Dicer1-mediated miRNA processing shapes the mRNA profile and function of murine platelets

Abstract

Human platelets contain microRNAs (miRNAs) and miRNA processing machinery, but their contribution to platelet function remains incompletely understood. Here, we show that murine megakaryocyte (MK)-specific knockdown of Dicer1, the ribonuclease that cleaves miRNA precursors into mature miRNAs, reduces the level of the majority of miRNAs in platelets. This leads to altered platelet messenger RNA (mRNA) expression profiles and mild thrombocytopenia. Fibrinogen receptor subunits Itga2b (αIIb) and Itgb3 (β3) mRNAs were among the differentially expressed transcripts that are increased in platelets lacking Dicer1. Argonaute 2 (Ago2), a member of the miRNA silencing complex, co-immunoprecipitated with αIIband β3mRNAs in wild-type platelets. Furthermore, co-immunoprecipitation experiments suggested reduced αIIb/β3/Ago2 complexes in miRNA-deficient platelets. These results suggested that miRNAs regulate both integrin subunits. Subsequent 3' untranslated region luciferase reporter assays confirmed that the translation of both αIIband β3mRNAs can be regulated by miRNAs miR-326, miR-128, miR-331, and miR-500. Consistent with these molecular changes, the deletion ofDicer1resulted in increased surface expression of integrins αIIband β3, and enhanced platelet binding to fibrinogen in vivo and in vitro. Heightened platelet reactivity, shortened tail-bleeding time, and reduced survival following collagen/epinephrine-induced pulmonary embolism were also observed in Dicer1-deficient animals. CombinedPf4-cre-mediated deletion of Drosha and Dicer1 did not significantly exacerbate phenotypes observed in single Dicer1 knockout mice. In summary, these findings indicate that Dicer1-dependent generation of mature miRNAs in late-stage MKs and platelets modulates the expression of target mRNAs important for the hemostatic and thrombotic function of platelets.

Figure 1

Deletion of Dicer1 reduces mature miRNA level in platelets. (A) Schematic representing Pf4-cre recombinase-mediated deletion of Dicer1. LoxP sites flanking the second RNase IIIb domain of Dicer1 mediate its removal in the presence of Cre recombinase, which is under control of the MK and platelet-specific Pf4 promoter. (B) Real-time PCR amplification of platelet RNA from Dicer1^Pf4Δ/Pf4Δ or Dicer^fl/fl mice. The primers are designed to amplify within the Dicer RNase IIIb domain (top) or flank the Dicer RNase IIIb domain (bottom). Note that, in the Dicer1^Pf4Δ/Pf4Δ mice, the transcript is present in platelets, but truncated. (C) Western blots for Dicer1 and Gapdh in platelets from Dicer^fl/fl (lanes 1 and 3) or Dicer1^Pf4Δ/Pf4Δ (lanes 2 and 4) mice. (D) Dicer activity assay to monitor the processing of ³²P-labeled human pre–let-7a into mature let-7a using platelet lysates (or no lysate [–]; lane 1) from Dicer1^Pf4Δ/Pf4Δ (lane 3) or Dicer1^fl/fl littermate controls (lane 2). (E) Heat map depicting the clustering and relative fold changes of mature miRNAs, changed twofold or more, in platelets from Dicer1^Pf4Δ/Pf4Δ compared with Dicer1^fl/fl littermates (n = 3 mice per group. Each lane represents data from a single mouse/array for a total of 6 microarrays).

Figure 2

α_IIb and β₃ expression is increased in Dicer1-deficient platelets. (A) RNA-seq reads from Dicer1^Pf4Δ/Pf4Δ or Dicer1^fl/fl platelets aligning to α_IIb (top) or β₃ (bottom) mRNAs. Refseq annotations are included underneath each tracing, where thick and thin lines represent exons and introns, respectively. In these depictions, reads are piled up at their alignment location, so that both the density and height of histogram peaks are indicative of relative abundance. RNA-seq was performed on a pool of 3 mice per group. (B-C) Real-time PCR expression analysis of α_IIb or β₃ mRNA in (B) platelets or (C) BM-derived MKs from Dicer1^Pf4Δ/Pf4Δ vs Dicer1^fl/fl mice. *P < .05; **P < .01; unpaired Student t test. n = 9 to 13 mice per group for (B) and n = 4 mice per group for (C). (D) Surface expression of α_IIb and β₃ protein. Platelets from Dicer1^Pf4Δ/Pf4Δ or Dicer1^fl/fl mice were isolated and stained in littermate pairs for flow cytometry analysis. The gMFI of surface staining of α_IIb and β₃ for each pair are shown. Light gray lines connect each matched littermate pair, and the horizontal lines represent the mean for each group (**P < .01; paired Student t test; n = 10 pairs).

