Loss of the F-BAR protein CIP4 reduces platelet production by impairing membrane-cytoskeleton remodeling

Megakaryocytes generate platelets through a highly coordinated process that requires membrane and cytoskeletal reorganization.¹  Important components for cytoskeletal reorganization are the RhoGTPase Cdc42, actin nucleator Wiskott-Aldrich Syndrome Protein (WASP), and actin-associated Arp2/3 complex.²  Wiskott-Aldrich Syndrome is characterized by microthrombocytopenia, the mechanism of which is only partially known and may include both autoimmunity³  and dysregulated platelet production.^4,5  The molecular machinery required for membrane remodeling in megakaryocytes that generate proplatelet protrusions is also mysterious.

The family of F-BAR (Fer/Cdc42 interacting protein 4 [CIP4] homology-bin, amphiphysin, Rvs]) domain-containing proteins bridge the membrane to the cytoskeleton. BAR domains sense and generate membrane curvature through interaction with membrane phospholipids.⁶  An elongated dimer formed by the antiparallel interaction of 2 α-helical coiled-coils,⁶  the banana-shaped F-BAR domain promotes membrane tubulation of large size (>100 nm).^6,7  The CIP4 gene encodes a protein with a C-terminal SRC homology 3 (SH3) domain, an N-terminal domain with homology to protein tyrosine kinase Fes/Fer, and a sequence that binds active Cdc42.⁸  We identified CIP4 in a yeast 2-hybrid screening with the Src kinase Lyn as bait.⁹  To determine the physiological role of CIP4, we generated CIP4-null mice by disrupting the gene.¹⁰  The CIP4-null mice appeared grossly normal, but displayed decreased endocytosis.¹⁰ 

Through its C-terminal SH3 domain, CIP4 binds to WASP and thus promotes actin cytoskeletal reorganization.^7,11  Because loss of function in WASP results in thrombocytopenia, we investigated whether deficiency of CIP4 affects platelet biogenesis by remodeling the membrane and the actin cytoskeleton. Indeed, we found that CIP4-null mice displayed thrombocytopenia with a similar severity to that of WAS-null mice. Loss of CIP4 conferred decreased cortical actin tension, similar to WASP deficiency. However, different than WASP-null mice, loss of CIP4 led to impaired proplatelet formation, reduced megakaryocyte platelet fields, and plasma membrane stiffening. Thus, CIP4 facilitates the cytoskeletal and membrane remodeling that leads to demarcation membrane system (DMS), known as the reservoir for proplatelet extensions and subsequent platelet release. These studies link platelet biogenesis to membrane biology.

CIP4 KO male C57BL/6 mice were aged 3 to 6 months, as previously described.¹⁰  Their ethical use was approved by the Northwestern Animal Care Use Committee. The WAS− mice¹²  were genotyped by flow cytometric quantification of WASP in blood lymphocytes. Rabbit anti-WASP polyclonal antibody was prepared against peptide (SSRYRGLPAPGPSPADKK) from murine WASP exon 7.

CHRF-288-11 cells¹³  were cultured in Iscove modified Dulbecco medium (Gibco) supplemented with Pen/Strep/Glutamax and 10% fetal bovine serum. Differentiation for proplatelet-like protrusions was obtained by adding PMA (Sigma-Aldrich) or fibronectin (Sigma-Aldrich) as described.¹⁴ 

Platelets were activated with 1 unit/mL thrombin for 5 minutes, lysed in 1% Triton X100, and then centrifuged at 13 000 g for 15 minutes at 4°C. The insoluble fraction was resuspended in radioimmunoprecipitation assay buffer (RIPA) spun at 15 400 g for 5 minutes and actin cytoskeleton was collected from the supernatant. The soluble fraction was spun at 100 000 g for 1 hour and the cytosolic fraction was collected. The pellet was resuspended in RIPA buffer, spun at 15 400 g for 5 minutes and the membrane cytoskeleton was collected from the supernatant.

