Cleavage Fragments of the Third Complement Component (C3

Cleavage fragments of the third complement component (C3) enhance stromal derived factor-1 (SDF-1)-mediated platelet production during reactive postbleeding thrombocytosis

M Wysoczynski1, M Kucia1, J Ratajczak1 and MZ Ratajczak1,2

1Stem Cell Biology Program, James Graham Brown Cancer Center, University of Louisville, Louisville, KY, USA and 2Department of Physiopathology, Pomeranian Medical University, Szczecin, Poland

We hypothesized that the third complement component (C3) The a-chemokine stromal-derived factor-1 (SDF-1) has been cleavage fragments (C3a and des-ArgC3a) are involved in stress/ proposed as a new regulator of Megs maturation and platelet inflammation-related thrombocytosis, and investigated their 11–16 potential role in reactive thrombocytosis induced by bleeding. formation. A recently published study clearly demonstrated We found that platelet counts are lower in C3-deficient mice in that SDF-1, together with fibroblast growth factor-4 (FGF-4), À/À À/À response to excessive bleeding as compared to normal restored thrombopoiesis in Tpo and Mpl mice by directing littermates and that C3a and des-ArgC3a enhance stromal- trans-localization of megakaryocytic progenitors positive for derived factor-1 (SDF-1)-dependent megakaryocyte (Megs) CXCR4, the receptor for SDF-1, from the endostial to the migration, adhesion and platelet shedding. At the molecular vascular niche, thereby promoting survival, maturation to level, C3a stimulates in Megs MAPKp42/44 phosphorylation, 17 and enhances incorporation of CXCR4 into membrane lipid megakaryocytes and platelet production. rafts increasing the responsiveness of Megs to SDF-1. We Furthermore, there is growing evidence that the responsiveness found that perturbation of lipid raft formation by statins of hematopoietic cells to SDF-1 is optimal when the CXCR4 decreases SDF-1/C3a-dependent platelet production in vitro receptor is included into membrane lipid rafts.18–20 The formation and in an in vivo model statins ameliorated post-bleeding of lipid rafts in vivo may be perturbed by cholesterol- thrombocytosis. Thus, inhibition of lipid raft formation could lowering drugs, for example, statins,18–20 which are effective find potential clinical application as a means of ameliorating in lowering LDL cholesterol and exert pleiotropic effects on some forms of thrombocytosis. 21 Leukemia (2007) 21, 973–982. doi:10.1038/sj.leu.2404629; mature platelets, for example, inhibit their activation. This published online 1 March 2007 latter effect is probably due to lowering of the cholesterol Keywords: megakaryopoiesis; platelets; complement; chemokines; content in the platelet membranes leading to disruption of CXCR4; lipid rafts membrane lipid raft formation, as lipid rafts are necessary for proper platelet activation and signaling21 and, as we hypothesized, platelet production. However, under steady-state conditions, statins do not influence thrombopoiesis, no direct experimental studies have been performed on their effects on platelet production in reactive thrombocytosis. We recently reported that the complement (C) system and Introduction third complement protein (C3) cleavage fragments C3a and C3a enhance the responsiveness of hematopoietic cells to Megakaryopoiesis is regulated by several factors that affect the des-Arg an SDF-1 gradient.22 C3a binds to G -protein-coupled seven proliferation and differentiation of megakaryopoietic cells.1,2 ai transmembrane receptor (C3aR) and we reported that C3aR is The fact that mice with ‘knockout’ of thrombopoietin (TPO) and expressed on human and murine hematopoietic stem/progenitor thrombopoietin receptor (c-mpl) show a 90% reduction in both cells (HSPC).22,23 In contrast, the receptor for C3a has not the number of megakaryocytes (Megs) in bone marrow (BM) and des-Arg been identified yet. As C is activated in several clinical circulating platelets indicates that TPO is a crucial regulator of conditions associated with high platelet counts (e.g., chronic megakaryopoiesis.3–5 Nevertheless, the fact that some Megs and inflammation, hemorrhage), we hypothesized that C3a and platelets are still present in these animals even in the total C3a are involved in stress/inflammation-related thrombo- absence of TPO suggests that some other factors can compen- des-Arg cytosis. Supporting this are the observations that C3-deficient sate for TPO deficiency and the cytokines that signal through mice, which have normal platelet counts in steady-state gp130 such as interleukin-6 (IL-6), IL-11, leukemia inhibitory conditions,23 display delayed recovery of platelets after sub- factor, cilliary neurotropic factor and oncostatin M have been lethal irradiation or hematopoietic transplants.23 suggested to be such factors.6–9 However, recent data show that We report that C3a and C3a modulate the responsiveness IL-6 and IL-11 do not induce platelet production in thrombo- des-Arg of Megs to SDF-1 and postulate that the crosstalk between cytopenic, TPO-deficient (TpoÀ/À) and TPO-receptor-deficient C3aR and CXCR4 receptors, we have newly identified, here (MplÀ/À) mice.10 Thus, it is unlikely that gp130 signaling plays an important and previously unrecognized role in stress/ cytokines are significantly involved in platelet production in inflammation-dependent Megs maturation and platelet formation. these animals.4–10

