Development of Platelet Inhibition by cAMP during Megakaryocytopoiesis*

Prostacyclin is a potent inhibitor of agonist-induced Ca2+ increases in platelets, but in the megakaryocytic cell line MEG-01 this inhibition is absent. Using human megakaryocytic cell lines representing different stages in megakaryocyte (Mk) maturation as well as stem cells and immature and mature megakaryocytes, we show that the inhibition by prostacyclin develops at a late maturation stage shortly before platelets are formed. This late appearance is not caused by insufficient cAMP formation or absent protein kinase A (PKA) activity in immature cells. Instead, the appearance of Ca2+ inhibition by prostacyclin is accompanied by a sharp increase in the expression of the catalytic subunit of PKA (PKA-C) but not by changes in the expression of the PKA-regulatory subunits Iα/β, IIα, and IIβ. Overexpression of PKA-C in the megakaryocytic cell line CHRF-288-11 potentiates the Ca2+ inhibition by prostacyclin. Thus, up-regulation of PKA-C appears to be a key step in the development of Ca2+ inhibition by prostacyclin in platelets.

A rise in cytosolic Ca 2ϩ ([Ca 2ϩ ] i ) is a key step in platelet activation. Platelet agonists raise [Ca 2ϩ ] i by the direct stimulation of Ca 2ϩ channels in the plasma membrane (1) and by stimulating G-protein-and tyrosine kinase-coupled receptors, which initiate signaling pathways that generate inositol 1,4,5trisphosphate (IP 3 ) 1 (2,3). IP 3 activates IP 3 receptors in the endoplasmic reticulum, releasing Ca 2ϩ into the cytosol. The depleted Ca 2ϩ stores trigger store-operated Ca 2ϩ entry across the plasma membrane (4). Plasma membrane Ca 2ϩ -ATPases and sarco/endoplasmic reticulum Ca 2ϩ -ATPases (SERCAs) then remove Ca 2ϩ ions from the cytosol and restore the basal [Ca 2ϩ ] i concentration of about 70 nM (5).
In the circulation, uncontrolled platelet activation is prevented by the platelet inhibitors prostacyclin (prostaglandin I 2 ) and nitric oxide produced by endothelial cells. Prostacyclin binds to IP receptors that start cAMP formation via the GTPbinding protein G s and adenylyl cyclase. The importance of this pathway is reflected by the occurrence of acute myocardial infarction, coronary artery disease, and angina in patients with impaired platelet cAMP production (6 -8).
The inhibitory effects of cAMP are mediated via protein kinase A (PKA), which phosphorylates and thereby inactivates numerous signaling elements in platelet-activating pathways such as the thromboxane A 2 (TxA 2 ) receptor type ␣ (9), phospholipase C␤ 3 (PLC-␤ 3 ) (10), and IP 3 receptors (11,12). Because a small rise in cAMP already leads to a strong activation of PKA, Ca 2ϩ signaling in platelets is extremely sensitive to increases in cAMP (13).
Because platelets are synthesized by megakaryocytes (Mks), one would expect a similar suppression of Ca 2ϩ responses by prostacyclin in these progenitor cells. There are indications that the regulation of Ca 2ϩ homeostasis is different in these cells. Megakaryocytic cell lines thought to mimic immature Mks respond to the stable prostacyclin analog iloprost with an expected rise in cAMP but also with a rise in [Ca 2ϩ ] i . Thus, in these cells Ca 2ϩ responses are insensitive to inhibition by cAMP (14,15). It is possible that the cause of this different [Ca 2ϩ ] i regulation must be sought in the action of PKA.
PKA is activated by cAMP through binding to its regulatory subunits. One PKA holoenzyme complex is composed of two catalytic subunits and two regulatory subunits. Three isoforms of the PKA catalytic subunit (PKA-C) have been characterized and designated the C␣, C␤, and C␥ isoforms, respectively. They appear on SDS-PAGE as proteins with molecular masses ranging from 41 to 50 (PKA-C␣), from 42 to 50 (PKA-C␤), and from 39 to 40 kDa (PKA-C␥) (16 -18). Four isoforms of the PKA regulatory subunit (PKA-R) have been found and named the RI␣, RI␤, RII␣, and RII␤ isoforms, respectively. They have apparent molecular masses of 49 (PKA-RI␣), 54 -55 (PKA-RI␤), 51 (PKA-RII␣), and 53 kDa (PKA-RII␤) (19). PKA is activated by the binding of two cAMP molecules to each of the regulatory subunits, thereby releasing and activating the catalytic subunits.
