- Email: [email protected]
A flow cytometric study of hydrodynamic damage to mammalian cells
Journal of Biotechnology, 31 (1993)161-177
161
© 1993ElsevierSciencePublishers B.V. All rights reserved0168-1656/93/$06.00
BIOTEC 00925
A flow cytometric study of hydrodynamic damage to mammalian cells M o h a m e d AI-Rubeai, A.N. Emery, S. Chal der and M.H. G o l d m a n SERC Centre for Biochemical Engineering, School of Chemical Engineering, University of Birmingham, Edgbaston, UK
(Received 12 October 1992;revisionaccepted 12 February1993)
Summary Flow cytometry has been used to study the mechanisms of damage to mammalian cells by hydrodynamic forces. Cell damage resulted from the stresses created both by bubble entrainment and by bubble bursting caused by vortex formation in highly agitated culture. Damage to the antigen molecules on the cell surface together with increasing leakage and passive transport were observed. Pluronic F-68 reduced the efflux of fluorescein out of cells suggesting the possibility of plugging damaged parts of the membrane or change in membrane molecular organisation. Surface-associated immunoglobulin molecules were also protected by Pluronic F-68. The loss of plasma membrane integrity was followed by the destruction of cytoplasmic matrix. The nuclei were last to be destroyed. The fraction of metabolically active ceils was drastically decreased by the intense hydrodynamic forces. However, the mitochondrial transmembrane potential, which is linked to the ATP requirements of ceils as well as the mean cell size of the live cell population, remained unaffected by the stressful hydrodynamic conditions. F-Actin content peaked during the early exponential phase and declined during the mid and late exponential and death phases of batch culture. The increase of actin during exponential phase was faster in stationary culture than in agitated culture. The increase was associated with the proliferative rate. Furthermore, the response to hydrodynamic forces was not related either to cell size difference or to the stage in the cell cycle. Correspondence to: Mohamed Al-Rubeai, SERC Centre for Biochemical Engineering, School of
Chemical Engineering,Universityof Birmingham, Edgbaston, BirminghamB15 2TT, UK
162 Flow Cytometry; Hydrodynamic damage; Surface antigen; Actin; Membrane integrity and function; Hybridoma
Introduction
There are manifest advantages in growing mammalian cells in suspension culture. At production scales it is however necessary to enhance mixing and transfer characteristics through the provision of mechanical agitation or gas sparging or through a combination of the two. Although it has been suggested that mammalian cells are susceptible to the hydrodynamic forces resulting from mechanical agitation, it has however been demonstrated that mechanical action alone, even at very high energy intensities, does not necessarily cause significant cell damage (Oh et al., 1989). If, however, bubbles are introduced into a mechanically agitated system then severe cell damage results. We are concerned with understanding the nature of such bubble/ agitator interactions and their effects on the ceils in the culture and we describe here the use of flow cytometry to provide understanding of the damage mechanisms. Flow cytometry is a powerful and rapid technique for analyses of cells and their components in suspension. Several interesting aspects of animal cell population dynamics can be revealed during cultivation in stressful hydrodynamic environments by flow cytometry, and quantitative information on the changes in individual cellular components in single cells can be obtained. Such analyses make flow cytometry a very valuable tool in achieving a better understanding of the biophysical mechanisms of 'shear' damage. There is an urgent need for determination of critical biophysical 'damage parameters' in order to construct models which relate reactor operations to the effect of stress on cells (Bliem, 1989). Furthermore, identification of the mechanism of cell damage and understanding of the underlying biological response (i.e. how cellular structures are responsible for maintaining cell shape and integrity) can help in establishing the appropriate design criteria for application of the higher levels of energy output required by intensified cell culture processes. We may also then understand better why some cell lines show a greater sensitivity to hydrodynamic action than others, and why such sensitivity may vary with culture conditions. In a series of experiments we have examined agitation parameters at levels well above those normally experienced in mammalian cell culture in order to accelerate cell damage processes. The work has been done in small reactors operating in fully developed turbulent conditions known to cause cell damage. The exact hydrodynamic determinants of this damage are not defined, and it must be accepted that the responses of cells to high shear stresses applied over a short time may possibly differ from those found in practical bioreactions. Such experimental restrictions, common also to studies using defined shear devices (Abu-Reesh and Kargi, 1989; McQueen et al., 1987; Petersen et al., 1988; Smith et al., 1987; Born et al., 1992),
163 have to be accepted to ensure that cell damage effects are isolated from those arising from cell aging and other physiological variations.
Materials and Methods
Cell culture Cell line T B / C 3 (a gift from Professor R. Jefferis, Department of Immunology, University of Birmingham, UK), an NSl-derived mouse hybridoma that produces IgG monoclonal antibody directed against human IgG, was routinely cultured in RPMI 1640 Medium containing 2 mg ml-1 glucose and 2 mM glutamine (GIBCO, Paisley, Scotland) supplemented with 5% (v/v) newborn bovine serum (NBS; Advanced Protein Products, Brierley Hill, England). Cells were maintained in T-flasks in a 5% COz-humidified incubator at 37°C. Viability and viable cell number of random samples were counted using trypan blue dye exclusion in a haemacytometer. Damage to 'antigen' receptor molecules The damage to antigen receptor molecules on the cell surface was characterised by measurement of surface associated IgG (slgG). slgG molecules are inserted permanently into the surface membrane where they act as specific antigen receptors and are never released to the growth medium (Roitt et al., 1990) They can be detected on the cell surface by staining the cell suspension with fluorescent IgG. Cells were exposed to steady pulses of a high agitation condition as follows: 50 ml of exponentially growing cells were inoculated into either T-flasks or 100 ml Duran bottles (60 mm in diameter; Fisons, Loughbourgh, UK) at an initial density of 3-4 x 105 m1-1 in RPMI 1640 + 5% (v/v) NBS. Pluronic F-68 (BASF Corp.) was supplemented in some experiments as required. In the first set of experiments two cultures were exposed to two different agitation conditions (600 rpm using a 25 mm x 5 mm magnetic follower, described as the high agitation condition and 400 rpm using a 40 mm x 8 mm magnetic follower, described as the very high agitation condition). Speed was determined stroboscopically. The observed vortex, which was 40 mm deep in the first culture, was just touching the follower thus leading to entrainment of large bubbles ( > 1 mm). In the second culture, due to the larger follower used, the resulting vortex was 55 mm in depth and in this case reached to the base of the vessel resulting in substantial entrainment from the gas-flooded magnetic follower leading to the formation of swarms of small bubbles ( << 1 mm). These are of course extremely high agitation rates for mammalian cell culture, but they were chosen specifically to enable us to study the effects of hydrodynamic force within a relatively short time scale (120 min) during which other growth or death effects could be assumed to be inconsequential. Cell samples were removed at intervals, centrifuged, washed with PBS, resuspended into 1 ml PBS and stained with 50 /zl of FITC-conjugated goat anti-mouse (H + L) IgG (1:10 dilution; The Binding Site, Birmingham, UK) for 30 min at room temperature. Cells were
164
washed twice with 0.5% Tween 20 (Sigma, St. Louis, MO, USA) in PBS and resuspended in PBS. Fifty/xg m1-1 propidium iodide (PI; Sigma, St. Louis, MO, USA) was added 5 min prior to analysis of cells in order to stain dead cells which would consequently be removed during analysis by flow cytometry. Fluorescence intensity and forward light scatter were measured at 30 min intervals in a FACS 440 (Beckton-Dickinson; Oxford, UK) flow cytometer.
Changes of permeability The permeability of the cell membrane was determined by measuring the effiux of fluorescein. The technique is based on the principle that fluorescein diacetate (FDA; Sigma, St. Louis, MO, USA), being an uncharged compound, passes readily into the cell and is hydrolysed into free fluorescein by esterases. This free fluorescein compound carries a net positive charge and cannot diffuse out of the cell as fast as its esters can enter. This consequently leads to a progressive increase in the free fluorescein concentration inside the cell. FDA was dissolved in acetone (Sigma, St. Louis, MO, USA) at a concentration of 1 mg ml-1 and 0.5 ml of this solution was added to 100 ml of cell suspension. In a separate control experiment no effect of 0.5% acetone on cell viability was observed. Two cultures (no Pluronic and 1% Pluronic) were exposed to a low agitation condition (200 rpm using a 25 mm x 5 mm magnetic follower) while a third culture was exposed to the high agitation condition (600 rpm with a 40 mm x 8 mm follower). The fluorescence intensity was measured at intervals in a FACS 440 flow cytometer (Becton-Dickinson). When the passive permeability of the plasma membrane increases, the effiux of intracellular fluorescein increases, resulting in lowered fluorescence.
Mitochondrial activity In another experiment employing similar exposures to hydrodynamic forces, 5 x 105 ceils per ml were incubated at 37°C in RPMI 1640 medium containing 0.5 ~g ml -x rhodamine 123 (R123; Sigma, St. Louis, MO, USA) for 30 min to analyse the mitochondrial transmembrane potential. After incubation cells were collected by centrifugation, resuspended in phosphate buffer saline (PBS), and 50/xg m1-1 propidium iodide (PI; Sigma, St. Louis, MO, USA) added 5 min prior to analysis of cells.
