Real-time monitoring of adhesion and aggregation of platelets using thickness shear mode (TSM) sensor

Available online at www.sciencedirect.com

Biosensors and Bioelectronics 23 (2007) 575–582

Real-time monitoring of adhesion and aggregation of platelets using thickness shear mode (TSM) sensor E. Ergezen a , M. Appel b , P. Shah a , J.Y. Kresh c , R.M. Lec a , D.M. Wootton b,d,∗ a

School of Biomedical Engineering, Science and Health Systems, Drexel University, United States b Mechanical Engineering & Mechanics, Drexel University, United States c Department of Cardiothoracic Surgery, Drexel University College of Medicine, United States d Albert Nerken School of Engineering, Cooper Union for the Advancement of Science and Art, United States Received 29 January 2007; received in revised form 23 May 2007; accepted 25 May 2007 Available online 31 July 2007

Abstract Hemostasis is required to maintain vascular system integrity, but thrombosis, formation of a clot in a blood vessel, is one of the largest causes of morbidity and mortality in the industrialized world. Novel clinical and research tools for characterizing the hemostatic system are of continued interest, and the object of this research is to test the hypothesis that clinically relevant platelet function can be monitored using an electromechanical sensor. A piezoelectric thickness shear mode (TSM) biosensor coated with collagen-I fibers to promote platelet activation and adhesion was developed and tested for sensitivity to detect these primary events. Magnitude and frequency response of the sensor were monitored under static conditions at 37 ◦ C, using platelet-rich plasma (PRP), and PRP with adenosine diphosphate (ADP), a clinical aggregation inhibitor (abciximab), or a collagen binding inhibitor. Sensors loaded with PRP exhibited a 3-stage response; no significant change in response for the first 20 min (Stage-1), followed by a larger drop in response (Stage-2) and subsequently, response gradually increased (Stage-3). Exogenous ADP stimulated an immediate Stage-2 response, while abciximab delayed and reduced the magnitude change of Stage-2. In the presence of collagen inhibitor, Stage-2 response was similar to that of control but was delayed by an additional 20 min. The obtained results, supported by epifluorescence and complementary SEM studies, demonstrated the selective sensitivity of TSM electromechanical biosensors to monitor platelet function and inhibition, particularly aggregation. © 2007 Elsevier B.V. All rights reserved. Keywords: Thickness shear mode; Blood platelet; Multi-frequency; Adenosine diphosphate; GPVI

1. Introduction Sensors to characterize hemostasis and thrombosis are of continued importance because of the growing use of anti-thrombotic agents to prevent heart attack, stroke and other clinical problems. Both hemostasis and thrombosis share three phases. They involve the formation of a loose temporary platelet aggregate at the site of injury. Platelets bind to collagen at the site of vessel wall injury, via integrin ␣2 ␤1 and glycoprotein VI (GPVI), and are activated by collagen binding, or by thrombin, adenosine 5 -diphosphate (ADP) and other soluble agonists released from neighboring activated platelets. Upon activation, platelets undergo a conformation change from a discoid to a rounded ∗ Corresponding author at: Albert Nerken School of Engineering, Cooper Union for the Advancement of Science and Art, 51 Astor Place, New York, NY 10003, United States. Tel.: +1 212 353 4393. E-mail address: [email protected] (D.M. Wootton).

0956-5663/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2007.05.009

shape and aggregate principally via activated platelet integrin ␣IIb ␤3 to form a hemostatic plug or thrombus. Platelet activity may be inhibited due to a variety of inherited defects such as von Willebrand disease that may need prompt diagnosis. Platelet activity is also routinely inhibited in antithrombotic therapy to reduce the risk of thrombosis during cardiovascular procedures and surgery, typically with a platelet ␣IIb ␤3 inhibitor such as abciximab (ReoPro). Oral antithrombotic therapy is routinely prescribed for patients at increased risk for heart attack or stroke, principally with platelet activation inhibitors such as aspirin and clopidogrel (Plavix) (Cadroy et al., 2000). Many patients are resistant to oral antiplatelet agents, and inexpensive sensor technology to diagnose or monitor antiplatelet therapy could have significant clinical impact (Altman et al., 2004). Thickness shear mode (TSM) sensors have been used to monitor biological processes such as adsorption/desorption (Wu, 1999) and adhesion (Gryte et al., 1993; Fredriksson et al., 1998) in organic and biological films. In addition, a significant number

