Effect of Shear Stress on Expression of a
Recombinant Protein by Chinese
Hamster Ovary Cells
Julian T. Keane,1 David Ryan,2 Peter P. Gray1
1
Department of Biotechnology, University of New South Wales, Sydney
2052, Australia; telephone: +61-29-385-2057; fax: +61-29-313-6710; e-mail:
p.gray@unsw.edu.au
2
CSL, Ltd., Melbourne, Australia
Received 13 January 2002; accepted 19 June 2002
DOI: 10.1002/bit.10472
Abstract: A flow chamber was used to impart a steady
laminar shear stress on a recombinant Chinese hamster
ovary (CHO) cell line expressing human growth hormone
(hGH). The cells were subjected to shear stress ranging
from 0.005 to 0.80 N m−2. The effect of shear stress on the
cell specific glucose uptake, cell specific hGH, and lactate
productivity rates were calculated. No morphological
changes to the cells were observed over the range of
shear stresses examined. When the cells were subjected
to 0.10 N m −2 shear in protein-free media without
Pluronic F-68, recombinant protein production ceased
with no change in cell morphology, whereas control cul-
tures were expressing hGH at 0.35 µg/106 cells/h. Upon
addition of the shear protectants, Pluronic F-68 (0.2%
[w/v]) or fetal bovine serum (1.0% [v/v] FBS), the produc-
tivity of the cells was restored. The effect of increasing
shear stress on the cells in protein-free medium contain-
ing Pluronic F-68 was also investigated. Cell specific
metabolic rates were calculated for cells under shear
stress and for no-shear control cultures performed in
parallel, with shear stress rates expressed as a percent-
age of those obtained for control cultures. Upon increas-
ing shear from 0.005 to 0.80 N m−2, the cell specific hGH
productivity decreased from 100% at 0.005 N m−2 to 49%
at 0.80 N m−2 relative to the no-shear control. A concur-
rent increase in the glucose uptake rate from 115% at
0.01 N m−2 to 142% at 0.80 N m−2, and decreased lactate
productivity from 92% to 50%, revealed a change in the
yield of products from glucose compared with the static
control. It was shown that shear stress, at sublytic levels
in medium containing Pluronic F-68, could decrease hGH
specific productivity. © 2002 Wiley Periodicals, Inc. Biotech-
nol Bioeng 81: 211–220, 2003.
Keywords: Chinese hamster ovary (CHO); human growth
hormone (hGH); recombinant protein productivity;
Pluronic F-68; shear protectants
INTRODUCTION
Cells in most bioreactor systems are subjected to fluid-
mechanical shear stress through the processes of agitation
Correspondence to: J. Keane
Contract grant sponsor: The Australian Government’s Co-Operative Re-
search Center Program
© 2002 Wiley Periodicals, Inc.
and aeration. For the rational design of reactor systems and
scale-up strategies, an understanding of the effect of shear
stress on cellular viability, metabolism, and recombinant
protein productivity is required. It has been well established
that sparging can damage cells by subjecting them to high
shear forces at the air/liquid interface (Bavarian et al., 1991;
Cherry and Hulle, 1992; Garcia-Briones and Chalmers,
1990; Handa et al., 1987; Handa-Corrigan et al., 1989; Jor-
dan et al., 1994; Trinh et al., 1994). Shear stress has been
found to have a range of effects depending on the cell type
and the severity and duration of the exposure. At high levels
of shear the viability and growth of cells can be affected by
the shear environment (Garcia-Briones and Chalmers, 1994;
Gregoriades et al., 2000; McDowell and Papoutsakis, 1998;
Michaels et al., 1996).
Increasing the amount of stress or exposure time to hy-
bridomas was found to increase damage and death rate of
the cells (Kretzmer et al., 1990; Peterson et al., 1988) and
increased cell detachment (Kretzmer and Schugerl, 1991).
The stage of growth of CRL-8018 hybridoma cells (Peter-
son et al., 1988) and exposure duration and magnitude of
shear stress can have an impact on the effect of shear stress
on the cell. Dividing baby hamster kidney (BHK) cells were
found to require more time for cell spreading compared with
static controls when subjected to shear stress (Ludwig et al.,
1992). Murine hybridoma (TB/C3), insect cells (Sf9), and
CHO cells in cell-cycle stages S1 and G2 were found to be
more susceptible to shear stress than those cells in G1 phase
when exposed to intense hydrodynamic forces in a turbulent
flow capillary tube and in separate experiments by con-
trolled agitation and aeration (Al-Rubeai et al., 1995b). It
was found that the shear forces damaged the antigen mol-
ecules on the cell surface, together with increasing leakage
of fluorescein and an increase in passive transport. The
presence of Pluronic F-68 in the media was found to protect
the surface-associated immunoglobulins from shear stress
as well as decreasing the efflux of fluorescein out of the
cells, highlighting the impact medium formulation has on
the cellular effect of shear stress. The mitochondrial mem-
brane potential, linked to the ATP requirements of the cell,
was unaffected by the shear stress.
Under a shear stress severe enough to cause gas entrain-
ment, cells were shown to lose the ability to maintain ion
gradients and passive transport increases, leading to loss of
cell viability (Al-Rubeai et al., 1993). This has also been
demonstrated by the affect on cytostolic pH, due to prefer-
ential leakage of small ions out of rat aortic endothelial cells
grown on glass capillaries (Ziegelstein et al., 1992). The
permeability enhancement for rat endothelial cells occurred
even at the low shear stress of 0.05 N m−2 over a duration
of 2 min. Loss of cell viability due to shear stress by low-
level shear stress (320 W m−3) has also been examined.
Murine hybridoma cells were subjected to sublytic forces
and were observed to cause cellular damage within the cell,
which, over time, caused the induction of the apoptotic
pathway and ultimately cell death (Al-Rubeai et al., 1995a).
Prediction of cellular damage due to shear stress is dif-
ficult due to cell-line variations and unknown mechanisms
for cell responses to shear stress. Micromanipulation tech-
niques have been used to measure the fragility of hybrido-
mas to shear stress by measuring the membrane-bursting
tension to predict the disruption of cells by laminar shear
stress (Born et al., 1992). Losses of cells could be predicted
within a maximum error of 30%, although this was only for
short exposure periods (180 s). These techniques have also
been used to assess the relative fragilities of murine hybri-
domas, myelomas, and insect cells. Hybridomas were found
to be the least fragile of these cell lines, and that the me-
chanical properties, and hence their susceptibility to shear,
changed during batch culture (Zhang et al., 1993).
