Technical Approaches


Design of Micro & Nanofluidic Devices

Intracellular signal transduction involves a variety of molecular-level events occurring dynamically in space and time. To understand the signaling processes that, when integrated, generate specific cellular response, one often needs to perform multiple experimental perturbations of signaling network.  The low throughput of traditional manual experimental technique is the limiting bottleneck in the collection of data necessary for constructing quantitative models of signaling transduction.  Commercial “robotics” is too expensive and space consuming for typical laboratories, and often only capable of performing a single task.

We therefore set out to design PDMS based microfluidic devices and develop the accompanying protocols for employing these devices to culture and extract data from a variety of biological model systems:  primary neuron cells, tumor cells, stem cell, immortalized cell lines, bacteria, yeast, and Dictyostelium.  The small scale and precise chemical delivery capability conferred by microfluidics enabled us to perform experiments with complex perturbation over the microenvironment the cells experience and thus revealing novel and biologically significant phenomenon.  We then proceed to probe the underlying fundamental mechanism by using our patented automation devices that allows rapid and inexpensive testing of a large number of stimulation conditions, resulting in data that facilitate the development of computational models for elucidating the complex dynamics of signal transduction.
Through close collaboration with Kahp-Yang Suh at Seoul National University, we have started to probe model systems at the nano scale with nanostructured polymeric substrate.  In addition, we have interfaced nanotopology with microfludic structures, creating cutting edge tools for investigating novel mechantransduction phenomenon on the nanometer scale.
Our research laboratory is situated inside the same building that houses a sophisticated microfabrication laboratory (see Facilities for further details).  Thus our turn around time from designing a new microfluidic chip to finished product is typically 1 work day.  We also possess an array of sophisticated microscopy systems that are coupled to our custom microfluidic setup (see Automated Image Acquisition & Analysis), along with cell culture laminar flow hoods for preparing our in-chip experiment.  Our uniquely integrated environment enables us to rapidly create microfluidic tools that closely fulfill the needs of the mainstream biologists.

Automated Image Acquisition & Analysis

Our lab has several fully automated epifluorescent deconvolution microscopes, ranging from x-y-z stage, filter and multi-camera control. The scopes are driven through 3I’s Slidebook and custom software. When coupled to our robust microfluidic platform, a single experiment can generate thousands of images with exquisite spatial-temporal detail. Following data acquisition, we seek to determine quantitative and significant relationships from our image data, measuring protein expression, cell morphology, and motility patterns. The lab is actively implementing and investigating a large variety of imaging algorithms, especially PDE based segmentation involving parametric and geometric active contours.

Integration of Model & Experiment

In our lab, experimental and mathematical approaches are integrated to explore the signaling pathways within a single cell and communications among cells. Microfluidic systems, combined with other techniques (eg. real time PCR, live cell imaging, Western blots, and immunocytochemistry), are utilized to better define the environment of cell culture and to characterize the molecular events more quantitatively in both spatial and temporal dimensions. Computational models are developed based on experimental results, providing better understanding of the underlying mechanisms and making predictions to be validated by further experiments. With the power of integrated experimentation and simulation in mathematical models, we have successfully developed insight into the roles of MAPK cascades in pheromone sensing of yeast and NF-kappaB pathway in inflammation. We expect this approach will also be fruitful in other areas we are studying, such as angiogenesis, axon guidance, tumor metastasis and cell morphogenesis.

Mathematical Model Analysis

We use a wide variety of different mathematical approaches to address many aspects of the biological questions we are currently investigating.  Systems of ordinary differential equations (ODEs) are used to predict the behavior of complex intracellular signaling pathways and validated through both conventional and novel experimental approaches.  For biological questions that are spatially dependent, such as chemotaxis and axonal guidance, we use both finite element and finite volume approaches to determine the significance of the polarization of signaling molecules. Depending on the question at hand we have also utilized a finer scale analysis involving Monte Carlo simulations tracking individual interacting molecules or agent based modeling aimed at describing multiple interacting cells within a population.  In addition, we also use network analysis to model regulation of gene transcription and validate results using molecular biosensors.

Molecular Biosensor Development

Analysis of intercellular signaling events requires development of novel probes for efficient tracking of localization and activity of various molecular players. We are using modern molecular biology tools to create a library of molecular biosensor that would allow us to image multiple signaling events, on a large scale (hundreds to thousands of cells) simultaneously. When coupled to our micro- and nano-scale device technology, utilization of the biosensors results in precisely controlled datasets providing a wealth of information on signaling events expressed in the form of imaging information with exquisite temporal and spatial resolution.