576

E. Ergezen et al. / Biosensors and Bioelectronics 23 (2007) 575–582

of TSM sensors have already been developed for medical applications, such as facilitating rapid detection of antibodies specific for HIV in human serum sample with sensitivity and specificity on par with a licensed HIV ELISA (Konig and Gratzel, 1995). TSM sensors were also used to detect human cells (Su et al., 1999), herbicides in drinking water (Yokoyama et al., 1995), bacteria such as Staphylococcus aureus (Pavey et al., 1999), drugs in human urine (Zhou et al., 2000), viruses such as hepatitis and African swine fever (Uttenthaler et al., 1998), and a variety of toxins (Rowe-Taitt et al., 2000). In all of those applications, the information about ambient processes or conditions was derived from the measured changes that are taking place in the thin films deposited on the surface of piezoelectric sensors. Several research teams have published promising experiments using thickness shear mode based sensors to monitor platelet adhesion (Kawakami et al., 1993; Matsuda et al., 1992). TSM sensors operating at 5 or 9 MHz, and coated with collagen, were shown to be sensitive to platelet adhesion using whole blood, platelet-rich plasma, or platelet/blood/cell suspensions. However, developing a sensor that is specific for platelet adhesion/aggregation using whole blood will not be simple, because plasma flowing over the sensor contributes to rapid changes of the sensor signal due to protein adsorption, the magnitude of which is similar to adherent platelets (Cavic et al., 2001). It is also unclear how these sensors can differentiate between inhibited platelet adhesion, activation, and aggregation. The long-term aim of the research discussed in this paper is to develop a unique class of easy-to-use, inexpensive, portable instruments for analysis of the hemostatic condition of an individual patient utilizing the feedback from a TSM sensor. Ideally, these instruments could be used in the clinic or at home without specialized training. The measurement technique utilizes acoustic sensor operating parameters to probe the interactions of blood with its surface at nanometer and micron length scales. Accordingly, such processes as plasma protein interaction, platelet glycoprotein binding, platelet adhesion, platelet aggregation, and fibrin coagulation could be monitored in real time using a very small blood sample. The potential research and clinical utility of an easy to use whole blood real-time sensor-based instrument for both platelet adhesion and aggregation are immense. Moreover, this technology may be applicable to monitor and study other cell–protein interactions. The focus of this study is to develop the experimental and theoretical foundations of the interactions of blood and platelets with collagen-coated TSM sensor. Gravity-driven platelet adhesion and aggregation in a static cell was used as an experimental model to demonstrate sensitivity to several platelet hemostatic mechanisms, particularly platelet aggregation and its inhibition by several distinct pathways. 2. Methods and materials 2.1. Chemicals and blood sample preparation Sensors were coated with insoluble fibrillar collagen type I (bovine Achilles tendon, Sigma) by adsorption for 60 min from acetic acid solution, using the methods of Folie et al. (1988).

Fresh whole porcine blood from commercial swine (Hatfield Meats, Hatfield, PA) was collected in 500 ml polyethylene bottles containing 3.2% (w/v) sodium citrate (1 part citrate to 8 parts blood). Platelet-rich plasma (PRP) was separated from red blood cells by centrifugation at 1400 rpm for 12 min at 22 ◦ C. Platelet-poor plasma (PPP) was prepared by centrifugation of PRP at 2500 rpm for 12 min. After plasma preparation, platelets were fluorescently labeled by incubation of the PRP with mepacrine (final concentration 10 ␮M; Sigma–Aldrich). ␣IIb ␤3 (GPIIb/IIIa) antagonist abciximab (ReoPro; Eli Lilly) was added to PRP at final concentration of 10 ␮g/ml. Putative inhibitors of ␣2 ␤1 and GPVI-mediated binding to collagen were added to PRP immediately prior to exposure to the sensor. Inhibitors were generously provided by Dr. Joel Bennett from University of Pennsylvania. 2.2. Sensor and measurement system 2.2.1. Thickness shear mode (TSM) sensor The TSM sensor is a piezoelectric-based sensor, which has the property that an applied alternating voltage (ac) induces mechanical shear strain and vice versa. By exciting the sensor with ac voltage, standing acoustic waves are produced within the sensor, and the sensor behaves as a resonator. The TSM sensor has been used since the 1950s to measure the deposited mass on the sensor in a vacuum (Sauerbrey, 1959), and is therefore often referred to as a quartz crystal microbalance (QCM), but for loading of the sensor with viscoelastic biological materials we refer to the sensor as a TSM to stress the sensitivity to medium viscosity and elasticity. The TSM sensor can be operated not only at the fundamental frequency, but also at odd harmonics. As the harmonic frequency of the piezoelectric plate increases, depth of penetration of the shear wave decreases (Reed et al., 1990; Shah et al., 2004). Therefore, by changing the frequency, one can control the distance at which the wave probes the liquid (Fig. 1a). For example, in phosphate buffered saline (PBS) the depth of penetration varies from 80 nm (35 MHz) to 250 nm (5 MHz). The magnitude of the TSM response in the vicinity of the fundamental resonant frequency is given in Fig. 1b, where the magnitude response of the TSM sensor is S21 (|S21 | = 20 log(100/(100 + Zt )), and Zt = total electromechanical impedance of the TSM sensor (Rosenbaum, 1998). When the TSM sensor is loaded with a biological media, there will be a shift in resonant frequency and a decrease in the magnitude, which can be correlated with changes in the mechanical properties of the medium. Depending on the changes in the mechanical properties of the sensor surface-medium interface, a positive and/or negative shift can be seen in the frequency response (Fig. 1b). The fluctuations on the frequency data are due to the final discrete resolution of frequency measurement system. In this paper, the changes in magnitude and frequency response were monitored in parallel and represented by α (α = αIR − αR ) and f (f = fRI − fR ) respectively.

