Keywords
collagen receptor - kinase inhibitors - Bruton's tyrosine kinase - protein kinase
C - secretion
Introduction
Platelets have crucial roles in hemostasis, thrombo-inflammation, infection, and cancer.[1]
[2] Membrane-proteins such as G-protein coupled receptors (GPCRs) and tyrosine (Y)-protein
kinase-linked receptors mediate the activation of platelets in response to numerous
agonists.[3]
[4] GPCRs include receptors for thrombin (PAR1, PAR4), thromboxane A2 (TP), and ADP (P2Y1, P2Y12)[2]
[5]; tyrosine kinase-linked receptors are glycoprotein VI (GPVI), C-type lectin receptor-2
(CLEC-2), GPIbα, FcγRIIA, and integrin αIIbβ3.[2]
[6] Since there is a growing need to control platelet hyperreactivity, new approaches
to inhibit platelets are sought.[2]
GPVI, a platelet-specific tyrosine kinase-linked collagen receptor, signals via the
Fc receptor γ-chain and the spleen tyrosine kinase (Syk), similar to the signaling
of the B cell receptor (BCR) and related immune receptors.[7]
[8] B cell studies originally established that a BCR-/immunoreceptor tyrosine-based
activation motif (ITAM)-induced and membrane-associated Src/Syk/PI3K/Btk/PLCγ2 signalosome
with additional signaling components (protein kinase C [PKC], Akt, calcineurin, mitogen-activated
protein kinase [MAPK] nuclear factors) controls final B cell responses.[9]
[10]
The BCR signalosome concept was readily transferred to platelets and the ITAM-based
signaling complexes induced by GPVI or CLEC-2 activation.[8]
[9] GPVI agonists including collagen, cross-linked collagen-related peptide (CRP-XL),
and the snake venom toxin convulxin, activate platelets via Src family kinases (SFKs),
which induce dual Y-phosphorylation of proteins with the ITAM.[9] This recruits the SH2 domain-containing Syk to the membrane followed by SFK-mediated
Syk Y352 phosphorylation and kinase activation associated with autophosphorylation
(Y525/526).[9]
[11] Syk substrates are linkers for activation of T cells (LAT) and Bruton's tyrosine
kinase (Btk), which stimulate Y-phosphorylation/activation of phospholipase Cγ 2 (PLCγ2)
and platelet activation.[12]
GPVI stimulation also activates phosphoinositide 3-kinases (PI3Ks) by their recruitment
to membranes via their SH2-domain binding to Y-phosphorylated proteins such as LAT
or by interaction with GPCR βγ subunits.[13] Activated PI3Ks via their product phosphatidylinositol (3,4,5)-trisphosphate (PIP3) also recruit PH-domain containing proteins (i.e., PLCγ2, Btk, Akt) to the membranes
and activate them.[13]
However, the sequential signaling events in ITAM-mediated responses, the interactions
of Y- with S/T-protein kinases, the crosstalk with other pathways, and the coupling
to specific functional responses are not well defined. Recent developments of selective
and potent kinase inhibitors of the ITAM signaling components Syk, PI3K, and Btk provide
novel approaches to study this pathway in human platelets.[2]
Syk inhibitors and first-/second-generation Btk inhibitors, ibrutinib and acalabrutinib,
respectively, inhibited various functions of human platelets and tyrosine phosphorylation
of downstream targets.[14]
[15]
[16]
[17]
[18] Earlier studies showed that ibrutinib strongly inhibits many tyrosine kinases including
SFKs and Tec in addition to Btk, whereas acalabrutinib is Btk selective.[19] Antiplatelet effects of several novel Btk/Syk inhibitors and reversible/irreversible
Btk inhibitors were compared,[20]
[21] but with limited phosphorylation data. In contrast, extensive phosphoproteomic data
were obtained with GPVI-activated human platelets[22] and compared with functional effects of several Syk and Btk inhibitors.[23] These studies provided substantial evidence for the important role of tyrosine kinases
and especially Btk in GPVI-induced platelet activation, but the phosphorylation data
had limitations. Most studies used only one time point after activation, indicating
that the data reported represent a static snapshot, and not a dynamic view on platelet
signaling. Most papers also primarily addressed Y-phosphorylation of Syk, Btk, and
their direct substrates LAT and PLCγ2, but not additional S/T protein kinases. PKC
activation was analyzed by phosphoantibodies against PKC consensus phospho-sites,[14]
[22] similarly also Akt and MAPK.[22] However, these procedures do not detect specific substrates.
Many tyrosine protein kinases are known to be regulated by S/T protein kinases, but
this has been rarely studied in human platelets. In our previous phosphoproteomic
approaches, we detected numerous ADP- and/or prostacyclin (PGI2)-regulated phosphoproteins
in human platelets and noted that ADP stimulated S/T phosphorylation of several tyrosine
protein kinases such as JAK3, Btk, TNK2, and Syk.[24] Previously, we focused on the regulation of the individual platelet tyrosine protein
kinases Syk and Btk. Analysis of Syk serine phosphorylation in response to GPIbα stimulation
by beads coated with the toxin echicetin, GPVI stimulation by the toxin convulxin,
and by ADP detected prominent Syk S297 phosphorylation preferentially dependent on
PKCα/β.[25]
[26] Syk S297 phosphorylation was negatively affected by PKA or the protein phosphatase
2A (PP2A) and correlated well with reduced Syk Y-phosphorylation/kinase activity.[26]
[27] We also compared the differential regulation of Syk and Btk by PKC, PKA, and PP2A
in human platelets and noted that PP2A does not directly affect Btk pS180.[28]
Now, we hypothesized that the analysis of GPVI-activated multisite protein phosphorylation
kinetics in human platelets identifies interactions of key Y-kinases (SFK, Syk, Btk)
and S/T-kinases (PKC, MEK1/2, Erk1/2, p38, Akt) within a signaling network, which
is required for granule exocytosis and αIIbβ3 integrin activation. To define time-dependent
dynamics of GPVI signaling in human platelets, we quantified relevant convulxin-regulated
phosphoproteins (14 phospho-sites) during an extended activation time of 10 to 300 seconds
and six different time points. Interactions of GPVI-stimulated Y- and S/T-protein
kinases and resulting functional effects on platelets, in particular regulation of
intracellular Ca2+ mobilization, degranulation, and αIIbβ3 integrin activation, were assessed by the
effects of selective inhibition of Syk, Btk, and PKC.
