Speaker
Description
Rational:
Modern research in Life Sciences is integrating diverse data from complex models to understand, at the molecular level, the mechanisms underlying the development, function and dysfunction of living organisms. For this, imaging technologies are playing a crucial role. Here, the aim is to record the living state (functional and dynamic) at the highest resolution possible (ultrastructural level). To achieve this, various imaging technologies can be correlated to study the same specimen. Correlative Light and Electron Microscopy (CLEM) is now an imaging field per se that covers a broad spectrum of applications on multiple biological domains and models. On animal models, correlating intravital imaging (by LM) to subcellular recordings (by EM) requires acute targeting precision in 3D with sufficient speed and robustness. In doing so, the enhanced experimental/imaging throughput is expected to enable quantitative analysis of biological phenomena.
We have recently demonstrated the power of multimodal correlative microscopy that greatly facilitates the correlation of intravital imaging to large volume electron microscopy of single metastatic cells (1). Reducing the CLEM workflow from 3-5 months to ~14 days, it is now feasible to study the cellular mechanisms, e.g. of cancer metastasis, in relevant tissue models in vivo. Working closely with numerous laboratories, we are now establishing workflows for various applications in the fields of cell biology, development biology, neurobiology and physiology in normal or pathological conditions.
We aim to enable routine and high-throughput CLEM studies, allowing targeting rare and transient phenomena in vivo. Although our recently developed approach has already demonstrated unprecedented throughput, our aim is to make it faster and, importantly, to make the workflow more accessible to non-specialized laboratories. In doing so, we will answer to the growing needs from the scientific community to link structure to function in biologically relevant multicellular models. Achieving this goal will only be possible with key technological development to offer new instrumentation and automation in the areas of image processing (software), of sample preparation (control and mechanics), of the large volume imaging (microscopic x-ray computed tomography – ‘microCT’) and of the high resolution electron microscopy step itself.
Project:
Our correlation strategy (1) relies i) on 3D image acquisition and registration, ii) on precise shaping of the samples to reveal the region of interest (ROI), and iii) on intelligent acquisition of the ROI volume by EM. We propose to develop automated solutions that are generally applicable to a large range of biological samples and dramatically improve throughput of the experiment.
i) The multimodal workflow relies on mapping the EM-processed sample in 3D using microCT, and correlate this volume to the in vivo dataset in 3D. Following EM preparation, the ROI observed in vivo is no longer visible with LM and its location within the sample is therefore difficult to retrieve. MicroCT enables the visualization of structural landmarks within the EM-processed sample that can be correlated to landmarks visible in the in vivo dataset (LM). 3D registration of these features then enables to predict the position of the ROI inside the EM-processed sample. Currently, the correlation between LM and microCT is performed manually, which is a tedious and time-consuming (2-3 days) task within the workflow. We will develop a software solution to automate the registration task. In parallel, we intend to innovate in the imaging of the EM sample by developing compact, low-cost, accessible instrumentation for this step, which should allow easier implementation in any research laboratory. ii) When the precise mapping of the EM sample is achieved, it is trimmed in order to expose the volume of interest for the 3D EM acquisition. This process demands highly-accurate instruments (ultra-microtomes), and a skilled operator. Since the current ultramicrotomes are not designed to achieve the required accuracy, careful approach to the ROI is needed, which unnecessarily prolongs the workflow with ~1-2 days. By developing a motorized solution, we intend to fully automate and further facilitate this process. Here, based on the 3D maps generated in i), the unwanted parts of the sample will be automatically removed and the progress to the ROI will be monitored by an automated feed-back loop. iii) Large volume imaging by EM is now made possible by automated serial imaging technologies based on scanning EMs. To overcome the current slow acquisition time, we intend to further improve the SEM scanning procedures, detectors and acquisition schemes. For our correlative workflows, we will also develop a software solution to fully automate imaging of the correlated sub-volume of interest. Such “intelligent” acquisition will be controlled using the 3D maps built from the other imaging modalities registered as described in i).
- Karreman, M. A. et al. Fast and precise targeting of single tumor cells in vivo by multimodal correlative microscopy. Journal of Cell Science 129, 444–456 (2016).