Correspondence: Dr MZ Ratajczak, Stem Cell Biology Program, James Materials and methods Graham Brown Cancer Center, University of Louisville, Louisville, KY 40202, USA. þ E-mail: [email protected] Human CD34 cells, megakaryoblasts and platelets Received 4 December 2006; revised 16 January 2007; accepted 25 Light-density BM mononuclear cells (BM MNC) were obtained January 2007; published online 1 March 2007 from consenting healthy donors and enriched for CD34 þ cells Complement as a new regulator of megakaryopoiesis M Wysoczynski et al 974 by immunoaffinity selection with MiniMACS paramagnetic on human Megs with mouse anti-human C3aR monoclonal beads (Miltenyi Biotec, Auburn, CA, USA) (purity 495%) and antibody (MoAb), clone no. 8H1 and a fluorescein isothiocya- were expanded in a serum-free liquid system, and growth of nate (FITC)-labeled goat anti-mouse antibody. Subsequently, the CFU-Meg was stimulated with recombinant human (rh) TPO cells were stained with phycoerythrin (PE)-labeled anti-CD41 (100 ng/ml) (R&D Systems, Minneapolis, MN, USA) as de- MoAb (Becton Dickinson, San Jose, CA, USA), washed, fixed in scribed.16 After incubation for 11 days at 371C, about 85% of 1% paraformaldehyde and subjected to FACS analysis using a the expanded cells were positive for the megakaryocytic- FACScan analyser (Becton Dickinson). specific marker CD41, and this population was further enriched to 495% purity by additional selection with immunomagnetic beads (Miltenyi Biotec) as described previously by us.16 Gel- Isolation of mRNA and RT-PCR for detection of C3aR filtered platelets (GFP) were prepared from four individuals as and C5L2 described previously.13,16 Total mRNA was isolated with the RNeasy Mini Kit (Quiagen Marrow aspiration and blood donation from normal volun- Inc., Valencia, CA, USA) as described.20 We employed teers was carried out with the donors’ informed consent following primers – for detection of C3aR forward 50-GCC obtained through the Institutional Review Board. GCC TGG AGA AAT GAA TGA TAG G-30 and reverse 50-AGA AAG ACA GCC ACC ACC ACG-30, and for detection of C5L2 forward 50-CCT GGT GGT CTA CGG TTC AG-30 and reverse CFU-Meg assay 50-GGG CAG GAT TTG TGT CTG TT-30. Amplified products þ CD34 BMMNC were resuspended in 1% methylcellulose in (10 ml) were electrophoresed on a 2% agarose gel. Iscove’s DMEM (Gibco BRL, Grand Island, NJ, USA) (104/ml) supplemented with 25% artificial serum as described.9,16 CFU- Meg growth was stimulated with a suboptimal (10 ng/ml) and Real-time PCR analysis of MMPs and VEGF optimal dose of rhTPO (100 ng/ml). Cultures were incubated at Detection of MMP2, MMP9, VEGF and b2-microglobulin 371C in a fully humidified atmosphere supplemented with 5% mRNA levels was performed by real-time RT-PCR using an CO2. Under these conditions, after 11 days, approximately ABI PRISM 7000 Sequence Detection System (ABI). Each 25 ml 9,16 100% of the colonies were glycoprotein aIIb/b3 positive. reaction mixture contained 12.5 ml SYBR Green PCR Master Mix, 10 ng of cDNA template and primer for MMP9 forward 50- GGA CGA CGT GGG CTA CGT-30, reverse 50-AAT CTC ACC Cell lines GAC AGG CAG CT-30, for MMP2 forward 50-TGG GAC AAG The MO7E cell line used in these studies was purchased from AAC CAG ATC ACA TA-30, reverse 50-TTT CGA GTC TCC ACG ATCC (Rockville, MD, USA) and maintained in Iscove medium CAT CTC-30, for VEGF forward 50-TGA GCG GCT CAT CTA CTT (Gibco BRL, Long Island, NY, USA) supplemented with 10% CTA TGT-30, reverse 50-CAC CGG CTG GCC CTC TA-30, for bovine calf serum (BCS) (Hyclone, Logan, UT, USA) and 5 ng/ml b2-microglobulin forward 50-TGA CTT TGT CAC AGC CCA of GM-CSF (R&D Systems). AGA TA-30, reverse 50-AAT GCG GCA TCT TCA AAC CT-30. Relative quantitation of MMP2, MMP9 and VEGF mRNA expression was calculated with the comparative threshold cycle Bleeding procedure 24 Mice were anesthetized by i.p. injection of Ketamine (100 mg/kg, (Ct) method described elsewhere. Sigma, St Louis, MO, USA) and bled from the retro-orbital plexus as recommended by the Animal Resource Center, Case Western Reserve University (http://labanimals.case.edu/ Apoptosis assay templates_retro_orbital_bleeding.html). Apoptosis of Megs was assessed by cytometric analysis after staining cells with FITC-Annexin V (apoptosis detection kit from R&D Systems) according to the manufacturer’s protocol as 9,16 ELISA on serum from murine BM described. Blood samples were collected into EDTA tubes and plasma was separated immediately by centrifugation at 2000 g for 15 min at 41C. For detection of des-ArgC3a in murine plasma, we employed Megs proliferation by MTT assay the rat/mouse ELISA kit according to the producer’s protocol The MTT assay was performed according to the manufacturer’s (Cedarline, ON, Canada). In brief, the microtiter plates were recommendations (Promega, Madison, WI, USA). Briefly, cells Â 4 m coated with a chicken anti- des-ArgC3a antibody and serial were seeded in 96-well plates at 5 10 /well in 100 l of RPMI dilutions of mouse serum (1:100 starting dilution in blocking medium containing 0.5% BSA plus 10 ng/ml or 100 ng/ml of buffer) were added and incubation for 1 h at 371C was carried TPO þ SDF-1 (300 ng/ml) (R&D Systems) þ /ÀC3a or des-ArgC3a m m out. Subsequently, polyclonal chicken anti-mouse des-ArgC3a (1 g/ml) (Calbiochem, San Diego, CA, USA). After 24 h, 20 lof (biotinlated, 4 mg/ml in blocking buffer, ICN Pharmaceuticals CellTiter 96 Aqueous One Solution reagent was added to each Inc., OH, USA) was mixed with extravidin-peroxidase and the well and the plates were incubated for 3–4 h. Subsequently, 9,16 mixture was then added to the wells for further incubation plates were read at 490 nm using an automated plate-reader. of 1 h. des-ArgC3a-containing complexes were detected by measuring the optical density in an ELISA reader (KC junior, BIO-TEK Instruments Inc., VT, USA) at 450 nm with a reference Calcium flux assay wavelength set at 630 nm. Briefly, the cells were incubated for 30 min at 301C with 1–2 mM Fura-2/AM (Molecular Probes, Carlsbad, CA, USA). After incubation, the cells were washed once, resuspended in loading FACS analysis buffer without BCS, stimulated with C3a (1 mg/ml), des-ArgC3a Expression of C3aR was evaluated by fluorescence-activated (1 mg/ml) or SDF-1b (300 ng/ml) and analyzed within 1 h as cell sorter (FACS) as described previously.22 C3aR was detected described previously.22

