Vol. 92, Issue 2, 691-700, February 2002
Expression of renal cell protein markers is dependent on
initial mechanical culture conditions
Nancy L.
Cowger1,
Edmund
Benes1,
Patricia L.
Allen1, and
Timothy G.
Hammond1,2,3
1 Nephrology Section, Department of Medicine, and
2 Tulane/VA Environmental Astrobiology Center, Center for
Bioenvironmental Research, Tulane University Medical Center, New
Orleans 70112; and 3 Veterans Affairs Medical Center, New
Orleans, Louisiana 70146
 |
ABSTRACT |
The rotating wall vessel is
optimized for suspension culture, with laminar flow and adequate
nutrient delivery, but minimal shear. However, higher shears may occur
in vivo. During rotating wall vessel cultivation of human renal cells,
size and density of glass-coated microcarrier beads were changed to
modulate initial shear. Renal-specific proteins were assayed after 2 days. Flow cytometry antibody binding analysis of vitamin D receptor
demonstrated peak expression at intermediate shears, with 30%
reduction outside this range. Activity of cathepsin C showed the
inverse pattern, lowest at midshear, with twofold increases at either
extreme. Dipeptidyl-peptidase IV had no shear dependence, suggesting
that the other results are specific, not universal, changes in membrane trafficking or protein synthesis. On addition of dextran, which changes
medium density and viscosity but not shear, vitamin D receptor assay
showed no differences from controls. Neither cell cycle,
apoptosis/necrosis indexes, nor lactate dehydrogenase release varied between experiments, confirming that the changes are primary, not secondary to cell cycling or membrane damage. This study provides direct evidence that mechanical culture conditions modulate protein expression in suspension culture.
suspension culture; vitamin D receptor; tissue engineering; cell
culture
| |
INTRODUCTION |
MOST DIFFERENTIATED
CELLS from diverse tissue sources lose their specialized features
and "dedifferentiate" when grown as a two-dimensional monolayer
culture (1, 13, 35). This greatly limits the utility of
these cultures for the study of tissue-specific receptors, metabolic
pathways, signal transduction mechanisms, and/or nuclear
transcriptional events. The classic cell biology approach to overcome
this problem and maintain many differentiated features of cells in
culture is to grow the cells in suspension culture (1).
Suspension culture can be performed in a diverse array of culture
vessels, often including beads or other matrixes to provide support for
adherent cells (1, 13, 35, 37).
To optimize mechanical culture conditions in suspension culture
vessels, the challenge is to minimize shear and turbulence, while
keeping the cell aggregates in suspension and preserving mass transport
of both nutrients and gases (10, 11, 13, 22, 27, 37).
These requirements are embodied in the rotating wall vessel (RWV): a
horizontally rotating cylindrical culture vessel with a coaxial tubular
oxygenator (13, 37).
The RWV has characteristic features that determine its utility. First,
the culture medium is gently mixed by rotation, avoiding the necessity
for stirring vanes that can damage cells with local turbulence. Second,
fluid flow is laminar. The outer vessel wall, the coaxial oxygenator,
and the fluid inside all rotate at the same slow (10-60 rpm) rate,
avoiding shear from differential rotation. Third, there are no bubbles
in the vessel because oxygenation is by diffusion across a silicone
membrane (13, 22, 37). Interaction with bubbles is
typically a potent source of cell damage (3, 37). In the
RWV, three-dimensional assembly and colocation of dissimilar cell types
are accommodated. The result is spheroid formation with increased
cell-cell and cell-matrix interactions and tissue differentiation
(4, 5, 13, 35). Aggregates of human prostate tumor cells,
for example, had stronger staining for select cytoskeletal proteins,
suggesting a more differentiated population than control cells grown in
spinner flask or as a monolayer (4).
The question now becomes whether the engineering optimization of shear
to extremely low levels reflects biological optimization as well.
Several lines of evidence suggest not only that many mammalian cells
live in a milieu of shear but that shear stress is necessary for normal
structure and function of these cells (12, 20, 25, 28,
29). Perhaps the best studied example of shear dependence of
gene and protein expression is the serial clusters of genes expressed
during exposure of vascular endothelial cells to shear stresses that
mimic flow in blood vessels (20, 25, 28, 29).
Kidney cells in vivo, especially renal proximal tubular cells, also
experience a shear stress environment far larger than the levels
present during optimized RWV culture (14). Shear forces
vary from ~1 to 5 dyn/cm2 over the normal range of flow
rates in the proximal tubule (14), whereas shear is
estimated at 0.5 dyn/cm2 or less for the RWV
(13). Three recent studies from our laboratory have
demonstrated the importance of mechanical culture conditions on gene
and protein expression for human renal cells (16, 17, 20).
Kaysen et al. (20) compared RWV cultures to stirred and static controls. The findings include the following: 1)
tissue-specific markers such as megalin protein and mRNA and also
cubilin protein expression increased dramatically in the RWV;
2) villin, the structural protein of microvilli, showed an
early mRNA increase in association with microvillar reformation; and
3) MnSOD, a shear-stress response gene, showed an
early mRNA decrease. Hammond et al. (16, 17) documented
large populations of genes influenced by mechanical culture
conditions by gene array analysis of 6-day steady-state human
renal cell expression of 10,000 known genes. More than 800 genes
changed at least twofold during RWV culture, and more than 1,600 genes
changed as shear approached zero in the microgravity of space
(16, 17).
Although a relationship between mechanical culture conditions and renal
cell gene and protein expression is established, no strict correlation
has been made in a controlled culturing environment. In the current
study, individual properties of suspension culture are varied,
specifically medium density, and size or density of microcarrier beads,
within the single environment of a RWV. As a result, the shear stress
for cell aggregates is varied within a small dynamic range. Our study
assists in answering the following questions: Which common renal cell
proteins respond to dynamic culture conditions? Are the cellular
responses highly specific? Is there an in vivo correlation for this response?
| |
MATERIALS AND METHODS |
General cultivation.