Another approach to predict damage to cells from shear
was to identify a critical shear stress level, which, when
exceeded, caused cellular viability to decrease rapidly. Us-
ing reactor and viscometer studies (Tramper et al., 1986),
insect cell viability decreased markedly when exposed to 1
N m−2 for >1 h. A critical shear stress level was also de-
termined for baby hamster kidney (BHK) cells using the
parameters of cell morphology and loss of viability. A criti-
cal shear stress level of 0.80 to 1.0 N m−2 for an exposure
of 24 h in flow chamber studies was determined (Ludwig et
al., 1992). These studies calculated critical shear stress lev-
els from cellular viability data, not metabolic characteris-
tics, and the cells in these studies were cultured in media
containing fetal bovine serum.
The ability of the cell to respond quickly to the shear
environment has been demonstrated by a number of studies.
Gudi measured the activation of GTP-binding proteins in
endothelial cells within seconds of the onset of flow at the
shear stress of 1.0 N m−2 (Gudi et al., 1996). Enhanced
arachidonic acid metabolism and altered protein synthesis
were detected in endothelial cells by Nollert at 2.5 N m−2
(Nollert et al., 1991). The phosphorylation of heat shock
proteins in vascular endothelial cells was observed to in-
crease by threefold in 30 min after exposure to a shear stress
of 1.6 N m−2 (Li et al., 1996).
The effect of sublytic levels of shear stress on the me-
212 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 81, NO. 2, JANUARY 20, 2003
tabolism of various cell types was established by examining
the induction of the transcriptional activator c-fos. In a range
of cell types of human and animal origin, Ranjan (1996)
examined the induction of c-fos by shear stress. Exposed to
a shear stress of 2.5 N m−2 for 1 h, a consistent response in
the cell types HUVEC, HeLa, BAEC, and Chinese hamster
ovary (CHO) was found to occur within minutes after the
onset of flow. The minimum exposure time required for
shear stress to induce c-fos protein expression in HeLa cells
was 1 min. This demonstrated the rapid response of cells to
shear stimuli, and the potential for cell metabolism to be
altered in response to shear stress.
The mechanism of transduction of the mechanical stress
stimuli to an intracellular response including secondary
messenger generation, protein activation, and modulation of
gene expression is not fully understood. The involvement of
intracellular Ca++ in a membrane event linked to the acti-
vation of the phospholipase C pathway was proposed as a
possible mechanism for signal transduction (Levesque et al.,
1989; Wiesner et al., 1997). Intracellular Ca++ levels were
not observed to increase during shear experiments with en-
dothelial cells and arachidonic acid uptake (Nollert et al.,
1991), suggesting that other secondary messengers, such as
inositol-(1,4,5)-triphosphate (IP3) or other intracellular ions
(e.g., H+), may be involved in signal transduction. A spe-
cific GTP-binding protein (G protein) has been identified as
being activated by shear stress within 1 s of onset of flow in
primary human umbilical vein endothelial cells at a shear
stress of 1.0 N m−2 (Gudi et al., 1996). This represented one
of the earliest signal transduction events detected due to
onset of shear. This study identified the G protein family as
an important messenger in the shear stress response.
Despite the existing knowledge on the effect of lethal
shear forces on hybridomas and endothelial cells and the
emerging information on sublytic shear stress on cardiovas-
cular cell types, very little information regarding the effect
of sublytic shear stress on recombinant protein production is
available. The expression of a recombinant protein in a
BHK-21 cell line expressing ␤-galactosidase has been ex-
amined under shear stress conditions. When exposed to low
shear stress, productivity increased, with a maximum at 1.3
N m−2, after which productivity of ␤-D-galactosidase de-
creased markedly (Ludwig et al., 1993). The relationship
between protein production by the cell and shear stress has
been examined for various cell types, including endothelial
cells. Prostacyclin is a major metabolite produced by human
umbilical vein endothelial cells and the effect on its pro-
duction was determined while subjected to a steady shear
stress of 2.4 N m−2 (Frangos et al., 1988). The onset of flow
stimulated prostacyclin production linearly with increasing
shear stress.
The effect of shear stress on CHO cells in a flow chamber
experimental system has been studied previously (Shi-
ragami et al., 1992). The viability decreased with increasing
shear stress; however, there was no measurement of cell
specific metabolic rates and a recombinant cell line was not
examined. For renin-secreting recombinant CHO cells, Mo-
tobu et al. (1988) found that exposure at 0.02 N m−2 and
0.082 N m−2 for 24 h caused inhibition of cell growth,
whereas renin mRNA increased for subconfluent CHO
cells. However, there was no change in renin protein pro-
ductivity at the levels of shear stress examined.
In this study, recombinant CHO cells were exposed to a
steady laminar shear stress in a fluid flow chamber. The
recombinant CHO cell line, secreting human growth hor-
mone, was subjected to shear stress and the metabolic re-
sponse assessed by measuring various parameters, including
the cell specific productivity of the secreted recombinant
protein. A profile of the recombinant CHO cells’ response
to shear stress in relation to cell specific productivity and
uptake rates was established.
MATERIALS AND METHODS
Cell Culture
The cell line used was an attachment-dependent (in growth
medium) recombinant CHO line that was capable of secret-
ing human growth hormone (hGH) under the control of the
human metallothionein IIA (hMTIIA) promoter, an induc-
ible promoter system induced by the presence of heavy
metal ions (Friedman et al., 1989; Gray et al., 1990).
Ophthalmic glass plates (Schott–Garscow) were used for
culturing of cells for all shear experiments, providing a
uniform flat surface for cell attachment. Plates 290 mm × 26
mm × 6 mm (length × width × height) were used in the flow
chamber. To enhance cell attachment, the glass plates were
subjected to Na2CO3 treatment prior to inoculation (Rapport
et al., 1959).