Actin analysis Two 50 ml batch cultures of TB/C3 cells were incubated in either T-flasks (stationary culture) or in Duran bottles agitated at 200 rpm (suspension culture). Samples were removed daily, counted, fixed with cold 70% ethanol, and stored at 4°C. The samples were centrifuged (1000 rpm, 5 min), washed twice with PBS and incubated with 0.5 ~g ml-1 FITC-phalloidin (Sigma) to visualise and quantify actin microfilaments, (Wieland, 1977) in the dark for 30 min. Cells were then washed twice in PBS and resuspended in PBS. 50 ml of 1 mg ml-1 PI was added to the cell suspensions which were kept at room temperature for 5 min before analysing them by flow cytometry.
165
Cell size and cell cycle analysis Hybridoma cells were inoculated into 100 ml Duran bottles as described above. One culture was agitated at 200 rpm with sparging at 800 ml min -1 of a i r / 5 % (v/v) CO 2. A second culture was agitated only to serve as a control. Samples were removed at intervals and analysed for light scattering characteristics without further delay. From another experiment operated under identical conditions cell samples were removed for cell cycle analysis, fixed with cold 70% ethanol and kept for at least 30 min at 4°C. The samples were centrifuged (1000 rpm, 5 min), washed twice with PBS and resuspended in 1 ml PBS. 50 ml of 1 mg ml-1 PI was added to the cell suspensions, which were kept at 4°C for 15 min before analysing them with flow cytometry. Cell size and cell cycle analysis were performed with the FACStar PLUS (Beckton-Dickinson, Oxford).
Flow cytometric analysis Flow cytometric analysis was carried out with a Becton-Dickinson FACS 440 using an argon ion laser operating at 200 mW light power with excitation at 488 nm. Duplicate samples and negative controls were analysed. Emission was collected using a 530 nm band pass filter for R123 and FITC and 620 nm band pass filter for PI. Ten to twenty thousand cells/particles were recorded and results were analysed on an IBM PC. Sorting on a cell size basis was carried out on a Coulter Elite (Coulter Electronics Ltd, Luton) at 3000 cells per s and four decisions, three drops per sort, using an argon ion laser at 488 nm. Sorting was based on forward angle light scatter and 90 ° light scatter.
Results and Discussion
Cell surface-associated immunoglobulins The cell surface is the organelle responsible for communication between the cell and its environment. Each cell expresses receptors for a wide variety of hormones, growth factors, growth substrates and other cells (Klaus, 1988). Each receptor on the cell surface requires a ligand for its activation and this receptor-ligand binding is essential to initiate the cell response to an external stimulus. An intense hydrodynamic environment may interfere with such binding processes or may damage the extracellular portion of the receptor. To quantify the magnitude of damage to the receptor-ligand interaction resulting from hydrodynamic effects we have used the binding and cross-linking of antigens on surface membrane glycoproteins as indicators of the early cell response. Figure 1 shows that the extent of ligand-receptor interaction (measured by FITC anti-mouse IgG staining) which takes place after 60 rain exposure to intense agitation is much reduced compared to that in unexposed cells and is further reduced after 120 min. No apparent change in cell size and viability is observed. There was no apparent difference in the shapes of the distributions of fluorescence or cell size which indicates a uniform decrease in the amount of sIgG across the entire population. Such a decrease in the labelled sIgG might be due to: (a) the
166 disruption or removal of the IgG antigen receptor or part of it (the epitope) on the cell surface, (b) a change in the concentration of sIgG, (c) a decrease in the rate of sIgG turnover. There is a possibility that some secreted IgG may bind to the FITC anti-mouse IgG. However, such binding should be minimised by the repeated washing of cells used. Moreover in previous work with this cell line it has been shown both that total protein synthesis is actually initially increased during intense agitation (Oh, 1991) and also that specific IgG production increases with increased agitation rate (AI-Rubeai and Emery, 1990). Such secreted IgG might also bind to Fc receptors, if such were to be expressed by these cells, but there is no evidence of this so far. Even were this so, it does not detract from the point that the cell surface interactions are reduced by agitation. The viable cell number was unchanged in the stationary culture but it was reduced in both agitation conditions, the more so in the very high agitation condition. It is clear that the number of non-viable ceils as measured by trypan blue staining does not increase at the same rate as the number of viable cells decreases. Under such intense hydrodynamic conditions the cells do not enter a relatively long intermediate 'damaged' non-viable state - but rather are disrupted directly into small cell fragments. This is consistent with the observation during the
Cell No. & Viability
Cell Size & IgG-FITC
10
120 110 100 90
4 80 70 0
I
TO
t
I
$60
$120
H60
H120
60 VH60VH120
Culture Condition m
Cell N o . ( * 1 0 s m1-1)
~
% Viability (-10)
Mean Cell Size
~
Mean IgG Receptors per Cell
Fig. 1. Effect of agitation on IgG receptors. Effect on the mean channel fluorescence distribution of IgG surface receptors of TB/C3 hybridoma cells. Cells were first incubated in stationary (S) and agitated (H, high; VH, very high) conditions for 120 min and stained with FITC-anti mouse IgG conjugate at the end of the time indicated. Cell size and IgG-FITCare in relative values. TO, time zero.
167
Cell Size & IgG-FITC
Cell No. & Viability 10
120 110
8
100 6 90 4
80
0
L
i
T0-P
*P
i
$120-P *P
M
H12()-P *P
70 i
60 VH120-P *P
Culture Condition i
Cell N o . ( O l 0 5 i n ] - 1 )
~-~
Mean Cell Size
~'~
Viability (o10) % Mean IgG Receptors per Cell
Fig. 2. Protection by Pluronic-F-68. Effect of Pluronic-F68 (P) supplementation to culture medium on the mean channel fluorescence distribution of IgG surface receptors. Cells were first incubated in stationary (S) and agitated (H, high; VH, very high) conditions for 120 min and stained with FITC-anti mouse IgG conjugate at the end of the time indicated. TO, time zero.
same experiment (see again Fig. 1) that the mean cell size of live cells remained unaffected by the different conditions used. Figure 2 shows that the addition of Pluronic F-68 had a protective effect on slgG. slgG receptor concentrations per cell were increased by the addition of Pluronic by 15% and 30% in high and very high agitation conditions respectively as compared to those in non-supplemented cultures. The loss of cell viability was also reduced by the presence of Pluronic. Viable cell number was only reduced by 15% in the presence of Pluronic in very high agitation conditions compared to a reduction of 36% in the absence of Pluronic. Here again the shapes of the distributions were similar (data not shown) indicating the absence of selective changes in specific subpopulations of cells. The protective effect of Pluronic may be related to the complex interaction between Pluronic and the cell surface membrane which has been discussed recently (Murhammer and Goochee, 1990; Smith et al., 1987). This interaction with the cell surface could also explain the so-called 'protective' effect on cell viability demonstrated by Pluronic (Handa-Corrigan et al., 1989; Murhammer and Goochee, 1988; Tramper et al., 1987). Marquis et al. (1989) suggested a plugging mechanism of damaged portions of the plasma membrane. However, Murhammer and Goochee (1988) hypothesized that the hydrophobic portion in the Pluronic could serve as an 'anchor' which allowed the
168 polyols to protect ceils from an intense hydrodynamic environment. Such interaction could result in changes in the m e m b r a n e molecular organization including receptor structure. Ramirez and Mutharasan (1990) suggested that the protective mechanism of Pluronic relies on its ability to decrease the plasma m e m b r a n e fluidity through direct interaction with it. Clearly, if such damage does occur to any of the many receptors and their ligands, it will certainly affect the response of the cell to external stimuli and hence its growth and other cellular activities.
Integrity and function of plasma membrane The cell surface is also the barrier between the cell and its environment which regulates the flow of both simple and complex molecules into and out of the cell. An attempt was made here to quantify the effect of hydrodynamic conditions on the integrity and function of the cell surface m e m b r a n e by studying the passive transport of a neutral molecule. Fig. 3 shows that the rate of effiux of fluorescein out of hybridoma cells was increased by high agitation. These trends have also been shown in a series of experiments not reported here. It was also shown that the forward and 90 ° light scatters for cells subjected to the various conditions were not changed. The increase measured in the effiux may well be due to an increase
110
F.I.(% of initial value)
90
70
50
30
0
I
I
I
I
20
40
60
80
100
Time (rain.) Low Agitation No Pluronlc
J Low Agitation 1% Pluronl¢
~
H i g h Agitation No Pluronic
Fig. 3. Efflux of fluorescein diacetate (FDA) out of hybridoma cells. Cells were incubated at low (with and without Pluronic F-68) and high agitation conditions. The relative fluorescein intensity (FI) represents the ratio of mean channel fluorescence distribution at the time indicated to that obtained initially.
