High Speed Computation Aids Bone Regeneration Research

Approximately 33,000 U.S. soldiers have been injured in Afghanistan and Iraq since the start of the “war on terrorism” in 2001.  Many have suffered severe injuries involving broken, fractured and shattered bones. Millions more Americans experience bone damage each year due to osteoporosis, arthritis, breaks and fractures. If doctors could enhance bone repair, these individuals could quickly return to their normal routines and reclaim their quality of life.

That aim is at the heart of an ongoing, multidisciplinary research project linking researchers at the UConn Health Center and Storrs campuses, who seek to advance our understanding of bone regrowth in mice as a crucial step toward stimulating human bone regeneration.

The team, led by David Rowe, M.D., Director of the Center for Regenerative Medicine & Skeletal Biology and a professor of Reconstructive Sciences in the School of Dental Medicine at the UCHC, was recently awarded $2.7 million from the U.S. Department of Defense (DOD) to conduct multi-pronged research that builds upon their previous studies.  The project involves long-time collaborator Dong-Guk Shin, a professor of Computer Science & Engineering (CSE), Seung-Hyun (Sean) Hong, an assistant professor-in-residence (CSE), Cato Laurencin, M.D., Director of the Institute for Regenerative Engineering and a professor of Chemical, Materials & Biomolecular Engineering, Alexander Lichtler, associate professor in the Center for Regenerative Medicine and Skeletal Development, and David Goldhamer, Director of the Center for Regenerative Biology at the Storrs campus.

The cross-disciplinary nature of the collaboration underscores the complex and evolving landscape of modern medical research. “By bringing together experts from different disciplines, perspectives and experiential backgrounds – and applying a variety of tools – we are able to take a more holistic and replicable approach to solving medical mysteries,” remarked Dr. Rowe.

With respect to the project challenge, he said, “We know that in a mouse, we can take healthy bone marrow cells, cryogenically freeze them, thaw them and insert them into the mouse to stimulate bone repair. If we could extract and freeze progenitor cells from soldiers before they depart for the battlefield, we could thaw and insert these cells into their original owner when the soldier returns injured, to supplement the body’s natural healing processes. Progenitor cells play a starring role in injury repair, but massive injuries require many more progenitor cells than the body can produce.  We need to find a new source of these cells.”

Computational Boost

Before human studies can be undertaken, however, researchers must understand how to predictably excite bone regeneration in lab mice.  An earlier DOD-funded project on which Drs. Rowe and Shin collaborated, demonstrated that mouse progenitor cells injected at the site of a bone injury would differentiate and rebuild the bone within mice.  A vital aspect of that research, Dr. Rowe noted, was the systematic, objective data analysis provided by Drs. Shin and Hong.  “Their computational algorithms were instrumental in helping us establish reliable, systematic and unbiased analyses of our test results across different approaches and time.  The computer is our ‘neutral observer’ and enabled us to build a reliable mouse model of the process.”

Dr. Rowe believes this approach was a decisive advantage in securing the second round of DOD funding.  The new project involves four integrated thrusts: the development and testing of an array of human progenitor cells; the development and testing of different “scaffold” materials to which the new bone cells will attach themselves; and the evaluation of different techniques for optimizing bone regrowth while minimizing the inappropriate formation of bone within tissues. At each stage, the team will rely on the computational expertise of Drs. Shin and Hong to provide rapid, uniform and objective analysis of the data.

While the previous study involved the use of mouse cells in mice, Dr. Rowe noted the new project will involve the insertion of human cells in lab mice that have been genetically altered to accept human cells without rejection. These mice also exhibit a characteristic fluorescence of certain cells when exposed to light, enabling researchers to differentiate the naturally-produced mouse cells from the human cells in bone regrowth.

Study Format

In laboratories at the UCHC, two holes will be made in the skull of each test mouse.  One site will serve as the control, while the second will serve as the experimental site.  At the experimental site, the researchers will test different combinations of scaffolding material as well as different variations of so-called “induced pluripotent stem cells” (iPSCs), a form of adult somatic cell that can be readily harvested and engineered to differentiate into the type of cell desired, such as bone or muscle.  Approximately two weeks after inserting the scaffold/iPSC combination – a period during which bone regrowth is typically concluding – thin sections of frozen mouse skulls will be sliced, microscopically scanned into images and transmitted electronically from the UCHC to Dr. Shin’s computer lab.

Dr. Shin said that traditionally, such images have been analyzed by specialists in a laborious manual process that is prone to human error and subjectivity.  The computational algorithms developed by Drs. Shin and Hong perform quantitative analyses of the microscopic images that are accurate and systematic across the entire array of images, ensuring reliable and objective analytical results.

Not only can the computers interpret the data more objectively than humans, they also can do so at relative lightning speed.  Dr. Shin explained that the transmitted images are translated into quantitative numbers that enable the researchers to measure the skull bone growth at both the control and test sites; to differentiate whether the new bone cells were produced by the mouse’s own progenitor cells or the human iPSCs; to measure the differences in cell growth among different scaffold structures; and to identify the errant development of bone cells within tissues outside of the skeleton.

Dr. Shin’s resources for high throughput computation received a much-need boost recently, thanks to competitive funding provided by UConn Provost Peter Nicholls and Engineering Dean Mun Choi, supplemented by Dr. Rowe.  The funding allowed Dr. Shin to purchase of a 144-core Dell cluster computer with 24 nodes. “This cluster will allow us to massively parallel process data transmitted from the Health Center in the form of images of thin sections of the mouse skulls showing the bone regrowth. Using the GFP mice, we can differentiate the host cells from those of the donor, and compare the bone regrowth in the control versus the experimental areas on the skull. Quantification is critical. We must be able to quantify the difference, repeatedly, between the iPSC treated site and the control site. This requires that we test the procedure on many mice to obtain repeatable, reliable results. This increases our confidence level.”

Before, remarked Dr. Shin, “One image might take the computer a week to translate into data points. With our new cluster, I hope to turn the data around in two days.”

With researchers located on two campuses located approximately 40 traffic-congested minutes apart by car, the teams maintain regular face-to-face contact through weekly e-meetings enabled by remote telecommunications equipment. Dr. Rowe noted, “The integration of disciplines in this project has been crucial to its success.  So, too, has been the availability of the teleconferencing network, which enables and leverages our communications, eliminating the disadvantage of distance and aiding our successful research collaborations.  Certainly, UConn’s support of these teleconferencing systems has been an important tool for us, and the University would be well served to expand its support for this highly effective communications technology.”

Categories: algorithm, award, emagination, Headline