Supplementary Information for

FOXA1 overexpression suppresses interferon signaling and immune response in cancer

Yundong He, Liguo Wang, Ting Wei, Yu-Tian Xiao, Haoyue Sheng, Hengchuan Su, Daniel P. Hollern, Xiaoling Zhang, Jian Ma, Simeng Wen, Hongyan Xie, Yuqian Yan, Yunqian Pan, Xiaonan Hou, Xiaojia Tang, Vera J. Suman, Jodi M. Carter, Richard Weinshilboum, Liewei Wang, Krishna R. Kalari, Saravut J. Weroha, Alan H. Bryce, Judy C. Boughey, Haidong Dong, Charles M. Perou, Dingwei Ye, Matthew P. Goetz, Shancheng Ren and Haojie Huang

This PDF file includes: Supplementary Methods; Supplementary Figures 1-19; Supplementary Tables 1-5.

Supplementary Methods Cell lines and cell culture LNCaP, VCaP, DU145, PC3, NCI-H660, MCF7, MDA-MB-231, RT4, TRAMP-C2, MyC-CaP and 293T cell lines were purchased from ATCC. LNCaP, VCaP and PC3 cells were maintained at 37°C and 5% CO2 in RPMI 1640 containing 10% fetal bovine serum (FBS) and 1% antibiotic/antimycotic (Thermo Fisher Scientific). NCI-H660 cells were maintained at 37°C in HITES medium (RPMI1640 medium plus 0.005mg/ml Insulin, 0.01mg/ml Transferrin, 30nM Sodium selenite, 10nM Hydrocortisone, 10nM beta-estradiol and 2mM L-glutamine) with 10% FBS and 1% antibiotic/antimycotic (Thermo Fisher Scientific). MCF7, DU145, TRAMP-C2, MyC-CaP and 293T cells were maintained in DMEM medium with 10% FBS. MDA-MB-231 cells were maintained at 37°C and 5% CO2 in Leibovitz's L15 containing 10% fetal bovine serum (FBS) and 1% antibiotic/antimycotic (Thermo Fisher Scientific). RT4 cells were maintained at 37°C and 5% CO2 in McCoy's 5a containing 10% fetal bovine serum (FBS) and 1% antibiotic/antimycotic (Thermo Fisher Scientific). Mycoplasma contamination was tested by the PCR Mycoplasma Detection Set (Takara, Otsu, Japan). All cell lines are negative for mycoplasma contamination.

RNA-seq analyses and real-time PCR For RNA-seq, libraries were prepared using Illumina TruSeq RNA prep kit and standard protocol. The RNA libraries were sequenced as 51 nt pair-end reads at one sample per lane of an Illumina HiSeq 4000, generating an average of 265 million reads per sample. All reads were aligned to the human reference genome (hg19/GRCh37) by TopHat 2.0.9 using default options. Gene expression counts were generated using RSeQC and expression differential analysis was conducted using edgeR (version 3.6.8) (1). Differentially expressed genes were determined based on the false discovery rate (FDR) threshold 1e-5 and log2 fold change threshold of 1.0. The Gene ontology biological process (GO-BP) analysis was performed by gene set enrichment analysis online (http://software.broadinstitute.org/gsea/msigdb/annotate.jsp) (2, 3). For RT-PCR, cells treated with or without 100 ng/L IFNα (SigmaAldrich, # I4276) for 24 h were harvested for RNA extraction using Trizol Reagent (Thermo Fisher Scientific, # 15596018). RNA was eluted in

RNase-Free H2O and reverse-transcribed to cDNA following the manufacturer’s protocol (Thermo Fisher Scientific, # FERK1672). Gene expression was determined by qPCR using Power SYBR Green (Thermo Fisher Scientific, # 4368708). Primer sequences used for RT- qPCR are listed in Supplementary Table 4.

