protection when cells are exposed to laminar shear stresses of 8 N.m^2 for less than 20
minutes. Eyl et al. (1992) reported that the PMF decreases during the batch growth phase
and increases during the batch death phase, which agrees well with the findings of
Petersen et al. (1988, 1990). With respect to Pluronic this does not agree with the results
of Michaels et al. (1991b), who found no effect of Pluronic on the shear sensitivity of
cells. Finally, Wu et al. (1997) actually measured the adsorption of Pluronic F68 and
PEG to cells at different concentrations. Pluronic adsorption followed the Langmuir
isotherm with maximum adsorption values of 3.74 μg for 10^6 Sf-9 cells and 4.54 μg for
106 Tn-5 cells. Half the maximum adsorption was reached at concentrations of 2.8 g.m−^3
(0.0003%) and 16 g.m−^3 (0.002%), respectively, which is far below concentrations
normally used (0.1%). For PEG a plateau value was not observed and at a concentration
of 50 g.m−^3 , 30 μg and 48.1 μg PEG was adsorbed per 10−^6 sf-9 and Tn-5 cells,
respectively. The main advantage of viscometer and capillary-tube studies over
bioreactor studies is that they are done in a well-defined and controllable shear field.
However, the shear sensitivity of the cells still cannot be separated from the interaction of
the cells with the shear field. Thus, addition of protective compounds may affect the
fragility of the cells as well as the interaction of the cells with the shear fields. Moreover,
the exposure of cells is not comparable to the exposure in real bioreactors. For instance,
exposure times are usually longer than the exposure occurring in bioreactors and besides
the laminar and turbulent flows also elongational flows may occur as well (Garcia-
Briones and Chalmers, 1994).
In the literature two methods are described to independently measure mechanical
properties of individual cells. One method is the aspiration technique in which individual
cells are sucked partly into a pipette (Sato et al., (1987), Needham et al., 1991).
Assuming animal cells can be modelled as incompressible liquid drops with a constant
cortical tension, Needham et al. (1991) calculated the cortical membrane tension, cell
volume and apparent cell viscosity. According to Zhang and Thomas (1993a), this
technique cannot be used to calculate the strength of individual cells, because it can only
cause small deformations. Therefore, Zhang et al. (1991) developed a new micro-
manipulation technique with which the force causing large deformations can be
measured. In short, a cell is captured between two parallel surfaces (optic fibres)
connected to a micro-manipulator. Next, the cell is squeezed by moving the surfaces
towards each other, while at the same time the force applied is measured. From curves of
the force applied versus the distance between the plates the bursting force and cell
diameter can be obtained. Again modelling the cell as an incompressible liquid drop
surrounded by a thin elastic membrane, the cell bursting membrane tension, bursting
pressure and the elastic area compressibility modulus may be calculated.
Measurements with these techniques are carried out on whole cells in a time span of
the order of 0.5 s. Interactions of cells with hydrodynamic forces may involve only part
of the membrane and occur within milliseconds. Whereas, from a rational point of view,
parameters like the bursting membrane tension are good measures of the strength of the
membrane, the translation to situations in a real bioreactor remains a problem. To solve
this problem, Born et al. (1992) developed models to describe the interaction of cells with
laminar-shear fields in viscometers, while Zhang et al. (1993b) and Thomas et al. (1995)
developed such models for turbulent-flow fields in capillary tubes. Upon shearing cells at
high levels (200–600 N.m−^2 ) for short times (3 minutes) model predictions were within
Lethal effects of bubbles in animal-cell culture 463