E. Ergezen et al. / Biosensors and Bioelectronics 23 (2007) 575–582

577

Fig. 1. (a) Penetration depth for fundamental and harmonic resonant frequencies, (b) typical frequency-dependent response curve for the TSM sensor in the vicinity of the fundamental resonant frequency; αR = initial maximum magnitude, fR = initial resonant frequency, and in the case of both positive and negative frequency shifts throughout the experiment, αIR , αIIR = instantaneous maximum magnitudes of loaded TSM sensor at time t1 and t2 respectively, fRI , fRII = instantaneous resonant frequencies of the loaded TSM sensor at time t1 and t2 respectively. (Inset) Resonant frequency and magnitude are monitored as a function of time.

2.2.2. Measurement system A 14 mm diameter, 0.33 mm thick, 5 MHz quartz crystal with bonded 7 mm gold electrodes was placed in a custom fabricated brass sensor holder. The sensor holder was placed in a humidified, 37 ◦ C incubator and connected to a network analyzer (NA) (HP4395A). A personal computer controlled the network analyzer and collected the data. Maximum magnitude and frequency corresponding to the maximum magnitude were monitored at 5, 15, 25 and 35 MHz. The results reported here are for the sensor operating at 5 MHz unless otherwise indicated. 2.3. Experimental procedure Following assembly, the TSM sensor was monitored continuously as a series of 200 ␮l samples were placed on the sensor (Fig. 2). Reference measurements were taken using (1) air (to test attenuation less than 1.5 dB), (2) deionized water, and (3) PBS. The water was removed from the well, and a collagen suspension added and incubated for 60 min to allow collagen coating of the sensor surface by adsorption (4). The coated sensor was gently rinsed with PBS, followed by a 5 min reference measurement of the collagen-coated sensor covered with 200 ␮l of PBS (5). Next, platelet-poor plasma (PPP) was placed on the sensor surface for 30 min (6) to monitor plasma protein adsorption, fol-

lowed by PBS rinse (7) and reference measurement. Afterwards, platelet-rich plasma (PRP) was placed on the sensor for 60 min (8). At the end of the monitoring period, the sensor was washed with PBS to remove non-adherent cells and a final reference measurement was made (9). 2.4. Imaging Optical images were taken to verify the acoustic signal response, and to better correlate magnitude and frequency responses with biological processes occurring during platelet aggregation and adhesion. At the conclusion of monitoring, the rinsed sensor was removed from the cell and submerged in a fixative buffer (2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, Cat. # 15960, Electron Microscopy Sciences) for 1 h to fix the platelets and platelet aggregates adherent to the sensor surface, followed by storage in PBS solution until imaging within 7 days. Epifluorescence imaging of mepacrine-labeled platelets was used to quantify surface coverage and visualize aggregate morphology and density. A 12-bit high-resolution cooled CCD camera (Photometrics CoolSNAP fx) mounted to an inverted epifluorescence microscope (Nikon Diaphot 300) with a 20× lens was used for imaging. Epifluorescence imaging was followed by critical-point drying from CO2 to prepare the sensor for scanning electron microscopy (SEM) imaging (CPD5701, SPI Supplies, West Chester, PA). Sensor surfaces were coated with platinum (Denton Vacuum Desk II) and imaged by an environmental SEM at magnifications of 650× to 3500× (FEI XL30 FE-ESEM, Drexel University). 3. Results 3.1. Signal signatures of sensor to PRP interactions

Fig. 2. Time-dependent response of TSM sensor to imposed experiment conditions; PBS, phosphate buffer saline; PPP, platelet poor plasma; PRP, platelet rich plasma.

