Platelet dense granule secretion plays a critical role in thrombosis and subsequentvascular remodeling in atherosclerotic mice

Abstract

Background

Platelet aggregation plays a critical role in myocardial infarction and stroke; however the role of platelet secretion in atherosclerotic vascular disease is poorly understood. Therefore, we examined the hypothesis that platelet dense granule secretion modulates thrombosis, inflammation, and atherosclerotic vascular remodeling after injury.

Methods and Results

Functional deletion of the Hermansky Pudlak Syndrome 3 gene (HPS3^−/−) markedly reduces platelet dense granule secretion. HPS3^−/− mice have normal platelet counts, platelet morphology, alpha granule number and maximal secretion of the alpha granule marker P-selectin; however, their capacity to form platelet-leukocyte aggregates was significantly reduced (p<0.05). To examine the role of platelet dense granule secretion in these processes, atherosclerosis-prone mice with combined genetic deficiency of ApoE and HPS3 (ApoE^−/−,HPS3^−/−) were compared to congenic, atherosclerosis-prone mice with normal platelet secretion (ApoE^−/−,HPS3^+/+). After 16–18 weeks on a high fat diet, both groups of mice had similar fasting cholesterols and body weight. Carotid arteries of ApoE^−/−,HPS3^+/+ mice rapidly thrombosed after FeCl₃-injury but ApoE^−/−, HPS3^−/− mice were completely resistant to thrombotic arterial occlusion (p<0. 01). Three weeks after injury, neointimal hyperplasia (from alpha smooth muscle actin+ cells) was significantly less (p <0.001) in the arteries from ApoE^−/−,HPS3^−/− mice. In ApoE^−/−, HPS3^−/− mice, there were also pronounced reductions in arterial inflammation, as indicated by a 74% decrease in CD45+ leukocytes (p< 0.01) and a 73% decrease in Mac-3+ macrophages (p<0.05).

Conclusions

In atherosclerotic mice, reduced platelet dense granule secretion is associated with marked protection from the development of arterial thrombosis, inflammation and neointimal hyperplasia after vascular injury.

Introduction

Vascular injury is a critical step in the pathogenesis of coronary artery disease. Platelets become activated at sites of vascular injury and secrete their alpha and dense granule contents. Platelet alpha granules are the primary storage organelle for adhesive and pro-inflammatory molecules such as P-selectin, CD40L, and RANTES,^{1, 2} but platelet dense granules also contain cell activating molecules (serotonin, histamine, adenine nucleotides, etc), that may be considered to be pro-inflammatory.³ Under pathophysiologic conditions, local delivery of adhesive and pro-inflammatory molecules released from platelet granules may contribute to atherosclerosis and neointima formation after injury.^{4, 5}

Previous studies have used antiplatelet agents to examine the role of platelets in arterial wall remodeling after injury. Aspirin, at doses that reduce platelet aggregation, had no effect on neointimal remodeling after acute injury in animal models.^{6, 7} Clinical studies using P2Y₁₂ receptor antagonists show that these agents do not reduce rates of restenosis after stent placement or angioplasty.^8–10 Data regarding the role of IIb/IIIa inhibitors remains controversial. Several clinical studies found treatment with IIb/IIIa inhibitors reduced the need for target vessel revascularization over conventional treatment,^{11, 12} while other studies found no benefit for restenosis.^{13, 14} While these antiplatelet agents may be used to determine the role of specific platelet activators or platelet aggregation (IIb/IIIa inhibitors), they cannot be used to directly address the role of platelet secretion in vessel remodeling after acute injury.

To determine the contributions of platelet dense granule secretion to thrombosis and atherosclerotic vascular remodeling, we used atherosclerosis-prone ApoE^−/− mice with defective platelet dense granule secretion due to a mutation in the Hermansky Pudlak Syndrome 3 (HPS3) gene that results in a functional gene knockout or deletion of HPS3 mRNA (i.e., HPS3^−/−).^15–17 HPS3-deficient (HPS3^−/−) mice have impaired platelet dense granule secretion due to defects in the biogenesis of these organelles.^15–17 We hypothesized that, when compared to atherosclerotic mice with normal platelet secretion (ApoE^−/−, HPS3^+/+), mice with impaired dense granule secretion (ApoE^−/−, HPS3^−/−) would have attenuated thrombotic occlusion, vascular inflammation and atherosclerotic remodeling after injury.