Luciferase reporter assay for interferon (IFN)-stimulated response activity To analyze the interferon-stimulated response activity, cells were transfected using lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions with the following plasmids: interferon-stimulated response element luciferase reporter (ISRE-luc, purchased from Promega) containing type Ⅰ and Ⅲ IFN response elements, IFN-γ-activated sequence luciferase reporter (GAS-luc, a gift from Dr. Avudaiappan Maran at Mayo Clinic) (4) containing type Ⅱ IFN response elements, Renilla-luc (phRL-TK, purchased from Promega) as internal control reporter, FOXA1-WT or FOXA1 mutants. At 24 h after transfection, the cells were treated with 100 ng/L IFNα (Sigma-Aldrich, # I4276) or IFNγ (Sigma-Aldrich, # SRP3058) for 5 hours. Renilla and firefly activities were measured with luminometry using the Dual-Luciferase Reporter Assay System (Promega) and the ratio was calculated. Results were shown as the ratio of firefly over Renilla luciferase activity.

Plasmid construction and mutagenesis Wild-type (WT) V5-tagged FOXA1 lentiviral plasmid was purchased from Addgene (# 70090) and cloned into the SFB-tagged pcDNA3.1 or Flag-tagged pcDNA3.1 or pTSIN lentiviral vector using Phusion High-Fidelity DNA Polymerase (New England Biolabs, # M0530L). FOXA1 hotspot mutations (H247Q, R261G and F266L) and FOXA1 truncation mutations were engineered from FOXA1-WT using the KOD-Plus-Mutagenesis Kit (TOYOBO, # KOD-201) according to the manufacturer’s instructions. To generate theFOXA1 luciferase reporter (FOXA1-luc) construct, six copies of the FOXA1 response element from the KLK3 (PSA) gene 5’- tcgaTGTTTACTTAcagtaTGTTTACTTTatccgTGTTTACATAgtctaTATTTACTTAccataTGTT TGCTTAgtcaTGTTTACTCA-3’ were inserted into pGL4.28 luc2CP/minP/hygro (Pomega). All plasmids were confirmed using Sanger sequencing. Mutant plasmids were further transfected in 293T cells to confirm the expression of the mutant proteins.

Cell transfection and RNA interference 293T cells were co-transfected with pTSIN-Vector or pTSIN-FOXA1 WT or mutant lentiviral plasmids along with packaging and envelope plasmids by Lipofectamine 2000 according to the manufacturer’s instructions. At two days post-transfection, virus particles were used to infect cells according to the protocol provided by Sigma-Aldrich. The interest cells were infected with a 1:1 mixture of fresh medium and virus supernatant supplemented with Polybrene (4 μg/ml final concentration) for 24 h. To knock down FOXA1, cells were transfected using Lipofectamine 2000 with 50 nM FOXA1 siRNA (5’-GAGAGAAAAAAUCAACAGC-3’ (siFOXA1#1, at the 3’ UTR region) or 5’-GCACUGCAAUACUCGCCUU-3’ (siFOXA1#2, at the CDS region)) or siSTAT2 (L-012064-00-0005, Dharmacon) or non-targeting control siRNA (siCon) 5'- UAGCGACUAAACACAUCAA-3' (synthesized from Dharmacon). For inducible shFoxa1 construction, the shRNA sequence against Foxa1 (shFoxa1#1 sense: 5’- TCGAGTGCCTTCTCAATGTTAATTTACTCGAGTAAATTAACATTGAGAAGGCATTTTT G-3’; shFoxa1#1 antisense: 5’- AATTCAAAAATGCCTTCTCAATGTTAATTTACTCGAGTAAATTAACATTGAGAAGGC A-3’; shFoxa1#2 sense: 5’- TCGAGACTCTCCTTATGGCGCTACCCTCGAGGGTAGCGCCATAAGGAGAGTTTTTTG- 3’; shFoxa1#2 antisense: 5’- AATTCAAAAAACTCTCCTTATGGCGCTACCCTCGAGGGTAGCGCCATAAGGAGAGTC -3’) was cloned into the lentiviral doxycycline-inducible pINDUCER10 vector (5) using XhoI and EcoR I sites. Then lentiviral plasmids along with packaging and envelope plasmids by Lipofectamine 2000 according to the manufacturer’s instructions. At two days post-transfection, virus particles were used to infect cells according to the protocol provided by Sigma-Aldrich. The MyC-CaP cells were infected with a 1:1 mixture of fresh medium and virus supernatant supplemented with Polybrene (4 μg/ml final concentration) for 24 h. The knockdown or transfection efficiency was determined using Western blotting analysis.