For knockdown in CHRF-288 cells, we used MISSION short hairpin RNA (shRNA) lentiviral transduction particles (Sigma) (for CIP4, the RNAi consortium [TRC] number TRCN0000063185; for WASP, TRCN0000029822; for TOCA1, TRCN0000129142) provided with nontargeting shRNA control transduction particles (SHC002V), according to the manufacturer’s instructions.

Complete blood cells were run on a HemaVet counter. Femoral bone marrow sections were stained with hematoxylin and eosin; pictures were taken on an Olympus microscope BX50 with an Olympus DP71.

Bone marrow was flushed from femurs and tibias, then filtered through a 70 um strainer. For colony forming unit-MK (CFU-MK) counting with the MegaCult-C assay, a total of 2.2 × 10e6 bone marrow mononuclear cells were used according to the manufacturer's instructions (StemCell Technologies), with 50 ng/mL human thrombopoietin (TPO) and 10 ng/mL murine IL-3 (PeproTech).

Bone marrow cells, flushed from femurs and tibias, were washed in phosphate-buffered saline (PBS)–EDTA–bovine serum albumin and incubated with anti-CD41-FITC (BD Biosciences). Cells were then washed again in PBS–EDTA–bovine serum albumin and fixed in 0.5% paraformaldehyde. Cells were washed, permeabilized in 70% ice cold methanol, ribonuclease-treated, incubated with propidium iodide, and then run on a BD-LSRII. Data were analyzed using FlowJo 7.6 (Tree Star Inc.).

Megakaryocytes from flushed bone marrows were cultured for 3 days in Iscove modified Dulbecco medium (Gibco) with 5% fetal bovine serum and 50 ng/mL TPO, separated on a bovine serum albumin gradient¹⁵  and plated in 96-well plates at a concentration of 3000 cells per well. On day 4, megakaryocytes displaying proplatelets were scored by counting at least 300 cells per well, in at least 3 wells for each condition, using a microscope (Fisher Scientific) with a magnification of 200×. For proplatelet-like protrusion counts and measurements in CHRF-288-11 cells, the cells were seeded at a concentration of 30 000/mL and then exposed to PMA at 10 ng/mL overnight. Protrusion length was measured with ImageJ software (NeuronJ plugin).

CHRF-288-11 cells were transfected with a human CIP4-GFP plasmid (provided by James Goldenring) using the Amaxa Nucleofector II (Lonza), according to manufacturer’s instructions, Kit L and program X-005.

Cells from culture plates were incubated for 2 hours at 37°C on Poly-L-Lysine-coated slides (Polysciences) or fibronectin (Sigma, 50 µg/mL). The cells were then fixed for 20 minutes with 4% paraformaldehyde, permeabilized with Triton-X100 at 0.2% for 3 minutes, washed in PBS, incubated with primary antibody at room temperature for 1 hour or at 4°C overnight, washed in PBS again, then incubated with secondary antibody at room temperature for 30 minutes, washed in PBS again, and then mounted with Vectashield with 4,6 diamidino-2-phenylindole. Antibodies were: rabbit anti-vWF antibody (Dako), mouse anti-CIP4 antibody (BD Biosciences), mouse anti-b-tubulin antibody (Sigma), rabbit anti-b-tubulin (Sigma), rabbit anti-WASP antibody (custom made; Covance), Alexa-conjugated secondary antibodies (Alexa-488 or 594; Invitrogen), and phalloidin-TRITC or phalloidin-FITC (Sigma). Images were taken on a Leica DM 4000B microscope with a Leica DFC320 camera and analyzed by LAS software (Leica). Confocal microscopy was performed on a Nikon Eclipse C1Si confocal microscope.

