Quantification of collagen fiber organization in biological tissues at cellular and molecular scales using second-harmonic generation imaging

Collagen is the most abundant structural protein found in the human body, and is responsible for providing structure and function to tissues. Collagen molecules organize naturally into structures called fibers on the scale of the wavelength of light and lack inversion symmetry, thus allowing for the process of second harmonic generation (SHG) when exposed to intense incident light. Consequently, SHG microscopy has been frequently employed in medicine and biology to obtain high-contrast 3D images of collagen fibers without the need for staining. Quantification of collagen fiber organization at the scales of the optical diffraction limit (cellular scales) and sub-diffraction limit (molecular scales) is quintessential for understanding structure and function, assessing the health of the tissue, and identifying traits of a specific disease or damage early. In this regard, we have developed two quantitative techniques: Fourier transform-second-harmonic generation (FT-SHG) imaging and generalized χ2 second-harmonic generation (χ2-SHG) imaging. FT-SHG imaging involves extracting quantitative metrics through the application of spatial Fourier analysis on the images of collagen-based tissues obtained from an SHG microscope (cellular scale). Simple metrics such as preferred orientation, maximum spatial frequency, and fiber spacing are defined and used to quantify differences in collagen fiber organization among several porcine tissues: ear cartilage, trachea, and cornea. Further, we quantitatively compare the information content between SHG images obtained from the forward and backward direction for three tissue types: porcine tendon, sclera, and ear cartilage. We find that both signal types yield consistent information on the preferred orientation of collagen fibers and several overlapping peaks in the magnitude spectrum for these tissues. This strengthens the application of backward SHG imaging for potential clinical diagnostics. In order to show that FT-SHG imaging can be used as a valuable diagnostic tool for real-world biological problems, we first investigate collagenase-induced injury in horse tendons. Clear differences in collagen fiber organization between normal and injured tendon are quantified. In particular, we observe that the regularly oriented organization of collagen fibers in normal tendons is disrupted in injured tendons leading to a more random organization. We also observe that FT-SHG microscopy is more sensitive in assessing tendon injury compared to the conventional polarized light microscopy. The second study includes quantifying collagen fibers in cortical bone using FT-SHG imaging and comparing it with scanning electron microscopy (SEM). Further, as an example study, we show how FT-SHG imaging could be used to quantify changes in bone structure as a function of age. Some initial work and future directions for extending FT-SHG to 3D are also discussed. The second technique, χ2-SHG imaging, takes advantage of the coherent nature of SHG and utilizes polarization to extract the second-order susceptibility (d elements) which provides information on molecular organization, i.e., it provides access to sub-diffractional changes “optically”. We use χ2-SHG in combination with FT-SHG imaging to investigate a couple of biological problems. First, we quantify differences in collagen fiber organization between cornea and sclera of the eye in order to investigate their properties of transparency and opacity, respectively. We find from χ2-SHG imaging that there is no statistical difference in the values of d elements between cornea and sclera, indicating that the underlying collagen structure generating SHG from the two is similar at the level of detection of SHG microscopy. However, the difference lies in the spatial organization of these collagen fibers as observed from FT-SHG imaging. We find that cornea contains lamellae with patches of ordered and uniform diameter collagen fibers with axial order, which could be the reason for its transparent behavior. Conversely, there are no lamellae in sclera (i.e., no axial order), and fibers are thicker, denser, have inconsistent diameters, and possess relatively inhomogeneous orientations, leading to its opaque nature. We also utilized the two techniques to assess differences in stromal collagen fibers for several human breast tissue conditions: normal, hyperplasia, dysplasia, and malignant. Using FT-SHG imaging, we note differences between malignant and other pathological conditions through the metric A.I. ratio. Using generalized χ2-SHG imaging, we observe structural changes in collagen at the molecular scale, and a particular d element showed a more sensitive differentiation between breast tissue conditions, except between hyperplasia and normal/dysplasia. We also find that the trigonal symmetry (3m) is a more appropriate model to describe collagen fibers in malignant tissues as opposed to the conventionally used hexagonal symmetry (C6). Furthermore, the percentage of abnormal collagen fibers could potentially be used as a metric for differentiating breast tissue conditions. We also introduce a technique for extending χ2-SHG to fibers with curvature which is useful for generating χ2-image maps (in terms of d elements) instead of the conventional SHG intensity images. The spatial variations in d elements will provide additional information. For example, in breast cancer tissues, it may help in observing how fibers change from normal to malignant spatially, especially around region of cancerous cells. Finally, we discuss some of the interesting immediate and later future work of quantitative SHG imaging we aim to carry out in our lab