Enhancing spatial and temporal resolution for isotropic volumetric imaging and 3D cell tracking

Optical microscopy is ubiquitous in biological sciences with fluorescence microscopy in particular being utilised to map specific labelled proteins and/or structures. While the majority of such research is performed on populations of cells growing on glass slides, increasingly there is an appreciation that more realistic environments are required to obtain relevant data on biological processes, and ultimately this means using live biological models. A range of small, optically accessible, organisms (e.g. zebrafish, nematode worms, etc) provide convenient live samples for studying biological processes in vivo. Such samples are inherently three-dimensional (3-D), zebrafish being <1 mm in diameter when under 16 days old, and therefore require 3-D discrimination to provide unambiguous positional/structural information. Most 3-D microscopy is undertaken with laser scanning microscopes that scan a spot of light through the sample, building up a map of fluorescence intensity point by point. Such scanning microscopes are typically optimised for higher magnifications (i.e. small fields of view), suffer from unequal resolution (transverse better than axial) and require significant financial investment (e.g. often >£150K). An alternative method of acquiring 3-D data is optical projection tomography (OPT), the optical equivalent to X-ray computed tomography, which can be implemented on a standard wide-field imaging microscope and can provide 3-D imaging at a fraction of the cost of point scanning systems. In OPT, wide-field images (either fluorescence or transmitted light) of a rotating sample are acquired at different orientations. These images can be used to reconstruct the 3-D distribution of fluorescence/absorption.
The standard approach to OPT imposes three key constraints: the requirement that at least the front half of the sample must be 'in focus', that the whole sample must stay in the field of view throughout the acquisition to prevent artefacts in the reconstruction process and that the sample must be non-scattering (i.e. transparent). The first two constraints limit the achievable spatial resolution, since they require the numerical aperture (NA) of imaging system to be small. This limit can be overcome by scanning the imaging lens, and therefore the focal plane, through the rotating sample while acquiring the angularly-resolved images. This produces an 'in focus' image of the whole sample that is superimposed on an out of focus "back-ground" signal that can be removed during image processing. We propose to extend this approach to yet higher resolution imaging of selected sub-volumes within the sample by incorporating a lateral scanning microscope stage to allow the motion of a "volume of interest" (VOI) inside a larger specimen (e.g. an organ) to be maintained in focus as the sample rotates. This VOI can then be modelled as a detailed structure within a larger 'unstructured' volume, to permit high resolution reconstruction without artefacts associated with parts of the sample entering/leaving the field of view. This would permit isotropic high resolution 3-D imaging of, e.g. immune cell distribution in specific organs in live zebrafish, which is currently not possible using the standard commercially available instruments.
To address the limits to temporal resolution, we would investigate a novel "orthogonal scanning approach", acquiring sequential images at right-angles with respect to each other, such that the 3-D structure/location of features within the sample could be determined much faster than the rotation period. This could be applied, e.g. to follow cell migration within a live zebrafish. Finally we will extend this system to simultaneously acquire two images at different wavelengths of light using a commercially available spectral image splitter. By analysing these two wavelength channels we will be able to indirectly probe the signalling events that control the immune response and occur within cells.

