Surface immobilisation and properties of smooth muscle cells monitored by on-line acoustic wave detector

Abstract

The attachment of rat aortic smooth muscle cells to various surfaces has been monitored by a thickness shear mode acoustic wave device incorporated into an on-line configuration. Using the total injection analysis method, laminin and fibronectin were adsorbed to the device surface, to be followed by introduction of cells into the system. The results of these experiments in terms of frequency and motional resistance measurements were also compared with those for cell attachment to the bare gold electrode of the sensor. The responses of the surface-bound cells to the introduction of various ions, depolarisation events and damage subsequent to exposure to hydrogen peroxide were also observed. Morphological changes in the cells, as confirmed by scanning electron microscopy , are correlated with results of the acoustic wave measurements.

---

Introduction

The behaviour of cells on surfaces is an important phenomenon from the perspectives of both fundamental science and medicine.^1,2 With regard to the latter, there is an obvious interest in the role of cell–substrate chemistry in terms of implants and other devices subject to issues of biocompatibility. From the pharmaceutical standpoint, there is increasing interest in the behaviour of single cells and populations of communicating cells following exposure to small molecules, in particular, drugs. The immobilisation of cells onto a substrate surface is often a precursor to studies of morphological and other changes. An additional related application is that where the cell–device combination acts as a sensing structure in its own right. An example of this approach is the so-called neural biosensor which has been employed as a detector for various chemical species.³

Following the initial cell–surface interaction, morphological changes such as cell spreading can occur.⁴ With respect to cell growth, it has been found that this oft-desired phenomenon is anchorage-dependent. Many factors are expected to play a role in cell growth, one example being the influence of the substrate–extracellular matrix (ECM) interaction. In view of the above, it is not surprising that there has been a significant effort to develop techniques for the study of cell–surface interactions, and fluorescence and electrochemical methods have figured prominently in recent times. However, it is clear that a limited number of techniques exist that are capable of label-free, real-time monitoring of such interactions.¹

The thickness shear mode acoustic wave device (TSM) offers sensitive and label-free detection, can be employed in a flow-injection system and has potential for miniaturization. It has been applied widely in the field of bioanalytical chemistry in recent years, largely owing to these properties.² The main applications of the TSM have been for the detection of nucleic acid hybridisation,⁶protein conformational changes,⁵immunochemistry ⁷ and other areas.^8,9 With respect to cells, studies have been performed on bacterial,¹⁰ cancerous,^8,11 endothelial,^12,13 and cardiac muscle cells,^14,15 and on blood platelets.¹⁶ A recent paper described TSM detection of both inter-neuron communication and response to drugs for surface-attached populations of hypothalamic neurons displaying different levels of cell surface coverage.¹⁷ Finally, single cell detection has been achieved by fabricating microfluidic channels on the surface of a TSM device.¹⁸

In the present work, we have used an on-line TSM system to measure rat aortic smooth muscle cell attachment to various surfaces in a quantitative fashion. The surfaces include bare gold and extracellular matrix species such as fibronectin and laminin. Additionally, ions (Ca²⁺ or Mg²⁺) were used to affect the cell morphology and a depolarisation event instigated by KCl was detected. The oxidant H₂O₂ is well known to cause cell morphological changes, so it was used as a probe to ascertain if this acoustic technique is capable of detecting such structural alterations. Correlation of the acoustic observations with surface electron microscopy (SEM ) was also conducted.

TSM as a cellular biosensor

The TSM device functions through the generation of a resonant shear wave in an electroded quartz wafer. The acoustic wave travels through the quartz with very little dissipation, and is reflected at the solid interface to maintain the standing wave. Changes at the substrate surface affect the apparent thickness and acoustic properties of the quartz wafer, and can be measured as changes in the resonant frequency of the shear wave. Biochemical events at the device surface can cause changes in wave properties that are monitored over time.

In the purely adsorptive limit, a decrease in the series resonant frequency f_s (measured in Hz) corresponds to an increase in the thickness of the attached layer, with no dissipation. However, when this device is operated in a liquid environment, a portion of the acoustic energy is not reflected at the solid/liquid interface, and is instead dissipated into the liquid. This dissipation is measured as an increase in the motional resistance (R_m). Both the decrease in f_s and increase in R_m due to the liquid interaction, to a first approximation, are related to the density and viscosity of the liquid. Since a liquid does not propagate a shear wave to any great distance, the acoustic wave will dissipate most of its energy within a short distance δ of the device surface. For 9 MHz quartz crystals in aqueous liquids, δ is on the order of 200 nm, and is shown schematically in Fig. 1.