Femurs were collected from the mice and bone marrow was flushed into glutaraldehyde 2.5% in PBS. CHRF-288-11 cells were collected and fixed in glutaraldehyde 2.5% in PBS. Fixed samples were kept at 4°C and shipped to the Hospital for Sick Children in Toronto or to the Institut Gustave Roussy in Villejuif, and were further processed as previously described.^5,16 

Cell extracts were obtained from lysis in Laemmli buffer with β-mercaptoethanol, or lysed in Triton-X100. Proteins were transferred to Immobilon membranes (Millipore) and incubated overnight at 4°C with the primary antibody, then incubated with peroxidase-conjugated secondary antibody. Antibodies used in western blot were mouse anti-CIP4 (BD Biosciences), goat anti-actin antibody (Santa Cruz), mouse anti-WASP (sc-5300 at 1/200), mouse anti-TOCA 1 (a gift from Dr Giorgio Scita, Milan), and rabbit anti-FBP17 (Bethyl).

Cells were resuspended in PBS and labeled with 1 uM of 1-(4-[trimethylamino]phenyl)-6-phenylhexa-1,3,5-triene (TMA-DPH, Invitrogen) in 37°C for 10 minutes¹⁷  and stimulated with PMA. Alternatively, cells were incubated on fibronectin-coated plates (50 µg/mL; Sigma-Aldrich) for 2 hours at 37°C, and were then washed in PBS. End point readings were taken to detect changes in plasma membrane fluidity. Each experiment was performed in triplicate. Fluorescence anisotropy values were recorded using a Spectramax M5 microplate reader at the Institute for BioNanotechnology in Medicine facility at Northwestern University. Excitation and emission wavelengths were taken at 360 nm and 430 nm, respectively. Polarizers were set at vertical (V) or horizontal (H) positions. Anisotropies (r) were calculated from the intensity maxima using the quantum yield and instrumental correction factor (G-factor) found by measuring cells without TMA-DPH in PBS (r = 0.01) and using the equation: G = IHV/IHH, where IHV is the intensity when excited with a horizontally polarized light and vertical emission detected, and IHH is the intensity with horizontally polarized excitation and emission detected in the same horizontal plane.¹⁸  Assuming G = 1, steady state fluorescence anisotropy was calculated using the equation: r = (IVV − IVH)/(IVV + 2 · IVH).

Molecular dynamics simulations were performed using NAMD (Not [just] Another Molecular Dynamics program) 2.8 (http://www.ks.uiuc.edu/Research/namd/) and run on the Texas Advanced Computing Center Ranger supercomputer cluster. The Chemistry at HARvard Macromolecular Mechanics (CHARMM) 22 force field with CMAP corrections¹⁹  was used for protein-protein interactions and the CHARMM 36 (Klauda et al²⁰ ) forcefield with CMAP corrections was used for lipid-lipid and protein-lipid interactions. The PIP₂ lipid parameters were used as described by Lupyan et al.²¹  A square membrane bilayer patch consisting of 55% 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine, 35% cholesterol, and 10% PIP₂ was created and relaxed using CHARMM-GUI²²  (http://www.charmm-gui.org/). The CIP4 F-BAR domain structure measured by Shimada et al²³  Protein Data Bank Identification code: 2EFK was used for the protein-membrane system and was manually placed on the membrane bilayer surface. Systems were solvated using Visual Moleuclar Dynamics, version 1.9²⁴  (http://www.ks.uiuc.edu/Research/vmd/). The temperature was heated to 310 K and held constant.²⁵  The pressure was held constant at 1 bar.²⁶ 

Cortical tension was measured according to Hochmuth.²⁷  Observations were made at 40× with a Nikon TiE microscope through a camera (GC1290, Prosilica). CHRF-288-11 cells were injected into an open-sided chamber. The micropipette aspiration pressure was controlled through a homemade manometer. Cortical actin layer tension pulls the cell into a spherical shape (with a radius [R_c]). By fine-tuning the suction pressure (P_p) relative to the pressure outside the micropipette (P₀), the cell is maintained in a spherical shape while the aspirated region is a hemisphere such that the aspiration length (L_p) is equal to the pipet radius (R_p). Therefore, the cortical tension (T_c, with units of force per length) can be calculated with the equation according to the law of Laplace.

The Student t test was used to compare the mean ± standard error of the mean, with the assumption of normal distribution, unless mentioned otherwise. Experiments were performed (n = 3, unless mentioned otherwise).

