MAL/SRF complex is involved in platelet formation and megakaryocyte migration by regulating MYL9 (MLC2) and MMP9

Serum response factor (SRF) is a widely expressed transcription factor required for the expression of immediate early, muscle-specific, and cytoskeletal genes.^1-4  SRF contains a MADS domain that mediates homodimerization and DNA binding, and that allows recruitment of transcriptional cofactors. SRF binds to a CArG box present in promoter/enhancer regions of SRF-regulated genes.⁵  Depending on cell lines, different extracellular stimuli activate SRF through 2 main signaling pathways: the MAP-kinase pathway through members of the ternary complex factor (TCF)^6,7  and the small GTPases pathway through the Rho family⁸  members regulating the myocardin-related transcription factors (MRTFs). The Rho-actin signaling pathway^9-12  stimulates SRF by 2 ubiquitous MRTFs, megakaryoblastic leukemia 1 (MAL; MKL1, MRTF-A, BSAC) and MAL16 (MKL2, MRTF-B).

MAL was initially identified in acute megakaryoblastic leukemia (AMKL, M7) as a chromosome 22 encoded protein fused in 3′ with RNA-binding motif protein 15 (RBM15; OTT) located on chromosome 1.^13-15  The translocation t(1,22)(p13;q13) leads to the in-frame fusion of the quasi-totality of OTT/RBM15 to the MAL gene. The OTT-MAL fusion protein is restricted to AMKL occurring de novo in infancy,^16,17  in children older than 1 year or, occasionally, in Down syndrome patients.^18,19 

The subcellular localization of MAL is regulated through its association with globular actin by its RPEL motifs in the N-terminal region. The modification of the actin treadmilling by the Rho pathway results in the nuclear accumulation of MAL, and actin binding may also control MAL transcriptional activity.^12,20  The 3 best-characterized Rho GTPase family members are Rho, Rac, and Cdc42. They regulate assembly and disassembly of actin cytoskeleton in response to a variety of extracellular stimuli. Rho regulates stress fibers and focal adhesion assembly, Cdc42 promotes filopodia, and Rac promotes lamellipodia formation. Different effectors of these 3 signaling pathways are involved in promoting actin polymerization. Except their direct role in actin cytoskeleton remodeling, Rho, Rac, and Cdc42 also control cell cycle, including G1 progression and cytokinesis, cell contraction, polarity and migration, vesicle transport (endocytosis, exocytosis), enzyme activation, and gene transcription via JNK, p38, nuclear factor-κB, and SRF.^21,22  Hill et al reported that Rac1 and Cdc42 activate SRF in a RhoA-independent manner⁸ ; however, several studies showed a crosslink between these pathways in SRF regulation.^23,24  In megakaryocytes (MKs), the platelet precursors, reorganization of actin cytoskeleton plays a crucial role not only in MK adhesion and migration²⁵  but also at different stages during maturation, ie, during endomitosis,²⁶  demarcation membrane formation, proplatelet formation, and platelet shedding.^27-32  Recently, it has been shown that thrombopoietin (TPO) better stimulates Rho activity in immature than in mature MKs and that the Rho/ROCK pathway is a negative regulator of proplatelet formation. On the other hand, the use of constitutively active Cdc42 led to increased proplatelet formation^33,34  (Y.C., unpublished data, February 2005). Interestingly, a recent study strongly supports the hypothesis that a local inhibition of the Rho/ROCK pathway is at the origin of MK ploidization by decreasing F-actin and myosin accumulation at midzone during cytokinesis.²⁶ 

The involvement of MAL in the transformation of megakaryoblasts and recent experimental evidence³⁵  strongly suggested a role in MK maturation. We therefore investigated its role in maturation, adhesion, migration, and proplatelet formation of MKs derived from neonate and adult CD34⁺ cells. The functional studies coupled to gene profiling after MAL depletion in MK revealed that the MAL/SRF complex regulates positively the known cytoskeletal genes and 2 new targets: metalloproteinase 9 (MMP-9) important for MK migration and MLC2 (MYL9) involved in platelet formation.

Cord blood and cytapheresis samples were obtained after informed consent in accordance with the Declaration of Helsinki. Approval was obtained from the Assistance Publique des Hôpitaux de Paris. CD34⁺ cells were isolated using an immunomagnetic beads technique³⁶  and grown in serum-free medium as previously reported.³⁷  The culture medium was supplemented with recombinant human thrombopoietin (TPO; 10 ng/mL) or with a cytokine cocktail containing TPO (10 ng/mL), interleukin-3 (IL-3; 100 U/mL; Immunex), interleukin-6 (IL-6; 10 ng/mL; Tebu), stem cell factor (SCF; 25 ng/mL), and fetal liver tyrosine kinase 3 ligand (FLT3-L; 1 ng/mL; Amgen).

