Biorheology
Subcommittee Minutes
23 July 2011
14:00-18:00
Room E
Chairman: Michael R. King (US)
Co-chairmen: Lawrence Brass (US), Thomas Diacovo (US), Johan Heemskerk (NL), Shaun Jackson (AU), Armin Reininger (DE), J. Zwaginga (NL)
The session was well attended, with 180 attendees in Part I and 130 attendees in Part II.
14:00-14:10 Michael King (Cornell Univ., USA): Welcome and Opening Comments
King welcomed attendees and briefly summarized the activity of the Biorheology subcommittee over the previous year, including three JTH publications. He mentioned that a major activity over the next year will be to consider flow dependent thrombin and fibrin generation. He introduced the two-part session to follow, and reminded speakers to keep formal presentations under 15 minutes in length to allow for discussion.
Part 1: In vitro thrombosis assays and standardization (Moderators: T. Diacovo and M. King)
14:10-14:30 Judith Cosemans (Maastricht Univ., NL): Standardization of the use of flow devices to measure thrombus formation.
Dr. Cosemans showed movies of platelet aggregation on collagen under flow, and categorized the different systems used in the parallel-plate geometry. Advantages of parallel plate geometry were presented. A microgrid robot machine is used to create different spots of protein, separated by a distance of 1 mm, which is a sufficient distance to avoid “cross-over” effects between spots. Readout from these experiments is automated on computer, and a morphological score from 0 to 5 assigned. New disposable flow chambers were presented, fabricated in PDMS in the form of 4 parallel flow channels. A multicenter study using this device will start in September 2011, with the goal to look at variability between subjects and labs. Dr. Diacovo asked about the negative control experiments to demonstrate the specific molecular interactions. Dr. Turrito questioned whether there is an upstream vs. downstream dependence on different stripes. King asked whether the stripe order is randomized. Cosemans answered that reversing the order gives the same result. Dr. Brass asked about the potential clinical use of the device. Cosemans expanded on the application for diabetic patients vs. normal as an example.
14:30-14:50 Keith Neeves (Colorado School of Mines, USA): Clinical evaluation of microfluidic flow assays.
Dr. Neeves described a clinical study at University of Colorado Denver and Colorado School of Mines involving over 1000 subjects (500 healthy; Hemo A,B; other rare disorders). All consenting adults >18 yrs of age were included. 100x500um PDMS channels were used with type I collagen surface, following SSC recommendations. Raw data videos were shown alongside the binary analysis videos. Lag time, slope, and final surface coverage were measured as a function of time. Donor variability as well as day-to-day variability of the same donor were examined. Area fraction was not found to correlate with platelet count or hematocrit, but did correlate with bleeding score. Sufficient overlap between normal and diseased samples prompted nanoscale examination of different collagen thin films (fibrillar vs. non). Microcontact stamping of different collagen spots has been used. Coagulopathy has been examined with blood from hemophilia A, thrombus height correlated with FVIII level. Alisa Wolberg from UNC Chapel Hill asked about the lipid combinations and concentrations of tissue factor in experiments. Willem Ouwehand from Cambridge asked about whether enrollment numbers of the diseased population will be increased from 75. Neeves hopes to eventually get up to 1000. The audience member advised increasing to 500-1000. Dr. Heemskerk asked about fibrin production with or without CTI (corn trypsin inhibitor). Neeves uses CTI, they find that it yields more reproducible results. Turrito asked about the differences between normals and diseased, and whether this depends on collagen density. Neeves briefly discussed pros and cons of “increasing the gain” on these experiments.
14:50-15:10 Owen McCarty (Oregon Health & Science Univ., USA): Development of an in vitro model of occlusive thrombus formation.
Dr. McCarty briefly reviewed typical blood cell counts and coag factor concentrations in blood. He showed high resolution movies of platelet spreading, and showed thrombus formation movies in re-calcified blood over collagen fibers. McCarty showed his system that uses gravity driven flow at constant pressure gradient. Flow rate is monitored by eye or using optical methods. Occlusion is formed somewhere in the tube, as visualized by loss of flow. The dependence of time to occlusion was measured as a function of tissue factor concentration, with TF introduced on microspheres. McCarty also presented the initiation of thrombosis by breast epithelial cell lines, motivated by the study of metastatic cancer. He hopes to scale down the assay volume in the future, to enable use of mouse blood. Neeves asked about variation along the tube length, and where coagulation occurs. In future experiments some optical methods will be used to monitor clot location. King asked about whether blood is added during the experiment to maintain the fluid height (it is, added manually by the experimentalist).
15:10-15:30 Attie Tuinenburg (Univ. Med Center Utrecht, NL): Perfusion chamber with static mixer: New possibilities to study the interplay between platelets and coagulation under conditions of flow.
