Mitofusin-2 Regulates Platelet Mitochondria and Function

Abstract

Background: Single-nucleotide polymorphisms linked with the rs1474868 T allele (MFN2 [mitofusin-2] T/T) in the human mitochondrial fusion protein MFN2 gene are associated with reduced platelet MFN2 RNA expression and platelet counts. This study investigates the impact of MFN2 on megakaryocyte and platelet biology.

Methods: Mice with megakaryocyte/platelet deletion of Mfn2 (Mfn2^-/- [Mfn2 conditional knockout]) were generated using Pf4-Cre crossed with floxed Mfn2 mice. Human megakaryocytes were generated from cord blood and platelets isolated from healthy subjects genotyped for rs1474868. Ex vivo approaches assessed mitochondrial morphology, function, and platelet activation responses. In vivo measurements included endogenous/transfused platelet life span, tail bleed time, transient middle cerebral artery occlusion, and pulmonary vascular permeability/hemorrhage following lipopolysaccharide-induced acute lung injury.

Results: Mitochondria was more fragmented in megakaryocytes derived from Mfn2^-/- mice and from human cord blood with MFN2 T/T genotype compared with control megakaryocytes. Human resting platelets of MFN2 T/T genotype had reduced MFN2 protein, diminished mitochondrial membrane potential, and an increased rate of phosphatidylserine exposure during ex vivo culture. Platelet counts and platelet life span were reduced in Mfn2^-/- mice accompanied by an increased rate of phosphatidylserine exposure in resting platelets, especially aged platelets, during ex vivo culture. Mfn2^-/- also decreased platelet mitochondrial membrane potential (basal) and activated mitochondrial oxygen consumption rate, reactive oxygen species generation, calcium flux, platelet-neutrophil aggregate formation, and phosphatidylserine exposure following dual agonist activation. Ultimately, Mfn2^-/- mice showed prolonged tail bleed times, decreased ischemic stroke infarct size after cerebral ischemia-reperfusion, and exacerbated pulmonary inflammatory hemorrhage following lipopolysaccharide-induced acute lung injury. Analysis of MFN2 SNPs in the iSPAAR study (Identification of SNPs Predisposing to Altered ALI Risk) identified a significant association between MFN2 and 28-day mortality in patients with acute respiratory distress syndrome.

Conclusions: Mfn2 preserves mitochondrial phenotypes in megakaryocytes and platelets and influences platelet life span, function, and outcomes of stroke and lung injury.

Keywords: blood platelets; ischemic stroke; megakaryocytes; mitochondria; neutrophils.

Figure 1:. Mitochondrial morphology in platelets and MK from Pf4-CRE conditional Mfn2 knockout mice

A. Transgenic Pf4 promoter-driven Cre recombinase mice were crossed with mice harboring homozygous floxed Mfn2 to generate platelet/MK specific Mfn2 knockout mice (called Mfn2^−/−). B. Representative western blot for Mfn2 in platelets from Mfn2^+/+ and Mfn2^−/− mice. C. Representative live immunofluorescent microscopy images of mitochondrial morphology in MKs derived from the bone marrow of Mfn2^+/+ and Mfn2^−/− mice isolated and cultured in pairs. MKs are marked with surface GPIbα staining (blue); mitochondria (mitotracker, green) and nuclei (Hoechts, magenta) are stained with live cell penetrating dyes. Bottom panels are magnified portions of the top panels outlined by the yellow boxes. D. Overall mitochondria morphology in individual MKs was scored for fusion on a scale of 1-10 by an observer blinded to genotype as described in detail in the methods. Briefly, higher scores are given for MKs with predominantly elongated and branched mitochondria and low scores for predominantly punctate, single mitochondria. On average 15 MKs were scored per sample and >250 MKs were scored in total (Wilcoxon paired sign rank test, N=9 culture pairs per group). E. Representative confocal immunofluorescent microscopy images of mitochondrial morphology in platelets from Mfn2^+/+ and Mfn2^−/− mice stained with mitotracker (green) and phalloidin (cytoskeleton, red). F. Representative TEM sections of platelets from Mfn2^+/+ and Mfn2^−/− mice. Arrows point to mitochondria. Most mitochondria were round or oval in shape, but occasional mitochondria were elongated (see Supplement Figure 3 for additional images and quantitation).