For MAL, the published small interfering RNA (siRNA) sequence was used.⁹  For MYL9, the siRNA sequence was designed using siSearch software (http://www.dharmacon.com/DesignCenter/DesignCenterPage.aspx). Small hairpin sequences were created as previously described.^38,39  Oligonucleotide MAL and MYL9 short hairpins (sh; supplemental Table 1A, available on the Blood website; see the Supplemental Materials link at the top of the online article) were synthesized (Eurogentec) and inserted into a pBlue Script vector containing the human H1 promoter. H1-shRNA specific for MAL (shMAL), H1-shMYL9, or H1-SCR (scramble control sequence) cassettes were inserted into a lentiviral vector (pRRLsin-PGK-eGFP-WPRE, Genethon). Lentiviral stocks were prepared and stored, and concentration was determined, as previously described.⁴⁰ 

CD34⁺ cells (10⁵/mL) were prestimulated for 24 hours with TPO, IL-3, SCF, and FLT3-L. Lentiviral particles were added at a concentration corresponding to 125 ng viral p24/100 μL for 12 hours, followed by a second transduction. Cells were then cultured in the presence of TPO alone.

Cells were incubated for 30 minutes at 4°C with an allophycocyanin-conjugated anti-CD41a monoclonal antibody and a phycoerythrin-conjugated anti-CD42a monoclonal antibody (BD Biosciences PharMingen). After washing, cells were incubated for 2 hours in 0.01 mM Hoechst 33342 (Sigma-Aldrich) at 37°C. To study endogenous MAL expression in the different differentiation classes, cells were sorted into CD34⁺41⁻; CD34⁻41⁺CD42⁻; CD34⁻41⁺CD42^low; CD34⁻41⁺CD42^high cell fractions for cord blood cultures, and into CD34⁺41⁻; CD34⁺41⁺CD42⁻; CD34⁺41⁺CD42 ⁺; CD34⁻41⁺CD42 ⁺cell fractions for cytapheresis cultures. To study the ploidy level, CD41⁺CD42⁺ cells stained with Hoechst were analyzed on a LSRII (BD Biosciences) with the CellQuest software package (BD Biosciences).

Primers and internal probes for quantitative reverse-transcribed polymerase chain reaction (RT-PCR) are listed in supplemental Table 1B. PCR was carried out in the ABI Prism GeneAmp 7500 Sequence Detection System (Applied Biosystems) using TaqMan Universal PCR Master Mix (ABI) containing the specific primers (1.2 mM) and the specific probe (0.1 mM). Expression levels of all genes were expressed relatively to hypoxanthine phosphoribosyl transferase (HPRT).

Western blot analyses were performed as previously reported.⁴¹  The following antibodies were used: anti-MRTF-A (C-19), anti-HDAC-1 (H-51; Santa Cruz Biotechnology), and anti-HSC70 (Stressgen). In addition of anti-HDAC and anti-HSC70 staining, coloration with Ponceau S solution (Sigma-Aldrich) after transfer was used to control the quantity of proteins.

CD41⁺CD42⁺GFP⁺ MK were sorted at day 8 of culture and plated in 96-well plates at a concentration of 2000 cells/well in serum-free medium containing TPO, and quantification of proplatelet MK was performed as previously described.⁴²  Images were obtained using a Zeiss inverted microscope at a magnification of 40×, and the proplatelet network area was measured using the Axio Vision 4.6 software.

The 440-bp and 550-bp promoter regions upper transcription initiation of MMP9 and MYL9, respectively, were amplified (primer listed in supplemental Table 1C) and subcloned into the pLuc-MCS vector (Stratagene). MAL cDNA was cloned into the vector pMegix (gift of Dr Morgenstern from Millennium Pharmaceuticals). HEK293T cells were transfected with pLuc constructs (pLuc_MCS, pLuc_MMP9, pLuc_MYL9, pLuc_SRE, or pLuc_FLT3), a Renilla luciferase vector (for normalization) and MAL expression vector using Lipofectamine 2000 (Invitrogen). Cells were cultured in 12-well plates and transfected with 50 to 200 ng DNA. The total amount of DNA was kept constant by the inclusion of appropriate amounts of empty vectors. Cells were lysed with Passive Lysis Buffer 48 hours after transfection, and the luciferase activity was evaluated using the Dual Luciferase Reporter Assay System (Promega).