Powdered media (Gibco BRL, Grand Island, NY) con-
sisted of a 1:1 mixture of Dulbecco’s modified Eagle’s me-
dium (DMEM) and Coon’s F12 medium. Upon hydration of
this medium, 2.5 g/L NaHCO3 was added to provide extra
buffering capacity, subsequently referred to as protein-free
(PF) media. Growth media (GM) consisted of PF media
supplemented with 10% (v/v) fetal bovine serum (CSL,
Ltd., Australia) and was used to propagate cells. ZnSO4
(May and Baker, Australia) was added at 80 ␮M to induce
the metallothionein promoter, controlling the expression of
the recombinant human growth hormone gene. Pluronic
F-68 (PE6800, BASF) was added to the PF medium at a
concentration of 0.2% w/v prior to filtration.
Cell Culture Assays
Glucose and lactic acid concentrations were determined us-
ing an analyzer (2300 Stat Plus, Yellow Springs Instru-
ments, Yellow Springs, OH). Human growth hormone
(hGH) assays were performed by reverse-phase high-
performance liquid chromatography (RP-HPLC). For se-
rum-free samples a Hamilton PRP-␣ 30 mm × 3.1 mm
column was used, and for serum-containing samples a Poly-
mer Labs PLRP-S 150 mm × 4.6 mm column was used,
KEANE ET AL.: EFFECT OF SHEAR STRESS ON A RECOMBINANT CHO CELL LINE
with both methods using a gradient elution with 0.10% (v/v)
trifluroacetic acid (TFA) in water and 0.08% (v/v) TFA in
acetonitrile.
Shear Stress Experimental Protocol
One control plate and one test plate were inoculated with the
cell suspension from one confluent T-75 flask into 200 mL
of 37°C GM + Zn. Both plates were gassed with a 5% CO2
(v/v in air) gas mixture (BOC Gases) and incubated at 37°C
for 3 days. Medium volume in both the test and control
plates was 200 mL, with sampling occurring at each time-
point from the bulk fluid, thus ensuring that the only dif-
ference between the control and test plate was the exposure
of the cells to the defined shear stress. The hold-up volume
of the shear stress experimental system was <5% of the total
medium volume and the maximum time required for a com-
plete recirculation of the fluid was <15 min, which is
equivalent to in excess of 15 cycles of the fluid in the
chamber to the bulk fluid between every sample point.
Static culture conditions were used to culture cells to con-
fluence and to serve as a no-shear control during shear
experiments. Once confluent plates were obtained 72 h after
inoculation, media changeover to SF + Zn + Pluronic F-68
media was performed and repeated 24 h later. Subsequent to
the second media change, control plates were cultured under
static conditions and test plates were subjected to shear
stress. During the experimental period, samples from con-
trol and test-plate supernatant were taken at 0 h, 4 h, 8 h, 12
h, 16 h, 24 h, 28 h, and 32 h, with time zero taken as the time
shear stress was initiated. Each experiment was repeated at
least three times.
Calculation of Dissolved Oxygen Uptake Rate
The dissolved oxygen (DO) uptake rate of the cell line
CB515 was measured in PF + Zn + Pluronic F-68 media by
monitoring the decrease in DO levels of the culture. In
addition to the measurement of the cells’ requirement for
oxygen, the exit DO in the chamber during the shear stress
experiments at various shear stress levels was measured.
Measurement of Cell Distribution in the
Flow Chamber
The cell density in the flow chamber was measured by
direct visualization. The cell counts were performed at ev-
ery timepoint and consisted of six separate positions across
the width of the chamber (1 to 6), each consisting of four
cell counts, and at five positions down the length of the
chamber (Fig. 1, areas A–E). Cell viability was determined
at the conclusion of each experiment by direct visualization
following the same method as for cell distribution. Cell
viability was determined by uptake of trypan blue by the
cell. Control plates and test-plate cell viability were deter-
mined using the same method. Trypan blue was introduced
213
 the test plate was housed within the flow chamber. The
shear stresses experienced by the cells were related to the
velocity of the fluid flow down the chamber. Dissolved
oxygen was monitored and temperature and pH were moni-
tored and controlled at 37°C and 7.3, respectively.

Description of the Flow Chamber

The flow chamber used parallel plate, channel flow geom-
etry, to provide a steady, uniform laminar flow. This was
achieved by designing the channel with a rectangular cross-
sectional, the height being much less than the chamber
length or width. This simple channel geometry with a rect-
angular cross-section provided well-defined fluid conditions
(White, 1974). For newtonian fluids, the velocity profile in
chambers of this geometry have the simple parabolic form:

u=
1 dp
∗ ∗
2␮ dx
h2 2
4
−y 冋 册 (1)

where dp/dx is the pressure gradient along the chamber, u is

the fluid velocity, ␮ is the fluid viscosity, y is the distance
from the centerline, and h is the height of the chamber. The
velocity of the fluid down the chamber has the simple para-
bolic form, with the maximum velocity at the centerline
Figure 1. Representation of the overall shear stress experimental system, (i.e., at y ⳱; 0). The maximum velocity is described as:
showing both the static (A) and shear stress (B) culture methods, and an
overhead view of the flow chamber (C), showing the positions of cell 1 dp
counts A to E and 1 to 6, and the fluid entry and exit points. DO probe umax = − ∗ ∗ h2 (2)
8␮ dx
refers to dissolved oxygen probe.
where dp/dx is the pressure gradient along the channel. The
mean velocity is described as:
to the culture surface in situ after washing the cells three
times with phosphate-buffered saline (PBS). 2
u = umax (3)
3
1 dp
Calculation of Relative Metabolic Values =− ∗ ∗ h2 (4)
12␮ dx
Dividing the cell specific metabolic rates for shear stress
data by the control data and expressing the result in per- The wall shear stress is related to the pressure gradient
centage terms calculated relative to metabolic values. This along the channel and the height of the channel, and is
method was necessary as each shear experiment contained described as:
eight sample points for both test and control plates and were h dp
performed at least three times, resulting in at least 48 sample ␶w = − ∗ (5)
2 dx
points for each shear stress level examined. This method
ensured accurate comparisons of temporally separated ex- For these relationships to be correct the fluid flow must
periments and allowed for easy comparison of a number of be laminar, two-dimensional, and fully developed. To de-
experiments. When the addition of a sequence of numbers termine whether the fluid flow is laminar, the Reynolds
with their associated error was required, a weighted mean number (Re) must be <2000, and can be determined from
was calculated and used to calculate the average relative the following relationship:
metabolic rates.