The collagen-coated sensor response to PRP exhibited three distinctive stages of signal attributes (Fig. 3a and b). The typical Stage-1 response lasted approximately 20 min after the addition of PRP; there was no change in α but a small increase in f. Stage-2 was characterized by a sharp decrease in both α and f responses over a period of 15–20 min. In Stage-3 there was

578

E. Ergezen et al. / Biosensors and Bioelectronics 23 (2007) 575–582

Fig. 3. Characteristic sensor signatures following exposure to collagen, plasma poor and plasma rich test samples partitioned into three distinct stages (Stage-1, Stage-2, and Stage-3): (a) α, (b) f, (c) fluorescence intensity images obtained at a specific time point (20, 25, 35, and 60 min), (d) average surface area coverage at time points 1, 2, 3, and 4.

an increase in both α and f responses, that was much more gradual than the Stage-2 responses. In particular while there was no change in f during Stage-1 at 5 MHz, there was a 200 Hz shift in the frequency response at 35 MHz. At the end of Stage-2, the 5 MHz frequency response reached minimum values at 45 min. In contrast, the 35 MHz minimum response was reached at 35 min. These differences indicate that the multi-resonance operation of the TSM sensor gives more robust information on the underlying biomechanics of the platelet aggregation and adhesion on collagen surface, and will be the scope of further theoretical study. 3.2. Correlation of PRP signatures with optical data In parallel experiments, fluorescence images were taken at different time points to allow visualization of the dynamics of the monitored platelet-sensor interactions. At set time point of 20, 25, 35, and 60 min following PRP application, the sensor was gently rinsed followed by fixation with glutaraldehyde. The fluorescence intensity gradually increased from 20 to 35 min, reached a peak at 35 min, and then decreased slightly during Stage-3 (Fig. 3c). Similarly, sensor surface area coverage as assessed by image analysis increased up to 35 min, and then decreased slightly (Fig. 3d). It appears from the images that the platelets fall onto the sensor surface, spread, and reach a critical point, which corresponds to an increase in area coverage during Stage-2. Some platelets contact adherent platelets to form small aggregates. When a maximum coverage has been reached, the platelet aggregates begin to contract. This corresponds to decreasing area coverage during Stage-3.

3.3. Initiation of platelet aggregation with adenosine 5 -diphosphate (ADP) ADP is one of the most important soluble physiological agonists, which activates platelets and triggers platelet aggregation (Colman et al., 2006). Ten milligrams of ADP (Sigma–Aldrich) was mixed in 10 ml PBS. Five minutes after PRP addition, 20 ␮l ADP solution was added to the PRP sample and briefly stirred with pipette tip. The sample was stirred again 1 min later to initiate aggregation of the activated platelets. Stage-2 of the sensor signal was observed immediately following the addition of ADP (Fig. 4a and b), while a control sensor with 20 ␮l PBS added and stirred at 1 min did not trigger an immediate Stage-2 response. With the ADP addition, there was slightly less of a decrease in the magnitude and frequency during Stage2, and flatter minima were observed, when compared to typical untreated PRP samples on collagen-coated sensors. Typical control and ADP added samples reached the minimum α and f 20 min after the platelet aggregation starts. This suggests that Stage-2 in the signature is caused partially by bulk aggregation followed by sedimentation to the sensor surface. Fluorescence images taken at 60 min showed that the exogenous ADP addition caused smaller platelet aggregates compared to that of the control experiment (Fig. 5a). 3.4. Inhibition of platelet aggregation with abciximab To further test the contribution of platelet aggregation to the sensor response characteristics, a treated blood sample with the clinical aggregation inhibitor abciximab (ReoPro; 10 ␮g/ml PRP) was compared to an untreated control. The