Materials and Methods

Animals

Mice (ApoE ^−/−, HPS3^−/−) were obtained from the Jackson Laboratory (Bar Harbor, ME) and are congenic with the C57/Bl6 line. Animals were housed and cared for in accordance with the Guide for the Care and Use of Laboratory Animals [DHEW (DHHS) Publication No. (NIH) 85–23, revised 1996, Office of Science and Health Reports, DRR/NIH, Bethesda, MD 20205]. Procedures involving laboratory animals were approved by the Institutional Animal Care and Use Committee of the Medical College of Georgia and the Standing Committee on the Use of Animals in Research and Teaching at Harvard University. For cholesterol assays, carotid injury model, and analysis of atherosclerosis, mice were placed on a high fat diet (TD.88137 Adjusted Calorie Diet, Harlan Teklad, 42% calories from fat) at 6 weeks of age. For all other assays, mice were maintained on normal chow.

Coagulation assays

Whole blood was anticoagulated using 1/10 volume of 3.8% sodium citrate. Citrated blood was centrifuged at 1500 g for 20 min to obtain plasma. A second centrifugation (8000 g for 5 min) was performed to remove any cellular debris. Centrifugations were carried out at room temperature. The sampleswere analyzed with the Dade Behring (Marburg, Germany)BFT II analyzer using reagents and standards designed for use in humanclinical testing. Prothrombin time (PT) assays were initiated using Dade Innovin Reagent (containing recombinant human tissue factor, thromboplastin and calcium ions) as described¹⁸. Activated partial thromboplastin time (aPTT) assays were performed using Dade Actin FS (purified soy phosphatides) and calcium chloride solution as described¹⁸. Dade Citrol Coagulation Control Level 1 and Level 2 were used as controls. The PT and aPTT assays were performed exactly as directed in the BFT II instruction manual, using 50 μl of undiluted mouse plasma. Differences between groups were assessed using the Mann Whitney U test.

Cholesterol Assays

Blood from animals fasted for 24 hours was obtained by retro-orbital sinus puncture and anticoagulated with 1/10 volume of sodium citrate (3.8%) and PPACK (10 μM). Blood was centrifuged at 1470 g at 4°C for 20 minutes to obtain plasma. Plasma was used immediately for analysis or frozen at −80°C. Cholesterol was measured using the Amplex Red Cholesterol Assay Kit (Molecular Probes, Eugene, OR) and detected on a Synergy HT Microplate Reader (BioTek, Winooski, VT) following the manufacturer’s protocol. Differences between groups were assessed using the Mann Whitney U test.

Platelet granule secretion assays

Blood was immediately collected from euthanized mice via cardiac puncture. Approximately 500–600 μl blood was drawn through a 27 gauge needle into a 1 ml plastic syringe with 70 μl 3.8% sodium citrate, 10 μM PPACK.

For measurement of secretion from alpha granules, whole blood was immediately diluted 6-fold with a modified Tyrode’s buffer (137 mM NaCl, 2.8 mM KCl, 1 mM MgCl₂, 12 mM NaHCO₃, 0.4 mM Na₂HPO₄, 5.5 mM glucose, 1 mM EDTA, 0.35% BSA, 10 mM Hepes, pH 7.4). P-selectin expression was monitored in diluted whole blood by flow cytometry (FACSCalibur, BD Biosciences, San Jose, CA) as described ¹⁹. Briefly, platelets were stimulated with agonists (as indicated) and incubated with biotinylated rat anti-mouse CD62P antibody or an isotype-matched control antibody followed by phycoerythrin-conjugated streptavidin to detect P-selectin by flow cytometry. Fluorescein isothiocyanate-conjugated rat anti-mouse CD41 antibody (BD Pharmingen, San Jose, CA) was used as a platelet identifier. Platelets were maximally stimulated with 1 mM PAR4p, 2 μM A23187 and 1 mM Ca-ion (RGDS (1.5 mM) and GPRP (0.8 mM) were also added to these samples to prevent aggregation and clot formation). In a similar fashion, leukocyte-platelet aggregates were assessed by flow cytometry in whole blood stimulated with various amounts of PAR4 peptide as described.²⁰ Blood (10 ul) was labeled with (10ul) of 1/100 diluted antibodies for 30 minutes at room temperature. Blood was diluted 1/10 in modified Tyrode’s buffer containing 0.35% BSA, 50 mM EDTA, 20 micromolar P-PACK. The samples were fixed with 1% paraformaldehyde. Platelets were identified with anti-CD41 antibodies (see above) and anti-CD45 antibodies (BD Pharmingen, San Jose, CA). Isotype-matched control antibodies were used to measure nonspecific binding. The percent platelet – leukocyte aggregates was determined by assessing the proportion of CD41 and CD45 positive events in 5,000 leukocyte-gated events using the formula: % Leukocytes in platelet-leukocyte events = 100 × [(CD41,CD45 double positive cells)/(CD45 positive cells+CD41,CD45 positive cells)]. Differences between groups were assessed by a Mann Whitney U test.