共u ∗ 共h兲兲
Re = (6)
v
Culture Conditions
where v is the kinematic viscosity (i.e., ␮/␳), and ␳ is the
Static culture conditions were used to both culture cells to fluid density.
confluence and to serve as a no-shear control during experi- For the fluid flow to be considered two-dimensional, the
ments. It consisted of glass plates housed in stationary glass width of the chamber must be much greater than the height
bottles incubated at 37°C with 5% CO2 in air headspace of the chamber. For the fluid flow to be considered fully
gassing. Shear culture conditions were experienced when developed, the length of the fluid entrance to the channel

214 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 81, NO. 2, JANUARY 20, 2003
must be small compared with the entire length of the chan-
nel (Levesque and Nerem, 1985). The flow chamber dimen-
sions were based on rectangular cross-sectional geometry,
as described in the literature (Bussolari and Dewey 1982;
Goldstein and DiMilla, 1996; Levesque and Nerem, 1985;
Ruel et al., 1995; Zhang et al., 1993), and satisfied all the
criteria for laminar two-dimensional laminar flow (Table I).
The channel in this study was longer than other reported
channels (270 mm × 23 mm × 0.3 mm with an entrance
length of 2 mm) to ensure that large numbers of cells were
exposed to the shear stress. Satisfying the criteria for rect-
angular channel flow, the parabolic form of fluid flow ex-
isted over at least 98% of the channel length (Levesque and
Nerem, 1985; Ruel et al., 1995).
The chamber was fabricated from 12-mm-thick polycar-
bonate (Lexan, GE Structural Products) and consisted of
two parts. The lower part contained the media inlet and
outlet ports and the machine-milled and polished channel on
which the glass plate with attached cells were positioned.
The upper part secured the glass plate in place. All materials
were sterilized by autoclaving (121°C 30 min), except for
the flow chamber where a liquid sterilization method (Oxo-
nia Active–Ecolabs, Australia) was used.
RESULTS
Parameters of the Flow Chamber
Specific parameters of importance in determining the fluid
flow through the flow chamber were measured and calcu-
lated and are presented in Table I. The error involved in the
experimental system was quantified by performing a series
of experiments in the flow chamber system, without cells
and hGH introduced from an external source. Replicate
samples were taken and assayed and the systematic error of
the experimental system was estimated (Table II). The error
values calculated were considered to be in an acceptable
Table I. Summary table of Reynolds numbers and residence times in the
flow chambers at various shear stress levels.
Shear stress Chamber height Reynolds number Mean residence
(␶w) Nm−2 (h) m (Re × 10−3) time (min)
Table II. Calculated standard errors as a percentage value for glucose
and hGH assay results for samples taken from the shear stress experimental
system.
Glucose Glucose hGH hGH
control plate test plate control plate test plate
3.77% ± 2.17 5.62% ± 3.97 3.34% ± 0.59 3.53% ± 1.98
Data represent the average of triplicate experiments, conducted accord-
ing to the shear stress experimental protocol.
range considering the complicated nature of the experimen-
tal apparatus.
Dissolved Oxygen Profile of the Flow Chamber
During Shear Stress Experiments
The oxygen uptake rate was measured by monitoring the
decrease in dissolved oxygen across the flow chamber dur-
ing culture and was calculated to be 0.6 mM/1010 cells/h.
This value was used in experiments to ensure that oxygen
was not limiting in the flow chamber during shear stress
experiments. To examine the oxygen uptake within the
chamber a series of experiments monitoring the culture pa-
rameters over a range of shear stress levels in various media
formulations were performed. The geometry of the chamber
included a low ratio of channel height to channel width that
ensured boundary layer formation was minimized, as well
as maintaining laminar flow down the chamber. The dis-
solved oxygen profile from flow exiting the chamber was
not below 80% of DO air saturation and was therefore con-
sidered nonlimiting.
To ensure glucose was not limiting in the flow chamber
during shear stress experiments, the glucose consumption
along the length of the chamber was calculated. Calculating
from a minimum glucose concentration of 1.5 g/L entering
the flow chamber and an exiting concentration of 0.5 g/L,
and a glucose uptake rate of 0.45 mg/106 cells per hour, a
maximum residence time for both chambers was calculated.
The maximum residence time for fluid in the flow chamber
before glucose became limiting was calculated to be 48.2
min. This was well in excess of the maximum residence
times experienced for the range of shear stress levels tested
(see Table I).

0.80 3 × 10−4 100.0 0.16

0.40 3 × 10−4 50.0 0.32 Cell Viability During Shear Stress Experiments
0.10 8 × 10−4 96.8 0.40
0.05 3 × 10−4 6.2 2.57 Direct measurement of cell viability at the completion of the
0.03 8 × 10−4 30.0 1.20
0.01 3 × 10−4 1.2 12.85
shear stress experiments were performed using the trypan
0.005 8 × 10−4 4.4 7.10 blue dye exclusion method. Culture viabilities for experi-
ments with media containing Pluronic F-68 were above 90
Median viscosity was measured at 1.10 × 10−3 kg−1 m−1, although media ± 7% for all shear stress levels tested. Cell viability levels
density was assumed to be that of water. Mean residence times refer to the
on the control and test plates were within the range of the
average time required for one complete pass through the flow chamber and
was an important criteria when considering nutritional limitations. Two error of the assay, and both were >90% and therefore con-
chamber heights were used to generate the range of shear stress forces sidered to be the same for the range of shear stress levels
examined. tested.

