Biotechnol. hog. 1993, 9, 105-108
105
Real-Time Measurement of Anchorage-Dependent Cell Adhesion Using a Quartz Crystal Microbalance David M. Gryte, Michael D. Ward,*and Wei-Shou Hu' Department of Chemical Engineering and Materials Science, Amundson Hall, 421 Washington Avenue, SE, University of Minnesota, Minneapolis, Minnesota 55455
An in situ technique based on the mass-sensitive piezoelectric quartz crystal microbalance (QCM) was applied t o continuously monitor the attachment and detachment of anchorage-dependent mammalian cells on a metal surface. Specifically, we have demonstrated t h a t the attachment of Vero cells t o metal surfaces on the piezoelectrically active area of the QCM results in decreases in the QCM resonant frequency t h a t can be monitored conveniently in real time. Lysis and detachment of Vero cells caused by vesicular stomatitis virus (VSV) infection can also be monitored readily. Our results demonstrate t h a t the QCM is a viable technique for monitoring anchorage-dependent cell attachment and detachment on surfaces (caused by stimulatory, inhibitory, or cytopathic effects).
Introduction Attachment, growth, and detachment of anchoragedependent mammalian cells on a surface often reflect the stimulatory, inhibitory, or cytopathic effects of some culture condition. Notable examples include the effect of surface properties or cell adhesion molecules on cell adhesion to a surface (Massia and Hubbel, 1990) and the cytopathiceffect caused by virus infection or by a biological or chemicalagent (Havelland Vilcek, 1972). Measurement of adhesion typically is performed by enumerating the cells that are attached or that are in suspension during the time course of the experiment (Himes and Hu, 1986) or by microscopicobservation of cell attachment to a surface. Cytopathic effects generally are evaluated by direct microscopic observation or chemical assay following the introduction of a cytopathic agent to adherent cells. These methods, however, are often tedious, and the interpretation of results is subjective. The need for new approaches for measuring cell adhesion prompted us to examine the feasibility of measuring cell adhesion with a mass-sensing quartz crystal microbalance (QCM) (Ward and Buttry, 1990). The QCM comprises an AT-cut piezoelectric quartz crystal sandwiched between two metal electrodes. Application of an alternating voltage potential across the quartz crystal by the two excitation electrodes on opposite sides of the crystal causes the crystal to oscillate a t a characteristic resonant frequency. If one assumes rigid layer behavior, the resonant QCM frequency depends upon the mass attached to the quartz crystal surface according to the Sauerbrey relationship
where Af is the measured frequency shift, fo is the initial resonant frequency of the quartz crystal, Am is the effective mass change, A is the piezoelectrically active area defined by the two gold electrodes, ps is the density of quartz (2.648 g cm-9, and ps is the shear modulus (2.947 X 10" dyn cm-2) (Sauerbrey, 1959). Recent advances allow QCM measurements to be performed directly in liquid media
* Authors to whom correspondence should be addressed. 8756-7938/93/3009-0 105$04.00/0
that is in contact with one side of the QCM. This had led to the demonstration of several QCM-basedimmunological and microbiological methods, including antibody-based immunoassay by direct or sandwich methods (Muramatsu et al., 1987;Ebersole and Ward, 19881,detection of nucleic acids (Ebersole e t al., 19901, glucose detection via hexakinase binding (Lasky and Buttry, 1989), real-time measurement of cell metabolism and division rates (Ebersole et al., 19911, and immunological detection of microbes (Muramatsu e t al., 1986; Muramatsu et al., 1989). Direct measurements of antibody-antigen binding using other piezoelectric transducers has also been reported (Roederer and Baastians, 1983). While QCM methods are generally configured and data interpreted on the basis of mass measurements, in liquid media the resonant frequency also depends upon the viscosity and density of the fluid in contact with the QCM (Kanazawa and Gordon, 1985a; Kanazawa and Gordon, 1985b;Glassford, 1978). The basis for this behavior is the effect of viscosity and density on the propagation of the shear wave that radiates from the resonator into the fluid. The shear wave propagation into the fluid can be described by