Clonetics (San Diego, CA) isolated human renal cortical cells from
kidneys unsuitable for transplantation (15, 20), and the
cell fractions consist of a natural mixture of cells from the renal
cortex. Flow cytometry analysis of the proximal tubular marker leucine
aminopeptidase demonstrated that the cells are >98% proximal tubular
cells after culture in selective media (data not shown). These cells
were stored frozen at passage 2 and, a few weeks before an
experiment, were thawed and brought up to passage 4 as
monolayer cultures. Experiments were conducted in RWVs, specifically
the 55-ml slow-turning lateral vessel (STLV) model, manufactured by
Synthecon (Houston, TX). The STLV has a cylindrical chamber (~6-cm
diameter by 2-cm length) with gases exchanged across a silicone rubber
membrane mounted on a smaller cylindrical core. Stationary control
cultures, included in some of the experiments, were grown in 100-ml
silicone SiCulture bags (TC Tech, Minneapolis, MN). Cultivation was in
a 37°C incubator with a 5% CO2 humidified atmosphere.
Culture conditions.
The culture medium was DMEM/F12, pH 7.4, supplemented with 30 mM sodium
bicarbonate and 30 mM HEPES (total), 10% fetal calf serum, and an
antibiotic cocktail (Ciprofloxacin and Fungizone) (15,
20). For the majority of experiments presented here, renal cells
were inoculated at ~3.5 × 105 cells/ml attached to
glass-coated microcarrier beads (Solohill, Ann Arbor, MI) at 9-18
mg/ml but at the same cell coverage of 8 × 104
cells/cm2 bead surface area (see
DISCUSSION). Experiments were conducted for 2 days
in the STLV, rotating at 17 rpm, or in the stationary bag. This time
period was considered long enough to observe differences in aggregation
and gene and protein expression while not requiring feeding (confirmed
by satisfactory pH and glucose values at harvest). Bead properties were
varied between experiments, using the following average diameters
(davg) and densities (
):
davg = 120 µm ( = 1.02, 1.03, and
1.04 g/ml); davg = 180 µm ( = 1.04 g/ml); davg = 275 µm ( = 1.04 g/ml). When the larger beads are used for cell culture
experiments, the same bead concentration cannot be used throughout
because the cell-to-bead ratio will increase. Instead, we chose the
constant parameter as bead coverage. The same value of cells per square
centimeter should result in the same potential for interactions between
cells over all experimental conditions.
For an additional set of experiments, 2% (wt/vol) dextran (70 kDa) was
added to the medium. Renal cells were inoculated at 7 ± 1 × 105 cells/ml attached to Cytodex-3 (Pharmacia, Uppsala,
Sweden) microcarrier beads at 2.5 mg/ml. The Cytodex-3 beads, made of
collagen-coated dextran, have a density of 1.04 g/ml and an average
diameter of 175 µm. These experiments were conducted for 5 days in
the STLV, rotating at 17 rpm. Conditioned medium was replaced with
fresh at 25% of the vessel volume on day 2 and 35% of the
volume on days 3 and 4. After 5 days, cells were
harvested and processed for membrane protein studies.
Calculation of mechanical properties.
In mathematically defining the mechanical culture conditions, an
initial condition was assumed of cells coating single beads. The
terminal velocity (Vt) of a bead inside the RWV
was determined by Eq. 1
|
(1)
|
where g is gravitational acceleration, R
is bead radius, p is bead density, f is
fluid density, and µ is fluid viscosity.
The density difference between bead and the surrounding fluid (
)
was varied in our experiments by changing either the density of the
bead or the density of the medium. In addition, the bead radius was
varied as a separate parameter affecting bead velocity.
The maximum shear stress (
max) at the surface of a bead
is a function of its Vt as shown by
|
(2)
|
These two equations allowed us to estimate relative values of
shear stress for all of our experimental conditions. These values are
shown in Table 1. The equations allowed
us to predict a concentration of dextran in the medium that would
reduce Vt with only a minor change in shear.
Statistical analysis.
Kolmogorov-Smirnov (KS) summation statistics are often selected as the
statistical method of choice for large data sets of independent
measurements such as flow cytometry files (38). When KS
statistics are applied, the 2,000 independent flow cytometry observations on individual cells or membrane vesicles are treated as
separate observations. This greatly increases the statistical power of
cell biology data in which experimental replicates have a practical
limitation. In addition to KS statistics, single-factor ANOVA was used,
as applicable, to compare means between experimental conditions.
Flow cytometry analysis of vitamin D receptor.
To quantify the protein content of vitamin D receptor (VDR) on cells
grown in the STLV and bag, indirect fluorescent antibody binding was
assayed by flow cytometry. Cell aggregates were harvested, washed in
phosphate-buffered saline (PBS), and lysed with a PowerGen 125 homogenizer (Fisher Scientific, Pittsburgh, PA). A postnuclear supernatant was prepared by spinning out bead debris and nuclei at 200 g. The supernatant containing membranes was pelleted at 100,000 g for 10 min in a Beckman ultracentrifuge. Membranes
were resuspended in mannitol buffer (300 mM mannitol + 10 mM
HEPES, pH 7.4), and nonspecific binding was blocked after treatment
with 50% clarified goat serum. Dilutions of rat monoclonal anti-VDR antibody (Chemicon, Temecula, CA) were incubated with aliquots of
membranes overnight at 4°C, the membranes were washed, and a goat
anti-rat FITC-conjugated (Sigma Chemical, St. Louis, MO) secondary
antibody was applied. With samples in fresh mannitol buffer, the bound
antibody was then assayed by flow cytometry analysis of 2,000 membranes.