Methods
Reagents and Antibodies
Convulxin and GF109203X were obtained from Enzo Life Sciences (Lausen, Switzerland).
PRT-060318 was from Selleckchem (Houston, Texas, United States). Acalabrutinib was
purchased from Abcam (Cambridge, United Kingdom). AR-C669931 (AR-C) was from the Medicines
Company (Parsippany, New Jersey, United States). Total Syk (4D10), total PLCγ2 (B-10),
and total Akt1 (B-1) antibodies were from Santa Cruz Biotechnology (Dallas, Texas,
United States). Phospho-Syk (S297, Y352, Y525/526), total Btk (D3H5), phospho-Btk
(S180, Y223, Y551), phospho-PLCγ2 (Y759, Y1217), phospho-LAT Y220, phospho-Akt (T308,
S473), phospho-MEK1/2 S217/S221, phospho-Erk1/2 T202/Y204, phospho-p38 T180/Y182,
total MEK1/2, α-actinin, and β-actin were provided by Cell Signaling Technologies
(Danvers, Massachusetts, United States). Secondary horseradish peroxidase (HRP)-conjugated
goat anti-rabbit/mouse antibodies were obtained from BioRad Laboratories (Hercules,
California, United States). FITC-conjugated mouse anti-human CD62P, CD63, and PAC-1
antibodies were from BD Biosciences (Heidelberg, Germany).
Blood Donors, Ethics Approval, and Informed Consent
Blood collection was performed as previously described.[28] This study was approved in accordance with the Declaration of Helsinki by the local
Ethics Committee of the University Medical Center Mainz (study no. 837.302.12; 25.07.12;
2018–13290_1; 27.07.2018).
Isolation of Human Platelets and Preparation of Fura-2-Loaded Platelets
For immunoblotting and flow cytometry, human platelets were washed and isolated as
previously described.[28] For the measurement of cytosolic Ca2+ rises, the protocol was slightly modified based on our previous publication.[29] Briefly, platelet-rich plasma (PRP) was prepared via centrifugation at 260 × g for 10 minutes at room temperature (RT), supplemented with 1:10 vol/vol acid citrate
dextrose (ACD) (80 mM trisodium citrate, 183 mM glucose, 52 mM citric acid). Platelets
in PRP were pelleted by centrifugation at 2,360 × g for 2 minutes and resuspended in HEPES buffer (136 mM NaCl, 2.7 mM KCl, 2 mM MgCl2, 5.5 mM glucose, 10 mM HEPES, 0.1% BSA, pH 6.6) in the presence of apyrase (1 U/mL)
and 1:15 vol/vol ACD. After a further centrifugation step, platelets were resuspended
in HEPES buffer (150 mM NaCl, 5 mM KCl, 1 mM MgCl2, 5 mM glucose, 10 mM HEPES, pH 7.5) and loaded with Fura-2 acetoxymethyl ester (3
µM) and Pluronic (0.4 µg/mL) for 30 minutes at RT. Fura-2-loaded platelets were adjusted
to a platelet concentration of 2 × 108/mL with HEPES buffer, pH 7.5.
Light Transmission Aggregometry
Light transmission aggregometry (LTA) was performed as previously described.[28] Briefly, using an Apact4S Plus aggregometer (DiaSys, Flacht, Germany), washed human
platelets (200 µL, 3 × 108/mL) were preincubated with vehicle control (0.1% DMSO), 1 µM PRT-060318, 5 µM acalabrutinib
or 5 µM GF109203X for 5 minutes at 37°C and then stimulated by 50 ng/mL convulxin
under stirring. Samples for immunoblotting were collected at 0, 10, 30, 60, 120, and
300 seconds by adding 100 µL of 3× Lämmli buffer (200 mM Tris/HCl, 15% (v/v) glycerol,
6% (w/v) SDS, 0.06% (w/v) bromophenol blue, 1:10 β-mercaptoethanol), and boiled for
10 minutes at 95°C with gentle shaking.
SDS-PAGE and Immunoblotting Analysis of Phosphoproteins
SDS-polyacrylamide electrophoresis (SDS-PAGE), immunoblotting, and phosphoprotein
analysis were performed as previously described.[28] Briefly, proteins in prepared samples were separated by 8% gels and transferred
to polyvinylidene fluoride membranes. After blocking membranes with 2% BSA in 1× TBS-T
for 1 hour at RT, the membranes were incubated overnight at 4°C with specific antibodies
with 2% BSA in 1× TBS-T. The incubated membranes were washed three times with 1× TBS-T
and incubated with relevant HRP-conjugated secondary antibody for 2 hours at RT with
2% BSA in 1× TBS-T, and then rewashed three times with 1× TBS-T. The membranes were
developed by electrochemiluminescence (ECL) detection. The antibodies used are listed
above. An alternative fluorescence-based detection system [IRDye 800 CW goat anti-rabbit
(LiCor); Bio-Rad ChemiDoc MP Imaging System] was also used with selected platelet
samples and compared with our regular ECL detection system, and found superiority
of the latter in terms of signal sizes and signal-to-noise ratios.
