362
Biotechnol. Prog. 1993, 9,362-365
Surface Area and Anchorage-Dependent Growth of Chinese Hamster Ovary Cells R.-C. Ruaan, G.-J. Tsai, and G . T. Tsao' School of Chemical Engineering and Laboratory of Renewable Resources Engineering, Purdue University, West Lafayette, Indiana 47907
Cell growth can be affected by many factors. For study of the cell responses under various environmental changes, the traditional method using batch cultures may not provide precise answers. It sometimes even leads to misapprehension. A flow system was therefore designed to eliminate the variations in chemical environment during the cell growth. An anchorage Chinese hamster ovary (CHO) cell line was chosen for study under this system. With the help of time-lapse video recording and image analysis, it was found that the exposed surface area (A,) of Chinese hamster ovary (CHO) cells was significantly affected by the serum concentration (S) in the medium. Cells appeared to be larger when cultivated in medium of low serum content. Comparing the growth rates of cells on different substrata, we have found that, although they were different at a given serum concentration, the growth curves could be correlated to a Monod type equation by replacing the substrate term with A$. Furthermore, when we studied the effect of density-dependent inhibition of cell growth, we found that there exists a minimum value of A,S required for stimulating cell proliferation. Once A,S falls below this value, cell proliferation stops. These observations seemed to indicate that the interaction between serum factors and cell surface, Le., A,S, is an important parameter of the growth of CHO cells.
Introduction The growth of animal cells is relatively sensitive to environmental factors. It can be affected physically by heat or shear stress, it can be affected chemically by medium composition and cell excretions, and it can be affected biologically through cell-cell interaction. Although some of these factors can be properly controlled inside the laboratory, there are still a lot of factors that cannot be studied independently. Most of the previous studies of animal cell growth were performed in batch cultures (1-3). Although some of the studies were performed with daily medium change, one cannot be sure whether a necessary factor was depleted. The problem is also complicated by the cell excretions. Whether growth retardation or acceleration is caused by the depletion of serum factors or by the excretion of metabolic waste or growth factors cannot be distinguished by batch operations, not even with daily medium change. A flow system was employed in this study to overcome these difficulties (4). Fresh medium was introduced continuously through the flask. The flow rate was kept fast enough to maintain a sufficient supply of medium and slow enough to avoid high shear stress. The complications caused by the excreted factors were removed by the flow. The effects of the amount of serum added to the medium and the substratum to which cells attach on cell growth were closely monitored by time-lapse video recording. To ensure a sufficient supply of nutrients and complete removal of metabolites, glucose and lactate concentrationswere periodicallymonitored. The medium flow rate was controlled such that the glucose level was maintained and the lactate concentration was immeasurably small. The shape of cells has been found to be closely related to the growth of anchorage cells ( 5 , 6 ) . It has been shown that cell shape is tightly coupled to DNA synthesis in 8756-7938/93/3009-0362$04.00/0
nontransformed cells (7). The change of cell shape has also been proposed to be responsible for the phenomenon of density-dependent inhibition of cell growth (8). To examine whether the controlling factor is the cell shape itself or the consequent surface area change, we performed experiments at various serum concentrations and monitored the growth of Chinese hamster ovary (CHO) cells on two different substrata. The attached cell surface area and the growth rate were carefully monitored. By comparing cells grown on tissue culture flasks with those grown on bacteriologicalPetri dishes, it was found that, although the growth rates were different at a given serum concentration, the growth curves coincided with each other when we plotted cell growth rate versus the product of surface area and serum concentration (A,$). It was also found in the confluent cultures that the saturation cell density and the cell surface for effective mass transfer (A,) were inversely proportional to serum concentration when the serum concentration (S) was low. This may imply that a minimum A$ value is required for cell proliferation. When the A,S value falls below a certain value, cell proliferation terminates. Our observations indicate that it is cell surface area instead of cell shape that is the important factor in the growth regulation of CHO cells.