k = (1FfoPLltJ'2 (2) where k is the propagation constant of the shear wave, VL is the viscosity of the fluid, and p~ is the density of the fluid. The decay length of the shear wave, which is equal to llk, is approximately 250 nm in water a t room temperature. Accordingly, the viscosity and density of the hydrodynamic layer must be taken into consideration when QCM measurements are performed in liquid media. These factors may be of considerable importance for the attachment of cells, which generally will have dimensions exceeding the decay length. We present herein the measurement of cell adhesion to a metal surface using QCM methods. Specifically,we have demonstrated that QCM is a convenient method for measuring in real time the attachment of African Green Monkey kidney (Vero) cells as well as their detachment that accompanies cell death caused by either NaOH addition or virus infection. Additionally, the observed QCM responses are consistent with attached cell layers
0 1993 American Chemical Society and American Institute of Chemical Engineers
Biotechnol. Frog., 1993, Vol. 9, No. 1
106
Figure 1. Schematic diagram of apparatus. A 5-MHz AT-cut crystal wafer was inserted into the feedbackloop of a broadband radio frequency amplifier via two gold electrodes which were evaporativelydepositedonto the facesof the wafer. A glass O-ring joint clampedto the upper face of the quartz crystal wafer served as the experimental chamber.
that behave as viscous films rather than rigid masses, illustrating that care must be exercised in the interpretation of QCM measurements involving cell adhesion or binding events.
Materials and Methods Apparatus. Piezoelectric measurements of cell adhesion were performed with a 1in. diameter, 5-MHz planoplano AT-cut quartz crystal (McCoyElectronics,Mt. Holly Springs, PA). Gold electrodes (2000 A thick) were deposited by electron-beam evaporation onto similarly prepared titanium underlayers (200 A), whose role is to provide better adhesion of the gold electrodesto the quartz crystal. The electrode patterns comprised a circularregion in the center of the crystal with two tabs leading to the crystal edge for electrical connection. An asymmetric electrode format was used with the larger gold electrode (A = 0.35 cm2) facing solution and the smaller electrode (A = 0.22 cm2)facing air. The quartz crystal was mounted between two PTFE-coated O-rings confined by standard glass joints and a metal clamp to insure a tight seal. The upper glass joint, with an inside diameter of 1.0 cm, was used as the cell culture chamber (Figure 1). A glass coverslipwas placed over the upper opening of the chamber to prevent evaporation and gas exchange with ambient air. The interior of the glass chamber was siliconized with Prosil-28 (PCR Inc., Gainesville, FL) to prevent cell attachment to the glass surface. All parts of the apparatus that were in contact with the cell culture were steamsterilized a t 121 "C for 20 min. The apparatus was placed in an incubation chamber at 37 "C. The quartz crystal was inserted into the feedback loop of a broadband radio frequency amplifier in a homemade oscillator. A Hewlett-Packard 6234A dual output power supply was used as the voltage source for the oscillator. The frequencywas monitored by a Hewlett-Packard 5384A frequency counter and was recorded on a Macintosh IIcx computer. Cells and Cell Culture. The cells used were African Green Monkey kidney (Vero) cells, originally obtained from American Type Culture Collections (Rockville,MD). The Vero cells were maintained in 25 cm2 polystyrene tissue culture flasks (Corning Glass Works, Corning, NY). The cell culture medium used was Dulbecco's modified Eagle's (DME)medium (GibcoLaboratories,Grand Island, NY) without sodium bicarbonate. The medium was supplemented with 10% (v/v) fetal bovine serum (Hyclone Laboratories,Logan, UT) and 31 mMN-(2-hydroxyethyl)piperazine-N'-2-ethanesulfonicacid (Hepes)(SigmaChemical Co., St. Louis, MO). The pH was adjusted to 7.3.