Fig. 1 Schematic representation of the shear wave penetration depth into a muscle cell on a TSM surface. Of interest is that the shear wave does not penetrate through the entire cell, but instead dissipates within the cytoplasm. Cell sizes and shear wave dissipation into the cytoplasm are not to scale. δ is the viscous decay length into the fluid.

It will become apparent from the results presented below that when this device is used as a cellular biosensor , it does not function purely as a deposition sensor, as is normally assumed. Instead, this technique is capable of detecting cell deposition by measuring the attachment of the cell membrane and the properties of the liquid cytoplasm in contact with the device surface.

Materials and methods

Materials

Dulbecco's phosphate buffered saline (DPBS, Sigma, St. Louis, USA) was used as the reaction medium throughout this work. Calcium chloride dehydrate and magnesium chloride hexahydrate (Sigma) were used as 1 mM solutions of CaCl₂ and MgCl₂ in buffer. Potassium chloride (J.T. Baker, New Jersey, USA) was used at 60 mM in DPBS buffer. Laminin and fibronectin (VWR, Pennsylvania, USA) were used at final concentrations of 50 µg ml^–1 in buffer. Trypsin (Sigma) was dissolved in DPBS buffer to a final concentration of 0.05% (w/v). Iscove's modified Dulbecco's medium (IMDM, Invitrogen Corp, Ontario, Canada) containing 10% fetal bovine serum (FBS, Invitrogen Corp) was used in the cell culture experiment. H₂O₂ (Sigma) was employed as a 200 µM solution. The mixture of collagenase and Trypsin was used in cell culture and they were obtained from Sigma and Invitrogen (Difco) separately. 2% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.3) and 0.1 M sodium cacodylate buffer with 0.2 M sucrose (pH 7.3) for fixing cells onto surfaces for SEM experiments were gifts from Mount Sinai Hospital, Toronto.

On-line TSM instrument

As described previously,⁵ the TSM instrument is mainly composed of two parts: AT-cut 9 MHz piezoelectric quartz crystals with gold electrodes and a flow-through system. The crystals were obtained from LapTech Corp. (Bowmanville, Canada). One side of the crystal was exposed to liquid in the flow-through cell, while the other side was flushed with nitrogen gas. A network analyzer (HP4195A Network/Spectrum Analyser, Hewlett Packard, Colorado Springs, USA) was employed to measure the impedance change on the gold electrodes of the sensor. Changes of series resonance frequency (f_s) and motional resistance (R_m) were recorded for analysis. All acoustic wave experiments were conducted inside an incubator held at 37 °C in order to keep the cells alive.

Cell culture

The aortic smooth muscle cells used in these experiments were taken from a three-month old male Lewis rat. The animal protocol for the study was approved by the Animal Care Committee of the University Health Network, Toronto, and all animal procedures were performed according to the Guidefor the Care and Use of Laboratory Animals.¹⁹ The cell culture procedure was similar to previous descriptions in the literature.²⁰ The rat was euthanised and the thoracic aorta was harvested by aseptic technique. Briefly, the rat was excised and the aorta was removed. After removing the surrounding connective tissues and inner endothelial layer, the aorta was minced and digested by a mixture of collagenase and trypsin. The harvested cells were cultured in IMDM containing 10% FBS and then incubated at 37 °C in a 5% CO₂ humidified incubator. In this study, the cells were subcultured and second to fourth passaged cells were utilised as experimental cells. For the cell introduction and surface detachment experiments, the subcultured cells were kept in DPBS immediately before injection into the TSM flow cell. In order to study the effect of ions and cell membrane damage, the cells were cultured directly on the surface of the TSM sensor. At full coverage, there should be approximately 4000 cells on the gold electrode surface.

Cell introduction and surface detachment

Different extracellular matrices were used to investigate the attachment preference of the aortic smooth muscle cells. Using the total injection analysis method, laminin and fibronectin were immobilised on the TSM surface by chemisorption. Cells in the DPBS buffer were injected subsequent to binding of the protein . Additionally, a bare electrode surface was also used for evaluation of the possibility of the cell attachment. Trypsin was employed in order to effect the detachment of cells from the various surfaces.

Ionic influence and cell depolarisation

Ca²⁺ and Mg²⁺ (both at 1 mM) were used separately to affect the attachment of the cells onto the surface. K⁺ (60 mM) was used to depolarise surface-attached cells. A TSM device with cultured cells in place was introduced into the flow-through system and the DPBS was injected until a stable baseline was achieved. Ions (Ca²⁺, Mg²⁺ and K⁺) in DPBS buffer were pumped over the sensor surface by flow injection for 15 min. DPBS buffer was employed in a final wash.

