Abstract

Morphometric imaging of organoids for toxicology and assessment
of therapeutic efficacy.

James Noble, Craig Russell, Nilofar Faruqui, Paula
Gomez, Elisabeth Fraser², Mark Treherne², Trevor²
Dale, Mike Shaw¹

¹Biometrology Group, National Physical
Laboratory, Hampton Road, Teddington, TW11 0LW, UK

²Cellesce Ltd, Cardiff Medicentre, Heath Park,
Cardiff, CF14 4UJ, UK

The use of patient derived
organoids in therapeutic assessment has shown a strong correlation with patient
treatment data [1]. Organoids, organ on a chip and other multicellular models
are also being assessed for a range of different toxicology and drug
development applications. Traditional in vitro cellular end-point, or biochemical
assays may not be applicable or representative of multicellular and
heterogeneous structures therefore additional methods to characterise these
models are required. Morphometric imaging describes the capture and quantitative
analysis of structural image data and can be applied to measure the properties
and responses of these model systems. This presentation describes a
collaboration between NPL and Cellesce in which Light Sheet Fluorescence Microscopy
[2] was used to capture images of colorectal tumour organoids treated with a
range of therapeutic compounds. Light sheet techniques are particularly well
suited to visualising 3D cell cultures, spheroids, organoids and model
organisms due to their capacity for fast, minimally invasive, volumetric
imaging of large samples. Bespoke computational image analysis pipelines were
developed to extract  shape and texture metrics
at the organoid and subcellular levels. Analysis of the resulting shape data
was used to derive dose-response relationships and supervised machine learning
was applied to classify the overall organoid response. Similar morphometric imaging
techniques have broad application in the study and analysis of other 3D model
system and are currently being applied to assess the respiratory toxicity used
3D lung models.

[1] Vlacchogiannis
G., et al., Science, 359, 920-926
(2018)

[2] Power,
R. M., et al. Nat Methods, 14, 360-373 (2017)