sialidase activity increases on the surface of circulating platelets as they age, a process that would facilitate sialic acid hydrolysis and removal from the circulation. To test this hypothesis, we directly injected the sialidase inhibitor 2-deoxy-2,3-dehydro-N-acetylneuraminic acid (DANA) into WT mice and determined endogenous platelet circulatory times. Platelet survival was prolonged by ∼30% (T1/2 of 62.0 ± 2.7 h) in DANA-treated mice, compared to that of mock-treated mice (T1/2 of 47.5 ± 4.3 h). DANA injections decreased terminal sialic acid loss on circulating platelets by ∼40% by day 2, compared to control platelets, as evidenced by binding of RCA-I lectin that specifically recognizes terminal β1-4 galactose moieties exposed by sialic acid removal. Freshly isolated, resting platelets from Asgr2-/- mice (AMR-platelets) were significantly smaller in size (22%) and had increased sialidase Neu1 (∼5 fold), but not Neu3 surface expression, when compared to WT platelets or St3gal4-/- platelets, as measured by flow cytometry. We next investigated if AMR-platelets age/deteriorate faster upon in vitro storage. Platelets were isolated from WT, Asgr2-/- and St3gal4-/- mice and stored for 24hrs at room temperature, and sialidase expression (Neu1 and Neu3) as well as microvesiculation were measured by flow cytometry. Although significant Neu1 and Neu3 surface expression increase was measured on platelets from all phenotype after storage, Neu1 and Neu3 surface expression was significantly higher in AMR-platelets (∼2 and 4 fold, respectively) when compared to WT and St3gal4-/- platelets. AMR-platelets, but not St3gal4-/- platelets microvesiculated upon storage, consistent with a faster deterioration of aged AMR-platelets. We next injected into WT and Asgr2-/- mice the BH3 mimetic, ABT-737, which binds and inhibits the pro-apoptotic Bcl-2, Bcl-xL and Bcl-w. After injection of ABT-737, platelets in the Asgr2-/- mouse were cleared more efficiently (∼20%) from the circulation when compared to those in WT mice. Collectively, our data show that blood borne sialidases contribute to loss of sialic acid during circulation to regulate platelet survival. Our data also suggest that platelet glycan degradation, i.e. sialic acid loss, may trigger the intrinsic apoptotic machinery in platelets, linking glycan degradation and intrinsic apoptotic machinery in the clearance mechanisms regulating platelet survival. Disclosures: No relevant conflicts of interest to declare.

Abstract 2170 The human body produces and removes 1011 platelets daily to maintain a normal steady-state platelet count, and the level of production can be greatly increased under conditions of platelet destruction. Here, we provide the experimental evidence that platelets with impaired Siaa2–3Galb1–4GlcNAc (LacNAc) structures are removed by the hepatic asialoglycoprotein receptor Asgr1/2, indicating that survival of platelets is intimately tied to surface glycans. Mice lacking Asgr2 subunit that is necessary to assemble a functional receptor have increased platelet survival (t1/2 = 49.5 ± 2h) compared to wild type mice (t1/2 = 31 ± 4h). Surprisingly, platelets from Asgr2-null mice have diminished surface sialic acid, as evidenced by lectins that bind exposed Gal moieties. Hence, desialylated platelets circulate in Asgr2-null mice. Besides “terminating” platelet circulation, the liver is the main source of thrombopoietin (TPO), the major hormone regulating platelet production. We hypothesized that desialylated platelet uptake by hepatic Asgr1/2 would affects TPO mRNA synthesis and have found that liver tissue from Asgr2-null mice has a 40% decrease in TPO mRNA levels compared to liver tissue from WT mice. In contrast, ST3Gal4-null mice, which have high rates of platelet turnover and increased desialylated platelet uptake by the Asgr2, have a 30% increase in TPO mRNA content in their livers. Both plasma TPO levels and platelet TPO contents are similarly altered in both mutant mice. In contrast, and in agreement with published data, antibody-mediated platelet clearance did not affect hepatic TPO mRNA levels. Taken together, these data show that the clearance of desialylated platelets by the hepatic Asgr1/2 regulates TPO homeostasis in vivo. Disclosures: No relevant conflicts of interest to declare.

Abstract 3257 The regulatory mechanisms of platelet homeostasis remain elusive. We investigated here the role of hepatic asialoglycoprotein receptor (a.k.a. Ashwell-Morell receptor) in platelet clearance. Mice lacking the hepatic asialoglycoprotein receptor Asgpr2 subunit had increased platelet survivals (T1/2 = 49.5±2h) when compared to wild type (WT, T1/2 = 31±4h) mice. Consequently, Asgpr2−/− mice had platelet counts increased by ∼20%, compared to WT, with increased terminal galactose exposure, as demonstrated using the galactose specific lectin RCA1. Bone marrow and spleen megakaryocyte numbers were reduced by ∼15% and ∼20% in Asgpr2−/− mice, compared to WT mice. Sialidase (NA, Clostidium perfringens, 50mU/mice) maximally desialylated circulating platelets when injected intravenously, as evidenced by increased RCA1 binding. Sialidase injection resulted in a ∼60% depletion of circulating platelets after 24h in Asgpr2−/− mice, compared to >90% in WT mice, indicating that desialylated platelets were partially removed by Asgpr1/2. In contrast to platelets, red blood cell counts were unaffected by sialidase treatment. Sialidase injection for 72h resulted in a 2.3-fold and 1.2-fold increase in megakaryocyte numbers in the spleen and bone marrow of WT mice, respectively, but not in Asgpr2−/− mice. In contrast to sialidase treatment, injections of rabbit anti-mouse platelet serum (RAMPS) depleted >95% of circulating platelets and increased by 70% bone marrow, but not spleen MK numbers in both WT and Asgpr2−/− mice. The data shows that removal of desialylated, i.e, senescent, platelets by the hepatic Ashwell-Morell receptor differs to that of antibody-mediated platelet clearance. Disclosures: No relevant conflicts of interest to declare.

Abstract 2025 The human body produces and removes more than a 100 billion of platelets every day. The mechanisms responsible for platelet homeostasis are subject to speculation since the 1950's. The most popular hypothesis to date has been antibody-mediated clearance, platelet consumption due to massive blood loss and an internal “senescence timer”. We and others have recently demonstrated that sialic acid deficient platelets due to external triggers such as sepsis or chilling are cleared by hepatic asialoglycoprotein receptors (ASGPR) independently of macrophages. Here, we investigated whether loss of sialic acid mediates platelet clearance in vivo. We show that 1) Injection of the specific sialidase inhibitor 2-deoxy-2,3-dehydro-N-acetylneuraminic acid (DANA) lengthened the survival of biotinylated platelets by ∼50% (T1/2 of 72h), compared to mock treated (PBS injected) control mice (T1/2 of 49h); 2) Similarly, biotinylated platelet survival in ASGPR-null mice was prolonged by ∼ 50% (T1/2 of 74h) compared to platelet survival in wild type (WT) mice (T1/2 of 48h); 3) ASGPR-null mice have significantly increased platelet counts, compared to WT (p=0.0004) and platelets isolated from ASGPR-null mice are ∼15% smaller than WT (p=0.03); 4) Platelets isolated from ASGPR-null mice showed significant increased in b-galactose exposure (∼50% increase, i.e. decrease of sialic acid), compared to WT, as evidenced by binding of the b-galactose specific lectin (RCA-I). These data show that the ASGPR not only removes desialylated platelets due to sepsis or chilling, but also regulates platelet homeostasis. Disclosures: No relevant conflicts of interest to declare.