169 in leakage and passive transport caused by changes in the overall state of the plasma membrane. However, the nature of this damage cannot be elucidated from this experiment alone and more work is needed to determine the relative rates of efflux of fluorescein as a function of the applied hydrodynamic condition and the presence of Pluronic in order to obtain a fuller understanding of the damage mechanism. The addition of Pluronic to the 'low' agitated culture also reduced the rate of efflux of fluorescein out of cells, although most of the reduction occurred during the first 5 min from commencing the agitation. This finding is in agreement with that of King et al. (1991) who have demonstrated that Pluronic can alter the rate of FDA uptake into yeast cells. They found that this effect occurred independently of changes in both membrane and soluble cell protein profiles. It is possible that Pluronic directly affects membrane integrity by either plugging damaged portions of the membrane reducing cell leakage or by coating the cells with a viscous layer of polymer which buffers them from hydrodynamic forces (Marquis et al., 1989). However, analysis of Pluronic-treated cells revealed that the coating action was not merely a physical effect, but that it involved direct interactions with the cell (King et al., 1991). It has also been suggested that Pluronic affects membrane strength or rigidity through alteration of the lipid-protein interactions (Helenius and Simons, 1975).
Rhodamine 123 labelling of mitochondrial activity In our flow cytometric analysis of hybridoma cells, three parameters were simultaneously measured in individual cells: mitochondria-associated fluorescence, DNA-associated fluorescence and scattered light. The two-dimensional relation between each pair of parameters was obtained and analysed. An example of the bivariate mitochondria/DNA distribution for exponentially growing hybridoma cells is shown in Fig 4. The closely packed contours seen at A in Fig. 4 represent the highly active (hence high green R123 fluorescence) cell fraction of the cell/particle population. These cells show low red PI fluorescence because they are viable and their plasma membranes are intact. However, the low green fluorescence region contains not only dead cells but also intact nuclei stripped of cytoplasmic material, as well as smaller nuclear fragments. It is apparent that the dead cell population contains cells which have died during all phases of the cell cycle. R123 stains only intact, viable and metabolically active cells. Its fluorescence intensity reflects changes in mitochondrial transmembrane potential which are linked to the process of energy metabolism (Johnson et al., 1981); thus changes in dye uptake by viable cells reflect changes in the metabolic and membrane integrity of the ceils. By inspection of the two-dimensional displays and using gating analysis we were able to classify the cell population into two different regions according to their fluorescence intensity. Fig. 5 shows the frequency of cells/particles in the bivariate R 123/PI distributions for hybridoma cells subjected to low, high and very high agitation rates for 90 min. These results show that the frequencies of intact and metabolically active cells were considerably reduced in the higher agitation conditions with substantial bubble entrainment.
170
/ ~ . ,.::~...,..
~"~';',~L~/~ i. "~','~ ~.
1".. .. . .. ..
:.
i. .¢-:,~~' ~
...'. ~ . , ' ~ . .
.
~.
~;-'.~,'
!
.
A.
.
,--
.
.
.
.
.
.
.
.
~. . . . . . . .
.
. . . . .
~, r , " ~ . % , ~ ~ " , : i :
' :. . . . . . . . . . . .
:L
.
"¢ . ~ . .'.:C" .~:~Y~ --." :,:.
.
.
.
.
9
.
t - -:'~,~."; :".--:'-:"""/. . . . . . . . . . . . . . . . . .
,~
y
.:'.~"
~........,-
.
.
.
.
!=
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.~!.;.~.:i..
'10
.¢~-
0
.--"
.i :~-~'G1
.
G 2
•..::~ ~:',. ~..,:. . . . . . . . . . . . . . .
IZ:
• .." -'¢. . ,,A . . . .
•
• .
" . . . . . . . . •. . . . . . . . .
.
.
. . . . . . ,
.
.~,
~.
. ( . . . . . . . . ~ . . . . . . .
. . ~ . . . . .
. . . . .
::
.
.
• .., . - . - : . : . . .
.
"~ • .':
: " .-~ ".. ~ i. i~' ~: ' I ",: . . . . . . . . ...... :,. . . . . .~...
.
.
~ . . . . . . . . . . . . . . . . . . . . . . ~. • ,-"L> ::. . . . . . . . . . . . "., . , ' ~ : J .~ . . . . . . . . . . . .
..-r-':
: .'~'.
.
.
.
.
.
.
:,.
.,.. : :.,'" " ' ; ; " ' - , ~ " .
;
. .
.
. .
.
~" . . . . . . . . . . . . .
.
.
.
.
. .
.
.
.
.
.
.
.
.
. .
.
.
.
.
.
.
.
~.:..i~;~.(, ... ~ ~ , h .
.; .
.
.
.
.
• . . . . . . . . . . . . . . .
.
.
: ....
.
.
.
.
. .
.
. .
.
. .
.
~.
.
.
~'~ ;,
~.
~. ,/
.
•
.s 4. :..i
" ' " "~' I . ~, ,
:....
V. .
.
•
1": : I - . ~ ' ; .".,-..~. :74.
.
.
, .
.
.
.
. . . . . . .
~:: . . . . . .
Propidium Iodide Fig. 4. Bivariate rhodamine 123 (R)/propidium iodide (PI) distribution measured for T B / C 3 hybridomas grown in stirred flask• Region A, contour of metabolically active live cells. G1 and G2, cell cycle phases of dead cells•
Such effects may be related to the stresses which result from either vortex instabilities a n d / o r the associated bubble entrainment (A1-Rubeai et al., 1990; Kunas and Papoutsakis, 1990). By contrast the mitochondrial t r a n s m e m b r a n e potential of the viable cell fraction was not affected even by acute exposure to such extreme environmental stress (Fig. 5). J o h n s o n et al. (1981) suggested a link between mitochondrial t r a n s m e m b r a n e potential and the A T P r e q u i r e m e n t of cells; m o r e o v e r we have r e p o r t e d that in batch culture the specific metabolic activity (measured by the M T T assay) after 24 h of cultivation was raised with increasing sparging rates in the range 10-100 ml min -1 in 1 1 culture volume (AloRubeai et al., 1990). It is possible then to suggest that the metabolic integrity
171 % of Cells
ReI.Rhodamine concn./cell 60
60
50
50
40
40
30
30
20
20
10
410
0
0 Low Agitation
V.High Agitation
High Agitation
Hydrodynamic Condition ~ B Rhodamine/Cell
~
% R*/PI- Cells
~
% R-/PI* Cells
Fig. 5. Effect of agitation on mitochondrial activity. Mean mitochondrial membrane potential (concentration of rhodamine per cell) and proportions (as %) of two different regions in the bivariate rhodamine 123 (R)/propidium iodide (PI) distributions measured for hybridoma cells cultivated for 90 min in various hydrodynamicenvironments. +, positive staining; - , negative or negligible staining. of those cells which withstand the impact on their plasma m e m b r a n e remains unaffected.
Cell cycle A selected example of the cell cycle distribution of cells subjected to an intense sparging pulse (800 ml m i n - 1 of a i r / 5 % CO z in 50 ml culture for 90 min) is shown in Fig 6. These distributions are based on the forward scattering and D N A fluorescence characteristics of intact cells. Debris were excluded f r o m analysis by gating. Further analyses were carried out on the cells which were mostly in the G2 and M phases of the cell cycle. These cells are characterized by their relatively larger cell size. Growth in size and cell division are linked processes since, if dividing cells were not to double their size in progressing from G1 to M, they would become progressively smaller at each division and eventually vanish (Baserga, 1985). Comparison of the cell cycle distribution of the gated cell population in the control and sparged cultures revealed no significant changes in either the distribution patterns nor the m e a n and coefficient of variation of fluorescence intensities. The absence of such changes indicates an absence of selective removal of cells of a particular size or in a particular phase of the cell cycle as a result of their sensitivity even to such extreme hydrodynamic effects. However, these results do not exclude a possible effect of hydrodynamic forces on mitotic ceils, a situation which is difficult to investigate because they only represent under 2% of the cell
172
|FIE #15:MOHAMOlOxFL2-xFluore$cence Two |
0
min
80
..I
O
04, I -
Fig. 6. Cell cycle distributions of hybridoma cells subjected to various times of air sparging. The histograms show the normalized values of red fluorescence intensity for PI stained intact cells only.
population during the exponential phase and also their DNA content and cell size are similar to those of G2 cells. Incidentally, we have observed normal cell cycle distributions in the populations of cells which have died as a result of agitation conditions (see Fig. 4). This observation again suggests the absence of differential sensitivity during different stages of the cell cycle.
Effect of agitation on F-actin content The growth curves for the stationary and agitated batch cultures are shown in Fig. 7. Both cultures show typical growth curves with the stationary culture reaching the highest cell density. Measurements of F-actin concentration per cell by flow cytometry during batch culture show a variation with time similar to that in the growth curves (Fig. 8). The change of F-actin is in fact slightly ahead of that of growth, suggesting a lack of immediate effect on viability. It is worth noting that the change in actin levels does not appear to be related to the total protein content in the cell when compared to values reported by us elsewhere (AI-Rubeai et al., 1991b). An important increase in the FITC-phalloidin concentration per cell was
173
Cell No. (m 10 5 ml - t ) 10
0 0
J
J
1
2
I
I
3
4
5
Time (day) Stationary Culture
÷
Agitated Culture
Fig. 7. Growth kinetics. Growth of T B / C 3 hybridomas in stationary and agitated cultures.
Fluorescence Intensity 100
80 60 40 4-
.
20
0
0
I
I
I
I
1
2
3
4
5
Time (day) Stationary Culture
+
Agitated Culture
Fig. 8. Actin. FITC-phalloidin analysis of fixed T B / C 3 hybridoma cells at various time during batch cultivation in stationary and agitated conditions.
