Uncovering the role of the Hsp40 family member cysteine string protein-α in mouse platelets

Abstract

Platelets modulate vascular microenvironments via the release of cargo molecules. Granule secretion is modulated by proteins called soluble N-ethylmaleimide sensitive factor attachment protein receptors (SNAREs). Secretion is complex and regulated by several protein-protein interactions; however, not all are characterized in platelets. We have identified cysteine string protein-α (CSPα; also known as, DNAJC5 or CLN4) as required for platelet secretion. CSPα is the only member from the CSP family present in platelets and has been proposed as a chaperone for the SNAP-23/25 t (Qb,c) SNAREs. To address CSPα's role, we analyzed platelets from CSPα-/- mice. The loss of CSPα significantly affected dense- and α-granule release with minimal effects on lysosomal secretion. Consistent with the secretion defects, in vivo and ex vivo assays showed that loss of CSPα caused significant bleeding and attenuated thrombosis under flow. Interestingly, loss of CSPα caused a reduction in glycoprotein VI (GPVI) levels and reduced αIIbβ3 activation, especially in response to GPVI-specific agonists. Deletion of CSPα did not affect proteins in the platelet secretory machinery, for example, the SNAP-23/25 proteins. Subcellular fractionation studies showed that CSPα, which is reported to be acylated, was present on membranes but not in lipid rafts. Immunofluorescence studies showed CSPα colocalized with α and lysosomal granule markers. CSPα-/- mice had reduced red blood cell, leukocyte, and megakaryocyte numbers, suggesting effects on bone marrow progenitor cells. Simultaneously, we detected increased collagen I deposition, but no fibrosis in the marrow of CSPα-/- mice. These results identify CSPα as another element of the platelet secretory machinery that significantly contributes to thrombosis and hemostasis.

Graphical abstract

Figure 1.

SNARE machinery proteins levels remain unchanged in CSPα^−/− platelets. (A) Washed platelets (50 x 10³/µL platelets per lane) were prepared from CSPα^+/+, CSPα^+/−, and CSPα^−/− mice (n = 4), and the indicated proteins were probed by western blotting. Data are representative of 4 independent experiments. Because of molecular weight overlaps between probed proteins and loading controls, different controls were on different blots. Blots for CSPα, α-synuclein, and RabGDI were on the same membrane (RabGDI as control). VAMP-8, SNAP-23, and RabGDI were on the same membrane (RabGDI as control). Syntaxin-17 and RabGDI were on the same membrane (RabGDI as control). VAMP-3, VMAT2, and syntaxin-11 were on the same membrane (syntaxin-11 as control). Fibrinogen and syntaxin-11 were on the same membrane (syntaxin-11 as control). Platelet factor 4 and actin were on the same membrane (actin as control). GPVI and actin were on the same membrane (actin as control). (B) Quantification of protein levels was performed using ImageLab, and data were plotted as the ratio of CSPα^+/− to CSPα^+/+ (gray bars) and CSPα^−/− to CSPα^+/+ (red bars). Statistical analyses were done using individual values and performed using the unpaired nonparametric Mann-Whitney U test. (C) Washed platelets (50 x 10³/µL) from CSPα^+/+, CSPα^+/−, and CSPα^−/− were resting and incubated with fluorescein isothiocyanate–conjugated anti-GPVI antibodies for 20 minutes at 37°C. Fluorescent intensities were measured by flow cytometry. Shown are representative data and geometric mean fluorescent intensity (GMFI) (mean ± standard error) of 5 independent experiments. Statistical analyses were performed using the Kruskal-Wallis multiple comparison test and corrected using the Dunn multiple comparison test. Significance: ns = P > .05; ∗P ≤ .05; ∗∗P ≤ .01; ∗∗∗P ≤ .001; ∗∗∗∗P ≤ .0001. All error bars represent the standard error of the mean. AU, arbitrary units; HET, heterozygous; KO, knockout; ns, not significant; RabGDI, rab guanosine diphosphate dissociation inhibitor; VMAT2, vesicle monoamine transporter 2.

Figure 2.

