Final Report Summary - SPIMTOP (Combined Selective Plane Illumination Microscopy and Optical Projection Tomography for in vivo quantitative imaging of the developing zebrafish vasculature.)
Final Publishable Summary Report (see also Periodic Report):
Our understanding of embryonic development relies fundamentally on the observation of morphogenesis in living biological organisms. Fluorescence light-sheet microscopy such as Selective Plane Illumination Microscopy (SPIM) has proven to be a powerful tool to image developmental processes in vivo with fast high-resolution optical sectioning over large tissue volumes. However, an intrinsic limitation of all fluorescence microscopy techniques, including SPIM, is their inability to image any structures beyond the labeled tissue. The fluorescence structures therefore appear out of context to the sample anatomy and, e.g. growth or migration of fluorescent tissue through the surrounding tissue may therefore be hard to interpret. It is desirable therefore to find a technique that delivers structural information of the unstained tissue to complement the fluorescence data, such as Optical Projection Tomography (OPT).
The objectives of the project were to develop a 3D optical imaging platform that combines SPIM and optical tomography. Although these two techniques are similar from the technical point of view (they can share several components, including the sample holder and the detection system), a multimodality system that fully takes advantage of the features of SPIM and OPT and combines the techniques to gain the maximum amount of information has not yet fully been exploited. The second objective was to develop a data analysis methodology able to provide a detailed dynamic map of the vascular system of zebrafish. The third objective was to validate and apply the new technology to quantitatively study the developing zebrafish embryo, looking at sample morphology and gene expression over a large volume of the organism.
We demonstrated that, without modifying any component of a typical SPIM setup, we can perform a multimodal acquisition that integrates the high-resolution SPIM fluorescence data with optical tomography, purely based on bright-field contrast. We obtained in vivo data of zebrafish that demonstrate that optical tomography is a valuable tool to observe the whole anatomy of translucent biological organisms, complementing and improving the analysis of SPIM data. We have demonstrated that our multimodal system offers new possibilities to precisely observe zebrafish embryogenesis over several hours or days.
OPT had been developed to exploit the bright-field contrast of the sample by acquiring several light transmission images (or projections) from different directions; the three-dimensional structure of the sample is then reconstructed using a back-projection algorithm. Unfortunately, a limitation when combining OPT and SPIM is that OPT requires a long depth of field, ideally spanning the entire depth of the specimen, whereas SPIM benefits from a shallow depth of field. To overcome this incompatibility hybrid OPT-SPIM instruments have recently been developed by reducing the numerical aperture (NA) of the detection unit to achieve the long depth of field needed for OPT. As a consequence the overall resolution and contrast in the OPT reconstructions are compromised. Moreover, SPIM is considered an ideal technique to image development in vivo, whereas tomographic imaging is commonly believed suitable only for fixed and chemically cleared samples and, indeed, only a few in vivo applications have been reported. We decided to look for alternative solutions that would achieve tomographic reconstructions of living specimens in a high-resolution, high-NA SPIM setup without any modifications to the hardware.
To integrate an optical tomography approach in a SPIM microscope we took advantage of three features of our existing SPIM setup: (i) fast image acquisition with high frame rate sCMOS cameras, (ii) multi-view capability, i.e. the sample can be quickly rotated, and (iii) LED for back illumination, which can provide transmission images of the specimen. Due to the high NA of the detection lens, the depth of field does not span the depth of a typical sample in SPIM of ca. 0.1-1mm. Therefore, a stack of transmission images was taken by sliding the sample through the detection objective's plane of focus. The in-focus information was extracted by high-pass filtering the images and a weighted average yielded a projection of the sample with enhanced depth of field. Since a projection represents an approximation of the line integral of light attenuation along a certain direction, by collecting multiple projections from several directions we created a dataset suitable for tomographic reconstruction: optically sectioned volumes of the samples were obtained by a filtered back-projection algorithm. No modification to the SPIM hardware and no calibration were needed for the tomographic reconstruction. This method can therefore be readily adopted by a large number of systems, including commercial light sheet microscopes. We developed a real time routine able to record only the relevant data for tomographic reconstruction with a short acquisition time that integrate well with the fast SPIM recording and make the measurement compatible with in vivo imaging and time-lapse observation of development. We found that the tomographic reconstruction is useful on its own, to observe zebrafish anatomy in vivo, to annotate, segment and quantitatively measure the volume of several organs and that the multimodal system provides a comprehensive visualization of the specimen that makes it well suited for in vivo localization analysis of gene expression in the entire zebrafish. We concluded that the multimodal acquisition offers the possibility to localize, track and register specific anatomical regions of the sample, which is not readily available in fluorescence modalities.
We believe that the developed technology represents a valuable tool for quantitative imaging of development, potentially useful for the study of several translucent organism, with a strong impact for the growing field of developmental biology and for the microscopy community, which represents one of the recognized as a strategic sector for the competitive edge of EU.
