We have developed SHG microscope tools to probe all levels of collagen architecture organization in human high grade serous ovarian cancer (HSOC). We have found pronounced differences using machine learning classification of the fiber morphology as well as alterations in macro/supramolecular structural aspects through polarization analysis. We have used multiphoton excited fabrication to create SHG image-based orthogonal models that represent both the collagen morphology and stiffness of normal ovarian stroma and HGSOC. We found the fiber morphology of HGSOC promotes motility through a contact guidance mechanism and that stiffer matrix further promotes these same processes through a mechanosensitive mechanism. We have also developed a machine learning approach using generative adversarial networks (GANs) to optimize the scaffold design. Collectively, this data provides insight into disease etiology and suggests future diagnostic approaches.
Collagen remodeling occurs in many prostate pathologies; however, the underlying structural architecture in both normal and diseased prostatic tissues is largely unexplored. Here, we use second-harmonic generation (SHG) microscopy to specifically probe the role of the proteoglycan decorin (Dcn) on collagen assembly in a wild type (wt) and Dcn null mouse (Dcn − / − ). Dcn is required for proper organization of collagen fibrils as it regulates size by forming an arch-like structure at the end of the fibril. We have utilized SHG metrics based on emission directionality (forward–backward ratio) and relative conversion efficiency, which are both related to the SHG coherence length, and found more disordered fibril organization in the Dcn − / − . We have also used image analysis readouts based on entropy, multifractal dimension, and wavelet transforms to compare the collagen fibril/fiber architecture in the two models, where all these showed that the Dcn − / − prostate comprised smaller and more disorganized collagen structures. All these SHG metrics are consistent with decreased SHG phase matching in the Dcn − / − and are further consistent with ultrastructural analysis of collagen in this model in other tissues, which show a more random distribution of fibril sizes and their packing into fibers. As Dcn is a known tumor suppressor, this work forms the basis for future studies of collagen remodeling in both malignant and benign prostate disease.
Both ovarian cancer and idiopathic pulmonary fibrosis (IPF) are accompanied by significant collagen remodeling in the respective extracellular matrix (ECM). These diseases have similar attributes and collagen alterations can be probed with the same methods. Remodeling can be reflected in increased collagen concentration, changes in alignment within fibrils and fibers and/or up-regulation of different collagen isoforms. We used pixel-based SHG polarization analyses to discriminate the macro/supramolecular collagen structure in human tissues by: i) determination of the helical pitch angle via the single axis molecular model, ii) dipole alignment within fibrils via anisotropy, and iii) chirality via SHG circular dichroism (SHG-CD). For ovarian cancer, the largest differences were between normal stroma and benign tumors, consistent with gene expression showing Col III is up-regulated in the latter. The tissues also displayed differing SHG anisotropies and SHG-CD responses, consistent with randomization of Col I alignment in fibrils in all tumors. These results collectively indicate the fibril assemblies are distinct in all ovarian tissues and likely result from synthesis of new collagen rather than remodeling of existing collagen. For IPF, the largest change was in the SHG-CD response, indicating the fibrotic collagen has different helical structure than that of normal tissues. Interestingly, for both diseases, no increase in Col IIII was found, in contrast to previous reports by immunostaining. We suggest these polarization-based metrics could form the basis of a new classification scheme and complement conventional classification based on genetic profiles and conventional histology for these diseases as well as other cancers and fibroses.
Remodeling of the extracellular matrix in human ovarian cancer can be manifested in increased collagen concentration, changes in alignment within fibrils/fibers and/or up-regulation of different collagen isoforms. We used pixel-based second harmonic generation (SHG) polarization microscopy analyses to probe these molecular changes in human ovarian tissues [normal stroma, benign tumors, and high-grade serous (HGS) tumors] by: (i) determination of the α-helical pitch angle via the single-axis molecular model, (ii) collagen alignment within fibrils via SHG anisotropy, and (iii) chirality via SHG circular dichroism (SHG-CD). Pixel approaches are required due to the complex structure of the matrix that lacks a high degree of fiber alignment. The largest differences in the helical pitch angle were between normal stroma and benign tumors, consistent with gene expression showing the Col III isoform is up-regulated in the latter. The data were not consistent with up-regulation of Col III in HGS tumors as previous reports have suggested. The different tissues also displayed differing SHG anisotropies and SHG-CD responses, consistent with either Col III incorporation or randomization of Col I alignment within benign and malignant tumors. Additionally, the high-grade tumors displayed higher collagen concentration, where this desmoplasia is consistent with the higher fiber density in these tissues. These results collectively indicate that the fibril assemblies are distinct in all tissues, where these differences likely result from the synthesis of collagen rather than remodeling of existing collagen. Importantly, these analyses are label-free and interrogate subresolution collagen structure on intact tissues, without the need for conventional structural biology tools.
Remodeling of the collagen architecture in the extracellular matrix (ECM) has been implicated in ovarian cancer. To quantify these alterations, we implemented a form of 3D texture analysis method based on textons to delineate the fibrillar morphology observed in 3D Second Harmonic Generation (SHG) microscopy image data of normal (1) and high risk (2) ovarian stroma, (3) benign ovarian tumors, low grade (4) and high grade (5) serous tumors, and endometrioid tumors (6). We developed a tailored set of 3D filters which extract textural features in the 3D image sets to build (or learn) statistical models of each tissue class. By applying k-nearest neighbor classification using these learned models, we achieved 83-91% accuracies for the six classes. The 3D method outperformed the analogous 2D classification by 10-15% on the same tissues, where we suggest this is due the increased information content available in 3D voxels. This classification based on ECM structural changes will complement conventional classification based on genetic profiles and can serve as an additional biomarker. Moreover, the texture analysis algorithm is quite general, as it does not rely on single morphological metrics such as fiber alignment, length, and width but their combined convolution with a customizable basis. We further discuss a new approach to achieve complete 3D SHG imaging, that is based on a rotating multiview platform. We show this visualizes axially oriented features missing in conventional en face imaging. The data sets are compatible with the texture analysis here and will further improve upon this approach.