Platelet aggregation assays

Aggregation studies were performed with isolated platelets or in whole blood using a lumi-aggregometer (Chrono-Log, Havertown, PA) according to the manufacturer’s instructions. To isolate the platelets from other blood cells, anticoagulated blood was diluted two-fold with modified Tyrode’s buffer and centrifuged at 150 g for 5 min. The platelet-rich upper phase was transferred to a fresh tube, and manual platelet counts were performed using a hemocytometer. Equivalent platelet counts were obtained by adding platelet poor plasma. The final platelet count was adjusted to 1.5×10⁸ cells/ml with modified Tyrode’s buffer. The samples were kept in a 37°C incubator and used within 2 hours. Chronolume reagent (Chrono-Log, Havertown, PA) was added to the samples to measure ATP secretion by luminescence while aggregation was measured by light transmission. Differences between groups were assessed using the Mann Whitney U test.

Electron Microscopy

Platelets (in platelet-rich plasma) were fixed with 4% paraformaldehyde/0.2% glutaraldehyde for 1 hr on ice. After washing with PBS, pellets were dehydrated with graded ethanol steps and embedded in LR White Resin (EMS, PA). Ultra-thin sections were stained with uranyl acetate and lead and viewed on a Jeol 1010 Transmission Electron Microscope (Jeol USA, Peabody, MA). Alpha granules were counted by an investigator blinded to the genotype of the specimen. Differences between groups were assessed using the Mann Whitney U test.

Carotid Injury Model

FeCl₃-induced arterial injury was induced according to published procedures.²¹ After the animal was adequately anesthetized with isoflurane (1.5–2%), a midline incision was made in the neck and the left common carotid artery was exposed by blunt dissection. Baseline blood flow was measured using a miniature Doppler flow probe (model 0.5PSB; Transonic Systems, Ithaca, NY) connected to a TS420 Flow Module (Transonic Systems, Ithaca, NY). The data were recorded and analyzed using the PowerLab 16/30 data acquisition unit and Chart software (AD Instruments, Colorado Springs, CO). After baseline blood flow was established, the carotid artery was carefully rinsed with sterile saline, prewarmed to 37°C, to remove any trace of acoustic couplant (SurgiLube, Fougera, Melville, NY). Thrombosis was induced by applying two pieces of filter paper (1×2 mm) saturated with 5% ferric chloride (Sigma, St. Louis, MO, USA). The pieces of filter paper were placed on opposite sides of the carotid artery (above and beneath) in contact with the adventitial surface of vessel. Care was taken to avoid regions with obvious atherosclerotic plaques. The filter paper was applied for 3 min and then removed. The carotid blood flow was monitored at time 0 (prior to the ferric chloride paper application), and continuously for 30 min following application of the filter paper. Differences in carotid blood flow between groups were assessed by a Mann Whitney U test. After the 30 min. monitoring period, vessels in the deeply anesthetized animals were perfusion-fixed with 4% paraformaldehyde.

For remodeling studies, the surgical opening was rinsed with sterile saline and the muscle and skin layers were closed using 6–0 sutures. The animal was kept on a 37°C heating pad until it was conscious and active (usually 2 hours), then it was placed in our animal housing facility and monitored daily. After 3 weeks, the animal was deeply anesthetized with isoflurane (1.5–2%) and vessels were perfusion-fixed as above.