E. Ergezen et al. / Biosensors and Bioelectronics 23 (2007) 575–582

579

Fig. 4. Magnitude (a) and frequency (b) response as a function of added ADP; shown with typical control experiments and measured at 5 MHz. (c) Representatives magnitude responses to the inhibition of platelet aggregation by abciximab, (d) typical magnitude response with inhibition of ␣2 ␤1 collagen binding site.

abciximab-treated sample sensor-response was much weaker than the control (Fig. 4c), showing a much-delayed Stage-2 with minimal to no decrease in either magnitude or frequency at all monitored frequencies. Fluorescence images from abciximab experiments showed a paucity of platelets adherent to the collagen-coated sensor sur-

face, with a few surface adherent aggregates and overall much lower surface coverage and few smaller aggregates compared to the control (Fig. 5b). 3.5. Inhibition of collagen binding sites Additional tests were run using a putative inhibitor of binding of collagen to the platelet integrin ␣2 ␤1 receptor, compared to the responses to an uninhibited control sample (Fig. 4d). The inhibited sample showed the typical sensor responses, but Stage-2 was delayed by almost 20 min, indicating a delay in the aggregation stage, presumably caused by the integrin inhibitor. Similar results were obtained in samples treated with a putative inhibitor to platelet collagen binding via GPVI, the other physiologically important platelet collagen receptor (not shown). Fluorescence images taken after 60 min in each test show brighter platelet aggregates and more surface coverage of the collagen-inhibited sample than the control sample (Fig. 5c). It should be noted that both tests were started and finished at the same time. The delay in Stage-2 is due to the collagen binding inhibitor and indicates a delay in activation and aggregation, but the subsequent Stage-2 signal appears similar to control. Thus the inhibited sample is more comparable to an untreated control sample observed at 35 min, which showed brighter platelet aggregates and denser surface coverage. 4. Discussion

Fig. 5. Fluorescence intensity images shown at 60 min time point (a) ADP treated and control samples; (b) abciximab treated and control samples; (c) inhibition of collagen binding sites and control sample.

Based on TSM sensor response signatures, we characterized different stages (Stage-1, Stage-2 and Stage-3) of platelet

580

E. Ergezen et al. / Biosensors and Bioelectronics 23 (2007) 575–582

adhesion and aggregation from PRP. In addition, stimulation of earlier aggregation by ADP and inhibition of platelet aggregation with abciximab revealed that the large response in both magnitude and frequency during Stage-2 might be attributed to aggregation of platelets. Selective inhibition of collagen binding to platelet receptors (␣2 ␤1 and GPVI) prolonged the time duration of Stage-1 while it had no effect on Stage-2 and Stage-3 for all the recorded harmonics. Real-time microscopy performed under similar experimental conditions (static platelet-rich plasma samples), showed that initially a small number of “vanguard” platelets bind to collagen and begin spreading rapidly (Patel et al., 2003). Follower platelets attached to the “vanguard” platelets via the expressed filopodia. Shortly after the “follower” platelets adhered to nearby collagen, they started spreading on collagen or directly on the platelets to which they were attached. This dynamic process resulted in sheet-like underlayers on the surface of the sensor. Patel et al. (2003) point out that platelets activated and aggregated in the bulk often subsequently sediment on, adhere to and spread on adherent platelets and aggregates. Treatment with monoclonal antibody inhibiting ␣IIb ␤3 receptors did not affect platelet adhesion to the collagen surface, but did prevent the interactions between platelets and thereby changed the typical sheet-like spreading of the platelets on collagen. In many ways the TSM sensor and the biofunctionalized active surface behave as a giant pseudo-platelet that can be made to attract and activate the neighboring platelets that come in contact with it. In our experiments, there was no significant change in α response for all the measured harmonics during sedimentation phase and initial adhesion of individual platelets on the collagen-coated sensor surface (Stage-1). From the sensor’s vantage point this suggests that these processes do not have any dissipative effect, which would correspond to changes in the viscous properties at the interface. In contrast, the f response showed distinctive differences between the responses, especially the first and seventh harmonics. While there is no visible change at 5 MHz, a positive shift was observed at 35 MHz during Stage-1. It has been shown that a positive frequency shift can be interpreted as an increase in the mechanical shear stiffness at the interface (Hong et al., 2006). Analysis of the PPP measurements demonstrated that the kinetics of the frequency and magnitude responses of the sensor to PPP loading were quite similar to Stage-1 after PRP loading (data is not shown). This similarity suggests that adsorption of plasma proteins to the collagen-coated sensor surface is a contributing factor to the sensor response during Stage-1. The more concentrated plasma protein adheres to the surface first, and then the other proteins present in the plasma that have higher affinity to collagen replace it. This type of sequential competitive adsorption to artificial surfaces (Vroman et al., 1980) is now known as the Vroman effect and causes a more elastic or more firmly bound protein layer at the surface. The difference between f response of the harmonics considered may stem from the fact that the penetration depth decreases with the increase in operating frequency of the sensor (Cote et al., 2003; Hieda et al., 2004). Our results indicated that the sensor response to Stage-1 is primarily due to platelet binding to both collagen and to pre-absorbed protein layers, and that the

