Adenovirus-induced thrombocytopenia: the role of von Willebrand factor and P-selectin in mediating accelerated platelet clearance

Maha Othman, Andrea Labelle, Ian Mazzetti, Hisham S. Elbatarny and David Lillicrap

Article Figures & Data


  • Figure 1

    Thrombocytopenia in adenovirus and LPS-treated mice. Platelet counts were assessed following tail vein injection of 1 × 1011 adenoviral particles into each of Balb/C and C57/ BL6 mice. The platelet count falls significantly at 24 hours (Balb/c: n = 8, P < .001; C57BL/6: n = 6, P = .002) and returns to normal at 1 week. In Balb/c mice receiving 0.4 ng LPS, the platelet count falls significantly as early as 1 hour (n = 3, P = .04) and remains low at 5 and 24 hours (n = 3, P = .02). Values are the mean ± SEM.

  • Figure 2

    Expression of platelet P-selectin (CD62p) in vitro and following intravenous administration of adenovirus to mice. (A) Murine washed platelets were incubated with adenovirus for 30 minutes at 37°C and analyzed for CD62p by flow cytometry. The index of platelet activation IPA+ (MCF × percentage of CD62+ events) is significantly increased in adenovirus (Ad)–treated platelets compared to resting platelets (average of 5 experiment). Thrombin-activated platelets are tested as a positive control. (B) Flow cytometry histogram representative of 5 in vitro experiments showing increased fluorescence in adenovirus-treated platelets and thrombin-activated platelets compared to resting platelets and the isotype control. (C) Platelet P-selectin expression during the first 24 hours following adenovirus or LPS intravenous administration in mice. Graph represents fold increase of percent of P-selectin positive platelets after adenovirus compared to preinjection levels. Significant P-selectin expression is significant in both LPS and adenovirus treated at 1 hour compared to the preinjection level (adenovirus-treated mice n = 4, P < .001). Values are the mean ± SEM.

  • Figure 3

    Adenovirus induces platelet-leukocyte aggregate formation and both platelet and leukocyte MPs following intravenous administration to mice. (A) Whole blood samples obtained from Balb/c mice during the first 24 hours following injection of adenovirus. Flow cytometry assessment of platelet-leukocyte aggregate was performed as follows: platelets were first gated based on the CD41 marker and the percentage of CD45+ cells (leukocytes) associated with platelets was assessed on another histogram. The upper right quadrants of the histograms show an evolving significant increase in platelet-leukocyte aggregates from before to various time points following adenovirus treatment. These data are representative of 3 independent in vivo experiments. (B) Graphs showing the average number of MPs in plasma obtained from Balb/c mice based on separate assessment of CD41 and CD61 (PMPs) and CD45 (LMPs). There is a significant increase in PMPs (n = 3, P = .02 based on CD41; P < .001 based on CD61 marker) and LMPs (n = 3, P = .03) at 2 hours following virus administration compared to the preinjection levels (Before). Values shown are the mean ± SEM.

  • Figure 4

    Adenovirus activates endothelial cells in vitro and induces release of ultra-large molecular weight VWF as well as EMPs following intravenous administration in mice. (A) BOECs were treated with 3 doses of adenovirus (V1-3) for 6 hours. Cells were then harvested, washed, and stained with fluorescent-labeled anti–human VCAM-1 and assessed by flow cytometry. BOECs treated with LPS were used as a positive control and LPS data were set at 100% (average of 3 experiments). (B) BOEC culture media was collected at 24 hours following adenovirus treatment for VWF quantitation. There was a non–dose-dependent increase in VWF levels when compared to the untreated cells. (C) Plasma samples from Balb/C mice were collected 1 to 2 hours following intravenous administration of adenovirus. VWF levels increased significantly (11.9-fold above preinjection level; upper panel) and ultra-large molecular weight VWF multimers appear on multimer analysis of murine Balb/c plasma (lower panel). The figure shows an increased multimer density (increased VWF) as well as ultra-large multimers in the 7 adenovirus-treated mice (lanes1-7) compared to control mouse plasma (C lanes, represented by the two arrows). The dotted line shows highest molecular weight multimer in normal mice. (D) The number of EMPs in plasma obtained from Balb/c mice after intravenous virus administration based on CD62E expression increases significantly following administration (n = 3, P = .03). Values shown are the mean ± SEM.

  • Figure 5

    VWF KO mice do not show significant thrombocytopenia, experience a level of P-selectin expression but do not form platelet leukocyte aggregates following adenovirus administration. (A) Platelet counts were assessed at 1 hour, 5 hours, 24 hours, and 1 week following intravenous administration in WT and VWF−/−C57BL/6 mice showing nonsignificant thrombocytopenia in VWF−/− following adenovirus (n = 6, P = .002 in WT versus P = .06 in VWF KO mice). (B) Similar to VWF+/+ mice, platelets from VWF−/− mice show significant P-selectin expression at 1 hour after adenovirus injection (n = 3, P < .001) compared to preinjection expression but significantly lower when compared to WT (P = .01). (C) Graph showing nonsignificant platelet leukocyte aggregates at 1 hour following virus administration to VWF KO mice compared to the preinjection level (n = 3, P = .12 in KO versus P = .007 in WT mice)

  • Figure 6

    Platelets bind adenovirus and express CAR on their surface. The 3H-labeled platelets were allowed to adhere on fibronectin-coated wells with or without the virus and the extent of adhesion was determined after lysis and measurement of radioactivity. (A) Graph showing reduced platelet adhesion in the presence of the virus compared to basal platelet adhesion (n = 4, P = .04) and significant interference with ADP potentiation of platelet adhesion (n = 4, P = .008). Error bars represent ± SEM. (B) Platelets and CAR-expressing HEK 293 cells were stained with mouse monoclonal anti–human CAR antibody and FITC rabbit anti–mouse IgG and analyzed with flow cytometry. The flow cytometry histogram shows that platelets are positive for CAR (graph representative of 4 experiments). (C) The presence of CAR on platelets was verified by human platelet mRNA analysis; 1.5% agarose gel showing RT-PCR of platelet-derived RNA amplifying 366-bp band from the CAR gene. Lane 1 is a 100-bp molecular weight ladder.