The present study was undertaken to clarify the factors that make cAMP an inhibitor of Ca 2ϩ responses in platelets. Earlier work has shown that immature megakaryocytic cell lines and Mks produce cAMP when stimulated with prostacyclin (15,20). The present results show that PKA is present and functional in stem cells and all stages of megakaryocytopoiesis but does not become an inhibitor of Ca 2ϩ increases before Mk maturation is almost completed and platelets are formed.
A rabbit polyclonal antibody against PKA-C subunits and a goat polyclonal antibody against vasodilator-stimulated phosphoprotein (VASP) were from Santa Cruz Biotechnology. Monoclonal antibodies against PKA-RI␣/␤, PKA-RII␣, and PKA-RII␤ were from BD Transduction Laboratories (Lexington, KY). Horseradish peroxidase-linked antirabbit antibody was from Cell Signaling Technologies (Beverly, MA), and goat-anti-mouse peroxidase and rabbit-anti-goat peroxidase were from DAKO.
Plasmids-EGFP-N1 was obtained from CLONTECH. pC␣EV and pC␤EV, encoding the mouse sequence of PKA-C␣ and PKA-C␤, respectively, under transcriptional control of the mouse metallothionein promoter and the empty vector Zem3, were kindly provided by G. S. McKnight (University of Washington, Seattle, WA) (16).
Platelet Isolation-Donors claimed not to have taken any medication during the preceding 10 days. After informed consent was obtained, freshly drawn venous blood from healthy volunteers was collected into 0.1 volume of 130 mM trisodium citrate. Citrated blood was centrifuged (150 ϫ g, 15 min, 20°C), and the platelet-rich plasma was collected and used for Ca 2ϩ measurements in intact platelets with Fura-2. For the preparation of washed platelets, platelet-rich plasma was supplemented with prostaglandin I 2 (10 ng/ml) and ACD (2.5 g of trisodium citrate, 1.5 g of citric acid, and 2.0 g of D-glucose in 100 ml of distilled water) for acidification to pH 6.5, centrifuged again (330 ϫ g, 15 min, 20°C), and resuspended in HEPES-Tyrode (HT) buffer (145 mM NaCl, 5 mM KCl, 2 mM MgCl 2 , 10 mM Na-Hepes, and 0.1% (w/v) glucose, pH 6.5).
Isolation of CD34 ϩ Stem Cells-Umbilical cord blood was collected during normal full-term deliveries with the informed consent of the mother and used within 48 h. CD34 ϩ cells were isolated as described previously (21). The purity of the CD34 ϩ cells after isolation was determined by FACS analysis (FACSCalibur, BD Biosciences Immunocytometry Systems) after labeling the cells with FITC-labeled antibody against CD34 (HPCA-2) and was between 95 and 99%.
Cultures of Megakaryocytic Cell Lines and Mks-MEG-01, DAMI, and CHRF-288-11 cells were cultured as described earlier (15). These cell lines represent different stages of Mk maturation. MEG-01 cells resemble immature Mks because they express CD41/CD61 but hardly any CD42b, which is a marker for more mature Mks (22). DAMI cells have the properties of MEG-01 cells but in addition contain CD42b, reflecting a further maturation stage (23). CHRF-288-11 cells represent mature Mks because they have secretion granules in addition to the surface markers present in the immature cell lines (24). Also, agonistinduced Ca 2ϩ rises and the expression of certain trimeric G-protein ␣-subunits show a sequential increase going from MEG-01 to DAMI and CHRF-288-11 (15). Freshly isolated CD34 ϩ cells were cultured in a 12-well culture plate at 37°C in a humidified atmosphere with 5% CO 2 . The initial cell density was 4 ϫ 10 5 cells/ml of medium. The composition of the medium was the same as described by Zauli et al. (25) with the omission of nucleosides. rhTPO and rhSCF were added to a final concentration of 20 and 50 ng/ml, respectively. After 3 days of culture, 1 ml of fresh medium supplemented with 20 ng/ml rhTPO and 50 ng/ml rhSCF (final concentrations) was added. After 7 and 11 days of culture, 1 ml of fresh culture medium containing 20 ng/ml rhTPO (final concentration) was added.