Cell interaction with hydrogen peroxide

In a similar protocol to that used for cell depolarisation, the device with attached cells was placed in the flow-through cell. Instead of using DPBS, the cell culture medium (IMDM) was injected until a stable baseline was achieved. Hydrogen peroxide was diluted to 200 µM by IMDM and pumped over the TSM surface. The pump was stopped for 15 min and the cell culture medium was then used to wash the device in a final step.

SEM

A scanning electron microscope was utilised to conduct a detailed study of the changes in cell morphology instigated by hydrogen peroxide. Two samples were analysed, these being surfaces before and after treatment with reagent. For this experiment, it was necessary to fix the cells onto the sensor surface. The following steps were taken for this purpose. First, a surface with cells attached was kept in 2% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.3) for 1.5 h to fix the cells. The device was then rinsed in 0.1 M sodium cacodylate buffer with 0.2 M sucrose (pH 7.3) for 10 min. It was then soaked in 70% and 90% ethanol, each for 10 min. The surface was then washed three times with 100% ethanol. Finally, the sensor was maintained in 100% ethanol. After the steps for dehydration, the samples were critical-point dried, attached to aluminium stubs, and gold-sputter coated.

Results

Surface characterisation and cell detection

Three different surface modifications were used to facilitate aortic smooth muscle cell attachment in this study. These were laminin, fibronectin, and bare gold. Fig. 2 shows a sample set of results for deposition of the surface modifications, adhesion of cells to the surface, and subsequent washing with buffer. These experiments were repeated in triplicate and similar frequency shifts were obtained in all cases. However, the changes in R_m on washing with buffer following cell adhesion were not always reproducible. The reason for this behaviour remains unclear.

Fig. 2 TSM signals for cell immobilisation on (a) laminin, (b) fibronectin and (c) bare gold surfaces, showing Δf_s (black lines) and R_m (grey lines). L, F, C, D represent the addition of laminin, fibronectin, cells, and DPBS buffer, respectively.

For the first two cases, there is a decrease in f_s and an increase in R_m when laminin or fibronectin were applied. The changes in R_m, while not large, do indicate some interfacial and viscoelastic effects in the monolayers. The shifts for laminin were Δf_s = –401.0 Hz and ΔR_m= 23.6 Ω, while those for fibronectin were Δf_s = –306.3 Hz and ΔR_m = 2.6 Ω.

Fig. 2a shows the device response for cell addition to the laminin-modified surface, which induces a small decrease in f_s and a large increase in R_m. However, upon washing with buffer, the signals returned to their original values, indicating that the cells attached on a temporary basis only and did not adhere strongly to the laminin surface. From this, laminin appears to be unsuitable for the attachment of smooth muscle cells on the device surface. It is unclear why cells do not adhere to the laminin surface.

The device response for cell deposition to a fibronectin-coated surface is shown in Fig. 2b. Addition of the cells resulted in a frequency decrease of –140 Hz and a resistance increase of 14 Ω. Subsequent washing with buffer resulted in net shifts of Δf_s = –110 Hz and ΔR_m = 4.3 Ω. This indicates that fibronectin is likely a suitable surface for aortic smooth muscle cell attachment.

The device response for cell adhesion to bare gold is shown in Fig. 2c. After cell deposition and subsequent washing with buffer, the net frequency and resistance shifts were –172 Hz and 4.0 Ω, respectively, indicating good attachment. This is consistent with previous reports for attachment of cells to bare gold,^10,11 although different cell types were used in our study.

Cell adsorption kinetics

Using the same experimental data, we can compare the attachment kinetics for these cells. This information reveals the dynamics of cell attachment, as well as providing information to study cell detachment using trypsin. The kinetics are determined from changes in the differential frequency and resistance shifts over time. Since the frequency shift is proportional to the surface adsorption of the cells, the magnitude of the differential value is linked to the attachment/detachment kinetics.²¹ As seen in Fig. 3, cell attachment to bare gold displays the fastest kinetics, while attachment to fibronectin has the slowest. From these results, the laminin surface has the least cell capacity, even though it has higher attachment kinetics (Fig. 3a). By contrast, the fibronectin surface showed much better cell capacity with lowest kinetics (Fig. 3b). Surprisingly, the gold surface has both the best cell capacity and the highest kinetics (Fig. 3c). The comparison is also shown in Table 1. Since fibronectin and gold have a better capacity for cell attachment than laminin, only these surfaces are used for the later experiments. As seen from Fig. 4, cell detachment instigated by trypsin from the gold surface shows lower kinetics than for the fibronectin surface, which further confirms the availability of the gold surface for cell attachment.