Prior to inoculation into the experimental apparatus, the Vero cells were detached from the tissue flask by treatment with 0.25 % (w/v) trypsin solution with 0.02 % ethylenediaminetetraacetic acid (EDTA) in phosphatebuffered saline (PBS). Viable cellswere enumerated using a hemacytometer (American Optical, Buffalo, NY) after staining a sample with 0.1 % (w/v) trypan blue (Sigma) in PBS. At the conclusion of the experiments, cells on the piezoelectrically active surface of the QCM were stained with fluorescein diacetate (Sigma) and ethidium bromide (Sigma) as described previously (Nikolai et al., 1991)and examined with a Zeiss epifluorescent microscope (Carl Zeiss, Inc., Thomwood, NY) using a standard FITC filter set. In the stained sample, the cytoplasm of viable cells stained green, and the nuclei of dead cells stained orange. Latex beads were used to replace the cells as a control to observe the QCM resonant frequency change due to mass settling on the piezoelectricallyactive surface. The beads were obtained from Duke Scientific (Palo Alto, CA). The beads were polystyrene microspheres with a mean diameter of 10.1 pm and a density of 1.05 g ~ m - ~ . Virus Infection. The virus used was vesicular stomatitis virus (VSV, Indiana strain), obtained from Dr. w. Thomas Shier of the Department of Medicinal Chemistry at the University of Minnesota. Prior to infection, Vero cells were inoculated into the QCM cell culture chamber and allowed to attach to the piezoelectricallyactive surface. To infect cells, the medium was withdrawn by suction, and a 60-pL suspension of virus medium was added. The multiplicity of infection (MOI) used was 5. The cells were incubated a t 37 "C for 1h. Afterward, 0.6 mL of DME was added, and the apparatus was returned to the 37 "C incubation chamber.
Results Cell Attachment. We initially investigated the attachment of African Green Monkey cells added to a medium in which one side of the QCM was submerged. After addition of 0.3 mL of media to the QCM cell culture chamber, the resonant frequency and temperature were allowed to stabilize for 1h. The glass coverslip was then removed from the opening of the QCM culture chamber and 0.5 mL of medium containing 4 X lo5 cells added. This corresponds to a surface coverage of 5 X lo5cell cm-2 on the piezoelectrically active surface of the QCM. The QCM cell culture chamber was then sealed with the glass coverslip,and the QCM resonant frequency was monitored and recorded every 90 s. Figure 2a depicts the change of resonant frequency after the addition of the cells. The instantaneous spike that occurred at inoculation was due to the disturbance of the system by the sample addition. The long-term decrease in resonant frequency that occurs after the spike reflects the attachment of cells to the gold electrode of the QCM. In this particular experiment, the overall frequencychange was -280 Hz. Microscopic examination or sampling of culture fluid during the course of the experiment was not possible as it would cause severe perturbation. At the conclusion of this experiment, it was confirmed microscopically that the piezoelectrically active surface was completely covered with cells. In a separate experiment, cells were inoculated under the same conditions to the gold electrode surface, and small aliquots of samples were taken periodically for measurement of unattached cells. The results are shown in Figure 2b. The frequency change was accompanied by a decrease in the unattached cells in the suspension. Parallel microscopicobservation revealed that cells had begun to spread on the surface 1 h after inoculation.
Biotechnol. Prog... 1993, Vol. 9, No. 1
107
(a) 50
,-
0
. - -Beads- - - /
-50
-100
N
E
-500 -200
c
-
-600 -
\
Inoculation
NaOH Addition I
-250
-
-0.2
Inoculation 0
0.2
0.4
0.6
Time (hr)
0.8
1
1.2
I
I
I
5
= I t
..
