Induction of cell processes by local force

Many cell phenomena are accompanied by morphological changes in the cells, including global (rounding of the cells during cell division and spreading) and local changes (formation of cell outgrowths) (Vasiliev, 1985; Trinkaus, 1985). It has been established that in all cases the cytoskeleton plays a key role in changing cell shape. However, the mechanical properties of the cell membrane and of the cortical layer might contribute to the cell deformability.

To investigate the mechanical properties of cell membranes several experimental methods have been developed (Evans and Skalak, 1979; Pasternak and Elson, 1985; Bray et al. 1986; Bo and Waugh, 1989). Using these methods, the mechanical properties of liposomes, sea urchin eggs, erythrocytes and lymphocytes have been investigated, but they are applicable mainly to suspended cells; the mechanical properties are expressed in terms of elastisity and viscosity, and are thought to be homogeneous over the cell surface. However, it is important to understand whether different areas of the cells with complex morphology differ locally in their mechanical properties. The existing methods are inapplicable to cells with complex cell surface structure (e.g. spread fibroblasts).

Previously (Popov and Margolis, 1988) we described a new experimental system based on the phenomenon of dielectrophoresis for applying a force directly to the plasma membrane of suspended cells. The force was sufficient to generate cell-specific membrane protrusions of various types. Inhibitors of cytoskeletal activity were not able to prevent generation of processes by an external force.

In this study we investigated the potential for different regions of mouse embryo fibroblasts (MEF) to form processes under the action of a membrane-applied force. The force, applied locally to the plasma membrane, was generated during the treatment of cells with an a.c. electric field.

Primary mouse embryo fibroblasts (MEF) were prepared according to the method of Domnina et al. (1982). Cells were maintained in Dulbecco’s modified Eagle’s medium containing 10% foetal calf serum and used after 1–3 passages.

After trypsinization, 100 μl of cell suspension containing 10³–10⁴ cells was transferred to the surface of a 25 mm×25 mm coverslip. The cells were incubated at 37 °C in a humidified atmosphere.

Cytochalasin D (Sigma) dissolved in dimethylsulphoxide (l mg ml⁻¹) was added to MEF immediately before incubation on the coverslip. The final concentration of cytochalasin D was l μg ml⁻¹. Carbonyl-m-chlorphenylhydrazon (Sigma) was dissolved in dimethylsulphoxide (l mg ml⁻¹); sodium azide was dissolved in water (200 mg ml⁻¹). The final concentration of these drugs in the medium were 1 μg ml⁻¹ and 2 mg ml^-1, respectively. In both cases 20 mM 2-deoxy-n-glucose (Sigma) was added to the medium. Carbonyl-m-chlorphenylhydrazon and sodium azide were added to MEF after a lh incubation of the cells on a coverslip at 37 °C.

Immediately before electrical field treatment, cells were washed three to four times with buffered sucrose (290 mM sucrose, 1 mM Hepes–NaOH, pH 7.4) and placed on the stage of inverted ‘Fluovert’ microscope (Leitz). Tungsten microneedles with 2–5 μm diameter tips were used as electrodes. Electrodes were placed by using Narishigi MO 303 micromanipulators 10–20 μm from the cell surface at an angle of 10° to the horizontal (Fig. 1). An a.c. electrical field was applied (voltage 7–8 V, frequency 1MHz). Cells were treated by the field no later than 10 min after replacement of the medium with buffered sucrose.

Previously it was shown that during the treatment of cells with an a.c. electrical field in a low-conducting sucrose solution a force is generated; it is applied to the plasma membrane and to a thin (10–20 Å; 1Å=0.1 nm) cytoplasmic layer and is directed outwards (Popov and Margolis, 1988).

Force applied to the surface of MEF was used to investigate: (1) cell process formation during cell spreading; (2) the role of the cytoskeleton in this phenomenon; and (3) the potential of morphologically different parts of the cell surface to generate processes.

After a 1 h incubation on the surface of a coverslip, cells adhered to the surface and spread, acquiring the characteristic shape with a circular lamella. A few seconds after application of the electrical field, processes were generated on that part of the membrane closest to the electrode (Fig. 2). Processes were cylindrical (∼0.5 μm in diameter), with a thickening at the distal edge of the protrusion (∼1 μm in diameter). The processes were directed towards the tip of the electrode and they reached it in approximately 30 s. The number of processes varied from 1 to 20 for different cells. When the electrodes were positioned so that they were not equidistant from the cell the processes were generated mainly at the membrane surface facing the nearest electrode.

Control cells incubated in cytochalasin D-containing medium and not treated with an electrical field attached to the coverslip after a 1 h incubation, However, practically no lamellae were observed. The processes were ∼1 μm in diameter, which was approximately twice that observed during spreading in cytochalasin-free medium.

Application of an electrical field to cytochalasin-treated cells resulted in the formation of processes directed towards the tip of the electrode like those formed by untreated cells in an electrical field (Fig. 3). The diameters of the processes did not exceed 0.5 μm.

Cells treated with sodium azide+2-deoxy-D-glucose for Ih also form processes after electrical field application (Fig. 4). Again, the processes are formed only at the electrode-facing part of the cell and reach the electrodes in a few dozen seconds. The diameter of the processes is ∼0.5 μm. Similar results were obtained when the time of incubation was increased to 3h and when another inhibitor of ATP synthesis, carbonyl-m-chlorphenylhydra-zon, was added instead of sodium azide.

During the early stages of spreading in normal physiological conditions the surface of some MEF were covered with microblebs (located mainly at the lamella). These cells easily formed processes (usually 10–20/cell) after electrical field treatment (Fig. 5) that quickly reached the electrode.