Fig. 3 Attachment kinetics on (a) laminin, (b) fibronectin and (c) bare gold surfaces, showing δ(Δf_s) (black lines) and δR_m (grey lines). L, F, C, D represent the addition of laminin, fibronectin, cells, and DPBS buffer, respectively. δ(Δf_s) is the finite differential of changes in Δf_s with respect to time, δ(Δf_n) = (Δf_n – Δf_n – 1)/(t_n – t_n – 1). The calculation for δR_m is similar.

Fig. 4 Detachment kinetics of the surface-attached cells by trypsin on (a) fibronectin and (b) bare gold surfaces, showing δ(Δf_s) (black lines) and δR_m (grey lines). F, C, D, T represent the addition of fibronectin, cells, DPBS buffer, and trypsin, respectively. The finite differentials were calculated as indicated in the caption to Fig. 3.

Table 1 Comparison of acoustic results upon cell attachment to different surface modifications. The frequency and resistance shifts are the differences in the signals before introduction of cells, and after washing with buffer. The frequency kinetics listed is the largest negative value calculated

Surface	Δfs/Hz	ΔRm/Ω	Kinetics/Hz s^–1
Laminin	ca. 0	ca. 0	–44
Fibronectin	–110	4.3	–25
Gold	–172	4.0	–72

---

Factors affecting cell morphology

To further test the device response, we studied the effect of divalent ions on cell attachment. Integrins are intramembrane proteins that are believed to play an important role in the attachment and spreading of animal cells. They require divalent cations in the extracellular medium to bind to specific recognition sites on ECMproteins .²² In this study, Ca²⁺ and Mg²⁺ (both at 1 mM) were used separately to affect the attachment of the smooth-muscle cells. Fig. 5 shows the effect of divalent ions. Binding of 1 mM CaCl₂ and MgCl₂ each resulted in decreases in f_s of –12.1 Hz (CaCl₂) and –18.7 Hz (MgCl₂). The R_m shifts were more difficult to determine, so a decaying exponential of the form y = Cexp(βx + α) + D was used to fit the wandering baseline. This yielded slight increases in R_m, of 0.7 Ω and 0.1 Ω for CaCl₂ and MgCl₂, respectively.

Fig. 5 Ion effect on cell attachment and depolarisation, showing Δf_s (black line) and R_m (grey line). Concentrations were 60 mM (KCl) and 1 mM (CaCl₂ and MgCl₂). The decaying exponential regression line used to fit the wandering baseline is not shown.

In addition to divalent ions, the monovalent KCl (60 mM) was used to depolarise the cell membrane. Both the frequency and resistance decreased, as seen in Fig. 5, which differs from the signals caused by the lower concentration divalent ions. There was a frequency decrease of –10.0 Hz, and a baseline-corrected resistance decrease of –1.7 Ω. This result, whereby the shifts of both f_s and R_m are in the same direction, is uncommon in TSM operation, although it has been observed previously and attributed to a number of different processes.^23–27 These results will be discussed in more detail below.

Exposure of the cells to H₂O₂ was studied to compare the acoustic signals from the TSM device to visual images obtained through microscopy. The microscopy work is described below. For the TSM study, cells were first cultured directly onto the bare gold surface and the signal was allowed to stabilise by injection of IMDM (approx. 25 min). Hydrogen peroxide was flowed over the cells, resulting in decreases in both f_s and R_m of –74.6 Hz and –9.9 Ω, respectively, as shown in Fig. 6. The device surface was then washed with culture medium (IMDM), leading to a return to near-baseline values for f_s and R_m.

Fig. 6 Hydrogen peroxide effect on cell attachment and depolarisation, showing Δf_s (black line) and R_m (grey line). H is the addition of H₂O₂ and M is the addition of medium.

Scanning electron microscopy

To test the explanations for the acoustic shifts observed with the TSM, SEM was used to monitor the cell morphology on exposure to H₂O₂. Hydrogen peroxide was used as an oxidant to damage the cell membrane and induce morphological changes to the cells. Fig. 7 shows the cells before (a) and after addition of 200 µM H₂O₂ (b,c). Before addition, Fig. 7a shows healthy, flat smooth muscle cells. However, the structure of the cells was changed after being interacted with hydrogen peroxide, shown in Fig. 7b and 7c. The membranes of these cells appear damaged and swollen.

Fig. 7 Images of surface-attached cells under SEM : (a) shows the cells' structure before interacting with H₂O₂; (b) and (c) show the morphology of the cells after the addition of the hydrogen peroxide.