Figure 3

Platelet α_IIb and β₃ mRNA associates with Ago2. Pooled lysates (5 to 7 mice per group) from Dicer1^Pf4Δ/Pf4Δ or Dicer1^fl/fl mouse platelets were immunoprecipitated with an antibody against Ago2 or against the isotype control FLAG followed by immunoblotting (IB) for the indicated protein. (A) Total Ago2 protein expression in platelets from Dicer1^Pf4Δ/Pf4Δ or Dicer1^fl/fl mice. Blot shown is representative of 2 independent experiments. (B) Immunoprecipitated Ago2 protein expression in platelets from Dicer1^Pf4Δ/Pf4Δ or Dicer1^fl/fl mice. n = 1 platelet pools (pooled from 5 mice per group). (C) Immunoprecipitated RNA was extracted from IP, and α_IIb or β₃ mRNA association to Ago2 was estimated by quantitative real-time PCR (in duplicates). The results are normalized with Gapdh mRNA to calculate a fold change vs FLAG IP. Results are representative of 2 independent experiments demonstrating enrichment of α_IIb or β₃ mRNA in Ago2 IP in pools (pooled from 5 to 7 mice per group) of Dicer1^fl/fl platelet lysates, and n = 1 platelet pools per group (pooled from 5 mice per group) for the comparison with Dicer1^Pf4Δ/Pf4Δ animals.

Figure 4

miRNAs that are decreased in Dicer1-deficient platelets regulate α_IIb and β₃ expression through their 3′UTR. HEK293 cells were transfected with a reporter gene construct that contains the 3′-UTR sequence of α_IIb (A) or β₃ (B), located downstream of the Rluc open reading frame, and vectors encoding pre–mmu-miR-128, pre–mmu-miR-500, pre–mmu-miR-331, pre–mmu-miR-326, or a nonrelevant short hairpin control. Rluc and Firefly luciferase signals were measured and normalized to the nonrelevant hairpin control vector, which is represented by the dashed line (set at 1). Shown are mean ± SEM. Analysis of variance: P < .05; *P < .05 vs control; unpaired Student t test. n = 5 to 6 experiments performed in duplicate.

Figure 5

In vivo fibrinogen uptake and endogenous fibrinogen levels are increased in Dicer1-deficient platelets. (A) Dicer1^Pf4Δ/Pf4Δ or Dicer1^fl/fl mice were IV injected with Alexa Fluor 488 labeled fibrinogen. After 48 hours, platelets were isolated, stained with an anti-CD41 antibody, and incorporation of the label was analyzed by flow cytometry. The histogram is representative of 6 independent samples isolated from Dicer1^Pf4Δ/Pf4Δ and Dicer1^fl/fl mice (difference in gMFI P < .05; unpaired Student t test; n = 6 samples per group). (B) Bar graph representing the amount of endogenous (not injected) fibrinogen in platelet lysates obtained from Dicer1^Pf4Δ/Pf4Δ vs Dicer1^fl/fl mice, as measured by ELISA (*P < .05; unpaired Student t test; n = 7 mice per group).

Figure 6

Dicer1-deficient platelets are hyper reactive and prone to clot. (A-B) Flow cytometry analysis of fibrinogen binding to in vitro activated mouse platelets. Washed platelets from Dicer1^Pf4Δ/Pf4Δ or Dicer1^fl/fl mice were activated with increasing concentrations of a PAR4 agonist for 20 minutes in the presence of Alexa Fluor 488 labeled fibrinogen. (A) Mean ± SEM of the percentage of fibrinogen labeled (FIB488⁺) platelets or (B) intensity of fibrinogen 488 staining (gMFI). *P < .05; **P < .01; paired Student t test; n = 8 mice per group. (C-D) Flow cytometry analysis of JONA binding to platelets in whole blood in the presence of increasing concentrations of a PAR4 peptide. (C) Mean ± SEM of the percentage of JONA-labeled platelets or (D) the gMFI of JONA antibody staining (*P < .05; **P < .01; unpaired Student t test; n = 4 Dicer1^fl/fl and n = 3 Dicer1^Pf4Δ/Pf4Δ mice per group). (E-F) Analysis of tail-bleed time (E) and blood loss (F) after tail resection in Dicer1^Pf4Δ/Pf4Δ vs Dicer1^fl/fl mice (*P < .05; **P < .01; unpaired Student t test). (G) PE was induced in anesthetized mice of the indicated genotypes by IV injection of a solution of 60 mg/kg epinephrine and 150 mg/kg collagen. Time to death was monitored for 30 minutes. Shown is the Kaplan–Meier survival curve. *P < .05; log-rank test; n = 5 Pf4-Cre and n = 4 Dicer1^Pf4Δ/Pf4Δ mice per group.