Defective nitrovasodilator-stimulated protein phosphorylation and calcium regulation in cGMP-dependent protein kinase-deficient human platelets of chronic myelocytic leukemia.

The presence and functional role of the cyclic nucleotide signal transduction system was investigated in platelets from patients with myeloproliferative disorders. Platelets from certain patients with chronic myelocytic leukemia showed decreased expression of cGMP-dependent protein kinase, and platelets from two such patients were studied in some detail. These platelets had very little if any cGMP-dependent protein kinase but a normal level of cAMP-dependent protein kinase. They also contained a normal level of VASP (vasodilator-stimulated phosphoprotein, a specific substrate of both cAMP- and cGMP-dependent protein kinase), as well as a functionally intact prostaglandin E1-stimulated cAMP-mediated VASP phosphorylation. In contrast, sodium nitroprusside-stimulated VASP phosphorylation was severely impaired in these cGMP-dependent protein kinase-deficient platelets, despite an exaggerated cGMP response to sodium nitroprusside. Furthermore, whereas selective activation of the cGMP-dependent protein kinase by 8-(4-chlorophenylthio)-cGMP strongly inhibited the ADP- or thrombin-evoked calcium mobilization from intracellular stores in normal platelets, this agonist-evoked calcium response was not inhibited by the cGMP analog in cGMP-dependent protein kinase-deficient platelets. The results demonstrate a defect in the nitrovasodilator-/cGMP-regulated signal transduction system in human platelets from some patients with myeloproliferative disorders, and underscore that a cGMP-dependent protein kinase regulatory system, distinct from that of cAMP-dependent protein kinase or other cGMP-dependent effectors is operative in normal human platelets.

The presence and functional role of the cyclic nucleotide signal transduction system was investigated in platelets from patients with myeloproliferative disorders. Platelets from certain patients with chronic myelocytic leukemia showed decreased expression of cGMP-dependent protein kinase, and platelets from two such patients were studied in some detail. These platelets had very little if any cGMP-dependent protein kinase but a normal level of CAMP-dependent protein kinase. They also contained a normal level of VASP (vasodilator-stimulated phosphoprotein, a specific substrate of both CAMPand cGMP-dependent protein kinase), as well as a functionally intact prostaglandin El-stimulated CAMP-mediated VASP phosphorylation. In contrast, sodium nitroprusside-stimulated VASP phosphorylation was severely impaired in these cGMPdependent protein kinase-deficient platelets, despite an exaggerated cGMP response to sodium nitroprusside. Furthermore, whereas selective activation of the cGMP-dependent protein kinase by 8-(4-chlorophen-y1thio)-cGMP strongly inhibited the ADP-or thrombin-evoked calcium mobilization from intracellular stores in normal platelets, this agonist-evoked calcium response was not inhibited by the cGMP analog in cGMP-dependent protein kinase-deficient platelets. The results demonstrate a defect in the nitrovasodilator-/cGMP-regulated signal transduction system in human platelets from some patients with myeloproliferative disorders, and underscore that a cGMP-dependent protein kinase regulatory system, distinct from that of CAMP-dependent protein kinase or other cGMP-dependent effectors is operative in normal human platelets.
