Microcirculation, tissue engineering, regenerative medicine, intravital microscopy, computational flow modeling
Profile
|
Our mission is to solve important problems in regenerative medicine and tissue engineering by applying first principles of structural biology, adaptive physiology and tissue engineering. The major focus of our laboratory is the development of techniques that allow us to study microcirculatory adaptations in inflammation and tissue repair. Imaging plays an important role in our approach to studying these complex processes because we believe that “seeing is understanding.” Imaging allows us to observe the complex interactions responsible for adaptive change.
Back to the top |
Projects
The mechanisms responsible for adaptive change interact on multiple levels--from small molecules to whole organs. Our laboratory is addressing this challenge in three ways:
- structural biology
- dynamic physiology
- computational modeling
Structural biology
Structural biology is a phrase often associated with the study of molecular shape. We consider structural biology in a broader sense; namely, the 3-dimensional arrangement of molecules, cells and tissues. The underlying structure is important because it defines the physical parameters, or constraints, of dynamic processes as well as establishes the initial conditions for computational modeling. An under-appreciated observation is that complex biological processes can involve dynamic changes in structure—for example, blood vessels can undergo marked inflammation-induced changes in shape and number within days to weeks. Understanding the process of structural adaptation holds significant promise for developing novel approaches to tissue injury and tissue engineering.
Our laboratory uses a variety of tools to investigate blood vessel structure. In addition to using standard 2-dimensional analysis of tissues (the classical “tissue section”), we believe that the 3-dimensional relationships of tissue elements (e.g. molecules, cells, networks) is essential to understanding complex biological processes. In close collaboration with Univ-Prof Moritz Konerding, Johannes-Gutenberg University, we use corrosion casting and 3-dimensional scanning electron microscopy to define the shape and size of microvessel segments and networks. Corrosion casting is a technique for creating plastic-like (polymer) casts of blood vessels from 6um to 100um (50um = diameter of a human hair). The surrounding tissue is “corroded” away leaving a pure cast of the blood vessel (Image 1). The cast not only provides a precise assessment of microvessel size and shape, but it provides a means for exactly defining the number of blood vessels per unit of tissue. This exacting assessment of blood vessels is crucial for studying the process of blood vessel growth (angiogenesis) as well as the results of potential treatments.
A limitation of corrosion casting is that the surrounding tissue is lost to subsequent testing. Particularly in inflammatory conditions, the identity and location of surrounding cells provide important clues to understanding the process of blood vessel growth. To provide this information, we have developed a technique of “vessel painting”; that is, labeling the lining of blood vessels with brightly fluorescent dyes. Using a specially designed microscope to acquire images at different levels throughout the tissue (so-called “optical sections”), and a computer to reconstruct the information into 3-dimensional images, we can obtain structural information that complements scanning electron microscopy (Video 2). Advances in the film and gaming industries has contributed to the development of very sophisticated hardware and software that we apply to imaging the microcirculation. Our imaging is stored as “voxel” data (derived from “volume” and “pixel”). By representing imaging data in 3-dimensional space, voxel data allows for complex analyses including perspective “fly-throughs” that allow for an entirely new understanding of spatial relationships in vessel networks and inflammation.
Although we analyze cell and tissue structure down to the level of individual molecules, this detailed analysis risks missing important network relationships. To illustrate, it is difficult to understand the development of rush-hour traffic jams by focusing on only one traffic light or even one street. To provide molecule-level detail in networks that extend over two millimeters or more (an imaging “zoom” equivalent to 20,000,000x), we analyze areas in high magnification and “stitch” them together using computer software algorithms. Similar to “panorama” views performed in commercial photography, these very large image files allow us to “zoom in” to identify potential causes of patterns we see on the network level (Image 3).