Leukemia Complement as a new regulator of megakaryopoiesis M Wysoczynski et al 975 Actin polymerization assay tion overnight at 41C, and CD41 þ cells (106/ml) were The cells were washed three times and resuspended in assay incubated for 30 min at 371C in serum-free Iscove DMEM in medium (Iscove’s DMEM with 0.5% BSA) at a concentration of the absence or presence of SDF-1 (1 mg/ml), C3a (1 mg/ml), 6 10 cells/ml. The ligands were added to the cell suspension and, des-ArgC3a (1 mg/ml), SDF-1 þ C3a or SDF-1 þ des-ArgC3a. Cell at indicated time points, 100 ml cell solution was transferred suspensions (100 ml) were applied to the wells and incubated for to the fixation solution (Cytofix/Cytoperm; Pharmingen). Subs- 10 min at 371C. The number of adherent cells was determined equently, cells were fixed and then centrifuged and resuspended using inverted microscopy. in 100 ml permeabilization reagent (Perm/wash buffer; PharMin- gen). After 5 min of incubation in this solution, 1 U/ml Alexa 488 phalloidin (Molecular Probes) was added to visualize the Transmigration and platelet formation assay F-actin. MFI was measured by FACScan (Becton Dickinson) Transwell chambers (8 mm pore size) were covered with an and examined using a BX51 fluorescence microscope (Olympus impermeable monolayer of human umbilical vein endothelial America, Melville, NY, USA) equipped with a charge-coupled cells (HUVEC) which were isolated as described previously. device camera (Olympus America). Separate pictures were HUVECs were activated overnight with IL-1b (10 U/ml) before merged using Image-Pro Plus software (Media Cybernetics, assay. Megs that had been expanded from CD34 þ cells were Silver Spring, MD, USA). added to the top transwell chambers, and aliquots of 650 ml serum-free medium containing SDF-1, C3a and des-ArgC3a, alone or in combination, were placed in the lower chambers. In the Chemotaxis assay platelet formation assay performed for 48 h, the number of Cells were resuspended in RPMI with 0.5% BSA and equili- CD41 þ platelets present in the lower chambers was determined 1 brated for 10 min at 37 C and chemotaxis assays were after staining with anti-CD41 MoAbs and using flow cytometry performed by employing Costar Transwell 24-well plates (Costar (FACSscan, Beckton Dickinson). Corning, Cambridge, MA, USA) with 8 mm-pore filters as described.20 The results are presented as a chemotactic index (the ratio of the number of cells that migrated toward the Isolation of lipid rafts medium containing C3a and/or SDF-1 to the number of cells 8 20 that migrated toward the medium alone). MO7E cells (1 Â 10 ) were lysed as described. Samples were overlain with 30 and 5% sucrose in MEB and were centrifuged at 100 000 g for 17 h. Fractions were gently removed from the Adhesion assay top of the gradient, and n-octylglucoside was added to each m Ninety-six-microtiter plates (Dynatech Labs, Chantilly, VA, fraction (60 M final) to solubilize rafts. For detection, Western USA) were covered with 4 mg/ml fibrinogen (Sigma) by incuba- blot analysis was carried out using standard techniques with a CXCR4 antibody (Serotec, Oxford, UK) and cholera toxin subunit B conjugated with horseradish peroxidase (HRP) (Sigma, Milwaukee, WI, USA).20 a 2000 1800 * * wt 1600 * C3 -/- 1400 A M 1 2 3 1 2 3 BLEEDING 1200 1000 800 PLT (K/uL) 600 1. C3aR 400 2. C5L2 200 33.. β-actin 0 -101234567891011121314151617 days after bleeding

4500 b wt Megs platelets 4000 * C3 -/- 3500 * * * BLEEDING B 3000 abc 10000 2500 14.8 66.5 100 100 80 80

ng/ml 2000 1000 60 60 1500 100

1000 FL2-H 40 40 % of Max % of Max 10 20 500 20 99.1% 30.3% 8.68 10.1 1 0 0 0 1 10 100 1000 10000 10 10 10 10 10 10 10 10 10 10 -1 4 9 14 19 24 29 FL1-H FL2-H FL2-H hrs after bleeding Figure 2 Human Megs and platelets express C3aR. (A) RT-PCR Figure 1 Thrombocytosis and complement activation during exces- analysis of mRNA expression for C3aR and C5L2 on human Megs and sive bleeding. (a) Kinetics of platelet counts after bleeding in wt versus platelets. M, marker; C3aR, lanes 1; C5L2, lanes 2 and b-actin, lanes 3. À/À C3 mice. (b) Changes in the level of des-ArgC3a in serum of wt mice (B)(a) double staining of human Megs with FITC-anti-C3aR and PE- and C3À/À mice (control) after bleeding were determined using ELISA. anti-CD41, (b) human platelets stained with PE-anti-CD41 and (c) with The data represent the average of three independent experiments FITC-anti-C3aR. The data represent the average of two (RT-PCR) and (using 6–8 mice for each experiment). Data are mean7s.d. three (FACS analysis) independent experiments. Representative results *Po0.00001 wt versus C3À/À mice. are shown.

Leukemia Complement as a new regulator of megakaryopoiesis M Wysoczynski et al 976 Phosphorylation of intracellular pathway proteins separated and analyzed for phosphorylation, of 44/42 MAPK as Western blot analysis was performed on extracts prepared from described.16 Equal loading in the lanes was evaluated by quiescent CD34 þ -derived Megs as described.16 Megs were then stripping the blots and reprobing them with the appropriate stimulated with 1 mg/ml of C3a or 1 mg/ml of des-ArgC3a for 1, 2, 5 monoclonal or polyclonal antibody p42/44 anti-MAPK antibody and 10 min at 371C and then lysed for 10 min on ice in M-Per clone 9102 (New England Biolabs). The membranes were lysing buffer (Pierce, Rockford, IL, USA) containing protease and developed with an ECL reagent (Amersham Life Sci.), dried and phosphatase inhibitors (Sigma). The extracted proteins were exposed to ﬁlm (HyperFilm, Amersham Life Sci.).

A a 1.6 b 1.55 1.55 1.5 1.5 C3a SDF-1 1.45 1.45 Ratio Ratio 1.4 1.4 1.35 1.3 1.35 1 50 99 148 1 50 99 148 time (sec) time (sec)

c 1.9 d 2.3 SB290157 pre-treated 1.7 2.1

C3a 1.9 1.5 1.7 1.3 1.5 C3a desArg SDF-1 Ratio Ratio SDF-1 1.1 1.3 1.1 0.9 0.9 0.7 0.7 1 50 99 148 197 246 295 1 50 99 148 197 246 time (sec) time (sec)

240 B SDF-1 160 SDF-1

220 C3a desArgC3a 150 SDF-1+C3a SDF-1 + C3a 200 desArg 140 180 130 160 % of control 140 % of control 120

120 110

100 100 0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 140 160 180 time (sec) time (sec)

C CONTROL SDF-1 C3a desArgC3a

Figure 3 Human Megs express functional C3aR. (A) C3a but not des-ArgC3a induces calcium flux in human Megs. Normal human Megs Fura-2- loaded were stimulated by (a) C3a (1 mg/ml), (b) SDF-1 (300 ng/ml), (c) des-ArgC3a (1 mg/ml) followed by SDF-1 (300 ng/ml) or (d) preincubated with SB290157 (C3aR-specific inhibitor) and then stimulated by C3a (1 mg/ml) and SDF-1 (300 ng/ml). All compounds were added at the indicated time points, and calcium flux was evaluated by spectrophotofluorimetry. Data presented are from a representative experiment, which was repeated three times with similar results. (B) FACS analysis of F-actin polymerization in human Megs. SDF-1, C3a and SDF-1 þ C3a but not des-ArgC3a induce F-actin polymerization. Data presented are from a representative experiment, which was repeated three times with similar results. Po0.0001. (C) Immunofluorescence analysis of F-actin polymerization in human Meg (green fluorescence, phalloidin; blue, DAPI). The experiment was repeated three times on cells from three independent donors with similar results. Representative staining is shown.