TSM sensor operating at 35 MHz is more sensitive to the molecular interactions such as the Vroman effect occurring in very close proximity to the sensor surface, compared to operating at 5 MHz. The duration of Stage-1 is closely related to the time required for sufficient platelets to reach the sensor surface and the time needed for collagen-mediated platelet activation. As soon as PRP is added onto the sensor surface, platelets which are close to the sensor surface will start binding to the collagen layer mainly via the ␣2 ␤1 and GPVI receptors, and then start releasing platelet agonists such as ADP and thromboxane A2 (Shah et al., 2001). Platelet activation and agonist release is strong if ␣2 ␤1 and GPVI receptors are clustered together, which occurs due to cytoskeletal changes when both receptors interact with collagen (Clemetson and Clemetson, 2001). Receptor density is also critically important in GPVI-collagen interactions and GPVI-mediated signaling (Chen et al., 2002). Supporting these hypotheses, inhibition of one of the main collagen binding receptors of the platelets (GPVI or ␣2 ␤1 ) doubled the duration of Stage-1 while no effect was seen on the durations and kinetics of Stage-2 and Stage-3. The inhibition of one of the receptors, for example ␣2 ␤1 , may delay the clustering process of the receptors, and require more time to recruit sufficient receptors to create a signal for platelet activation and aggregation. This theory emphasizes the importance of the receptor density dependence of the platelet aggregation and the similar roles of GPVI and ␣2 ␤1 receptors in the signaling process. Interestingly, to date no study has showed the equal importance of these receptors in the signaling. Some studies showed that GPVI is the main receptor mediating both adhesion and the signaling response (Nieswandt et al., 2001; Masberg et al., 2003), while others reported the importance of ␣2 ␤1 in adhesion and aggregation of platelets (Saelman et al., 1994). Therefore our results may bring a new perspective to the recruitment of the platelet receptors in the signaling pathway under no-flow conditions. We showed that platelet aggregation is responsible for the large change in both magnitude and frequency response in Stage2. Platelet aggregation was initiated by introducing ADP 5 min after PRP addition. ADP normally acts as a secondary agonist to recruit additional platelets in the formation of a stable platelet plug (Woulfe et al., 2001). As soon as the ADP was added to PRP, a typical Stage-2 was observed in both α and f responses. ADP addition did not change the duration of Stage-2 (∼15 min) but interestingly it did decrease the change in magnitude response while no significant difference was observed between f response magnitudes of control and treated sample. This suggests that while the α response is a combination of multiple processes such as plasma proteins binding to collagen and platelets bonding to collagen, Stage-2 f response is caused mainly by platelet aggregation. From the sensor point of view, these interactions somehow created a less lossy medium, but no change is seen in mass loading and elastic component of the medium. Platelet aggregation was inhibited by using monoclonal antibody abciximab, which binds to ␣IIb ␤3 receptors on platelets. These receptors support platelet aggregation by binding plasma adhesive proteins such as fibrinogen and vWF

E. Ergezen et al. / Biosensors and Bioelectronics 23 (2007) 575–582

581

Table 1 Model of platelet-sensor interactions Stage-1

Stage-2

Stage-3

Sedimentation of platelets, initial adhesion and spreading of platelets on collagen. Clustering of receptors. Vroman effect due to the blood plasma proteins Vroman effect → positive shift in f (at 35 MHz) Clustering of receptors → delay time Sedimentation → delay time

Platelet aggregation. Aggregate sedimentation. Continuous increase in surface coverage area Negative shift in the f and α responses

Contraction of platelet aggregates. Formation of isolated islands. Decrease in surface coverage area