KEANE ET AL.: EFFECT OF SHEAR STRESS ON A RECOMBINANT CHO CELL LINE 215
Cell Distribution in Flow Chamber
To assess the distribution of cells in the flow chamber dur-
ing shear stress experiments, cell counts were performed
across the width and length of the chamber. At all shear
stress levels tested, the attached cell density for the duration
of the experiment was measured by performing four cell
counts across the width of the chamber at each of the five
positions down the length (A to E) of the chamber, and the
positions numbered 1 to 6, as indicated in Figure 1. This
resulted in a total of 120 cell counts performed for each
timepoint, for eight consecutive sampling timepoints. No
change in cell density relative to static control was observed
and cell viability was greater than 81 ± 4.5% in static or
shear stress culture at all positions from 0 to 80 h. For all
shear stress levels tested there was no morphological change
to the cells. The results show that the attached cell density
(ACD) levels in the flow chamber were maintained at high
confluence, with a slow decrease, equal to that of the static
control with uniform cell distribution over the entire plate.
Cell densities were maintained at 2.5 to 3.0 × 105 cells/cm2
for both the test and static plate, with no washout of cells
occurring at the entry or exit points from the flow chamber.
Shear Stress Experiments
A series of shear experiments were performed to develop an
accurate and reproducible method for the assessment of cel-
Figure 2. Cell specific hGH productivity (␮g/106 cells per hour) for control plate sand test plates at 0.10 N m−2 shear stress in protein-free (PF) medium
with and without Pluronic F-68 addition. Note the absence of hGH productivity in PF medium without Pluronic F-68.
216 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 81, NO. 2, JANUARY 20, 2003
lular metabolic rates under varying laminar shear stress con-
ditions. The principle parameters examined were cell spe-
cific recombinant protein productivity, attached cell density,
and cell specific glucose uptake rate. Initially, an experi-
ment comparing PF medium and PF + Pluronic F-68 me-
dium at 0.1 N m−2 shear stress with no shear control was
performed; the cell specific hGH productivity values are
presented in Figure 2.
Initially, a series of experiments examining the effect of
shear stress on the recombinant protein production of the
cells were performed in PF + Zn medium. At 0.10 N m−2
shear stress, no hGH production was detected from the cells,
whereas glucose was taken up by the cell at a rate similar to
the static control. To establish whether hGH production
could be recovered by the addition of shear protectants,
experiments were performed under the same shear condi-
tions with the addition of either 1% FBS (v/v) or 0.2%
Pluronic F-68 (w/v). The results of these experiments are
presented in Figure 3. Shear stress was shown to have a
major inhibitory effect on recombinant protein production
in medium without shear protectants. The addition of shear
protectants Pluronic F-68 or 1% FBS enabled restoration of
recombinant protein productivity, but at the expense of in-
creased glucose consumption. Protection of the cells from
the effect of shear stress was not attributed to the production
of hGH, because the hGH molecule in cell culture has no
Figure 3. Average metabolic rates for relative attached cell density (rela-
tive ACD), relative hGH productivity (relative qhGH), and relative glucose
Figure 4. Average relative values for metabolic rates measured during
uptake (relative qglucose). Note the absence of hGH productivity in protein-
shear stress experiments in protein-free + zinc + Pluronic F-68 media.
free medium without shear protectants. P.F., protein-free medium; Pl,
These values represent averages from eight sample points from at least
Pluronic F-68; FBS, fetal bovine serum.
three experiments from each shear stress condition. Errors were calculated
from the SEM for each set of experiments.
known shear protectant properties. Similarly, the hGH mol-
ecule was considered to not be preferentially degraded when
subjected to the level of shear stress examined in this study with a concomitant decrease in recombinant hGH produc-
(Maa and Hsu, 1996, 1997). tion represented a major shift in the yield of recombinant
Following this experiment, a series of experiments over a protein from glucose when the cell was subjected to increas-
range of shear stress levels was performed. The attached cell ing shear stress. The relationship between the yield of hGH
density, glucose uptake, and hGH and lactate productivity from glucose (Rel qhGH/Rel qglucose) when subjected to
of the cells were calculated over the shear stress range of shear stress is presented in Figure 5.
0.005 to 0.80 N m−2 in PF medium supplemented with Also of major interest was the change in the yield of
Pluronic F-68. Each shear stress experiment consisted of lactate from glucose with changing shear conditions. These
eight sample points, for which cell specific metabolic rates experiments have shown that, when the recombinant CHO
were calculated and the average of these points, with asso-
ciated error, was determined. For each shear stress exam-
ined, experiments were performed in triplicate, enabling the
calculation of an average relative metabolic rate at each
shear stress examined. The results from this series of ex-
periments are presented in Figure 4 and represent a sum-
mary of 25 experiments, each with test and control data
sampled at eight timepoints over a 32-h culture period.
These results show the relationship between increasing
shear stress and the relative metabolic rates of the cell for
the recombinant CHO cell line producing recombinant hu-
man growth hormone (hGH).
During these experiments, the average ACD was main-
tained close to 100%. As shear stress increased, cell specific
glucose uptake rate increased and hGH and lactate produc-
tivity decreased. This showed a major shift in cell metabo-
lism, with increased uptake and decreased expression due to
Figure 5. The relative yield of hGH from glucose from cells exposed to
exposure to shear stress. This result was also of interest in shear stresses ranging from 0.005 to 0.80 N m−2. A Rel qhGH/Rel qglucose
light of earlier results of shear experiments in PF + Zn value of 1 represents the identical yield as that obtained from the static
media (see Fig. 3). The increased glucose uptake by the cell control. Error bars derived from the SEM of three experiments.

KEANE ET AL.: EFFECT OF SHEAR STRESS ON A RECOMBINANT CHO CELL LINE 217

Figure 6. The relative yield of lactate from glucose from cells exposed to
shear stress ranging from 0.005 to 0.80 N m−2. A Rel qlactate/Rel qglucose
value of 1 represents a yield identical to that obtained from the static
control. Error bars derived from the SEM of triplicate experiments.
cell line was subjected to shear stress, the yield of lactate
from glucose changed. The relationship between the yield of
lactate from glucose (Rel qlactate/Rel qglucose), when sub-
jected to shear stress, is presented in Figure 6.
DISCUSSION
An experimental system was developed that allowed ap-
proximately 2 × 107 cells to be subjected to a range of
defined shear stress in protein free medium for an extended
culture period, enabling cell specific productivities and up-
take rates to be calculated. To ensure calculations were
accurate, multiple cells counts were performed for both test
and control plates for each sampling point. To increase ac-
curacy, eight sample points were performed approximately
every 4 h, over the duration of each shear experiment, each
performed in triplicate. By using these techniques, the num-
ber of data points used in each calculation of cell specific
metabolism was increased and, subsequently, accuracy and
reproducibility were maximized.
