Benchmark ultra

Tissue sections were deparaffinized with xylene and rehydrated using a graded ethanol series. Heat-induced antigen retrieval was performed in a high-pH antigen retrieval buffer (BenchMark ULTRA; Roche, Basel, Switzerland). Endogenous peroxidase was blocked by incubation at 36 °C with 3% H₂O₂ for 4 min. Sections were subsequently labeled using the horseradish-peroxidase-labeled polymer method (Ventana ultraView DAB Universal Kit; Roche) and an automated immunostaining system (BenchMark ULTRA; Roche). Immunostained sections were counterstained with hematoxylin, dehydrated in ethanol, and cleared in xylene. Finally, sections were stained with anti-CCR7 (clone P-007, 1:100,000 dilution; NB Health Laboratory, Sapporo, Japan), anti-E-cadherin (clone 36, RTU, RRID: AB_397580; Ventana/Roche), and anti-vimentin (clone V9, RTU, RRID: AB_306239; Ventana/Roche) mouse monoclonal primary antibodies.

Formalin-fixed paraffin-embedded tissue underwent hematoxylin and eosin staining as well as immunohistochemistry to establish the tumor’s histopathological classification and mismatch repair status. Rabbit polyclonal primary antibodies (Roche Ventana, Tucson, AZ) were used in the detection of ACTH, LH, FSH, TSH, prolactin, growth hormone, P53, and Ki-67 on a Benchmark Ultra (Roche Ventana, Tucson AZ). Monoclonal antibodies from clone N1665 (Thermo Fisher Scientific, Rockford, IL) were used to detect SF-1; monoclonal antibodies from clone CL6251 (Sigma-Aldrich, St. Louis, MO) were used to detect T-PIT/TBX19; monoclonal antibodies from clone 8B6.1 (EMD Millipore, Burlington, MA) were used to detect PIT-1; monoclonal antibodies from clone ES05 (Leica Biosystems, Wetzlar, Germany) were used to detect MLH1; monoclonal antibodies from clone G21-91129 (Cell Marque, Rocklin, CA) were used to detect MSH2; monoclonal antibodies from clone EP49 (Agilent Dako, Santa Clara, CA) were used to detect MSH6; and monoclonal antibodies from clone A16-4 (BD Bioscience, San Diego, CA) were used to detect PMS2 on a Bond-III IHC and ISH stainer (Leica Biosystems, Wetzlar, Germany). In the treatment-refractory cohort, Ki-67 from the diagnostic resection was extracted from pathology reports when banked tissue was unavailable due to the age of these samples.

The 3 µm thick sections of TMA blocks were immunostained with a rabbit anti-HER2/neu (4B5) monoclonal antibody (Ventana Medical Systems, Tucson, AZ, USA) using an autoimmunostainer (BenchMark ULTRA, Ventana Medical Systems) [13 (link)]. An OptiView DAB detection kit (Ventana Medical Systems) was used for the detection of immunoreactions against HER2 [14 (link)]. In addition, HER2-positive breast cancer tissues were immunostained in parallel as a positive control.
Two pathologists (WYK and JHP) independently performed the immunohistochemical staining. Discrepant results in controversial cases were reviewed and determined in a common session using a multiview microscope.
HER2 protein expression on TMA blocks was evaluated by a four-tier system based on the DAKO HercepTest guidelines (DAKO, Glostrup, Denmark) [13 (link),15 (link)]. The immunohistochemical staining was scored as follows: 0, no staining or membrane staining in less than 10% of the tumor cells; 1+, faint/barely perceptible membrane staining in >10% of the tumor cells; 2+, weak-to-moderate complete membrane staining in more than 10% of the tumor cells; and 3+, strong complete membrane staining in more than 10% of the tumor cells [13 (link),15 (link)]. Scores of 2+ and 3+ were considered to indicate positive HER2 expression.

We processed tissue slides in a BenchMark ULTRA platform instrument (Ventana Medical Systems, Roche, Tucson, AZ, USA) and stained them with SP263 antibody, which we prediluted according to the manufacturer’s instructions. Following the established recommendations, we quantified the percentage of cells with membrane positivity (partial or complete expression) for each tumor. We did not assess tumor necrosis areas and discarded cases in which at least 50 viable cells were not available [33 (link)]. We determined the tumor proportion score by calculating the percentage of tumor cells membrane staining at any intensity. We considered PD-L1 expression in tumor cells positive if ≥ 1% of tumor cells had membranous staining of any intensity and high if > 50%.

Immunohistochemical staining was performed using an automated immunostainer (Benchmark ULTRA, Ventana). The antibodies against the following proteins were: SMARCA2 (EPR23103-44, Abcam, 1:400), ARID1A (EPR13501, Abcam, 1:1000), ARID1B (2D2, Abcam, 1:500), BRG1 (E8V5B, Origene, Ready-to-use), INI1 (25, Origene, Ready-to-use), Napsin-A (IP64, Origene, Ready-to-use), SOX-2 (EP103, Origene, Ready-to-use), P53 (DO-7, Ibp, Ready-to-use), TTF-1 (SPT24, Ibp, Ready-to-use), Ki-67(MyM1-Ki67, Ibp, Ready-to-use), ALK (D5F3, Ventana, Roche), and PD-L1 (SP263, Ventana, Roche).

University of California-Irvine, Pathology Services used Ventana BenchMark Ultra protocols for immunostaining of brain sections. To determine the localization of Tau in the human AD brain sections, neighboring slices were immunostained with the DAKO polyclonal antibody, which binds to total Tau which detects all 6 six isoforms of Tau (dilution 1: 3000, A0024; Agilent, CA, USA). Immunostained sections were scanned using the Ventana Roche instrumentation and the images were analyzed using QuPath software version 0.4.3.