By western blot, we demonstrated the presence of CIP4 in primary megakaryocytes and the megakaryocytic CHRF-288-11 cells (Figure 1A). The staining of CIP4 was diffuse throughout the cytosol and proplatelets (Figure 1B-C). Partial colocalization of CIP4 and WASP occurred in both the cytoplasm and the proplatelet swellings (Figure 1C). CIP4 translocated from the membrane cytoskeleton to the actin cytoskeleton, as did WASP (Figure 1D).

We analyzed C57Bl/6 male mice aged 3 to 6 months and found that they displayed thrombocytopenia to a similar degree as C57Bl/6 WAS− male mice (Table 1). CIP4^−/− mice showed neither lymphocytopenia nor splenomegaly, adenopathy, or runting. There was no difference in the circulating half-life of in vivo NHS-Biotin-labeled platelets from wild-type (WT) or Cdc42 interacting protein 4-null mice (performed as described,²⁸  data not shown), which argues against immune-mediated destruction of platelets. The platelet size did not differ between CIP4^−/−, WAS⁻ mice, and their WT littermates.

Histologic sectioning of femurs showed no obvious features for megakaryocytes (Figure 2A). To study megakaryocyte progenitors, bone marrow cells were cultured from CIP4^−/− and WAS⁻ mice and control littermates. On day 6, cells were stained with cholinesterase, and the number of CFU-MKs was counted. There was no difference between control and CIP4-null mice (Figure 2B) and no difference in ploidy distribution (Figure 2C). By flow cytometry, the number of CD41+ cells did not differ between CIP4-null and controls (1.7% ± 0.7 vs 1.5% ± 0.3). This indicated that CIP4 does not affect early megakaryocytopoiesis and suggested that the thrombocytopenia more likely reflects defective platelet biogenesis.

Bone marrow cells from CIP4-null mice or their WT littermates were grown in suspension with TPO. On day 4, the percentage of proplatelet-forming megakaryocytes was measured (Figure 3A). The fraction of megakaryocytes displaying proplatelet protrusions was decreased, which was not seen in WASP KO megakaryocytes (Figure 3A). The decrease in percentage of CIP4-null megakaryocytes demonstrating proplatelets (40%) correlated with the decrease in platelet number (44%). Because CIP4 was previously shown to interact with tubulin,^11,29  we looked at the ability to form tubulin-driven loops at proplatelet tips, which is a critical component for proper proplatelet formation.¹  Proplatelet protrusions derived from CIP4-null megakaryocytes appeared normal in structure and were stained with tubulin (Figure 3B). We did not notice an abnormal structure in proplatelets from WASP KO megakaryocytes from mice, in line with previous findings.^4,5 

We confirmed the reduction of proplatelet protrusions with loss of CIP4 in CHRF-288-11, a human megakaryocytic cell line that forms proplatelet-like protrusions with PMA treatment.¹⁴  Compared with controls, cells transduced with shRNA to human CIP4 demonstrated marked silencing (79% reduction) of CIP4 (Figure 4A) and a 50% decrease in proplatelet protrusions and in their length (Figure 4B-D), which was not observed with knockdown of WASP or TOCA1 (another F-BAR protein highly similar to CIP4).