Oxygen Uptake Rate (OUR)
To ensure oxygen was not limiting in the flow chamber the
OUR was measured, and to confirm these calculations, in-
line DO monitoring was performed pre- and post-chamber.
The OUR was estimated to be 0.6 mM/1010 cells per hour.
This value was in good agreement with the OUR calculated
for a hybridoma batch culture of 2.0 mM/1010 cells per hour
(Zupke and Stephanopoulos, 1995), but is lower than that
derived for the same CHO cell line in microcarrier airlift
culture of 0.5 ␮M/106 cells per hour (Lovrecz and Gray,
1994). It was assumed that the OUR of the cells was con-
stant in the range of 30% to 100% DO (air saturation) for a
stable, nondividing cell population, and that the concentra-
tion of oxygen in cell culture media at 100% air saturation
218 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 81, NO. 2, JANUARY 20, 2003
was 0.25 mM. The OUR of 0.6 mM/1010 cells per hour was
used in the validation of the flow chamber to ensure that at
no time was the culture oxygen limiting. For the range of
shear stress levels tested in this series of experiments the
minimum time required for the DO level to fall below 30%
DO was calculated to be 24.2 min. This value was well in
excess of the maximum residence times experienced in the
chamber of 12.85 min. In addition, DO profiles at the entry
and exit points of the chamber were measured. Entry DO
levels were always in excess of 80% DO, and minimum exit
DO levels were 80% as measured in the bulk flow imme-
diately after exit to the chamber.
CHO Response to Shear Stress
In this study it has been demonstrated that, at sublytic shear
stress levels, recombinant protein productivity could be re-
covered with the addition of shear protectants FBS and
Pluronic F-68 in the media. The effect of shear stress on the
rhGH molecule was considered insignificant (Maa and Hsu,
1996, 1997). The recovery of hGH productivity with the
addition of shear protectants suggests that the cessation of
hGH productivity in PF medium was due to the effect of
shear stress on the cell. There was no change in cell viability
of cells subjected to shear stress as compared with the static
control, and no morphological change of the cells was ob-
served.
In PF media supplemented with the shear protectant
Pluronic F-68, relative cell specific hGH productivity de-
creased rapidly from 100.4% at 0.005 N m−2 to 65.8% at
0.10 N m−2. The decrease continued, although at a slower
rate, from 0.10 to 0.80 N m−2, where relative hGH produc-
tivity decreased to 49.7% of the static control. This repre-
sented a substantial decrease in recombinant protein pro-
ductivity due to shear stress. Throughout these changes in
relative hGH productivity, the relative ACD was steady,
although there were small decreases in ACD at 0.40 N m−2
and 0.80 N m−2.
An increase in the relative qglucose of the cell occurred
over the range of shear stresses tested. This is important as
it illustrates an increase in glucose uptake by the cell that
was associated with increasing shear stress. With increasing
shear stress from 0.005 to 0.80 N m−2, the relative qlactate of
the cell was observed to decrease in similar magnitude as
that for relative qhGH. At 0.005 N m−2, the relative qlactate
was calculated to be 92.4%, which steadily decreased to
50.4% at 0.80 N m−2.
The decrease in hGH and lactate production and increase
in qglucose also indicated a change in the yield of both hGH
and lactate from glucose that was related to the shear stress
environment. Even at the lowest shear stress of 0.005 N m−2
the relative yield of hGH and relative yield of lactate was
lower than that of the static control (0.74 and 0.68, respec-
tively), indicating that, even at this low shear stress, in PF
media supplemented with Pluronic F-68 the metabolism of
the cell is influenced by the immediate shear stress envi-
ronment.
 The decrease in yield of lactate from glucose was of
interest as this value is normally high for continuous cell
lines. The change in yield also showed a decrease in pro-
duction of recombinant protein due to the shear environ-
ment. These results were obtained from cells that, under a
shear stress environment, were showing no morphological
changes or observed changes in confluence. Due to the large
media volume to cell number ratio, the culture contained
excess glucose, as was confirmed from glucose assays at the
completion of experiments. It was assumed that other nu-
trients utilized at lower rates than glucose would also be in
excess due to the large media volume to cell ratio and the
culture time of approximately 32 h. In this environment, an
increased glucose uptake rate and decreased lactate produc-
tivity was observed with increasing shear stress. This was of
interest because it had previously been reported that glucose
uptake rates in cultured animal cells were typically sub-
strate-concentration-dependent and that increased uptake
was associated with an increase in substrate concentration
(Doverskog et al., 1997; Zeng and Deckwer, 1995). In ad-
dition, uptake rates exceeding the cellular need for precur-
sors and energy resulted in increase flux of glycolysis, in
turn leading to overflow metabolism and the formation of
excess byproducts such as lactate (Doverskog et al., 1997;
Linz et al., 1997). In transformed mammalian cells, the
large majority of glucose that proceeds down the glycolytic
pathway is converted to lactate, with very little glucose-
derived carbon entering the TCA cycle (Neermann and
Wagner, 1996). This is in contrast to the material balance
data findings with hybridomas (Xie and Wang, 1996) show-
ing that the majority of ATP produced (60% to 76%) was
obtained from glucose, and that the TCA cycle provided
51% to 68% of the total ATP production. The amount of
glucose entering the TCA cycle was dependent on the nu-
tritional environment of the cell and glucose was used less
effectively by the cell when present in excess (Xie and
Wang, 1996). In this study, the cells were in an environment
of excess glucose, which, from the current literature, would
result in overflow metabolism and the accumulation of lac-
tate in the medium. However, in the presence of a defined
shear stress, a cellular response of increased glucose uptake
and reduced lactate productivity was observed. This indi-
cates a cellular response to shear stress culminating in re-
duced glucose- derived carbon converted to lactate, and it
appears that this response increased with increasing shear
stress.
In this study, the uptake of glucose from the media and
formation of hGH and lactate was carefully monitored.
