|
|
     |
|
|
|
by Neil Graf
Evolution's most complex and versatile achievement1, the human nervous system, provides a challenging research frontier for scientists. The nascent interdisciplinary field of neuroscience seeks to study the development, structure, and function of the nervous system in humans and other organisms. Although traditionally based in the life sciences, as the field of neuroscience matures, the need for quantitative modeling and insights from the physical and computational sciences grows according to Professor Timothy J. Ebner, head of the University's neuroscience department.
To foster training in this area, the National Science Foundation funds the Integrative Graduate Education and Research Training Program (IGERT) at the University. This program combines the knowledge, skills, and expertise of three existing resources on the Minneapolis campus: the Graduate Program in Scientific Computation, the Graduate Program in Neuroscience, and the Minnesota Supercomputing Institute. Participating faculty from many departments across the University have that include molecular modeling of receptors and ion channels, imaging and eigenvalue problems, single neuron models, motor control neurophysiology, modeling visual information processing and perception, neural networks and evolutionary algorithms, and problems in information control. IT graduate students in departments with participating faculty--biomedical engineering, chemistry, physics, chemical engineering, computer science, or math--are eligible to apply for the IGERT fellowships, as these departments have participating faculty.
One specific example of IGERT research in computational neuroscience is the work being done by Dr. Tony Varghese in the lab of neuroscience assistant professor Linda Boland. Varghese studies the kinetics of potassium ion channels using the mathematical theory of Markov models. This research requires skills in molecular biology, neuroscience and physiology, chemistry, and mathematics.
Ion channels are proteins that span a cell membrane to permit specific ions to enter and leave the cell. This flow of ions triggers the electrical activity that transmits signals along cells. Last December Varghese and his colleagues published a study of the effect of premature stimulation of HERG K+ channels2. HERG is a human gene, which manifests itself as a potassium (K+) ion channel in the brain and heart. The HERG K+ channel demonstrates some unusual kinetics that are consistent with their putative role of providing people protection by suppressing cardiac arrhythmias sparked by premature heart beats. According to this study, ventricular arrhythmias are responsible for over 300,000 deaths per year in the United States. These statistics provide a sense of the enormous benefit this research could bring to medicine.
Varghese and his collaborators compared experimental findings about the effects of premature stimulation on HERG K+ channels with computer simulations using Markov models. Markov modeling is a branch of statistics that describes a physical system that can exist in a number of different states. The transition between states is memory-less, the future state only depends on the current state.
The researchers used cutting edge molecular biology techniques to transfect a human HERG gene into a cultured Chinese hamster ovary cell. Then they employed a technique used extensively by neuroscientists called patch clamping. Patch clamping can be used to study the current flow, or ionic activity, of a single ion channel or a large number of channels in a single cell. A very fine hollow glass micro-pipette tip filled with a solution that closely mimics intracellular fluid is pressed against the cell's surface where a negative pressure differential is applied to form a pressure seal between a small patch of cell membrane and the pipette tip. The patch is then ruptured to allow an electrical connection between the inside of the cell and the pipette.
Neuroscientists can then use sophisticated electronic equipment to monitor ionic currents through the channels. Currents as small as a few picoamps can be measured, and the technique exhibits microsecond time resolution. With such a high degree of resolution for current and time, scientists can observe how fast the channel opens and closes, how frequently it opens, and how long it stays open. One can get a sense of how remarkable this technique is by observing Figure 1, which shows the response of HERG K+ channels during delivery of different cardiac signals to the channels.
For the study's modeling component, researchers set-up a Markov state model as a system of non-autonomous ordinary differential equations, with an algebraic equation to represent the conservation of states property of Markov chains. Figure 2 illustrates the differential equations being used in the model. In section A of this diagram, the circles with C0, C1, C2, O, and I represent the different states or three dimensional protein shapes of the HERG ion channel. The Greek symbols above the arrows represent rate constants of the forward and reverse configurations. Anyone who has taken a second semester, general chemistry course can appreciate the similarity of the representation in Fig. 2A showing different ion channel states to that of a chemical reaction in equilibrium.
The Markov model computer simulations were found to agree with the data obtained in the patch clamp portion of the experiment. Varghese is presently following up on this report with a computational study of changes in HERG kinetics when the channel interacts with other proteins with disease mutations.
