Phenochart

Opal Multiplex immunohistochemistry images were first traced, and selected tissues were unmixed in Phenochart (AKOYA Biosciences, USA). For all staining, the ipsilateral (left) dorsal and ventral horns were identified, and the area of positive staining quantified for IBA-1, P2RY12, GFAP, S100A10, LYZ and CD206 via thresholding in either HALO (Indica Labs, Albuquerque, NM, USA) or ImageJ (Bethesda, MD, USA). Each image was thresholded to remove the background and converted to an 8-bit grey-scale image. The area of pixels above the threshold was measured as a percentage area of the region of interest. For GFAP and S100A10 colocalisation, the object colocalisation module in HALO (Indica labs, USA) was used, whilst CD163+, GFAP+C3+ and CD11c+IGF1+ cells were counted manually using the cell counter plugin in ImageJ (Bethesda, USA). The experimenter was blinded to groups during analysis.

Indirect detection by fluorescence was based on the Opal^TM Multiplex IHC method (Akoya), and performed on the Autostainer Leica Bond RX. Samples were probed for RNAscope^® LS 2.5 Probe - Hs-ICOSLG - Homo sapiens inducible T-cell co-stimulator ligand (ICOSLG) transcript variant 1 mRNA or Negative control probe DapB (of Bacillus subtilis strain) following manufacturer’s instructions (ACDBio). Subsequent staining was performed with antihuman CD11c (Abcam) and PDPN (Abd Serotech) antibodies. Staining was performed using the Opal^TM 7-Color Automatic IHC Kit (Akoya) according to the manufacturer’s recommendations. Signal amplification and covalent binding of fluorophore was achieved by using a tyramide signaling amplification reagent (included in the Opal kit) that is conjugated with a different fluorophore for each cycle. Multispectral images were acquired at ×20 magnification using the Vectra Polaris Automated Quantitative Pathology Imaging System (Akoya). MoTIF settings were used for multispectral image acquisition. Multispectral image processing of multiplex IHC stains was performed using Phenochart (version 1.0.11/Akoya) and inForm Image Analysis Software (version 2.3, Akoya).

The mRNA expression levels for CXCL11, GBP1, and IDO1 were investigated using the RNAscope Multiplex Fluorescent Reagent Kit v2 and the RNAscope 4-Plex Ancillary Kit for Multiplex Fluorescent Kit v2, Advanced Cell Diagnostics, Inc. (ACD, Hayward, CA). The following probes were used: Hs-CD68-C4, Hs-CXCL11-C3, Hs-GBP1, Hs-IDO1-C2. For fluorescent detection, the label probes were conjugated to Opal 520 (CD68, 1:800), Opal 570 (CXCL11, 1:1500), Opal 620 (GBP1, 1:1500), and Opal 690 (IDO1, 1:800) from Akoya Biosciences/PerkinElmer (Waltham, MA, USA). The assay was performed according to the manufacturer’s instructions. Briefly, FFPE tissue slides were deparaffinizing and pre-treated before progressing onto hybridization with target probes. For the hybridization, slides were covered in a HybEZ™ Humidity Control Tray and placed in the HybEZ™ Oven and incubated at 40 °C for 2 h. Hybridization with target probes, preamplifier, amplifier, and label probes were performed according to manufacturer’s instructions. Nuclear staining was performed using DAPI. Images were acquired using the Akoya Vectra 3.0 Automated Quantitative Pathology Imaging System (P/N CLS142568) and analyzed using with Phenochart™ and inForm (Akoya Biosciences, Menlo Park, CA, USA). Adjustment of brightness and color merging of TIFF format files was performed using ImageJ.

Tissue sections were visualized using the confocal quantitative image cytometer CQ1 (Yokogawa) or the Aperio Scan Scope. Quantification of Iba1⁺ macrophages/MG and T‐cell populations was performed with Aperio ImageScope (v12.4.0.5043) using either a nuclear counting algorithm or a pixel counting algorithm depending on the complexity of cell shape. Multiplex stainings were acquired on Vectra Polaris (Akoya Biosciences Inc.) using Motif technology. Multispectral image analysis was performed with Phenochart, InForm Image analysis software (Akoya Biosciences Inc.), and HALO software (Indica labs).

The Polaris mIHC platform (Akoya) was used to measure spatial single-cell protein expression with 9-color antibody panels in mBC patient primary tumors and bone metastases. Slide staining and imaging was performed by the Human Immune Monitoring Shared Resource at the University of Colorado Cancer Center. Multispectral image FOVs were selected with Phenochart (Akoya v1.0.9), with up to 20 regions selected per tissue. FOVs were selected for tumor and stroma interaction while avoiding adjacent bone and bone marrow. Image analysis was performed with InForm (Akoya v2.6.0) to spectrally unmix images, segment tissue, segment cell compartments, and phenotype cells. Nearest-neighbor spatial analysis and complex phenotype evaluation was then performed using phenoprReports (Akoya). For each patient, the average cell phenotype count was calculated.

For multispectral immunophenotyping, the murine-specific kit from Akoya (NEL840001KT) was used as previously described (45 (link)), with additional markers that were evaluated in two separate panels. In the first panel, the following primary antibodies were used: anti-FOXP3, anti-CD4 and anti-CD8; the second panel included: anti-F4/80, anti-CD86 and anti-Arginase-1. The source and dilution of the antibodies as well as other experimental details can be found in Supplementary Table S4. Sample scanning, spectral unmixing and quantification of signals were conducted with the Vectra Polaris Automated Quantitative Pathology Imaging System (Akoya), using the Phenochart (Akoya) and QuPath-0.4.3 software. Data were given as number of cells with a specific immunophenotype/total number of cells.