Purification of Mks-To obtain immature and mature Mks, CD34 ϩ cells were cultured with rhTPO and rhSCF for 7 and 14 days, respectively, as described above. Cells were harvested and washed twice (5 min, 125 ϫ G, 20°C) in phosphate-buffered saline containing 0.5% BSA and 2.5 mM EDTA (buffer A). Subsequently, the cells were labeled for 20 min at 4°C with mouse-anti-human IgG directed against CD61 (antibody 7F12) for the isolation of immature Mks and against CD42b (antibody 6.20) for the isolation of mature Mks. After washing with buffer A, cells were labeled (20 min, 4°C) with goat-anti-mouse magnetic microbeads and washed again with buffer A. The cell suspension was brought on a ferromagnetic column, type MS ϩ (Miltenyi Biotec), and after three washing steps with 700 l of buffer A each, the retained cells were eluted with buffer A. The purity of the final immature (CD61 ϩ ) and mature (CD42b ϩ ) Mk suspensions was determined by flow cytometry and was typically between 94 and 98%. These populations have been characterized thoroughly in previous studies (26,27). Measurement of Calcium Responses in Suspension Cells-Plateletrich plasma was supplemented with 3 M Fura-2/AM and incubated for 45 min at 37°C in the dark. The platelet-rich plasma was acidified with ACD (2.5 g of trisodium citrate, 1.5 g of citric acid, and 2.0 g of D-glucose in 100 ml of distilled water) to pH 6.5 and centrifuged again (330 ϫ g, 15 min, 20°C). Platelets were resuspended in HT buffer (0.1% (w/v) glucose, pH 6.5) to a concentration of 2 ϫ 10 9 /ml and stored at room temperature in the dark. Immediately before the measurement of [Ca 2ϩ ] i , the platelet suspension was diluted 10 times in prewarmed (37°C) HT buffer (0.1% glucose, pH 7.4), and GRGDS peptide (100 M, final concentration) was added to prevent platelet aggregation.
Megakaryocytic cell lines, freshly isolated stem cells, and purified immature and mature Mks were pelleted (5 min, 125 g, 20°C) and resuspended in HT buffer, pH 7.4, supplemented with 0.1% (w/v) BSA and 0.1% (w/v) glucose at a concentration of 5 ϫ 10 5 cells/ml. Fura-2/AM was added to a final concentration of 3 M, and cells were loaded for 1 h at 37°C in the dark. Cells were spun down, resuspended in HT buffer without BSA to a final concentration of 1.6 ϫ 10 6 cells/ml, and stored at room temperature in the dark. Immediately before the analysis of [Ca 2ϩ ] i , the cells were resuspended in prewarmed (37°C) HT buffer to a final concentration of 2 ϫ 10 5 /ml. Measurements were performed at 37°C with mild stirring (50 rpm) on a Hitachi F-4500 fluorescence spectrophotometer (Hitachi Ltd., Tokyo, Japan). Fura-2 fluorescence was measured at 340-(F1) and 380-nm (F2) excitation and 510-nm emission wavelength. [Ca 2ϩ ] i was calculated using the formula of Grynkiewicz et al. (28). Calculations were based on a dissociation constant of the Fura-2-Ca 2ϩ complex of 224 nM.
Measurement of IP 3 -induced Ca 2ϩ Mobilization-IP 3 -induced Ca 2ϩ mobilization was determined according to a method described previously (29) with slight modifications. Washed platelets were resuspended in Ca 2ϩ -free HT buffer (0.2% glucose, pH 6.5) at a concentration of 5 ϫ 10 9 platelets/ml. MEG-01 and CHRF-288-11 cells were washed and resuspended in the same buffer at a concentration of 15 ϫ 10 7 and 7.5 ϫ 10 7 cells/ml, respectively. Immediately before the measurements, the 60-l cell suspension was diluted with 240 l of a buffer composed of 20 mM HEPES, 100 mM KCl, 1.4 mM MgCl 2 , 100 mM sucrose, 1.25 mM NaN 3 , 0.6 g/ml oligomycin, and 1 M Fluo-3, pH 7.4. Cells were prewarmed at 37°C, treated with 1 M iloprost for 1 min, and subsequently permeabilized with 50 (platelets) or 65 g/ml (MEG-01 and CHRF-288-11 cells) saponin at 37°C for 2-5 min until a stable base line was obtained. When necessary, EGTA was added to lower the Ca 2ϩ concentration to 350 -400 nM, which is an optimal concentration for the measurement of IP 3 -induced Ca 2ϩ release (29). Subsequently, IP 3 (50 nM, final concentration) was added, and changes in fluorescence inten-sity were monitored at 488-nm excitation and 525-nm emission wavelengths in a Hitachi F-4500 fluorescence spectrophotometer.