Analysis was performed on a Becton Dickinson FACS Vantage flow
cytometer with excitation at 488 nm using a Coherent 5W argon-ion laser. For each particle, emission was measured using photomultipliers at 530 ± 30 nm. After collection of data as list mode files, KS statistics (38) were performed by using Cell Quest
software (Becton Dickinson, Franklin Lakes, NJ).
Flow cytometry analysis of aminopeptidases.
Cellular activities of the enzymes cathepsin C and dipeptidyl-peptidase
IV (DPP-IV) were measured by using the Gly-Pro and Pro-Arg derivatives,
respectively, of 4-methoxy-
-naphthylamine (4-MNA; Enzyme System
Products, Livermore, CA). The enzymes specifically cleave these
substrates, liberating free 4-MNA. In the presence of
2-nitrosalicylaldehyde, at pH 
6.0, free 4-MNA is trapped and
precipitated at the site of formation (6, 15). This
product fluoresces when excited at 488 nm, with a broad emission
spectrum from 510 to 680 nm. For our assay, the cellular membrane
fraction was obtained as detailed above, and we incubated the following for 1 h at room temperature: a 75-µl aliquot of membranes with equal volumes of 1 mM dipeptide derivative of 4-MNA, 1 mM
2-nitrosalicylaldehyde, and 0.1 M sodium acetate, pH 5. The fluorescent
product was then assayed immediately by flow cytometry, and FL2 channel
(575 ± 26 nm) data were compared by use of KS statistics.
Metabolic measurements.
After harvest of cells at the conclusion of each experiment, metabolic
measurements were made on the cell-free conditioned medium, including
pH and glucose. In addition, the activity of lactate dehydrogenase
(LDH) was assayed with a spectrophotometric end-point kit (Sigma Chemical).
Other cellular assays.
Flow cytometry was used to perform cell cycle analysis, after propidium
iodide (PI) nuclear staining, and apoptotic indexes, using an
annexin V kit (Trevigen, Gaithersburg, MD), on cell samples from each
experiment. A 5-ml cell sample, containing ~1.8 × 106 cells, was spun down, washed in PBS, and treated with
trypsin solution at 37°C for 5-10 min. Cells were separated from
beads by passing the solution through 70-µm nylon mesh. The cells
were washed once more in PBS and then used for two different assays, cell cycle and apoptosis. Approximately 70% of the original
cell sample was stained with 200 µl of 50 µg/ml PI. Fluorescence
emission was collected through a 575 ± 26 nm band-pass filter.
Data for cell-cycle analysis, 2,000 events, were acquired with Cell
Quest and analyzed by using WinMDI shareware (Scripps Research
Institute, La Jolla, CA) to gate out debris and doublets from a plot of
signal width vs. area.
Approximately 20% of the original sample was used for the
apoptosis assay. First, the cells were incubated with annexin V reagent conjugated with FITC and a PI counterstain. Note that because
annexin V binds a phospholipid, the earlier trypsin proteolysis is not
thought to interfere with the assay. With the sample in binding buffer,
fluorescence is measured, and FL2 vs. FL1 dot plots indicate the
separate groups of viable, early apoptotic, and late apoptotic
cells. Necrotic cells may also appear as a PI-positive,
annexin-negative population.
Finally, relative total protein content was measured on membrane
aliquots from each vessel using the bicinchoninic acid protein assay (Pierce, Rockford, IL).
Transmission electron microscopy.
Cell aggregate samples were gently withdrawn from the vessel at the end
of the cultivation period and washed in PBS. Cells were fixed in 2.5%
glutaraldehyde at 4°C for 30 min, followed by postfix in 1% osmium
tetroxide at 4°C for 30 min. Fix and postfix solutions were in 0.1 M
cacodylate buffer, and this buffer was also used for multiple rinses
after both steps. Finally, samples were sent, in the cacodylate buffer,
to University of Wisconsin-Madison Electron Microscopy Facility.
| |
RESULTS |
Morphology.
The appearance of renal cell aggregates after 2 days culture in the
STLV or bag is documented in Table 2.
There were obvious differences between the vessel with the largest
beads [specific gravity (sg) 1.04] and the other vessels. Aggregates
had fewer beads, and there were also cell aggregates not attached to
beads. The stationary bag cultures generally averaged larger
aggregates, and these were often asymmetrical, even chainlike, compared
with more spherical ones in the STLV cultures. Note that small, medium, and large designation for beads refers to those of average diameter 120, 180, and 275 µm, respectively. Any morphological differences between vessels with 1.03-sg small beads and the 1.04-sg medium beads
were subtle.
Transmission electron microscopy demonstrated the presence of
microvilli and intracellular desmosomes in each of the RWV cultures examined (Fig. 1). With the exception of
dextran in lysosomes in dextran-treated cultures, examination of the
images demonstrates no demonstrable differences. In scanning electron
micrographs (not shown), cells are flattened as they spread over the
bead surface. This likely does not mimic the structural organization of
these proximal tubule cells in vivo.
View larger version (152K):
[in this window]
[in a new window]
|
Fig. 1.
Transmission electron micrographs of renal cortical epithelial
cells. A and B: slow-turning lateral vessel
(STLV) cultures with 2% dextran added to the medium. Arrows indicate
an intracellular desmosome (A) and lysosomes with dextran
(B). C: STLV control cells without dextran; arrow
points to microvilli in cross section.
|
|
Cell cycle and apoptosis/necrosis.
Typical results of cell cycle and apoptosis measurements are
shown in Table 3. The cell cycle
distribution varies little with the experimental condition, with means
of 48 ± 2% G1, 48 ± 3% S, and 5 ± 1%
G2 + M phase. The percentages of
apoptotic and necrotic cells are small and also have limited
variation, at 1.7 ± 0.7% apoptotic and 5.0 ± 1.3%
necrotic.
Metabolism and lysis.