Studies have been performed assessing animal cell metabo-
lism by metabolic flux analysis (Nyberg et al., 1999) and by
material balance analysis (Xie and Wang, 1996). These
techniques could be applied to determine metabolic changes
occurring in cells exposed to shear stress. Two-dimensional
electrophoresis could also be used to examine the full pro-
tein profile of cells exposed to shear stress and to identify a
specific protein or group of proteins induced by exposure to
shear stress.
KEANE ET AL.: EFFECT OF SHEAR STRESS ON A RECOMBINANT CHO CELL LINE
CONCLUSIONS
A shear stress experimental system capable of measuring
cell specific recombinant protein productivity accurately
and reproducibly was developed. A flow chamber was con-
structed that was able to expose attached cells to steady
laminar shear in controlled culture conditions. Shear stress
was shown to have a major affect on recombinant protein
productivity in protein-free media. At a shear stress of 0.10
N m−2, hGH production ceased. The inhibition of recombi-
nant protein productivity by the cell under shear stress was
relieved with the addition of the shear protectants Pluronic
F-68 or 1% (v/v) FBS to the media.
With increasing shear stress, the cell specific hGH pro-
ductivity of the cell decreased from 100.4% at 0.005 N m−2
to 49.7% at 0.80 N m−2 relative to the static control. These
results were obtained with cells of high confluence, showing
no morphological changes, in protein-free medium with
Pluronic F-68 (0.2% w/v) added as a shear protectant. These
data show that low levels of shear stress had a negative
impact on recombinant protein productivity for this cell
line. An increase in glucose uptake by the cell from 114.8%
at 0.01 N m−2 to 141.8% at 0.80 N m−2 also occurred over
this range of shear stresses. Concomitant with the increase
in glucose uptake was a decrease in lactate productivity by
the cell. At 0.005 N m−2, lactate productivity decreased
from 92.4% to 50.4% at 0.80 N m−2 relative to the static
control. The change in glucose uptake rate and recombinant
protein (hGH) and lactate productivity indicated a change in
the yield of cellular products from glucose, and that this
change was shear-related. This represents a shift in glucose
metabolism of the cell associated with increasing shear
stress. These results highlight the importance of low-shear-
stress operating procedures and design parameters for ani-
mal-cell bioreactors for optimal recombinant protein pro-
duction.
The authors thank Dr. Peter Phillips and Neil Kitchen for their
invaluable help with the fermentation data.
References
Al-Rubeai M, Emery AN, Chalder S, Goldman MH. 1993. A flow cyto-
metric study of hydrodynamic damage to mammalian cells. J Biotech-
nol 31:161–177.
Al-Rubeai M, Goldman SMH, Emery AN. 1995a. Death mechanisms of
animal cells in conditions of intensive agitation. Biotechnol Bioeng
45:463–472.
Al-Rubeai M, Singh RP, Emery AN, Zhang Z. 1995b. Cell cycle and cell
size dependence of susceptibility to hydrodynamic forces. Biotechnol
Bioeng 46:88–92.
Bavarian F, Fan LS, Chalmers JJ. 1991. Microscopic visualization of insect
cell–bubble interactions. I: Rising bubbles, air–medium interface, and
the foam layer. Biotechnol Progr 7:140–150.
Born C, Zhang Z, Al-Rubeai M, Thomas CR. 1992. Estimation of disrup-
tion of animal cells by laminar shear stress. Biotechnol Bioeng 40:
1004–1110.
Bussolari SR, Dewey CF. 1982. Apparatus for subjecting living cells to
fluid shear stress. Rev Sci Instrum 53:1851–1854
219
Cherry RS, Hulle CT. 1992. Cell death in the films of bursting bubbles.
Biotechnol Progr 8:11–18.
Doverskog M, Ljunggren J, Ohman L, Haggstrom L. 1997. Physiology of
cultured animal cells. J Biotechnol 59:103–115.
Frangos JA, McIntire LV, Eskin SG. 1988. Shear stress induced stimula-
tion of mammalian cell metabolism. Biotechnol Bioeng 32:
1053–1060.
Friedman JS, Cofer CL, Anderson CL, Kushner JA, Gray PP, Chapman
GE, Stuart MC, Lazarus L, Shine J, Kushner PJ. 1989. High expression
in mammalian cells without amplification. Biotechnology 7:359–363.
Garcia-Briones M, Chalmers JJ. 1990. Cell–bubble interactions. Mecha-
nisms of suspended cell damage. Ann NY Acad Sci 219–229.
Garcia-Briones M, Chalmers JJ. 1994. Flow parameters associated with
hydrodynamic cell injury. Biotechnol Bioeng 44:1089–1098.
Goldstein AS, DiMilla PA. 1996. Application of fluid mechanic and kinetic
models to characterize mammalian cell detachment in a radial-flow
chamber. Biotechnol Bioeng 45:616–629.
Gray PP, Marsden WL, Crowley J, Gebert C, Jirasripongpun K, Lovrecz G,
Mitchell C, Mahler SM. 1990. Mammalian cell culture at the Univer-
sity of New South Wales. Austral J Biotechnol 4:121–123.
Gregoriades N, Clay J, Ma N, Koelling K, Chalmers JJ. 2000. Cell damage
of microcarrier cultures as a function of local energy dissipation cre-
ated by a rapid extensional flow. Biotechnol Bioeng 69:171–182
Gudi SRP, Clark CB, Frangos JA. 1996. Fluid flow rapidly activates G
proteins in human endothelial cells. Involvement of G proteins in
mechanochemical signal transduction. Circ Res 79:834–839.
Handa A, Emery AN, Spier RE. 1987. On the evaluation of gas–liquid
interfacial effects on hybridoma viability in bubble column bioreac-
tors. Dev Biol Stand 66:241–253.
Handa-Corrigan A, Emery AN, Spier RE. 1989. Effect of gas–liquid in-
terfaces on the growth of suspended mammalian cells: Mechanisms of
cell damage by bubbles. Enzyme Microb Technol 11:230–235.
Jordan M, Sucker H, Einsele A, Widmer F, Eppenberger HM. 1994. In-
teractions between animal cells and gas bubbles: The influence of
serum and Pluronic F68 on the physical properties of the bubble sur-
face. Biotechnol Bioeng 43:446–454.