Determination of PKA Activation-For the measurement of PKA activation, cells were resuspended in Ca 2ϩ -free HT buffer (pH 7.4, 0.1% glucose) to a final concentration of 4 ϫ 10 5 cells/ml and incubated with iloprost for 10 min at 37°C with or without preincubation with H89 (15 min, 10 M, 37°C). The activity of PKA at different time points was deduced from the mobility shift of VASP on a Western blot caused by PKA-mediated phosphorylation on Ser 157 as described elsewhere (30).
Western Blotting-Cells were pelleted and lysed in a buffer consisting of 1% (v/v) Nonidet P-40, 0.5% (w/v) octyl glucoside, 5 mM EDTA, 0.1% SDS supplemented with 2 mM sodium orthovanadate, and protease inhibitors (1% inhibitor mixture from Sigma; catalog no. P8340). A sample was taken for determination of the protein content with the BCA assay, and a Laemmli electrophoresis sample buffer was added (final concentrations: 0.001% (w/v) bromphenol blue, 2% (w/v) SDS, 5% (v/v) ␤-mercaptoethanol, and 10% (v/v) glycerol in 62.5 mM Tris, pH 6.8). Samples were boiled, and 5 g of protein per lane was subjected to SDS-PAGE (10% gel) and electroblotted onto a nitrocellulose membrane. The blots were stained with Ponceau S and scanned, and bands were quantified using ImageQuant from Molecular Dynamics. In all experiments the amount of protein per lane, based on the pixel density of all of the proteins, varied less than 10%. Subsequently, blots were destained with water and blocked in Tris-buffered saline with 0.2% (v/v) Tween 20 (TBST) supplemented with 2% (w/v) Protifar and 0.5% BSA for 1 h at room temperature. Incubation with antibodies against PKA-C, PKA-RI␣/␤, PKA-RII␣, or PKA-RII␤ was performed overnight in TBST at 4°C. The membranes were washed six times with TBST and incubated with horseradish peroxidase-linked anti-rabbit, anti-goat, or antimouse antibody for 1 h at 4°C in TBST. Bands on blot were visualized by enhanced chemiluminescence, and the intensity of the bands on the film was quantified using ImageQuant. Relative intensities were calculated, and the value for MEG-01 cells (for comparison of different cell lines) or stem cells (for comparison of different stages of megakaryocytopoiesis) was set at 1.00.
Transient Transfection and FACS Sorting of Transfected CHRF Cells-CHRF cells (1 ϫ 10 7 ) were resuspended in 300 l of Fischer's medium (without serum and penicillin/streptomycin/L-glutamine), transferred to a Gene Pulser cuvette (0.4-cm electrode gap; Bio-Rad), and incubated for 17 min with 12 g of EGFP-N1 and 41 g of pC␣EV while gently mixing after 10 min. For mock transfection, cells were incubated with 12 g of EGFP-N1 and 41 g of Zem3. Electroporation was performed at room temperature in a Bio-Rad Gene Pulser II at 200 V, 950 microfarads, 20 -30 ms. Cells were subsequently cultured for 24 h in medium supplemented with 100 M ZnSO 4 to induce PKA-C␣ expression. EGFP-positive cells were sorted on a FACSCalibur and analyzed for PKA-C expression by Western blotting.
Measurement of Ca 2ϩ Responses in Transfected CHRF Cells-Ca 2ϩ responses were measured in transfected CHRF cells by single cell fluorescence imaging microscopy as described in Ref. 27. In short, transfected CHRF cells were adhered to anti-GPIIIa (7F12) antibodycoated coverslips and loaded with Fura-2/AM. Fura-2 fluorescence was recorded upon stimulation of the cells with iloprost (1 M) and thrombin (1 unit/ml) as indicated. Subsequently, EGFP fluorescence was determined with a 485-nm excitation filter (half-bandwidth, 22 nm), a 505-nm dichroic long pass filter, and a 530-nm emission filter (halfbandwidth, 30 nm). Ca 2ϩ tracings of EGFP-positive cells were analyzed and averaged (n ϭ 17).