Figure 2:. Mfn2^−/− impairs mitochondrial function.

A. Mitochondrial membrane potential in platelets (gated on FSC/SSC and CD41+) from Mfn2^+/+ and Mfn2^−/− mice as measured by flow cytometry analysis of TMRM staining. Measurements were taken before and after treatment with CCCP which depolarizes mitochondria. See Supplement Figure 4A for representative flow cytometry histograms. t-test, N=22 per group (10 males, 12 females). B. Mitochondrial load measured by Mitotracker in Mfn2^+/+ and Mfn2^−/− platelets (N=9 Mfn2^+/+ and N=10 Mfn2^−/−) C. Representative experiment (of 5 independent experiments with a total of n=14 independent mice per group, see Supplement 4C for pooled, normalized data from all experiments) for Oxygen Consumption Rate (OCR) in untreated Mfn2^+/+ and Mfn2^−/− platelets as measured by Seahorse at baseline, and following treatment with Oligomycin (ATP synthase inhibitor to identify cellular ATP production); FCCP (uncoupling agent that disrupts membrane potential to determine maximal respiration); Rotenone and Antimycin A (Complex I and III inhibitors to define mitochondrial specific OCR). D. summary of seahorse measurements in C. See Supplement Figure 2B for how these measurements are derived. t tests, N=3 per group. E. OCR in Mfn2^+/+ and Mfn2^−/− platelets treated with 1 U/mL thrombin and treatments with mitochondrial modulators (described for B) at the time points indicated in C. F. Summary of Seahorse measurements in E. t tests, N=3 per group. G. Representative western blot of an index mitochondrial complex I nuclear encoded subunit Ndufb8 (labile when complex I is not assembled) compared to beta tubulin (cytoskeletal) and Vdac (nuclear encoded mitochondrial channel protein). H. Densitometry analysis of complex I/Vdac protein levels in platelets Mfn2^+/+ and Mfn2^−/− mice. Mann-Whitney test, N=5 per group. I. Mitochondrial complex I activity measured in lysates from male Mfn2^+/+ and Mfn2^−/− platelets. Mann-Whitney test, N=4 per group. J. Stored ATP content in Mfn2^+/+ and Mfn2^−/− platelets before and after stimulation with thrombin or thrombin plus apyrase (to separate intracellular from stored/released ATP).

Figure 3:. Conditional platelet/MK deletion of Mfn2 reduces platelet counts and platelet lifespan through accelerated apoptosis of older platelets.

