Anti cd163 clone 10d6

A bulk GE dataset (GSE132348) of 103 cHL patients from Gene-Expression Omnibus (GEO) was used to validate the results^{10 (link)}. It was obtained using the same panel and methods as for the discovery dataset, without the 30 extra custom genes. After two cycles of ABVD, patients were classified following interim FDG-PET (iPET) response, where positive is equivalent to U (4 or 5 on the Deauville-five-point scale), and negative is equivalent to F. Early response to ABVD assessed with iPET is prognostic for cHL patients and has been used to identify GE profiles related to outcome^{10 (link)}.
Selected markers were validated using IHC on TMAs constructed with duplicate 0.6-mm tissue cores from tumor-rich selected areas of archival FFPE tumor blocks. Primary antibodies were anti-PTPN2 (clone 2A1D1, Fisher Scientific), monoclonal anti-CD123 (clone 7G3, BD Pharmingen), anti-CD163 (clone 10D6, Leica Biosystems), anti-CD68 (clone 514H12, Leica Biosystems), anti-CD8 (clone 4B11, Leica Biosystems), and anti-granzyme B (GrB) (clone 11F1, Leica Biosystems). Staining was done with the BOND RX automated Stainer system (Leica Biosystems). IHC counts were manually scored by one of the authors (J.L.S.R.). Finally, the most significant variables from ssGSEA that were correlated with a bad prognosis were included in a prognostic validation model.

Expression of the immune checkpoint molecules in CLL and RS lymph node samples was analyzed by IHC staining using antibodies specific for PD-L1 and PD1. Immune cell infiltration in CLL and RS lymph node samples was assessed by IHC staining using antibodies specific for CD3, CD8, FOXP3, and CD163. IHC staining was done on 4-μm FFPE sections on DAKO Autostainer Plus (Agilent, Santa Clara, CA) using standard protocol^{17 (link),20 (link)}. The antibodies used include anti-PD-L1 (clone SP263, Ventana Medical Systems, Inc., Tucson, AZ), anti-PD1 (clone NAT105, Abcam, Inc., Cambridge, MA), anti-CD3 (clone LN10, Leica Biosystems, Newcastle Upon Tyne, UK), anti-CD8 (clone C8/144B, Dako, Carpenteria, CA), anti-FOXP3 (clone 236A/E7, Abcam, Inc., Cambridge, MA), and anti-CD163 (clone10D6, Leica Biosystems, Newcastle Upon Tyne, UK). IHC images were taken using whole slide imaging technology with MoticEasyScan Pro (Motic digital pathology, San Francisco, CA), and saved in tagged image file format. Percentage of expression for each individual antigen was calculated by dividing number of cells with positive staining by number of total cells in the image using the Image-Pro premier 3D 9.1.4 software (Media Cybernetics, Silver Spring, MD).

Additional spatial immune profiling was performed with multiplex immunofluorescence staining in a subset of the tumors (n = 30) by a previously described method (10 (link)). In short, the OPAL 7-color fluorescence immunohistochemistry (IHC) kit (Akoya Biosciences, USA) was used following the manufacturer’s instructions to stain for human cytokeratin (anti-CK, clone AE1/AE3 (Dako)), CD8+ cells (Anti CD8 clone C8/144B (Dako)), CD3+ cells (anti-CD3 polyclonal (Dako)), FoxP3 + cells (anti-FoxP3 clone 236A/E7 (Abcam)), CD163+ cells (anti-CD163 clone 10D6 (Novocastra)) and Ki67+ cells (anti-Ki67 clone SP6 (Abcam)). Slides were stored at 4°C until imaging. Whole slide and multispectral imaging were done using the Vectra^® Polaris™ multispectral scanning microscope (Akoya Biosciences, USA). Multispectral images were unmixed and analyzed per tumor case in INFORM^® (Akoya Biosciences, USA). All data was exported for analysis with the phenoptrReports package (Akoya Biosciences, USA) in RStudio (RStudio, Inc., Boston, MA, USA).