Occlusive thrombi arise in mammals but not birds in response to arterial injury: evolutionary insight into human cardiovascular disease

Abstract

Mammalian platelets are small, anuclear circulating cells that form tightly adherent, shear-resistant thrombi to prevent blood loss after vessel injury. Platelet thrombi that form in coronary and carotid arteries also underlie common vascular diseases such as myocardial infarction and stroke and are the target of drugs used to treat these diseases. Birds have high-pressure cardiovascular systems like mammals but generate nucleated thrombocytes rather than platelets. Here, we show that avian thrombocytes respond to many of the same activating stimuli as mammalian platelets but are unable to form shear-resistant aggregates ex vivo. Avian thrombocytes are larger than mammalian platelets, spread less efficiently on collagen, and express much lower levels of the α(₂b)β₃ integrin required for aggregate formation, features predicted to make thrombocyte aggregates less resistant than platelets are to the high fluid shear forces of the arterial vasculature. In vivo carotid vessel injury stimulates the formation of occlusive platelet thrombi in mice but not in the size- and flow-matched carotid artery of the Australian budgerigar. These studies indicate that unique physical and molecular features of mammalian platelets enable them to form shear-resistant arterial thrombi, an essential element in the pathogenesis of human cardiovascular diseases.

Figure 1

Avian thrombocytes express genes associated with platelet function and signal in response to platelet agonists collagen and thrombin. (A) Real-time quantitative PCR in thrombocytes and lymphocytes of select genes with a known role in platelet function. Shown is the log₂ difference in gene expression between chicken thrombocytes and lymphocytes. (B-E) 5-HT release from platelets or thrombocytes was measured after loading with [³H] 5-HT. Cells were stimulated with increasing doses of bovine thrombin (B) or equine type-I collagen (C). (D) Chicken thrombocytes were stimulated with collagen (3 μg/mL) in the presence of increasing doses of PP2. (E) PP2 (3μM) was used to inhibit 1.0 U/mL thrombin- or 3.0 μg/mL collagen-induced 5-HT release in either human platelets or chicken thrombocytes. An equivalent volume of DMSO solvent was added in no inhibitor controls. Mean ± SD is shown for all panels; n = 3 experiments per condition.

Figure 2

Avian thrombocytes adhere to collagen but fail to form 3-dimensional aggregates under flow. PPACK-anticoagulated human or chicken whole blood was perfused through a tapered-wall parallel plate flow chamber for 5 minutes over a collagen-coated glass slide. (A) Representative fluorescent images captured at areas corresponding to shear rates of 1300, 1100, 700, and 400 seconds⁻¹ after 5 minutes perfusion of whole blood labeled with AP-2 mAb. (B) Representative images of thrombocyte or platelet adhesion to collagen after whole blood flow for 5 or 15 minutes at 1300 seconds⁻¹. Slides were fixed, permeabilized, and stained with Alex-Fluor 594–conjugated phalloidin to detect actin filaments. Nuclear staining of thrombocytes with DAPI is also shown. (C) Time traces of percent surface coverage of collagen surface by platelets or thrombocytes. Shown is mean ± SD (gray zone); n = 7-10 experiments for each condition. Percent collagen surface area coverage (D) and mean aggregate area (E) of human platelets and chicken thrombocytes after perfusion of whole blood for 5 minutes was determined by analysis of fluorescent images. Shown are the mean ± SEM, n = 7-10 experiments for each condition. (F) Scanning electron microscopy is shown after perfusion of human and chicken blood over collagen at a shear rate of approximately 1300 seconds⁻¹. Scale bars indicate 50 μm (500×), 20 μm (1000×), and 5 μm (5000×). Shown are representative images from 3 and 5 human and chicken flow experiments, respectively.

Figure 3

Avian thrombocytes spread less efficiently than platelets and express low levels of α_2bβ₃ integrins. Platelet and thrombocyte spreading on collagen was measured over 10 minutes with RICM. (A) The fold increase in surface area covered by adherent platelets and thrombocytes is shown. (B) The absolute area covered by adherent platelets and thrombocytes is shown. Mean ± SEM is shown; n = 25 for each group. (C) Levels of α_2bβ₃ integrin expression on washed human platelets and washed chicken thrombocytes were determined by staining with FITC-conjugated AP-2 mAb staining and analyzed by flow cytometry. The forward and side scatter of values of cells in platelet-rich plasma (PRP) and thrombocyte-rich plasma (TRP) are shown, and gates denote platelets and thrombocytes, respectively. Staining of human platelets and chicken thrombocytes for the α_2bβ₃ integrin with the use of AP-2 mAb binding is shown. (D) The histogram depicts the mean fluorescence intensity (MFI) of AP-2 binding for platelets and thrombocytes. Error bars indicate SD; n = 3 in each group. (E) Scanning electron microscopy images are shown after perfusion of human and chicken blood over collagen at a shear rate of approximately 1300 seconds⁻¹. The α_2bβ₃ integrin antagonist eptifibatide was added at 10μM. Scale bars indicate 10 μm. Shown are representative images from 3 and 5 human and chicken flow experiments, respectively.

Figure 4

Arterial vessel wall injury results in thrombotic occlusion in mice but not in birds. The carotid arteries of anesthetized mice and Australian budgerigars were exposed to filter pads soaked in the indicated concentrations of FeCl₃ for 5-20 minutes. (A) Time to occlusion after removal of the FeCl₃ pad was determined by measuring arterial blood flow using a Doppler flow probe. N/O indicates no occlusion, defined as normal blood flow after 20 minutes afterFeCl₃ pad removal, the termination point of the experiment. Zero minutes indicates occlusion before the time of Doppler probe measurement. Diamonds and circles represent individual mouse and budgerigar carotids, respectively. Horizontal bars indicate the mean time to occlusion for mouse carotid vessels at a given concentration and time of FeCl₃ administration. Budgerigar carotid occlusion was observed only after repeated administration of the highest concentration of FeCl₃ (40% FeCl₃ for 10 minutes × 2). (B) Histologic examination of thrombus formation in FeCl₃-injured mouse and budgerigar carotid arteries. Transverse sections of injured carotid arteries were stained with H&E or Prussian blue (to detect FeCl₃). The occlusive mass of eosin-staining anuclear cells that occlude the mouse arterial lumen is composed of platelets (PLTs). Budgerigar thrombocytes (TCs) are elongated nucleated cells that do not stain strongly for Prussian blue. Occlusive thrombi failed to form in injured budgerigar carotids in the absence of FeCl₃ penetration into the vessel lumen. Scale bar = 50 μm. n = 1-5 experiments for each condition, as indicated by the number of diamonds and circles in panel A.