Leukemia Complement as a new regulator of megakaryopoiesis M Wysoczynski et al 977 Statistical analysis Likewise, we also observed that C3a but not des-ArgC3a Arithmetic means and standard deviations were calculated on a induced F-actin polymerization in human Megs (Figure 3b). Macintosh computer powerbase 180, using Instat 1.14 (Graph- However, the F-actin polymerization induced by C3a alone was pad, San Diego, CA, USA) software. Statistical significance was shorter than after addition of SDF-1, and addition of C3a to defined as Po0.01. Data were analyzed using the Student’s SDF-1 significantly enhanced the SDF-1-mediated effect. The t-test for unpaired samples. fact that des-ArgC3a did not induce F-actin polymerization (Figure 3b) and SB290157-inhibited C3a-induced F-actin polymerization (data not shown) indicates that C3aR is responsible Results for the polymerization of F-actin. Our FACS-based F-polymer-

Extensive bleeding in C3-deficient mice leads to a lower platelet count and in normal mice causes a 10000 activation of C3 * Given that excessive bleeding leads to reactive thrombocyto- 8000 25,26 * sis, and C becomes activated in several stress-related 6000 * * situations,27 our aim was to determine whether C3 cleavage 4000 fragments might be implicated in post-bleeding thrombocytosis. % of control To address this, we first investigated the kinetics of platelet 2000 counts after excessive bleeding (20% of blood vol) in wild-type 0 c- (wt) and C3-deficient mice and, in parallel, measured C C3a C3a C3a C3a+SB activation by detecting the level of C3 cleavage fragment SDF-1 low SDF-1+C3a desArg desArg SDF-1 high (des-ArgC3a) in the peripheral blood of bled animals (Figure 1). SDF-1+C3a+SB desArg Figure 1a shows that both wt and C3-deficient mice display a SDF-1+ biphasic increase in peripheral blood counts at days 1–2 and SDF-1+ 6–10. Although the first elevation in platelet counts may reflect b 700 the supply of platelets by Megs located in the BM endothelial * niche, the second elevation is likely related to generation of new 600 * Megs from a pool of CFU-Meg that are attracted to this niche 500 from the osteoblastic niche.17 Although C3-deficient mice have normal platelet counts in normal steady-state conditions,23 they 400 generated fewer platelets after excessive bleeding as compared 300

to wt littermates (Figure 1a). Next, using an ELISA assay % of control ** 200 (Figure 1b), we found that C3 is cleaved in peripheral blood during bleeding as seen by an increase in des-ArgC3a in the blood 100 of normal mice. As expected, however, no C3a was des-Arg 0 detectable in C3-deﬁcient control animals. SDF-1 control SDF-1 C3a SDF-1+C3a desArgC3a desArgC3a

C3aR is expressed on normal Megs and platelets Next, we evaluated whether normal human Megs and platelets express receptors that bind C3 soluble cleavage fragments. c 230 Using RT-PCR, we observed that highly puriﬁed normal human 210 * * Megs obtained from expansion of BM-derived CD34 þ cells as 190 well as peripheral blood platelets express mRNA for C3aR 170 150 (Figure 2a). More importantly, we detected C3aR protein 130 expression by FACS on megakaryocytic cells derived from ex 110 þ % of control vivo expanded BM-derived CD34 cells and puriﬁed peripheral 90 blood platelets (Figure 2b). Normal Megs and platelets also 70 28 express mRNA for the C5L2 receptor, an orphan receptor that 50 SDF-1 may potentially bind des-ArgC3a (Figure 1a). control SDF-1 C3a SDF-1+C3a desArgC3a desArgC3a

Figure 4 C3a and des-ArgC3a enhance SDF-1-dependent chemotaxis, C3aR is functional on normal human Megs adhesion and platelet production. (a) C3a and des-ArgC3a enhance the The finding that C3aR is expressed on normal human Megs chemotactic response of human Megs to a low threshold dose of SDF-1 (10 ng/ml), in a C3aR-independent manner as shown after suggests a role for the C3 cleavage fragments in regulating employing SB290157 (C3aR-specific inhibitor). As positive control, normal human megakaryopoiesis. Highly purified human Megs SDF-1, was used in high dose (300 ng/ml). The data represent the were exposed to C3a or des-ArgC3a, and assayed for calcium flux, average of three independent experiments. Data are mean7s.d. F-actin polymerization and chemotactic responses. As positive *Po0.00001 versus SDF-1 low. (b) Both C3a (1 mg/ml) and des-ArgC3a controls, we employed SDF-1 as a stimulant (Figure 3). (1 mg/ml) enhance SDF-1 (1 mg/ml)-dependent adhesion of Megs to Figure 3a shows that C3a, like SDF-1, stimulated calcium flux fibrinogen. Lower part – representative pictures (inverted microscope). The data represent the average of three independent experiments. Data in megakaryocytic cells. The fact that C3a did not des-Arg are mean7s.d. *Po0.00001 versus SDF-1 low. (c) C3a (1 mg/ml) and stimulate calcium flux and that the C3aR inhibitor SB290157 des-ArgC3a (1 mg/ml) increase SDF-1 (300 ng/ml)-dependent production inhibited C3aR but not SDF-1-mediated calcium flux supports of platelets. The data represent the average of four independent the notion that this effect is specific to C3aR. experiments. Data are mean7s.d. *Po0.00001 versus SDF-1 alone.

