Nanozoomer digital pathology

Biopsies were processed, paraffin embedded, sectioned into 3 μm sections and stained with haematoxylin and eosin. Morphometry was carried out using NanoZoomer Digital Pathology (Hamamatsu Photonics, Hamamatsu, Japan) as previously described (Kelly et al., 2004 (link), Kelly et al., 2016 ) (Supplementary Fig. S1). The following measurements were made: villus height (VH) and width (VW), crypt depth (CD), villus mucosal perimeter (VP) as a measure of epithelial surface area and related to length of muscularis mucosae, and villus cross-sectional area (VA) as a measure of villus volume. VH/CD and villus surface area/volume ratio (SAVR) (Kelly et al., 2016 ) were derived from these. Claudin-4 immunostaining was performed on 3 μm sections (see Supplementary material).

Luxol fast blue (LFB) staining for myelin was used in 14 μm thick sections. The cryostat sections were processed according to the cryo-nerve myelin fast blue staining kit protocol (Genmed Scientifics Inc., USA). We adopted the experimental protocol as reported previously [32 (link)]. Images were captured under a NanoZoomer Digital Pathology (Hamamatsu). Three blinded experimenters calculated the staining density to quantify the myelin using the Image J [33 (link)]. Three animals per group were examined at each time point. Three spinal cord cross sections per animal from the same levels in the injured regions were analyzed.

Approximately 3-4 mm² pieces of tissue were collected at the same sites sampled for gene expression analysis and tissues were immediately embedded in OCT (Optimal cutting temperature compound) and stored in sealed containers at -80°C. Tissue sections (8 μm) were cut for hematoxylin and eosin (H&E) staining, acid fast staining, and immunohistochemical staining at the Plateforme d’histologie et microscopie électronique, Faculté de médecine, Université de Sherbrooke.
For immunohistochemistry, the primary antibodies were polyclonal rabbit anti-human CD3 (clone A0452, 1:400 dilution) (DAKO, Glostrup, Denmark), anti-bovine CD11c mAb (clone BAQ153A, 1:50 dilution)] (Kingfisher Biotech, Saint Paul, MN, USA), mouse anti-bovine CD172a mAb (clone CC149, 1:500 dilution) (Bio-Rad Laboratories, Mississauga, ON, Canada), mouse anti-Ki67 (clone MIB-1, 1:50 dilution) (Agilent Technologies, Mississauga, ON, Canada), and a purified mouse IgG1 isotype control (product code MG100, 1:50 dilution) (Life Technologies, by Thermo Fisher Scientific, Mississauga, ON, Canada). Histological images of the stained tissues were acquired using a slide scanner (NanoZoomer Digital Pathology, Hamamatsu Photonics, Boston, MA, USA) followed by viewing with NDP.view2 viewing software U12388-01 (Hamamatsu Photonics, Boston, MA, USA).

Placental samples were embedded in paraffin. Sections (7 μm) were stained with hematoxylin/eosin using an automat (Varistain, Thermofisher, Waltham, MA, USA) for stereological analysis and scanned using NanoZoomer Digital Pathology^® (Hamamatsu Photonics, Hamamatsu, Japan). Surface densities (Sv) and Volume fractions (Vv) of the different components of the allantochorion, including microcotyledons and allantois, were quantified using One stop stereology with the Mercator^® software (ExploraNova, version 7.9.8, La Rochelle, France [47 (link)]). The measured components included haemotrophic and histiotrophic trophoblasts, microcotyledonary and allantoic vessels, microcotyledonary and allantoic connective tissues, and microcotyledons as the sum of all microcotyledonary components. The chorionic mesoderm and allantoic connective tissue were combined as allantoic connective tissue. Vv and Sv were then multiplied by the total volume of the placenta to estimate the absolute volume (cm³) and surface area (cm²) of the components of the allantochorion, as previously described [45 (link)].
For each placenta, 12 allantoic arterioles were randomly selected. The surface area of the lumen (calculated using transverse diameters) and the thickness of the vascular wall were measured using the ImageJ^® software (version 1.8.0_345, National Institute of Health, Bethesda, MD, USA).

Sections stained for T- and B-cell markers were scanned with a slide scanner (NanoZoomer Digital Pathology, Hamamatsu Photonics) at ×200 magnification. Cells per mm² were counted manually using the Hamamutsu NDP.view program. The selected regions ranged from 7.6 to 111 mm². To quantify multilabelled cells, fluorescent staining was scanned with the Vectra Polaris Automated Quantitative Pathology Imaging system from Perkin Elmer and quantified semi-automatically with Qupath. This software provides a machine learning tool for object classification. As a first step, the algorithm had to be trained. Therefore, the image was split in different channels, one for each marker. Cells were detected by nuclear staining (DAPI). Next, the object classifier was trained and checked manually for accuracy. Subsequently, the object classifier was saved, a region of interest (from 4 to 102 mm²) selected and the algorithm applied. To quantify multilabelled cells, the classifiers were applied at the same time. Finally, to obtain the numbers of cells per mm², the absolute numbers of quantified cells were divided by the analysed areas.

The fixed eyes were rinsed, dehydrated and embedded in paraffin, and 3 µm thickness sections on glass slides were stained with haematoxylin and eosin. Eight cross sections of the retina through the optic disc, taken at 45 µm intervals, were prepared, and three out of these eight sections were randomly selected for histological evaluation. The light microscopic images of the retinas were obtained with a fully automated digital slide scanner (NanoZoomer Digital Pathology™, Hamamatsu Photonics, Sizuoka, Japan). For each image, the IPL thickness was measured and the number of cells in the GCL was determined within an approximately 800 µm expanse of the retina, starting at a distance of 700 µm from the centre of the optic disc. Data from three sections were averaged and used as the representative value for each eye.