Histology

For histological analysis, paraformaldehyde-fixed carotid vessels were processed and embedded in paraffin. Cross-sections (4 μm) were cut from paraffin blocks and mounted on treated slides (Superfrost Plus, VWR Scientific Products, Suwanee, GA). Equally spaced cross sections were stained with Verhoef-van Giesen’s stain. Digital images were captured on a Zeiss AXIOPHOT upright microscope with Axiocam camera (Carl Zeiss Microimaging, Thornwood, NY). Analysis was performed by an investigator blinded to the genotypes of the specimens. Intima was defined as the area bounded by the endothelium and internal elastic lamina. Media was defined as the area bounded by the internal and external elastic laminae. Intima and media areas of each cross section (typically 10 sections) were quantified using NIH ImageJ software, and summarized into a mean areas for each artery. The means of the mean intimal and medial areas for arteries in each group were compared by an unpaired Student’s t-test.

Immunohistochemistry

Cross-sections (4 μm) were cut from paraffin blocks and mounted on treated slides (Superfrost Plus, VWR Scientific Products, Suwanee, GA). Slides were air dried over night, then placed in a 60°C oven for 30 minutes. Slides were de-paraffinized and run through graded alcohols to distilled water. Slides were then pretreated with Target Retrieval Solution, pH 6.0 (Dako Corp., Carpinteria, CA) using a steamer, followed by a distilled water rinse. Endogenous peroxidases were quenched with 0.3% H₂O₂ for 5 minutes followed by distilled water for 2 minutes, and 1x PBS for 5 minutes.

Slides were incubated with primary antibody at the indicated dilutions for 30 minutes at room temperature: anti-P-selectin, 1:100, (Santa Cruz Biotechnology, Santa Cruz, CA), anti-CD45, 1:100 (BD Pharmingen, San Jose, CA), anti-α smooth muscle actin, 1:100 (Sigma, St. Louis, MO), anti Mac-3, 1:100 (BD Pharmingen, San Jose, CA). Antibody incubation was followed by two rinses with 1x PBS. Slides were then incubated with peroxidase-conjugated AffiniPure(F(ab′)) fragment of the appropriate species specificity (Jackson ImmunoResearch Laboratories, West Grove, PA) at 1:100 dilution for 1 hour, followed by two rinses with PBS. Bound antibody was detected with DAB substrate kit (Dako, Carpinteria, CA). Slides were then counterstained with hematoxylin (Richard-Allan Scientific, Kalamazoo, MI), mounted, and cover-slipped. Negative control slides were treated with an identical protocol but the primary antibody was replaced with negative control serum (Dako, Carpinteria, CA). The average immunochemical staining was assessed from multiple samples from each artery. The means of the mean immunochemical staining for arteries in each group were compared by an unpaired Student’s t-test.

Data Analysis

Statistical analyses are indicated above and differences between groups were considered to be significant if P < 0.05. The number of animals analyzed, n, is given in the figure legends. Data are reported as mean ± SEM.

Results

Previous studies have established that the secretion of platelet dense granule contents such as ATP and serotonin are markedly reduced in HPS3^−/− mice. ¹⁷ In confirmation of these findings, we found that the secretion of ATP was reduced (30-fold) in HPS3^−/− vs. wild-type congenic (HPS3^+/+) platelets (p<0.05). To examine whether HPS3^−/− mice retained normal reactivity to ADP (another dense granule marker) we measured platelet aggregation. When stimulated with ADP doses that do not induce secretion in wild-type HPS3^+/+ platelets,²² HPS3^−/− platelets showed equivalent aggregation (Fig. 1A–C) indicating that responsiveness to ADP was retained.

Figure 1. Platelet aggregation, secretion and morphology.

Platelet aggregation was induced by doses of ADP (5 uM) in HPS3^−/− mice (black bars) and congenic wild-type (HPS3+/+) mice (gray bars). The area under the curve (A), slope (B) and amplitude (C) of platelet aggregation curves is shown (N= 4 mice in each group). Transmission electron micrographs of resting platelets isolated from HPS3^−/− mice (D) or congenic controls (E, CTL). Scale bars = 500 nm. F. Leukocyte-platelet aggregates induced by different dose of Par4 peptide in HPS3^−/− mice (black bars) and congenic HPS3^+/+ mice (gray bars). The percent leukocyte platelet aggregates (LPA %) was assessed by flow cytometry (N=4 pairs of mice). G. Secretion of the alpha granule protein P-selectin by platelets after stimulation in whole blood with the calcium ionophore A23187 (0 or 2 uM). Platelet P-selectin expression (percent activation) was measured by flow cytometry in blood from HPS3^−/− mice (black bars) and congenic, HPS3^+/+ mice (gray bars, N= 4 pairs of mice). *p < 0.05).