The DMS³⁰  is rich in phosphoinositides, phospholipids that BAR proteins preferentially bind.⁷  DMS comprises an intracellular membrane system of branching tubules and flattened cisternae.³⁰  Because F-BAR proteins polymerize around the lipid tubules,⁶  we speculated that loss of CIP4 would not promote extensive tubulation. Transmission electron microscopy (TEM) identifies distinct regions within the cytoplasm of megakaryocytes that have been termed platelet territories,³¹  or platelet-regions containing distinct α-granules surrounded by DMS. Although it has been proposed that platelet territories give rise to platelets via fragmentation, more recent studies suggests that platelets are formed via microtubule-containing projections called proplatelets.¹  TEM of bone marrow showed (Figure 5A), reduced to missing platelet territories, suggesting defective DMS formation, in CIP4-null compared with their WT littermates (>50 megakaryocytes visualized). DMS membranes separating platelet territories were either poorly defined or missing in CIP4-null megakaryocytes, opposite to what was described for WASP deficiency.^4,5  Occasionally, TEM revealed some platelet shedding, as had been described for WASP KO mice by Sabri et al.⁵  However, the abnormal DMS was the predominant abnormality observed, which likely leads to reduced proplatelet formation in CIP4-null megakaryocytes. Thus, in vivo, in vitro, and ultrastructural studies of megakaryocytes suggested that CIP4 affected proplatelet formation. Interestingly, some signs of membrane disruption of the plasma membrane and vacuole membrane were evident in the CHRF cells with knockdown compared with control cells (Figure 5B). Platelets from CIP4-null mice were not ultrastructurally different from their WT littermates (data not shown).

As CIP4 links membrane and actin cytoskeleton,⁶  we hypothesized that loss of CIP4 would affect membrane and cytoskeletal remodeling that leads to proplatelet formation. Cell membrane fluidity is inversely proportional to anisotropy values.³²  We measured anisotropy on cells labeled with TMA-DPH, a molecular probe that localizes specifically in plasma membrane.¹⁸  Treatment with either PMA or fibronectin, agents that promote proplatelet protrusions in CHRF-288-11 cells induced decreased membrane fluidity. PMA-induced membrane fluidity increased by 4% (range 2% to 6%) in control cells. However, no effect of PMA was observed in CIP4-deficient cells (Figure 6A). Fibronectin also increased membrane fluidity by 6 ± 1% in control CHRF-288-11 cells. In contrast, CIP4-deficient cells displayed a 3 ± 1% increase in membrane rigidity (Figure 6B). The effect of CIP4 loss was greater on membrane fluidity than that seen with knockdown of WASP. The knockdown of WASP showed less plasma membrane fluidity in response to PMA, with the increase in fluidity being only 1.8 ± 0.7%. However, this effect was not found with fibronectin. We also performed simulations based on nuclear magnetic resonance (NMR) deuterium studies. NMR deuterium order parameters, S_cd, were calculated from atomistic molecular dynamics simulations for the unsaturated phosphatidylcholine acyl lipid tails (Figure 6C-D). The order parameters were calculated for a 12.0 nm by 3.0 nm region of lipids near the CIP4 protein (Figure 6E) and a membrane patch with no protein. The order parameter is a function of the angle formed by a carbon-deuterium bond on the lipid acyl chain and the vector perpendicular to the membrane surface (Figure 6F-G). A lower order parameter value is associated with increased disorder and fluid character of the acyl chains.³³  As seen in Figure 6C, the order parameters of the plateau region are lower for acyl chains near the protein than in the protein-free system, corresponding to an increase in fluid character of the acyl chains near the protein. The order parameter calculations show a 2.34 ± 0.02% difference between both systems. The simulation results showed that individual F-BAR domains increase lipid disorder in their local region as measured by NMR order parameters. Collectively, this behavior can lead to increased membrane fluidity.

Cell cortical rigidity depends on the connections between the actin and the membrane.³⁴  Through its SH3 domain and recruitment of WASP, F-BAR proteins regulate actin polymerization,³⁵  a main regulator for cortical rigidity.³⁶  To measure stiffness of the actin cortex that provides support for the plasma membrane, we performed micropipette aspiration (Figure 7A-F). Compared with control CHRF-288-11 cells, knockdown of CIP4 or WASP resulted in reduced cortical tension in all conditions tested (baseline, PMA or fibronectin) (Figure 7B), suggesting that the anchorage of actin fibers with the cellular membrane was weakened. These studies revealed that the actin-rich cellular cortex is softer (due to decreased actin polymerization) in CIP4 or WASP-deficient cells. These differences were maintained even when cortical tension increased after PMA stimulation. Sometimes, the transparent region turned into blebs (Figure 7C-F). In controls, the transparent region was aspirated farthest into the micropipette and blebbing was rare. With CIP4 knockdown, cell swelling was observed under the isotonic condition, and repeated blebbing was observed in the inner transparent region inside the pipette (measurements shown on Figure 7B were performed before blebbing occurred).