Inhibition of Ca 2ϩ Signaling by Iloprost in Platelets and
Mks- Fig. 1A illustrates the rapid increase in [Ca 2ϩ ] i in platelets after stimulation with thrombin (1 unit/ml). Following a subsequent slight decrease, a stable level was maintained for at least 2 min. The addition of the prostacyclin analog iloprost (1 M) 30 s after the thrombin addition immediately lowered the elevated Ca 2ϩ level to almost basal levels. When platelets were first incubated with iloprost for 1 min and thereafter stimulated with thrombin, the thrombin-induced Ca 2ϩ increase was almost completely abolished (Fig. 1B). These data illustrate the potent inhibition of [Ca 2ϩ ] i increases by iloprost and confirm earlier observations (5). Similar experiments in megakaryocytic cell lines ( Fig. 2A), stem cells, and immature and mature Mks (Fig. 2B) revealed that the effect of iloprost on thrombin-induced Ca 2ϩ signaling was different in these cells. When megakaryocytic cell lines were first stimulated with thrombin (1 unit/ml) and after 3 min treated with iloprost (1 M), different responses were observed among MEG-01, DAMI, and CHRF 288-11 cells. In MEG-01 cells, thrombin induced a Ca 2ϩ increase of 153 Ϯ 50 nM (mean peak height of Ca 2ϩ increase Ϯ S.D.; n ϭ 3). The subsequent addition of iloprost failed to reduce [Ca 2ϩ ] i but instead triggered a further rise in the Ca 2ϩ level of 165 Ϯ 30 nM. As we described in detail in another report (27), this iloprost-induced Ca 2ϩ increase is also found in stem cells. It is caused by an increase in cAMP and further signal generation to the sarco/endoplasmic reticulum via a mechanism that is independent of PKA. In DAMI cells, a first addition of thrombin raised [ compared with stem cells, which was followed by a 30% decrease induced by iloprost. Thus, these results in Mk cultures were qualitatively similar to those found in megakaryocytic cell lines and show that the inhibition of Ca 2ϩ signaling by iloprost appears at a late stage in the maturation of Mks.
To investigate the inhibition by iloprost prior to stimulation with thrombin, the different cell populations were first treated with iloprost (1 M) and 1 min later stimulated with thrombin (1 unit/ml). In MEG-01 and DAMI cells (Fig. 3A) and stem cells and immature Mks (Fig. 3B), iloprost induced a rise in [Ca 2ϩ ] i . The subsequent addition of thrombin (1 unit/ml) induced a further elevation of [Ca 2ϩ ] i . Thus, iloprost failed to inhibit the thrombin-induced Ca 2ϩ increase in these immature cells. In contrast, in CHRF-288-11 cells and mature Mks there was a weak inhibition of thrombin-induced Ca 2ϩ increases by iloprost amounting to 16 Ϯ 5 and 18 Ϯ 6%, respectively (Fig. 3, A and  B). A stronger inhibition by iloprost was observed at lower concentrations of thrombin (not shown). Thus, iloprost inhibited Ca 2ϩ increases both before and after thrombin stimulation in CHRF-288-11 cells and mature Mks but not in the immature stages of Mk maturation.

Iloprost Inhibition of IP 3 -induced Ca 2ϩ Release during Mk Maturation-Multiple steps in agonist-induced increases in
[Ca 2ϩ ] i are known to be inhibited by cAMP-mediated activation of PKA. Common steps in the signaling pathways that trigger the mobilization of Ca 2ϩ ions from storage sites in the sarco/ endoplasmic reticulum are the generation of IP 3 and the subsequent activation of IP 3 receptors, which form tetrameric ligand-gated Ca 2ϩ channels that release Ca 2ϩ from intracellular stores upon the binding of IP 3 . To investigate the contribution of these steps in the development of iloprost-inhibition, MEG-01 cells, CHRF-288-11 cells, and platelets were permeabilized with saponin to enable the entry of IP 3 , and the release of Ca 2ϩ was determined as described elsewhere (29). The addition of IP 3 (50 nM) to MEG-01 and CHRF-288-11 cells and platelets induced an increase in the initial Ca 2ϩ concentration of 210 Ϯ 18 nM and 500 Ϯ 34 nM per 3 ϫ 10 7 MEG-01 and 1.5 ϫ 10 7 CHRF-288-11 cells in 1 ml, respectively, and of 110 Ϯ 9 nM per 1 ϫ 10 9 platelets in 1 ml. A preincubation with iloprost (1 M, 1 min) had no effect on the IP 3 -induced Ca 2ϩ increase in MEG-01 cells (Fig. 4A). In CHRF-288-11 cells, pretreatment with iloprost suppressed the IP 3 -induced Ca 2ϩ increase by 21 Ϯ 5%, and in platelets this inhibition was 52 Ϯ 6% (Fig. 4, B and C). The results for platelets correspond well with findings by other investigators (12). Thus, the increased inhibition of Ca 2ϩ mobilization is one of the mechanisms that contribute to the development of Ca 2ϩ inhibition by iloprost.