When compared to wild-type mice, HPS3^−/− mice had equivalent platelet counts (1.25 ± 0.18 ×10⁹ and 1.35 ± 0.11 ×10⁹ platelets per mL, respectively). Electron microscopic analysis of ultrathin sections revealed that HPS3^−/− mice also had comparable numbers of alpha granules (4.65± 1.72 granules per platelet, Fig. 1D) with a morphology similar to congenic control mice (4.32± 2.09 granules per platelet, Fig. 1E). Consistent with this expression of the alpha granule protein P-selectin expression was the same in wild-type and HPS3^−/− platelets after maximal cell activation by calcium ionophore (Fig. 1G) or by thrombin receptor peptide and calcium ionophore, which are known to induce maximal platelet secretion (not shown).²³ However, leukocyte-platelet interactions were reduced in HPS3^−/− mice when compared to wild-type mice (Fig. 1F). Although previous reports have indicated that bleeding times are prolonged in HPS3^−/− mice,¹⁷ coagulation parameters were the same in HPS3^−/− and normal mice as assessed by measurements of prothrombin times (8.0 ± 0.3 vs. 7.9 ±0.3 s, respectively, n=7) and partial thromboplastin times (23.3 ± 2.0 vs. 23.8 ± 2.1 s, respectively, n=10).

To examine the effect of platelet dense granule secretion on thrombosis and vascular remodeling in atherosclerosis, HPS3^−/−, ApoE^−/− mice were compared to ApoE^−/−, HPS3^+/+mice. Mice were placed on a high fat diet beginning at 6 weeks of age to accelerate the development of atherosclerotic lesions. ApoE^−/−, HPS3^−/− mice exhibited similar cholesterol levels and weight as control mice (ApoE^−/−, HPS3^+/+) (Table 1) after 10–13 weeks on a high fat diet. Microscopic analysis confirmed that the mice developed atherosclerosis after 16–18 weeks on this high fat diet.

Table 1.

Parameters after 16–18 weeks on a high fat diet

	ApoE−/−, HPS3+/+	ApoE−/−, HPS3−/−
Body weight (g), females	27 ± 1	26 ± 2
Body weight (g), males	37 ± 8	37 ± 6
Fasting Cholesterol (mg/dL)	950 ± 140	940 ± 150

To examine the effects of platelet secretion on arterial thrombosis, carotid arteries were treated with FeCl₃ after mice had been on the high fat diet for 16–18 weeks. Occlusion of the carotid artery occurred in less than 10 min in all ApoE^−/−, HPS3^+/+ mice (Fig 2c). Microscopic examination confirmed that this was due to thrombotic occlusion (Fig. 2a, left panel). In contrast, none of the ApoE^−/−, HPS3^−/− mice experienced a decrease in blood flow due to thrombotic occlusion during the 30 minute study period (Fig 2c). The area under the curve of the relative Doppler blood flow vs. time showed that ApoE^−/−, HPS3^−/− mice were protected from arterial thrombotic occlusion (Fig 2c, P<0.01). No thrombotic occlusion was obtained in ApoE^−/−, HPS3^−/− mice even with high dose (15%) FeCl₃ (Fig. 2c). Histologic examination confirmed the absence of an occlusive thrombus in the arteries of ApoE^−/−, HPS3^−/− mice (Fig. 2b, right panel).

Figure 2. Reduced acute thrombotic occlusion after FeCl3-induced injury in ApoE^−/−, HPS3^−/− mice.

Arterial thrombosis was monitored with a Doppler flow probe after 5% ferric chloride application to the carotid arteries of anesthetized ApoE^−/−, HPS3^+/+ (n=6) and ApoE^−/−, HPS3^−/− mice (n=6). Representative arterial cross-sections from ApoE^−/−, HPS3^+/+ (A) and ApoE^−/−, HPS3^−/− (B) carotid arteries after 5% ferric chloride at 200× original magnification. C. Relative changes in blood flow (mean ± SEM) over time after application of the 5% FeCl₃ (ApoE^−/−, HPS3^+/+, gray boxes; ApoE^−/−, HPS3^−/−, black diamonds). *P<0.001. For comparison the relative blood flow is also shown for ApoE^−/−, HPS3^−/− mice (n= 3) after high dose (15%) ferric chloride application (open boxes).

