(142bn) A Peg-DA Microfluidic Device for the Study of Cellular Chemotaxis

A PEG-DA
MICROFLUIDIC DEVICE FOR THE STUDY OF

CELLULAR
CHEMOTAXIS

Mahama A. Traore¹, Bahareh Behkam^{1, 2}

1. Mechanical Engineering Department, Virginia Tech, Blacksburg, VA
2. School of Biomedical Engineering and Sciences, Virginia Tech, Blacksburg, VA

Corresponding Author: Bahareh Behkam, Email: Behkam@vt.edu

 
INTRODUCTION

Microfluidic devices have been extensively used for
the study of both prokaryotic and eukaryotic cells under well-defined chemical
and mechanical conditions. In many cases, the establishment of steady-state
(temporally invariant) spatially varying chemical gradients within which cells
can be studied over extended periods of time and in absence of any fluid flow is
of outmost importance.

In
this work, we have microfabricated a microfluidic device made of the
biocompatible and photopolymerizable polyethylene glycol diacrylate (PEG-DA). We
have demonstrated that PEG-DA of different molecular weight can be used in
order to achieve customizable chemical gradients. This microfluidic device is
composed of three channels fabricated using photolithography methods and
operates under the principle of diffusion as shown in Fig 1.

Figure 1. Schematic
showing a cross-section of microdevice and SEM image of a 2000 Da molecular
weight porous gel

A solution
of a chemical agent of choice (e.g. casamino acids) in a buffer (e.g. PBS) is flowed
in one of the outer channels while just the buffer solution is flowed in the
other outer channel. Cells (i.e. subject of study) reside in the buffer
solution in the center channel. The chemical agent diffuses through the porous
gel wall to the center channel, generating a well-defined spatially varying
concentration field. The gel wall thickness and porosity can be designed to
accommodate for the required chemical diffusion rate. COMSOL multiphysics
simulations, diffusion coefficient measurements by optical spectrophotometry and
diffusion characterization using fluorescence microscopy were employed to
develop custom-devices with desired gel wall characteristics to determine the
threshold chemical gradient at which Escherichia coli (E. coli)
strain RP437 begin to exhibit a chemotactic behavior towards casamino
acids.
 
METHODOLOGY

Fabrication of microfluidic device

The microfluidic device is fabricated using a two-step
photolithography process (Fig 2). Briefly, the glass surface was first rendered
hydrophilic using oxygen plasma for duration of 5 minutes and then
functionalized using a TPM (3-(Trichlorosilyl)propyl
methacrylate) solution. PEG-DA gels of a
desired molecular weight (700 Da, 2000 Da, 6000 Da or 10000 Da) were exposed to
UV light (EFOS Ultracure,
UV spot lamp, Mississauga, ON, Canada) at
a wavelength of 365nm for 20 s through a photomask with the chosen pattern.

The
fabricated gel pattern is assembled with the other layers of the device; Figure
3 shows an assembled microfluidic chemotaxis assay device.

Diffusion properties characterization

The diffusion coefficients of the desired chemical
agent through the fabricated gels were determined using a Franz diffusion cell¹
as seen in Fig. 4. Known concentration of a chemical agent in buffer present in
the donor compartment diffuses over time across the porous PEG-DA membrane and
into the acceptor chamber. Samples from the acceptor compartment are collected over
time and their absorbance was measured using a UV spectrophotometer. For
chemical agents that cannot be detected by spectrophotometry, the chemical gradient
in the center channel of the microfluidic device was quantified using a
different method². A fluorescent dye solution of known concentration
with molecular weight close to the molecular weight of the chemical agent of
interest was flown in the outer channel and diffused through the PEG-DA porous
walls to the center channel. Fluorescence microscopy images were used to find
the chemical concentration profiles within the center channel and thus the
diffusion coefficient values.  A picture of the fluorescence profile within the
center channel can be seen in Fig 4.

Figure 4. Franz diffusion
cell used for diffusion coefficient measurements.

Figure 5. Fluorescent
gradient profile across microfluidic channel using a 700 Da molecular weight
gel

  RESULTS

Table
1. 
Diffusion coefficient of casamino acid through PEG-DA gel as a function of the
gel molecular weight.

COMSOL simulations were run
to determine the chemical gradient profile in the center channel where the
cells will be inserted. The chemical gradient profile in the center channel
depends on the dimensions of the device, the molecular weight of the gel used,
and the initial chemical concentration in the source channel. In these
simulations, all other parameters were kept the same except the molecular
weight of the PEG-DA gel used to fabricate the device.

The initial chemical
concentration of the chemical diffusing from the source channel was 3.84×10^-4
M, and the chemical concentration in the buffer channel was maintained at zero.
The chemical concentration gradient in the center channel after 1 hour is shown
in Fig. 5. As expected the more porous gels have a higher chemical gradient
slope. This can be attributed to the fact that the chemical in the source channel
diffuses at a faster rate to the center channel, thus creating a steeper linear
gradient profile. 

Figure
6.
Chemical concentration across center channel of microfluidic device

The chemical gradient values
obtained above were verified experimentally using a fluorescein solution of
identical initial concentration in the source channel. Fluorescent images of
the center channel of the device were taken and analyzed using Image J (free
software provided by the NIH). Fluorescein gradients within the center channel
were then determined and compared within 20% of the chemical concentration
gradients obtained through simulations. Table 2 demonstrates the results from
the simulations and the experiments. The differences between the experimental
and theoretical values can be attributed to the fact that the molecular weight
of fluorescein (332 Da) used in the fluorescent quantification of the diffusion
coefficient is different from the molecular weight of the casamino acids (250
Da).

Table 2. Comparison
between experimental and simulation chemical gradients in the center channel.

A chemotaxis sensitivity
assay was conducted using E. coli bacteria strain RP437 (fluorescent
bacteria) towards casamino acids. The molecular weight of the gel used in this
experiment was 700Da. The lowest molecular weight gel was chosen for this
experiment because of the low chemical gradient it generates in the center
channel. E. coli bacteria positively responded to the chemical
attractant present in the source channel. Fig 6. shows the
E. coli bacteria response towards casamino acids in the center channel.

Figure
7.
Response of E. coli bacteria strain RP437 in center channel to a
gradient of casamino acids.

DISCUSSION

                The
fabrication of a microfluidic device with different gradient profiles can be
achieved using different molecular weight gels. Bacterial
chemotaxis studies usually require smaller chemical gradients to be established
in the center channel. The use of the 700 Da molecular weight PEGDA gel can
accommodate for the chemotaxis study of bacteria. The ability to control the
chemical gradient in the center channel can help determine the threshold
chemical gradient needed to provoke the chemotactic behavior of a strain of
bacteria as well as the saturation chemical concentration.  

CONCLUSIONS

In this work, multiple microfluidic devices were
fabricated using a photopolymerizable hydrogel PEG-DA at different molecular
weights. Different gel structures lead to significantly different diffusion
coefficients and therefore distinctly different chemical gradient field within
the main channel. The capability of realizing well-defined and customizable steady-state
chemical gradient fields in absence of fluid flow opens many opportunities for
the study of various behaviors of bacteria and mammalian cells within a
controlled environment.     REFERENCES

1. Wenquan Liang et al.
Journal Chem Phys, 2006, 125, 0447071-0447077.

2. Mingming Wu et al. Lab
on a Chip, 2007, 7, 763-769.