Identification of quantitative second-harmonic generation imaging metrics for collagenous tissue

Second-harmonic generation (SHG) imaging has high specificity to collagen, sub-micrometer resolution, and optical sectioning capability. It is applied to various unstained collagenous specimens to explore the microstructure. Since collagen is one of the most abundant proteins in biological tissue, understanding its spatial organizational structure could lend insight into its intrinsic mechanical properties and thus its potential function in various biological systems. It therefore is important to extract information on collagen organization from the obtained SHG images quantitatively. Previously, our group, the PROBE Lab at Illinois, developed a technique using (spatial) Fourier transform combined with SHG imaging (FT-SHG) to measure the collagen fiber orientation quantitatively. Recently, I expanded this approach by adapting new parameters on two-dimensional (2D) images and significantly broadening the three-dimensional (3D) analysis of 3D-SHG images. Ultimately, these parameters such as circular and spherical variance become the building blocks for a catalog of useful quantitative SHG metrics. In this thesis, I identified measures and introduced the corresponding quantitative metrics for both 2D and 3D analysis of volumetric SHG images of collagen structure in various tissues. This work was applied to quantitatively categorize collagen fiber crimping, examine collagen fiber damage, growth and where possible, find correlations of the collagen structure with the mechanical properties of the tissue. In an initial study, I quantitatively characterized crimp patterns occurring in ligament tissue using SHG microscopy. I developed a simple algorithm using FT-SHG to quantitatively classify the fiber crimp patterns in ligament into three classifications based on the collagen fiber alignment and crimp direction. The algorithm considered the non-collagenous regions and the collagen fiber alignment in the 2D-SHG images of the 3D stack. This work revealed the 3D structural variation and the underlying helicity in the crimp patterns. In another study, I applied quantitative 3D-SHG imaging of collagen fibers to assess potential damage induced by electron-beam (e-beam) irradiation from environmental-scanning electron microscopy (ESEM). In this case, the 2D metrics employed to analyze the structural alteration of tissues were obtained by measuring the SHG signal intensity and the intensity distribution in the spatial-frequency domain. SHG images obtained before and after ESEM e-beam exposure in low-vacuum mode, disclosed evidence of cross-linking of collagen fibers. Meanwhile, wet-mode ESEM appeared to radically alter the structure from a regular fibrous arrangement to a more random fiber orientation. This work provided insight on both the limitations of the ESEM for tissue imaging, and the potential opportunity to use as a complementary technique when imaging fine features in the non-collagenous tissue samples. Lastly, I expanded the quantitative SHG analysis in three-dimensions to assess the 3D collagen organization in a manner that is consistent with direct observation from images, and applied this analysis on complex 3D collagenous structures. The algorithm that analyzes the 3D fiber organization simulated five classifications of fiber organization that are in natural tissues. The quantitative metrics used were based on 3D fiber orientation and spread to differentiate each classification in a repeatable manner. As a validation process, I applied SHG images of tendon tissue cut in specified orientations and found a strong agreement between the classification algorithm and the physical fiber structure. My 3D fiber analysis approach was further expanded to mouse bile ducts at different stages of growth. It was found that the obtained 3D parameters gradually changed with age, indicating as the fibers grow, the collagenous volume and the amount of fiber crimping increases. In yet another study, I have adapted my 3D-SHG analysis method to analyze localized regions of rat cervix tissue, the results of which have been co-registered with the corresponding measured mechanical properties obtained via nanoindentation. The 3D-SHG parameters obtained were found to be significantly different in distinctive spatial regions of the cervix, and the results from the SHG data were found to be positively correlated with the mechanical properties. This work has the potential to contribute to understanding of collagen remodeling in cervix during pregnancy, and thus this could lead to developing improved methods for preterm birth prediction. Finally, I discussed the future directions of the quantitative 3D-SHG image analysis, and the potential of the analysis method to be applied to other imaging modalities.U of I OnlyAuthor requested U of Illinois access only (OA after 2yrs) in Vireo ETD syste