Dynamic physiology
Dynamics is a branch of classical physics concerned with the motion of bodies. We use the phrase dynamic physiology to describe the space and time relationships of cells in the blood and tissues. In most cases, we are interested in the motion of cells in the blood and their fate in inflammatory tissues. In other cases, we also concerned with the forces that may have an effect on cell motion and blood vessel structure. For example, tracking the motion of cells can help us understand why inflammatory cells exit the blood in certain vessel structures, whereas understanding the forces involved in inflammation-induced changes in blood flow may allow us to mimic these forces in treatment strategies.
To study blood flow and inflammation in living tissues, our laboratory has developed microscope techniques that allow us to examine tissues even when they are swollen and inflamed. Using a variety of multi-colored fluorescent labels, we can track inflammatory events in “real-time.” Some of our labels are invisible to the human eye (e.g. infra-red labels), but have long wavelengths allowing the light to penetrate thick tissues. These labels are detectable by very sensitive cameras that convert the light to visible images. These images can be recorded at very high “frame rates”---in many cases, more than 3 times faster than a standard movie. Thousands of these images can be recorded by special computers designed to rapidly store large amounts of information.
Because of the large size of the data streams produced by imaging data, a basic approach of our laboratory is to “mine” the data. Software-based data mining approaches can provide simple tools to facilitate data analysis. For example, a computationally-derived summary of the flow patterns in a vascular network can be superimposed on a video stream to provide a summary of flow patterns (over time). Areas of increased flow and absent flow can be readily identified and the potential reasons for these differences can be studied (Video 4). More sophisticated tools, allowing large amounts of data to be analyzed, require the application of spatial statistical tools. This new and important area of biostatistics (see our colleague Chris Paciorek’s website) is a strong interest of our laboratory.
Computational modeling
Computational modeling is a scientific approach that uses mathematical models and extensive computational resources to study a complex system. Computational modeling has become crucial to modern science because traditional methods of studying one variable at a time (formalized by John Stuart Mill in 1843) cannot cope with multiple interactive variables. In complex systems, it is important to study the relationships of multiple components of a system (or network)—primarily because interventions often have unexpected consequences. For example, a small “input” into a system may produce a large “output.” Conversely, a large “input” may produce a small “output.” These complex systems (so-called “nonlinear systems”) can be illustrated by vascular networks such as the colon mucosal plexus.
Working with Dr. Akira Tsuda, Harvard School of Public Health, we have developed computational approaches to studying the inflammation-induced events in the microcirculation. Computational modeling of blood flow is challenging. One of our approaches is to treat blood as a continuous homogeneous medium. The idea of a continuous medium allows blood to be studied using traditional mathematical models (e.g. Navier-Stokes equations). The limitation of this approach is that it treats “blood” as a homogeneous medium and often “fudges” the behavior of individual blood components (one example is the idea of “apparent viscosity”). A complementary approach to continuum-based models is called “discrete particle dynamics” (DPD). DPD treats blood components as micron-sized (mesoscopic) particles that are small enough that the motion of each constituent of the blood can be tracked.
An valuable scientific contribution of computational modeling is that a good model makes unexpected, but testable, predictions. Some of our model’s predictions have been difficult to test in the living animal, so we have developed an “artificial capillary” system. Using the remarkable silicon-based polymer PDMS (polydimethylsiloxane), we can create complex geometric structures to test our theoretical predictions (Video 5). PDMS, used in applications from contact lenses to complex microfluidics devices, has remarkable physical properties that allow us to study the and behavior of fluids in very controlled and reproducible conditions.
Image 1
Scanning elecron micrograph of a corrosion cast of the mouse colon. The honeycomb network (luminal surface) is the mucosal plexus. (Back to project description)
Video 2
3D structured illumination confocal reconstruction of a small area of the colon microcirculation. The vessels were labeled with a green lipophilic fluorescent dye (DiO), acquired in a “whole mount” of the colon wall, then reconstructed into a 3D model using rendering software. (Back to project description)
Image 3
Region of the colon mucosa “mapped” by high resolution images of the mucosal crypts (blue) and microcirculation (green). The use of “mosiacized” images of the mucosa allows for the investigation of spatial patterns and the application of spatial statistics. (Back to project description)
Video 4
Send Feedback to: 