174
S*G2M (%) 40
30
20
10
0
0
i
I
1
2
Time Stationary Culture
i
I
3
4
5
(day) +
Agitated Culture
Fig. 9. Cell cycle. Percentages of cells in S and G 2 + M phases of T B / C 3 hybridomas grown in stationary and agitated cultures.
seen in the lag and early exponential growth phases of this cell line. These phases are characterised by larger cell size (AI-Rubeai et al., 1991a). The requirement for an increased microfilament network as a cell enlarges is not unreasonable. Polymerisation of actin filaments at membrane-actin connections is likely to occur simultaneously with an increase in cell membrane size during cell growth. However, the difference in actin contents between the two culture conditions is not due to differences in cell size as the cell size distribution was similar in each culture. The higher actin content in stationary culture is surprising because it has been found that cells cultured with low levels of fluid stress (T-flask and slowly stirred spinner cultures) are subsequently more sensitive to applied shear than cells from rapidly agitated cultures, showing that cells respond and adapt, to some extent, to their environment (Petersen et al., 1988). This disparity might be attributed to such low agitation rates not producing sufficient adaptive pressure. The results given in Fig. 9 also suggest that the observed changes in actin content are not a consequence of the changes in cell cycle distributions that occur during batch cultivation. However, the differences in cell proliferative rate with the associated metabolic changes are seemingly influencing the cellular actin content. A1-Rubeai et al. (1990) have shown that inhibition of DNA synthesis actually does occur during exposure of suspension culture to high agitation. If actin plays any role in determining shear sensitivity the results should indicate that ceils from the lag phase are more shear sensitive than cells from the exponential phase. Indeed, Petersen et al. (1988) found such a relationship between growth phase and shear sensitivity. Furthermore, Zhang et al. (1992) showed that the mean bursting
175
tension of this cell line rose from the lag phase and peaked after 2 d in exponential growth. Recently, Goldman (1992), in this laboratory, found that Cytochalasin B (a fungal metabolite which strongly inhibits the rate of actin polymerisation) decreased the cell resistance to shear forces while the chemotactic peptide N-formylmethionylleucylphenylalanine (FMLP), which induces a rapid and transient polymerisation of actin, was found to increase the cell's resistance to shear forces. Such results also provide evidence of the role of F-actin filaments in the shear sensitivity of cells. It remains to be seen whether cells can change their F-actin concentration and cell cycle duration during long term adaptation to a higher shear environment.
Conclusions In summary then, hydrodynamic damage, at least at low levels of energy dissipation, initially results in increasing leakage and passive transport together with damage to receptors on the extracellular surface of the membrane. This is followed by the loss of plasma membrane integrity as indicated by trypan blue exclusion tests. Once membrane damage exists, it may be suggested that secondary injury is inflicted on the cell, thus leading to metabolic death. More catastrophic damage occurs at higher levels of energy dissipation (higher rates of bubble entrainment/disengagement) through rupturing of plasma membranes and internal structures and even shearing of DNA. In this case the nuclei may be ruptured at the same time as the plasma membranes. Forward scatter analyses clearly support such a description of cell destruction. At any time during high intensity agitation a bimodal rather than a normal distribution of cell/debris population prevails, indicating that the cell is not gradually destroyed by the loss of some cellular material thereby reducing cell size but rather that it experiences perhaps just after the onset of cell death a violent destruction thus shattering it into very small fragments. The insignificant change in viability as measured by trypan blue exclusion provides support to the above postulation. In these conditions viable cells which withstand the impact of high energy dissipation still, however, endure changes to their extracellular surfaces.
Acknowledgement We gratefully acknowledge the help of R. Bird for the FACS analysis and the Science and Engineering Research Council, UK for support of this work.
References Abu-Reesh, I. and Kargi, F. (1989) Biological responses of hybridoma cells to defined hydrodynamic shear stress. J. Biotechnol. 9, 167-178.
176 A1-Rubeai, M., Oh, S.K.W., Musaheb, R. and Emery, A.N. (1990) Modified cellular metabolism in hybridomas subjected to hydrodynamic and other stresses. Biotechnol. Lett. 12, 323-328. AI-Rubeai, M. and Emery, A.N. (1990) Mechanisms and kinetics of monoclonal antibody synthesis and secretion in synchronous and asynchronous hybridoma cell cultures. J. Biotechnol. 16, 67-85. AI-Rubeai, M., Chalder, S., Bird, R. and Emery, A.N. (1991a) Cell cycle, cell size and mitochondrial activity of hybridoma cells during batch cultivation. Cytotechnology 7, 179-189. A1-Rubeai, M., Emery, A.N. and Chalder, S. (1991b) Flow cytometric study of cultured mammalian cells. J. Biotechnol. 19, 67-82. Baserga, R. (1985) The Biology of Cell Reproduction. Harvard University Press, Cambridge, MA. Bliem, R. (1989) A need for systematic investigations into the material properties of cultured animal cells. Tibtech 7, 197-200. Born, C., Zhang, Z., AI-Rubeai, M. and Thomas, C.R. (1992) Estimation of disruption of animal cells by laminar shear stress. Biotechnol. Bioeng. 40, 1004-1010. Goldman, M.H. (1992) The role of the actin filament network in the shear sensitivity of cultured animal cells. B.Sc. thesis, Univ. Birmingham. Handa, A., Emery, A.N. and Spier, R.E. (1987) On the evaluation of gas-liquid interfacial effects on hybridoma viability in bubble column bioreactors. Dev. Biol. Standard, 66, 241-253. Handa-Corrigan, A., Emery, A.N. and Spier, R.E. (1989) Effects of gas-liquid interfaces on the growth of suspended mammalian cells: mechanisms of cell damage by bubbles. Enz. Microb. Technol. 11, 230-235. Helenius, A. and Simons, K. (1975) Solubilization of membrane by detergents. Biochim. Biophys. Acta 415, 29-79. Johnson, L., Walsh, M., Bockus, B.J. and Chen, L. (1981) Monitoring of relative mitochondrial membrane potential in living cells by fluorescence microscopy. J. Cell Biol. 88, 526-535. King, A.T., Davey, M.R., Mellor, I.R., Mulligan, B.J. and Lowe, K.C. (1991) Surfactant effects on yeast cells. Enz. Microb. Technol. 13, 148-153. Klaus, G.G.B. (1988) Mechanisms of receptor signalling in B-lymphocytes. In: Human Disease. Oxford University Press, Oxford, pp. 126-147. Kunas, K.T. and Papoutsakis, E.T. (1990) Damage mechanisms of suspended animal cells in agitated bioreactors with and without bubble entrainment. Biotechnol. Bioeng. 36, 476-483. Marquis, C.P., Low, K.S., Bafford, J.P. and Harbour, C. (1989) Agitation and aeration effects in suspension mammalian cell cultures. Cytotechnology 2, 163-170. McQueen, A., Meilhoc, E. and Bailey, J.E. (1987) Flow effects on the viability and lysis of suspended mammalian cells. Biotechnol. Lett. 9, 831-836. Murhammer, D. and Goochee, C. (1988) Scale up of insect cell cultures: protective effects of Pluronic F-68. Bio/Technol. 6, 1411-1418. Murhammer, D. and Goochee, C. (1990) Structural features of non-ionic polyglycol polymer molecules responsible for the protective effects in sparged animal cell bioreactors. Biotechnol. Prog. 6, 142-148. Oh, S.K.W. (1991) A study of mammalian cell behaviour in stirred bioreactors. Ph.D. thesis, Univ. Birmingham. Oh, S.K.W., Nienow, A.W., A1-Rubeai, M. and Emery, A.N. (1989) The effects of agitation intensity with and without continuous sparging on the growth and antibody production of hybridoma cells. J. Biotechnol. 12, 45-62. Petersen, J.F., Mclntire, L.V. and Papoutsakis, E.T. (1988) Shear sensitivity of cultured hybridoma cells (CRL 8018) depends on mode of growth, culture age and metabolite concentration. J. Biotechnoi. 7, 229-246. Ramirez, O.T. and Mutharasan, R. (1990) The role of the plasma membrane fluidity on the shear sensitivity of hybridomas grown under hydrodynamic stress. Biotechnol. Bioeng. 36, 911-920. Roitt, M., Brostoff, J. and Male, D. (1990) Immunology. Gower Medical Publishing Ltd, London. Smith, C.M., Hebbel, R.P., Tukey, D.P., Clawson, C.C., White, J.G. and Vercellotti, G.M. (1987) Pluronic F-68 reduces the endothelial adherence and improves the rheology of liganded sickle erythrocytes. Blood 69, 1626-1631.
177 Tramper, J., Joustra, D. and Vlak, J.M. (1987) Bioreactor design for growth of shear-sensitive insect cells. In: Webb, C. and Mavituna, F. (Eds.) Plant and Animal Cells: Process Possibilities. Ellis Horwood, Chichester, pp. 125-136. Wieland, H. (1977) Modification of actins by phallotoxins. Naturwissenschaften 64, 303-309. Zhang, Z., AI-Rubeai, M. and Thomas, C.R. (1992) Mechanical properties of hybridoma cells in batch culture. Biotechnol. Lett. 14, 11-16.
