Combined deficiency of RAB32 and RAB38 in the mouse mimics Hermansky-Pudlak syndrome and critically impairs thrombosis

Abstract

The biogenesis of lysosome related organelles is defective in Hermansky-Pudlak syndrome (HPS), a disorder characterized by oculocutaneous albinism and platelet dense granule (DG) defects. The first animal model of HPS was the fawn-hooded rat, harboring a spontaneous mutation inactivating the small guanosine triphosphatase Rab38 This leads to coat color dilution associated with the absence of DGs and lung morphological defects. Another RAB38 mutant, the cht mouse, has normal DGs, which has raised controversy about the role of RAB38 in DG biogenesis. We show here that murine and human, but not rat, platelets also express the closely related RAB32. To elucidate the parts played by RAB32 and RAB38 in the biogenesis of DGs in vivo and their effects on platelet functions, we generated mice inactivated for Rab32, Rab38, and both genes. Single Rab38 inactivation mimicked cht mice, whereas single Rab32 inactivation had no effect in DGs, coat color, or lung morphology. By contrast, Rab32/38 double inactivation mimicked severe HPS, with strong coat and eye pigment dilution, some enlarged lung multilamellar bodies associated with a decrease in the number of DGs. These organelles were morphologically abnormal, decreased in number, and devoid of 5-hydroxytryptamine content. In line with the storage pool defect, platelet activation was affected, resulting in severely impaired thrombus growth and prolongation of the bleeding time. Overall, our study demonstrates the absence of impact of RAB38 or RAB32 single deficiency in platelet biogenesis and function resulting from full redundancy, and characterized a new mouse model mimicking HPS devoid of DG content.

Graphical abstract

Figure 1.

Expression of RAB38 and RAB32 proteins in MKs and platelets. (A, left panel) Western blots of human, mouse, and rat total platelet (Plts) lysates (10 µg) showing expression of RAB38 in all 3 species but not of RAB32 in rat platelets. Actin was present as a loading control. Representative blots from 3 independent experiments. (A, right panel) Reverse transcription polymerase chain reaction amplification of mRNA from rat lung, spleen, and platelets and mouse platelets, indicating absence of RAB32 mRNA in rat platelets. Rat plts 1: OFA rat; rat plts 2: Wistar rat. (B) Confocal images of immunolabeled mouse MKs. Left panels show RAB32 or RAB38 labeling (green), 5-HT or VWF labeling (red), and the merged images. Right panels show for each labeling a line scan of the fluorescence intensity along the drawing line visualized in the merged images. Bar, 5 µm. Images are representative of at least 3 independent labeling experiments.

Figure 2.

Impact of Rab32 and/or Rab38 inactivation on coat pigmentation. (A) Western blots showing the absence of RAB32 and RAB38 in Rab32^−/− and Rab38^−/− mouse platelets, respectively, and in DKO mouse platelets, and partial RAB32 and RAB38 expression in [Rab38^−/−;Rab32^+/−] and [Rab38^+/−;Rab32^−/−] mouse, respectively. Representative of 3 independent western blots. (B) Different coat color dilution according to the mouse genotype.

Figure 3.

Impact of Rab32 and/or Rab38 inactivation on platelet granule content. (A) Quantification of DG constituents in mouse whole platelet lysates: 5-HT measured by ELISA (left); ADP (middle) and ATP (right) measured by HPLC separation. Results are the mean ± SEM in at least 3 mice; **P < .01, ***P < .0001 using 1-way analysis of variance (ANOVA). (B) Quantification of α granule constituents in mouse whole platelet lysates: PF4 (left) measured by ELISA, P-selectin (middle) and fibrinogen (right) measured by western blotting (mean ± SEM, n = 3; not significant [ns] using Student t test). (C) Quantification of β-hexosaminidase in mouse sera as a measure of lysosome content (mean ± SEM, n = 5, *P < .05 using Student t test).

Figure 4.

Decreased number and abnormal DGs in Rab32/Rab38 DKO mouse platelets. (A) TEM images showing the ultrastructure of WT and DKO platelets. Right, the close-up views illustrate the various abnormal DG morphologies in DKO mice. Images are representative of at least 3 independent platelet preparations. Scale bar, 1 µm. (B) Evaluation of mepacrine uptake by WT and DKO platelets. Bar graph represents the percentage of mepacrine-positive resting or degranulated platelets (prestimulation with 1 U/mL thrombin) following incubation with 10 µM mepacrine for 30 minutes (n = 3 independent experiments). ***P < .001 using 1-way ANOVA. (C) FIB/scanning electron microscopy images and 3D reconstructions of whole platelets. (Right panel) Dark spots represent DGs reconstructed by drawing the granule membranes on each slice. Note the heterogeneity of the size of the DGs in both genotypes. (Left panel) Quantification of the number of DGs per platelet using 3D FIB/scanning electron microscopy reconstructions (n = 25). (D) TEM images showing the presence of normal MVBs in DKO mouse MKs differentiated in culture from Lin⁻ cells compared with WT. Images are representative of at least 20 MKs. Scale bars, 500 nm.

