Specific Gtpase Dynamin Isoforms Regulate Megakaryocyte Membrane Remodeling and The Formation Of Multivesicular Bodies and Microparticles

Abstract

Cdc42 interacting protein 4 (CIP4) is a membrane-associated BAR protein, which also forms a complex via its SH3 domain with the dynamins (DNMs) and Wiskott-Aldrich Syndrome (WAS) protein. Thus, CIP4 remodels the plasma membrane and cortical actin cytoskeleton. To determine its physiological function, we generated CIP4-null mice. They displayed thrombocytopenia similar to that of WAS-null mice and have abnormal megakaryocytes (MKs) with decreased proplatelet formation and underdeveloped demarcation membrane system (DMS) (Chen et al, Blood 2013). The DMS is an extensive network of membrane tubules which serves as a membrane reservoir for proplatelet formation. The membranes are enriched for polyphosphoinositides that are docking sites for BAR proteins and for pleckstrin homology domain-containing proteins such as the dynamins. Still, the formation of the DMS is poorly understood. Dynamins are cell vesicle trafficking proteins that possess a GTPase domain. They induce neck vesicle constriction and scission from the plasma membrane. When the GTPase activity is abrogated, vesicle scission does not occur; instead, the plasma membrane invagination induced by the BAR proteins results in deep plasma membrane tubulations. Of the three dynamin isoforms, DNM3 participates in MK development including DMS formation (Reems et al, Exp Hematol 2008; Wang et al, Stem Cells Dev 2011). Moreover, a recent genome-wide association study suggested that an MK-specific DNM3 isoform might play a role in human platelet size determination (Nürnberg et al, Blood 2012). However the exact mechanism for dynamin’s participation in DMS formation is unclear. A double knockout for dynamin 1 and dynamin 3 in neurons causes accumulation of long invaginations from the plasma membrane (Ferguson and De Camilli, Nat Rev Mol Cell Biol 2012). We initially hypothesized that CIP4’s association with DNM3 contributes to the DMS development during platelet biogenesis and wanted to test for functional redundancy with other dynamins present in MKs and platelets. To determine if CIP4 interacts with dynamin in the MK lineage, we found that following either phorbol ester (PMA) or fibronectin stimulation in the human MK cell line CHRF-288, CIP4 co-precipitated with DNM3 and colocalized by confocal microscopy. To determine dynamin’s effect on membrane biophysical properties, we measured the fluorescence anisotropy, which reflects the disorder of membrane lipids due to movement and indicate membrane rigidity. Compared with controls in CHRF-288 cells, shRNA-mediated knockdown (KD) of DNM2 or DNM3 resulted in higher membrane rigidity in response to PMA. The strongest effect was seen in double KD cells with decreased fluidity by 2.6 ± 0.3%, which is similar to what was observed with CIP4 KD and is physiologically significant (Chen et al Blood 2013). KD of DNM2 resulted in aberrant morphology, greater cell diameter, and electron microscopy (EM) showed formation of new multivesicular bodies (MVBs) which are sorting compartments during α- and dense granules formation. Single DNM3 KD cells had no observable phenotype. EM imaging of DNM2 and DNM3 double KD cells revealed plasma membrane tubulation that resembles the DMS. While control CHRF-288 cells, with high DNM3 protein expression, do not have a DMS at baseline, MK cell lines Meg-01 and L8057, with respectively lower or no dynamin-3 protein expression, both have a DMS (Battinelli et al PNAS 2001; Ishida Y et al, Exp Hematol 1993). Platelet microparticles (MPs) are known to mediate a prothrombotic state in patients. Having previously found that CIP4-null mice show reduced levels of platelet MPs, we measured MPs in dynamin knockdown cell supernatant by flow cytometry and CD41/Annexin V staining. Surprisingly, we found that microparticle levels were increased 2.9-fold in DNM2 KD cells and 3.8-fold in double DNM2 and DNM3 KD cells. Our findings suggest that: 1) there is only partial functional redundancy between DNM2 and DNM3 in platelet biogenesis, 2) DNM2 controls MVB formation and MP release in MK cells, and 3) the CIP4-dynamin pathway contributes to DMS formation. It is possible that CIP4’s interaction with dynamins restrains their spatial and temporal activity to allow for long invaginations to accumulate in the DMS. Dynamin depletion might also increase surface membrane availability for MP formation. Dynamins are thus potential targets to modulate thrombotic state and platelet biogenesis. Disclosures: No relevant conflicts of interest to declare.