Chromatin immunoprecipitation (ChIP) assays were performed with a ChIP assay kit (Upstate Biotechnology) using the anti-SRF antibody (clone H-65; Santa Cruz Biotechnology). Assays were performed using chromatin prepared from human MKs. Quantification of precipitated DNA fragments was carried out on an ABI PRISM 7500 sequence detection system using SYBR green (Applied Biosystems) in duplicate. Primers for MYL9, MMP9, and THSB1 are described in supplemental Table 1D. Fold enrichment is expressed as the ratio of target antibody signal to IgG signal calculated by extrapolation from a standard curve of DNA dilutions.

Cell migration was assessed with BD Bio-Coat Matrigel Invasion Chamber (BD Biosciences) according to the protocol provided by the manufacturer. Briefly, MKs were washed and prepared to a final concentration of 10⁶ cells/mL. MK aliquots (100 μL) were seeded into each 8-μm pore transwell inserts, which were placed into the wells of a 24-well plate and incubated for 22 hours at 37°C. The lower compartment contained 600 μL serum-free media with or without 100 ng to stromal cell-derived factor 1α (SDF-1α; Abcys). The percentage of migrated CD41⁺CD42⁺GFP⁺ MK was determined by flow cytometry.

Slides were coated overnight at 4°C with convulxin (kindly provided by M. Jandrot-Perrus, Hôpital Bichat, France) at 15 μg/mL in 1× PBS or with collagen type I (Nycomed) at 50 μg/mL. pRRL-SCR or pRRL-shMAL GFP⁺-sorted cells were seeded on coated slides for 2 hours (convulxin) or 1 hour (collagen type I) at 37°C. After a gentle wash, adherent cells were fixed in 2% paraformaldehyde for 10 minutes, permeabilized with 0.1% Triton X-100 for 5 minutes, and washed with 1× PBS for 5 minutes. Primary antibodies were subsequently applied to the slides at a concentration of 1 μg/mL and incubated for 1 hour at room temperature. Slides were then washed 3 times with 1× PBS for 5 minutes before and after application of the secondary antibody (20 μg/mL) for 30 minutes at room temperature. Slides were mounted using either Vectashield with 4,6-diamidino-2-phenylindole (Invitrogen) and observed under an epifluorescence microscope (Nikon Eclipse 600) equipped with a 60× objective (Nikon) or TOTO-3 iodide (Molecular Probes) and observed under a Zeiss LSM 510 laser scanning microscope (Carl Zeiss) with a 63×/1.4 numeric aperture oil objective. Images were processed using the Adobe Photoshop 6.0 and LSM 5 Image Browser software, respectively.

Primary antibodies were: anti-MRTF-A (clone C-19; Santa Cruz Biotechnology), anti-phalloidin–tetramethylrhodamine isothiocyanate (TRITC) (P-1951; Sigma-Aldrich). The secondary antibody was anti-goat TRITC (Jackson ImmunoResearch).

Samples were washed in 1× PBS, then fixed in 1.5% glutaraldehyde for 1 hour, and washed 3 times in 0.1 M phosphate buffer, pH 7.4. For morphologic examination, samples were postfixed in 1% osmic acid, dehydrated in ethanol, and embedded in Epon by standard methods. Samples were counterstained and were observed on a Philips Tecnai electron microscope.

We used Agilent long (60-bp) oligonucleotide microarrays and the dual-color analysis method in which probes from specimens and from the reference are differentially labeled with Cyanine 5 and Cyanine 3, as described previously.⁴³  We performed a set of 4 dye swap experiments to compare RNAs obtained from MKs derived either from cytapheresis or cord blood CD34⁺ cells after transduction by shMAL. GFP⁺ cells were sorted at day 9 of culture. The purity of MKs derived from cytapheresis was approximately 97% (supplemental Figure 1). RNAs from scramble-lentivirus–transduced cells were used as the reference. From each of the 4 combined experiments, signatures (supplemental Table 2) were obtained by an analysis of variance test with P value of 10⁻³, and annotated with updated databases as Entrez-Gene. Data were analyzed using Resolver software, and gene selection was obtained by an analysis of variance test with P < 10⁻³ value.

All data obtained by microarray analyses have been submitted to ArrayExpress⁴⁴  with the accession number E-TABM-640.