The general in vitro model for studying primary hemostasis was briefly reviewed, along with previous ex vivo experiments directly from the donor. Challenges in re-calcifying blood (improper mixing) were presented. The new microfluidic device with static mixer has no moving parts, fluids are mixed by flowing through the device passing a certain repeating geometry. The static mixer splits the flow stream in two, rotates the flow, and then combines the two streams. With every cycle (split, rotate, combine) the number of layers increases exponentially. Between the layers molecular diffusion occurs. The injection molding manufacture of the disposable device was presented. A coated coverslip is held in contact with the multi-layer device using vacuum. The two parts of the device are pressed together using an aluminum frame. Video results in the prototype chamber were shown (calcium (end concentration after mixing) at 8 mM; shear 1600 1/s) and robust thrombus formation over 15-20 minutes. At lower shear (500 1/s), more fibrin in a dense network was observed. Validation of the final model is ongoing, as is optimization of the experiments. A more accurate in vitro analysis will be possible, that includes coagulation with well mixed calcium. Turrito asked about the shear conditions upstream within the mixing elements (e.g., is there higher shear than in the main channel?). Dr. Tuinenburg will investigate this. Shear rate in the static mixer is lower than in the channel. The dimensions of the static mixer are 1mmx1mm, whereas the height of the parallel perfusion channel is 125 micrometer and the width is 1 mm.Brass asked about potential clinical questions to address with this model, and next steps. Tuinenburg answered that first, the will look at normal blood, and then move on to different disease states.
15:30-15:50 Johan Heemskerk (Maastricht Univ., NL) and Peter Lenting (Univ. Paris-Sud, France): Questionnaires and literature review on mouse models and genes in arterial thrombus formation in vivo and in vitro.
Dr. Lenting presented the questionnaire distribution and number of responses asking about in vivo thrombosis models used by various research labs. Different models: laser, mechanical injury, photochemical, FeCl3 are used, and most labs are satisfied with their own models. A large range in protocols (e.g., conc. Of ferric-chloride or Rose-Bengal) exists between labs. Laser wavelength also varies, as well as optical configuration. After reviewing the inventory, new recommendations were presented, related to the mouse, vessels, and measurement parameters, that are model independent. Then, standardization recommendations for different models were given: ferric-chloride, photochemical, and laser. Some parameters are difficult to standardize. An audience member from France asked a good question about which model is best for which scientific questions. Lenting agreed that this is a good question. Heemskerk then presented the standardization paper and questionnaire results for in vitro assays. 25 questionnaires were returned, 27 questions with more than one answer possible for each question. Strengths and weaknesses of the different custom made vs. micro-capillary vs. commercial chambers were presented. Most labs utilize human blood, but also mouse blood or cell lines. Different coatings and anticoagulants are used. Also, differences in blood storage, inhibitors, imaging and quantification are used. Finally, recommendations of the Biorheology SSC were presented, related to the flow chamber, surface coating, blood collection and storage, image recording. An inter-laboratory study comparing the various chambers is proposed.
15:50-16:00 Break
Part 2: In vivo thrombosis models (Moderators: K. Neeves and J. Heemskerk)
16:00-16:20 Shaun Jackson (Monash, Univ., AU): New approaches to investigate biomechanical platelet activation.
Dr. Jackson introduced the different zones of flow disturbance in arterial stenosis. Discoid platelet aggregation and membrane tethers as biomechanical sensors were the main topics of this talk. He showed evidence that changes in vessel geometry induce platelet aggregation, and that the flow shows shear micro-gradients, shear acceleration and deceleration. A microfluidic device with fixed stenosis was created to mimic the needle manipulation in mouse. Movies were shown that demonstrate that platelet aggregation begins right after the tip of the stenosis. There are three key variables regulating discoid platelet aggregation: rate of flow acceleration, peak shear, and rate of flow deceleration. Video evidence of membrane tethers forming in vivo was presented, to introduce the corresponding microfluidic experiments. TIRF (total internal reflection fluorescence) microscopy was used in vitro to image membrane tethers very close to the surface. Calcium spikes in spreading platelets has been visualized. Tethers restructure and contract to sustain the adhesion of discoid platelets under flow, by getting thicker via actin polymerization. Brass asked whether tether retraction is really passive. Jackson answered that cell activation is local, with no global P-selectin expression observed. Turrito asked about the shear gradients, and whether the magnitude of shear attained is the cause instead of the gradient. He proposed varying gradient and magnitude separately. Jackson explained that they have already done this, by varying the angle of the microfluidic stenosis, and also answered how they have proposed that both gradient and peak shear magnitude are important.
16:20-16:40 Lawrence Brass (Univ. of Pennsylvania, USA): Using intravital confocal fluorescence microscopy to probe the microheterogeneity and permeability of the platelet response to laser and mechanical injury.