Western blot The cells were lysed and equal amounts of protein (50–100 μg) from cell lysate were denatured in sample buffer (#NP0007, Thermo Fisher Scientific), subjected to SDS-polyacrylamide gel electrophoresis (PAGE), and transferred to nitrocellulose membranes (Bio-Rad). The membranes were immunoblotted with specific primary antibodies, horseradish peroxidase (HRP)-conjugated secondary antibodies, and visualized by SuperSignal West Pico Stable Peroxide Solution (# 34577, Thermo Fisher Scientific). The primary antibodies used in this study include AR (dilution 1:1000; #sc-816, Santa Cruz Biotechnology), FOXA1 (dilution 1:2000; # ab23738, Abcam), FOXA1 (dilution 1:1000, # sc-101058, Santa Cruz Biotechnology), FOXA2 (dilution 1:1000; #8186, Cell Signaling Technology), FOXA3 (dilution 1:500, # sc-74424, Santa Cruz Biotechnology), STAT1 (dilution 1:1000; #14994S, Cell Signaling Technology), STAT2 (dilution 1:1000; # 72604S, Cell Signaling Technology), Phospho-STAT1 (Tyr701) (dilution 1:1000; # 9167S, Cell Signaling Technology), Phospho-STAT2 (Tyr690) (dilution 1:1000; # 88410S, Cell Signaling Technology), IRF9 (dilution 1:1000; # 76684S, Cell Signaling Technology), ISG15 (dilution 1:500; # sc-166755, Santa Cruz Biotechnology), PARP1 (dilution 1:1000; # 9532S, Santa Cruz Biotechnology), HSP70 (dilution 1:1000; #4873S, Santa Cruz Biotechnology), MHC class I (MHC-I) (dilution 1:500; # sc-55582, Santa Cruz Biotechnology), Flag (dilution 1:1000; # F1804, Sigma-Aldrich) and V5 (dilution 1:1000; # A190-120A, Bethyl Laboratories), HA (dilution 1:1000; # MMS-101R, Covance), HDAC2 (dilution 1:1000; # sc- 7899, Santa Cruz Biotechnology) and ERK2 (dilution 1:2000; # sc-1647, Santa Cruz Biotechnology).

Co-immunoprecipitation (co-IP), protein purification and GST pulldown assay The NE-PER Nuclear and Cytoplasmic Extraction Kit (# 78835, Thermo Fisher Scientific) was used for extraction of nuclear and cytoplasmic proteins from cells according to the manufacturer’s instruction. For extraction of whole cell lysate, we used IP buffer (50 mM Tris- HCl pH7.5, 150 mM NaCl, 1% NP40, 0.5% Sodium Deoxycholate). Protein A/G agarose (# 20421, Thermo Fisher Scientific) was used for immunoprecipitation of FOXA1 (# ab23738, Abcam) and Flag-tag (# F1804, Sigma-Aldrich). Monoclonal anti-HA agarose (# A2095, Sigma- Aldrich) was used for HA-tag Co-IP. For glutathione-S-transferase (GST) pulldown assay, GST- tagged STAT2 fragment expressed at E. coli was purified by Glutathione Sepharose 4B beads (# 84-239, Genesee Scientific) and incubated with lysate from 293T cells expressing interested proteins or in vitro transcribed/translated proteins followed by GST pulldown assays. For protein co-IP, samples eluted in IP butter were incubated with agarose beads and antibodies overnight at 4 °C, followed by wash for 6 times with IP buffer in the following day. Samples were boiled for 10 min in 50 µl sample buffer (2% SDS, 10% glycerol, 10% β-mercaptoethanol, bromophenol blue and Tris-HCl, pH 6.8) and subjected to Western blotting analysis.