Figure 5.

Defective hemostasis and thrombus growth in DKO mice. (A) Bleeding time measured as the time to the first cessation of bleeding. For DKO mice, bleeding was manually stopped at 1800 seconds; n = 10-12 mice, scatter plot with mean ± SEM, each dot corresponds to an individual mouse, ***P < .0001 with Kruskal-Wallis test and Dunn’s multiple comparison test. (B) TF-induced thromboembolism experiments. (Left panel) Percentage survival following tissue factor injection, n = 10 mice; statistics using a log-rank (Mantel-Cox) test. (Right panel) Individual and mean percentage of platelet consumption, n = 8-9 mice; **P < .01 using Student t test. (C) Laser-induced mesenteric arteriole injury. (Left panel) Curves representing the mean ± SEM thrombus area at each time point of 10 to 13 vessels observed from 7 to 9 mice. (Right panel) Bar graph of the area under the curve (AUC) of the individual curves corresponding to the graph shown in left. **P < .001, Kruskal-Wallis test. (D) FeCl₃-induced carotid artery injury. FeCl₃ was applied to the lateral side of the carotid for 150 seconds (arrow), after which thrombus growth was visualized from the top. (Upper left panel) Curves representing the mean ± SEM thrombus area at each time point, n = 10 mice. (Upper right panel) Representative top view showing the fluorescent platelet accumulation at the peak thrombus formation in WT mice (748 seconds) (upper image) and the absence of platelet accumulation in DKO mice (lower image). (Lower panel) Scatter bar graph of the AUC; **P < .001, ***P < .0001 using Kruskal-Wallis test and Dunn’s posttests.

Figure 6.

Defective platelet aggregate formation in DKO mice. (A) Whole blood anticoagulated with hirudin was perfused through glass capillaries coated with collagen I fibers for 5 minutes at a shear rate of 1500 seconds⁻¹. Scanning electron microscopy images that are representative of at least 3 independent experiments. (B) Aggregation tracings of WT (red) and DKO (blue) washed platelets stimulated with various concentrations of agonists as indicated. Representative of at least 4 independent experiments.

Figure 7.

Thrombin-induced integrin αIIbβ3 activation, P-selectin exposure and LAMP1 exposure. (A-F) Washed platelets were stimulated with increasing thrombin concentrations and periods of time (numbers in graph legend indicate the final thrombin concentration) and analyzed by flow cytometry. Red symbols, WT; blue symbols, DKO. All graphs represent mean ± SEM for n = 3 independent experiments. Statistical analyses were performed using 2-way ANOVA, Bonferroni posttest, to compare differences between WT and DKO in similar stimulation conditions. (A) αIIbβ3 integrin activation as measured by JonA-PE labeling and represented as mean fluorescence intensity. At each time point and for all thrombin concentrations, DKO platelet JonA-PE labelings were significantly different from WT ones (P < .001), except after stimulation with 0.05 U/mL thrombin for 10 seconds, where there was no difference. (B) αIIbβ3 integrin activation in the presence (dotted lines) or absence (straight lines) of ADP (10 µM). The presence of ADP significantly increases JonA-PE labeling in DKO mice compared with absence of ADP for all conditions (P < .001). (C) P-selectin exposure at the surface of platelets. No significant difference was observed between WT and DKO platelets, except after stimulation with 0.05 U/mL thrombin for 300 seconds (P < .01). (D) P-selectin exposure in the presence (dotted lines) or absence (straight lines) of ADP (10 µM). The presence of ADP did not increase P-selectin exposure. No significant difference between WT and DKO platelets. (E, left) LAMP1 platelet surface exposure. LAMP1 exposure is significantly decreased in DKO platelets compared with WT (P < .001) except for 0.1 U/mL thrombin at 30 seconds. (F) Exogenous addition of 10 µM ADP (dotted lines) only partially rescued DKO platelet LAMP1 exposure in response to 1 U/mL thrombin.