CSPα^−/− platelets have defective dense- and α-granule secretion and integrin activation. (A-H) Washed platelets (50 x 10³/µL) from CSPα^+/+, CSPα^+/−, and CSPα^−/− mice were stimulated with 0.1 U/mL thrombin or 100 ng/mL convulxin for 2 minutes and then incubated with fluorescein isothiocyanate (FITC) anti–P-selectin (A,E), phycoerythrin (PE)-conjugated LAMP-1 (B,F), PE-conjugated Jon/A (C,G), or FITC anti-CD41/61 (D,H) antibodies for 20 minutes at 37°C. Fluorescent intensities were measured by flow cytometry. Shown are representative data and GMFI (mean ± standard error of mean) of 5 independent experiments. Statistical analyses were performed using the Kruskal-Wallis multiple comparison test and corrected using the Dunn multiple comparison test. The significant P values are indicated. (I-K) Platelet-rich plasma (PRP) was isolated and adjusted to a concentration of 100 x 10³/µL from CSPα^+/+, CSPα^+/−, CSPα^−/− mice, and Unc13d^Jinx mice. PRP was stimulated with 0.05 U/mL thrombin (I,J) or 100 ng/mL convulxin (K) over a 10-minute period at 37°C to measure ATP release from the platelets at 2 minutes increments. Data are mean ± standard error of the mean of triplicate measurements and are representative of 3 independent experiments. Statistical analyses were performed using 2-way analysis of variance multiple comparisons and corrected using the Tukey multiple comparison test. Significance: ∗(red) = P ≤ .05 WT vs KO and ∗(black) = P ≤ .05 WT vs HET. HET, heterozygous; KO, knockout.

Figure 3.

CSPα^−/− platelets have defects in signaling pathways and defective hemostasis. (A) Washed platelets (200 x 10³/µL) from CSPα^+/+ and CSPα^−/− mice were kept in a resting state (R) or stimulated with 0.1 U/mL thrombin (T) or 100 ng/mL convulxin (C) for 10 minutes and were probed by western blotting to look at phospho-Akt and phospho- myosin light chain (MLC) levels. (B) Quantification of protein levels was performed using ImageLab and data were plotted as either the ratio of phospho-Akt to Akt, or phospho-MLC to β-tubulin. (C) Washed platelets (200 x 10³/µL) from CSPα^+/+ and CSPα^−/− mice were kept in a resting state (R) or stimulated with 0.1 U/mL thrombin (T) or 100 ng/mL convulxin (C) for 10 minutes and were probed by western blotting to determine phosphorylation of PKC substrates, which was interpreted as PKC activity. (D) Quantification of protein levels was performed using ImageLab and data were plotted as the ratio of PKC activity to β-tubulin. (E) Tail bleeding assay was performed to analyze thrombus formation in vivo. CSPα^+/+ (n = 27; male = 16, and female = 11), CSPα^+/− (n = 83; male = 40, and female = 43), and CSPα^−/− (n = 9; male = 4, and female = 5) mice were used to perform the experiment. Statistical analyses were performed using the Kaplan-Meier method using the log-rank test. Akt, phosphatidylinositol 3-kinase (PI3K)/protein kinase B signaling pathway; phosphor, phosphorylated.

Figure 4.

CSPα^−/− mice have defective thrombosis under flow at low shear rates. A BioFlux microfluidics system was used to examine thrombus formation at a low shear rate of 10 dyn/cm² (A-E) or high shear rate of 35 dyn/cm² (F-J) over immobilized collagen for CSPα^+/+, CSPα^+/−, and CSPα^−/− mice. Historic data from the laboratory for Unc13d^Jinx mice was added for reference. (A) Representative images of thrombus formation were taken for CSPα^+/+ (n = 5), CSPα^+/− (n = 7), and CSPα^−/− (n = 4) mice at low shear rates during postperfusion washing. Quantitative analysis of platelet surface area coverage (B), morphological score (C), contraction score (D), and multilayer score (E) were measured at a low shear rate. (F) Representative images of thrombus formation were taken for CSPα^+/+ (n = 5), CSPα^+/− (n = 6), and CSPα^−/− (n = 5) mice at high shear rates during postperfusion washing. Quantitative analysis of platelet surface area coverage (G), morphological score (H), contraction score (I), and multilayer score (J) were measured at the high shear rate. Scaled from 0 (no thrombus formation) to 5 (fully formed contracted and multilayered thrombi) for morphological score. Contraction and multilayer scores were scaled from 0 (no thrombus formation) to 3 (fully formed contracted and multilayered thrombi). Statistical analyses were performed using individual values and the Kruskal-Wallis multiple comparison test and corrected using the Dunn multiple comparison test. The P values are indicated. Scale bar, 50 μm.