Role of PKA in the Inhibition of Ca 2ϩ Signaling by Iloprost-The observation that iloprost fails to inhibit thrombin-induced Ca 2ϩ signaling in stem cells and the early stages of megakaryocytopoiesis might have different causes. First, iloprost might fail to induce cAMP accumulation because of the absence of IP receptors, the trimeric G-protein G s , or adenylyl cyclase. Second, cAMP accumulation might be prevented by enhanced removal by phosphodiesterases. Third, PKA or certain PKA subtypes might not have been fully expressed in these immature cells. In earlier studies, we showed that the megakaryocytic cell lines MEG-01, DAMI, and CHRF-288-11 respond to iloprost with a rise in cAMP. Iloprost raised the cAMP concentration in MEG-01 and DAMI cells from 10 to 50 pmol/10 6 cells (15). This is in the range found in platelets that are 1000ϫ smaller and accumulate 90 pmol cAMP/10 9 cells following stimulation with prostaglandin E 1 (PGE 1 ), which also activates G s (13). At this cAMP level in platelets, thrombin-induced increases in [Ca 2ϩ ] i are almost completely abolished (29). Thus, differences in iloprost sensitivity between different maturation stages cannot be explained by insufficient cAMP accumulation.
To investigate a possible defect in the expression or function of PKA, the different cell populations were screened for their capacity to phosphorylate VASP, which is phosphorylated on Ser 157 by PKA, leading to a mobility shift on SDS-PAGE (30). All cell populations revealed a mobility shift of VASP upon the addition of iloprost (1 M) as shown in Fig. 5A for MEG-01 and CHRF-288-11 cells, stem cells, and mature Mks. Iloprost-induced VASP phosphorylation in these cells was rapid, reaching 50% after 1 min and remaining high for at least 10 min. A similar degree of VASP phosphorylation has been observed in prostaglandin E 1 -treated platelets, and this is accompanied by the complete inhibition of thrombin-induced Ca 2ϩ increases (13,29). Furthermore, the mobility shift of VASP was strongly inhibited by H89, a specific inhibitor of PKA. These results are in agreement with the inhibition of VASP-Ser 157 phosphorylation by H89 in platelets (31). Together, these data indicate that PKA is present and functional at all stages of Mk maturation.
To investigate how PKA contributed to the Ca 2ϩ inhibition by iloprost, the effect of the PKA inhibitor H89 on thrombininduced rises in [Ca 2ϩ ] i was studied in cells showing a partial (CHRF-288-11 cells and mature Mks) or complete (platelets) inhibition by iloprost. Both CHRF-288-11 cells (Fig. 5B) and mature Mks (not shown) fully recovered from a pretreatment with iloprost in the presence of H89, reaching increases in [Ca 2ϩ ] i of 180 Ϯ 24% of cells stimulated in the absence of iloprost and H89. A complete recovery of thrombin-induced Ca 2ϩ signaling was found in platelets reaching 100 Ϯ 13% of untreated samples (Fig. 5C). Thus, the inhibition of Ca 2ϩ by iloprost in these more mature cell populations was mediated by PKA.
Changes in PKA Subunit Expression during Megakaryocytopoiesis-To understand why PKA was present and functional in stem cells and all stages of Mk maturation but inhibited Ca 2ϩ signaling only in mature Mks and especially in platelets, the expression of PKA subunits was analyzed by Western blotting. As shown in Fig. 6, A and B, PKA-C subunits with molecular masses of 39 and 42 kDa were present in all cell types. A 44-kDa PKA-C isoform was weakly expressed in MEG-01, DAMI, and CHRF-288-11 cells but was abundantly present in platelets. Quantitation of the total pixel density of the three PKA-C subunits revealed a 1.4-fold increase in the megakaryocytic cell lines and a further increase to 1.8-fold in platelets. In the cultured megakaryocytes this expression was more or less constant, but platelet formation was accompanied by a 1.9-fold rise in the expression of PKA-C. Hence, the appearance of Ca 2ϩ inhibition by prostacyclin observed especially in platelets was accompanied by up-regulation of PKA-C.
An antibody against the PKA-R subunits I␣ and I␤ revealed a band at an apparent molecular mass of 49 kDa in cell lines, cultured Mks, and platelets, showing that they express the PKA-RI␣ subunit. PKA-RI␣ expression varied little among these cell populations (Fig. 6, A and B). Immature Mks expressed the 54 -55-kDa PKA-RI␤ isoform, but this isoform was not present at more mature stages. Also, the expression of PKA-RII␣ remained more or less the same. In contrast, the expression of PKA-RII␤ showed a stepwise increase when cell lines and Mks were compared at different maturation stages, reaching a 15-fold increase in CHRF-288-11 cells compared with MEG-01 cells and a 4-fold increase in mature Mks compared with stem cells. No further increase was observed in platelets. Thus, changes in PKA-RII␤ expression did not correlate with the appearance of Ca 2ϩ inhibition by prostacyclin.