A. Platelet counts in 2-month-old male and female Mfn2^+/+ and Mfn2^−/− littermate mice (t-test, N=10-12 per group). B. Thiazole orange staining of reticulated platelets as measured by flow cytometry in 5-month-old Mfn2^+/+ and Mfn2^−/− littermate mice (t test, N=16-17 per group). C. Mice were injected with Dylight 488-GPIbα and the fraction of Dylight labeled platelets of CD41+ platelets was assessed by flow cytometry in whole blood obtained by repeated tail pricks in the same mouse at the indicated time points starting 4 h after injection (set as T0). D. Calculated platelet half-life from C (C-D: Mann-Whitney tests, N=6-7 per group). E. Schematic summary and representative flow plots of platelet transfusion mixing experiment. Platelets in Mfn2^−/− and Mfn2^+/+ mice were differentially prelabeled by intravenous injection of Dylight 488 or Dylight 649 anti-platelet fab fragment. Two hours later blood was drawn, and Mfn2^−/− and Mfn2^+/+ PRP were mixed 1:1 and injected intravenously into recipient mice. Blood was drawn at 2 h for baseline (T0), then 24, 48, and 96 h later and the percent of Mfn2^−/− vs Mfn2^+/+ platelets determined by flow cytometry. Shown are representative flow plots showing the percent of Mfn2^−/− and Mfn2^+/+ platelets at T0 and T96. F. Summary of the fraction of Mfn2^−/− (red) and Mfn2^+/+ (blue) remaining at different time points post injection. Shown is a representative of 3 independent experiments. Mfn2^−/− platelets disappeared faster than Mfn2^+/+ platelets after transfusion regardless of whether the labels were swapped (not shown) or recipients were Mfn2^+/+ or Mfn2^−/− mice: Experiment 1: 4 Mfn2^+/+ and 4 Mfn2^−/− (2 each male/female) mixed and transfused into 4 Mfn2^+/+ and 4 Mfn2^−/−, p=0.08; Experiment 2: 7 Mfn2^+/+ males and 7 Mfn2^−/− males mixed and transfused into 3 Mfn2^+/+ and 4 Mfn2^−/− mice, p=0.02; Experiment 3: 7 Mfn2^+/+ and 7 Mfn2^−/− females mixed and transfused into 7 recipients, p=0.003. G. Mfn2^+/+ and Mfn2^−/− platelets differentially labeled and mixed as in C were stained with annexin V and analyzed by flow cytometry immediately (T0) or following ex vivo co-culture for 16 h at 37° C. The dotted lines indicate the respective Mfn2^+/+ and Mfn2^−/− pairs that were mixed in the same tube (Wilcoxon paired sign rank test, N=7). H. Mitochondrial potential (TMRM) in ex vivo cultured platelets. I-J. Western blot and densitometry analysis for cleaved caspase 3 and procaspase 3 in Mfn2^+/+ and Mfn2^−/− platelets after 14h ex vivo at 37° C (Mann-Whitney test, N=7 per group) K. Dylight was injected into mice 72 h prior to isolation to distinguish old from new platelets. Annexin V staining assessed by flow cytometry at 0 (T0), 8 h (T8), and 14 h (T14) on new (Dylight negative) and old (Dylight positive) platelets (Mann-Whitney tests for AUC, N=4-6 per group).

Figure 4:. Mfn2^−/− alters platelet activation, hemostasis, and thrombosis.

A-B. Flow cytometry analysis of platelet activation assessed by surface binding of A. Jon/A (activated αIIb/β3) or B. anti-p-selectin antibody (degranulation) at baseline or after 10 min incubation with indicated doses of PAR4 (thrombin receptor) or CRP (GPVI receptor) agonists (t test, N=6 per group). C. Aggregation of washed platelets, normalize to counts, in response to collagen (Mann-Whitney test, N=7 per group). D. Platelet-neutrophil aggregates in Mfn2^+/+ and Mfn2^−/− whole blood at baseline and in response to convulxin (Mann-Whitney test, N=3-4 per group). E. Flow cytometry analysis of PS exposure (annexin V binding) on the surface of platelets at baseline and after activation with dual agonist thrombin (0.1 U/mL) + convulxin (500 ng/mL) (Mann-Whitney test, N=6 per group). F. ROS generation in Mfn2^+/+ and Mfn2^−/− platelets in response to thrombin (0.1 U/mL) + convulxin (500 ng/mL) (Mann-Whitney test, N=6 per group). G. Ca²⁺cyt transients in Mfn2^+/+ and Mfn2^−/− platelets after stimulation with thrombin (0.1 U/mL) + convulxin (500 ng/mL). N=9 per group. H. Time to cessation of bleeding after tail resection in Mfn2^+/+ and Mfn2^−/− mice. Each point represents an individual mouse. (Mann Whitney test, N=20-21 per group). I. Representative images of brains from mice after ischemia/reperfusion injury in Mfn2^+/+ (left) and Mfn2^−/− (right) mice. Infarcts are outlined with dotted lines. J. Summary of infarct volume. t-test, N=13-15 mice per group.

Figure 5:. Mfn2^−/− increases lung vascular permeability and aggravates inflammatory bleeding in LPS-induced acute lung injury model.