Different studies suggest a redundant role for 2 ubiquitous MRTFs, MAL and MAL16. Therefore, we first studied their relative expression in culture of MKs derived from CD34⁺ cells isolated from cord blood or cytapheresis and grown for 9 days in the presence of TPO. MAL mRNA level was approximately 7-fold higher than MAL16 mRNA in both types of MKs, taking HPRT as an internal control (Figure 1A). Because of the high expression of MAL and its direct involvement in megakaryoblastic transformation, we focused our study only on the role of MAL in megakaryopoiesis. To determine whether MAL expression varies through MK maturation during ontogenesis, CD34⁺ cells isolated from cord blood and cytapheresis were cultured in the presence of TPO and sorted at day 6 according to the expression of 3 differentiation markers (“Cell sorting, flow cytometric, and cytologic analysis”). No difference in MAL expression was found during ontogenesis. MAL transcripts remained stable during the first stages of differentiation in both neonate and adult-derived MKs but clearly increased in the most mature fraction (CD34⁻CD41⁺CD42⁺; Figure 1B). We next investigated the protein expression at different days of culture in the presence of TPO (until day 12 for cord blood and until day 9 for cytapheresis), as well as in platelets isolated either from cord blood or cytapheresis samples. Consistent with MAL mRNA level, the protein level increased continuously until day 12 of culture for cord blood-derived MKs and until day 9 of culture for cytapheresis-derived MKs (∼ 65% of CD41⁺CD42⁺ MKs; Figure 1C). In our culture conditions, proplatelet-forming MKs first appeared at day 12 for adult MKs and at day 15 for neonate MKs. Thereafter, MAL protein level decreased with the appearance of proplatelet-forming MKs and was undetectable in platelets (data not shown).

As previously described,^8,12  we observed that MAL localized in the cytoplasm of serum-starved NIH3T3 cells and accumulated in the nucleus after serum stimulation (Figure 2A). This is mediated through small GTPase activation.^8,12  To examine the effect of RhoA and Cdc42 activation on MAL localization in cord blood and cytapheresis-derived MKs, MKs were cultured for 9 days in serum-free medium supplemented by TPO and then allowed to adhere onto different extracellular matrix. After 10 minutes on polylysine, a substrate that does not allow Rho or Cdc42 activation,³³  MAL remained cytoplasmic (Figure 2B). After 2 hours on collagen type I, a substrate that activates principally Rho/ROCK leading to fiber stress formation,³³  MAL accumulated in the MK nucleus. After 2 hours onto convulxin, a substrate inducing mainly filopodia formation through activation of the Cdc42 pathway,³³  MAL also accumulated in the MK nucleus (Figure 2B). The nuclear accumulation of MAL after Rho and Cdc42 pathway activation is observed in MKs derived either from cord blood or cytapheresis CD34⁺ cells.

To investigate further the role of MAL in MKs, we developed a lentivirus encoding an shMAL under the H1 promoter. The vector efficiency was first tested in the MO7e leukemic cell line. Two days after transduction, MAL level was determined by RT-PCR and Western blot in cells sorted on the expression of GFP. The shMAL reduced MAL transcripts approximately 13-fold compared with the control shRNA (Figure 3A left panel), and the protein expression was similarly decreased (Figure 3A right panel). Then, CD34⁺ cells obtained from cord blood or cytapheresis were prestimulated in culture by IL-3, FLT3-L, SCF, and TPO and transduced at days 1 and 2 with the control shRNA or shMAL lentiviruses. Thereafter, cells were cultured in TPO alone. At day 9 of culture, GFP⁺ cells were sorted. The MK purity was approximately 90% in both cases (data not shown). As shown in Figure 3B, shMAL led to a major decrease of MAL protein level compared with the control shRNA.

Recently, it has been reported that inhibition of the Rho/Rock pathway leads to an increased MK ploidy level.²⁶  Therefore, we examined whether the Rho effector MAL could be involved in the endomitotic process. To study the effect of MAL repression on MK ploidization, CD34⁺ cells isolated from cytapheresis were transduced at days 1 and 2 of culture and analyzed at day 9. The ploidy level of transduced (GFP⁺) cells was analyzed in the gate of mature CD41⁺CD42⁺ MKs. As illustrated in Figure 3C, MAL knockdown resulted in a moderate (21.3% ± 9%) increase in mean ploidy level that did not reach significance (n = 3, P ≤ .09), compared with the untransduced population.