PKA-C␣ Overexpression in CHRF-288-11 Cells Enhances Ca 2ϩ Inhibition by Iloprost-To confirm that increases in PKA-C expression contribute to the inhibition of Ca 2ϩ increases by iloprost, PKA-C␣ and PKA-C␤ were transiently overexpressed in CHRF-288-11 cells, and Ca 2ϩ responses were determined. Co-expression of EGFP and Ca 2ϩ analysis in single, immobilized cells allowed specific screening of transfected (EGFP ϩ ) cells, which constituted only 10 -15% of the total cell population.
The transfection of CHRF-288-11 cells with PKA-C␣ introduced a 39-kDa PKA-C subtype (Fig. 6, inset). Together with the basal expression of a 42-kDa subtype, this resulted in a 1.5-fold increase in total PKA-C. A similar transfection with PKA-C␤ had less effect and together with the 42-kDa subtype led to an expression that was only 1.2-fold higher than in mock-transfected cells (data not shown). Strikingly, the introduction of PKA-C␣ led to a 50% increase in Ca 2ϩ inhibition by prostacyclin (Fig. 6). No change in Ca 2ϩ inhibition was induced by PKA-C␤ overexpression, possibly because of the minor increase in total PKA-C observed in these cells. These findings support the concept that an increase in PKA-C expression causes Ca 2ϩ inhibition by prostacyclin. DISCUSSION The present study shows that the inhibition of Ca 2ϩ increases by prostacyclin develops at a late stage during megakaryocytopoiesis. The inhibition is absent in MEG-01 and DAMI cells and in CD34 ϩ stem cells and immature Mks but becomes apparent in CHRF-288-11 cells, mature Mks, and especially in platelets. Each of these cell populations is capable of raising cAMP levels and contains PKA capable of inducing VASP phosphorylation. Thus, the signaling elements required for the activation of PKA are present in these immature cell populations, including receptors for prostacyclin (IP receptors), the stimulatory G-protein G s , and adenylyl cyclase. PKA is present and induces the phosphorylation of VASP, one of its major substrates (13). Apparently this is not sufficient to suppress Ca 2ϩ rises. Two explanations may account for this discrepancy. First, the PKA-mediated inhibition is present but is overruled by the cAMP-induced Ca 2ϩ increase in these cells. Second, the inhibition by PKA is too weak because of the incomplete expression of the components of the PKA complex. Evidence for the first explanation comes from the observation that in the presence of the PKA inhibitor H89, cAMP-and thrombin-induced Ca 2ϩ increases are higher than in the absence of this inhibitor (Ref. 27 and the present study). Evidence for the second explanation comes from the observation that the maturation of Mks and platelet formation are accompanied by changes in the expression of the PKA catalytic and regulatory subunits.
The present data show that the appearance of Ca 2ϩ inhibi- tion by prostacyclin is accompanied by a sharp up-regulation of PKA-C. Western blots show the presence of two major PKA-C subtypes with molecular masses of 39 and 42 kDa and a minor expression of a 44-kDa subtype. All cell populations contain the three subtypes, but platelet formation is accompanied by a strong up-regulation of the 44-kDa subtype. The combined expression of the three subtypes varies slightly in the megakaryocytic cell lines and during megakaryocyte maturation. In contrast, platelet formation shows a 1.7-1.8-fold increase in the combined expression of PKA-C subtypes compared with MEG-01 cells and stem cells. At the same time Ca 2ϩ inhibition by prostacyclin increases sharply, changing from a minor inhibition in CHRF-288-11 cells and mature Mks to complete inhibition in platelets. This suggests that these observations are causally related. Indeed, the transfection of CHRF-288-11 cells with PKA-C␣ results in a 1.5-increase in total PKA-C expression and a 50% stronger Ca 2ϩ inhibition by prostacyclin.
The overexpression of both PKA-C␤ and PKA-C␣ resulted in an increase of the 39-kDa PKA-C subunit, which is slightly lower than the reported molecular mass of these subunits of 41-50 kDa (16 -18). The overexpression of PKA-C␤ was lower than that of PKA-C␣. This corresponds with previous findings by Uhler and McKnight (16), who transfected these PKA constructs in NIH 3T3 cells and found a 1.5-fold lower expression of PKA-C␤ than of PKA-C␣. The low PKA-C␤ overexpression in CHRF-288-11 cells did not affect Ca 2ϩ signaling, suggesting that PKA-C expression must exceed a certain threshold to make PKA an inhibitor of Ca 2ϩ signaling.