The experiments with bead variations were conducted for only 2 days
without feeding. At the conclusion of these experiments, glucose ranged
from 130 to 250 mg/dl and pH from 7.2 to 7.6. Within each set of
experiments (that is, different conditions run at the same time),
glucose concentration varied by <15%. We also measured LDH activity
in cell-free culture medium at the conclusion of the 2-day runs. Other
researchers (13, 26, 29) have used this assay as a measure
of cell lysis. We recorded LDH values in enzyme units per milliliter by
comparison with a standard curve. Because basal levels in the inoculum
culture varied, however, we report LDH values normalized by the average
for a set of vessels started with the same inoculum (Fig.
2). The higher LDH level for stationary
bag cultures was not statistically significant.
View larger version (34K):
[in this window]
[in a new window]
|
Fig. 2.
Activity of lactate dehydrogenase (LDH) in cell-free
medium at the end of the cultivation period for STLV and bag cultures;
n = 8, 8, 4, 4, 3, and 3, respectively. Units of enzyme
activity have been converted to percentages weighted by the average
activity for a batch of experiments performed on the same day. There is
no statistical significance between means by single-factor ANOVA when
comparing each data pair in the group of glass-coated beads nor when
comparing the final data pair in which Cytodex beads were used with or
without dextran in the medium.
|
|
The experiments with medium variations, with or without dextran, were
conducted for 5 days with identical feeding schedules. The glucose
consumption showed almost identical patterns for dextran-treated and
control vessels, and the pH was 
7.0 throughout. The last two
bars in Fig. 2 show the LDH levels at the end of 5 days for these
cultures, and they were also quite similar.
Comparison of protein expression.
VDR, specifically the membrane-bound fraction, was one of the proteins
tested for possible variation between experimental conditions. Figure
3 demonstrates with antibody binding
curves the relative levels of VDR for renal cell cultures in the STLV with the only difference between vessels being the size and/or density
of the microcarrier beads. Replicate experiments were conducted on
selected experimental conditions as noted, for the 1.03-sg small,
1.04-sg medium, and 1.04-sg large beads, to measure culture properties
within a range of shear environments. Figure 4 shows the peak VDR antibody binding for
these conditions. For cells attached to 1.04-sg large beads, binding
was lower, at 80 ± 10 compared with means of ~110 ± 20 for the other STLV conditions. Analysis using KS statistics reveals
that the bag culture had significantly lower antibody binding compared
with the STLV cultures.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 3.
Average vitamin D receptor (VDR) antibody binding curves
for renal cells attached to beads in the STLV and bag. The
membrane-bound fraction of VDR was isolated, and indirect fluorescent
antibody binding was assayed by flow cytometry. Fluorescent units of
antibody binding are shown vs. antibody dilution on a log scale. Solid
curves represent the following samples with replication number as
indicated: STLV with 1.03-sg small glass-coated beads
( ), n = 5; STLV with 1.04-sg medium
beads ( ), n = 6; STLV with 1.04-sg
large beads ( ), n = 4; and stationary
bag with an equal mixture of these 3 types of beads ( ),
n = 2. Shown also as dotted-line curves are 2 additional samples with n = 1 each: STLV with
1.02-specific gravity (sg) glass-coated bead of average diameter 120 µm ( ) and STLV with 1.04-sg bead of average diameter
120 µm ( ).
|
|
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 4.
Peak VDR antibody binding in the STLV and bag.
Replication numbers are given with Fig. 3. Here, to aid comparison
between experiments, fluorescence units were converted to percentages
weighted by the average fluorescence for a batch of experiments
performed on the same day. When comparing each data pair by
single-factor ANOVA within the group of glass-bead conditions, only the
STLV with 1.04-sg large beads (*) was significantly different from the
other cases, with 0.025 < P < 0.08. The bag
culture (**) showed reduced binding from all other conditions on the
basis of Kolmogorov-Smirnov (KS) statistics ( < 0.05). On
comparing the final data pair, in which Cytodex beads were used with or
without dextran in the medium, KS statistics showed lower antibody
binding in the dextran-treated cultures (8 of 9 pair comparisons,
0.01).
|
|
Also shown in Fig. 4 are the peak VDR levels for another set of
experiments in which the medium density was changed by the addition of
dextran. With dextran, the fluid density approaches that of the
microcarrier beads, but fluid viscosity is increased, resulting in a
fairly minor change in shear stress. By KS statistical analysis, VDR
antibody binding to renal membranes was significantly lower in the
dextran-treated group.
Another membrane protein measured in our experiments was
DPP-IV. Although detected by flow cytometry, it was assayed
functionally instead of the antibody binding being used to measure
VDR. Figure 5 shows the fairly
constant activity level of this protein over all the experimental
conditions. No statistical difference was found.
View larger version (35K):
[in this window]
[in a new window]
|
Fig. 5.
Activity of dipeptidyl-peptidase IV (DPP-IV) in the STLV
and bag cultures, as measured by flow cytometry after proteolysis of a
specific dipeptide to form a fluorescent product; n = 6, 6, 4, and 2, respectively. No significant difference was shown
between means.
|
|
The third protein measured in our experiments was cathepsin C, or
dipeptidyl-peptidase I. Like DPP-IV, it was measured with a functional
assay and detected by flow cytometry. Figure
6 demonstrates that, for one experimental
condition, the cathepsin C activity was always significantly lower,
when using either KS statistics or single-factor ANOVA. This condition
was cells in the STLV attached to 1.04-sg medium beads.
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 6.
Activity of cathepsin C in the STLV and bag cultures, as
measured by flow cytometry after proteolysis of a specific dipeptide to
form a fluorescent product; n = 6, 6, 4, and 2, respectively. One condition, the STLV with 1.04-sg medium beads (*),
had significantly lower activity than all of the other experimental
conditions, with 0.0002 < P < 0.002.
|
|
Total protein was measured on the cellular membrane fraction for each
sample isolated for the specific protein assays. Those results (Fig.
7) demonstrate that total protein was
statistically similar except for the STLV with 1.04-sg medium beads.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 7.
