Platelet secretion is kinetically heterogeneous in an agonist-responsive manner

Abstract

Platelets release numerous bioactive molecules stored in their granules enabling them to exert a wide range of effects on the vascular microenvironment. Are these granule cargo released thematically in a context-specific pattern or via a stochastic, kinetically controlled process? Here we sought to describe the platelet exocytosis using a systematic examination of platelet secretion kinetics. Platelets were stimulated for increasing times with different agonists (ie, thrombin, PAR1-agonist, PAR4-agonist, and convulxin) and micro-ELISA arrays were used to quantify the release of 28 distinct α-granule cargo molecules. Agonist potency directly correlated with the speed and extent of release. PAR4-agonist induced slower release of fewer molecules, whereas thrombin rapidly induced the greatest release. Cargo with opposing actions (eg, proangiogenic and antiangiogenic) had similar release profiles, suggesting limited thematic response to specific agonists. From the release time-course data, rate constants were calculated and used to probe for underlying patterns. Probability density function and operator variance analyses were consistent with 3 classes of release events, differing in their rates. The distribution of cargo into these 3 classes was heterogeneous, suggesting that platelet secretion is a stochastic process potentially controlled by several factors, such as cargo solubility, granule shape, and/or granule-plasma membrane fusion routes.

Figure 1

Time course of release from activated platelets. Human platelets (1.2 × 10⁹/mL) were prepared as described and were stimulated with thrombin (0.3 U/mL, A), convulxin (0.3 μg/mL, B), PAR1-agonist (50μM, C), or PAR4-agonist (500μM, D) for the indicated times. Thrombin stimulation was stopped with hirudin (0.6 U/mL) followed by centrifugation; the rest of the reactions were stopped by centrifugation, and the releasates were recovered. Release from the 3 classes of granules was measured using the marker cargo molecules: [³H]-serotonin for dense granules (■), PF4 for α-granules (▴), and β-hexosaminidase for lysosomes (▾). Percent release was calculated using the equation [(Releasate)/(Pellet + Releasate)] × 100. Each time point was measured in triplicate, and the averages and SDs are indicated.

Figure 2

Cargo release in response to different agonists. Human platelets (1.2 × 10⁹/mL) were prepared as described and were stimulated with thrombin (0.3 U/mL, T; n = 3), convulxin (0.3 μg/mL, C; n = 2), PAR1-agonist (50μM, P-1; n = 2), or PAR4-agonist (500μM, P-4; n = 2) for 5 minutes. Releasates were then probed using the micro-ELISA arrays, and the presence (gray square) or absence (white square) of a given cargo molecule was recorded. Cumulative data indicate whether a cargo was released at least once in any of the secretion trials performed with the indicated agonist. Single donor indicates the cargo released by platelets from a single donor in response to different agonists. PAR1/PAR4 indicates a direct comparison of the data for cargo release in response to stimulation of either the PAR1 or the PAR4 receptor. Relative abundance indicates the number of molecules of each cargo per resting platelet and was measured using the micro-ELISA arrays.

Figure 3

Probability density function (pdf) of the frequency distribution of K_ex values. Human platelets (1.2 × 10⁹ /mL) were prepared as described and were stimulated with thrombin (0.3 U/mL; n = 3), convulxin (0.3 μg/mL; n = 2), PAR1-agonist (50μM; n = 2), or PAR4-agonist (500μM; n = 2) for increasing times. Releasates were probed using the micro-ELISA arrays, and the rate constant, K_ex of release was calculated for each cargo molecule. The pool of K_ex values with MSE < 150 was used for constructing the pdf. The best-fit curve is seen in black corresponding to the existence of 3 different subclasses. The dashed blue, green and red curves are the Gaussian functions that represent fast, intermediate, and slow classes, respectively.

Figure 4

Distribution of K_ex values. The rate constants, K_ex, were calculated as described, and the average was obtained for each from multiple trials involving various agonists for stimulating the platelets. The values are divided into 3 different classes represented by different colors based on their distribution in pdf in Figure 3. The cargo in gray, white, and dark gray correspond to the fast, intermediate, and slow classes, respectively.

Figure 5

Cargo characteristics from modeling operator variance. The distribution of the simulated K_ex values (open symbols) assuming 3 (A,B,D) or 2 (C) cargo classes with K_ex values equaling the mean values obtained from the 3-Gaussian fit shown in Figure 3. The closed symbols represent the 90 K_ex values obtained from our experimental measurements. The Index is simply the ordinal number of the K_ex value. The rate constants and the operator variances used for the analysis are varied in different panels to get a better overlap with the obtained experimental K_ex values. The description of the variable in each panel is as follows: (A) Three cargo classes, K_ex = 0.010, 0.067, and 0.148; σ_op = 0.2. (B) Three cargo classes, K_ex = 0.010, 0.033, and 0.148; σ_op = 0.2. (C) Two cargo classes, K_ex = 0.010 and 0.148; σ_op = 0.2. (D) Three cargo classes, K_ex = 0.010, 0.033, and 0.148; σ_op = 0.1.

Figure 6

Frequency distribution of K_ex resulting from activation by different agonists. The rate constants, K_ex, were calculated as in “Methods.” The gray bars represent the frequency distribution of K_ex values obtained on stimulation of platelets with various agonists as indicated. Superimposed dashed curves represent the Gaussian functions shown in Figure 3. The number of K_ex values used in making the pdfs incorporate 43 K_ex values for thrombin (A), 19 K_ex values for PAR1-agonist (B), and 17 K_ex values for convulxin (C).