The present results agree well with observations in other cell types showing that signal transduction via cAMP is sensitive to changes in PKA-C expression. The treatment of cytotoxic Tlymphocytes with PKA-C␣ antisense oligonucleotide reduced both the basal and cAMP-inducible PKA activation, leading to increased cytotoxicity, inhibition of T-cell receptor-triggered ␥-interferon secretion, and inhibition of expression of ␥-interferon mRNA (32). Chinese hamster ovary cells have a 1.7-fold higher level of PKA-C␣ mRNA than the mutant 10260 cell line and an 80% higher activity of PKA (33).
Also, the expression of PKA-RII␤ changes during Mk maturation, but the up-regulation starts at a much earlier stage and appears to precede the development of Ca 2ϩ inhibition by prostacyclin. Nevertheless, up-regulation of PKA-RII␤ expression might be important for PKA activity in platelets. PKA-RII␤ has been shown to mediate the targeting of the PKA holoenzyme to different subcellular compartments via the binding to protein kinase A anchoring proteins (AKAPs) (34). Type II PKA containing the regulatory subunit II␤ binds to the sarco/endoplasmic reticulum via D-AKAP1 (34 -37). This might be a prerequisite for PKA to come in close proximity to IP 3 receptors and facilitate their phosphorylation and suppression of Ca 2ϩ mobilization. Also, the steps in signaling cascades initiated by surface receptors may be subject to PKA inhibition via anchorage to AKAPs. In myometrial cells, the anchoring of PKA to the plasma membrane via AKAP79 was necessary to inhibit phosphatidylinositide turnover (38). PKA is targeted to phospholipase C␤ via its regulatory subunit II␤ and AKAP79 in myometrial cells (38) and to receptors in the plasma membrane via AKAP250/gravin in endothelial cells (39,40).
Apart from changing PKA subunit composition as part of the formation of an important inhibitory mechanism in platelets, alterations in subunit expression may serve a role in the proliferation of Mks. In human neuroblastoma cells, the overexpression of PKA-RII␤ correlated with the inhibition of cell growth (41). By analogy, up-regulation of PKA-RII␤ might be a mechanism by which immature Mks shift from a proliferating to a maturating phenotype.
Multiple steps in signaling cascades from surface receptors to Ca 2ϩ mobilization are inhibited by PKA. A common downstream step is the release of Ca 2ϩ ions from the sarco/endoplasmic reticulum via the activation of IP 3 receptors. Our experiments in permeabilized cells indicate that the Ca 2ϩ inhibition by prostacyclin observed in mature Mks and especially in platelets is also found at the level of IP 3 -induced Ca 2ϩ release. These results correspond with studies on Mks derived from Wistar rats, which showed 80% reduction in IP 3 -induced Ca 2ϩ increases by the prostacyclin analog carbacyclin (42). This indicates that in fully mature Mks cAMP-mediated inhibition of IP 3 -induced Ca 2ϩ release has completely developed. Thus, the mobilization of Ca 2ϩ from internal stores might be at least one of the steps responsible for the changes in Ca 2ϩ signaling seen during Mk maturation and a direct consequence of the upregulation of PKA-C. In addition, the appearance of iloprost inhibition might reflect a change in IP 3 receptor subtypes. Three types of IP 3 receptors have been identified in platelets and are known as IP 3 -RI, -RII, and -RIII (43). Type I IP 3 receptors are better phosphorylated by PKA than type II or type III receptors (44), making this receptor the most sensitive target for cAMP-mediated inhibition. Expression studies in DAMI cells and platelets revealed up-regulation of IP 3 -RI and down-regulation of IP 3 -RIII with IP 3 -RII expression being the same in the two cell types (45). Other observations are in conflict with these data and show down-regulation of IP 3 -RI and up-regulation of IP 3 -RIII (46). Thus, a better insight in the role of IP 3 receptor subtypes in the development of Ca 2ϩ inhibition by PKA awaits more detailed studies in Mks at different stages of maturation.
In conclusion, our data reveal that the appearance of Ca 2ϩ inhibition by prostacyclin starts at a late stage of megakaryocyte maturation when platelets are formed. This inhibition is accompanied by a sharp up-regulation of PKA-C expression, which appears to be a key step in the development of Ca 2ϩ inhibition by prostacyclin in platelets.