Materials and Methods CultureConditions. Wild-type Chinese hamster ovary (CHO) cells were used for all of the experiments in this study. Cells were cultivated in minimum essential medium (MEM) containing 1 % penicillin-streptomycin and varying amounts of heat-inactivated fetal calf serum (GIBCO Laboratories). Before each continuous run, cells were precultured in the same medium for 30 h. Precultured cells were trypsinized and suspended in the same medium for inoculation. For each continuous run, cells were allowed to settle for 2 h before introduction of the flow.
0 1993 American Chemical Society and American Institute of Chemical Engineers
EbtmhmL pmp., 1883. Vol. 8, No. 4
Medium flow rate was adjusted according to the cell density, and the liquid holdup was kept at 10 mL. For each run, the flow rate was started at 0.5 mL/h while the cell density was about 20-30 celld0.25 mm2 and was adjusted every 12 h according to cell density. The final flow rate was kept at 5 mL/h when cells reached confluence. System Setup. Cells were cultivated on the surface of atissue culture flask or a bacteriological Petri dish. Fresh medium was flowed into the culture vessel and the waste was continuously withdrawn. The attachment, spreading, and dividing of cells were constantly monitored through an invertedmicroscope(OlympusIMT-2),and theimages were recorded by a video cassette recorder, digitized by a framegrabber (PCVISIONplus, Imaging Technologies), and analyzed by image analysis software (JAVA, Jandel Video Analysis Software, Jandel Scientific). Glucose and Lactate Concentrations. Sampleswere taken from the flask every 12 h. Glucose was assayed with aYSIglucose analyzer Model 27,and lactate concentration was measured with a Sigma lactate assay kit 826-10. Cell Density Measurement. Cells growing in an area of 0.25 mm2 were monitored under the microscope. The cell count was done by image analysis software and was compared with the result from manual counting. The number of cells in each frame was calculated by averaging the cell count results from two different persons. The percentage deviation of the cell count of each frame was always within 4%. Specific Growth Rate Measurement. Specific growth rate wasmeasuredhyaveragingthedoublingtime between the first and second cell divisions. These results were found to be comparable with the specific growth rates obtained bytakingslopesofthegrowthcurvesattheregion between 20 and 50 h after inoculation (early exponential phase). Projective Cell Area. The projective cell area was estimated by image analysis software. The edge of each cell image was trimmed during image processing. Therefore, the projective area of each cell obtained by this method was smaller than the actual size. However, the procedures for image processing were standardized and applied to every frame. The percentage of size decrease through the procedures was assumed to he the same. The estimations were performed by two different persons, and the deviations were within 17%. Cell Volume Estimation. Cells were cultured in media containing varying amounts of fetal calf serum. Thirty minutes after inoculation, each culture was trypsinized and cells were transferred into a bacteriological Petri dish. Twenty minutes after cell transfer, all of the cells had settled down and appeared to be spherical in shape. The image of each cell was taken, and the projective area was calculated by image analysis. The cell diameter (d,) was calculated by A, = 7rdC2/4,and the average projective area was obtained by averaging the projective areas of 30 cells. Cell Shape Characterization. CHO cells spreading on surfaces were assumed to have the shape of a hemiellipsoid. The long axis and the short axis were calculated from the projective area and perimeter information. If we denote the long axis as a and the short axis as b, the projective area (A,) of a hemiellipsoid is equal to Tab, and the perimeter (P)can be approximated as (a2 b2)1/27r/2. Since the projective area and perimeter of each cell can be obtained from image analysis, the long and short axes are easily calculated. Estimation of Exposed Surface Area of CHO Cells. Supposethat CHO cells have the shape of ahemiellipsoid, of which the three axes have lengths a, b, and c. The
+
(b)
Figure 1. Shape of CHO cells on different substrata: (a) CHO cells in a tissue culture flask; (h) CHO cells in a bacteriological Petridish. Cellsare cultivated inmedium containing 1.5%heatdeactivated fetal calf serum.
parametric representation of an ellipsoid is x = a sin.$cosB
y = bsin .$sin8
z = ccosfl
where 0