Measurement of Calibrated Changes in Cytosolic Ca2+ Rises
Using a high-throughput FlexStation 3 device (Molecular Devices, San Jose, United
States), the elevation of intracellular Ca2+ in Fura-2-loaded platelets was measured in 96-well plates as previously described.[30] Briefly, 200 µL of platelets/well (2 × 108/mL) were pretreated with PRT-060318, acalabrutinib, or GF109203X for 10 minutes at
37°C. The platelet suspension was supplemented with vehicle (HEPES buffer, pH 7.5)
or 0.1 mM EGTA as needed before starting the measurement. Convulxin (final concentration
50 ng/mL) was injected by automatic pipetting at a high rate of 125 µL/s for maximal
platelet response. For each column, fluorescence responses were monitored for 10 minutes
at 37°C, recording 510 nm emission fluorescence at two excitation wavelengths (340
and 380 nm). The Fura-2 fluorescence ratio of each well was acquired every 4 seconds.
The calibration wells contained Fura-2-loaded platelets and 0.1% Triton-X-100 in the
presence of either 2 mM CaCl2 or 1 mM EGTA/Tris for determining Rmax and Rmin values, thus resulting in nanomolar
changes in intracellular Ca2+. Duplicate time traces capturing nanomolar changes in intracellular Ca2+ concentration were subjected to floating point averaging via an Excel script. These
traces were subsequently assessed for the area under the curve (AUC; expressed in
nM × s) over a 10-minute period.[31] To facilitate comparisons across different experimental days involving various blood
donors, curve parameters were normalized against the control condition, specifically
when the agonist was administered with a vehicle medium and no inhibitor. This control
condition was standardized at 100%. The subsequent analysis of inhibitor effects involved
expressing these effects as percentage changes relative to the established control
condition. Notably, the normalization process was independently conducted for experimental
runs featuring either CaCl2 or EGTA.
Flow Cytometry
After incubation with PRT-060318, acalabrutinib, or GF109203X for 5 minutes or with
AR-C669931 for 15 minutes at 37°C, washed human platelets (2 × 108/mL) were stimulated with 50 ng/mL convulxin for 5 minutes at RT. Platelets were stained
with PAC-1-FITC (recognizing activated αIIbβ3 integrin), anti-CD63-FITC, or anti-CD62P-FITC
antibodies for 10 minutes at RT and then fixed with formaldehyde in HEPES buffer (final
concentration 0.5%) for 30 minutes at RT, followed by adding 1 mL HEPES buffer (pH
7.4) to stop fixation. After centrifugation at 800 × g for 10 minutes, platelets were resuspended in 500 µL of HEPES buffer (pH 7.4) and
analyzed by flow cytometry using a BD FACSCANTO II and FACS DIVA software (BD biosciences,
Heidelberg, Germany) as previously described.[32]
Statistical Analysis
Data are presented as means ± standard deviation (SD), from n ≥ 3 independent experiments with platelets from at least three healthy donors. Statistical
analysis was performed using GraphPad Prism 9.5.1 (GraphPad Software, San Diego, California,
United States). One-way or two-way ANOVA, followed by a Sidak's multiple comparison
test, was used for the comparison of more than two groups. A p-value <0.05 was considered significant.
Results
Convulxin Induces Very Rapid, Reversible Y-phosphorylation, Followed by Delayed, Also
Reversible Multisite S/T-phosphorylation of Syk, Btk, and Downstream Targets
For the kinetic experiments, we preferred to use the tetrameric C-type lectin convulxin,
since it dimerizes and further clusters very rapidly the GPVI receptor molecules in
contrast to the cross-linked CRP polypeptide. Using convulxin, we were able to distinguish
between very early, intermediate and later protein phospho-sites and also to detect
reversible protein phosphorylation within 5 minutes.
Convulxin induces phosphorylation of Syk (Y352, Y525/526, S297) and Btk (Y551, Y223,
S180) in human platelets.[28] These earlier studies used a limited activation window time, a small phospho-site
spectrum, and primarily addressed Syk and Btk. Here, we significantly extended the
time courses to compare early phosphorylation changes (at 10 seconds and 30 seconds
and later events up to 300 seconds). Furthermore, we enlarged the spectrum of analyzed
phospho-sites to other protein kinases with suspected involvement in long-term GPVI
signaling. These include the MAPK-related kinases, MEK1/2, Erk1/2 and p38, and the
PI3K-dependent kinase Akt.[13]
[33]
[34] We analyzed phospho-sites known to be clearly linked to activation of the defined
protein kinases and/or to serve as a direct substrate of these kinases ([Supplementary Table S1], available in the online version).
In human platelets stimulated with the GPVI agonist convulxin, we determined, with
several time points within 300 second activation, the phosphorylation effects using
a panel of 14 specific phospho-antibodies, recognizing the following phosphoproteins:
Syk (S297, Y352, Y525/526), LAT (Y220), Btk (S180, Y223, Y551), PLCγ2 (Y759, Y1217),
MAPKs (MEK1/2 S217/221, Erk1/2 T202/Y204, p38 T180/Y182), and Akt (T308, S473). We
found an equally rapid and strong convulxin-induced Y-phosphorylation of Syk and Btk,
which was maximal within 10 seconds, reversible, and preceded their S-phosphorylation
([Fig. 1A, B]). Next to PLCγ2 and LAT, also the MAPKs MEK1/2, Erk1/2, and p38 showed a reversible
pattern of phosphorylation ([Fig. 1A, B]). In contrast, the Akt T308 and S473 phosphorylation was slower in onset and more
persistent ([Supplementary Fig. S1], available in the online version). Overall, this kinetic analysis indicated that
the Y-phosphorylation of Syk, LAT, Btk, and PLCγ2 preceded the S/T phosphorylation
of Syk, Btk, MEK1/2, Erk1/2, p38, and Akt. At longer time points, all phosphorylation
events, except for Akt, reversed within the 5-minute stimulation period, indicating
the activity of Y/S/T protein phosphatases.
