PROGRESS REPORT - YEAR 1
(September 15, 1995-September 15, 1996)
This first year we ordered and received the needed computers, software, video equipment, and basic supplies, and focused on recruiting students. The following students were supported under this grant throughout this period of the grant: Ms. Shannon Hill (senior), Ms. Erin Hines (junior), Mr. Todd Stevens (junior), Ms. Rebecca Weingard (sophomore), Mr. Mark DeSantis (sophomore), Ms. Rebecca Weber (sophomore), Mr. Matthew Smetanick (junior), Mr. Peter Larson (junior).
Computers and video equipment took longer to arrive than we anticipated, arriving in February 1996. Consequently, during the first part of the period we mainly concentrated on adjusting techniques and strengthening the students' theoretical background and understanding of the project.
Adjusting Techniques
The first step is to create computer reconstructions. There are at least five parts to the problem of creating accurate computer reconstructions from serial-sectioned tissues: 1) adequately preserve the tadpole cartilage and muscle, 2) create and mount a complete set of serial sections of embedded tissues, 3) stain the sections to reveal cartilage and muscle, 4) capture and align the separate images of sections in the same registration as in vivo, and 5) display the reconstructed images in a ways that is meaningful to humans. We made progress in all these issues.
1)Preservation: Standard 4% formaldehyde fixation has provided adequate preservation of tadpole chondrocrania and muscle tissues. This is an advantage because formaldehyde is readily available to us and to colleagues. Specimens of species collected at remote sites can be fixed and stored indefinitely.
2) Serial Sections: Serial sections of entire tadpoles skulls were easily obtained from paraffin-embedded tissues. A more difficult problem was ensuring that each section mounted on a slide was retained during processing for staining, and not dislodged and washed off. The best procedure we found was to coat slides with a solution of 2% aminopropy-triethoxylane (Sigma) in acetone. Slides are dipped in this solution for 30 seconds, rinsed in acetone followed by water, and air dried. Sections mounted on slides prepared this way are secure.
3) Staining: We tested several stains to find a combination that would simultaneously enhance contrast of cartilage and muscle. The best is a modified Milligan's trichrome stain, which consist of acid fuchsin, orange G, and fast green FCF or aniline blue, using phophomolybdate and potassium dichromate and mordants (Milligan, 1946, as described in Humanson, 1975). We also found that whole mounted tissues previously stained with alcian blue to reveal cartilage could subsequently be embedded, and sectioned, and the stain persist to show cartilage. Thus, we have a simple method to correlate the appearance of cartilage in a whole specimen with the appearance of the same cartilage reconstructed from serial sections.
4) Aligning: Aligning and registering the captured images of separate sections is a difficult problem for which we have not yet found a perfect solution. The goal is to create a set of fiducial marks independent of the specimen. Most investigators try to co-embed straight and easily sectioned materials such as pine needles, hairs, silk threads, or uniform stripes of liver. None of these proved reliable for us.
We have decided instead to exploit the regular tissue architecture of the central nervous system. Since we are examining the skull, and in these animals the brain occupies most of the cranial volume, the brain is present in nearly every section. Because it is bilaterally symmetrical, corresponding points on lateral halves maintain a regular spatial relationship with each other. Finally, the brain outline does not change shape radically from section to section, allowing adjacent sections to be aligned consistently. Of course, the brain shape depends largely on interior skull shape, so our fiducial marks are not entirely independent of our specimen.
5) Reconstructions: Capturing section images and displaying the 3D reconstruction depends largely on the software available. As we expected, we are using the public domain software NIH Image to capture and align digital images of serial sections. We initially captured complete grayscale images of each section and used NIH Image to create the reconstruction, called a "stack". However, it was too difficult to observe the cartilage separately from the over- and underlying tissues. There was too much information in each image to extract the shape we wanted to see.
To avoid this problem, we created another set of images that contained just the outlines of the cartilage elements. Outlines were created by a talented and dedicated team of undergraduates who used a computer mouse to painstakingly trace each cartilage element as it appeared on screen, and erase all the parts of the image that were not cartilage outline. This second set of cartilage outlines was then used to create a second stack.
Reconstructions from outlines are clearer than those from whole images. the downside is that the spatial accuracy of the reconstructed borders depends on the accuracy of the outline procedure. Students had to use their judgment and dexterity to decide where to place the border.
Reconstructions were displayed in one of three software programs. NIH Image has limited ability to display stacks, so for most purposes we have used either MacStereology (Ranfurly Microsystems) or Slicer (Fortner Research). MacStereology is relatively easy to use and manipulates images well, but suffers from a limit of 200 sections per stack. Some of our reconstructions contain 300-400 sections or more. Slicer is state-of-the-art display software for Macintosh but requires large amount of RAM. Consequently, we needed to upgrade our computers to at least 32-64 MB to display the smallest of stacks. It now appears that Slicer will be our software of choice for 3D image display.
Strengthen Theoretical Background
Students needed to be brought up to speed in their theoretical backgrounds to be able to effectively work on the project. Two areas were emphasized: 1) mathematical descriptors and 2) anatomical elements.