Platelet activation and aggregation is inhibited by CAMPelevating prostaglandins, such as prostaglandin El (PG-E1),' * This work was supported in part by the Deutsche Forschungsgemeinschaft  and the Wilhelm Sander-Stiftung. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.   'The abbreviations used are: PG-E1, prostaglandin E,; cA-PK, CAMP-dependent protein kinase; cG-PK, cGMP-dependent protein kinase; CML, chronic myelocytic leukemia; 8-pCPT-cGMP, 8-(4-chloropheny1thio)-cGMP; PRP, platelet-rich plasma; VASP, vasodilator-stimulated phosphoprotein; SNP, sodium nitroprusside. prostacyclin (PG-I*), and their stable analogs, and also by cGMP-elevating nitrovasodilators, such as endothelium-derived relaxing factor, sodium nitroprusside (SNP), and other NO-liberating agents (1)(2)(3)(4). Considerable evidence suggests that the effects of CAMP-and cGMP-elevating platelet inhibitors are mediated by CAMP-and cGMP-dependent protein kinases (cA-PK, cG-PK), respectively (5-7). Our laboratory has characterized a 46/50 kDa (dephospho and phospho form, respectively) protein termed VASP (vasodilator-stimulated phosphoprotein) and demonstrated that vasodilator-stimulated phosphorylation of VASP in intact cells is mediated by either cA-PK or cG-PK (5-9). Additional studies showed that the latter correlates very well with agonist-induced inhibition of platelet activation (3,(8)(9)(10). VASP, originally purified from platelets from patients with myeloproliferative disorders, is present at high concentrations in platelets from both healthy individuals and patients with myeloproliferative diseases (11-13). Recently, VASP was identified as a novel microfilamentassociated protein present in cell-cell and cell-matrix junctions (focal contacts) of a wide variety of cell types (14). Nitrovasodilator stimulation of cG-PK in intact human platelets results in increased VASP phosphorylation and in inhibition of agonist-evoked calcium mobilization from intracellular stores (7).
However, nitrovasodilators and cGMP are known to regulate several distinct signal transduction systems including cGMP-gated ion channels, cGMP-inhibited or -stimulated phosphodiesterases, cGMP-dependent protein kinases, and perhaps even CAMP-dependent protein kinases (1,4,(15)(16)(17). Therefore it has been difficult to conclusively establish a functional role of cG-PK in platelet inhibition and other cellular functions. We now have studied the presence and regulation of cGMP-dependent protein kinase and its function in platelets from patients with myeloproliferative disorders including chronic myelocytic leukemia (CML). Patients with myeloproliferative disorders have been reported to exhibit distinct, but variable platelet abnormalities, such as changes in platelet morphology, reduced glycoprotein Ib levels, deficient a-adrenergic, prostaglandin Dz, or Fc receptors, altered arachidonic acid metabolism, and often abnormal platelet coagulant activity, all of which may cause both bleeding and thrombotic complications (18). CML is a chronic disease of bone marrow stem cells of variable duration which usually evolves into a more aggressive leukemia characterized by an accelerated phase and ultimately the often terminal blast crisis (19)(20)(21). The initial cell transformed in the chronic phase of CML is thought to be a pluripotent stem cell since all myeloid and often even lymphoid cell lineages are more or less affected. The Philadelphia (Ph') chromosomal anomaly 13526 cGMP-dependent Protein Kinase-deficient Platelets 13527 involving reciprocal translocation of parts of the long arm of chromosomes 9 and 22, and more specifically the translocation of the c-ABL protooncogene from chromosome 9 to the bcr region of chromosome 22, has been identified as the pathogenetic molecular lesion responsible for the chronic phase of CML (22)(23)(24). In contrast, the molecular lesions causing the accelerated phase and/or blast crisis have not been unequivocally elucidated, although alterations in the p53 anti-oncogene and n-RAS genes have been detected (20, 21). We investigated whether platelets from patients with myeloproliferative disorders have an intact cyclic nucleotide-regulated signal transduction system. Moderate alterations with respect to the content and regulation of CAMP, cGMP, and their respective protein kinases were frequently detected (13), however, severe reductions of cG-PK levels were observed in platelets from two patients with CML. Comparison of normal platelets and cG-PK-deficient platelets enabled us to investigate the functional role of cG-PK in agonist regulation of platelet protein phosphorylation and calcium responses. The data demonstrate that cG-PK is an essential component of nitrovasodilator and cGMP actions in human platelets. A preliminary account of this work has been presented in abstract form (13).

FIG. 1.
Autoradiographs showing Western blot analysis of cG-PK, catalytic subunit (C) of cA-PK, or VASP in purified standards or human platelet homogenates. Platelets were obtained from a control person (F. F., Co) or patients with CML (chronic myelocytic leukemia, patients F. B. and H. F.). The protein content of purified standards (ng) is indicated. Platelet protein homogenates made from 1 X lo', 2 X lo', or 4 X 10' platelets were used for the analysis of VASP, cG-PK, or catalytic subunits of cA-PK, respectively. Platelets were obtained from patient F. B. on two different dates as indicated.