Discussion

Surface characterisation

We begin by analysing the deposition of the surface modification agents, laminin and fibronectin. While a full analysis of these binding events is not the focus of this article, we can obtain a very rough estimate of the extent of surface binding by using the Sauerbrey equation,²⁸ and then dividing the result by a factor of two.^29,30 The Sauerbrey formula is given by where 2/Z_q ≈ 2.26 × 10^–6 cm² g^–1 Hz^–1, the fundamental frequency f₀ is 9.0 MHz and the area is 0.28 cm². The molecular weights of laminin and fibronectin are 600 and 500 kDa, respectively, so the shifts of 400 Hz for laminin (Fig. 2a) and 300 Hz for fibronectin (Fig. 2b) correspond to surface coverages of 1.6 and 1.8 pmol cm^–2. This is similar to values reported for other protein –surface interactions.^31,32

Acoustic detection of cell adsorption

We attempt a similar analysis for the adsorption of cells onto the fibronectin and bare gold surfaces (we do not consider laminin, since the frequency shift was negligible). The frequency shifts for fibronectin (–110 Hz) and bare gold (–172 Hz) correspond to mass depositions of 167 ng and 261 ng, respectively, as calculated from eqn (1). Assuming the dry weight of a single cell is 100 pg,³³ and the live weight is ten times that value,³⁴ these shifts correspond to 167 and 261 cells on the fibronectin and bare gold surfaces.

The number of cultured cells on the TSM surface with full, 100% coverage is approximately 4000. Optical microscopy results (not shown) reveal that 50% cell coverage was achieved on both surfaces, which corresponds to 2000 cells. The values calculated from eqn (1) are lower than this figure by a factor of ten, indicating that the device is not measuring cell mass.

The TSM is sensitive to many more factors than adsorption, and is affected by a wide range of surface characteristics, including material acoustic properties, interfacial effects, and the presence of a contacting liquid. With cells present on the device surface, all of these factors come into play, most notably that the acoustic wave only penetrates a few hundred nanometres into the contacting liquid, as shown schematically in Fig. 1. As mentioned above, this distance is on the order of 200 nm. Since aortic smooth muscle cells are 0.5–1 µm thick on the surface and are comprised mostly of liquid cytoplasm, the acoustic device is only probing the ECM, the cell membrane, and some distance into the cytoplasm. Consequently, the device does not ‘feel’ the entire cell, but instead detects changes in acoustic properties directly above the sensor surface. In fact, the increases in R_m values indicate that changes in dissipation into the contacting liquid are occurring. Since the increases are similar for both fibronectin and bare gold, this likely corresponds to changes in the viscosity and density of the liquid near the surface, in this case a change from buffer to cytoplasm. As well, upon cell adhesion, the acoustic wave must now travel through the thickness of the cell membrane.

Therefore, the frequency shift due to cell adhesion should be caused by a combination of the increase in thickness on the surface from adhesion of the cell membrane, and the change in (ρη)^1/2 as the liquid in contact with the surface changes from buffer to cytoplasm. This situation is displayed schematically in Fig. 8.

Fig. 8 Schematic representation of the acoustic response due to the adsorption of a cell to the TSM surface. In the case on the left, the bare device surface is exposed to buffer, so the entire frequency shift is due to the contacting liquid and can be modelled with eqn (3). In the case on the right, the device is now covered with a cell, and the frequency shift is a combination of the thickness of the cell membrane [eqn (2)] and the liquid effect from the cytoplasm [eqn (3)]. The shift in frequency caused by the buffer and that caused by the cytoplasm will be different, due to the differences in density (ρ) and viscosity (η).

The observed frequency shift can be modelled as in two separate portions: a surface adsorption effect and a liquid interaction effect, Δf_s ≈ Δf_a + Δf_liq.³⁵ The surface adsorption Δf_a could be estimated from eqn (1); however, since we are interested in the thickness of the layer, we can use an alternate form, where the Δm/A term is replaced by ρ_ah_a, the density-thickness product of the surface-adsorbed cell membrane:²⁵

The shift due to the liquid can be modelled with the Kanazawa and Gordon equation³⁶where ρ_liq and η_liq are the density and viscosity of the contacting liquid. Similarly, the change in resistance is given by³⁵where K² = 0.00774 is a dimensionless electromechanical coupling coefficient and C₀ ≈ 10 pF is the parallel capacitance for a 9 MHz quartz device. Note that the resistance shift in eqn (4) is determined solely from dissipative effects in the contacting liquid, and also is proportional to the liquid-induced frequency shift from eqn (3). Therefore, any change in R_m, assumed to be due to changes in the viscosity and density of the liquid, will be accompanied by a proportional change in Δf_liq. Any remaining shift in frequency is then due to adsorption onto the surface, such as the presence of the cell membrane or the deposition of ECM, and this change is estimated from eqn (2).