The effect of arterial injury on atherosclerotic remodeling after was assessed three weeks after the injury. In the injured arteries, ApoE^−/−, HPS3^+/+ mice displayed significantly greater intimal hyperplasia (Fig. 3A) than the ApoE^−/−, HPS3^−/− mice (Fig. 3B, p < 0.001). In contrast, minimal if any intimal hyperplasia was observed in the contralateral, uninjured (control) arteries of either the ApoE^−/−, HPS3^+/+ or the ApoE^−/−, HPS3^−/− mice (Fig. 3A,B). There was no significant difference in the medial areas between the two groups. When assessed by intima to media ratios, FeCl₃-induced injury induced a 14-fold increase in the intima to media ratio in the ApoE^−/−, HPS3^+/+ mice (Fig 3C). However, injury-induced arterial remodeling, assessed by intima to media ratios, was reduced > 65% in ApoE^−/−, HPS3^−/−mice compared to ApoE^−/−, HPS3^+/+ mice (P < 0.05, Fig 3C).

Figure 3. Reduced arteriosclerotic vessel remodeling 3 weeks after FeCl₃-induced injury in HPS3^−/− mice.

A. Representative Verhoef-Van Gieson stained cross-sections from injured carotid arteries or uninjured contra-lateral control arteries taken 21 days after injury. Images originally taken at 200× magnification B. Ratios of intima to media from cross sections of FeCl₃-injured arteries and contralateral, control arteries 21 days after experimental injury. N= 5 mice per group.

Microscopic examination revealed that the cell population in the neointima was heterogeneous in both mouse genotypes, as shown by immunostaining for smooth muscle cells (α-smooth muscle actin), macrophages (Mac-3), and leukocytes (CD45). Although the lesions differed in size, smooth muscle cells were abundant in the neointima for both ApoE^−/−, HPS3^−/− and ApoE^−/−, HPS3^+/+ mice (Fig. 4). Immunostaining for the mouse common leukocyte antigen, CD45, showed that there was significantly less inflammatory cell accumulation within the neointima in ApoE^−/−, HPS3^−/− mice. In the ApoE^−/−, HPS3^+/+ mice, 30.5 ± 4.0 % of neointimal cells were CD45-positive, while only 8.0 ± 3.6% were CD45-positive in ApoE^−/−, HPS3^−/− lesions (n=5, p <0.01). Immunostaining with the activated macrophage marker, Mac-3 revealed that the inflammatory cells consisted primarily of macrophages. The neointima consisted of 25.2 ± 7.7 % Mac-3-positive cells in the ApoE^−/−, HPS3^+/+ mice, and only 6.9 ± 1.9% Mac-3-positive cells in ApoE^−/−, HPS3^−/− mice (n=5, p <0.05).

Figure 4. Reduced inflammation after FeCl3-induced injury in ApoE^−/−, HPS3^−/− mice.

Cross sections of carotid arteries 21 days after FeCl₃ injury were immunostained for smooth muscle αactin, macrophages (Mac3), and leukocytes (CD45). All images were initially taken with 400× magnification. For the ApoE^−/−, HPS3^−/− mice, the neointimal lesions were small and arrows indicate the internal elastic lamina to demarcate the boundaries of the lesions. For the ApoE^−/−, HPS3^+/+ mice the large neointimal lesions completely fill the figure.

Discussion

Mice with ApoE gene deletion develop atherosclerosis, that is marked by a mild pro-thrombotic phenotype, with lesions similar to that seen in humans.^{24, 25} After ferric chloride injury, ApoE^−/− mice, especially those on a high fat diet, have enhanced neointimal hyperplasia which contributes to luminal stenosis. ^25–27 This process involves migration of smooth muscle cells from the media across the internal elastic lamina, proliferation in the neointima, and deposition of extracellular matrix.²⁸ Inflammatory cells (leukocytes, macrophages, etc.) also contribute to the atherosclerotic, vascular remodeling process. Because platelets secrete potent cell activating agents at sites of vascular injury, we hypothesized that impaired dense granule secretion might protect against arteriosclerosis after arterial wall injury. Consistent with this notion, we found that thrombotic occlusion and atherosclerotic remodeling were markedly attenuated in the injured vessels from mice with defective platelet dense granule secretion, due to deficiencies of HPS3. The reduced atherosclerotic remodeling was associated with marked reductions in neointimal size and content of alpha actin-positive smooth muscle cells, CD45-positive leukocytes and Mac-3 positive macrophages.

