Optical coherence tomography for REtinal Organoid imaging – OREO

Setup: A new multimodal (fluorescence, static and dynamic FFOCT) setup will be developed for use in a clean room environment at the Vision Institute. An efficient sectioning of the fluorescence channel with multiwavelength excitation and detection will be developed, along with real time dynamics acquisition capability via rapid computing. We will engineer an enclosing box to control the environment and operate our microscope in conditions suitable for organoid development. We plan to design a procedure to allow automated motorized image stack acquisition in standard format transwell culture plates.
Quantification: We aim to display the dynamic FFOCT signal with a truly quantitative scale, and automate cell classification. While we know that movements are occurring inside cell cytoplasm and nuclei, we are not yet fully aware of the precise biological processes involved in generating the dynamic signal: for use as a biomarker of cell activity we need to understand the intracellular mechanisms that create dynamic signal contrast. Validation against ground-truth data will identify the FFOCT signatures for specific cell types or behaviors. Quantitative display in a hue/saturation/value color scale will be refined accordingly to facilitate interpretation of frequency, bandwidth and amplitude of activity. For non-invasive differentiation of the identified specific cell types and behaviors, we will develop and implement signal processing and machine learning algorithms.
Application: A 3D retinal organoid model of RP (testing both Class I mutated RHO P347L and Class II mutated RHO P23H) and its isogenic control will be developed and followed with the installed FFOCT setup. The degenerative profile will be analyzed and characterized, and therapeutic solutions will be tested on these organoid-based models with the hope of detecting a recuperation of cell function following therapy.

Validation du signal D-FFOCT : nous sommes en train de valider quelles organelles sont responsables du signal DFFOCT sur des cultures de cellules épithéliales rétiniennes en comparant les images DFFOCT et l’immunohistochimie sous condition de stress et sans stress, ainsi que sur les organoïdes rétiniennes en cours de développement. Une échelle de couleur quantitative a été développée pour pouvoir comparer les échantillons entre eux.
Microscope pour la salle de culture : le microscope DFFOCT destiné à la salle de culture est construit, complet et en phase de test à l’Institut Langevin. Nous avons amélioré le design initial pour optimiser son utilisation par le plus grand nombre dans une salle de culture, notamment en l’intégrant dans une base de microscope commercial (Olympus). Le montage sera déplacé à l’Institut de la Vision en été 2021 dans une nouvelle salle de culture dédiée.
Modèle de rétinite pigmentaire: Pour la validation du signal des organoïdes rétiniens (OR) ont été générés à partir de cellules souches induites à la pluripotence (iPS) humaines saines clone 5f (iPS-5f). Les âges des OR actuellement produites en flux tendu varient entre 28 jours (J28) et J42. Néanmoins, des OR plus vieux sont également générées (J150-200). L’analyse biochimique du profile dégénératif des OR mutés devrait débuter cet été (2021). L’objectif de démontrer de l’évidence d’action thérapeutique va démarrer en 2023.
Classification de cellules automatique : nous avons développé des algorithmes pour la classification automatique des cellules par deux méthodes différentes (par Feature engineering et par Convolutional neural networks). Ces algorithmes ont donné des résultats très prometteurs avec des précisions supérieures à 90%. Nous pourrons ensuite modifier ces algorithmes pour l’analyse automatique des organoïdes et la classification des cellules qui les composent.

OREO’s device and disease model development aims to move towards new therapeutic development, which will have a health impact, and therefore a societal and cultural impact, by helping to reduce blindness.

Indeed, prevalence of retinal degeneration including RP is reported to be 1/3,000 to 1/5,000, which represents high economic cost in terms of care. Improved disease understanding and therapeutic development could have a strong impact on improving care and even reducing prevalence. Again touching on the economic impact, improving efficacy of drug screening will lead to more cost effective drug development.

In the wider context, the ability to see biological processes in real time will have scientific impact by furthering fundamental understanding of subcellular dynamics in health and disease.