Using this strategy, we can perform a semi-quantitative analysis of cell binding. If we assume that all the change in R_m is due to dissipation into the contacting liquid, any remaining shift in f_s is due to a change in the location of the effective acoustic reflecting boundary. Referring to Table 1 and eqns (3) and (4), the changes in f_s and R_m from cell-adsoprtion to fibronectin are –110 Hz and 4.3 Ω. This shift in R_m corresponds to a Δf_liq of approximately –69 Hz, which is proportional to changes in the density–viscosity product of the contacting liquid. Assuming that the density of the plasma membrane is comparable to that of water,³⁷ the remaining frequency shift Δf_a = –41 Hz corresponds, from eqn (2), to a thickness increase of 2.2 nm. This is likely due to the presence of the cell membrane on the device surface. While this is an estimate, and does not account for viscoelastic properties of the membrane or interfacial effects between the membrane and the cytoplasm, it is close to the expected value for cell membrane thickness (the lipid bilayer is likely between 2 and 8 nm thick³⁸). This shows that, by measuring storage and dissipation parameters, the sensor is a suitable semi-quantitative technique for measuring cell deposition onto a surface. In addition, since 2.2 nm is a typical value for the thickness of a cell membrane, it follows that the cell likely does not deposit a significant amount of extracellular matrix above the fibronectin.

On the bare gold surface, the change in R_m due to cell adhesion was 4.0 Ω, a similar value to the fibronectin study. This indicates that the sensor probably detects similar cytoplasm-induced frequency and resistance shifts as in the fibronectin case. The 4 Ω shift corresponds to a frequency decrease of –64 Hz, so the remaining thickness-based shift is –108 Hz, corresponding to a thickness increase of 5.9 nm. This value is larger than that found for fibronectin, but still within the range of expected values for the cell membrane thickness, providing further evidence that this sensor is detecting the presence of the cytoplasm and cell membrane on surface adhesion.

As mentioned above, the difference between the calculated fibronectin and bare gold thickness changes could be due to the ECM deposited by the cells upon adhesion to the bare gold, or viscoelastic effects near the surface. However, further study is required to fully characterise the liquid properties of the cytoplasm and the thickness and density of the cell membrane on the device surface.

Detection of cell morphology changes

Fig. 5 shows the signals for additions of mono- and di-valent salts to study whether cell attachment and other morphology changes could be observed with this experimental setup. As mentioned above, integrins are membrane proteins that are believed to be responsible for cell attachment. They require divalent ions, such as Ca²⁺ and Mg²⁺, to bind to specific recognition sites on the ECM. Decreases in f_s and slight increases in R_m could be due to a variety of factors. Most likely, the shifts are due to an increase in the number of focal attachments from the increased integrin activity on the cell membrane. The change in f_s could be due either to an increase in the surface thickness from the added cell membrane attachment, or to a change in viscosity of the contacting liquid, coupled to the R_m shift, that results from more cytoplasmic contact as more focal points are formed. It is, however, not immediately apparent which of these factors is more prevalent.

Future experiments to confirm this will employ RGDS (Arg-Gly-Asp-Ser) tetrapeptides, which are believed to be a common recognition sequence, as an anchor for cell attachment to the ECM surface.²² Adding this tetrapeptide to the surface might act to detach those attached cells, which should be easily detected by the TSM.

The acoustic signals induced by 60 mM KCl (Fig. 5) and those caused by hydrogen peroxide (Fig. 6) are surprisingly similar with respect to the direction of the changes. While the signals caused by H₂O₂ are much greater than those from KCl, they are likely due to related factors. It is known that membrane depolarisation can initiate oxidant formation in the endothelial cell.³⁹ It is then possible that KCl-induced membrane depolarisation also generates small amounts of H₂O₂ or other oxidants, leading to similar chemical or morphology changes at the surface, as detected by the TSM. The oxidation caused by KCl would not generate as much hydrogen peroxide as was used in the H₂O₂ studies, so the decreases from KCl are not as large. An alternative explanation is that the disruption of the cell membrane by H₂O₂ could appear as a massive depolarisation, yielding a much larger event than that caused by KCl. Both descriptions are consistent with our observations and future study is needed to clarify the precise mechanism.