Average total membrane protein concentration for STLV and
bag cultures; n = 6, 6, 4, and 2, respectively.
Comparison between the first 2 data points shows that STLVs with the
1.04-sg medium beads (*) had higher levels of total membrane protein
than STLVs with the 1.03-sg small beads, with P < 0.03. This finding does not affect any other results.
|
|
Correlation with shear.
By using estimates of shear stress, as presented in Table 1, the
relative concentration or enzyme activity of the three measured proteins was plotted vs. shear to examine the relative relationships. Figure 8 shows three very different shear
dependencies for measurements of these three proteins. In the following
section, we will discuss the potential biological implications of this
result.
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 8.
Plots of enzyme activity or peak antibody binding related
to shear stress associated with each experimental condition. Terminal
velocities were calculated from Eq. 1 and maximum shear
stress on a single cell-coated bead from Eq. 2 in
METHODS. This yields not an absolute but a relative value
of shear stress suitable for purposes of comparison between
experiments. The 3 different renal-specific proteins tested show vastly
different patterns of expression or activity with shear, indicating
that these effects are specific. Polynomial curves or lines were
included in the figure to show the approximate trends in protein
expression with shear for VDR ( ), DPP-IV
(), and cathepsin C ().
|
|
| |
DISCUSSION |
Although an old approach, suspension culture only increases in
importance over time, as the demand for cell cultures maintaining differentiated features increases for both academic and industrial applications (1, 17). Suspension culture continues to be useful for production of many bioproducts, from antibodies to hormones
(1, 35). The availability of cell culture models, which
are so easily modulated and studied, with an intact tissue-specific repertoire of signal transduction and metabolic pathways, would represent a dramatic biotechnology advancement. Engineering
optimization of suspension culture was largely undertaken by National
Aeronautics and Space Administration engineers to model culture
conditions in spaceflight (13, 22, 37) but may find its
greatest utility in the carryover to ground-based applications
(13, 16, 35).
Engineering optimization of the RWV to minimize shear while maintaining
laminar flow has been successfully modeled and experimentally validated
(10, 11, 13, 22, 27, 32, 37). However, the engineering is
so effective that residual shear in the vessel is likely to be far
below the in vivo levels experienced by many tissues such as vascular
endothelium, renal proximal tubules, and blood cells. One interesting
question is how this environment will affect the cultured cell's
biological behavior in terms of protein expression.
RWV culture provides a complex set of dynamic culture conditions, in
which the quantitative importance of each physical parameter to the
observed biological effects remains unknown. Understanding the effects
of the RWV has commonly been approached by comparing cells grown in the
vessel to diverse control conditions ranging from conventional
two-dimensional monolayer cultures, through stirred fermentors, and
nonadherent culture bags (13, 16, 17, 35). To begin to
dissect the mechanistic issues, we chose to change a single parameter
(medium density, bead size, or bead density) in comparing cell behavior
in a set of otherwise identical RWVs. This approach begins to bring
these experiments into a much more controlled and interpretable
setting, with defined variables to modify and observe.
Terminal velocity and shear stress in the RWV.
The RWV suspends cells by rotation, which introduces centrifugal and
Coriolis pseudoforces. A particle in the RWV traces a small circular
path while it is rotating with the fluid in a much larger circular
path. A force balance on the particle shows that in the radial
direction centrifugal force and gravity counteract each other, whereas
in the tangential direction gravity and Coriolis force are additive.
The particle's velocity components are functions of its
Vt and the rotational speed of the vessel
(37). The terminal or sedimentation velocity
(Vt) is defined by Eq. 1 in
MATERIALS AND METHODS as the equilibrium velocity at which
a particle moves under the influence of gravity. Equation 1
suggests methods of reducing Vt under unit
gravity; for example, driving the density of particle and fluid closer
together, decreasing particle size, or increasing fluid viscosity. When
the vessel is operated at low rotational speeds, the gravitational
effect dominates and particles tend to settle. When the vessel is
operated at high rotational speeds, centrifugal effects dominate,
accumulating with time, so that particles move out toward the wall of
the vessel. Typically, aggregate size will increase as cells grow over
time in the vessel. Thus rotational speed may need to be adjusted to promote ideal suspension conditions and avoid settling or excessive wall impacts.
Equation 2 in MATERIALS AND METHODS describes
the maximum shear stress (max) as a function of the
Vt. Both Eqs. 1 and 2
apply to creeping flow around a solid sphere (2), which
adequately describes the movement of a bead particle inside the RWV.
This is true when the Reynolds number (Re = 2RfVt/µ) is less
than 0.1 (2), which is the case for all of our
experimental conditions except for the largest beads, the 1.04-sg beads
with an average diameter of 275 µm. For this condition, the Reynolds
number is ~0.3. There are other expressions for estimating the
Vt for higher Reynolds numbers. However, for
Re < 2, the present expression is more accurate (2).
Hence, it is reasonable to apply the same equations to all of our
experimental conditions.
Gao et al. (11), in their model of bead motion in the RWV,
defined a relative velocity (Vrel) between the
bead and the fluid and used Eq. 2 to calculate shear stress
at the bead surface, with Vrel in place of
Vt. They show the periodic nature of the relative
velocity, and hence shear stress, but because the oscillations are
slight these values may be assumed constant for a given set of
conditions. The authors define a system in terms of the following physical parameters: fluid viscosity (µ), density difference () between particle and fluid, microcarrier bead radius (R),
vessel rotational speed (
), and initial particle position
(r0). They found that the maximum shear stress increases
linearly with particle radius and with the density difference. Maximum
shear stress is constant, however, with respect to fluid viscosity and
vessel speed. The authors suggest that the opposing effect of viscosity on relative velocity cancels its effect on shear stress. Although max is not dependent on or µ, another important
parameter is, i.e., the time until wall impact, which varies inversely
with and increases linearly with µ. Thus there will be more wall impacts at higher vessel speeds and lower fluid viscosities. From this
analysis, we can conclude that the major determinants of shear on a
cell aggregate in the RWV are density difference and particle radius,
and these are indeed the variables that are permuted in the current work.
