Immunohistochemistry for RAD51/Geminin (GMNN) double staining

FFPE samples taken 24–48 h after the last dose of rucaparib were cut, deparafinised, rehydrated and stained with hematoxylin and eosin (H&E) and double stained with GMNN and RAD51. Pre- and post-radiotherapy-induced squamous cell carcinoma were used as negative and positive controls.

Antigen retrieval was performed and RAD51 primary antibody (mouse monoclonal, Genetex, GTX70230) was diluted 1/200 in Dako antibody diluent (K8006) and applied. Slides were incubated with Dako Envision Flex HRP (K8002). Geminin (GMNN) antibody (rabbit polyclonal, Proteintech 10802-1-AP) diluted 1/1500 in Dako antibody diluent was applied and incubated with Dako Envision Flex HRP (K8002). Sections incubated with Vector TMB blue (SK4400) and counterstained with Gills 1 hematoxylin, air dried and dehydrated in xylene before mounted and cover slipped with Vectamount.

Five random fields at 40× magnification were identified and marked in PathXL. GMNN staining was identified with blue/green staining and RAD51 was identified by the presence of brown nuclear foci. Scoring was done by 2 scorers blinded to each other, time point and clinical details.

The number of tumour cells, GMNN-positive cells and RAD51-positive cells were counted. Cells with 5 or more RAD51 foci were classified as RAD51 positive. A minimum of 300 tumour cells and 30 GMNN-positive cells were a minimum requirement for inclusion. Raw data were collected and the proliferation fraction ((no. of GMNN-positive cells / total number of tumour cells) × 100) and RAD51 score ((no. of RAD51-positive cells / no. of GMNN-positive cells) × 100) was calculated.