Fig. 1 Time-dependent phosphorylation of multiple protein kinases and substrates in human
platelets induced by convulxin. Washed human platelets were stimulated with 50 ng/mL
convulxin (cvx) at 37°C under stirring, and activation was stopped after 0, 10, 30,
60, 120, or 300 seconds with Lämmli buffer. (A) Representative blots showing convulxin-stimulated protein phosphorylation of: Syk
S297, Syk Y352, Syk Y525/526, LAT Y220, Btk S180, Btk Y223, Btk Y551, PLCγ2 Y759,
PLCγ2 Y1217, MEK1/2 S217/221, Erk1/2 T202/Y204, and p38 T180/Y182. Antibodies against
total Syk, total Btk, total PLCγ2, total MEK1/2, and α-actinin were used as loading
controls. (B) The phosphorylation of Syk (i), Btk (ii), LAT (iii), PLCγ2 (iv), MEK1/2 S217/221
(v), Erk1/2 T202/Y204, and p38 (vi) was analyzed, and compared with staining for total
Syk, total Btk, α-actinin, total PLCγ2, total MEK1/2, and α-actinin, respectively.
Quantitative data are represented as mean ± SD from three independent experiments
with platelets from three healthy donors (n = 3). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, 0 second versus other time points of convulxin-treated platelets. SD, standard
deviation.
Inhibition of Syk, Btk, or PKC Differentially Affects the Convulxin-Induced Multisite
Protein Phosphorylation
To further investigate the hierarchy of protein kinases in the GPVI signaling pathway,
we then systematically studied the effects of Syk, Btk, or PKC inhibitors on protein
phosphorylation events in the GPVI signalosome.
In human platelets, the compound PRT-060318 (1 µM) has been characterized as a selective
Syk inhibitor.[25]
[35]
[36] We first confirmed that PRT-060318 blocked the convulxin-induced Syk Y525/526 phosphorylation,
as an indicator of Syk activity ([Fig. 2]). However, it did not inhibit, even increased Syk Y352 phosphorylation, an established
SFK substrate sequence. Syk inhibition also strongly suppressed Syk S297 phosphorylation
and other Y/S/T phospho-sites, i.e., LAT, Btk, PLCγ2, MEK1/2, Erk1/2, p38, and Akt
([Fig. 2], [Supplementary Fig. S1], available in the online version). Overall, this indicated that these phosphorylation
events are downstream of the activated Syk kinase.
Fig. 2 General suppression of convulxin-induced phosphorylation by Syk inhibitor PRT-060318
(PRT) except for Syk Y352. Washed human platelets were treated with 0.1% DMSO or 1
µM PRT-060318 (PRT) for 5 minutes, prior to stimulation with 50 ng/mL convulxin (cvx)
under stirring, and activation was stopped after 0, 10, 30, 60, 120, or 300 seconds
with Lämmli buffer. (A) Representative blots show convulxin-stimulated protein phosphorylation with or without
PRT-060318 preincubation including Syk S297, Syk Y352, Syk Y525/526, LAT Y220, Btk
S180, Btk Y223, Btk Y551, PLCγ2 Y759, PLCγ2 Y1217, MEK1/2 S217/221, Erk1/2 T202/Y204,
and p38 T180/Y182. Antibodies against total Syk, total Btk, total PLCγ2, total MEK1/2,
and α-actinin were used as loading controls. (B) Phosphorylation of Syk (i–iii), Btk (iv–vi), LAT (vii), PLCγ2 (viii–ix), MEK1/2 S217/221
(x), Erk1/2 T202/Y204 and p38 T180/Y182 (xi–xii) was analyzed, and compared with total
Syk, total Btk, α-actinin, total PLCγ2, total MEK1/2, and α-actinin, respectively.
Quantified data are mean ± SD from three independent experiments with platelets from
three healthy donors (n = 3). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, DMSO versus PRT-060318-treated platelets in response to convulxin at the
same time points. Note, original blots of convulxin-induced protein phosphorylation
are identical to those of [Fig 1]. SD, standard deviation.
Second, we compared the roles of Syk and Btk in the GPVI-signalosome by using the
Btk inhibitor, acalabrutinib. Platelet treatment with acalabrutinib abolished the
phosphorylation of Btk Y223 ([Fig. 3]), a well-known Btk autophosphorylation site representing Btk activity ([Supplementary Table S1], available in the online version), under basal and convulxin-stimulated conditions.
Acalabrutinib also strongly inhibited the convulxin-stimulated phosphorylation of
PLCγ2 Y759/Y1217, MEK1/2 S217/221, Erk1/2 T202/Y204, p38 T180/Y182, Syk S297, Btk
S180 ([Fig. 3]), and Akt T308/S473 ([Supplementary Fig. S2], available in the online version), but not of Syk Y352, Y525/526, LAT Y220, and
Btk Y551 ([Fig. 3]). Interestingly, acalabrutinib, similar to PRT, moderately increased Syk Y352 phosphorylation,
which was statistically not significant. These data hence indicate that Btk, but not
Syk, controls PLCγ2 Y-phosphorylation/activation, Syk S297/Btk S180 phosphorylation,
and MEK1/2, Erk1/2, p38, Akt activation. However, also Syk may regulate these functions
indirectly via Btk.
Fig. 3 Partial suppression of convulxin-induced phosphorylation by Btk inhibitor acalabrutinib,
but not Syk Y352, Y525/526, LAT Y220, and Btk Y551. Washed platelets were treated
with 0.1% DMSO or 5 µM acalabrutinib for 5 minutes, prior to stimulation with 50 ng/mL
convulxin (cvx) under stirring, and activation was stopped after 0, 10, 30, 60, 120,
or 300 seconds with Lämmli buffer. (A) Representative blots showing convulxin-stimulated protein phosphorylation with or
without acalabrutinib preincubation: Syk S297, Syk Y352, Syk Y525/526, LAT Y220, Btk
S180, Btk Y223, Btk Y551, PLCγ2 Y759, PLCγ2 Y1217, MEK1/2 S217/221, Erk1/2 T202/Y204,
and p38 T180/Y182. Antibodies against total Syk, total Btk, total PLCγ2, total MEK1/2,
and α-actinin were used as loading controls. (B) Phosphorylation of Syk (i–iii), Btk (iv–vi), LAT (vii), PLCγ2 (viii–ix), MEK1/2 S217/221
(x), Erk1/2 T202/Y204, and p38 T180/Y182 (xi–xii) was analyzed and compared with total
Syk, total Btk, α-actinin, total PLCγ2, total MEK1/2, and α-actinin, respectively.