161
© 1993ElsevierSciencePublishers B.V. All rights reserved0168-1656/93/$06.00
BIOTEC 00925
A flow cytometric study of hydrodynamic damage to mammalian cells M o h a m e d AI-Rubeai, A.N. Emery, S. Chal der and M.H. G o l d m a n SERC Centre for Biochemical Engineering, School of Chemical Engineering, University of Birmingham, Edgbaston, UK
(Received 12 October 1992;revisionaccepted 12 February1993)
Summary Flow cytometry has been used to study the mechanisms of damage to mammalian cells by hydrodynamic forces. Cell damage resulted from the stresses created both by bubble entrainment and by bubble bursting caused by vortex formation in highly agitated culture. Damage to the antigen molecules on the cell surface together with increasing leakage and passive transport were observed. Pluronic F-68 reduced the efflux of fluorescein out of cells suggesting the possibility of plugging damaged parts of the membrane or change in membrane molecular organisation. Surface-associated immunoglobulin molecules were also protected by Pluronic F-68. The loss of plasma membrane integrity was followed by the destruction of cytoplasmic matrix. The nuclei were last to be destroyed. The fraction of metabolically active ceils was drastically decreased by the intense hydrodynamic forces. However, the mitochondrial transmembrane potential, which is linked to the ATP requirements of ceils as well as the mean cell size of the live cell population, remained unaffected by the stressful hydrodynamic conditions. F-Actin content peaked during the early exponential phase and declined during the mid and late exponential and death phases of batch culture. The increase of actin during exponential phase was faster in stationary culture than in agitated culture. The increase was associated with the proliferative rate. Furthermore, the response to hydrodynamic forces was not related either to cell size difference or to the stage in the cell cycle. Correspondence to: Mohamed Al-Rubeai, SERC Centre for Biochemical Engineering, School of
Chemical Engineering,Universityof Birmingham, Edgbaston, BirminghamB15 2TT, UK
162 Flow Cytometry; Hydrodynamic damage; Surface antigen; Actin; Membrane integrity and function; Hybridoma
Introduction
There are manifest advantages in growing mammalian cells in suspension culture. At production scales it is however necessary to enhance mixing and transfer characteristics through the provision of mechanical agitation or gas sparging or through a combination of the two. Although it has been suggested that mammalian cells are susceptible to the hydrodynamic forces resulting from mechanical agitation, it has however been demonstrated that mechanical action alone, even at very high energy intensities, does not necessarily cause significant cell damage (Oh et al., 1989). If, however, bubbles are introduced into a mechanically agitated system then severe cell damage results. We are concerned with understanding the nature of such bubble/ agitator interactions and their effects on the ceils in the culture and we describe here the use of flow cytometry to provide understanding of the damage mechanisms. Flow cytometry is a powerful and rapid technique for analyses of cells and their components in suspension. Several interesting aspects of animal cell population dynamics can be revealed during cultivation in stressful hydrodynamic environments by flow cytometry, and quantitative information on the changes in individual cellular components in single cells can be obtained. Such analyses make flow cytometry a very valuable tool in achieving a better understanding of the biophysical mechanisms of 'shear' damage. There is an urgent need for determination of critical biophysical 'damage parameters' in order to construct models which relate reactor operations to the effect of stress on cells (Bliem, 1989). Furthermore, identification of the mechanism of cell damage and understanding of the underlying biological response (i.e. how cellular structures are responsible for maintaining cell shape and integrity) can help in establishing the appropriate design criteria for application of the higher levels of energy output required by intensified cell culture processes. We may also then understand better why some cell lines show a greater sensitivity to hydrodynamic action than others, and why such sensitivity may vary with culture conditions. In a series of experiments we have examined agitation parameters at levels well above those normally experienced in mammalian cell culture in order to accelerate cell damage processes. The work has been done in small reactors operating in fully developed turbulent conditions known to cause cell damage. The exact hydrodynamic determinants of this damage are not defined, and it must be accepted that the responses of cells to high shear stresses applied over a short time may possibly differ from those found in practical bioreactions. Such experimental restrictions, common also to studies using defined shear devices (Abu-Reesh and Kargi, 1989; McQueen et al., 1987; Petersen et al., 1988; Smith et al., 1987; Born et al., 1992),
163 have to be accepted to ensure that cell damage effects are isolated from those arising from cell aging and other physiological variations.
Materials and Methods
Cell culture Cell line T B / C 3 (a gift from Professor R. Jefferis, Department of Immunology, University of Birmingham, UK), an NSl-derived mouse hybridoma that produces IgG monoclonal antibody directed against human IgG, was routinely cultured in RPMI 1640 Medium containing 2 mg ml-1 glucose and 2 mM glutamine (GIBCO, Paisley, Scotland) supplemented with 5% (v/v) newborn bovine serum (NBS; Advanced Protein Products, Brierley Hill, England). Cells were maintained in T-flasks in a 5% COz-humidified incubator at 37°C. Viability and viable cell number of random samples were counted using trypan blue dye exclusion in a haemacytometer. Damage to 'antigen' receptor molecules The damage to antigen receptor molecules on the cell surface was characterised by measurement of surface associated IgG (slgG). slgG molecules are inserted permanently into the surface membrane where they act as specific antigen receptors and are never released to the growth medium (Roitt et al., 1990) They can be detected on the cell surface by staining the cell suspension with fluorescent IgG. Cells were exposed to steady pulses of a high agitation condition as follows: 50 ml of exponentially growing cells were inoculated into either T-flasks or 100 ml Duran bottles (60 mm in diameter; Fisons, Loughbourgh, UK) at an initial density of 3-4 x 105 m1-1 in RPMI 1640 + 5% (v/v) NBS. Pluronic F-68 (BASF Corp.) was supplemented in some experiments as required. In the first set of experiments two cultures were exposed to two different agitation conditions (600 rpm using a 25 mm x 5 mm magnetic follower, described as the high agitation condition and 400 rpm using a 40 mm x 8 mm magnetic follower, described as the very high agitation condition). Speed was determined stroboscopically. The observed vortex, which was 40 mm deep in the first culture, was just touching the follower thus leading to entrainment of large bubbles ( > 1 mm). In the second culture, due to the larger follower used, the resulting vortex was 55 mm in depth and in this case reached to the base of the vessel resulting in substantial entrainment from the gas-flooded magnetic follower leading to the formation of swarms of small bubbles ( << 1 mm). These are of course extremely high agitation rates for mammalian cell culture, but they were chosen specifically to enable us to study the effects of hydrodynamic force within a relatively short time scale (120 min) during which other growth or death effects could be assumed to be inconsequential. Cell samples were removed at intervals, centrifuged, washed with PBS, resuspended into 1 ml PBS and stained with 50 /zl of FITC-conjugated goat anti-mouse (H + L) IgG (1:10 dilution; The Binding Site, Birmingham, UK) for 30 min at room temperature. Cells were
164
washed twice with 0.5% Tween 20 (Sigma, St. Louis, MO, USA) in PBS and resuspended in PBS. Fifty/xg m1-1 propidium iodide (PI; Sigma, St. Louis, MO, USA) was added 5 min prior to analysis of cells in order to stain dead cells which would consequently be removed during analysis by flow cytometry. Fluorescence intensity and forward light scatter were measured at 30 min intervals in a FACS 440 (Beckton-Dickinson; Oxford, UK) flow cytometer.
Changes of permeability The permeability of the cell membrane was determined by measuring the effiux of fluorescein. The technique is based on the principle that fluorescein diacetate (FDA; Sigma, St. Louis, MO, USA), being an uncharged compound, passes readily into the cell and is hydrolysed into free fluorescein by esterases. This free fluorescein compound carries a net positive charge and cannot diffuse out of the cell as fast as its esters can enter. This consequently leads to a progressive increase in the free fluorescein concentration inside the cell. FDA was dissolved in acetone (Sigma, St. Louis, MO, USA) at a concentration of 1 mg ml-1 and 0.5 ml of this solution was added to 100 ml of cell suspension. In a separate control experiment no effect of 0.5% acetone on cell viability was observed. Two cultures (no Pluronic and 1% Pluronic) were exposed to a low agitation condition (200 rpm using a 25 mm x 5 mm magnetic follower) while a third culture was exposed to the high agitation condition (600 rpm with a 40 mm x 8 mm follower). The fluorescence intensity was measured at intervals in a FACS 440 flow cytometer (Becton-Dickinson). When the passive permeability of the plasma membrane increases, the effiux of intracellular fluorescein increases, resulting in lowered fluorescence.
Mitochondrial activity In another experiment employing similar exposures to hydrodynamic forces, 5 x 105 ceils per ml were incubated at 37°C in RPMI 1640 medium containing 0.5 ~g ml -x rhodamine 123 (R123; Sigma, St. Louis, MO, USA) for 30 min to analyse the mitochondrial transmembrane potential. After incubation cells were collected by centrifugation, resuspended in phosphate buffer saline (PBS), and 50/xg m1-1 propidium iodide (PI; Sigma, St. Louis, MO, USA) added 5 min prior to analysis of cells.