Then, we analyzed the effect of MAL knockdown on maturation of MKs issued from cord blood and cytapheresis samples. Cells were transduced as at days 1 and 2, and GFP⁺ cells were analyzed at day 7 of culture for the expression of 2 MK markers, CD41 and CD42. As shown in Figure 3D, the percentage of CD41⁺CD42⁺ mature MKs slightly increased in cytapheresis samples (top panel, 26% ± 13.6%, n = 4, P ≤ .004) but decreased in cord blood samples (bottom panel, 16.8% ± 1.5%, n = 3, P ≤ .08) after MAL repression. The minor difference in the percentage of CD41⁺CD42⁺ MK in both cases after MAL repression and the slight increase in ploidy level suggest that MAL does not play an essential role in MK ploidization and during the early differentiation steps.

MAL and MAL16 are necessary for the formation of stress fibers and focal adhesion in NIH3T3 cells.⁴  Here, we investigated whether MAL had a similar role in MK. Control shRNA- or shMAL-transduced MKs were sorted at day 8 of culture on GFP⁺ and allowed to adhere for 1 hour on collagen type I or for 2 hours on convulxin. As illustrated in Figure 4A, the percentage of filopodia and stress fibers forming MKs was markedly decreased after MAL repression (n = 2, P ≤ .016; n = 2, P ≤ .05, respectively). These results suggested that, like in NIH3T3 cells, MAL acts as a mediator of actin-polymerized structures by regulating transcription of several cytoskeletal genes.

Electron microscopy studies revealed structural abnormalities of maturing MK after MAL knockdown (Figure 4B). Compared with the control shRNA (Figure 4Bi), the cytoplasm was disorganized with numerous vacuoles (Figure 4Bii-iii); the demarcation membrane system was heterogeneously dilated and abnormally distributed with its presence at the periphery of MKs (Figure 4Bii-iii); moreover, the α granules were scattered heterogeneously in the cytoplasm (Figure 4Bii).

To test whether MAL was involved in the process of proplatelet formation, control shRNA- and shMAL-transduced MKs were cultured in a liquid serum-free medium containing TPO. CD41⁺CD42⁺GFP⁺ MKs were sorted at day 9 and seeded in 96-well plates at 2 × 10³ cells per well. At day 13 of culture, the percentage of proplatelet-forming MKs was 68% reduced in MAL knockdown MKs derived both from cytapheresis (n = 3, P ≤ .003) and cord blood (n = 2, P ≤ .1; Figure 4C). Moreover, we measured the total surface area of the proplatelet network in cells transduced with control or shMAL and found that it was decreased 3-fold after MAL inhibition (Figure 4D). Ten cells were analyzed in each group (P ≤ .008). These results suggest that the MAL/SRF complex plays a major role in terminal maturation, proplatelet formation, and adhesion of MK by transcriptionally regulating genes implicated in these processes.

MAL is thought to regulate transcription through its interaction with SRF. To characterize MAL/SRF target genes in MKs derived from cord blood or cytapheresis samples, we compared control shRNA- and shMAL-transduced MKs by microarray analyses. CD34⁺ cells from cord blood or cytapheresis were transduced as described in the previous paragraph, cultured in the presence of TPO alone after transduction and GFP⁺ cells were sorted at day 9 of culture. The purity of MKs was approximately 97%, as illustrated for cytapheresis-derived MKs in supplemental Figure 1. Gene expression profiles were obtained using Agilent Human Whole Genome 44K oligonucleotide arrays allowing a direct comparison of 2 populations labeled by 2 different dyes (Cy3 and Cy5). Cell populations transduced by the control lentivirus were used as reference for transduction-induced changes. The signal from 1929 and 1436 probe sets (corresponding to 1037 and 773 genes, respectively) was significantly deregulated by the expression of the shMAL in MKs derived from cytapheresis and cord blood, respectively (Figure 5A). A total of 613 probe sets were common between cytapheresis- and cord blood–derived MKs (supplemental Table 2). The signal of 589 probe sets was found to be modified similarly in both cell types, and the expression of 24 probe set genes showed an opposite variation in cord blood and cytapheresis MKs (supplemental Table 3). Figure 5A shows the 20 most down-regulated genes found in MKs derived from cytapheresis and cord blood. They are listed according to the fold change observed in the experiment using cytapheresis MKs.

To confirm the microarray data, 12 genes were selected, most with a key role in cytoskeleton organization. Three of them (zyxin, profilin 1, and N-wasp) play a major role in actin polymerization and 3 others (β-actin, vinculin, and tropomyosin 2) are major structural proteins of the cytoskeleton and thus could be implicated in the defects in fiber stress and filopodia formation seen in MAL knockdown MKs. Gene expression was analyzed by quantitative RT-PCR (Figure 5B). For some genes (zyxin, thrombospondin, β-actin, and vinculin), a discrepancy was found between microarray (Figure 5Bi) and RT-PCR analyses (Figure 5Bii) that could be explained by methods sensitivity. Indeed, saturation of highly expressed genes may lead to their exclusion in microarray analyses.