The actual acoustic mechanism leading to concurrent frequency and resistance decreases is quite complex. It is likely caused by a combination of factors, including a decrease in the number of focal attachments to the device surface, coupled with changes in the density and viscosity of the cytoplasm due to the depolarisation. Structural changes in the cell membrane, morphology of the cell membrane in contact with the device surface, and the density and viscosity of the contacting liquids could also be responsible. Inner slip has also been proposed as a mechanism that can also lead to this type of signal.⁴⁰

While it is currently impossible to determine the precise mechanism for these signals, we can infer a phenomenological description by referring to the SEM images of the cells, before and after addition of hydrogen peroxide. Before the addition of H₂O₂, the cells appear flat and healthy on the gold surface (Fig. 7a). After the addition of H₂O₂ (Fig. 7b and 7c), the cell surface has been disrupted, and the cells have swelled and become detached slightly from the surface. This could result in numerous effects, of which two are likely to be important. First, the cell medium can penetrate closer to the substrate surface and, since the medium likely has a lower viscosity than the cytoplasm, f_s would increase and R_m would decrease, both proportionally to (ρη)^1/2 [eqns (3) and (4)]. This is a main contributor to the decrease in R_m. Second, it appears from Fig. 7b and 7c that the surface has swelled and become thicker and rougher. This essentially results in a longer average effective wavelength for the acoustic wave, corresponding to a decrease in frequency. These two effects appear to occur concurrently, and the frequency increase due to the change in viscosity is offset by the increase in the effective thickness of the surface film. Another possible contribution to the frequency decrease may arise from the thickening of the cell membrane due to the swelling of the cells.

When the surface is washed with IMDM medium (Fig. 6), R_m returns to baseline, while f_s increases to a value above baseline. It is likely that most cells return to their original configuration once the cell-growth medium was restored. However, some cells likely died and detached on exposure to H₂O₂, resulting in the increase in frequency due to the loss in cell mass at the surface. Further acoustic and SEM studies are required to validate this result.

Conclusions

The transverse-shear mode acoustic wave biosensor has been used to measure the adsorption of rat aortic smooth muscles cells, and their subsequent response to various chemical stimuli. This work has demonstrated that this acoustic biosensor can detect cell deposition, morphological changes due to ion effects, and membrane depolarisation and denaturation.

Along with the bare gold surface, two common extracellular matrixproteins , laminin and fibronectin, were tested as possible surface modifications. Cells were easily washed from the laminin surface, indicating poor binding, but both fibronectin and the bare gold surface performed well. However, further study is required to confirm the results and elucidate the underlying mechanisms.

The Sauerbrey equation alone is unable to describe cell binding, and in our opinion is not a suitable model for studying complex biological surfaces. Instead, measurements of separate storage and dissipation parameters (in our case f_s and R_m) are required to fully characterise the acoustic system. By employing both measurements, we have constructed a semi-quantitative model of adsorption behaviour, based on corroboration with SEM images, that explains the observed shifts in terms of cell behaviour. As opposed to using the TSM device as a pure mass detector, this device has the capability to detect and measure a wide variety of cell functions in an on-line format.

At present, cell behaviour on surfaces remains a significant problem for study and, as such, we can only present a partial description of the complex mechanisms that contribute to the observed acoustic signals. Full characterisation of the sensor will require more complete models and predictors of cell structure under given perturbations, both mechanical and chemical, to be able to link the observed signals to cellular properties. However, with improved models and a better understanding of the complex behaviour of cells on surfaces, in the presence of high frequency acoustic fields, we can use this device as an on-line, label-free measurement technique to test cell deposition and metabolism throughout its lifecycle, including morphological changes, cell death, and interactions with drugs and other small molecules.

Real-time information on cell behaviour, including adsorption, kinetics, and morphological changes, is of great interest to molecular biologists and in drug development. Acoustic biosensing, in conjunction with imaging techniques such as SEM , promises to become a useful bio-analytical technique for the detection of morphological and metabolic changes in cells in response to a wide range of stimuli.

Acknowledgements

The authors are grateful to the Natural Sciences and Engineering Council of Canada for support of this work. This research was also supported by a grant from the Canadian Institutes for Health Research (MOP 62698) to R.-K. L. and by an OGSST grant to J. E. In addition, the authors would like to thank Doug Holmyard (Mount Sinai Hospital, Toronto) and Neil Coombs (University of Toronto) for assistance with SEM measurements.

