Real-time assessment of three-dimensional cell
aggregation in rotating wall vessel bioreactors in vitro
Gregory P Botta
1
, Prakash Manley
2
, Steven Miller
3
& Peter I Lelkes
1
1
Laboratory of Cellular and Tissue Engineering, Drexel University, New College Building, 245 N. 15th St., Philadelphia, Pennsylvania, USA.
2
Valleylab, Tyco Healthcare
Group, Mail Stop A23, 5920 Longbow Drive, Boulder, Colorado 80301, USA.
3
I. Miller Precision Optical Instruments, Inc., 35 N. Second St., Philadelphia, Pennsylvania
19106, USA. Correspondence should be addressed to P.I.L. (pilelkes@drexel.edu).
Published online 14 December 2006; doi:10.1038/nprot.2006.311
Until now, tissue engineering and regenerative medicine have lacked non-invasive techniques for monitoring and manipulating
three-dimensional (3D) tissue assembly from specific cell sources. We have set out to create an intelligent system that automatically
diagnoses and monitors cell–cell aggregation as well as controls 3D growth of tissue-like constructs (organoids) in real time. The
capability to assess, in real time, the kinetics of aggregation and organoid assembly in rotating wall vessel (RWV) bioreactors could
yield information regarding the biological mechanics of tissue formation. Through prototype iterations, we have developed a versatile
high-resolution ‘horizontal microscope’ that assesses cell–cell aggregation and tissue-growth parameters in a bioreactor and have
begun steps to intelligently control the development of these organoids in vitro. The first generation system was composed of an
argon-ion laser that excited fluorescent beads at 457 nm and fluorescent cells at 488 nm while each was suspended in a high-aspect
rotating vessel (HARV) type RWV bioreactor. An optimized system, which we introduce here, is based on a diode pumped solid
state (DPSS) green laser that emits a wavelength at 532 nm. By exciting both calibration beads and stained cells with laser energy
and viewing them in real time with a charge-coupled device (CCD) video camera, we have captured the motion of individual
cells, observed their trajectories, and analyzed their aggregate formation. Future development will focus on intelligent feedback
mechanisms in silico to control organoid formation and differentiation in bioreactors. As to the duration of this entire multistep
protocol, the laser system will take about 1 h to set up, followed by 1 h of staining either beads or cells. Inoculating the bioreactors
with beads or cells and starting the system will take approximately 1 h, and the video-capture segments, depending on the aims
of the experiment, can take from 30 s to 5 min each. The total duration of a specific experimental protocol will also depend
on the specific cell type used and on its population-doubling times so that the required numbers of cells are obtained.
INTRODUCTION
Rotating wall vessel bioreactors
Rotating wall vessel (RWV) bioreactors have become a common
low-shear culture venue for studying cellular aggregation and the
formation of three-dimensional (3D) tissue-like assemblies (orga-
noids). These organoids are demonstrably more differentiated,
have increased size and exhibit behavior more similar to actual in
vivo tissue than those cultured in either two-dimensional (2D)
flasks or other bioreactors
1,2
. By mimicking natural tissue with
higher fidelity, culture in RWV bioreactors permits more accurate
evaluations of protein interactions, gene expression and signaling
pathways
3–6
. The in vitro generation of these organoids recalls their
ontogenesis in vivo, supporting the notion that RWV bioreactors
might serve as suitable locations for in vitro embryology experi-
ments. In regards to the mechanism of action of the RWV, the solid
body rotation in the enclosed fluid medium allows cellular particles
to be suspended continuously by the equalization of gravity-
induced sedimentation against the constant upward forces pro-
vided by the rotation of the vessel
7
. Over time, the close proximity
(colocation) of particles moving along with the fluid in the RWV
facilitates the formation of cellular aggregates. These aggregates
rapidly increase in size and mass, forming tissue-like assemblies,
many of which can become up to three orders of magnitude larger
than those in stationary flask cultures
8
. As they increase in mass,
each organoid gains a more forceful gravitational pull on its
structure, increasing shear stress and drag on the nascent tissue.
So, without increasing the speed of rotation and offsetting the
continually increasing gravitational pull, the organoids no longer
remain in the venue-specific rotatory suspension culture as
characterized by ‘stationary’ free-fall. Without appropriate and
continual adjustment of the rotational speed, the growing cellular
aggregates tumble and collide with one another, as well as with the
walls of the RWV. The increasing collision frequency with the wall
eventually results in the destruction of the nascent tissue-like
assemblies. In current methodologies, the increase in velocity
to perpetuate the continual free-fall of growing tissues has been
manually or empirically coordinated from either past experience
or from algorithmic knowledge of bodies tumbling in this
environment
9
.
Current techniques to determine aggregation kinetics in 3D
Past aggregation models have relied on inanimate particles or
microcarrier beads seeded with cells in RWVs and have based
their schemes on Smoluchowski’s population-balance equa-
tions
10,11
. These current experimental procedures provide the
‘real world’ input for quantitative-aggregation models of micro-
carriers in RWVs, yet are confronted with the major disadvantage
of static interference
9
. In other words, at each of the known or
calculated time points during cellular aggregation, the bioreactor is
regularly stopped and drained of its enclosed fluid medium in an
effort to extract the growing cellular mass. After this step, the
cellular body is studied for membrane linkages, measurement of its
development and reaction to certain growth factors. Although these
studies are well founded in their observational analysis, the deve-
loped tissue is often subjected to disruptive shear forces during the
p u o r G g n i h s i l b u P e r u t a N 6 0 0 2 © natureprotocols / m o c . e r u t a n . w w w / / : p t t h
2116 | VOL.1 NO.5 | 2006 | NATURE PROTOCOLS
PROTOCOL