Actin analysis Two 50 ml batch cultures of TB/C3 cells were incubated in either T-flasks (stationary culture) or in Duran bottles agitated at 200 rpm (suspension culture). Samples were removed daily, counted, fixed with cold 70% ethanol, and stored at 4°C. The samples were centrifuged (1000 rpm, 5 min), washed twice with PBS and incubated with 0.5 ~g ml-1 FITC-phalloidin (Sigma) to visualise and quantify actin microfilaments, (Wieland, 1977) in the dark for 30 min. Cells were then washed twice in PBS and resuspended in PBS. 50 ml of 1 mg ml-1 PI was added to the cell suspensions which were kept at room temperature for 5 min before analysing them by flow cytometry.
165
Cell size and cell cycle analysis Hybridoma cells were inoculated into 100 ml Duran bottles as described above. One culture was agitated at 200 rpm with sparging at 800 ml min -1 of a i r / 5 % (v/v) CO 2. A second culture was agitated only to serve as a control. Samples were removed at intervals and analysed for light scattering characteristics without further delay. From another experiment operated under identical conditions cell samples were removed for cell cycle analysis, fixed with cold 70% ethanol and kept for at least 30 min at 4°C. The samples were centrifuged (1000 rpm, 5 min), washed twice with PBS and resuspended in 1 ml PBS. 50 ml of 1 mg ml-1 PI was added to the cell suspensions, which were kept at 4°C for 15 min before analysing them with flow cytometry. Cell size and cell cycle analysis were performed with the FACStar PLUS (Beckton-Dickinson, Oxford).
Flow cytometric analysis Flow cytometric analysis was carried out with a Becton-Dickinson FACS 440 using an argon ion laser operating at 200 mW light power with excitation at 488 nm. Duplicate samples and negative controls were analysed. Emission was collected using a 530 nm band pass filter for R123 and FITC and 620 nm band pass filter for PI. Ten to twenty thousand cells/particles were recorded and results were analysed on an IBM PC. Sorting on a cell size basis was carried out on a Coulter Elite (Coulter Electronics Ltd, Luton) at 3000 cells per s and four decisions, three drops per sort, using an argon ion laser at 488 nm. Sorting was based on forward angle light scatter and 90 ° light scatter.
Results and Discussion
Cell surface-associated immunoglobulins The cell surface is the organelle responsible for communication between the cell and its environment. Each cell expresses receptors for a wide variety of hormones, growth factors, growth substrates and other cells (Klaus, 1988). Each receptor on the cell surface requires a ligand for its activation and this receptor-ligand binding is essential to initiate the cell response to an external stimulus. An intense hydrodynamic environment may interfere with such binding processes or may damage the extracellular portion of the receptor. To quantify the magnitude of damage to the receptor-ligand interaction resulting from hydrodynamic effects we have used the binding and cross-linking of antigens on surface membrane glycoproteins as indicators of the early cell response. Figure 1 shows that the extent of ligand-receptor interaction (measured by FITC anti-mouse IgG staining) which takes place after 60 rain exposure to intense agitation is much reduced compared to that in unexposed cells and is further reduced after 120 min. No apparent change in cell size and viability is observed. There was no apparent difference in the shapes of the distributions of fluorescence or cell size which indicates a uniform decrease in the amount of sIgG across the entire population. Such a decrease in the labelled sIgG might be due to: (a) the
166 disruption or removal of the IgG antigen receptor or part of it (the epitope) on the cell surface, (b) a change in the concentration of sIgG, (c) a decrease in the rate of sIgG turnover. There is a possibility that some secreted IgG may bind to the FITC anti-mouse IgG. However, such binding should be minimised by the repeated washing of cells used. Moreover in previous work with this cell line it has been shown both that total protein synthesis is actually initially increased during intense agitation (Oh, 1991) and also that specific IgG production increases with increased agitation rate (AI-Rubeai and Emery, 1990). Such secreted IgG might also bind to Fc receptors, if such were to be expressed by these cells, but there is no evidence of this so far. Even were this so, it does not detract from the point that the cell surface interactions are reduced by agitation. The viable cell number was unchanged in the stationary culture but it was reduced in both agitation conditions, the more so in the very high agitation condition. It is clear that the number of non-viable ceils as measured by trypan blue staining does not increase at the same rate as the number of viable cells decreases. Under such intense hydrodynamic conditions the cells do not enter a relatively long intermediate 'damaged' non-viable state - but rather are disrupted directly into small cell fragments. This is consistent with the observation during the
Cell No. & Viability
Cell Size & IgG-FITC
10
120 110 100 90
4 80 70 0
I
TO
t
I
$60
$120
H60
H120
60 VH60VH120
Culture Condition m
Cell N o . ( * 1 0 s m1-1)
~
% Viability (-10)
Mean Cell Size
~
Mean IgG Receptors per Cell
Fig. 1. Effect of agitation on IgG receptors. Effect on the mean channel fluorescence distribution of IgG surface receptors of TB/C3 hybridoma cells. Cells were first incubated in stationary (S) and agitated (H, high; VH, very high) conditions for 120 min and stained with FITC-anti mouse IgG conjugate at the end of the time indicated. Cell size and IgG-FITCare in relative values. TO, time zero.
167
Cell Size & IgG-FITC
Cell No. & Viability 10
120 110
8
100 6 90 4
80
0
L
i
T0-P
*P
i
$120-P *P
M
H12()-P *P
70 i
60 VH120-P *P
Culture Condition i
Cell N o . ( O l 0 5 i n ] - 1 )
~-~
Mean Cell Size
~'~
Viability (o10) % Mean IgG Receptors per Cell
Fig. 2. Protection by Pluronic-F-68. Effect of Pluronic-F68 (P) supplementation to culture medium on the mean channel fluorescence distribution of IgG surface receptors. Cells were first incubated in stationary (S) and agitated (H, high; VH, very high) conditions for 120 min and stained with FITC-anti mouse IgG conjugate at the end of the time indicated. TO, time zero.
same experiment (see again Fig. 1) that the mean cell size of live cells remained unaffected by the different conditions used. Figure 2 shows that the addition of Pluronic F-68 had a protective effect on slgG. slgG receptor concentrations per cell were increased by the addition of Pluronic by 15% and 30% in high and very high agitation conditions respectively as compared to those in non-supplemented cultures. The loss of cell viability was also reduced by the presence of Pluronic. Viable cell number was only reduced by 15% in the presence of Pluronic in very high agitation conditions compared to a reduction of 36% in the absence of Pluronic. Here again the shapes of the distributions were similar (data not shown) indicating the absence of selective changes in specific subpopulations of cells. The protective effect of Pluronic may be related to the complex interaction between Pluronic and the cell surface membrane which has been discussed recently (Murhammer and Goochee, 1990; Smith et al., 1987). This interaction with the cell surface could also explain the so-called 'protective' effect on cell viability demonstrated by Pluronic (Handa-Corrigan et al., 1989; Murhammer and Goochee, 1988; Tramper et al., 1987). Marquis et al. (1989) suggested a plugging mechanism of damaged portions of the plasma membrane. However, Murhammer and Goochee (1988) hypothesized that the hydrophobic portion in the Pluronic could serve as an 'anchor' which allowed the
168 polyols to protect ceils from an intense hydrodynamic environment. Such interaction could result in changes in the m e m b r a n e molecular organization including receptor structure. Ramirez and Mutharasan (1990) suggested that the protective mechanism of Pluronic relies on its ability to decrease the plasma m e m b r a n e fluidity through direct interaction with it. Clearly, if such damage does occur to any of the many receptors and their ligands, it will certainly affect the response of the cell to external stimuli and hence its growth and other cellular activities.
Integrity and function of plasma membrane The cell surface is also the barrier between the cell and its environment which regulates the flow of both simple and complex molecules into and out of the cell. An attempt was made here to quantify the effect of hydrodynamic conditions on the integrity and function of the cell surface m e m b r a n e by studying the passive transport of a neutral molecule. Fig. 3 shows that the rate of effiux of fluorescein out of hybridoma cells was increased by high agitation. These trends have also been shown in a series of experiments not reported here. It was also shown that the forward and 90 ° light scatters for cells subjected to the various conditions were not changed. The increase measured in the effiux may well be due to an increase
110
F.I.(% of initial value)
90
70
50
30
0
I
I
I
I
20
40
60
80
100
Time (rain.) Low Agitation No Pluronlc
J Low Agitation 1% Pluronl¢
~
H i g h Agitation No Pluronic
Fig. 3. Efflux of fluorescein diacetate (FDA) out of hybridoma cells. Cells were incubated at low (with and without Pluronic F-68) and high agitation conditions. The relative fluorescein intensity (FI) represents the ratio of mean channel fluorescence distribution at the time indicated to that obtained initially.