References

1. K. A. Marx, Biomacromolecules, 2003, 4, 1099–1120.
2. B. A. Cavic, G. L. Hayward and M. Thompson, Analyst, 1999, 124, 1405–1420.
3. L. E. Cheran and M. Thompson, Chem. Soc. Rev., 2007 , submitted.
4. J. J. Ramsden, S. Y. Li, J. E. Prenosil and E. Heinzle, Biotechnol. Bioeng., 1994, 43, 939–945.
5. X. M. Wang, J. S. Ellis, E. L. Lyle, P. Sundaram and M. Thompson, Mol. BioSyst., 2006, 2, 184–192.
6. X. D. Su, Y. J. Wu and W. Knoll, Biosens. Bioelectron., 2005, 21, 719–726.
7. J. Zhang, X. D. Su and S. J. O'Shea, Biophys. Chem., 2002, 99, 31–41.
8. S. J. Braunhut, D. McIntosh, E. Vorotnikova, T. Zhou and K. A. Marx, Assay Drug Dev. Technol., 2005, 3, 77–88.
9. K. A. Marx, T. A. Zhou, A. Montrone, H. Schulze and S. J. Braunhut, Biosens. Bioelectron., 2001, 16, 773–782.
10. S. Rozhok and R. Holz, Talanta, 2005, 67, 538–542.
11. X. E. Jia, Z. W. Zhang, L. Tan, Y. Y. Zhang, Q. J. Xie, Z. M. He and S. Z. Yao, Chin. Chem. Lett., 2006, 17, 509–512.
12. S. J. Braunhut, K. A. Marx, T. Zhou, M. C. Warren and H. M. Schulze, Mol. Biol. Cell, 1999, 10, 64A–64A.
13. K. A. Marx, T. Zhou, A. Montrone, D. McIntosh and S. J. Braunhut, Anal. Biochem., 2005, 343, 23–34.
14. M. Pax, J. Rieger, R. H. Eibl, C. Thielemann and D. Johannsmann, Analyst, 2005, 130, 1474–1477.
15. A. Alessandrini, M. A. Croce, R. Tiozzo and P. Facci, Appl. Phys. Lett., 2006, 88, 083905.
16. B. A. Cavic, J. Freedman, Z. Morel, M. Mody, M. L. Rand, D. C. Stone and M. Thompson, Analyst, 2001, 126, 342–348.
17. L. E. Cheran, S. L. Cheung, A. Al Chawaf, J. S. Ellis, D. D. Belsham, W. A. MacKay, D. Lovejoy and M. Thompson, Analyst, 2007, 132, 242–255.
18. P. C. H. Li, W. J. Wang and M. Parameswaran, Analyst, 2003, 128, 225–231.
19. Guide for the Care and Use of Laboratory Animals, Institute of Laboratory Animal Resources, US Department of Health and Human Services, Public Health Service, National Academy Press, Washington, D.C., 1996.
20. T. Yasuda, R. D. Weisel, C. Kiani, D. A. G. Mickle, M. Maganti and R. K. Li, J. Thorac. Cardiovasc. Surg., 2005, 129, 904–911.
21. I. Sasaki, H. Tsuchiya, M. Nishioka, M. Sadakata and T. Okubo, Sens. Actuators, B, 2002, 86, 26–33.
22. J. Wegener, C. R. Keese and I. Giaever, Exp. Cell Res., 2000, 259, 158–166.
23. J. S. Ellis and M. Thompson, Phys. Chem. Chem. Phys., 2004, 6, 4928–4938.
24. E. L. Lyle, G. L. Hayward and M. Thompson, Analyst, 2002, 127, 1596–1600.
25. G. McHale, R. Lucklum, M. I. Newton and J. A. Cowen, J. Appl. Phys., 2000, 88, 7304–7312.
26. N. G. Tassew, Ph.D. Thesis, University of Toronto, Canada, 2003.
27. S. Ghafouri and M. Thompson, Electroanalysis, 2000, 12, 326–336.
28. G. Sauerbrey, Z. Phys., 1959, 155, 206–222.
29. X. C. Zhou, L. Q. Huang and S. F. Y. Li, Biosens. Bioelectron., 2001, 16, 85–95.
30. T. Hianik, X. Wang, S. Andreev, N. Dolinnaya, T. Oretskaya and M. Thompson, Analyst, 2006, 131, 1161–1166.
31. L. M. Furtado, H. B. Su, M. Thompson, D. P. Mack and G. L. Hayward, Anal. Chem., 1999, 71, 1167–1175.
32. N. Tassew and M. Thompson, Anal. Chem., 2002, 74, 5313–5320.
33. H. J. Yardley and D. J. Goldstein, Br. J. Dermatol., 1976, 95, 621–626.
34. I. S. Park and D. I. Kim, Biotechnol. Tech., 1993, 7, 627–630.
35. S. J. Martin, V. E. Granstaff and G. C. Frye, Anal. Chem., 1991, 63, 2272–2281.
36. K. K. Kanazawa and J. G. Gordon, Anal. Chim. Acta, 1985, 175, 99–105.
37. M. S. Kang, J. Auyoung and E. Cabib, J. Biol. Chem., 1985, 260, 2680–2684.
38. B. Alberts, A. Johnson, J. Lewis, M. Raff, K. Roberts and P. Walter, Molecular biology of the cell, Garland Science, New York, 4th edn, 2002.
39. A. B. AlMehdi, H. Ischiropoulos and A. B. Fisher, J. Cell. Physiol., 1996, 166, 274–280.
40. J. S. Ellis and M. Thompson, Chem. Commun., 2004, 1310–1311.

---