1) Mathematical Descriptors: The focus was on theoretical background material relevant to the description of shape, starting with shapes enclosed by simple closed curves in the plane and discussing continuos and discrete methods for computing the perimeter of the shape and its enclosed area. Students were given a proof of the isoperimetric inequality to discuss the so-called isoperimetric deficiency index which gives a measure of the deviation of the shape from the circular.
Considerable time was spent on the representation of plane curves as envelopes of either families of lines or families of circles. We derived formulas for the points on the envelope when the family is given and discussed the fact that our problem involves finding a family of circles when we already know the envelope. Students spent time working with envelopes of planes trying to extend their formulas and intuition to three dimensions and read papers concerning the existence of bitangent spheres for surfaces in 3-space. Particular attention was placed on a foundational paper ("Biological shape and visual science" Blum, 1973) for shapes termed "Blum ribbons" which consist of a symmetric axis and a Blum radius function along the axis. The idea is to represent a given shape as the envelope of a family of circles whose centers lie along the symmetric axis and whose changing radii are given by the Blum radius function. Part of this reading was to have students go from formalizing the statements of Blum's results to proving them; this allowed the students to develop connections between different ideas.
One student proved that Blum's symmetric axis is an axis of infinitesimal bilateral symmetry for the boundary of a shape and another student worked to compare the complexity of shape descriptions using the symmetric axis as opposed to radial Fourier series. The latter derived a formula for the Blum radius function of a shape whose radial Fourier series consists only of cosine functions; using complex number, he also worked through a simple criterion for a curve given as a radial Fourier series to possess a finite group of rotational symmetries. Other papers related to the symmetric axis description of shape, applications of the symmetric axis in archeology and medicine. These help the students to understand the symmetry-curvature duality theorem, extensions of the symmetric axis idea to three dimensions, and how different measures of lengths and widths are used in biology.
2) Anatomical elements:
It was immediately clear that for the students to interpret and understand the serial cross-sections, they needed to have a solid understanding of the three dimensional, whole mounted structures before they were cross sectioned. This required training each student in processing a second set of tadpoles. This second set of tadpoles was not cross sectioned but was cleared and double-stained for bone and cartilage with alizarin red and alcian blue respectively. Students learned to determine the developmental stage of the tadpoles by comparing the external morphology of their specimens to the Gosner Table of Normal development (Gosner, 1960). After staging them, each tadpole was measured (total length and body length), skinned and eviscerated. Subsequently, a standard procedure was followed with each specimen to clear the muscle and to stain bones and cartilages. Once the specimens were cleared and stained, students spent time analyzing and drawing the chondrocranial anatomy. This required them to read and compare the skeletal structures that they were observing with those reported in the literature for other anurans. Furthermore, whole mounted tadpoles were prepared in different stages of development for each of the species to allow the students to visually see and understand the morphological changes that take place during development. This is important because when the students are reconstructing images, they need to have a clear idea of the basic anatomy at that particular stage, since not all specimens cross sectioned are at the same developmental stage.
Presentations and Publications
The work done during the first year of this project allowed us to present and submit several papers for publications.
Rebecca Weingard was able to give a partial account of her work on circle-preserving transformations at the Meeting of the American Mathematical Society held in Seattle, WA, August 9-12, 1996, (abstract attached).
Shannon Hill worked throughout the year on describing the skull anatomy and skeletogenesis of a poison arrow frog (Dendrobates auratus). She was able to present the results of her work in two different occasions. Shannon presented her preliminary data at the Student Research Symposium, University of Richmond (April 1996) and a completed project at 39th Annual Meeting Society for the Study of Amphibians and Reptiles (July 1996) (abstract attached).
Shannon Hill is co-authoring a manuscript with Rafael de Sá. The manuscript and illustrations are completed and will be submitted for publication in the Journal of Herpetology, in late December 1996. The specimens used have been deposited at the Smithsonian Institution and we are waiting to receive the catalog number assigned to them to submit the manuscript.
Rafael de Sá worked on and submitted a second manuscript co-authored with Esteban Lavilla. The manuscript, entitled "The tadpole of Pseudis minuta (Anura: Pseudidae), an apparent case of heterochrony," was submitted for publication to Amphibia-Reptilia in August, 1996, and it has been accepted for publication.
Current and Pending Support of Senior Personnel
Since we applied to the C-RUI, Dr. Gary Radice and Dr. Rafael de Sá have each completed previous grant projects. Dr. Radice research was supported by an NIH project, Grant # R15 AR41567-01 "Myosin Isoforms: Fiber types and thyroid regulation" from 5/92 to 5/95, and extended to 5/96. This project is now completed.
Dr. Rafael de Sá had received an NSF-ILI, Grant # DUE-9451845 "Molecular Evolution: Integration of Theory and Practice" from 7/94-12/96. This project is now completed and a final report has been submitted to NSF.
In addition, Dr. Rafael de Sá has submitted an RUI proposal entitled
"Evolutionary relationships within the Neotropical frog genus Leptodactylus".
Funding for this proposal is pending. This project is a DNA sequence analysis
of the genus Leptodactylus and it is completely unrelated to the
C-RUI project. Furthermore, if funded, it will not detract from Dr. de
Sá's commitment to the present project, particularly since the submitted
proposal includes a reduction in teaching load which has been approved by
the University administration.
Last modified January 9, 1998.
Return Home