Our studies demonstrate that loss of CIP4 resulted in thrombocytopenia, which was associated with decreased proplatelet formation. Biophysical studies revealed that loss of CIP4 produced a stiffer membrane and a softer actin cortical tension. Ultrastructural studies revealed poorly defined or missing DMS in CIP4-deficient megakaryocytes. Our findings identify CIP4 as a new component in the molecular machinery that remodels megakaryocyte membrane and generates proplatelets, and suggests a new mechanism for the tubulation required for formation of the DMS.

The production of platelets from megakaryocytes requires a series of highly coordinated processes involving membrane/cytoskeletal remodeling unique in the body.¹  Two events are critical: first, the megakaryocyte gains mass and then the megakaryocyte buds off proplatelets.³⁷  In vivo imaging of megakaryocytes in the bone marrow microenvironment niche demonstrated extension of proplatelets into the vasculature, where blood flow provides shear forces that generate pre-platelet/platelet release into the circulation.³⁸  To date, only a few molecules have been identified that drive the dynamic remodeling of membrane and cytoplasm in platelet biogenesis. Our studies have now identified a new component, F-BAR protein CIP4, and a new mechanism for proplatelet formation involving CIP4-directed membrane/actin remodeling.

The F-BAR subfamily of proteins is diverse with 6 subgroups, each possessing different properties, such as tyrosine kinase activity, GTPase function, or phosphatase interaction.³⁹  The F-BAR protein most closely related to CIP4 is TOCA1. When TOCA1 was silenced by shRNA, no decrease in proplatelet formation was found, suggesting that CIP4 provides a unique function. Other types of BAR proteins exist: N-BAR proteins and I-BAR proteins, such as IRSp53. When platelet biogenesis was studied in IRSp53-deficient cells, no abnormalities were found.⁴⁰  As a scaffolding protein, CIP4 interacts with the activated form of Cdc42, inactive state of WASP, Src kinases, and the negatively charged phosphoinositides residing in the inner leaflet of the plasma membrane. The F-BAR domain dimers enable oligomerization and extensive lateral interactions.⁶  This tight packing on curved membranes promotes membrane tubulation. In addition, CIP4 brings together actin cortex with these plasma membrane deformations. We recently reported that invadopodia, a form of cellular protrusion, were decreased in breast cancer cells when CIP4 is knocked-down by small interfering RNA.⁴¹  Here, we observed decreased platelet fields in CIP4-null megakaryocytes when analyzing their ultrastructure. Because proplatelet formation depends on the membrane and cytoskeleton of the megakaryocyte, reduced DMS likely impairs proplatelet formation. Therefore, not surprisingly we also observed decreased proplatelet protrusions per megakaryocyte in primary megakaryocytes from CIP4-null mice. We reasoned that loss of CIP4 affected membrane remodeling. Indeed, fluorescence anisotropy and micropipette aspiration assays demonstrated a more rigid plasma membrane and a softer cell cortex. The relationship between 2H-NMR order parameters and elastic properties of the membrane has been experimentally established.³²  As the flexibility of the acyl chain decreases, the membrane adopts a more gel-like phase. Interestingly, a previous work showed that TOCA1, a protein closely related to CIP4, binds preferentially to the more gel-like, ordered membrane regions; however, the gel phase was not an absolute perquisite for TOCA1-induced filopodia-like structure formation.⁴²  In contrast to spontaneous membrane undulations, a comparatively high level of energy is required to cause membrane bending.⁴³  Our findings of decreased cortical tension in cells with knockdown of CIP4 or WASP suggest that CIP4 provides a localization cue for engaging WASP-triggered cortex actin network contributing to platelet biogenesis; however, besides cortical tension measurements, CIP4 behave differently than WASP in regard to proplatelet formation, DMS formation, and effect on plasma membrane order. Even though the difference in plasma membrane order induced by CIP4 depletion was small in our experimental data, it was statistically significant in experimental data and was supported by theoretical computer-based simulation. This new finding supports that a change in membrane order at the molecular level will have repercussion in plasma membrane recycling at the cellular level, with eventually perturbing processes as membrane invagination (DMS) and subsequent proplatelet protrusion.

