DOE Low Dose Radiation Research Program Workshop I Abstracts November 10-12, 1999, Washington, D.C.

6. Mechanisms of Tissue Response to Low Dose Ionizing Radiation Exposures: Bioinformatic Tools for Multiparametric Image Analysis

M.H. Barcellos-Hoff, B. Parvin, M.J. Bissell, C. Park, and F. Wang
Building 74, Room 355, Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, CA 94720
mhbarcellos-hoff@lbl.gov

Summary: This project uses highly sensitive and precise cell biology techniques and newly developed computer-assisted tools to microscopically map complex patterns of radiation-induced proteins.

Abstract: Tissue response to radiation, and hence risk, is a composite of genetic damage, cell loss and induced gene products. We have defined a series of events that occur from 1 hr and 4 weeks after whole body exposure to doses of 0.1 Gy to 5 Gy 60Co-g radiation in liver and mammary gland, which indicate that remodeling is a general and rapid consequence of irradiation but that the pattern and composition of the radiation-induced ECM is tissue specific. These studies suggest that cell type and interactions play a prominent role in determining the character and composition of radiation-induced remodeling, which in turn represents a fundamental process initiated by radiation exposure that may promote tissue dysfunction. Alternatively, such events may also be necessary for tissue recovery in a as yet unappreciated way.

Aim 1: Quantitative comparison of radiation-induced changes in ECM composition, growth factor expression and activation, and integrin using immunofluorescence localization as a function of dose (0.1 -5 Gy) and time (1 hr-14 days) of mice from resistant (C57/bl6) and sensitive (Balb/c) genetic backgrounds.

We have analyzed radiation-induced proteins in the C57/bl6 mice that are heterozygotes for the transforming growth factor-ß 1 (TGF-ß ) gene. This strain of mice is known to be relatively resistant to mammary carcinogenesis but sensitive to radiation-induced fibrosis. Interestingly, qualitative assessment of radiation induced proteins such as collagen III and plasminogen activator shows that TGF-ß heterozygotes appear similar to Balb/c mice, which are known to be relatively sensitive to mammary carcinogenesis but resistant to radiation-induced fibrosis.

Aim 2: Determine whether low dose radiation exposure affects the ability of human mammary epithelial cells to maintain normal interactions with the microenvironment by studying integrin and E-cadherin expression as a function of dose (0.1 -5 Gy) and time (1 hr-14 days) in 3-dimensional cell culture.

We have examined radiation-induced changes in integrins in HMT-3522 S1 human breast epithelial cells cultured within an extracellular matrix. Confocal microscopy revealed that immunoreactive 6 and 4 integrins localized on the basal surface of are perturbed following g - radiation exposures of less than 2 Gy. The cells that survive radiation exposure show persistent dysregulation of these integrins and a potential ligand, collagen type IV. TGF-ß amplifies this effect.

Aim 3: Develop a bio-informatics framework of integrated image acquisition, annotation, and hierarchical image abstraction (i.e. data model) to create a database that registers localization and intensity information about multiple targets along with positional references and morphological features. Statistical and visualization tools will be integrated to allow hypothesis testing and data mining.

This multidisciplinary approach will result in the construction of phenotypic database of radiation- induced proteins in several animal models and human epithelial cell culture. The identification of critical biological responses to low dose radiation exposure can then be integrated into computational models of radiation risk. Image bio-informatics will also be valuable for functional genomics aspects of the Human Genome Project by providing integrated analysis of complex biological information in the relevant tissue context.