169 in leakage and passive transport caused by changes in the overall state of the plasma membrane. However, the nature of this damage cannot be elucidated from this experiment alone and more work is needed to determine the relative rates of efflux of fluorescein as a function of the applied hydrodynamic condition and the presence of Pluronic in order to obtain a fuller understanding of the damage mechanism. The addition of Pluronic to the 'low' agitated culture also reduced the rate of efflux of fluorescein out of cells, although most of the reduction occurred during the first 5 min from commencing the agitation. This finding is in agreement with that of King et al. (1991) who have demonstrated that Pluronic can alter the rate of FDA uptake into yeast cells. They found that this effect occurred independently of changes in both membrane and soluble cell protein profiles. It is possible that Pluronic directly affects membrane integrity by either plugging damaged portions of the membrane reducing cell leakage or by coating the cells with a viscous layer of polymer which buffers them from hydrodynamic forces (Marquis et al., 1989). However, analysis of Pluronic-treated cells revealed that the coating action was not merely a physical effect, but that it involved direct interactions with the cell (King et al., 1991). It has also been suggested that Pluronic affects membrane strength or rigidity through alteration of the lipid-protein interactions (Helenius and Simons, 1975).
Rhodamine 123 labelling of mitochondrial activity In our flow cytometric analysis of hybridoma cells, three parameters were simultaneously measured in individual cells: mitochondria-associated fluorescence, DNA-associated fluorescence and scattered light. The two-dimensional relation between each pair of parameters was obtained and analysed. An example of the bivariate mitochondria/DNA distribution for exponentially growing hybridoma cells is shown in Fig 4. The closely packed contours seen at A in Fig. 4 represent the highly active (hence high green R123 fluorescence) cell fraction of the cell/particle population. These cells show low red PI fluorescence because they are viable and their plasma membranes are intact. However, the low green fluorescence region contains not only dead cells but also intact nuclei stripped of cytoplasmic material, as well as smaller nuclear fragments. It is apparent that the dead cell population contains cells which have died during all phases of the cell cycle. R123 stains only intact, viable and metabolically active cells. Its fluorescence intensity reflects changes in mitochondrial transmembrane potential which are linked to the process of energy metabolism (Johnson et al., 1981); thus changes in dye uptake by viable cells reflect changes in the metabolic and membrane integrity of the ceils. By inspection of the two-dimensional displays and using gating analysis we were able to classify the cell population into two different regions according to their fluorescence intensity. Fig. 5 shows the frequency of cells/particles in the bivariate R 123/PI distributions for hybridoma cells subjected to low, high and very high agitation rates for 90 min. These results show that the frequencies of intact and metabolically active cells were considerably reduced in the higher agitation conditions with substantial bubble entrainment.
170
/ ~ . ,.::~...,..
~"~';',~L~/~ i. "~','~ ~.
1".. .. . .. ..
:.
i. .¢-:,~~' ~
...'. ~ . , ' ~ . .
.
~.
~;-'.~,'
!
.
A.
.
,--
.
.
.
.
.
.
.
.
~. . . . . . . .
.
. . . . .
~, r , " ~ . % , ~ ~ " , : i :
' :. . . . . . . . . . . .
:L
.
"¢ . ~ . .'.:C" .~:~Y~ --." :,:.
.
.
.
.
9
.
t - -:'~,~."; :".--:'-:"""/. . . . . . . . . . . . . . . . . .
,~
y
.:'.~"
~........,-
.
.
.
.
!=
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.~!.;.~.:i..
'10
.¢~-
0
.--"
.i :~-~'G1
.
G 2
•..::~ ~:',. ~..,:. . . . . . . . . . . . . . .
IZ:
• .." -'¢. . ,,A . . . .
•
• .
" . . . . . . . . •. . . . . . . . .
.
.
. . . . . . ,
.
.~,
~.
. ( . . . . . . . . ~ . . . . . . .
. . ~ . . . . .
. . . . .
::
.
.
• .., . - . - : . : . . .
.
"~ • .':
: " .-~ ".. ~ i. i~' ~: ' I ",: . . . . . . . . ...... :,. . . . . .~...
.
.
~ . . . . . . . . . . . . . . . . . . . . . . ~. • ,-"L> ::. . . . . . . . . . . . "., . , ' ~ : J .~ . . . . . . . . . . . .
..-r-':
: .'~'.
.
.
.
.
.
.
:,.
.,.. : :.,'" " ' ; ; " ' - , ~ " .
;
. .
.
. .
.
~" . . . . . . . . . . . . .
.
.
.
.
. .
.
.
.
.
.
.
.
.
. .
.
.
.
.
.
.
.
~.:..i~;~.(, ... ~ ~ , h .
.; .
.
.
.
.
• . . . . . . . . . . . . . . .
.
.
: ....
.
.
.
.
. .
.
. .
.
. .
.
~.
.
.
~'~ ;,
~.
~. ,/
.
•
.s 4. :..i
" ' " "~' I . ~, ,
:....
V. .
.
•
1": : I - . ~ ' ; .".,-..~. :74.
.
.
, .
.
.
.
. . . . . . .
~:: . . . . . .
Propidium Iodide Fig. 4. Bivariate rhodamine 123 (R)/propidium iodide (PI) distribution measured for T B / C 3 hybridomas grown in stirred flask• Region A, contour of metabolically active live cells. G1 and G2, cell cycle phases of dead cells•
Such effects may be related to the stresses which result from either vortex instabilities a n d / o r the associated bubble entrainment (A1-Rubeai et al., 1990; Kunas and Papoutsakis, 1990). By contrast the mitochondrial t r a n s m e m b r a n e potential of the viable cell fraction was not affected even by acute exposure to such extreme environmental stress (Fig. 5). J o h n s o n et al. (1981) suggested a link between mitochondrial t r a n s m e m b r a n e potential and the A T P r e q u i r e m e n t of cells; m o r e o v e r we have r e p o r t e d that in batch culture the specific metabolic activity (measured by the M T T assay) after 24 h of cultivation was raised with increasing sparging rates in the range 10-100 ml min -1 in 1 1 culture volume (AloRubeai et al., 1990). It is possible then to suggest that the metabolic integrity
171 % of Cells
ReI.Rhodamine concn./cell 60
60
50
50
40
40
30
30
20
20
10
410
0
0 Low Agitation
V.High Agitation
High Agitation
Hydrodynamic Condition ~ B Rhodamine/Cell
~
% R*/PI- Cells
~
% R-/PI* Cells
Fig. 5. Effect of agitation on mitochondrial activity. Mean mitochondrial membrane potential (concentration of rhodamine per cell) and proportions (as %) of two different regions in the bivariate rhodamine 123 (R)/propidium iodide (PI) distributions measured for hybridoma cells cultivated for 90 min in various hydrodynamicenvironments. +, positive staining; - , negative or negligible staining. of those cells which withstand the impact on their plasma m e m b r a n e remains unaffected.
Cell cycle A selected example of the cell cycle distribution of cells subjected to an intense sparging pulse (800 ml m i n - 1 of a i r / 5 % CO z in 50 ml culture for 90 min) is shown in Fig 6. These distributions are based on the forward scattering and D N A fluorescence characteristics of intact cells. Debris were excluded f r o m analysis by gating. Further analyses were carried out on the cells which were mostly in the G2 and M phases of the cell cycle. These cells are characterized by their relatively larger cell size. Growth in size and cell division are linked processes since, if dividing cells were not to double their size in progressing from G1 to M, they would become progressively smaller at each division and eventually vanish (Baserga, 1985). Comparison of the cell cycle distribution of the gated cell population in the control and sparged cultures revealed no significant changes in either the distribution patterns nor the m e a n and coefficient of variation of fluorescence intensities. The absence of such changes indicates an absence of selective removal of cells of a particular size or in a particular phase of the cell cycle as a result of their sensitivity even to such extreme hydrodynamic effects. However, these results do not exclude a possible effect of hydrodynamic forces on mitotic ceils, a situation which is difficult to investigate because they only represent under 2% of the cell
172
|FIE #15:MOHAMOlOxFL2-xFluore$cence Two |
0
min
80
..I
O
04, I -
Fig. 6. Cell cycle distributions of hybridoma cells subjected to various times of air sparging. The histograms show the normalized values of red fluorescence intensity for PI stained intact cells only.
population during the exponential phase and also their DNA content and cell size are similar to those of G2 cells. Incidentally, we have observed normal cell cycle distributions in the populations of cells which have died as a result of agitation conditions (see Fig. 4). This observation again suggests the absence of differential sensitivity during different stages of the cell cycle.
Effect of agitation on F-actin content The growth curves for the stationary and agitated batch cultures are shown in Fig. 7. Both cultures show typical growth curves with the stationary culture reaching the highest cell density. Measurements of F-actin concentration per cell by flow cytometry during batch culture show a variation with time similar to that in the growth curves (Fig. 8). The change of F-actin is in fact slightly ahead of that of growth, suggesting a lack of immediate effect on viability. It is worth noting that the change in actin levels does not appear to be related to the total protein content in the cell when compared to values reported by us elsewhere (AI-Rubeai et al., 1991b). An important increase in the FITC-phalloidin concentration per cell was
173
Cell No. (m 10 5 ml - t ) 10
0 0
J
J
1
2
I
I
3
4
5
Time (day) Stationary Culture
÷
Agitated Culture
Fig. 7. Growth kinetics. Growth of T B / C 3 hybridomas in stationary and agitated cultures.
Fluorescence Intensity 100
80 60 40 4-
.
20
0
0
I
I
I
I
1
2
3
4
5
Time (day) Stationary Culture
+
Agitated Culture
Fig. 8. Actin. FITC-phalloidin analysis of fixed T B / C 3 hybridoma cells at various time during batch cultivation in stationary and agitated conditions.