Figure 5.

CSPα is membrane-associated and present on both lysosomes and α-granules. (A) WT platelets were immunostained for CSPα (red) and LAMP-1 (green), and imaged using 3-dimensional structured illumination microscopy (3D-SIM). The white lines in the merged images indicate where the profile line analyses were performed. Profile line analyses are shown below the images. (B) WT platelets were immunostained for CSPα (red) and P-selectin (green), and imaged using 3D-SIM. Scale bar, 10 μm for widefield images; 5 μm for cropped images. (C) Resting platelets from CSPα^+/+ and CSPα^−/− mice were isolated and fixed in their resting state. Transmission electron microscopy micrographs were analyzed for platelet shape and granule distribution. Scale bar, 1 μm. (D) Resting platelets from CSPα^+/+ and CSPα^−/− mice were isolated, labeled with 1 μM mepacrine for 30 minutes, and imaged by epifluorescence microscopy to count dense-granule numbers. Scale bar, 5 μm. The number of mepacrine-positive granules was counted and graphed (Mann-Whitney U test, P = .4617). AU, arbitrary units.

Figure 6.

CSPα is membrane associated but not present in lipid rafts. (A) Lysates were prepared from washed human platelets subjected to 5 freeze-thaw cycles and centrifuged to separate membrane and cytosol (S1) fractions. The membrane fraction was treated with 1% Triton X-100 to generate Triton X-100–soluble (S_TX) and Triton X-100–insoluble (I_TX) fractions, which were separated by ultracentrifugation. The fractions were analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and probed by western blotting with the indicated antibodies. Data are representative of 2 independent experiments and quantification is based on the mean of the representative experiment. (B) Resting or 0.1 U/mL thrombin–stimulated human platelets were lysed with 2× lysis buffer and layered under a sucrose gradient. The samples were centrifuged and collected in 1-mL fractions. The fractions were analyzed by SDS-PAGE and probed by western blotting for the indicated antibodies syntaxin-11 and CSPα. Data are representative of 2 independent experiments. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; R, resting; S, stimulated.

Figure 7.

CSPα^−/− mice have reduced bone marrow megakaryocyte numbers and increased collagen I deposition. (A) Representative images of femur bone marrow from CSPα^+/+ and CSPα^−/− mice (aged 5-6 weeks). Collagen I (yellow), CD105 (purple), DAPI (blue), and CD41 (cyan). Scale bar, 20 μm. (B) Representative scan images of the femur from CSPα^+/+ and CSPα^−/− mice stained for CD41 (megakaryocyte). (C) Reticulin staining of marrow from CSPα^−/− mice. Scale bar, 50 μm. Supplemental Figure 8 for WT marrow and fibrotic liver controls. (D) Quantification of megakaryocytes per area, megakaryocyte size, and collagen I intensity in the femurs of CSPα^+/+, CSPα^+/−, and CSPα^−/− mice. Bones from CSPα^+/+ (n = 4; male = 3, and female = 1), CSPα^+/− (n = 3; male = 2, and female = 1), and CSPα^−/− (n = 4; male = 2, and female = 2) mice were evaluated. Megakaryocytes per area and megakaryocyte size were evaluated by examining the number and size of CD41⁺ cells. Quantification of collagen I fluorescence intensity was measured within each image. Statistical analyses were performed using individual values and the Kruskal-Wallis multiple comparison test and corrected using the Dunn multiple comparison test. The P values are indicated. Scale bar, 20 μm.