Schematic diagram

Physical processes

Characterization factors

(Scarborough et al., 1999). Blocking the ␣IIb ␤3 receptors with abciximab completely removed the Stage-2 f response and significantly decreased α. Stage-3 is characterized by positive increase in both α and f responses. Surface area coverage decreases during this stage, as platelets spreading at different parts of the surface contacted each other, then contract towards each other. Patel et al. (2003) also observed similar results in studies of platelet-type 1 collagen interactions by video phase contrast microscopy. It has been shown that the changes in frequency and magnitude responses are linearly proportional to the percent surface coverage (Redepenning et al., 1993; Marx et al., 2003; Wegener and Janshoff, 1998). Therefore, the decrease in the surface coverage may also contribute to gradual decrease in frequency and magnitude responses. The experiments performed to date suggest a model relating the observed biological processes to the sensor characteristics (Table 1). During Stage-1, individual platelets reach the sensor surface via sedimentation, where they adhere and spread. Simultaneously, plasma protein adsorption affects the sensor signal, particularly at the highest frequencies, where higher affinity binding may increase resonant frequency. During Stage2, platelets aggregate on the sensor surface, and activated platelets and aggregates sediment to the sensor surface, which causes a rapid decrease in magnitude and resonant frequency. In Stage-3 aggregate sedimentation ends, and aggregates are consolidated on the sensor surface, leading to increased frequency and magnitude and the observed decrease in surface coverage. According to this model, the TSM sensor is sensitive to inhibition or promotion of platelet activation mechanisms (e.g., ADP or collagen-induced activation) by the duration of Stage-1, and is sensitive to aggregation mechanisms (e.g., ␣IIb ␤3 -mediated aggregation) by the Stage-2 magnitude or frequency change. The experiments also suggest that plasma protein affinity for either protein coatings or biomaterials can be evaluated based on Stage-1 response, particularly at high frequencies.

Positive shift in the f and α responses

5. Conclusion The experiments presented demonstrate TSM sensitivity to the different phases of platelet adhesion, activation and aggregation on collagen-coated surfaces, and corresponding physical phenomena. The duration of Stage-1 was determined by the activation kinetics due to collagen binding and the sedimentation time. Stable α response during Stage-1 suggested that protein adsorption and platelet adhesion on collagen have elastic but not dissipative effect. The large change in α and f responses during Stage-2 is primarily due to platelet aggregation, but the plasma protein-collagen interactions continue to affect the α response. Positive shifts in α and f responses in Stage-3 are mainly due to the formation of distinctive islands causing a decrease in the coverage area. These experiments show that the technique is sensitive to detect the effect of interference with specific molecular mechanisms, which are involved in the platelet adhesion and aggregation on collagen surface. Multi-resonant operation of TSM sensor provided insight and exposed unique features of blood plasma protein’s absorption on the collagen-coated surface. Finally, the TSM sensor may provide a real time, miniature sensing platform that functionally mimics many of the attributes of an activated pseudo-platelet making it a particularly useful assay system to be used in the development of novel antithrombotic drugs, the diagnosis of platelet disorders, or the monitoring of antiplatelet therapy. Acknowledgements We are thankful to Dr. Joel Bennett, Chang-Beom Kim, and Robert Weisbein for their help throughout the project. References Altman, R., Luciardi, H., Muntaner, J., Herrera, R., 2004. J. Thromb. 2 (1), 1.