FOXA1 luciferase reporter assay and electrophoretic mobility shift assay (EMSA) For FOXA1 transcriptional activity analysis, 293T and other cells were transfected with SFB- tagged pcDNA3.1 vector (control) or different mutants of FOXA1, KLK3 enhancer luciferase reporter and Renilla-luc (phRL-TK, purchased from Promega, as internal control reporter). At 48 h after transfection, the renilla and firefly luciferase activities were measured with luminometry using the Dual-Luciferase Reporter Assay System (Promega) and the ratio was calculated. Results were expressed as the ratio of firefly over Renilla luciferase activity. For EMSA, The nuclear extract was from DU145 cells expressing exogenous FOXA1 and other proteins as indicated. We used sixty base pairs of forkhead response element in the KLK3 enhancer (centered at the FOXA1 consensus binding motif 5’-GTAAACAA-3’: 5’- ACATATTGTATCGATTGTCCTTGACAGTAAACAAATCTGTTGTAAGAGACATTATCT TTA-3’) and the ISRE probe (5’-CTCCCCTGAGTTTCACTTCTTCTCCCAACTTG-3’). These oligoes were synthesized through the Integrated Device Technology (IDT) and labelled with biotin using Biotin 3’-End DNA labelling kit (# 89818, Thermo Fisher Scientific) and annealed to generate a labelled double stranded DNA duplex. Binding reactions were carried out in 20 μl volumes containing 2 μl of the nuclear lysates, 50 ng/μl poly(dI.dC), 1.25% glycerol, 0.025%

Nonidet P-40 and 5 mM MgCl2. Biotin labelled KLK3 enhancer probe (10 fmol) was added and incubated for 1 h at room temperature, size-separated on a 6% DNA retardation gel at 100 V for 1 h in 0.5× TBE buffer, and transferred on Biodyne Nylon membrane (# 77015, Thermo Fisher Scientific) and crosslinked to the membrane using the UV light at 120 mJ/cm2 for 2 min. Biotin- labelled free and protein-bound DNA was detected using HRP-conjugated antibody and developed using Chemiluminescent Nucleic Acid Detection Module Kit (# 89880, Thermo Fisher Scientific) according to the manufacturer’s protocol.

Biotin-labeled DNA pulldown assay followed by co-IP assay To explore the effect of FOXA1 on the DNA binding ability of STAT2 in vitro, biotin labeled- ISRE DNA was used to pull down STAT2 proteins (from chromatin-free nuclear extract treated with IFNα in FOXA1 negative DU145 cells) in the presence or absence of FOXA1 proteins (produced by in vitro transcription/translation system). The biotin labeled-ISRE DNA was generated by biotin using Biotin 3’-End DNA labelling kit as mentioned above. The chromatin- free nuclear extract from DU145 cells with or without IFNα treatment was prepared by Subcellular Protein Fractionation Kit (Thermo Fisher Scientific, #78840) following the manufacturer’s protocol. The FOXA1 and its mutant proteins were produced by TNT Quick Coupled Transcription/Translation System (Promega, # L1170) following the manufacturer's instructions. For the DNA pulldown assay, 10 pmol biotinylated DNA duplex was incubated with the nucleoplasm from DU145 (1×106 cells) and different dose of FOXA1 proteins in binding buffer (10 mM Tris-Cl pH7.5, 100 mM NaCl, 0.01% NP40 and 10% glycerol) overnight at 4˚C. Streptavidin-coupled sepharose beads (Cell Signaling Technology, #3419) were added the next day, and incubated for another 2 hours with the samples. The unbound supernatant was collected and anti-Flag M2 beads (Sigma-Aldrich, #A2220) were added for further Flag IP. The beads were collected and washed with binding buffer five times. The beads were subsequently boiled in 1× loading buffer and analyzed by western blotting.