FIG. 2. Autoradiographs showing the regulation of VASP phosphorylation by SNP or PG-E, in platelets from a healthy control person ( A ) or the CML patient F. B. ( B ) . PRP was incubated without additions (basal)
, 100 p~ SNP, or 5 pM PG-E, for 5 min. Platelet VASP phosphorylation was analyzed by Western blot detection of the dephospho-VASP to phospho-VASP shift (see "Experimental Procedures"). Longer exposures than demonstrated here revealed small amounts of phospho-VASP in the SNP-treated CML platelets.
Characterization of CML Patients-Diagnosis and classification of patients (based on standard clinical, hematological, and pathological criteria) were determined by our departments of medicine, transfusion medicine, and pathology. Peripheral blood from healthy donors and patients was collected and analyzed after obtaining informed consent.
Isolation of Human Platelets and Measurement of cG-PK, Catalytic Subunit of cA-PK, and VASP-Isolation of human platelets was performed as described in detail previously (12). The content of cyclic nucleotide-dependent protein kinases and their substrate VASP in platelets was determined by the Western blot technique (12). Radioactively labeled proteins were localized by autoradiography, and autoradiographs were analyzed with a laser densitometer (Pharmacia, Freiburg, Germany).
Analysis of VASP Phosphorylation and Cyclic Nucleotide Level9 in Intact Human Platelets-Blood was obtained by cubital vein puncture and anticoagulated with a final concentration of 0.38% sodium citrate and 3 mM EDTA. Aliquots of 2 ml were centrifuged for 20 s at 2940 X g at room temperature using an Eppendorf centrifuge. The supernatant (platelet-rich plasma (PRP)) was transferred to a microcentrifuge tube and kept a t 37 "C for subsequent incubations. The PRP was either not treated (basal condition) or incubated with 5 pM PG-E, or 100 PM S N P for 5 min. This PRP was then centrifuged for 10 s a t 8160 X g, and the pellet was resuspended in SDS containing stop solution or in 10% trichloroacetic acid solution for analysis of VASP phosphorylation or cyclic nucleotide levels, respectively. Determination of VASP phosphorylation by the Western blot technique or  cyclic nucleotide levels by radioimmunoassay were carried out as described previously (12).
Fluorescence Measurements with Fura-2-loaded Platelets-Experiments were performed as described in detail previously (7). Briefly, PRP was incubated with 4 PM fura-2 at 37 "C for 45 min, then platelets were pelleted and resuspended in physiological buffer (pH 7.4, containing 10 mM Hepes and 0.1 unit/ml apyrase) to yield a final density of 1 X 1 0 ' cells/ml. Samples were stirred and kept at 37 "C throughout the following experiments. Platelet aliquots were preincubated without additions (control) or with 0.5 mM &pCPT-cGMP, 100 p~ sodium nitroprusside, or 10 p M prostaglandin E, for 10 min and then stimulated with agonists (20 p M ADP or 0.1 unit/ml thrombin). Fura-2 fluorescence was analyzed with a Perkin-Elmer luminescence spectrophotometer using excitation and emission wavelengths of 340 and 510 nm, respectively.