After a few hours of incubation on the surface of a coverslip cells polarized with the formation of active and stable edges. By placing the electrodes close to certain parts of the membrane, we were able to apply the force locally to the lamella region of the cell and to the stable edge. Under electrical field treatment the processes were formed only at the active edge. No processes were formed at the stable edge (Figs 6, 7).

Some cells in normal physiological conditions have processes ∼0.5 μm in diameter. Usually they are Located at the lamella region of the cell. These pre-existing processes easily elongated after electrical field application (Fig. 8).

Their diameter (∼0.5 μm) did not change after electrical field application.

Migrating fibroblasts in normal physiological conditions as a rule have retraction fibers at the rear of the cell, morphologically similar to the processes at the leading edge. Retraction fibers could not be elongated by electrical field treatment.

Previously we developed an experimental system for applying force to spherical cells brought in contact with a planar electrode (Popov and Margolis, 1988). We have found that the surface of the cells is deformed due to the membrane-applied force that is generated during electrical field treatment. In this paper the method is modified to investigate the mechanical potential of different parts of cells with surfaces of complex morphology.

The principle of force generation is based on the different conductivities of the cytoplasm and the extracellular solution (Zimmermann, 1982; Pohl, 1978; Engelhardt et al. 1984). After application of the electrical field an undermembrane diffuse layer 10–20 ∼ (1∼=0.1nm) thick is formed to compensate the field inside the cell (Landau and Lifshits, 1960). This phenomenon is equivalent to the polarization of a dielectric in a vacuum in an electrical field. A force applied to the surface of the dielectric and directed outwards is generated.

In our experiments the medium outside is conducting the current and a diffuse layer on the outer surface of the membrane tends to form. This outer layer would balance the inner diffuse layer and no force would be generated. But owing to the low conductivity of the extracellular solution the time of formation of the outer diffuse layer (10⁻⁴s) is much longer than the period of the applied field (1 0⁻⁶ s). So an outer diffuse layer does not form and a force applied to the thin membrane layer and directed outwards is generated. For more details see Margolis and Popov (1988) and Pastushenko et al. (1985).

The method presented in this article permits application of a local force to different parts of the membrane and investigation of morphological changes in cells, spread on a glass surface, induced by these forces.

After electrical field application processes were formed at the electrode-facing parts of the cells. Light microscopic data demonstrated that these processes were morphologically normal (∼0.5 μm in diameter, uniform in thickness with a thickening at the end of the process).

Previously (Popov and Margolis, 1988) we have shown that it is a membrane-applied force that generates the processes. Other possible effects of an electrical field (heating, electrical breakdown, electrophoresis etc.) do not contribute to the observed phenomenon.

In normal physiological conditions cell process formation is substantially inhibited by cytochalasin D (inhibitor of polymerization of G-actin into microfilaments) and by metabolic poisons. In the experimental conditions used in this work these substances did not prevent protrusion formation in the electrical field. Therefore, membrane-applied force per se is sufficient to generate processes with normal morphology (see also Popov and Margolis, 1988).

The method of application of local force to the cell surface used in this work permits investigation of the potential of different regions of the cell surface to form processes under the action of this force.

MEF were studied at the stage of radial spreading and after cell polarization was completed. At the stage of radial spreading MEF formed symmetrical lamella. Cell processes at this stage could be formed at any region of the lamella.

The surface of parts of the cells at the stage of radial spreading was covered with microblebs. These cells were very ‘deformable’: under the action of membrane-applied force numerous cell processes were formed, reaching the tip of the electrode within a few seconds. This result is in agreement with the previous suggestion about the higher ‘deformability’ of blebing cells (Trinkaus, 1985).

During polarization, the cell surface is divided into active and stable domains (Vasiliev, 1985), with processes forming only at the active region. Vasiliev et al. (1970) showed that the treatment of cells with colcemide or other microtubule-disrupting agents leads to the loss of cell polarization: all of the cell edge becomes active. It is proposed that the polarized form of the cell is determined by preferential orientation of microtubules along the stable edge (Goldman, 1971). Material transported along the microtubules is inserted at the active edge (Hollenbeck, 1989). It is here that the protrusive activity of the cell surface is localized. This model, however, does not explain the independence of polarization of fibroblasts in early primary cultures on microtubule integrity (Middleton et al. 1989).

The results of this paper suggest that, at least in part, the absence of protrusive activity at the stable edge can be explained by its ‘rigidity’: membrane-applied force does not generate processes. This rigidity is most probably determined by the structure of the cell membrane and/or by the underlying cortical layer (Zand and Albrecht-Buehler, 1989).

The experimental system used in this study provides information about the organization of the cortical layer at different regions of the cell surface. Protrusions at the active edge and retraction fibers react differently to the applied force. The processes at the active edge are easily elongated while retraction fibers are not.

The poorly understood internal organization of the blebs also can be studied by the method described here. The commonly held point of view about the absence of a cortical layer beneath the membrane of the bleb is based on three observations: blebs are spherical; the coefficient of lateral diffusion of proteins in the bleb membrane, measured by fluorescence recovery after photobleaching, is higher than in the cell body (Tank et al. 1982); electron microscopy does not reveal actin beneath the bleb membrane. According to the results presented here it can be proposed that the membrane of tbe bleb does not have uniform mechanical properties. The nonuniformities have the same diameter as that of the process (∼0.5 μm) and are determined by the cortical structures beneath the bleb membrane. The results are in agreement with the hypothesis of Yechiel and Edidin (1987) about the existence of micrometer-scale domains in the fibroblast plasma membrane.

Thus the method of local application of mechanical force to the membrane of substratum-attached cells gives an insight into the fine organization of the cell membrane and the underlying cortical layer.

We thank Drs Y. A. Chizmadzev and L. V. Chemomordik for useful discussions.