Quantitative data are represented as the mean ± SD from three independent experiments
with platelets from three healthy donors (n = 3). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, DMSO versus acalabrutinib-treated platelets in response to convulxin at the
same time points. SD, standard deviation.
Third, we examined the role of PKC activity in the GPVI-signalosome. Previously, we
characterized GF109203X (5 µM) as a potent and selective human platelet pan-PKC inhibitor
and reported that PKC isoforms (most likely PKC α/β) regulate Syk S297 and Btk S180
phosphorylation as a negative feedback mechanism.[27]
[28] With the extended time courses and the broader spectrum of analyzed phosphoproteins,
we now monitored effects of GF109203X on convulxin-induced protein phosphorylation
of Syk, LAT, Btk, PLCγ2, Akt, and MAPKs. GF109203X clearly enhanced the convulxin-induced
Y-phosphorylation of Syk Y525/526, LAT Y220, and PLCγ2 Y759/1217, moderately increased
Syk pY352, while it strongly reduced S-phosphorylation of Syk S297 and Btk S180 ([Fig. 4]). Importantly, GF109203X also abolished the phosphorylation of MEK1/2 S217/221 and
Erk1/2 T202/Y204, but not of p38 T180/Y182 ([Fig. 4]). GF109203X inhibited convulxin-induced Akt S473 phosphorylation, especially at
5 minutes, but not of Akt T308 ([Supplementary Fig. S3], available in the online version).
Fig. 4 Differential suppression of convulxin-induced S/T phosphorylation by PKC inhibitor
GF109203X, leaving p38 T180/Y182 unchanged. Washed human platelets were treated with
0.1% DMSO or 5 µM GF109203X (GFX) for 5 minutes at 37°C prior to stimulation with
50 ng/mL convulxin (cvx) under stirring, and stopped after 0, 10, 30, 60, 120, or
300 seconds with Lämmli buffer. (A) Representative blots showing convulxin-stimulated protein phosphorylation with(out)
GF109203X preincubation for: Syk S297, Syk Y352, Syk Y525/526, LAT Y220, Btk S180,
Btk Y223, Btk Y551, PLCγ2 Y759, PLCγ2 Y1217, MEK1/2 S217/221, Erk1/2 T202/Y204, and
p38 T180/Y182. Antibodies against total Syk, total Btk, total PLCγ2, total MEK1/2,
or α-actinin were used as loading controls. (B) Phosphorylation of Syk (i–iii), Btk (iv–vi), LAT (vii), PLCγ2 (viii–ix), MEK1/2 S217/221
(x), Erk1/2 T202/Y204 and p38 T180/Y182 (xi–xii) was analyzed and compared with total
Syk, total Btk, α-actinin, total PLCγ2, total MEK1/2 and α-actinin, respectively.
Quantitative data are represented as mean ± SD from three independent experiments
with platelets from three donors (n = 3). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, DMSO versus GF109203X-treated platelets in response to convulxin at the same
time points. SD, standard deviation.
At the functional level of platelet activation, we confirmed our results[25]
[26]
[28] that both 1 µM PRT-060318 and 5 µM acalabrutinib abolished the convulxin-induced
platelet aggregation, while 5 µM GF109203X had an only slightly inhibitory effect
(data not shown). We then extended these studies to the analysis of convulxin-stimulated
intracellular Ca2+ levels, granule secretion, and αIIbβ3 integrin activation.
Inhibition of Btk but Not PKC Suppresses Convulxin-Induced Intracellular Ca2+ Rises
The GPVI-dependent activation of PLCγ2 requires multisite Y-phosphorylation and recruitment
to the membrane via its SH2- and/or PH domain.[13] Activated PLCγ2 catalyzes the conversion of phosphatidylinositol 4,5-trisphosphate
(PIP2) to IP3 resulting in the mobilization of Ca2+ from the dense tubular system to the cytoplasm with concomitant platelet activation.[12] The data so far indicated that the convulxin-induced PLCγ2 Y-phosphorylation (Y759,
Y1217) was inhibited by PRT-060318 and acalabrutinib but was increased by GF109203X.
Strong Inhibition of GPVI-activated PKC (likely α/β isoforms) by GF109203X was demonstrated
by the downregulation of Syk S297 and Btk S180 phosphorylation.
To further clarify this phenomenon, regulation of cytosolic Ca2+ in Fura-2-loaded platelets was monitored using a high-throughput FlexStation 3 robot
system. Experiments were performed in the absence and presence of EGTA to elucidate
the effects on intracellular Ca2+ mobilization. Acalabrutinib (0.3–10 µM) dose-dependently inhibited the Ca2+ rise upon convulxin stimulation in either condition, as visualized by AUC analysis
([Fig. 5A]). Even 1 µM acalabrutinib had strong inhibitory effects. In contrast, GF109203X
caused a different pattern of agonist response. At low doses, GF109203X (0.3–1 µM)
slightly increased, whereas at high doses it did not reduce the Ca2+ mobilization induced by convulxin ([Fig. 5B]).
Fig. 5 Acalabrutinib, not GFX109203X, suppresses convulxin-induced Ca2+ mobilization in platelets. Fura-2-loaded human platelets were stimulated by 50 ng/mL
convulxin (cvx), after preincubation with acalabrutinib or GF109203X. Calibrated intracellular
Ca2+ rises were recorded for 10 minutes in 96-well plates at 37°C. Area under the curve
(AUC) within 10 minutes of convulxin-stimulated intracellular Ca2+ rises was normalized to 100%. Shown are normalized AUC of intracellular Ca2+ rises in response to convulxin in the presence of acalabrutinib (A) or GF109203X (B). Mean ± SD from at least three independent experiments with platelets from healthy
donors (n ≥ 3). ****p < 0.0001, ***p < 0.001, DMSO versus acalabrutinib or GF109203X-treated platelets in response to
convulxin. SD, standard deviation.