RESULTS
Western Blot Analysis of cG-PK, Catalytic Subunit of cA-PK, and VASP in Platelets from Healthy Control Persons or CML Patients-Specific well characterized antisera against cG-PK, catalytic subunit of cA-PK, and VASP were used to quantitate these proteins in platelets from healthy control persons and CML patients. Antisera specificity and the establishment of Western blot conditions for determining antigen concentrations in human platelets have been reported previously (12). Western blot analysis of cG-PK, catalytic subunit of cA-PK, and VASP present in platelets from a control person (Co) and CML patients (F. B. and H. F.) is shown in Fig. 1. The amount of cG-PK in platelets (prepared and analyzed on two different dates) from CML patient F. B. was very low, almost undetectable, although the amount in platelets from CML patient H. F. was quite similar to that of a normal healthy donor (F. F., Co). The cG-PK level in platelets from two additional CML patients was in one case normal (patient R. R.) and in another very low (patient C. S.) (data not shown). In contrast to the reduced level of cG-PK in platelets from the CML patient F. B., levels of catalytic subunit of cA-PK or VASP were normal or occasionally even slightly elevated (Fig. 1, middle and lower panels), as was the case for the other CML patients investigated (data not shown). These results indicate the loss of cG-PK in platelets from a subgroup of CML patients. Cyclic Nucleotide-regulated VASP Phosphorylation and Cyclic Nucleotide Levels in Platelets from Healthy Control Persons and CML Patients-VASP phosphorylation in response to various platelet inhibitors is mediated by either cG-PK or cA-PK in intact human platelets (9,12). The mobility change of VASP in SDS-polyacrylamide gel electrophoresis after phosphorylation (9) and the availability of a specific antiserum which recognizes both the dephospho form (46-kDa protein) and phospho form (50-kDa protein) of VASP enabled us to investigate cGMP-and CAMP-dependent VASP phosphorylation in normal platelets from control persons and in cG-PK-deficient platelets from CML patients F. B. and C. S. In these experiments, platelets were incubated in the absence or presence of 100 pM sodium nitroprusside or 5 pM PG-E1 for 5 min. In Fig. 2, typical Western blot results of such experiments are demonstrated. In untreated platelets from both healthy donors and CML patients, most of VASP was detected as the 46-kDa dephospho form, but was converted to the 50-kDa phospho form by PG-E1 which elevated platelet CAMP levels about 10-fold without affecting cGMP

cGMP-dependent Protein Kinase-deficient Platelets 13529
A) Norm81 PI8telcta  (Figs. 2 and 3). In normal platelets, a 5-min incubation with sodium nitroprusside caused significant VASP phosphorylation (up to 30-40%) accompanied by a severalfold elevated cGMP level (Figs. 2 and 3). In contrast, sodium nitroprusside had very little or only moderate effects on VASP phosphorylation in the cG-PK-deficient platelets from CML patient F. B. (independently analyzed on two different dates) despite a very increased SNP-induced cGMP response (Figs.  2 and 3). Results similar to those demonstrated here with platelets from the CML patient F. B. were also observed with the cG-PK-deficient platelets from the CML patient C. S., whereas the effects of PG-El and sodium nitroprusside on cyclic nucleotide levels and VASP phosphorylation in platelets from the CML patients H. F. and R. R (which had a normal cG-PK content) were similar to the responses observed for control platelets (data not shown). 8-

pCPT-cGMP or PG-El in Platelets from the CML Patient F. B. and a Control Person-It has been reported previously (7,
25, 26) that the ADP-or thrombin-induced calcium mobilization from intracellular stores, but not the ADP-gated cation channel, can be inhibited by CAMP-and cGMP-elevating agents and cell membrane-permeant cAMP and cGMP ana-logs in intact human platelets. The cGMP analog 8-pCPT-cGMP is a potent and selective activator of cG-PK in intact cell preparations (6,7). Therefore, we used this cGMP analog to investigate the regulation of calcium mobilization in control platelets and in the cG-PK-deficient platelets from the CML patient F. B. The ADP-induced calcium response of fura-2loaded platelets from the control person, but not from CML patient F. B., was strongly inhibited when platelets were preincubated for 10 min with 0.5 mM 8-pCPT-cGMP (Fig. 4). Similarly, a 10-min pretreatment with 8-pCPT-cGMP prevented the thrombin (0.1 unit/ml)-induced calcium response in normal platelets but not in the platelets from CML patient F. B. (Fig. 5). In contrast, a 10-min preincubation with the cAMP elevating PG-EI prevented the ADP-evoked calcium response both in normal platelets as well as in the cG-PKdeficient platelets from CML patient F. B. (data not shown). It is also interesting to note that the increased fura-2 fluorescence after thrombin and especially after ADP stimulation declined more rapidly in control platelets than in the cG-PKdeficient platelets from CML patient F. B. Unfortunately, more extensive experiments analyzing the regulation of the calcium response in platelets from the two CML patients F. B. and C. S. were not possible due to limited platelet samples.