The collagen architecture in all human ovarian cancers is substantially remodeled, where these alterations are manifested in different fiber widths, fiber patterns, and fibril size and packing. Second harmonic generation (SHG) microscopy has differentiated normal tissues from high-grade serous (HGS) tumors with high accuracy; however, the classification between low-grade serous, endometrioid, and benign tumors was less successful. We postulate this is due to known higher genetic variation in these tissues relative to HGS tumors, which are genetically similar, and this results in more heterogeneous collagen remodeling in the respective matrix. Here, we examine fiber widths and SHG emission intensity and directionality locally within images (e.g., 10×10 microns) and show that normal tissues and HGS tumors are more uniform in fiber properties as well as in fibril size and packing than the other tissues. Moreover, these distributions are in good agreement with phase matching considerations relating SHG emission directionality and intensity. The findings show that in addition to average collagen assembly properties the intrinsic heterogeneity must also be considered as another aspect of characterization. These local analyses showed differences not shown in pure intensity-based image analyses and may provide further insight into disease etiology of the different tumor subtypes.
Remodeling of the extracellular matrix in human ovarian cancer, can be reflected in increased collagen concentration, changes in alignment and/or up-regulation of different collagen isoforms, including Col III. Using fibrillar gel models, we demonstrate that Col I and Col III can be quantitatively distinguished by 3 distinct SHG polarization specific metrics: i) determination of helical pitch angle via the single axis molecular model, ii) dipole alignment via anisotropy, and iii) chirality via SHG circular dichroism (SHG-CD). These sub-resolution differentiations are possible due to differences in the α helix angles of the two isoforms, which co-mingle in the same fibrils. We also investigated the mechanism of the SHG-CD response and show that unlike conventional CD, it is dominated by electric dipole interactions and is consistent with the two state SHG model. We further applied these 3 polarization resolved analyses to human normal, high risk, benign tumors, and malignant human ovarian tissues. We found that these tissues could all be differentiated by these metrics, where high grade tissues had analogous α-helical pitch angles to the in the Col I/Col III gel model. This confirms literature suggestions based on immunofluorescence and gene expression that Col III is up-regulated in high grade ovarian cancers. The different tissues also displayed differing anisotropies, indicating the fibril assemblies are distinct and likely do not result from remodeling of existing collagen but synthesis of new collagen. Importantly, these SHG polarization methods provide structural information not otherwise possible and can serve as label-free biomarkers for ovarian and other cancers.
Ovarian cancer remains the most deadly gynecological cancer with a poor aggregate survival rate. To improve upon this situation, we utilized collagen-specific Second Harmonic Generation (SHG) imaging microscopy and optical scattering measurements to probe structural differences in the extracellular matrix of normal stroma, benign tumors, endometrioid tumors, and low and high-grade serous (LGS and HGS) tumors. The SHG signatures of the emission directionality and conversion efficiency as well as the optical scattering are related to the organization of collagen on the sub-micron size. The wavelength dependence of these readouts adds additional characterization of the size and distribution of collagen fibrils/fibers relative to the interrogating wavelengths. We found strong wavelength dependent dependencies of these metrics that were different between the different tumors that are related to respective structural attributes in the collagen organization. These sub-resolution determinations are consistent with the dualistic classification of type I and II serous tumors. However, type I endometrioid tumors have strongly differing ECM architecture than the serous malignancies. Moreover, our analyses are further consistent with LGS and benign tumors having similar etiology. We identified optimal wavelengths for the SHG metrics as well as optical scattering measurements. The SHG metrics and optical scattering measurements were then used to form a linear discriminant model to classify the tissues, and we obtained high accuracy (~90%) between the tissue types. This delineation is superior to current clinical performance and has potential applicability in supplementing histological analysis, understanding the etiology, as well as development of an in vivo screening tool.
Here, we examine ovarian cancer extracellular matrix (ECM) modification by measuring the wavelength dependence of optical scattering measurements and quantitative second-harmonic generation (SHG) imaging metrics in the range of 800-1100 nm in order to determine fibrillary changes in ex vivo normal ovary, type I, and type II ovarian cancer. Mass fractals of the collagen fiber structure is analyzed based on a power law correlation function using spectral dependence measurements of the reduced scattering coefficient μs′ where the mass fractal dimension is related to the power. Values of μs′ are measured using independent methods of determining the values of μs and g by on-axis attenuation measurements using the Beer-Lambert Law and by fitting the angular distribution of scattering to the Henyey-Greenstein phase function, respectively. Quantitativespectral SHG imaging on the same tissues determines FSHG/BSHG creation ratios related to size and harmonophore distributions. Both techniques probe fibril packing order, but the optical scattering probes structures of sizes from about 50-2000 nm where SHG imaging – although only able to resolve individual fibers – builds contrast from the assembly of fibrils. Our findings suggest that type I ovarian tumor structure has the most ordered collagen fibers followed by normal ovary then type II tumors showing the least order.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.