The defective secretion of HPS3-deficient platelets is attributed to a dense granule storage pool disorder because these platelets lack adenine nucleotides and other molecules typically found in dense granules.^{15, 17} However, ultrastructural studies suggest that platelet alpha granule numbers and morphology are normal in HPS3-deficient platelets, as is their capacity to secrete the alpha granule marker P-selectin (Fig. 1). In addition, coagulation parameters also are normal (PT and PTT) in HPS3-deficient mice. Still, HPS3-deficiency markedly inhibits thrombosis in response to FeCl₃- injury. Indeed, unlike previous studies with ADP-receptor antagonists, aspirin or other agents, we could not overcome the anti-thrombotic effects of HPS3-deficiency even with high doses of FeCl₃ up to 15%.²¹

Diminished secretion of platelet dense granule contents may specifically affect vascular remodeling after injury. Molecules secreted from platelet dense granules may act directly on cell types that are involved in the growth of neointima after acute injury. In vitro, molecules found in dense granules have been shown to have direct effects on cells involved in atherosclerotic remodeling. Serotonin has a mitogenic effect on endothelial cells²⁹ and SMC³⁰; histamine and serotonin increase vascular permeability; ADP increases the agonist-induced oxidative burst in PMNs³ and amplifies SMC proliferation³⁰. Platelet granule secretion may accelerate atherosclerotic remodeling by delivering high concentrations of these molecules to the cells that form the neointima.

The thrombus itself also may play a role as an organizing matrix in neointima formation. Fibrin is an important determinant of neointima formation, providing a network to support SMC migration³¹. Endothelial denudation activates platelets to form a thrombus. However, only minimal, non-occlusive mural thrombus formed in ApoE^−/−, HPS3^−/− mice, probably because HPS3-deficiency interferes with amplification of platelet activation that normally occurs with dense granule secretion. The adhesive receptors on activated platelets present in the thrombus play a role in recruiting inflammatory cells (neutrophils, monocytes, lymphocytes) to the injured area through P-selectin secretion and other mechanisms; this has been shown to accelerate atherosclerotic lesion development ^32–35. Leukocyte-platelet interactions are reduced in ApoE^−/−, HPS3^−/− mice. In addition, as noted above, the diminished dense granule secretion in ApoE^−/−, HPS3^−/− mice may also decrease leukocyte activation and this may also affect vascular remodeling. Once recruited into the vessel wall, macrophages contribute to local inflammation through the secretion of inflammatory mediators including adhesion molecules, cytokines, chemokines, matrix metalloproteases and oxygen radicals^{36, 37}. Therefore a lack of thrombus, and/or alterations in cellular signaling or other mechanisms, could prevent the recruitment of inflammatory cells and migration of smooth muscle cells in the injured vessel wall and, contribute to the diminished vascular remodeling in the ApoE^−/−, HPS3^−/− mice.

Only small numbers of humans have been described with decreased platelet dense granule secretion. Some individuals with altered platelet dense granule secretion alone have been found to have increased circulating levels of serotonin with pulmonary hypertension.^{38, 39} However, among humans and mice with decreased dense granule secretion in the setting of Hermansky-Pudlak Syndrome, there is significant variation according to the genetic cause. Individuals with HPS-1 appear to be at risk for restrictive pulmonary disease.^{16, 40} However, humans and mice with HPS3-deficiency appear to have minimal in any other physiologic abnormalities.^{16, 41}

Our findings indicate that platelet secretion reduces thrombosis and subsequent vascular remodeling after injury. Whether platelet secretion affects vascular remodeling through mechanisms independent of its effects on thrombosis can’t be determined solely through vascular injury models. However, an early study suggested that some mice with platelet secretion disorders had reduced susceptibility to diet-induced atherosclerosis.⁴² It will be important to further elucidate these mechanisms and to explore the potential utility of platelet dense granule secretion as a therapeutic target in atherothrombotic disease.