Validity of comparison.
The engineering calculation of Vt and shear is
based on a simplified initial condition of cells coating single beads.
As the culture matures and larger aggregates form, the
Vt and shear evolve dynamically. We have
examined the outcome at the end of 2 days in culture. Future studies
may need to follow the time course and/or quantify aggregate size over
time for a more accurate assessment of experimental conditions in the vessel.
Because of the body of evidence that shear stress modifies gene and
protein expression in vascular endothelial cells, shear stress is often
quoted as a biologically active culture parameter in the RWV (13,
20, 22). To test the importance of shear in our system, we
devised an initial set of experiments in which, by adding dextran to
the vessel, we cut the Vt of the cell aggregates by more than half but changed shear relatively little. Perhaps as
significant is that wall impacts were theoretically reduced with the
addition of dextran because the fluid viscosity increased. In this
setting, there was no effect on the assay of VDR. This provides direct
evidence that Vt per se, and perhaps wall
impacts as well, has little effect on at least one renal-specific
protein. In addition, this provides indirect evidence that shear is
important in the changes observed in subsequent experiments.
There was at least one physical difference between culture conditions,
that of aggregate morphology. The variable that we could control was
bead coverage, so that all cultures were inoculated with the same
number of cells per square centimeter of bead surface. At the
conclusion of the experiment, the number of beads per aggregate was
less for the 1.04-sg large-bead condition, which could have affected
the potential for cell-cell contacts. Interactions between cells are
generally accepted as important for modulating control of the cell
cycle and apoptosis (33). However, we found no
evidence that the cell cycle, cell death, or lysis varied to any extent for the different conditions in this study. Given the large size of the
beads in question, the reduced bead number does not necessarily correspond to smaller aggregate size. Again, it is aggregate diameter that affects Vt, shear, and wall impacts.
Monitoring the dynamic changes in aggregate diameter would be a logical
extension to this work.
The protein expression differences we observed are apparently not
related to modulations in total protein. Total protein levels were
similar between conditions, with the exception of higher protein in the
STLV with 1.04-sg medium beads compared with the smaller diameter
beads. However, lower levels of cathepsin C for the 1.04-sg medium-bead
condition cannot be attributed to lower total protein. And, for the
other two proteins studied, VDR and DPP-IV, levels were not higher
solely for this vessel condition.
Many changes in mechanical culture conditions can also change mass
transport (10, 32). However, unlike cells growing in the
quiescent conditions of spaceflight (17, 32, 37),
aggregates spinning in a rotating vessel have adequate mass transport
because of bulk flow (5, 11, 35). Similarly, it is
unlikely that the observed changes are nonspecific toxic effects
because there was no effect on cell cycle, cell death, or lysis. This
is particularly important given evidence of enhanced cell death by
apoptosis and necrosis for insect cells in a shaker flask, with
slightly higher shear, compared with those in a RWV (5).
Our results are consistent with mechanical forces mediating the
observed changes.
Molecular markers.
We chose to examine the biological effect of changing the mechanical
parameters, and R, in terms of the expression of select proteins
in our system. These were the VDR, and two aminopeptidases, cathepsin C
(dipeptidyl-peptidase I) and DPP-IV. They were chosen on the following
bases: the ability to measure these proteins by simple, reliable assay;
their physiological importance as renal proteins; and previous
demonstration of basal expression high enough to assay combined with
changes in gene or protein expression observed in our laboratory's
earlier RWV studies (16, 20).
VDR binds the active form of the vitamin D hormone,
1,25-(OH)2 vitamin D3, to create a
transcriptional complex. It is one of a family of nuclear receptors
and, besides kidney, is present in other target tissues of vitamin D
(19). Much is known about the DNA and ligand binding
activities of this receptor but not about its potential associations
with other proteins (19). A portion of the total VDR is
partitioned to the membrane as we are able to measure specific binding
from the membrane fraction of cultured renal cells. Cathepsin C is a
predominantly lysosomal aminopeptidase that is widely distributed in
mammalian tissues, but especially in the kidney and spleen
(34). Its main function is protein degradation in the
lysosomes. DPP-IV is a glycoprotein found on the brush border membrane
of cells of the proximal tubule and glomerulus and is also found in
liver, parotid, and salivary glands. It has a transmembrane segment,
but 95% is extracellular. Among its myriad functions are proteolysis
of bradykinin and substance P and binding to the extracellular matrix
components collagen and fibronectin (36).
Relating biological results to mechanical properties.
The pattern of changes in VDR, cathepsin C, and DPP-IV with changes in
shear has instructive characteristics. First, although abundantly
expressed, DPP-IV activity did not change with any modification
in culture conditions. Because DPP-IV is an apical brush border
membrane enzyme, this suggests that the other changes observed are
highly specific changes and are not due simply to a nonspecific or
general change in all protein production or membrane cycling. The rapid
biological turnover of DPP-IV (21) suggests that these
changes are valid and not simply secondary to a long membrane residence
time of this marker. Second, there is a relationship between VDR
expression and shear, such that optimal expression occurs at the
midrange of dynamic shear conditions examined and not the lowest shear
possible. Third, cathepsin C had an inverse relationship to shear,
being suppressed at mid levels of shear. Hence, the pattern of change
with modulation of shear is consistent and reproducible for each
protein examined, but the protein changes demonstrated diverse patterns
and specificity of response.