.-C
Inoculation
Vero cells, after trypsinization and detachment, are spherical. After settling on a surface, they develop attachment plaques and attach to the surface. It was not clear whether the settlement of cells on the piezoelectrically active surface, without the development of firm attachment, causes a change in the resonant frequency. When latex beads of diameter and density similar to the Vero cells described above were inoculated, very little change in resonant frequency was observed (Figure 2) compared to that observed during Vero cell attachment. The results of this experiment indicate that only the cell mass that actually attaches to the piezoelectrically active surface is detected. Cell Detachment. We have also investigated using a QCM to detect cell detachment from the piezoelectrically active surface. Loss of viability in Vero cells is often accompanied by their detachment from a surface, which can be induced by a lethal dose of NaOH. As in the previous experiments, 0.3 mL of cell media was initially added to the cell culture chamber and the resonant frequency was allowed to stabilize. After the resonant frequency remained stable for 1 h, 4 X lo5 mL of DME were inoculated into the cell culture chamber. The time required for completion of cell attachment, as surmised from the cessation of the frequency decrease, was similar to that of the experiment in Figure 2. The overall frequency change of -460 Hz, however, was somewhat larger, although the experimental conditions were identical. The variability in Af values was not unusual, generally ranging between these two values. While we do not understand the source of this variability, it is likely
I
1
I
I
I
Blotechnol. Prog., 1993, Vol. 9, No. 1
108
k30°
600
-i? 400 E
200
for continuous measurements, thereby providing kinetic information of cell response to these reagents. The use of multiple-well culture plates containing a piezoelectric resonator in each well can provide for rapid, convenient, and simultaneous measurements of cell response to different reagents. When combined with automated pattern recognition methods, such a configuration could provide rapid diagnosis of pathogens and screening of biologically active molecules.
i t
t1
Acknowledgment 4
6
a
10
Time after VSV Infec!ion (hr)
12
Figure 4. Observed resonant frequency changeof a 5-MHz AT-
cut quartz resonator due to cell lysis caused by infection of a Vero cell layer attached to the QCM with vesicular stomatitis virus at an MOI of 5. Monitoring of the resonant frequency than< ? began 4 h after VSV infection.
on the QCM electrode surface (Muramatsu et al., 1986; Muramatsu et al., 1989), in which the frequency changes were interpreted on the basis of a rigid layer using eq 1. We have observed some variability in the overall frequency changes during attachment of Vero cells, with Af values ranging from -280 to -460 Hz. While the source of this behavior remains unidentified, we tentatively attribute it to variability in cell culture media and the metal surface that may affect the mode of attachment of the cells to the gold electrode surface of the QCM. The frequency increases observed during detachment induced by cell lysis are consistent with an effective decrease in mass on the resonator surface. Notably, the data clearly indicate that the time course of cell attachment can be readily followed with the QCM. In order for these frequency changes to be interpreted on the basis of eq 1, the foreign layer must be assumed to behave as a rigid elastic mass. That is, the acoustic impedances of the crystal and the layer are assumed to be and identical. If we assume a cell density of 1.03 g a mean cell volume of 4000 pm3 (Griffiths and Riley, 1985), the calculated amount of mass in a cell monolayer on the 0.22 cm2piezoelectrically active surface was 450 pg. This mass corresponds to a frequency change during attachment of -116 000 Hz according to eq 1, substantially different from the observed value of -280 Hz in Figure 2 or -460 Hz in Figure 3. At the end of the experiment, microscopic observation confiimed coverage of the gold electrodes with cells. Cell adhesion to the siliconized glass surfaces of the culture chamber was negligible. The apparent discrepancy between the expected rigid mass value and the observed value therefore is not attributed to incomplete cell attachment. The observed frequency shifts therefore suggest that the mammalian cells do not behave as rigid fibs. Under these conditions the frequency change during cell attachment may be dictated by the viscosity of the attached celllayer and the accompanying changes in shear wave propagation according to eq 2. The simple technique described here allows one to continuously monitor attachment and detachment of anchorage-dependent mammalian cells on metal surfaces. The observed responses of piezoelectric quartz crystal microbalance suggest that the attached cell layer does not behave as a rigid mass, but rather approaches the behavior expected for a viscous fluid. The relative contribution of simple cell adhesion and subsequent spreading on the frequency is not clear and needs further study. Nevertheless, the simplicity of the QCM method makes it an attractive potential biological sensor for detection of cell mass changes on a surface caused by stimulatory, inhibitory, and cytopathic reagents. This approach can allow
This work was supported, in part, by the National Science Foundation through grants to W.S.H. (BSCS8552670 and M.D.W. (CTS-9111000). We thank MadhuSudan V. Peshwa for preliminary work.