We next focused our efforts on MMP9 and MYL9 (the myosin light chain associated with MYH9), 2 genes with a specific role in megakaryopoiesis.

The MAL/SRF complex binds DNA at specific sites through the SRF DNA-binding domain. Using the ChipMapper software, we identified putative SRF-binding sites in both MMP9 and MYL9 promoters. To determine whether MMP9 and MYL9 are targets of the MAL/SRF complex, we cloned 420 bp of the MMP9 promoter (including one SRF-binding site) and 650 bp of the MYL9 promoter (including 3 SRF-binding sites) upstream of the firefly luciferase gene (Figure 6A) and tested the ability of MAL to modulate expression in transient reporter assays using HEK293T cells. Transcription from both MMP9 and MYL9 promoters was highly stimulated by MAL coexpression (Figure 6B). The baseline activity of MMP9 and MYL9 promoters compared with minimal promoter with or without SRF consensus-binding sites (pLUC_SRE and pLuc_MCS, respectively) or to a control promoter region without SRF consensus binding site cloned into the same luciferase vector (pLuc_FLT3) is shown in supplemental Figure 2.

To demonstrate that SRF binds to the MMP9 and MYL9 promoters in vivo, ChIP assays were performed using human MKs derived in culture from CD34⁺ cells. Thrombospondin, a previously described target of SRF,⁴⁵  was used as a positive control for a ChIP experiment. Quantitative RT-PCR with primers amplifying the regions encompassing the predicted C (Figure 6A) and D (supplemental primer Table 1d) SRF-binding sites in MMP9 promoter and 3 SRF-binding sites in MYL9 promoter (Figure 6A) were performed. ChIP experiments revealed a high MYL9 promoter region enrichment and a slight but significant MMP9 promoter region enrichment using SRF antibody (Figure 6C). Our results confirmed that MMP9 and MYL9 are direct targets for MAL/SRF complex in MK.

To confirm that the drop in MYL9 expression could be responsible for the defect in proplatelet formation observed on the knockdown of MAL, we constructed a lentivirus expressing an shRNA specific for MYL9 (shMYL9). The shMYL9 was expressed in primary MKs derived from cytapheresis, as described for MAL knockdown. Two independent experiments showed a 50% decrease in the MYL9 transcript level (n = 2, P ≤ .08; Figure 7A). At day 13 of culture, knockdown of MYL9 elicited an approximately 35% decrease in the percentage of proplatelet-forming MKs (n = 4, P ≤ .003; Figure 7B). Moreover, the total surface area of the proplatelet network in cells transduced with shMYL9 was decreased approximately 2-fold after shMYL9 transduction compared with control-transduced cells. Ten cells were analyzed in each group (P ≤ .009; Figure 7C).

MYL9 knockdown was less efficient in affecting the proplatelet formation than MAL knockdown. This is probably the result of the lower efficiency of the shMYL9 leading only to 50% reduction of MYL9 transcript, whereas shMAL resulted in a nearly complete abolition of MAL expression. Nevertheless, the data confirm that MAL/SRF complex is involved in proplatelet formation via a transcriptional regulation of MYL9.

MKs migrate through bone marrow endothelial cells to release platelets within the sinusoidal space in response to SDF-1.^46,47  Because SRF was reported to be involved in cell migration,³  we investigated the effect of MAL knockdown on MK migration. MKs were transduced and grown as described in “Cell transduction,” and chemotactic assays in response to SDF-1 were performed on day 8 of culture. No difference was detected between control shRNA- or shMAL-transduced MKs derived from cytapheresis after migration through noncoated or collagen I–coated transwells (data not shown). However, when Matrigel-coated transwells were used, a 2-fold decrease in MK migration in response to SDF-1 was seen in shMAL-expressing cells (n = 2, P ≤ .08; Figure 7D). This defective MK migration through an extracellular matrix similar to the endothelial basement membrane suggests a deficiency in MMP activity.⁴⁸  It could be explained by the marked decrease of MMP9 expression in shMAL-expressing cells (Figure 5A), which encodes for an important proteinase involved both in MK migration through bone marrow endothelial cells in response to SDF-1 and in the release of platelets into the blood flow.⁴⁸ 