We propose to use focal plane scanning OPT, where a high NA imaging lens is scanned axially during the acquisition at each projection angle to acquire high resolution information from all depths of the specimen replicating a parallel projection, to achieve isotropic high resolution volume of interest (VOI) imaging in whole intact samples. Due to the short depth of field (a result of using high NA optics) regions outside the VOI appear significantly blurred and can be accounted for in the reconstruction process using appropriate spatial frequency filtering to suppress artefacts. To realise this system the VOI will have to be tracked during the acquisition as the sample rotates, achieved by appropriate control of lateral and axial motorized translation stages. We will apply this to imaging the spatial distribution of immune cells with respect to the gut epithelium in 6-16 day old zebrafish (available transgenic models) in response to inflammation instigated by a high cholesterol diet (HCD). Moreover we will demonstrate the capability of this OPT system to acquire time-lapse tracking information with a temporal resolution at a fraction of the total acquisition time by employing an orthogonal angle scanning procedure. By acquiring two sequential images at orthogonal angles, the approximate 3-D position of objects (i.e. cells) can be calculated. Therefore over the full acquisition of the OPT data set (i.e. many pairs of orthogonal images) the objects can be tracked. We will use this to extend our investigation to the spatio-temporal dynamics of immune cells in the same HCD zebrafish. Finally we propose to investigate the cellular signalling processes that govern the immune response using a FRET biosensor for the activation of Caspase-1. A spectral image splitter will be incorporated into the system to simultaneously acquire donor and acceptor fluorescence images, which can subsequently be analysed to indicate the relative Caspase-1 activity in this HCD model.

This project aims to develop a novel high resolution 3-D imaging modality aimed at enabling volume of interest (VOI) imaging on a micron scale and rapid 3-D feature tracking within intact live organisms, such as zebrafish embryos, of mm size. The proposed techniques would be widely applicable to objectives across biology, biomedical research and beyond with potential impact in basic biology research, drug discovery, healthcare, food security and materials science.
Besides the potential impact on academic research, the proposed new imaging capabilities would also benefit drug discovery. Pharmaceutical companies would benefit through more precise assays of the action of their compounds within specific organs in live zebrafish up to 16 dpf, with particular impact in studies of responses in whole organisms and phenotypes associated with cell migration (e.g. metastasis) and recruitment (e.g. inflammation) as well as tumour development. In general, zebrafish can serve as a cost effective and genetically tractable disease model in pharmaceutical development, e.g. providing initial in vivo screens for drug efficacy and toxicity before the expensive mammal testing phase - thereby addressing the "3Rs" agenda - as well as having the potential to measure/observe 'off-target' effects. Thus, this project would highlight new opportunities for instrumentation manufacturers and software developers in the UK and elsewhere - particularly in drug discovery sector but also for general microscopy and imaging applications.
The proposed system could be provided as an OPT 'add-on' to existing commercial microscopes and would provide a more cost effective route to 3-D imaging for general biological/biomedical research compared to advanced laser scanning confocal/multiphoton microscope systems. For drug discovery, zebrafish-based assays would require the development of automated animal handling and imaging/analysis systems and there would be considerable potential to extend the readouts, e.g. to spectroscopic measurements of relevant fluorescent labels. The potential impact on healthcare and society follows from advances in the understanding of fundamental biological processes, disease mechanisms and in more effective discovery and testing of new therapies. A further application would be non-destructive 3-D imaging of intact samples (e.g. histopathology) since high resolution OPT would enable whole complex samples to be studied without sectioning, which could help preserve fragile pathologies. This could have clinical benefit as well as impact on preclinical studies. A further key impact on healthcare could be in regenerative medicine where the development of techniques to promote and control cell and tissue growth for medical therapies could be imaged using high resolution OPT - potentially in a nondestructive manner to enable the development of cell/tissue growth to be monitored. The 3-D imaging capabilities could also be applied to a range of other extended samples including for the characterisation of biomaterials and also for materials science outside biomedicine, e.g. to study or evaluate polymer structures or the outcomes of 3-D (lithographic) "printers". 3-D imaging of intact (live) samples is also important for plant biology and OPT is finding increased application to study crop development and the action of diseases, chemicals and pests that impact food security. This field would particularly benefit from higher resolution 3-D imaging capabilities.
The results of this project would be disseminated through open access publications and presentations at such as BiOS Photonics West and CHI-HCA in the USA and ECBO/Laser Munich in Europe. Imperial has strong links with pharma and these would be exploited to reach companies like AstraZeneca, GSK and GE Healthcare, with whom the Photonics Group has already collaborated extensively. We would also explore opportunities for plant imaging with local plant biologists and companies like Syngenta.