174
S*G2M (%) 40
30
20
10
0
0
i
I
1
2
Time Stationary Culture
i
I
3
4
5
(day) +
Agitated Culture
Fig. 9. Cell cycle. Percentages of cells in S and G 2 + M phases of T B / C 3 hybridomas grown in stationary and agitated cultures.
seen in the lag and early exponential growth phases of this cell line. These phases are characterised by larger cell size (AI-Rubeai et al., 1991a). The requirement for an increased microfilament network as a cell enlarges is not unreasonable. Polymerisation of actin filaments at membrane-actin connections is likely to occur simultaneously with an increase in cell membrane size during cell growth. However, the difference in actin contents between the two culture conditions is not due to differences in cell size as the cell size distribution was similar in each culture. The higher actin content in stationary culture is surprising because it has been found that cells cultured with low levels of fluid stress (T-flask and slowly stirred spinner cultures) are subsequently more sensitive to applied shear than cells from rapidly agitated cultures, showing that cells respond and adapt, to some extent, to their environment (Petersen et al., 1988). This disparity might be attributed to such low agitation rates not producing sufficient adaptive pressure. The results given in Fig. 9 also suggest that the observed changes in actin content are not a consequence of the changes in cell cycle distributions that occur during batch cultivation. However, the differences in cell proliferative rate with the associated metabolic changes are seemingly influencing the cellular actin content. A1-Rubeai et al. (1990) have shown that inhibition of DNA synthesis actually does occur during exposure of suspension culture to high agitation. If actin plays any role in determining shear sensitivity the results should indicate that ceils from the lag phase are more shear sensitive than cells from the exponential phase. Indeed, Petersen et al. (1988) found such a relationship between growth phase and shear sensitivity. Furthermore, Zhang et al. (1992) showed that the mean bursting
175
tension of this cell line rose from the lag phase and peaked after 2 d in exponential growth. Recently, Goldman (1992), in this laboratory, found that Cytochalasin B (a fungal metabolite which strongly inhibits the rate of actin polymerisation) decreased the cell resistance to shear forces while the chemotactic peptide N-formylmethionylleucylphenylalanine (FMLP), which induces a rapid and transient polymerisation of actin, was found to increase the cell's resistance to shear forces. Such results also provide evidence of the role of F-actin filaments in the shear sensitivity of cells. It remains to be seen whether cells can change their F-actin concentration and cell cycle duration during long term adaptation to a higher shear environment.
Conclusions In summary then, hydrodynamic damage, at least at low levels of energy dissipation, initially results in increasing leakage and passive transport together with damage to receptors on the extracellular surface of the membrane. This is followed by the loss of plasma membrane integrity as indicated by trypan blue exclusion tests. Once membrane damage exists, it may be suggested that secondary injury is inflicted on the cell, thus leading to metabolic death. More catastrophic damage occurs at higher levels of energy dissipation (higher rates of bubble entrainment/disengagement) through rupturing of plasma membranes and internal structures and even shearing of DNA. In this case the nuclei may be ruptured at the same time as the plasma membranes. Forward scatter analyses clearly support such a description of cell destruction. At any time during high intensity agitation a bimodal rather than a normal distribution of cell/debris population prevails, indicating that the cell is not gradually destroyed by the loss of some cellular material thereby reducing cell size but rather that it experiences perhaps just after the onset of cell death a violent destruction thus shattering it into very small fragments. The insignificant change in viability as measured by trypan blue exclusion provides support to the above postulation. In these conditions viable cells which withstand the impact of high energy dissipation still, however, endure changes to their extracellular surfaces.
Acknowledgement We gratefully acknowledge the help of R. Bird for the FACS analysis and the Science and Engineering Research Council, UK for support of this work.
References Abu-Reesh, I. and Kargi, F. (1989) Biological responses of hybridoma cells to defined hydrodynamic shear stress. J. Biotechnol. 9, 167-178.
176 A1-Rubeai, M., Oh, S.K.W., Musaheb, R. and Emery, A.N. (1990) Modified cellular metabolism in hybridomas subjected to hydrodynamic and other stresses. Biotechnol. Lett. 12, 323-328. AI-Rubeai, M. and Emery, A.N. (1990) Mechanisms and kinetics of monoclonal antibody synthesis and secretion in synchronous and asynchronous hybridoma cell cultures. J. Biotechnol. 16, 67-85. AI-Rubeai, M., Chalder, S., Bird, R. and Emery, A.N. (1991a) Cell cycle, cell size and mitochondrial activity of hybridoma cells during batch cultivation. Cytotechnology 7, 179-189. A1-Rubeai, M., Emery, A.N. and Chalder, S. (1991b) Flow cytometric study of cultured mammalian cells. J. Biotechnol. 19, 67-82. Baserga, R. (1985) The Biology of Cell Reproduction. Harvard University Press, Cambridge, MA. Bliem, R. (1989) A need for systematic investigations into the material properties of cultured animal cells. Tibtech 7, 197-200. Born, C., Zhang, Z., AI-Rubeai, M. and Thomas, C.R. (1992) Estimation of disruption of animal cells by laminar shear stress. Biotechnol. Bioeng. 40, 1004-1010. Goldman, M.H. (1992) The role of the actin filament network in the shear sensitivity of cultured animal cells. B.Sc. thesis, Univ. Birmingham. Handa, A., Emery, A.N. and Spier, R.E. (1987) On the evaluation of gas-liquid interfacial effects on hybridoma viability in bubble column bioreactors. Dev. Biol. Standard, 66, 241-253. Handa-Corrigan, A., Emery, A.N. and Spier, R.E. (1989) Effects of gas-liquid interfaces on the growth of suspended mammalian cells: mechanisms of cell damage by bubbles. Enz. Microb. Technol. 11, 230-235. Helenius, A. and Simons, K. (1975) Solubilization of membrane by detergents. Biochim. Biophys. Acta 415, 29-79. Johnson, L., Walsh, M., Bockus, B.J. and Chen, L. (1981) Monitoring of relative mitochondrial membrane potential in living cells by fluorescence microscopy. J. Cell Biol. 88, 526-535. King, A.T., Davey, M.R., Mellor, I.R., Mulligan, B.J. and Lowe, K.C. (1991) Surfactant effects on yeast cells. Enz. Microb. Technol. 13, 148-153. Klaus, G.G.B. (1988) Mechanisms of receptor signalling in B-lymphocytes. In: Human Disease. Oxford University Press, Oxford, pp. 126-147. Kunas, K.T. and Papoutsakis, E.T. (1990) Damage mechanisms of suspended animal cells in agitated bioreactors with and without bubble entrainment. Biotechnol. Bioeng. 36, 476-483. Marquis, C.P., Low, K.S., Bafford, J.P. and Harbour, C. (1989) Agitation and aeration effects in suspension mammalian cell cultures. Cytotechnology 2, 163-170. McQueen, A., Meilhoc, E. and Bailey, J.E. (1987) Flow effects on the viability and lysis of suspended mammalian cells. Biotechnol. Lett. 9, 831-836. Murhammer, D. and Goochee, C. (1988) Scale up of insect cell cultures: protective effects of Pluronic F-68. Bio/Technol. 6, 1411-1418. Murhammer, D. and Goochee, C. (1990) Structural features of non-ionic polyglycol polymer molecules responsible for the protective effects in sparged animal cell bioreactors. Biotechnol. Prog. 6, 142-148. Oh, S.K.W. (1991) A study of mammalian cell behaviour in stirred bioreactors. Ph.D. thesis, Univ. Birmingham. Oh, S.K.W., Nienow, A.W., A1-Rubeai, M. and Emery, A.N. (1989) The effects of agitation intensity with and without continuous sparging on the growth and antibody production of hybridoma cells. J. Biotechnol. 12, 45-62. Petersen, J.F., Mclntire, L.V. and Papoutsakis, E.T. (1988) Shear sensitivity of cultured hybridoma cells (CRL 8018) depends on mode of growth, culture age and metabolite concentration. J. Biotechnoi. 7, 229-246. Ramirez, O.T. and Mutharasan, R. (1990) The role of the plasma membrane fluidity on the shear sensitivity of hybridomas grown under hydrodynamic stress. Biotechnol. Bioeng. 36, 911-920. Roitt, M., Brostoff, J. and Male, D. (1990) Immunology. Gower Medical Publishing Ltd, London. Smith, C.M., Hebbel, R.P., Tukey, D.P., Clawson, C.C., White, J.G. and Vercellotti, G.M. (1987) Pluronic F-68 reduces the endothelial adherence and improves the rheology of liganded sickle erythrocytes. Blood 69, 1626-1631.
177 Tramper, J., Joustra, D. and Vlak, J.M. (1987) Bioreactor design for growth of shear-sensitive insect cells. In: Webb, C. and Mavituna, F. (Eds.) Plant and Animal Cells: Process Possibilities. Ellis Horwood, Chichester, pp. 125-136. Wieland, H. (1977) Modification of actins by phallotoxins. Naturwissenschaften 64, 303-309. Zhang, Z., AI-Rubeai, M. and Thomas, C.R. (1992) Mechanical properties of hybridoma cells in batch culture. Biotechnol. Lett. 14, 11-16.