Design of an integrated microfluidic system for culturing patterned tissue

1
Design of an integrated microfluidic system for
culturing patterned tissue Allyson Fry1, Bryan
Gorman1, Jonathan Lin2, William Wong1 Advisors
John Wikswo1, Dmitry Markov1, Lisa McCawley3,
Phil Samson1 Departments of 1Biomedical
Engineering, 2Electrical Engineering, and 3Cancer
Biology at Vanderbilt University.
Introduction
Microfluidic pH control in LabVIEW
Characterization of tissue culture model
Bioreactors are commonly used biotechnology
equipment in which small organisms and other
biological components are grown. In this
project, a bioreactor is being used to grow a
three dimensional organotypic cell culture
comprised of primary endothelial cells and
fibroblasts embedded in a collagen matrix. We
have shown that this tissue culture model
produces a capillary network. One of the main
goals of this project is to show that the growth
of this capillary network can be guided using a
Matrix Enabled Capillary Scaffold (MECS).
Perfusion is then introduced using a microfluidic
network so that the cultured cells can receive
nutrients and be rid of wastes in a controlled
environment. The flow of this perfusion network
is monitored for pH content with a LabView
control system.
An organotypic culture is any cell culture that
mimics the in vivo tissue in form and function.
The organotypic culture in use in this project is
adapted from a model used by Manuela
Martins-Green at University of California-Riversid
e1. Our cultures contain two layers of human
dermal fibroblasts embedded in a collagen matrix
and separated by a layer of human microvascular
endothelial cells. This culture is constructed
in a transwell chamber in a normal 12-well
culture plate. The culture is fed by placing
media below the chamber, allowing it to perfuse
upward to the culture through a polycarbonate
membrane, as well as being placed on top of the
entire culture and seeping downward.
Compare pH value with desired pH range
Detect pH
Outside of range?
NO
YES
Do not change pumping rates

• Model a complete microvascular network
• Support cellular processes
• Allow evaluation of molecular mechanisms
• Provide environment for endothelial cell
  differentiation
• Facilitate formation of stable tubular structures
• Support a flow of perfusate

Find magnitude of error e(t)
Project Goals
The tissue culture model used as the basis for the design of the bioreactor and for assessing the angioconductive ability of scaffold designs
Changes in pH are detected by a pH sensor and
processed by a computer. Appropriate commands are
then sent to the syringe pumps to compensate for
these changes by altering pumping rates. Overall
flow rates are kept constant by close
coordination between the changes in pumping rates
for each pump.
Process error using PID equation C(t) P I D

1. Martins-Green M, Li QJ, and Yao M. FASEB 2005
   19222-224.

Upper collagen and fibroblast layer
Social Impact
a
In order to study the functions and mechanisms of
a system, scientists commonly use animal models
such as mice, rats, and rabbits. However this
practice has several shortcomings. First, the
use of animals for research is fraught with
ethical implications concerning the worth of
animal life. Second, although animal studies can
provide a reasonable basis to answer vital
questions, results can never full predict the
response of human tissue to treatment. Finally,
an animal system is relatively difficult to
manipulate, and often much of the control must be
done on the molecular level by reconstructing the
genome. The manipulations frequently cannot
carry over from one study to another, leading to
a redesigning of the entire model. Our project
overcomes these shortcomings by creating a
completely human model, with many levels of
control, and easy adaptability.
Change pumping rate of high pH pump
Lower collagen and fibroblast layer
The Proportional, Integral, Differential (PID)
control theory equation is used to control the
pumps. All three terms contribute to the
resulting control signal using different pieces
of information. The proportional term (P) takes
into account the immediate presence of error. The
integral term (I) helps compensate for
persistent, steady-state error. The differential
term (D) compensates for trends in error.
a
Polycarbonate Filter
Change pumping rate of low pH pump by equal and
opposite amount
a
a
Overall Bioreactor Design
The inspiration for the scaffold design is the
fractal pattern of vasculature occurring in
nature. Cells grown in fractal flows are exposed
to similar nutrient, waste, and shear stress
levels. The purpose of the bioreactor is to add
another spatial dimension to transwell plate
cocultures. In this design, a scaffold is seeded
with endothelial cells and placed between two
layers of collagen seeded with fibroblasts.
Parallel supply networks above and below perfuse
the tissue culture with nutrients via diffusion
through supporting nano-pore filters.
Physical constraints for pH control
Arrows point to endothelial tubes. For reference,
the scaffold is 10 um in thickness.
Cells Media needed (ul/cell/hr) Flow rates (ul/hr) Volume of bioreactor (ul) Desired pH range
45,000 4e-4 5-20 1-10 7.2-7.6
Endothelial tubes as seen using HE staining on a developed tissue culture with no scaffold added.
Fabrication of scaffolds
Integration of scaffolds into tissue culture

• We have demonstrated the fabrication of three
  different types of tissue culture scaffold
• The PDMS Egg-crate scaffold
• Advantage supports arterial and venous branches
  in rigid network
• Disadvantages difficult to construct, not
  biodegradable
• The patterned Matrigel scaffold
• Advantage biodegradable, easy fabrication
• Disadvtantage poor mechanical features
• The SU-8 Lift-off scaffold
• Advantage good mechanical features, easy
  fabrication
• Disadvantage poor biocompatibility

Endothelial cells were seeded into the three
different tissue culture scaffolds, which were
then placed between two layers of collagen
interspersed with fibroblasts using the organ
culture model characterized above. Three trials
using each type scaffold were conducted. The
results are forthcoming. Based on these results
the ability of each type of scaffold to induce
endothelial alignment and vascular tube formation
will be assessed with iimmunofluorescence and
confocal imaging.
An 80 um PDMS scaffold resting on a glass slide
Conclusion
Perfusion of the bioreactor occurs by a fluidics
system that both injects media to support the
cells, and also maintains stable pH and oxygen
levels. Two syringe pumps inject low- and high-pH
RPMI media into separate flow lines that unite
and feed fluid through an oxygenator. After
passing through the bioreactor, the flow line is
split and fed into two dedicated output syringes
mounted onto the same pumps as the input
syringes, so as not to re-use any fluid. Changes
in pH are detected by a pH sensor and processed
by computer. Appropriate commands are then sent
to the syringe pumps to raise or lower pH while
keeping flow rate constant.
Dimensions of the PDMS egg-crate scaffold
In conclusion, we have characterized an
angiogenic tissue culture model, fabricated three
types of capillary scaffolds, and tested those
scaffolds directly into the tissue culture model.
Additionally, we have written pH control
software in LabVIEW and set up a system for
controlling the pH of a microenvironmnent. Now,
we must make experimental measurements of pH
control and analyze the results of the
scaffold-patterned tissue cultures.
Cost and market analysis
The appeal of this product will be the high level
of microenvironment sophistication and control,
relative to current products.  The cost will also
be relatively low.  Although the pump, sensors,
and computer hardware will likely cost over
10000 alone, they represent a one-time
purchase.  However, the plastic microfluidic
chips are disposable and can be batch-produced
off a single mold for less than several dollars
each. 
Acknowledgements
Jason Greene contributed suggestions for device
design and experimental testing. This project is
a continuation of a senior design project started
in 2004-2005 by Barnett, Garrett, Harvill, Mayer,
and McClintock. This project was made possible
by use of resources provided by the Vanderbilt
Institute for Integrative Biosystems Research and
Education (VIIBRE) directed by Dr. John Wikswo.
Dr. Paul King is the course instructor for BME
272.
A Two-layer SU-8 photolithography mold
An image of the PDMS scaffold in cross-section