Leukemia Complement as a new regulator of megakaryopoiesis M Wysoczynski et al 978 ization data were subsequently confirmed at the single-cell level response to a low, ‘threshold’ dose of SDF-1 (10 ng/ml) by direct F-actin staining (Figure 3c). (Figure 4a). In the presence of C3a or des-ArgC3a, the chemotactic response of Megs to a threshold dose of SDF-1 was comparable to that at a high dose of this chemokine (300 ng/ml). Interest- C3a and des-ArgC3a do not affect CFU-Meg proliferation and Megs survival ingly, this priming effect of C3a on responsiveness of Megs to an Next, we investigated whether C3a and des-ArgC3a could influence SDF-1 gradient was not inhibited by the C3aR antagonist, the proliferation/survival of normal human Megs. Highly purified SB290157, suggesting the involvement of a receptor other than þ CD34 cells were cultured in a serum-free liquid cultures or C3aR that is probably activated both by C3a and des-ArgC3a. 16 serum-free methylcellulose cloning medium with various doses of Next, as SDF-1 enhances adhesion of Megs to fibrinogen, we focused on the effect of C3 cleavage fragments on SDF-1- C3a or des-ArgC3a (0.1–10 mg/ml) and stimulated to grow CFU- Meg colonies with sub-optimal (10 ng/ml) and optimal dependent adhesion of Megs to fibrinogen (Figure 4b) and found (100 ng/ml) doses of TPO alone or TPO þ SDF-1. In another set that C3a and des-ArgC3a significantly enhanced SDF-1-dependent of experiments, we incubated Megs for 24 h in serum-free adhesion of Megs (Figure 4b). At the same time, adhesion of Megs was also slightly enhanced in the presence of C3a alone, medium in the presence or absence of C3a or des-ArgC3a (0.1–10 mg/ml) and estimated cell viability/Annexin V binding but not des-ArgC3a. This C3a-induced adhesion of Megs after this period of time. In all these experiments, however, no correlated with the fact that C3a employed alone was able to effect of C3a or des-ArgC3a on the proliferation/survival of Megs enhance F-actin polymerization in these cells (Figure 3c). was observed (data not shown). Thus, C3a and C3a do not It is widely accepted that Megs produce platelets by extending des-Arg 12 affect the proliferation/survival of hematopoietic cells. proplatelet protrusions between endothelial cells, and it had been recently reported that SDF-1 enhances platelet formation by Megs that are in contact with endothelium.12 To address C3a and des-ArgC3a increase responsiveness of Megs to whether this process could be affected by C3a and des-ArgC3a, an SDF-1 gradient and enhance SDF-1-dependent we evaluated platelet production in a transwell-migration migration, adhesion and platelet production assay.29 In this assay, transwell membranes were covered by Although C3a and des-ArgC3a alone did not chemoattract Meg in confluent IL-1b activated HUVEC (Figure 4c), and megakar- a transwell system, we observed that they did increase their yocytic cells derived from day 12 expansion cultures were

a C3a desArgC3a c- 1 2 5 10 min c- 1 2 5 10 min phospho-MAPK42/44

total-MAPK42/44

b FRACTION FROM TOP 1234567891011 CONTROL

C3a CXCR4

desArgC3a RTFA NO RAFT GM1

c CXCR4 GM1 MERGE

CONTROL

C3a

desArgC3a

Figure 5 Western blot analysis of C3a- and des-ArgC3a-activated signaling pathways in Megs. (a) Both C3a (1 mg/ml) and des-ArgC3a (1 mg/ml) activate MAPK42/44. Experiments were performed three times with similar results. Representative blots are shown. (b) Analysis of the colocalization of CXCR4 in various fractions of cell membranes enriched in lipid rafts (fractions 3–5) and depleted of lipid rafts (fractions 9–11). Meg cell line MO7E cells were stimulated by C3a (1 mg/ml) or des-ArgC3a (1 mg/ml) or not stimulated (control). CXCR4 was detected in these membrane fractions by Western blot along with ganglioside M1 (GM1), a marker of lipid rafts. Experiments were performed three times with similar results. A representative blot is shown. (c) Lipid raft formation on Megs. Human Megs were cultured for 6 h in serum-free medium and than non-stimulated (control) or stimulated by C3a (1 mg/ml) or des-ArgC3a (1 mg/ml). The primary antibodies used for raft analysis are cholera toxin b- subunit conjugated with FITC and mouse monoclonal anti-hCXCR4 IgG. The stained cells are examined using a BX51 ﬂuorescence microscope (Olympus America) equipped with a charge-coupled device camera (Olympus America). Separate pictures are merged using Image-Pro Plus software (Media Cybernetics Inc., Silver Spring, MD, USA). Lipid raft formation was analyzed on samples from three different G-CSF-mobilized patients. A representative study is shown. Colocalization of GM1 and CXCR4 is shown as yellow patchy staining.

Leukemia Complement as a new regulator of megakaryopoiesis M Wysoczynski et al 979 a 350 placed in the upper chambers; C3a, des-ArgC3a and SDF-1 control together or alone were placed in the lower chambers. After 48 h, 300 5mM MBCD the number of platelets present in the lower chambers was * 250 * evaluated by FACS after staining with CD41 antibodies Furthermore, in control experiments, platelets were evaluated 200 both morphologically by microscopy and functionally by 150 % of control evaluating the increase in expression of CD62P after stimulation 100 with thrombin as described.12 We found that C3a and 50 des-ArgC3a, if combined with SDF-1, significantly enhanced control SDF-1 C3a SDF-1+ desArgC3a SDF-1+ SDF-1 platelet formation by human Megs (Figure 4c). C3a desArgC3a high As migration of Megs in the BM environment is dependent on 16,29 the secretion of MMP-9 and VEGF-mediated interaction b 1800 * with endothelium in the BM endothelial niche,12 we examined * * 1600 * contol the influence of C3a and des-ArgC3a on MMP-9 and VEGF statin expression in these cells. We found that C3 cleavage fragments 1400 BLEEDING * alone only slightly upregulated the expression of MMP-9 in 1200 normal megakaryocytic cells; however, when C3a or des-ArgC3a 1000 were used together with SDF-1, expression of MMP-9 was 800 upregulated in a synergistic manner B 9-fold (data not shown). PTK (K/uL) 600 Furthermore, C3a and des-ArgC3a increased the production of VEGF by human Megs fivefold ( þ /À1) and sixfold ( þ /À2), 400 respectively (data not shown). 200 0 5 10 15 20 25 30 35 40 45 50 55 60 days of feeding