Brass started by stating that microheterogeneity (variations in time and space of activation state of indiv. Platelets in the thrombus) and porosity (the ability of soluble molecules to pass in and out of the thrombus) are two important terms for his talk. He showed in vivo movies with platelets in red and P-selectin (activated platelets) in green. Red shows up minutes before green, but even after 1 hour there are still P-selectin (negative) platelets. Brass briefly presented on laser vs. mechanical injury (previous data was using the laser injury model), and qualitatively they behave the same. Fluorescent albumin shows that plasma leaks out before and after the clot has formed, and heterogeneity in the permeability of the clot. Fluorescent dextran (10 kD) has also been used. Brass reviewed previous electron microscopy data showing close contacts between platelets. The last results related to an inner core of the thrombus in which the porosity has been reduced. Heemskerk asked a question on secretion of factors by platelets. Brass noted that there is a need for additional in vivo markers of platelet activation. Jackson asked about platelet activation and porosity. There was a question on where does the P-selectin come from (i.e., maybe not from granules), but soluble P-selectin binding to platelets is unlikely because it does not show up on unactivated cells.
16:40-17:00 Tom Diacovo (Columbia Univ., USA) and Michael King (Cornell Univ., USA): The utility of a humanized mouse system for multiscale analysis of thrombus growth, platelet convection, and drug efficacy.
Diacovo began the talk by presenting the motivation for developing the humanized mouse model, and other (ex vivo) means to measure human platelet function. He briefly described the new mouse model that can utilize human platelets that form clots, whereas due to a transgenic mutation mouse platelets do not participate: thus the first in vivo studies using human platelets. The published work was cited for further details. Diacovo showed clinical data that demonstrates that the humanized mouse model is sensitive enough to differentiate diseased cells from normal cell function in thrombus growth. King gave the second half of the talk, focusing on detailed analysis of the fluid mechanics in Diacovo’s in vivo system, and computational models of platelet adhesion. Models that cover the scale of whole vessel, individual cells, and adhesion receptors (molecular scale) were presented, including the “platelet adhesive dynamics” model which recreates the stochastic behavior of platelet rolling. Some preliminary sensitivity results were shown, indicating how the model predictions of platelet rolling depend on shear rate, or an increase or decrease in the receptor number. Neeves asked a question on whether platelet deformation would be important in the model. King answered that the capabilities are there, as well as the effect of tether deformation as has been done for leukocytes. Another question was asked about modeling the effects of drugs and inhibitors, both King and Diacovo answered that it is possible and planned in both the computer model and mouse model.
17:00-17:20 Atsushi Yamashita (Univ. of Miyazaki, Japan): Disturbed blood flow induces plaque erosion and thrombus formation.
Dr. Yamashita gave background on plaque rupture vs. plaque erosion, and shared histological examples of both. Balloon injury of rabbit femoral artery was used to induce SMC-rich neointima within 3 weeks, with higher TF activity. Disturbed blood flow was introduced by occluding the vessel via external constriction. Neointima SMCs align parallel with blood flow. Erosive injury occurs even within 15 minutes after narrowing. More TUNEL-positive apoptotic cells were found in the neointima layer only after narrowing. Fibrin deposition occurred in the SMC-rich neointima (after narrowing), but not in the normal intima vessels. Thrombotic occlusion occurred in the injured+narrowed vessels within 3 hours. The blood itself showed no differences in in vitro assays.
17:20-17:40 William Olbricht (Cornell Univ., USA): In vivo imaging of cerebral circulation in mouse models of polycythemia vera and essential thrombocythemia.
Dr. Olbricht gave some brief background on the two clinical conditions with elevated hematocrit and platelet count. The clinical treatment is to reduce hematocrit via phlebotomy. The focus of this talk was on microvascular occlusions in brain. The classic viscosity dependence on hematocrit measured in rheometers is not necessarily applicable in the complex capillary bed in the brain. In vivo two-photon microscopy in mouse cortex was used to measure differences in blood flow under these conditions. RBC speed can be measured in capillaries by fluorescently labeling the plasma (no signal within the RBC due to exclusion). “Stalled” capillaries are observed using this technique. Transgenic mouse models are available that recreate the human PV and ET diseases, or EPO can be injected in normal mice to elevate hematocrit without increasing leukocytes or platelets. The different mice show different fractions of capillaries stalled, over hundreds of vessels analyzed. In the PV mice, stalls are often due to microthrombi or leukocyte plugs, not so in control mice. RBC speed in nonstalled capillaries is different in experimental mice. A question was asked about reducing hematocrit in the EV mice, but it has not been done yet. Turrito asked about the implications of the PV mouse for clinical observations.
