Eslide manager

PDX tumor samples were collected and fixed in 10% neutral buffered formalin and processed routinely for histopathology. Tissues were embedded in paraffin and sectioned at 4 µm. Multiplex staining was performed using Opal 4-Color IHC Automation kit plus opal 620 (Akoya Biosciences) on a Leica Bond RXm autostainer. See antibody tables below. After staining, slides were rinsed in DI water and Vector TrueVIEW Auto fluorescence Quenching Kit was used per manufacturer’s instructions. Slides were hand-coverslipped and imaged at 20x using a Leica Versa 8 fluorescent digital scanning microscope system. Digital slides were accessed through Leica eslide manager and opened in Imagescope software. Leica digital image analysis software was utilized for quantification. A cellular immunofluorescence algorithm was tuned for each panel. Quantitative data was exported into excel spreadsheets. Antibodies used for tissue staining are included in Supplemental Table 1B.

Mouse colon tissue “Swiss-rolls” were fixed in 4% paraformaldehyde for 4 h on ice, then processed with stepwise washings to the final step with 70% ethanol. The fixed tissue samples were dehydrated and embedded in paraffin according to standard histological procedures [36 (link)]. The tissue sections were then mounted on glass slides. Subsequently, after dewaxing and rehydration, the slides were stained with hematoxylin and eosin. The histological images were obtained using a Leica Aperio CS2 scanner with the software eSlideManager (Leica Biosystems, Vista, CA, USA). Histomorphological scoring of intestinal inflammation and the lesion was performed according to guidelines described by Erben et al. [37 (link)].

All animals were implanted with an osmotic pump containing BrdU prior to exposure. Hepatic S-phase DNA synthesis was determined using BrdU immunohistochemistry for identification of BrdU incorporation into nuclear DNA using the method outlined by Eldridge and Goldsworthy (1996) (link). Upon euthanasia, a section of liver was recovered as noted above, processed by standard techniques, and mounted on glass slides. Tissue was stained for BrdU using the manufacturers’ protocol (BD Biosciences, San Diego, California) and modified heat-induced antigen retrieval. A small section of duodenum from each animal was also processed to serve as a control for confirming systemic availability of the BrdU as well as confirmation of BrdU immunohistochemistry by visual inspection. BrdU was not scored for the duodenum. All slides were digitized using a Versa whole-slide scanner (Leica Biosystems) and the original digital image files stored on a secure GLP compliant server (e-Slide Manager; Leica Biosystems). A labeling index (BrdU-positive hepatocytes/BrdU-positive + BrdU-negative hepatocytes x 100) was calculated for each of three hepatocellular regions: centrilobular, midzonal, and periportal using a semi-automated image analysis system (Halo; Indica Labs). At least 1,000 cells/region/animal in centrilobular, midzonal, and periportal regions were evaluated.

Immunohistochemically and fluorescently stained sections were scanned at a magnification of ×20 using the Aperio Digital Pathology Scanner (Leica Biosystems Inc., Germany) to capture a digital whole slide image. Digital image analysis for quantification of the staining for S100A9, CD34, CD56, CD20, CD3, CD15 and CD68 was measured using Aperio ScanScope AT Turbo, eSlide Manager and ImageScope software (V12.3.3.7014; Leica Biosystems, Vista, CA, USA) in accordance with the manufacturer’s recommendations. The 3 most representative high-power fields were captured for each tumor and nontumor region in all specimens. All measurements were performed by the same researcher, who was blinded to the histologic and patient survival data, to prevent interoperator variability. Moreover, the protein expression level of S100A9 (cells/mm²) in the immunohistochemical staining of the tissue microarray is shown in Supplementary Table 5.

All immunohistochemistry slides were scanned and analyzed with the digital pathology system, including Aperio ScanScope CS, eSlide manager, and image analysis tools, made by Leica Biosystem (Chicago, Illinois). Quantification of IGSF9 used a macro built from the pixel counting algorithm (version 1), which reports the percentage of total number of positively stained pixels over total number of pixels analyzed. Cellular quantification nuclear stain algorithm (version 9) was used to make a macro to count positive cell percentage of total number of tumor cells analyzed in Ki67 immunostains.