An eventual extension of the technology to in vivo imaging implies a low cost clinical imaging of therapy in situ, allowing cost effective clinical trials with short term outcome measures, i.e. monitoring of cellular changes in vivo.

- Couturier A, Blot G, Vignaud L, Nanteau C, Slembrouck-Brec A, Fradot V, Acar N, Sahel JA, Tadayoni R, Thuret G, Sennlaub F, Roger JE, Goureau O, Guillonneau X, Reichman S. Reproducing diabetic retinopathy features using newly developed human induced-pluripotent stem cell-derived retinal Müller glial cells. Glia. 2021. PMID: 33683746.

- Slembrouk-Brec, Rodrigues A, Rabesandratana O, Gagliardi G, Nanteau C, Fouquet S, Thuret G, Reichman S, Orieux G, Goureau O. Reprogramming of Adult Retinal Müller Glial Cells into Human-Induced Pluripotent Stem Cells as an Efficient Source of Retinal Cells. Stem Cells Int. 2019. PMID: 31396286.

- Scholler, J., Groux, K., Goureau, O, Sahel JA, Fink M, Reichman S, Boccara A, Grieve K. Dynamic full-field optical coherence tomography: 3D live-imaging of retinal organoids. Light Sci Appl 9, 140 (2020). doi.org/10.1038/s41377-020-00375-8

- Scholler, J., Grieve, K., Thouvenin O., et al., Automatic diagnosis and biopsy classification with dynamic Full-Field OCT and machine learning, submitted to Nat. Medicine in march 2021

Induced Pluripotent Stem Cells (iPS) can generate 3D retinal organoids that contain all major types of retinal cells and a distinct stratification close to the in vivo morphology. These organoids are therefore very promising for disease modeling purposes, as well as possible cell therapy strategies in patients with retinal degeneration. In particular, the use of organoids is envisaged to facilitate understanding of the cellular mechanisms leading to retinitis pigmentosa (RP), a group of inherited neurodegenerative diseases that cause selective cell death of photoreceptors. The mechanisms leading to cell death remain unknown and no adequate treatment for RP is currently available. Retinal degeneration, and in particular RP, can be targeted via “disease in a dish” modeling methods. The idea is to create a degenerative process that resembles RP, with progressive apoptosis of photoreceptors. Although 2D RP models have been proposed, an organoid-based 3D RP model would be closer to reality. Drugs can then be tested on this RP model to monitor their effectiveness and identify potential drug candidates for clinical trials.

High-resolution imaging tools are essential for performing non-destructive quality control on these 3D cell cultures. At present, the development of organoids is ensured by cultivating a series of samples, a number of which are sacrificed at each control stage. The imaging process requires fixing, slicing and performing multiple fluorescent labeling procedures on organoids to determine the many cell types present, and the correct development of the 3D culture. If non-destructive live imaging were available for these samples, a single organoid could be monitored over time, allowing direct observation of the development and behavior of each sample. However, real-time imaging techniques for use in regenerative medicine are currently limited to 2D cultures.
Full Field Optical Coherence Tomography (FFOCT) is a non-invasive, high-resolution approach to Optical Coherence Tomography (OCT). FFOCT, developed by members of the project group, allows non-invasive imaging of ocular tissues with three-dimensional micrometric resolution. We recently added a complementary contrast mechanism to FFOCT to access functional information. Dynamic FFOCT detects intracellular movement to display a contrast based on cellular metabolism, an indicator of cell viability. In addition, our technique can combine fluorescence and FFOCT channels in superposition to allow the localization of fluorescent cells in their 3D microenvironment.

We hypothesize that we will be able to i) create a 3D organoid-based model of RP, and ii) detect photoreceptor degeneration and recovery in this model, non-invasively and label-free, with FFOCT, thus reproducing the quantitative information generally available only by immunofluorescence.

The multidisciplinary OREO project will push the limits of current practice on three fronts - technological, computational and biological - by developing a multimodal FFOCT imaging tool for use in a clean room environment; developing computational tools for the quantification and automatic classification of cells; and developing a new 3D model of RP and its isogenic control that will be tracked over time with FFOCT.