Tumor cell injection and treatment For TRAMP-C2 model, TRAMP-C2 cells (3 × 106) transfected with Vector, FOXA1-WT, FOXA1 αH3 or FOXA1 SBR were injected subcutaneously into the right flank of 8-week old

male C57BL/6△ (purchased△ from Jackson Laboratories (Bar Harbor, ME, USA)) on day 0. When the average tumor volume reached about 100 mm3, PolyI:C (2.5 mg/kg, 100 μl) purchased from Sigma-Aldrich (# P1530) or vehicle (PBS, 100 μl) was administered by intratumoral injection every two days for two weeks and tumors were measured twice per week until the tumor volume reached the maximum allowed size. The tumor volume was calculated using a formula of length × width × width × 0.5. For CyTOF, immunofluorescence cytochemistry and RT-qPCR experiments, mice were euthanized at 48 h post the last administration of PolyI:C. The protocol for the animal experiment was approved by the Mayo Clinic Institutional Animal Cure and Use Committee (IACUC). For MyC-CaP model, MyC-CaP cells (3 × 106) transfected doxycycline-inducible pINDUCER10-shFoxa1 (MyC-CaPIN-shFoxa1#1) were injected subcutaneously into the right flank of 6-week old male FVB mice (purchased from Jackson Laboratories (Bar Harbor, ME, USA)). When the average tumor volume reached about 120 mm3, mice were randomized into groups subsequently received 5% sucrose only (Dox-) or 5% sucrose plus 1 mg/mL doxycycline (Sigma, #D-9891) (Dox+) receptively. All water bottles were freshly prepared every second day, protected from light. Meanwhile, mice were treated with antibodies delivered using 100ul volumes and injected every four days for the following: non-specific control IgG (350 μg/dose, #BE0083, BioXcell) or combination of anti-CTLA4 (100 μg/dose, #BE0131, BioXcel) and anti- PD1 (250 μg/dose, #BE0146, BioXcel) (6) receptively. Tumors were measured with calipers and mice weighed twice a week. Mice whose tumors reached 500 mm3 were euthanized. For Flow cytometry analysis, immunofluorescence cytochemistry and RT-qPCR experiments, mice were euthanized at 48 h post the last administration of the antibodies. The protocol for the animal experiment was approved by the Mayo Clinic Institutional Animal Cure and Use Committee (IACUC).

Flow cytometry analysis (FCA) Tumor samples were digested to single cells with tumor dissociation kit (Miltenyi Biotec, # 130- 096-730) following the manufacturer’s protocol and then samples were fixed with FOXP3 Fixation/Permeabilization 1× working solution by diluting 4× Fixation/Permeabilization Concentrate (eBioscience, # 00-5123-43) with Fixation/Permeabilization Diluent (eBioscience, # 00-5223-56) at 1:4 dilution for 45 min at room temperature. Samples were stained with anti- FoxA1 (1:200 dilution; # ab23738, Abcam), Alexa Fluor 647 anti-CD8a (1:200 dilution; # 100724, BioLegend) and Alexa Fluor 594 anti-granzyme B (1:200 dilution; #ABIN6826724, Antibodies-online) and subsequently stained with secondary antibodies cross-adsorbed Alexa Fluor 488 goat anti-rabbit IgG (1:500 dilution; Invitrogen, # A-11037) for 30 min. Flow data were collected on BD FACSCanto™ flow cytometer and analyzed using FlowJo v10.