Although the specific signal transduction mechanisms from mechanical
culture conditions to changes in gene and protein expression are
incompletely understood, there are several important clues. First,
there are precedents for other mechanical culture conditions changing
gene expression (8, 9, 12, 13, 17, 25, 28, 30, 31, 39). In
particular, heat shock proteins may transduce changes in temperature
into molecular cell processes by heat-dependent conformational changes
in chaperone proteins (7, 24, 39). Similarly, both
vibration and gravity mediate specific gene responses in osteoblasts
and Jurkat cells, including but not limited to changes in heat shock
proteins and c-Fos (8, 9, 18, 23, 24). This
increase in c-Fos is cAMP but not protein kinase C dependent
(9). Several lines of evidence, including the
identification of transcriptional binding sites and candidate
transcription factors, suggest that shear stress can mediate select
changes in cellular gene expression in vascular endothelial cells
(12, 24, 28, 30). The central role of the cytoskeleton in
cell scaffolding, load bearing, and transport of vesicles continues to
make it a popular candidate for transducing physical stimuli into other
cellular processes (23, 24, 32). There is some molecular
evidence for this process, including concomitant cytoskeletal
alterations characterized by diffuse shortened microtubules, increased
apoptosis, and time-dependent elevation in Fas/APO-1 protein in
spaceflown human lymphocytes (23, 24).
There is some uncertainty as to the actual level of shear experienced
by the renal cell. A recent study by Guo et al. (14) models the response of renal cells to flow inside the proximal tubule.
Their model shows that, although shear stress varies from 1 to 5 dyn/cm2 over the normal range of tubule flow rates, shear
at the base of the microvilli is several hundred times smaller. That
is, most of the drag force from fluid flow affects only the tip of the microvillus. In our own experimental system, the cells may not have the
same structural organization as in the proximal tubule. Therefore, we
must relate the observed biological differences in our experiments to
relative values of shear stress.
This study provides direct evidence that mechanical culture conditions
modulate protein expression in suspension culture. When shear and
Vt are dissociated by changing the fluid
viscosity, Vt alone has no effect on protein
expression or activity. The responses of protein expression or activity
to changes in the shear environment were highly specific for each
protein. Future studies may examine the effect on other proteins
important to renal cell culture.
| |
ACKNOWLEDGEMENTS |
We are grateful to Randall Massey, Managing Director of the
University of Wisconsin-Madison EM Facility, for expertise.
| |
FOOTNOTES |
This work was supported by National Aeronautics and Space
Administration NRA Grants 9-811 Basic and NAG 8-1362 (to T. G. Hammond). The Department of Veterans Affairs provided equipment and facilities.
Address for reprint requests and other correspondence: N. L. Cowger, StelSys, LLC, 1450 South Rolling Rd., Baltimore, MD 21227 (E-mail: cowgern{at}stelsysllc.com).
The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement"
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Received 2 August 2000; accepted in final form 10 October 2001.
| |
REFERENCES |
1.
Birch, JR,
and
Arathoon R.
Suspension culture of mammalian cells.
Bioprocess Technol
10:
251-270,
1990[Medline].
2.
Bird, RB,
Stewart WE,
and
Lightfoot EN.
Transport Phenomena. New York: Wiley, 1960, p. 56-60.
3.
Cherry, RS,
and
Hulle CT.
Cell death in the thin films of bursting bubbles.
Biotechnol Prog
8:
11-18,
1992[Medline].
4.
Clejan, S,
O'Connor KC,
Cowger NL,
Cheles MK,
Haque S,
and
Primavera AC.
Effects of simulated microgravity on DU 145 human prostate carcinoma cells.
Biotechnol Bioeng
50:
587-597,
1996.
5.
Cowger, NL,
O'Connor KC,
Lacks DJ,
Hammond TG,
and
Navar GL.
Characterization of bimodal cell death of insect cells in a rotating-wall vessel and shaker flask.
Biotechnol Bioeng
64:
14-26,
1999[Web of Science][Medline].
6.
Dolbeare, FA,
and
Smith RE.
Flow cytometric measurement of peptidases with use of 5-nitrosalicylaldehyde and 4-methoxy-beta-naphthylamine derivatives.
Clin Chem
23:
1485-1491,
1977[Abstract/Free Full Text].
7.
Feder, ME,
and
Hofmann GE.
Heat-shock proteins, molecular chaperones, and the stress response: evolutionary and ecological physiology.
Annu Rev Physiol
61:
243-282,
1999[Web of Science][Medline].
8.
Fitzgerald, J,
and
Hughes-Fulford M.
Gravitational loading of a simulated launch alters mRNA expression in osteoblasts.
Exp Cell Res
228:
168-171,
1996[Web of Science][Medline].
9.
Fitzgerald, J,
and
Hughes-Fulford M.
Mechanically induced c-fos expression is mediated by cAMP in MC3T3-E1 osteoblasts.
FASEB J
13:
553-557,
1999[Abstract/Free Full Text].
10.
Gao, H,
Ayyaswamy PS,
and
Ducheyne P.
Numerical simulation of global diffusive mass transfer in a rotating wall vessel bioreactor.
Advances in Heat and Mass Transfer in Biotechnology American Society of Mechanical Engineers, HTD 335/BED
37:
59-67,
1997.
11.
Gao, H,
Ayyaswamy PS,
and
Ducheyne P.
Dynamics of a microcarrier particle in the simulated microgravity environment of a rotating wall vessel.
Microgravity Sci Technol
X/3:
154-165,
1997.
12.
Gimbrone, MA, Jr.
Vascular endothelium, hemodynamic forces, and atherogenesis.
Am J Pathol
155:
1-5,
1999[Free Full Text].
13.
Goodwin, TJ,
Prewett TL,
Wolf DA,
and
Spaulding GF.
Reduced shear stress: a major component in the ability of mammalian tissues to form three-dimensional assemblies in simulated microgravity.
J Cell Biochem
51:
301-311,
1993[Web of Science][Medline].
14.
Guo, P,
Weinstein AM,
and
Weinbaum S.
A hydrodynamic mechanosensory hypothesis for brush border microvilli.
Am J Physiol Renal Physiol
279:
F698-F712,
2000[Abstract/Free Full Text].
15.