Kretzmer G, Jammrich U, Schugerl K. 1990. In: Spier RE, Griffith JB,
editors. Advances in animal cell biology and technology for biopro-
cesses. London: Butterworths. p 172–174.
Kretzmer G, Schugerl K. 1991. Response of mammalian cells to shear
stress. Appl Microbiol Biotechnol 34:613–616.
Levesque MJ, Nerem RM. 1985. The elongation and orientation of cultured
endothelial cells in response to shear stress. J Biomech Eng 107:
341–347.
Levesque MJ, Sprague EA, Schwartz CJ, Nerem RM. 1989. The influence
of shear stress on cultured vascular endothelial cells: The stress re-
sponse of an anchorage-dependent mammalian cell. Biotechnol Progr
5:1–8.
Li S, Piotrowicz RS, Levin EG, Shyy YJ, Chien S. 1996. Fluid shear stress
induces the phosphorylation of small heat shock proteins in vascular
endothelial cells. Am J Physiol (Cell Physiol) 40:C994–C1000.
Linz M, Zeng A, Wagner R, Deckwer W. 1997. Stoichiometry, kinetics
and regulation of glucose and amino acid metabolism of a recombinant
cell line in batch and continuous cultures. Biotechnol Progr 13:
453–463.
Lovrecz G, Gray PP. 1994. Use of on-line gas analysis to monitor recom-
binant mammalian cell cultures. Cytotechnology 14:167–175.
Ludwig A, Tomeczkowski J, Kretzmer G. 1992. Influence of shear stress
on adherent mammalian cells during division. Biotechnol Lett 14:
881–884.
Ludwig A, Weideman R, Tomeczkowski J, Kretzmer G. 1993. Surface
coating — No improvement of shear stress resistance of anchorage
dependent mammalian cells. Biotechnol Techn 7:411–416.
220 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 81, NO. 2, JANUARY 20, 2003
Maa Y-F, Hsu C. 1996. Effect of high shear on proteins. Biotechnol Bioeng
51:458–465.
Maa Y-F, Hsu C. 1997. Protein denaturation by combined effect of shear
and air–liquid interface. Biotechnol Bioeng 54:503–512.
McDowell C, Papoutsakis ET. 1998. Increased agitation intensity increases
CD13 receptor surface content and mRNA levels, and alters the me-
tabolism of HL60 cells cultured in stirred tank bioreactors. Biotechnol
Bioeng 60:239–250.
Michaels JD, Mallik AK, Papoutsakis ET. 1996. Sparging and agitation-
induced injury of cultured animal cells: Do cell-to-bubble interactions
in the bulk liquid injure cells? Biotechnol Bioeng 51:399–409.
Motobu M, Wang P, Matsumura M. 1998. Effect of shear stress on re-
combinant Chinese hamster ovary cells. J Ferment Bioeng 85:
190–195.
Neermann J, Wagner R. 1996. Comparative analysis of glucose and glu-
tamine metabolism in transformed mammalian cell lines, insect and
primary liver cells. J Cell Physiol 166:152–169.
Nollert MU, Diamond SL, McIntire LV. 1991. Hydrodynamic shear stress
and mass transport modulation of endothelial cell metabolism. Bio-
technol Bioeng 38:588–602.
Nyberg GB, Balcarcel RR, Follstad BD, Stephanopoulos G, Wang DIC.
1999. Metabolism of peptide amino acids by Chinese hamster ovary
cells grown in a complex medium. Biotechnol Bioeng 62:324–335.
Peterson JF, McIntire LV, Papoutsakis ET. 1988. Shear sensitivity of cul-
tured hybridoma cells (CRL-8018) depends on the mode of growth,
culture age and metabolic concentration. J Biotechnol 7:229–246.
Ranjan V, Waterbury R, Xiao Z, Diamond SL. 1996. Fluid shear stress
induction of the transcriptional activator c-fos in human and bovine
cells, HeLa and Chinese hamster ovary cells. Biotechnol Bioeng 49:
383–390.
Rapport C, Poole JP, Rapport HP. 1959. Growth of mammalian cells in
synthetic medium. Exper Cell Res 20:465–479.
Ruel J, Lemay J, Dumas G, Doillon, Charara J. 1995. Development of a
parallel plate flow chamber for studying cell behavior under pulsatile
flow. ASAIO J 41:876–883.
Shiragami N, Oda H, Honda H, Unno H. 1992. Stimulation of animal cell
metabolism by shear stress. In: Murakami H, et al., editors. Animal
cell technology: Basic and applied aspects. The Netherlands: Kluwer
Academic Publishers. p 271–277.
Tramper J, Williams JB, Joustra D, Vlak JM. 1986. Shear sensitivity of
insect cells in suspension. Enzyme Microb Technol 8:33–36.
Trinh K, Garcia-Briones M, Hink F, Chalmers JJ. 1994. Quantification of
damage to suspended insect cells as a result of bubble rupture. Bio-
technol Bioeng 43:37–45.
White FM. 1974. Viscous fluid flows. New York: McGraw-Hill.
Weisner TF, Berk BC, Nerem RM. 1997. A mathematical model of the
cytosolic-free calcium response in endothelial cells to fluid stress. Proc
Natl Acad Sci 94:3726–3731.
Xie L, Wang DIC. 1996. Energy metabolism and ATP balance in animal
cell cultivation using a stoichiometrically based reaction network. Bio-
technol Bioeng 52:591–601.
Zeng AP, Deckwer WD. 1995. Mathematical modelling and analysis of
glucose and glutamine utilization and regulation in cultures of con-
tinuous mammalian cells. Biotechnol Bioeng 47:334–346.
Zhang Z, Al-Rubeai M, Thomas CR. 1993. Comparison of the fragilities of
several animal cell lines. Biotechnol Techn 7:177–182.
Ziegelstein RC, Cheng L, Capogrossi MC. 1992. Flow dependent cytosolic
acidification of vascular endothelial cells. Science 258:656–659.
Zupke C, Stephanopoulos G. 1995. Intracellular flux analysis in hybrido-
mas using mass balances and in vitro 13C NMR. Biotechnol Bioeng
45:292–303.