Inhibition of Syk, Btk, and PKC Affects Convulxin-Induced Degranulation and αIIbβ3
Integrin Activation
Platelet granule secretion is tightly regulated by multiple platelet agonists.[37] Since both elevated Ca2+ level and PKC activity are essential for platelet degranulation,[37]
[38] we determined effects of the same Syk, Btk, and PKC inhibitors on the convulxin-induced
degranulation and, in comparison, αIIbβ3 integrin activation. At 1 µM PRT-060318,
1 to 5 µM acalabrutinib and 5 µM GF109203X strongly inhibited, while at 1 µM GF109203X
only partially reduced the expression of CD63 ([Fig. 6A]) and CD62P ([Fig. 6B]), surface expression markers for δ- and α-granule exocytosis, respectively. Similar
inhibitory effects were observed for the convulxin-stimulated PAC-1 antibody binding,
as a marker for αIIbβ3 integrin activation ([Fig. 6C]). These data indicate that both GPVI-induced platelet granule secretion and αIIbβ3
integrin activation require all three protein kinases investigated, Syk, Btk, and
PKC.
Fig. 6 Inhibitors of Syk, Btk, and PKC strongly down-regulate convulxin-induced platelet
granule secretion and αIIbβ3 integrin activation. Washed platelets were pretreated
with 0.1% DMSO, 1 µM PRT-060318 (PRT), 1–5 µM acalabrutinib, 1–5 µM GF109203X (GFX)
for 5 minutes, before stimulation with 50 ng/mL convulxin (cvx) for 5 minutes. Surface
expression of CD63 (A) and CD62P (B), and FITC PAC-1 antibody binding (C) were assessed by flow cytometry. The mean fluorescence intensity in control condition,
established when the convulxin was administered with a vehicle medium and no inhibitor,
served as 100%. Subsequent analysis of inhibitor effects entailed expressing these
effects as percentage changes relative to this established control condition. Percentage
data are shown, with convulxin condition normalized to 100%. Mean ± SD from four donors
(n = 4). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, indicated inhibitor versus DMSO. SD, standard deviation.
P2Y12 Receptor Blockade Partially Impairs Convulxin-Stimulated Granule Secretion
The release of ADP from the δ-granules and subsequent activation of the platelet ADP
receptors P2Y1 and especially P2Y12 are positive feedback mechanism to enhance the functional effects of GPVI agonists.[33]
[39] We confirmed that the P2Y12 receptor antagonist AR-C669931 dose-dependently (2.5–1,000 nM) reduced the convulxin-stimulated
CD63 and CD62P surface expression, and the PAC-1 antibody binding ([Fig. 7]). Interestingly, treatment of platelets with 500 nM AR-C669931 resulted in partial
reduction of δ-granule (CD63) ([Fig. 7A]) and α-granule (CD62P) ([Fig. 7B]) release of approximately 50%, whereas αIIbβ3 integrin activation was nearly abolished
([Fig. 7C]). The data indicate a differential regulation, at least quantitatively, of α/δ-granule
secretion and integrin activation by secreted ADP in response to GPVI stimulation.
Fig. 7 The P2Y12 receptor blocker AR-C669931 suppresses convulxin-induced platelet granule secretion
and αIIbβ3 integrin activation. Washed platelets were pretreated with 0.1% DMSO, 2.5–1,000 nM
AR-C669931 (AR-C) for 15 minutes before stimulation with 50 ng/mL convulxin (cvx)
for 5 minutes. Surface expression of CD63 (A) and CD62P (B), and FITC PAC-1 antibody binding (C) were analyzed by flow cytometry. The control condition, established when the convulxin
was administered with a vehicle medium and no inhibitor, served as 100%. Subsequent
analysis of inhibitor effects entailed expressing these effects as percentage changes
relative to this established control condition. Results are shown as %; stimulation
by convulxin was normalized to 100%. Mean ± SD from three healthy donors (n = 3). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, DMSO versus AR-C669931-treated platelets in response to convulxin. SD, standard
deviation.
Discussion
In this study, GPVI-stimulated multisite protein phosphorylation in human platelets
demonstrates a multiple-phased kinetic pattern and interactions between tyrosine-
and serine/threonine protein kinases with reversible activation and feedback regulation,
thereby controlling degranulation and αIIbβ3 integrin activation.
GPVI Stimulation by Convulxin Induces Reversible Phosphorylation of S/T-kinases after
Initial Y-phosphorylation of the GPVI-LAT Signalosome in a Partly PKC-Dependent Manner
The revealed transient, reversible nature of convulxin-induced Y/S/T-phosphorylation
indicates a powerful role of both Y- and /S/T-protein phosphatases in human platelets.[40]
[41]
[42] For instance, the prominent tyrosine phosphatase TULA-2 dephosphorylates Y-phosphorylated
Syk and antagonizes GPVI-signaling,[43] whereas the ST protein phosphatase PP2A dephosphorylates Syk pS297 and components
of MAPK signaling.[27]
[42] Interestingly, platelet tyrosine phosphorylation induced by GPVI and CLEC-2 activation
was sustained for 50 minutes, when aggregation was prevented by eptifibatide,[18] suggesting that dephosphorylation can also be controlled.
Most phospho-sites studied here are located within the kinase domains and are indicators
for their activation, namely Syk Y525/526, Btk Y551, MEK1/2 S217/221, Erk1/2 T202/Y204,
p38 T180/Y182, and Akt T308/S473. Other sites are located within regulatory domains,
which are essential for kinase activation (Syk Y352) or have regulatory effects (Syk
S297, Btk S180, Btk Y223). The Syk and Btk phospho-sites are well studied at the kinase
level,[9]
[10]
[11] and documented in the PhosphoSitePlus database.