MK differentiation is characterized by DNA endoreduplication, cytoplasmic maturation, and expansion. In addition, MKs must migrate in the bone marrow from the endosteal to the vascular niches, and then form long cytoplasmic extension that crosses the endothelial barrier to form platelets under the blood flow. During differentiation, MKs must progressively accumulate the main platelet proteins localized in the membrane or in secretory granules but also develop a complex cytoskeleton network to allow polyploidization, proplatelet formation, and adhesion and aggregation of future platelets. MK development is controlled by the concerted action of transcription factors, which form complexes to specifically activate or repress gene expression. The transcriptional regulation of platelet-specific genes is now well understood. It essentially depends on GATA-1/FOG-1 transcription complexes and on members of the Ets family, especially FLI-1.⁴⁹  In contrast, the transcriptional regulation of cytoskeleton components as well as its relationship with the microenvironment-mediated signal transduction are still poorly understood. The role of this transcriptional regulation has been underscored by the knockout of p45^NF-E2, which leads to a lethal thrombocytopenia with the absence of proplatelet formation. This profound defect is in large part the result of a defect in β1-tubulin, a direct target of p45^NF-E2.⁴⁹  These data have emphasized the major role of microtubules in the process of proplatelet formation.²⁹  However, there is increasing evidence that the actomyosin cytoskeleton plays also an important function in efficient platelet production, not only by regulating proplatelet branching and amplification of platelet production but also by regulating MK polyploidization,²⁶  demarcation membrane distribution, MK migration, and probably also platelet release.^32,50 

SRF is a major factor that transcriptionally regulates many cytoskeleton components. Transcriptional regulation of this set of genes by SRF depends on the myocardin family coactivators. We have used an shRNA strategy to knockdown MAL in human MKs derived from CD34⁺ cells and to show that MAL plays a major role in regulating target genes involved in proplatelet formation and MK migration.

In fibroblasts, MAL/SRF activation is dependent on the Rho pathway and controls many Rho-dependent cytoskeleton processes.⁵¹  Our data indicate that it might be the same in MK differentiation. We showed that, in suspension cultures, in the presence of TPO, a condition in which the Rho and Cdc42 pathways are only slightly activated, MAL is mainly distributed in the MK cytoplasm. In contrast, when the RhoA or Cdc42 pathways are activated after adherence on collagen 1 and convulxin, respectively, MAL accumulates in the nucleus. This result strongly suggests that MAL in MK, as in fibroblasts, is regulated by the actin treadmilling cycle through RhoA and Cdc42 effectors. Consequently, knockdown of MAL impairs adhesion, stress fiber, and filopodia formation in MK. This could be explained by down-regulation of β-actin, thrombospondin 1 (THSB1), and some proteins involved in a multicomponent focal adhesion complex (zyxin and vinculin). These cytoskeleton components are regulated by the MAL/SRF complex in other cell types, such as fibroblasts⁴  and vascular smooth muscle cells.⁵²  THSB1 is an angiogenetic factor, which may negatively regulate MK progenitors,⁵³  although THSB1 knockout mice have no MK/platelet defect.⁵⁴  The expression of other focal adhesion proteins, such as talin, involved in GPIIb-IIIa–dependent platelet activation⁵⁵  and focal adhesion kinase, a negative regulator of megakaryopoiesis,⁵⁶  were not modified after MAL inhibition.

Recent reports performed in cell lines and CD34⁺ cells showed that MAL overexpression promotes MK differentiation and polyploidization.³⁵  These results differ from ours in which knockdown of MAL has no significant effect on human MK differentiation, as judged by the expression of CD41 and CD42 and polyploidization. Furthermore, a tendency to increase ploidy was observed in MAL knockdown MKs, an expected result because inhibition of RhoA activity also leads to a slight increase in MK ploidy.²⁶  It is possible that overexpression of MAL in MKs performed by Cheng et al³⁵  modifies the SRF complexes and interferes with its association with TCFs. TCFs are important mediators of the MAP kinase pathway,^6,7  which plays a major role in MK differentiation.^57,58  Thus, a part of the effect of MAL overexpression might be related to an alteration in this pathway. However, Cheng et al reported also a decrease in ploidy level in bone marrow MKs of MKL1^−/− mice.³⁵  We cannot exclude that the regulation of MK polyploidization between mouse and human is different. However, it will be interesting to confirm these data in bone marrow MK from another MKL1^−/− model developed by Li et al.⁵⁹ 

In contrast, our study demonstrates a major effect of MAL knockdown on proplatelet formation and MK migration. This result is in agreement with the phenotype of MAL knockout mice,³⁵  which have a decreased platelet count suggesting a defect in platelet biogenesis.