Hammond, TG.
Analysis and isolation of renal tubular cells by flow cytometry.
Kidney Int
42:
997-1005,
1992[Web of Science][Medline].
16.
Hammond, TG,
Benes E,
O'Reilly KC,
Wolf DA,
Linnehan RM,
Taher A,
Kaysen JH,
Allen PL,
and
Goodwin TJ.
Mechanical culture conditions effect gene expression: gravity induced changes on the space shuttle.
Physiol Genomics
3:
163-173,
2000[Abstract/Free Full Text].
17.
Hammond, TG,
Lewis FC,
Goodwin TJ,
Linnehan RM,
Wolf DA,
Hire KP,
Campbell WC,
Benes E,
O'Reilly KC,
Globus RK,
and
Kaysen JH.
Gene expression in space.
Nat Med
5:
359,
1999[Web of Science][Medline].
18.
Hughes-Fulford, M,
and
Lewis ML.
Effects of microgravity on osteoblast growth activation.
Exp Cell Res
224:
103-109,
1996[Web of Science][Medline].
19.
Jones, G,
Strugnell SA,
and
DeLuca HF.
Current understanding of the molecular actions of vitamin D.
Physiol Rev
78:
1193-1231,
1998[Abstract/Free Full Text].
20.
Kaysen, JH,
Campbell WC,
Majewski RR,
Goda FO,
Navar GL,
Lewis FC,
Goodwin TJ,
and
Hammond TG.
Select de novo gene and protein expression during renal epithelial cell culture in rotating wall vessels is shear stress dependent.
J Membr Biol
168:
77-89,
1999[Web of Science][Medline].
21.
Kehlen, A,
Gohring B,
Langnew J,
and
Riemann D.
Regulation of the expression of aminopeptidase A, aminopeptidase N/CD13 and dipeptidylpeptidase IV/CD26 in renal carcinoma cells and renal tubular epithelial cells by cytokines and cAMP-increasing mediators.
Clin Exp Immunol
111:
435-441,
1998[Web of Science][Medline].
22.
Kleis, SJ,
Schreck S,
and
Nerem RM.
A viscous pump bioreactor.
Biotechnol Bioeng
36:
771-777,
1990.
23.
Lewis, ML,
Reynolds JL,
Cubano LA,
Hatton JP,
Lawless BD,
and
Piepmeier EH.
Spaceflight alters microtubules and increases apoptosis in human lymphocytes (Jurkat).
FASEB J
12:
1007-1018,
1998[Abstract/Free Full Text].
24.
Lewis, ML,
and
Hughes-Fulford M.
Regulation of heat shock protein message in Jurkat cells cultured under serum-starved and gravity-altered conditions.
J Cell Biochem
77:
127-134,
2000[Web of Science][Medline].
25.
Malek, AM,
and
Izumo S.
Control of endothelial cell gene expression by flow.
J Biomech
28:
1515-1528,
1995[Web of Science][Medline].
26.
McQueen, A,
Meilhoc E,
and
Bailey JE.
Flow effects on the viability and lysis of suspended mammalian cells.
Biotechnol Lett
9:
831-836,
1987.
27.
Meaney, DF,
Johnston ED,
Litt M,
and
Pollack SR.
Experimental and numerical investigations of microcarrier motions in simulated microgravity.
Advances in Heat and Mass Transfer in Biotechnology American Society of Mechanical Engineers, HTD
63:
21-27,
1998.
28.
Nerem, RM,
Alexander RW,
Chappell DC,
Medford RM,
Varner SE,
and
Taylor WR.
The study of the influence of flow on vascular endothelial biology.
Am J Med Sci
316:
169-175,
1998[Web of Science][Medline].
29.
Petersen, JF,
McIntire LV,
and
Papoutsakis ET.
Shear sensitivity of cultured hybridoma cells (CRL-8018) depends on mode of growth, culture age, and metabolite concentration.
J Biotechnol
7:
229-246,
1988.
30.
Resnick, N,
and
Gimbrone MA, Jr.
Hemodynamic forces are complex regulators of endothelial gene expression.
FASEB J
9:
874-882,
1995[Abstract].
31.
Tjandrawinata, RR,
Vincent VL,
and
Hughes-Fulford M.
Vibrational force alters mRNA expression in osteoblasts.
FASEB J
11:
493-497,
1997[Abstract].
32.
Todd, P.
Physical Forces and the Mammalian Cell. San Diego: Academic, 1992, p. 347-381.
33.
Tomei, LD.
Apoptosis: The Molecular Basis of Cell Death. New York: Cold Spring Harbor Laboratory Press, 1991, p. 279-316.
34.
Turk, B,
Dolenc I,
and
Turk V.
Handbook of Proteolytic Enzymes. San Diego: Academic, 1998, p. 378-382, p. 631-634.
35.
Unsworth, BR,
and
Lelkes PI.
Growing tissues in microgravity.
Nat Med
4:
901-907,
1998[Web of Science][Medline].
36.
Vlahovic, P,
and
Stefanovic V.
Kidney ectopeptidases: structure, functions, and clinical significance.
Path Biol
46:
779-786,
1998.
37.
Wolf, DA,
and
Schwarz RP.
Analysis of Gravity-Induced Particle Motion and Fluid Perfusion Flow in the NASA-Designed Rotating Zero-Head-Space Tissue Culture Vessel. Houston, TX: NASA, 1991. (NASA Technical Paper 3143)
38.
Young, IT.
Proof without prejudice: use of the Kolmogorov-Smirnov test for the analysis of histograms from flow systems and other sources.
J Histochem Cytochem
25:
935-941,
1977[Abstract].
39.
Yura, T,
and
Nakahigashi K.
Regulation of the heat-shock response.
Curr Opin Microbiol
2:
153-158,
1999[Web of Science][Medline].
J APPL PHYSIOL 92(2):691-700
TOP
ABSTRACT
INTRODUCTION