Literature Cited Ebersole, R. C.; Ward, M. D. Amplified Mass Immunosorbent Assay with aQuartz Crystal Microbalance. J.Am. Chem. SOC. 1988,110,8623.
Ebersole, R. C.; Miller, J. A.; Moran, J. R.; Ward, M. D. SpontaneouslyFormedFunctionallyActive Avidin Monolayers on Metal Surfaces: A Strategy for Immobilizing Biological Reagentsand Design of PiezoelectricBiosensors. J.Am. Chem. SOC. 1990,112, 3239.
Ebersole,R. C.; Foss, R. P.;Ward,M.D. PiezoelectricCellGrowth Sensor. BiolTechnology 1991,9, 450-454. Glassford, A. P. M. Response of a Quartz Crystal Microbalance to a Liquid Deposit. J. Vac. Sci. Technol. 1978,15,1836. Griffiths, J. B.; Riley, P. A. Cell Biology: Basic Concepts. In Animal Cell Biotechnology;Spier,R. E.; Griffiths, J. B., Eds.; Academic Press: New York, 1986;Vol. I, pp 17-48. Havell, E. A.; Vilcek, J. Production of High Titered Interferon in Cultures of Human Diploid Cells. Antimicrob. Agents Chemother. 1972,2,476-484. Himes, V. B.; Hu, W. S. Attachment and Growth of Mammalian Cellson Microcarrierswith Different Ion Exchange Capacities. Biotechnol. Bioeng. 1986,29, 1155-1163. Kanazawa, K. K.; Gordon, J. G., I1 Frequency of a Quartz Microbalance in Contact with Liquid. Anal. Chem. 1986a, 57, 1771-1772.
Kanazawa, K. K.; Gordon, J. G., I1 The Oscillation Frequency of a Quartz Resonator in Contact with a Liquid. Anal. Chim. Acta 1986b,175,99. Lasky, S. J.; Buttry, D. A. Sensors Based on Biomolecules Immobilizedon the PiezoelectricQuartz CrystalMicrobalance. Detectionof GlucoseUsing Hexokinase. ACSSymp. Ser. 1989, 403, 237.
Massia, S. P.; Hubbell, J. A. Covalent Surface Immobilizationof Arg-Gly-Asp- and Tyr-Ile-Gly-Ser-Arg-Containing Peptides to Obtain Well-Defined Cell-Adhesive Substrates. A d . Biochem. 1990,187,292-301. Muramatsu,H.; Kajiwara,K.; Tamiya,E.; Karube,I. Piezoelectric Immuno Sensor for the Detection of Candida albicans Microbes. Anal. Chim. Acta 1986,188,257-261. Muramatsu, H.; Dicks,J. M.; Tamiya,E.; Karube, I. Piezoelectric Crystal Biosensor Modifiedwith Protein Afor Determination of Immunoglobulins. Anal. Chem. 1987,59,2760. Muramatsu, H.; Watanabe, Y.; Hikuma, M.; Ataka, T.; Kubo, I.; Tamiya, E.; Karube, I. PiezoelectricCrystal Biosensor System for Detection of Escherichia Coli. Anal. Lett. 1989,422,2155. Nikolai,T. J.;Peshwa, M. V.; Goetghebeur,S.;Hu, W. S. Improved MicroscopicObservationof MammalianCellson Microcarriers by Fluorescent Staining. Cytotechnology 1991,5,141-146. Roederer, J. E.; Baastians, G. J. MicrogravimetricImmunoassay with Piezoelectric Crystals. Anal. Chem. 1983,55,2333. Sauerbrey, G. Z. The Use of Quartz Crystal Oscillators for Weighing Thin Layers and for Microweighing Applications. 2.Phys. 1969,155, 206. Ward, M. D.; Buttry, D. A. In Situ Interfacial Mass Detection with PiezoelectricTransducers. Science 1990,249,1000-1007. Accepted October 13, 1992.