Because research in this area is diverse and multi-faceted, it is rare to find a researcher capable of carrying out every aspect of an experiment like the one described above. Instead, groups of highly talented individuals with expertise in disciplines such as molecular biology, neuroscience and physiology, chemistry, and mathematics, come together to understand complex biological phenomena.
"The Computational Neuroscience Program at the University is aimed at producing a group of Ph.D's each of whom will have interdisciplinary training and research experience," says Varghese. "The fellows in the program are the best in the country. The opportunities before them are tremendous. They are bound to make a tremendous impact on the field of Neuroscience. We should see the shock waves in the next three to five years."
Another interdisciplinary research project underway within the computational neuroscience program can be found in Professor Timothy J. Ebner's lab. Ebner, a neurophysiologist and head of the neuroscience department, studies the cerebellum and its role in an organism's movements. Characterizing and elucidating the spatiotemporal patterns of neuronal population activity in neuronal networks is crucial to understanding the cerebellum, as well as the nervous system as a whole. Understanding the awesome parallel wiring patterns formed by neurons in the brain will help physicians to better treat patients with brain lesions and other nervous system abnormalities.
One aspect of Ebner's research is to study the activity of specific neuronal architecture within the cerebellum of rodents in response to peripheral stimulation, using optical imaging techniques. In one study3, researchers used tungsten microelectrodes placed in the cerebellar cortex, brainstem, and the ipsilateral vibrissa (whisker) pad to provide stimulation. The neuronal activity resulting from this stimulation changes the pH within the affected cells. The researchers used a pH sensitive dye called Neutral Red to map the location of active neurons. A sophisticated setup with a fast, high-resolution charge-coupled device (CCD) camera recorded the changes in pH. The optics was an integral aspect of the signal recording technique used in the experiment.
The general schematic of the basic recording technique is shown in Figure 3. The setup is based on the optical principle of epi-fluorescence. Epi-fluorescence is the process of emitting at one wavelength of light, and observing a different wavelength being emitted. Neutral Red's optimal excitation wavelength is 546 nm. After excitation, the signals pass through a dichroic mirror and an emission filter for wavelengths greater than or equal to 620 nm, which contains the peak emission for Neutral Red. The signal finally enters the CCD camera, which records rapid digital sequential images before, during, and after the stimulation. Researchers extract the optical response to the stimulation by subtracting a non-stimulus frame from frames obtained during and after the stimulation. To quantify the optical responses, a two-dimensional fast Fourier transform analysis removes undesired horizontal artifacts like blood vessels, and high spatial frequency noise, leaving the desired signal intact for scrutiny. Researchers then use various computational techniques to characterize the optical images; for example, principle component analysis helps define the independent spatial patterns of activity present.
More recently, Ebner's group assembled an electronic device to drive a piezoelectric bimorph load. This set-up will provide a more natural means of stimulating rodent whiskers. A piezoelectric bimorph is basically a capacitor that will deflect if a voltage is applied between the two plates. If mounted as a cantilever, the bimorph will bend such that it acts as an actuator, capable of moving rodent whiskers in a periodic fashion. It turns out that a rodent's cerebellum responds best with stimulations to the whisker pad delivered at a frequency between 6-8 Hz. To drive the bimorph at this frequency, the researchers designed a virtual function generator on LabView, a computer using the software package sold by National Instruments.
Computational neuroscience is a highly dynamic and interdisciplinary area of research that spans engineering, mathematics, and the physical and biological sciences. Though researchers bring diverse expertise, they are all united by the common goal of solving scientific problems that would be extremely difficult to pursue as individuals. The computational neuroscience program is an ideal program for those graduate students whose interests lie in multiple areas and wish to conduct research that benefits people's lives.
Kenneth Reinert, a graduate student in the program, agrees, "The Computational Neuroscience Program is probably the most exemplary of the NSF IGERT training grant programs."
FOR MORE INFORMATION:
www.compneuro.umn.edu
Program's Administrator, Kathleen Clinton
(1) Crossman A.R., Neary D., Neuroanatomy 2nd, Churchill Livingstone, 2000.
(2) Yu, Smith, Varghese, et.al., Journal of Physiology, Effects of premature stimulation on HERG K+ channels, 537-3, pp.843-51, 2001.
(3) Hanson C.L., Chen G., Ebner T.J., Neuroscience, Role of climbing fibers in determining the spatial patterns of activation in the cerebellar cortex to peripheral stimulation: an optical imaging study, 96-2, pp.317-331, 2000.