Our understanding of embryonic development relies fundamentally on the observation of morphogenesis in living biological organisms. Fluorescence light-sheet microscopy such as Selective Plane Illumination Microscopy (SPIM) has proven to be a powerful tool to image developmental processes in vivo with fast high-resolution optical sectioning over large tissue volumes. However, an intrinsic limitation of all fluorescence microscopy techniques, including SPIM, is their inability to image any structures beyond the labeled tissue. The fluorescence structures therefore appear out of context to the sample anatomy and, e.g. growth or migration of fluorescent tissue through the surrounding tissue may therefore be hard to interpret. It is desirable therefore to find a technique that delivers structural information of the unstained tissue to complement the fluorescence data, such as Optical Projection Tomography (OPT).
The objectives of the project were to develop a 3D optical imaging platform that combines SPIM and optical tomography. Although these two techniques are similar from the technical point of view (they can share several components, including the sample holder and the detection system), a multimodality system that fully takes advantage of the features of SPIM and OPT and combines the techniques to gain the maximum amount of information has not yet fully been exploited. The second objective was to develop a data analysis methodology able to provide a detailed dynamic map of the vascular system of zebrafish. The third objective was to validate and apply the new technology to quantitatively study the developing zebrafish embryo, looking at sample morphology and gene expression over a large volume of the organism.
We demonstrated that, without modifying any component of a typical SPIM setup, we can perform a multimodal acquisition that integrates the high-resolution SPIM fluorescence data with optical tomography, purely based on bright-field contrast. We obtained in vivo data of zebrafish that demonstrate that optical tomography is a valuable tool to observe the whole anatomy of translucent biological organisms, complementing and improving the analysis of SPIM data. We have demonstrated that our multimodal system offers new possibilities to precisely observe zebrafish embryogenesis over several hours or days.
OPT had been developed to exploit the bright-field contrast of the sample by acquiring several light transmission images (or projections) from different directions; the three-dimensional structure of the sample is then reconstructed using a back-projection algorithm. Unfortunately, a limitation when combining OPT and SPIM is that OPT requires a long depth of field, ideally spanning the entire depth of the specimen, whereas SPIM benefits from a shallow depth of field. To overcome this incompatibility hybrid OPT-SPIM instruments have recently been developed by reducing the numerical aperture (NA) of the detection unit to achieve the long depth of field needed for OPT. As a consequence the overall resolution and contrast in the OPT reconstructions are compromised. Moreover, SPIM is considered an ideal technique to image development in vivo, whereas tomographic imaging is commonly believed suitable only for fixed and chemically cleared samples and, indeed, only a few in vivo applications have been reported. We decided to look for alternative solutions that would achieve tomographic reconstructions of living specimens in a high-resolution, high-NA SPIM setup without any modifications to the hardware.
To integrate an optical tomography approach in a SPIM microscope we took advantage of three features of our existing SPIM setup: (i) fast image acquisition with high frame rate sCMOS cameras, (ii) multi-view capability, i.e. the sample can be quickly rotated, and (iii) LED for back illumination, which can provide transmission images of the specimen. Due to the high NA of the detection lens, the depth of field does not span the depth of a typical sample in SPIM of ca. 0.1-1mm. Therefore, a stack of transmission images was taken by sliding the sample through the detection objective's plane of focus. The in-focus information was extracted by high-pass filtering the images and a weighted average yielded a projection of the sample with enhanced depth of field. Since a projection represents an approximation of the line integral of light attenuation along a certain direction, by collecting multiple projections from several directions we created a dataset suitable for tomographic reconstruction: optically sectioned volumes of the samples were obtained by a filtered back-projection algorithm. No modification to the SPIM hardware and no calibration were needed for the tomographic reconstruction. This method can therefore be readily adopted by a large number of systems, including commercial light sheet microscopes. We developed a real time routine able to record only the relevant data for tomographic reconstruction with a short acquisition time that integrate well with the fast SPIM recording and make the measurement compatible with in vivo imaging and time-lapse observation of development. We found that the tomographic reconstruction is useful on its own, to observe zebrafish anatomy in vivo, to annotate, segment and quantitatively measure the volume of several organs and that the multimodal system provides a comprehensive visualization of the specimen that makes it well suited for in vivo localization analysis of gene expression in the entire zebrafish. We concluded that the multimodal acquisition offers the possibility to localize, track and register specific anatomical regions of the sample, which is not readily available in fluorescence modalities.
We believe that the developed technology represents a valuable tool for quantitative imaging of development, potentially useful for the study of several translucent organism, with a strong impact for the growing field of developmental biology and for the microscopy community, which represents one of the recognized as a strategic sector for the competitive edge of EU.