1. Appendix Supplementary data (online only) to: Marleen Kok, Wilbert Zwart, Caroline Holm, Renske Fles, Michael Hauptmann, Laura J. Van ’t Veer, Lodewyk F.A. Wessels, Jacques Neefjes, Olle Stål, Sabine C. Linn, Göran Landberg, Rob Michalides PKA-induced phosphorylation of ERα at serine 305 and/or high PAK1 levels predict tamoxifen resistance in the majority of ER positive breast cancer cases

2. Figure A1. Staining for the catalytic subunit of PKA phosphorylated at threonine 197. A) Representative example of positive pPKA staining. B) Absence of pPKA expression after dephosphorylation via incubation with lambda phosphatase on the tumor slide prior to immunohistochemistry. ABAB

3. 564 randomized patients Start of trial 70 excluded (no FFPE tissue and no data on ERα) 494 with ERα data 384 ERα >10% 110 excluded (ERα ≤10%) 231 with ERα305-P, pPKA or PAK1 data 153 excluded (no ERα305-P, pPKA and PAK1 data) 262 events 32 events 230 events 55 events 175 events 71 events 104 events Figure A2. Flow diagram of patients Flow of patients through the study including number of patients in each stage. Reasons for dropout and number of events (recurrence) in each subgroup are listed.

4. N=231 Log-rank=0.020 HR=0.63, 95% CI 0.42-0.93, p=0.021 N=384 Log-rank=0.003 HR=0.64, 95% CI 0.47-0.86, p=0.004 Figure A3. Overall tamoxifen benefit in patients included in the current translational study (n=231) compared to tamoxifen benefit in all ER-positive breast cancer patients included in the trial (n=384). Kaplan-Meier analysis of recurrence-free survival according to randomization in A) 384 patients of whom tumor material was available and the ER was expressed in >10% of tumor cells and in B) 231 patients of whom tumor material was available, the ER was expressed in >10% of tumor cells and data on PAK1, PKA and ER305-P were available. ---- tamoxifen, --- no adjuvant systemic treatment

5. NKI-AVL (training series) ProteinSupplierCatalogusnrDilutionPretreatmentScoringReference ERαLabvision/NeomarkersMS-750-S50xcitratenuclear, percentage of cellsHannemann et al. PRImmunologicACD 14400xcitratenuclear, percentage of cellsKok et al. HER2Thermo ScientificRM-9103-S25xcitratemembranous, 0/1/2/3Van de Vijver et al. PAK1Cell Signaling2602.0025xcitrateintensity 0-5, nuclear 0/1Holm et al. Phospho-ERα-S305 Millipore 124-9-420xcitratenuclear, percentage of cellsHolm et al. Phospho-PKACell Signaling478140xcitrateintensity 0-5, nuclear 0/1x Lund (validation series) ProteinSupplierCatalogusnrDilutionPretreatmentScoringReference ERαVentana Medical Systemsclone 6F11predilutedcitratenuclear, percentage of cellsRyden et al. Ki-67DakoM7240200xcitratenuclear, percentage of cellsJirstrom et al. PAK1Cell Signaling2602.0025xcitrateintensity 0-5, nuclear 0/1Holm et al. Phospho-ERα-S305 Millipore 124-9-420xcitratenuclear, percentage of cellsHolm et al. Phospho-PKACell Signaling478140xcitrateintensity 0-5, nuclear 0/1x Table A1. Details immunohistochemistry

6. Table A2. Distribution of pPKA according to clinicopathological parameters (training series NKI)

7. Insufficient data for ScoreData for Score available VariableCategoryn%n%P-value Total153231 Age at surgeryMedian in years (range)45(27-57)45(26-57)0.25¶ Grade*I/II III Unknown 101 39 13 72% 28% 151 77 3 66% 34% 0.25 Lymph node statusNegative Positive 39 114 25% 75% 55 176 24% 76% 0.72 WHO subtypeInvasive Ductal Carcinoma Invasive Lobular Carcinoma Unknown or other type 104 14 35 88% 12% 196 23 12 89% 11% 0.72 Size≤ 20 mm > 20 mm Unknown 64 88 1 42% 58% 97 134 42% 58% 1.00 Progesterone receptor (IHC)≤ 10% > 10% Unknown 11 94 48 10% 90% 22 196 13 10% 90% 1.00 Ki67 (IHC)≤ 25% > 25% Unknown 86 14 53 86% 14% 165 38 28 81% 19% 0.34 Table A3. Randomized controlled trial (Lund) used as validation series. Distribution of clinico-pathological parameters in the excluded cases (ERα>10%, (no data to generate PAK1-PKA/ ER  S305-P Score, n=153) vs included patients in the validation series (n=231). * According Nottingham Grading system (Elston et al.).

8. Table A4. Distribution of prognostic factors in the subgroup of tumors classified as sensitive by the PAK1-PKA/ ER  S305-P Score in the validation set (PAK1 and/or PKA/ER305-P negative) according to systemic treatment received (tamoxifen versus no adjuvant systemic treatment). * According Nottingham Grading system (Elston et al.). † Mann-Whitney U test ‡ Fisher’s Exact test Abbreviations: LN lymph node status, ER estrogen receptor, PR progesterone receptor, IDC invasive ductal carcinoma, ILC invasive lobular carcinoma. TamoxifenControlP-value VariableCategoryN=78N=91 Age at surgeryMedian(range)45 (36-57)46 (26-52)0.79† LNNegative Positive 24 (31%) 54 (69%) 18 (20%) 73 (80%) 0.11‡ Size<= 20 mm > 20 mm 28 (36%) 50 (64%) 38 (42%) 53 (58%) 0.53‡ Grade*I/II III Unknown 48 (62%) 29 (38%) 1 61 (68%) 29 (32%) 1 0.52‡ ER levels>10,<75% >75% Unknown 30 (39%) 47 (61%) 1 36 (41%) 52 (59%) 3 0.87‡ PR<=10% >10% Unknown 9 (13%) 62 (87%) 7 9 (10%) 77 (90%) 5 0.80‡ WHOIDC ILC Unknown/other 66 (89%) 8 (11%) 4 73 (86%) 12 (14%) 6 0.63‡