Molecular analysis of signaling pathways activated Figure 6 Platelet secretion by Megs is lipid raft-dependent. (a) by C3a and des-ArgC3a in normal megakaryoblasts Platelet production in a trans-endothelial migration assay by normal Having demonstrated that both C3a and des-ArgC3a enhances human Megs (control) and Megs that were preincubated for 1 h with SDF-1-dependent migration and adhesion of megakaryocytic 5mM of methyl-b-cyclodextran (MBCD). Platelets were identified and cells, we turned our attention to signaling pathways that could analyzed by FACS. Data are pooled from quadruplicate samples from three independent experiments. *Po0.0001. (b) Mice were fed with regulate these processes, for example, MAPK p42/44.9,16 First, 16 statins (Lipitor – atorvastatin calcium) or not (control) for 24 days. As we observed that, like SDF-1, C3a and des-ArgC3a induced shown Lipitor (750 mg/mouse/day) ameliorates post-bleeding throm- phosphorylation of MAPK p42/44 in normal human Megs bocytosis. Platelet counts in peripheral blood were assayed by (Figure 5a). However, although maximal activation of MAPK employing an automatic analyzer (Hemavet). The data represent the p42/44 after stimulation by C3a occurred during first 1–2 min, average of three independent experiments (6–8 mice were pooled for B each experiment). Data are mean7s.d. *Po0.00001 versus control. des-ArgC3a activated MAPK p42/44 maximally 5 min latter. In our previous investigations on hematopoietic stem/progenitor cells, we reported that C3a and des-ArgC3a may enhance in vivo model of excessive bleeding in mice which had received 20,23 incorporation of CXCR4 into membrane lipid rafts, thereby high doses of statins for 25 days before bleeding and found a allowing this receptor to form an optimal association with significant decrease in platelet counts in response to acute downstream signaling proteins. To determine whether a similar bleeding as a result of the high statin diet (Figure 6b). mechanism plays a role in megakaryocytic lineage cells, we evaluated the incorporation of CXCR4 into membrane lipid rafts in the human Meg line MO7E stimulated with C3a or des-ArgC3a (Figure 5b). As expected, Western blot analysis revealed that Discussion both molecules enhanced incorporation of CXCR4 into membrane lipid rafts. Mounting evidence suggests that complement plays a role in Finally, direct confocal analysis of normal ex vivo expanded hematopoiesis, in particular under stressed situations where it human Megs confirmed that C3a and des-ArgC3a enhance acts to enhance the responsiveness of hematopoietic cells to an incorporation of CXCR4 into membrane lipid rafts in these cells SDF-1 gradient.20,22,23 On the basis of previously published (Figure 5c). observations that SDF-1 is involved in Megs migration/matura- tion11–16 and platelet formation,29 and our data that C3-deficient mice as compared to wt littermates have a delayed recovery of Platelet formation is lipid raft dependent platelet counts after sublethal irradiation or transplantation of C3 cleavage fragments enhance incorporation of CXCR4 into bone marrow cells,23 we hypothesized that C3 cleavage membrane lipid rafts allowing optimal SDF-1 signaling by this fragments are implicated in stress-induced platelet production, receptor.18,20 Thus, we determined whether perturbation of lipid as seen in reactive post-bleeding thrombocytosis. raft formation (by depletion of membrane cholesterol) affects We found that human Megs express functional C3aR and SDF-1-dependent platelet formation.18,20 When human Megs confirmed its expression by platelets.30 Furthermore, we report were preincubated for 1 h in 5 mM MbCD, which depletes here for the first time that C3 cleavage fragments appear to cholesterol from cell membranes,20,23 we found that perturba- enhance SDF-1-dependent platelet production. It has also been tion of lipid raft formation by MbCD significantly inhibited postulated that platelets may play a role in activation and platelet formation dependent on both SDF-1 þ C3a and propagation of the C system.31 Thus, inflammation and SDF-1 þ des-ArgC3a (Figure 6a). thrombosis are two processes that are tightly linked together Finally, we obtained proof of our hypothesis that lipid rafts and these data point to the existence of a direct link between the play an important role in platelet formation when we used an complement system and platelets.

Leukemia Complement as a new regulator of megakaryopoiesis M Wysoczynski et al 980 Our ﬁndings also demonstrate that C3 cleavage fragments, as ing.26 Our data provide a molecular explanation for this effect they do with other hematopoietic cells (e.g., CD34 þ cells), and suggest that statins have potential use in preventing some increase the responsiveness of Megs to an SDF-1 gradi- forms of reactive thrombocytosis. ent.20,22,23 Megs similarly as CD34 þ cells in the presence of Furthermore, as elevated platelet counts and levels of C3a or des-ArgC3a show more robust chemotaxis to SDF-1 and circulating platelet-derived microvesicles in blood are important adhere better to ﬁbrinogen. We also observed that as in other risk factors in cancer,34–38 evidence suggesting that statins may cell types, C3 cleavage fragments increase the incorporation of lower stress-related platelet counts and inhibit shedding of CXCR4 into membrane lipid rafts. CXCR4 þ cells respond much prometastaic platelet-derived microparticles points to their better to SDF-1 stimulation when incorporated into membrane potential use as new adjuvant drugs for the treatment of various lipid rafts,18,20 suggesting that the enhanced responsiveness of malignancies.39 In fact, statins were recently successfully Megs to SDF-1 mediated by C3 cleavage fragments is membrane employed in an animal model to inhibit the metastasis of lipid raft-dependent. prostate cancer.40 Lipid rafts have been reported to be essential to ADP- We found that both C3a and des-ArgC3a activate MAPKp42/44 mediated platelet activation21 and to play a role in the shedding and enhance SDF-1-dependent production of MMP-9 and VEGF of microparticles from platelets.32 Our current work suggests by human Megs, which are crucial for transendothelial that the release of platelets from proplatelets is also lipid raft- migration of Megs and platelet release. Supporting this, recently dependent. In fact, we have demonstrated that inhibition of lipid published studies documented that the incubation of mature raft formation by depletion of cholesterol from cell membranes, Megs with a synthetic MMP inhibitor, 5-phenyl-1,10-phenan- using both methyl-b-cyclodextran (MBCD) and statins, inhibits throline, resulted in the inhibition of platelet formation, platelet formation in vitro as well as in vivo, as shown in a suggesting that in fact the expression of MMPs is critical not model of thrombocytosis induced by excessive bleeding. only for megakaryocyte migration but also for platelet release.29 This latter observation suggests that statins, drugs known to Similarly, an important step in the interaction of Megs with inhibit platelet activation,33 may also ameliorate stress-induced endothelial cells in the BM endothelial niche is secretion of thrombocytosis. Supporting this contention is a recent report VEGF by Megs,12 which in turn upregulates the expression of that a high dose of statins effectively reduced postoperative E-selectin on endothelium, a process crucial for the formation of thrombocytosis in patients after coronary artery bypass graft- proplatelets by these cells.12 Giving greater credence to this, we

stress C3 cleavage Platelets

adhesion – interaction

C3a desArgC3a with endothelium

Endothelium SDF-1, VEGF

Meg – endothelial niche (Rafi S. J Exp Med. 1998) chemotaxis (Platelets shedding)