Insights have come from the analyses of hereditary thrombocytopenias, such as Wiskott-Aldrich syndrome, in which the platelet size is small, or those with giant platelet syndromes, including mutations of Myh9,^44,45  filamin-A,^46,47  and beta1-tubulin.⁴⁸  Because murine mean platelet volume (MPV) is half that of human MPV, WAS-deficient platelets in mice are already small. Our studies here suggest that loss or mutation in the CIP4 gene would result in a non-X-linked form of congenital thrombocytopenia, but the mouse strain cannot predict whether the platelets would be smaller.

Here, we report that gene disruption of CIP4 resulted in thrombocytopenia to a similar degree observed in WAS-null mice. The SH3 domain of CIP4 interacts with WASP.^11,29  We observed that loss of either CIP4 or WASP resulted in decreased cortical tension in megakaryocytes, suggesting at least one common mechanism in defective platelet biogenesis. Although CIP4 contributes to T-cell adhesion and migration⁴⁹  and to NK cell cytotoxicity,²⁹  we have not observed in CIP4 mice such a profound immunodeficiency that is similar to that of WAS-null mice.⁵⁰  WAS-null mice and CIP4-null mice also differ in regard to proplatelet and DMS formation in primary megakaryocytes. Thus, we conclude that the effects of CIP4 are due to cortical tension-related mechanisms involving WASP, and also WASP-independent mechanisms on plasma membrane remodeling and membrane recycling. The role of CIP4 in megakaryocyte interaction with matrix could be subsequently studied.

In conclusion, we have found that lack of CIP4, a scaffolding protein that interacts with Cdc42, Src kinases, and WASP, results in murine thrombocytopenia. F-BAR protein CIP4 remodels both the plasma membrane and the cortical cytoskeleton. Loss of CIP4 affects both megakaryocyte membrane fluidity and its cortical tension. Recent theoretical models support that modulation of biophysical cortical forces is crucial for platelet formation,⁵¹  with a new finding that depletion of a protein-modifying plasma membrane order will result in thrombocytopenia. Our findings should stimulate new investigations into the biophysical events driving cytoskeletal-membrane remodeling, as well as new candidate genes in inherited thrombocytopenias.

The authors thank Giorgio Scita, James Goldenring, and William Miller for providing reagents, and Amanda Prislovsky for genotyping WASP KO mice.

This work was supported by grants from the National Heart, Lung and Blood Institute, National Institutes of Health (NIH), Health and Human Services (HHS) Mentoring Program RO1HL080052 and R21HL106462 (S.J.C.); the American Heart Association Grant-in-Aid (S.J.C.); the National Institute of Allergy and Infectious Diseases, NIH HHS R21AI079757 (T.S.S.) and RO1AI071163 (D.J.R.); the Canadian Institutes of Health Research MOP-81208 (W.H.A.K.); the National Institute of General Medical Sciences, NIH HHS RO1GM063796 (G.A.V.), and the American Heart Association Post-Doctoral Fellowship (Y.C.).

Contribution: Y.C., J.A., S.K., E..B., B.C., A.A., S.B., N.D., D.R.M., R.T., D. Reece, L.J., Z.L.W., H.C., W.K, T.S.S., W.A.L., C.Z., F.V., G.A.V., and S.J.C. designed experiments, performed research, and/or analyzed data; D. Rawlings provided critical reagents; and Y.C., S.K., W.H.A.K., W.A.L., C.Z., F.X.V., and S.J.C. wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interest.

Correspondence: Seth J. Corey, Lurie 5-107, Robert H. Lurie Comprehensive Cancer Center, 303 E Superior St, Chicago, IL 60611; e-mail: coreylab@yahoo.com.