582

E. Ergezen et al. / Biosensors and Bioelectronics 23 (2007) 575–582

Cadroy, Y., Bossavy, J.P., Thalamas, C., Sagnard, L., Sakariassen, K., Boneu, B., 2000. Circulation 101, 2823–2828. Cavic, B.A., Freedman, J., Morel, Z., Mody, M., Rand, M.L., Stone, D.C., Thompson, M., 2001. Analyst 126, 343–348. Chen, H., Locke, D., Liu, Y., Liu, C., Kahn, L.M., 2002. J. Biol. Chem. 277 (4), 3011–3019. Clemetson, K.J., Clemetson, J.M., 2001. Thromb. Haemost. 86, 189–197. Colman, R.W., Clowes, A.W., George, J.N., Goldhaber, S.Z., Marder, V.J., 2006. In: Colman, R.W., Marder, V.J., Clowes, A.W., George, J.N., Clowes, A.W., Goldhaber, S.Z. (Eds.), Hemostasis and Thrombosis: Basic Principles and Clinical Practice, 5th ed. JB Lippincott, Philadelphia, pp. 1–17. Cote, G.L., Lec, R.M., Pishko, M., 2003. IEEE Sens. J. 3 (3), 251. Folie, B.J., McIntire, L.V., Lasslo, A., 1988. Blood 72, 1393. Fredriksson, C., Kihlman, S., Rodahl, M., Kasemo, B., 1998. Langmuir 14 (2), 248–251. Gryte, D.M., Ward, M.D., Hu, W.S., 1993. Biotechnol. Prog. 9 (1), 105–108. Hieda, M., Garcia, R., Dixon, M., Daniel, T., Allara, D., Chan, M.H.W., 2004. Appl. Phys. Lett. 84 (4), 628–630. Hong, S.J., Ergezen, E., Barbee, K., Lec, R., 2006. Biomaterials 27, 5813– 5820. Kawakami, K., Harada, Y., Sakasita, M., Nagai, H., Handa, M., Ikeda, Y., 1993. ASAIO J. 39, M558–M560. Konig, B., Gratzel, M., 1995. Anal. Chim. Acta 309 (1/3), 19–25. Marx, K.A., Zhou, T., Warren, M., Braunhut, S.J., 2003. Biotechnol. Prog. 19, 987–999. Masberg, S., Gawaz, M., Gruner, S., Schulte, V., Konrad, I., Zohlnhofer, D., Heinzmann, U., Nieswandt, B., 2003. J. Exp. Med. 197 (1), 41–49. Matsuda, T., Kishida, A., Ebato, H., Okahata, Y., 1992. ASAIO J. 38 (3), M171–M173. Nieswandt, B., Brakebusch, C., Bergmeier, W., Schulte, V., Bouvard, D., Mokhtari-Nejad, R., Lindhout, T., Heemskerk, J., Zirngibl, H., Fassler, R., 2001. EMBO J. 20 (9), 2120–2130.

Patel, D., Vaananen, H., Jirouskova, M., Hoffmann, T., Bodian, C., Coler, B.S., 2003. Blood 101 (3), 929. Pavey, K.D., Ali, Z., Olliff, C.J., Paul, F., 1999. J. Pharm. Biomed. Anal. 20 (1/2), 241–245. Redepenning, J., Schlesinger, T.K., Mechalke, E., Puleo, D.A., Bizios, R., 1993. Anal. Chem. 65, 3378–3381. Reed, C.E., Kanazawa, K., Kaufman, J., 1990. J. Appl. Phys. 68, 1993. Rosenbaum, J.F., 1998. Bulk acoustic Wave Theory and Devices. Artech House Publishers, Boston, MA. Rowe-Taitt, C.A., Golden, J.P., Feldstein, M.J., Cras, J.J., Hoffman, K.E., Lgler, F.S., 2000. Biosens. Bioelectron. 14.10 (11), 785–794. Saelman, E.U., Nieuwenhuis, H.K., Hese, K.M., de Groot, P.G., Heijnen, H.F., Sage, E.H., Williams, S., McKeown, L., Gralnick, H.R., Sixma, J.J., 1994. Blood 83, 1244–1250. Sauerbrey, G., 1959. Z. Phys. 155, 206. Scarborough, R.M., Kleiman, N.S., Philips, D.R., 1999. Circulation 100, 437–444. Shah, B.H., Rasheed, H., Rahman, I.H., Shariff, A.H., Khan, F.L., Rahman, H.B., Hanif, S., Saeed, S.A., 2001. Exp. Mol. Med. 33 (4), 226–233. Shah, P., Lec, R., Kwoun, S., 2004. Medical Devices and Biosensors, 2nd IEEE/EMBS International Summer School, June 26–July 2, pp. 12–16. Su, X.D., Chew, F.T., Li, S.F.Y., 1999. Anal. Biochem. 273 (1), 66–72. Uttenthaler, E., Kosslinger, C., Drost, S., 1998. Biosens. Bioelectron. 13 (12), 1279–1286. Vroman, L., Adams, A.L., Fischer, G.C., Munoz, P.C., 1980. Blood 55, 156–159. Wegener, J., Janshoff, A., 1998. Eur. Biophys. J. 28, 26–37. Woulfe, D., Yang, J., Brass, L., 2001. J. Clin. Invest. 107 (12), 1503–1505. Wu, T.Z., 1999. Biosens. Bioelectron. 14 (1), 9–18. Yokoyama, K., Ikebukuro, K., Tamiya, E., Karube, I., Ichiki, N., Arikawa, Y., 1995. Anal. Chim. Acta 304 (2), 139–145. Zhou, A.H., He, D.L., Nie, L.H., Yao, S.Z., 2000. Anal. Biochem. 282 (1), 10–15.