DISCUSSION
Elevation of cGMP and stimulation of cG-PK is closely associated with the inhibition of an early step in the human platelet activation cascade, most likely at the level of phospholipase C activation and subsequent mobilization of Ca2+ from intracellular stores (1-4,7,27,28). However, in addition to cG-PK, cGMP is capable of regulating several other distinct targets including ion channels and phosphodiesterases (1, 4, 15-17) which may contribute to the overall cellular response to cGMP. The present study was initiated to obtain more conclusive experimental evidence for the physiological role of cG-PK as a mediator of agonist-stimulated cGMP in human platelets.
Modern genetic approaches to elucidating the physiological significance of proteins include production and analysis of mutants, or development of transgenic or targeted gene disruption organisms. Obviously these approaches for studying cG-PK roles in human platelets are not possible. In the study reported here, it was, however, possible to identify certain patients with chronic myeloproliferative disorders who demonstrated a reduction, almost elimination, of the normal platelet cG-PK. It was then possible to investigate whether this deficit led to the expected consequences, alteration of cellular functions ascribed to cG-PK. Platelets from CML patients are known to develop variable but distinct defects at the level of receptors, integrins, and metabolism (18), but the expression and function of protein kinases and their substrates have not been previously analyzed. Initial screening of platelets from 10 control individuals and 24 patients with myeloproliferative diseases for their content of cyclic nucleotides, cA-PK, cG-PK, and VASP (12, 13) revealed platelets from two CML patients which had defective cGMP/cG-PK signal transduction. Platelets from CML patients F. B. and C. S. contained a very low, almost undetectable level of cG-PK, but in contrast, normal levels of cA-PK catalytic subunit and the substrate protein VASP (Fig.  1, see "Results"). Perhaps all platelets from these two patients expressed a very low level of cG-PK, or alternatively, there may have been a small number of normal healthy platelets among a majority of platelets deficient in cG-PK.
Defective cG-PK expression does not appear to be directly involved in the cause of CML disease, which occurs by c-ABL protooncogene activation following a reciprocal translocation of parts of chromosomes 9 and 22 (22-24), since the human gene for type I cG-PK has been recently localized to chromosome 10 (29). However, the underlying chromosomal anomaly in CML could nevertheless result in defective cG-PK expression in platelets from some CML patients. Whether Unfortunately, extensive additional studies were not possible defective cG-PK expression might be an indicator of the due to the limited quantities of human CML platelets from development of the accelerated phase and/or blast crisis of individual patients. CML needs to be investigated in greater detail in a prospective Our present results emphasize the important role of the clinical study.
cG-PK signal transduction system in mediating the effects of For the purposes of the present study, evaluation of plate-cGMP-elevating platelet inhibitors, and also demonstrate lets from the two CML patients permitted a correlation to be that the cG-PK-dependent mechanism is distinct from that made between platelet cG-PK deficiency and absent cellular of cA-PK or other cGMP-dependent effectors. Furthermore, functions. CML platelets, in contrast to normal platelets, our results show that normal human platelets are not only demonstrated considerably reduced nitrovasodilator-stimu-useful models for the elucidation of signal transduction syslated cGMP-mediated VASP phosphorylation despite an ex-tems, but that CML platelets can function as valuable adaggerated cGMP elevation in response to sodium nitroprus-juncts for substantiation of proposed physiological roles of side (Figs. 2 and 3). In comparison, the PG-El-stimulated specific components of these regulatory systems. CAMP-mediated VASP phosphorylation in CML platelets was essentially intact. It is possible that the low level of SNP-Acknowledgments-We thank Prof. Dr. K. Kochsiek for continuous induced VASP phosphorylation was due to the activation of support, the hematology laboratory for unlimited access to their cA-PK by cGMP since SNP increased the intracellular level ing this work.
facilities, and Dr. S. M. Lohmann for her critical comments concernof cGMP in these CML platelets from 1 to 47-85 pmol/lOg cells (Fig. 3). Based on a platelet cell volume of 5.2 fl and