Gene expression profiling indicates that MAL knockdown affects the expression of 2.5% of the genes common between adult and neonate MK. To further work, we selected 2 of the 20 most down-regulated genes, MMP9 and MYL9. We have demonstrated by luciferase and ChIP experiments that the MAL/SRF complex directly regulates both MMP9 and MYL9. Very recently, Medjkane et al reported the decrease of MYL9 and MYH9 expression in breast carcinoma and melanoma cell lines as a result of siRNA-mediated MAL knockdown.⁶⁰  Our study extends this report in another cell model and further supports the hypothesis that MAL/SRF regulates genes involved in actomyosin contractility.

MYL9 is a myosin light chain that is associated with MYH9, the only isoform of nonmuscle myosin II present in platelets.⁶¹  Phosphorylation of MYL9 activates MYH9 and negatively regulates proplatelet formation in normal MKs.⁶²  In contrast, deficiency in myosin IIA leads to a defect in proplatelet formation in the MYH9 syndrome, and use of a shMYH9 induces a decrease in the percentage of proplatelet-forming MKs (Y.C. and N.D., unpublished results, February 2008). Again, a difference between mouse and human model should be noted because, although MYH9-deficient mice exhibited a marked thrombocytopenia, an increase in proplatelet-forming MKs was observed. The cytoplasm of murine MYH9 knockout MK is totally disorganized compared with WT MKs. The demarcation membrane system is abnormally dilated with a vacuole-like appearance.⁶³  Interestingly, very similar ultrastructural abnormalities were observed in human MKs treated with the shMAL, strongly suggesting that basal regulation of MYL9 by MAL/SRF is an important determinant in the regulation of megakaryopoiesis by this transcriptional complex. This hypothesis was supported by MYL9 depletion in cultured MKs leading to inhibition of proplatelet formation.

In addition, we show that MAL is required for MK migration in response to SDF-1 through Matrigel-coated transwell but not through noncoated transwell (data not shown). This phenotype may be related to the MYL9 deficiency, which is implicated in cell motility, but even more by the marked decrease in MMP9 expression observed in MAL-deficient MKs. Indeed, MMP9 has been demonstrated to be necessary for in vitro MK migration through the basement membrane in response to SDF-1.⁴⁸  In this study, the authors suggested that MMP9 may be also required for platelet release. MMP9 plays an important role in platelet functions, and its antiaggregant effect may be at the origin of hemorrhagic transformation of acute stroke. However, the large redundancy between various MMPs renders the study of their respective role in platelet functions difficult.⁶⁴  Altogether, the defect in MYL9 and MMP9 induced by MAL knockout could explain the MK/platelet phenotype of MAL knockout mice recently described.³⁵ 

The oncogenic OTT-MAL fusion protein has been shown to interfere with the activity of SRF in model cell lines.⁶⁵  Acting as a dominant negative form of MAL, it could block the terminal stages of MK maturation affecting the expression of MMP9 and MYL9. Thus, it will be interesting to assess the role of these 2 new target genes of MAL in the phenotype of the t(1;22) acute megakaryoblastic leukemia, for example, in the context of recently developed knock-in OTT-MAL mouse model of AMKL.⁶⁶ 

The authors thank F. Wendling for helpful suggestions on the manuscript, Kirin Brewery (Tokyo, Japan) for human recombinant thrombopoietin, Dr Morgenstern for Megix vector (Millennium Pharmaceuticals, Inc, Cambridge, MA), M. Jandrot-Perrus (Hôpital Bichat, France) for the gift of convulxin, and T. Mercher for MAL cDNA (Hôpital Necker, France) and genomic platform of IGR for microarrays.

This work was supported by grants from Fondation de France and the Agence Nationale de la Recherche (ANR contrat blanc). L.G. was supported in part by the Association pour la Recherche sur le Cancer (ARC). D.B. was supported by a postdoctoral fellowship from ANR.

Contribution: L.G. performed experiments, analyzed data, and wrote the paper; D.B. performed experiments and analyzed data; S.B., Y.Z., T.R., and N.D. performed experiments; Y.C. and O.A.B. discussed the results; P.D. analyzed data; W.V. designed the work and wrote the paper; and H.R. designed and supervised the experiments and wrote the paper.

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

Correspondence: Hana Raslova, Inserm, U790, 39 rue Camille Desmoulins, Villejuif, France; e-mail: hraslova@igr.fr.