Single-cell mass cytometry (CyTOF) analysis Single tumor cells were isolated using the mouse tumor dissociation kit (Miltenyi Biotec, # 130- 096-730) following the manufacturer’s protocol. CyTOF staining panels are detailed in Supplementary Table 5. The separated single cells were filtered through 40 µm filters into RPMI-1640 and centrifuged at 300 g for 5 min at 4 °C. All single cells were depleted of erythrocytes by hypotonic lysis for 1 min at room temperature. Cells were washed once with PBS and incubated with 0.5 mM of Cell-IDTM Cisplatin (Fluidigm, # 201064) for 5 min. 5 × 106 or fewer cells per tumor were blocked with FcR Blocking Reagent (Miltenyl Biotec, # 130- 059-901) for 10 min and incubated with surface antibody mix for 45 min at room temperature. Cells were washed with MaxPar Cell Staining Buffer (Fluidigm, # 201068). For intracellular staining, cells were incubated with FOXP3 Fixation/Permeabilization 1× working solution by diluting 4× Fixation/Permeabilization Concentrate (eBioscience, # 00-5123- 43) with Fixation/Permeabilization Diluent (eBioscience, # 00-5223-56) at 1:4 dilution for 45 min at room temperature (keep in dark). Cells were washed twice with 1× working solution of Permeabilization Buffer (eBioscience, # 00-8333-56) and centrifuged at 800g for 5min. The supernatant was carefully aspirated and the pellet was resuspend in 500 µL CyPBS diluted with 10× PBS (Rockland Immunochemicals, # MB-008) in Maxpar water (Fluidigm, # 201069). Samples were fixed with 500 µl of 2× fixation solution 4% Paraformaldehyde diluted from 16% Paraformaldehyde Aqueous Solution (Electron Microscopy Sciences, # 915710S) in CyPBS and incubated at 4°C overnight. Cells were washed with 1 ml Maxpar Cell Staining Buffer (Fluidigm; # 201068) and centifuged at 800 × g for 5 min at room temperature and resuspended in 1 ml 12.5 nM intercalation solution by diluting 125 µM intercalator stock (Cell-IDTM lntercalator-Ir- 125 µM, Fluidigm, Part No. 201192A) at 1:10,000 with Maxpar Fix and Perm Buffer-100 mL (Fluidigm; # 201067). The samples were washed with 1 ml of CyPBS and EQ beads (Fluidigm; # 201078) were added. Cells were counted on Countess II and re-suspended to approximately 5×105 cells/mL. Samples were filtered through 35 µm blue cap FACS tube (Falcon, #352235) and were analyzed with CyTOF instrument (Fluidigm) in the Flow Cytometry and Cellular Imaging Core Facility at Mayo Clinic. Data were analyzed with PhenoGraph by following the instruction (7). R studio (Version 1.0.136) was downloaded from the official R website (www.r- project.org/). The cytofkit package (Release 3.6) was downloaded from Bioconductor (https://www.bioconductor.org/packages/release/bioc/html/cytofkit.html) and opened in R. The gated singlet (19Ir + 193Ir +), viable (195Pt +) events were imported into cytofkit, subjected to PhenoGraph analysis, and clustered on the basis of markers. The files were clustered with the PhenoGraph algorithm and tSNE was selected as the visualization method. 44 unique clusters were identified using PhenoGraph. The unique clusters identified with PhenoGraph were visualized via the R package “Shiny,” where labels, dot size, and cluster color were customized. Clusters were colored according to the phenotype based on the median expression of various markers. The frequency of each cluster was determined via csv files generated by the algorithm. The Percentage of each cell population was analyzed with the FlowJo and GraphPad Prism 8 software.