Understanding the hierarchy, interactions, and functional impact of the protein kinases
requires information on the human platelet proteome and kinome,[41]
[44] properties of the protein kinase inhibitors and on their effects on platelets.[2] Because of the important role of Syk (abundant in human platelets, ∼0.78 µM) in
inflammation and immune cell diseases, several potent Syk inhibitors have been developed
for clinical use. The Syk inhibitor PRT-060318 strongly inhibited purified Syk (IC50 4 nM), and the activation and function of Syk in murine and human platelets.[25]
[35]
[36] It was reported that Syk, Btk, and PKC inhibitors did not block convulxin-induced
phosphorylation of Syk Y352, a SFK-specific phospho-site essential for Syk activation,[11] indicating that SFKs are still operative under these conditions. We also found that
PRT-060318 (1 µM) even prolongs the GPVI-induced Syk Y352 phosphorylation, suggesting
that Syk inhibition enhances SFK activation, and Syk activation down-regulates the
SFK-increased Syk pY352. It was reported that acalabrutinib-induced Btk inhibition
caused Src potentiation in human platelets,[14] but this was less apparent in our experiments. On the other hand, PRT-060318 abolished
Syk Y525/526 and other phospho-sites studied, indicating that Syk acts upstream of
LAT, Btk, PLCγ2, PKC, and MAPKs. In immune cells, LAT is phosphorylated primarily
by Syk at four conserved Y-sites (Y161, Y200, Y220, Y255), which serve as docking
sites for SH2-domain-containing proteins (e.g., PLCγ2/1, PI3K, Btk).[45]
Another important tyrosine kinase within platelet GPVI signaling is Btk, a member
of the Tec family.[12] Btk deficiency or dysfunction causes X-linked agammaglobulinemia (XLA), characterized
by a severe impairment of B cell development and function.[10] XLA platelets showed an only moderate impairment of GPVI signaling, likely due to
a redundant role of Tec.[12]
[15] Since Btk is crucially involved in B cell differentiation and malignancies, several
inhibitors of human Btk have been developed and clinically validated. Ibrutinib, a
first-generation Btk inhibitor, strongly and irreversibly inhibited Btk (targeting
Cys 481), but it also showed off-target effects on other tyrosine kinases including
SFKs. The second-generation inhibitor acalabrutinib is more specific for Btk and Tec
(Btk IC50 5 nM, Tec IC50 83 nM, no effect on SFKs) and also irreversibly binds to Btk Cys 481.[19] In this landmark clinical study, acalabrutinib plasma levels of 1.2 to 1.4 µM were
detected in individuals taking 100 mg of acalabrutinib twice daily. This was accompanied
by high Btk target occupancy in peripheral blood monocytes (>90%) with concomitant
reduction of Btk Y223 autophosphorylation.[19] Based on this and the considerable expression levels in human platelets (Btk ∼1.76
µM, Tec ∼0.21 µM), we analyzed the effects of acalabrutinib on platelet protein phosphorylation.
Acalabrutinib (5 µM) strongly and specifically inhibited phosphorylation of Btk Y223,
PLCγ2, and MAPKs. An earlier report on B cells showed for PLCγ2 that phosphorylation
of four conserved Y-sites (Y753, Y759, Y1197, Y1217) is essential for full PLCγ2 activation,
mediated primarily by Btk.[46] In the present case, acalabrutinib abolished the GPVI-induced phosphorylation of
PLCγ2 Y759 and Y1217, but not of Syk Y352 (SFKs site), Syk Y525/526 (Syk autophosphorylation),
LAT Y220 and Btk Y551 (Syk site). Thus, the GPVI-induced PLCγ2 Y759/Y1217 phosphorylation
is directly mediated by Btk, and indirectly controlled by Syk and SFKs, which is consistent
with another report.[15]
Our results on a GPVI-induced sequential phosphorylation in platelets are summarized
in [Supplementary Table S1] (available in the online version) and compared with database information. Useful
protein kinase activity markers appear to be for: (1) SFK activity: Syk pY352; (2)
Syk activity: Syk pY525/526, LAT pY220, Btk pY551, and (3) Btk activity: Btk pY223,
PLCγ2 pY759/Y1217. It appears that the other S/T phospho-sites studied, inhibited
by acalabrutinib or GF109203X, are downstream of both Btk and PKC. Furthermore, p38
pT180/Y182 and Akt pT308 events are downstream of Btk, but not of PKC.
Markers for Tyrosine and Serine/Threonine Protein Kinase Activities in the Platelet
Btk-PLCγ2-PKC Signalosome
In B cells, Btk activates PLCγ2 followed by the activation of PKCβ, MAPKs, calmodulin/calcineurin,
Akt, and several transcription factors, which are all important for B cell development
and functions.[10] Much less is known about related pathways in platelets. In platelets, PLCγ2 activation
induces IP3-mediated intracellular Ca2+ mobilization from the dense tubular system with subsequent Ca2+-dependent responses and also increases diacylglycerol formation with activation of
PKC isoforms.[3]
[4] In agreement with this, acalabrutinib (5 µM) abolished the GPVI-induced intracellular
Ca2+ elevation and the Syk S297 and Btk S180 phosphorylation. Previously, we reported
that the GPVI-induced Syk S297 and Btk S180 phosphorylation is mediated by one of
the conventional PKC isoforms, and that this represents a possible inhibitory feedback
mechanism of GPVI signaling.[26]
[28] Our present data show that the PKC inhibitor GF109203X (likely via PKCα/β) abolishes
GPVI-induced phosphorylation of Syk pS297 and Btk pS180, but enhances the Y-phosphorylation
of Syk, LAT, and PLCγ2, as well the intracellular Ca2+ rise. These data are consistent with B cell studies showing that the PKCβ-mediated
phosphorylation of Syk S297 and Btk S180 attenuates the membrane localization and
activation of these protein kinases.[47]
[48]
Interestingly, Akt decreased the activity of several BCR signaling targets, including
Btk, Blink, and Syk.[49] In platelets, the convulxin-induced and PKC-mediated phosphorylation of Syk S297
and Btk S180 is rapid (maximal within 30 seconds and closely follows Btk/PLCγ2 activation),
and is therefore a useful marker for GPVI-induced PKC activity. Inhibition of PKC
with GF109203X strongly inhibited the phosphorylation of more distal PKC targets such
as MEK1/2 S217/221, Erk1/2 T202/Y204, and Akt S473, but not p38 T180/Y182 or Akt pT308.