Megakaryocyte

chemotaxis

Meg – hematopoietic niche (Meg development) CFU -Meg Megakaryoblast

TPO SDF-1

Osteoblasts/Stromal fibroblasts

Figure 7 C3a and des-ArgC3a modulate SDF-1-dependent attraction of Megs from bone niche to endothelial niche and platelet production. The concept of developmental allocation of Megs from the bone niche to the endothelial niche has been proposed by S. Raﬁi et al.17 This allocation is tightly regulated by an SDF-1 gradient and we envision that during stressed situations (e.g., chronic infection, hemorrhage), C3 cleavage fragments are important coregulators of this process. Accordingly, we propose that C3 is cleaved/activated in BM and as result of this C3 cleavage fragments (C3a and des-ArgC3a) modulate/enhance the responsiveness of maturing Megs to an SDF-1 gradient. As shown, Megs are chemoattracted by SDF-1 expressed by marrow endothelial cells and reallocate from the bone-marrow niche to the endothelial niche where proplatelets are formed and platelets shed from the Meg surface. At the molecular level, we envision that C3 cleavage fragments sensitize responsiveness of Megs to SDF-1 by enhancing its inclusion into membrane lipid rafts. CXCR4 if included into membrane lipid rafts regulates Megs migration more efﬁciently as well as adhesion to endothelium and secretion of MMP-9 and VEGF. All of these processes are crucial for proplatelet formation and platelet shedding and are enhanced by C3 cleavage fragments during reactive thrombocytosis.

Leukemia Complement as a new regulator of megakaryopoiesis M Wysoczynski et al 981 noticed that both C3 cleavage fragments increase expression of 8 Neben TY, Loebelenz J, Hayes L, McCarthy K, Stoudemire J, VEGF in human Megs. Schaub R et al. Recombinant human interleukin-11 stimulates On the basis of a proposal by Rafii et al. regarding maturation- megakaryocytopoiesis and increases peripheral platelets in normal related allocation of Megs progenitors from the bone niche into and splenectomized mice. Blood 1993; 81: 901–908. 17 9 Majka M, Ratajczak J, Villaire G, Kubiczek K, Marquez LA, the endothelial niche, we postulate the following scenario Janowska-Wieczorek A et al. Thrombopoietin, but not cytokines for enhanced platelet production during excessive-bleeding binding to gp130 protein-coupled receptors, activates MAPKp42/ (Figure 7), in which during stress situations (e.g., chronic 44, AKT, and STAT proteins in normal human CD34+ cells, infection, hemorrhage), C3 cleavage fragments are important megakaryocytes, and platelets. Exp Hematol 2002; 30: 751–760. coregulators of SDF-1–CXCR4 signaling. As shown in Figure 7, 10 Gainsford T, Nandurkar H, Metcalf D, Robb L, Begley CG, Megs are chemoattracted from the bone niche to the endothelial Alexander WS. The residual megakaryocyte and platelet production in c-mpl-deficient mice is not dependent on the actions of niche by SDF-1 expressed by marrow endothelial cells. In the interleukin-6, interleukin-11, or leukemia inhibitory factor. Blood endothelial niche, proplatelets are formed and subsequently 2000; 95: 528–534. 17 shed platelets. We envision that C3 cleavage fragments 11 Wang JF, Liu ZY, Groopman JE. The alpha-chemokine receptor sensitize the responsiveness of Megs to SDF-1 by enhancing CXCR4 is expressed on the megakaryocytic lineage from its inclusion into membrane lipid rafts. Further, CXCR4 progenitor to platelets and modulates migration and adhesion. incorporated into membrane lipid rafts results in more efficient Blood 1998; 92: 756–764. migration of Megs, adhesion to endothelium and secretion of 12 Hamada T, Mohle R, Hesselgesser J, Hoxie J, Nachman RL, Moore MA et al. Transendothelial migration of megakaryocytes in MMP-9 and VEGF, all processes that are crucial for proplatelet response to stromal cell-derived factor 1 (SDF-1) enhances platelet formation and platelet shedding. formation. J Exp Med 1998; 188: 539–548. In conclusion, both C3a and des-ArgC3a clearly have a role in 13 Kowalska MA, Ratajczak J, Hoxie J, Brass LF, Gewirtz A, Poncz M modulating the responsiveness of Megs to an SDF-1 gradient. et al. Megakaryocyte precursors, megakaryocytes and platelets The crosstalk between C3aR- and CXCR4-G-protein-coupled express the HIV co-receptor CXCR4 on their surface: determina- receptors we identified here with respect to Megs plays an tion of response to stromal-derived factor-1 by megakaryocytes and platelets. Br J Haematol 1999; 104: 220–229. important and previously unrecognized role in stress and 14 Riviere C, Subra F, Cohen-Solal K, Cordette-Lagarde V, Letestu R, inflammation-dependent platelet formation. On the basis of this Auclair C et al. Phenotypic and functional evidence for the new information, we foresee that inhibitors of the CXCR4–SDF-1 expression of CXCR4 receptor during megakaryocytopoiesis. axis have clinical potential as drugs to control certain forms of Blood 1999; 93: 1511–1523. thrombocytosis (e.g., as described in postoperative thrombocy- 15 Hodohara K, Fujii N, Yamamoto N, Kaushansky K. Stromal cell- tosis in patients after coronary bypass grafting).26 We are aware derived factor-1 (SDF-1) acts together with thrombopoietin to enhance the development of megakaryocytic progenitor cells that statins show several pleiotropic effects on platelets (e.g., (CFU-MK). Blood 2000; 95: 769–775. inhibit their activation). However, given that platelet formation 16 Majka M, Janowska-Wieczorek A, Ratajczak J, Kowalska MA, is dependent on membrane lipid formation as we have shown Vilaire G, Pan ZK et al. Stromal-derived factor 1 and thrombo- here, statins, which perturb lipid raft formation, may ameliorate poietin regulate distinct aspects of human megakaryopoiesis. stress-related thrombocytosis and could be considered as Blood 2000; 96: 4142–4151. potential drugs to prevent bleeding-related post-trauma or 17 Avecilla ST, Hattori K, Heissig B, Tejada R, Liao F, Shido K et al. postoperative thrombocytosis. Chemokine-mediated interaction of hematopoietic progenitors with the bone marrow vascular niche is required for thrombopoiesis. Nat Med 2004; 10: 64–71. 18 Nguyen DH, Taub D. CXCR4 function requires membrane Acknowledgements cholesterol: implications for HIV infection. J Immunol 2002; 168: 4121–4126. This work was supported by an NIH Grant R01 DK074720-01 19 Gu Y, Filippi MD, Cancelas JA, Siefring JE, Williams EP, Jasti AC to MZR. et al. Hematopoietic cell regulation by Rac1 and Rac2 guanosine triphosphatases. Science 2003; 302: 445–449. 20 Wysoczynski M, Reca R, Ratajczak J, Kucia M, Shirvaikar N, Honczarenko M et al. Incorporation of CXCR4 into membrane References lipid rafts primes homing-related responses of hematopoietic stem/ progenitor cells to an SDF-1 gradient. Blood 2005; 105: 40–48. 1 Long MW, Hoffman R (eds). Hematology: Basic Principles and 21 Quinton TM, Kim S, Jin J, Kunapuli SP. Lipid rafts are required in Practice. Churchill Livingstone: New York, 1999, p245. Galpha(i) signaling downstream of the P2Y12 receptor during 2 Gewirtz AM. Megakaryopoiesis: state of the art. Thromb Haemost ADP-mediated platelet activation. J Thromb Haemost 2005; 3: 1995; 74: 204–209. 1036–1041. 3 Gurney AL, Carver-Moore K, de Sauvage FJ, Moore MW. 22 Reca R, Mastellos D, Majka M, Marquez L, Ratajczak J, Franchini S Thrombocytopenia in c-mpl-deficient mice. Science 1994; 265: et al. Functional receptor for C3a anaphylatoxin is expressed by 1445–1447. normal hematopoietic stem/progenitor cells, and C3a enhances their 4 Gurney AL, de Sauvage FJ. Dissection of c-Mpl and thrombopoie- homing-related responses to SDF-1. Blood 2003; 101: 3784–3793. tin function: studies of knockout mice and receptor signal 23 Ratajczak MZ, Reca R, Wysoczynski M, Kucia M, Baran JT, transduction. Stem Cells 1996; 14 (Suppl 1): 116–123. Allendorf DJ et al. Transplantation studies in C3-deficient animals 5 Bunting S, Widmer R, Lipari T, Rangell L, Steinmetz H, Carver- reveal a novel role of the third complement component (C3) in Moore K et al. Normal platelets and megakaryocytes are produced engraftment of bone marrow cells. Leukemia 2004; 18: 1482– in vivo in the absence of thrombopoietin. Blood 1997; 90: 1490. 3423–3429. 24 Ratajczak MZ, Kucia M, Reca R, Majka M, Janowska-Wieczorek 6 Sasaki H, Hirabayashi Y, Ishibashi T, Inoue T, Matsuda M, Kai S A, Ratajczak J. Stem cell plasticity revisited: CXCR4-positive cells et al. Effects of erythropoietin, IL-3, IL-6 and LIF on a murine expressing mRNA for early muscle, liver and neural cells ‘hide out’ megakaryoblastic cell line: growth enhancement and expression of in the bone marrow. Leukemia 2004; 18: 29–40. receptor mRNAs. Leuk Res 1995; 19: 95–102. 25 Schafer AI. Thrombocytosis and thrombocythemia. Blood Rev 7 Burstein SA, Mei RL, Henthorn J, Friese P, Turner K. Leukemia 2001; 15: 159–166. inhibitory factor and interleukin-11 promote maturation of murine 26 Christenson JT. Preoperative lipid-control with simvastatin reduces and human megakaryocytes in vitro. J Cell Physiol 1992; 153: the risk of postoperative thrombocytosis and thrombotic complica- 305–312. tions following CABG. Eur J Cardiothorac Surg 1999; 15: 394–399.