A. Protein concentration in the BALF of Mfn2^+/+ and Mfn2^−/− mice 24 h post-intratracheal saline or LPS-challenge (1.6 mg/Kg body weight). Kruskal Wallis test with Dunn’s multiple (3) comparisons, n=8-10 mice per group. B. Representative images of lungs from thrombocytopenic Mfn2^+/+ and Mfn2^−/− mice 24 h post-intratracheal LPS-challenge (1.6 mg/Kg body weight). C. Bleeding score of lungs from thrombocytopenic Mfn2^+/+ and Mfn2^−/− mice 24 h post-intratracheal LPS-challenge. The lungs were scored for bleeding by two observers blinded to genotype and averaged. Higher scores were given to lungs with increased bleeding. Mann Whitney test, n=5-7 mice per group. D. Representative images (N=11) of BALF from thrombocytopenic Mfn2^+/+ and Mfn2^−/− mice 24 h post-intratracheal LPS-challenge (1.6 mg/Kg body weight). E. Protein concentration in the BALF of Mfn2^+/+ and Mfn2^−/− thrombocytopenic mice 24 h post-intratracheal LPS-challenge (1.6 mg/Kg body weight). Mann-Whitney test, n=11-12 per group. F. BALF hemoglobin as a measure of RBC infiltration in the lungs of thrombocytopenic Mfn2^+/+ and Mfn2^−/−. Mann-Whitney test, n=11-12 per group.

Figure 6:. pMFN2 eQTL affects MFN2 expression and mitochondrial morphology in human cord blood (CB) megakaryocytes (MKs).

A. pMFN2 mRNA normalized to GAPDH mRNA in day13 CB MKs from rs1474868 C/C, C/T and T/T individuals as measured by real time PCR (nonparametric bootstrap test, 1000 resamples), N=5 C/C or C/T and N=4 T/T). B. Representative live immunofluorescent microscopy images of mitochondria (mitotracker, green) and nuclei (Hoechts, blue) in day 13 CB MKs of the indicated rs1474868 genotype, showing more elongated mitochondria in C/C MKs and more punctate mitochondria in T/T MKs. Bottom panels are magnified portions of the top panels outlined by the white boxes. C. Mitochondrial fusion scores for CB MKs with the indicated rs1474868 genotype were scored for fusion on a scale of 1-10 by an observer blinded to genotype as described in detail in the methods. Representative images of mitochondria from MKs with different scores are shown on the left. Individual MK scores are shown in the violin and average scores per cord are shown in D (Mann-Whitney test, N=4 C/C and N=10 T/T).

Figure 7:. pMFN2 eQTL affects MFN2 protein and ex vivo survival of human platelets.

A. Real time PCR analysis of pMFN2 expression in platelets from individuals with rs1474868 C/C, C/T or T/T genotypes normalized to GAPDH mRNA (Mann Whitney test, C/C or C/T N=13; T/T N=8). B. Top: representative western blot of MFN2 protein in platelets from individuals with the indicated rs1474868 genotypes. Beta Tubulin is used as loading control. Bottom: densitometric analysis of all samples (t test, C/C or C/T N=13; T/T N=15). C. Mitochondrial potential (ΔΨ_M) measured as TMRM MFI in washed platelets (gated on FSC/SSC and CD41) from rs1474868 C/C or C/T versus T/T individuals (t test, C/C or C/T N=13; T/T N=8). D. Loss of mitochondrial potential (ΔΨ_M) measured as % TMRM negative (-) platelets from rs1474868 C/C or C/T versus T/T individuals (t test, C/C or C/T N=13; T/T N=8). E. Resting PS exposure measured as % annexin V on washed platelets from rs1474868 C/C or C/T versus T/T individuals at time 0 (T0) and after 24 h (T24) incubation at 37° (t test, N=18 C/C or C/T and N=13 T/T). Results in panels C-E are from a cohort of donors recruited in 2018. See Supplement Figure 9 for TMRM, PS exposure, and Caspase-3 activity over 48 hours in an independent cohort recruited in 2023. Meta-analysis of combined p-values for both cohorts (P_s, Stouffer’s method, 2 tailed): median TMRM P_s=0.040 (P₁=0.093, P₂=0.125); % TMRM- P_s=0.023 (P₁=0.018, P₂=0.31); Annexin at T24 P_s=0.048 (P1=0.038, P2=0.35).