PKC-Dependent Protein Phosphorylation Phases Induced by GPVI-Mediated Btk Activation
Are Required for Granule Exocytosis and αIIbβ3 Integrin Activation
Based on the observed effects of the protein kinase inhibitors PRT-060318 (Syk), acalabrutinib
(Btk), and GF109203X (likely PKCα/β isoforms) on GPVI-stimulated multisite phosphorylation,
we got evidence that the interactions of Syk, Btk, and PKC control platelet granule
secretion and αIIbβ3 integrin activation. However, the PKC inhibitor GF109203X at
1 µM showed less strong inhibitory functional effects. Possibly, inhibitor concentrations
saturating for PKC were not reached, as the intraplatelet concentration of the conventional
PKCα and PKCβ isoforms can be calculated as approximately 2.9 µM. Increasing the concentration
of GF109203X to 5 µM resulted in a similarly strong inhibition of these responses
as with PRT-060318 or acalabrutinib. While PKC and Ca2+-regulated pathways are known to regulate platelet granule release,[37] other protein kinases such as MAPKs and Akt may also contribute. In a previous phosphoproteomic
study, we observed that multiple proteins of the platelet secretory machinery are
regulated at the phosphorylation level by ADP.[24] This is relevant in the present context since ADP is released from the δ-granules
and subsequently stimulates the P2Y1 and P2Y12 receptors.[39] We found that the P2Y12 blocker AR-C669931, at 1 µM, reduced the convulxin-induced δ-granule and α-granule
exocytosis up to 50%, whereas this abolished the αIIbβ3 integrin activation, suggesting
that the convulxin-stimulated platelet degranulation is only partly controlled by
the P2Y12 pathway. Interactions of P2Y12 and GPVI signaling are clinically relevant, as ADP enhances GPVI signaling.[50]
Taken together, we identified a GPVI-induced multi-phased interactive network of Y-
and S/T-protein kinases in human platelets, which is characterized by regulative multisite
phosphorylation patterns ([Fig. 8]). Syk down-regulates SFK activity indicated by decreased Syk pY352. PKC not only
activates MAPKs, but also directly phosphorylates Syk S297 and Btk S180 as negative
feedback regulation. GPVI-induced granule exocytosis and αIIbβ3 integrin activation
are primarily mediated by PKC and involve autocrine ADP release which enhances this
process, but additional protein kinases are likely to participate. Acalabrutinib,
an approved irreversible Btk/Tec inhibitor, specifically inactivates human blood cell
and platelet Btk kinase activity (chemical kinase knockout) at clinically observed
concentrations. There is increasing evidence that not linear signaling pathways, but
rather rapid phosphorylation–dephosphorylation cycles of multiple proteins underlie
the regulation of complex and important biological functions.[51] Of clinical relevance, the human GPVI signaling network contains multiple Y/S/T
protein kinases, which are already targeted by an increasing number of Y- and S/T-protein
kinase inhibitors already used in patients with malignant and/or inflammatory diseases,
often in combination. There is a need to enhance the understanding of the GPVI-controlled
human platelet signaling network and the possible benefits of novel interventions.
Fig. 8 Model showing bi-directional links between GPVI signalosome effects and platelet
functions. Central (green background area) is the signalosome network of multiple
tyrosine kinases (SFK, Syk, Btk) and serine/threonine kinases (PKC, MEK1/2, Erk1/2,
p38, Akt), which are (in)activated by multisite phosphorylation changes. The phosphorylation
of PLCγ2, MEK1/2, and Erk1/2, similarly as platelet functional responses, is Btk-
and PKC-dependent, which includes mechanisms of feedback inhibition (Syk S297 and
Btk S180 phosphorylation) and feedforward enhancement (P2Y12 stimulation by secreted ADP). Indicated pathways were revealed by specific inhibition
of Syk, Btk, or PKC. While this model represents the data obtained, additional proteins,
phospho-sites, and alternative signaling pathways can be involved as well. For further
details, see text.
What is known about this topic?
-
GPVI is a platelet-specific tyrosine kinase-linked collagen receptor, which signals
via the ITAM-containing Fc receptor γ-chain.
-
Src-family kinases (SFKs), spleen tyrosine kinase (Syk), and Bruton's tyrosine kinase
(Btk) control proximal GPVI-induced tyrosine phosphorylation events.
-
Multiple tyrosine as well as serine/threonine protein kinases may play a role in GPVI-induced
platelet responses.
What does this paper add?
-
Analysis of GPVI-stimulated multisite protein phosphorylation in human platelets demonstrates
a multiple-phased kinetic pattern and interactions between tyrosine- and serine/threonine
protein kinases with reversible activation and feedback regulation.
-
After initial tyrosine phosphorylation of GPVI-LAT signalosome components, delayed
phases of serine/threonine phosphorylations occur, partially PKC-dependent (SykS297, BtkS180, MEK1/2S217/221, ErkT202/Y204, AktS473) and partially PKC-independent (p38T180/Y182, AktT308).
-
PKC-dependent protein phosphorylation phases induced by GPVI-mediated Btk activation
are required for granule exocytosis and αIIbβ3 integrin activation.