Leukemia Complement as a new regulator of megakaryopoiesis M Wysoczynski et al 982 27 Goldfarb RD, Parrillo JE. Complement. Crit Care Med 2005; 33: patients with esophageal carcinoma. J Am Coll Surg 2004; 198: 482–484. 737–741. 28 Kalant D, Cain SA, Maslowska M, Sniderman AD, Cianﬂone K, 35 Aoe K, Hiraki A, Yamazaki K, Nakamura Y, Murakami T, Maeda T Monk PN. The chemoattractant receptor-like protein C5L2 binds et al. Thrombocytosis as a useful prognostic indicator in patients the C3a des-Arg77/acylation-stimulating protein. J Biol Chem with lung cancer. Respiration 2004; 71: 170–173. 2003; 278: 11123–11129. 36 Janowska-Wieczorek A, Wysoczynski M, Kijowski J, Marquez- 29 Lane WJ, Dias S, Hattori K, Heissig B, Choy M, Rabbany SY et al. Curtis L, Machalinski B, Ratajczak J et al. Microvesicles derived Stromal-derived factor 1-induced megakaryocyte migration and from activated platelets induce metastasis and angiogenesis in lung platelet production is dependent on matrix metalloproteinases. cancer. Int J Cancer 2005; 113: 752–760. Blood 2000; 96: 4152–4159. 37 Ratajczak J, Wysoczynski M, Hayek F, Janowska-Wieczorek A, 30 Polley MJ, Nachman RL. Human platelet activation by C3a and Ratajczak MZ. Membrane-derived microvesicles (MV): important C3a des-arg. J Exp Med 1983; 158: 603–615. and underappreciated mediators of cell to cell communication. 31 Del Conde I, Cruz MA, Zhang H, Lopez JA, Afshar-Kharghan V. Leukemia 2006; 20: 1487–1495. Platelet activation leads to activation and propagation of the 38 Ratajczak J, Miekus K, Kucia M, Zhang J, Reca R, Dvorak P et al. complement system. J Exp Med 2005; 201: 871–879. Embryonic stem cell-derived microvesicles reprogram hemato- 32 Biro E, Akkerman JW, Hoek FJ, Gorter G, Pronk LM, Sturk A et al. poietic progenitors: evidence for horizontal transfer of mRNA and The phospholipid composition and cholesterol content of platelet- protein delivery. Leukemia 2006; 20: 847–856. derived microparticles: a comparison with platelet membrane 39 Demierre MF, Higgins PD, Gruber SB, Hawk E, Lippman SM. fractions. J Thromb Haemost 2005; 3: 2754–2763. Statins and cancer prevention. Nat Rev Cancer 2005; 5: 930–942. 33 Colli S, Werba JP, Tremoli E. Statins in atherothrombosis. Semin 40 Zhuang L, Kim J, Adam RM, Solomon KR, Freeman MR. Vasc Med 2004; 4: 407–415. Cholesterol targeting alters lipid raft composition and cell survival 34 Shimada H, Oohira G, Okazumi S, Matsubara H, Nabeya Y, in prostate cancer cells and xenografts. J Clin Invest 2005; 115: Hayashi H et al. Thrombocytosis associated with poor prognosis in 959–968.

Leukemia