Immunofluorescence cytochemistry Formalin-fixed paraffin-embedded TRAMP-C2 and MyC-CaP tumor samples were deparaffinized, rehydrated and subjected to heat-mediated antigen retrieval. Sections were incubated with 1% Sudan Black (dissolved in 70% ethanol) for 20 min at room temperature to reduce autofluorescence. Slides were washed with 0.02% Tween 20 and incubated with 0.1 M Glycine for 10 min. And immersed slides in 10 mg/ml Sodium Borohydride in ice cold Hanks Buffer on ice for 40 mins. After washing with two times PBS, slides were blocked by 1% BSA in PBS for 30 mins and incubated with FOXA1 antibody (1:1000 dilution; Abcam, # ab170933) and anti-mouse CD8 (1:1000 dilution; Santa Cruz Biotechnology, #sc-1177) at 4°C overnight. The sections were washed three times in 1× PBS and treated with secondary antibodies cross- adsorbed Alexa Fluor 594 goat anti-rabbit IgG (1:500 dilution; Invitrogen, # A-11037) or cross- adsorbed Alexa Fluor 488 goat anti-mouse IgG (1:500 dilution; Invitrogen, # A11001) for 30 min. Prior to imaging, sample were mounted with VECTASHIELD Antifade Mounting Medium with DAPI (Fisher Scientific, #NC9524612). Samples were imaged using Nikon spinning disk confocal.

Tumor mutational burden (TMB) analysis TMB was calculated according to previously defined methods (8, 9). We calculated TMB using exome data, including only non-silent mutations, as defined below. We only consider as somatic mutations: Frameshift Deletions, Frameshift Insertons, In-Frame Deletions, In-Frame Insertions, Missense Mutations, Nonsense Mutations, Nonstop Mutations, Splice Sites, Translation Start Sites and we ignore the following silent variants: 3’Flanks, 3’UTRs, 5’Flanks, 5’UTRs, IGRs, Introns, RNAs and Splice Regions. We quoted standard human exome length as 38Mb. The base substitutions and indels in the coding region of targeted genes, including synonymous mutations, were filtered for both oncogenic driver events according to COSMIC (Catalogue of Somatic Mutations in Cancer) and germline status according to dbSNP (The Single Nucleotide Polymorphism Database) and ExAC (The Exome Aggregation Consortium) databases.

Analysis of FOXA1 mRNA expression, IFN activity and immune cell markers in prostate, breast and bladder cancers of patients treated with immunotherapy or chemotherapy To evaluate the mRNA expression level of FOXA1, CD3E, CD8A and GZMB as well as IFN activity in patients with pathological complete response (pCR with residual cancer burden (RCB) 0 or I) and no pathological complete response (No-pCR with RCB II or III) to neoadjuvant chemotherapy (NAC) in breast cancer, RNA-seq data from Breast Cancer Genome-Guided Therapy (BEAUTY) (10) project at Mayo Clinic was analyzed. To further validate the data from the Mayo Clinic cohort, RNA-microarray data from NAC-treated breast cancer from independent cohorts (GSE106977; GSE41998; GSE34138; GSE21974) (11-14) were also analyzed similarly. To evaluate the impact of Foxa1 on immunotherapy response, IFN activity and cytotoxic T cell activities in murine breast cancer, the mRNA expression level of Foxa1, CD3e, CD8a and Gzmb as well as IFN activity from RNA-seq data (GSE124821) (15) was analyzed according to the responsiveness to anti-PD1 and anti-CTLA-4 combination therapy (resistant versus sensitive). To evaluate the impact of FOXA1 on the clinical outcome, IFN activity and cytotoxic T cell activities in metastatic castration-resistant prostate cancer (CRPC) patients treated with personalized peptide vaccination (PPV) (16), the mRNA expression level of FOXA1, CD3E, CD8A and GZMB as well as IFN activity from RNA microarray data (GSE53922) and clinical outcome was analyzed according to the FOXA1 expression level. Similarly, we also evaluated the correlations between FOXA1 and the clinical outcome, IFN activity as well as cytotoxic T cell activities in bladder cancers from different cohorts which treated with Bacillus Calmette- Guerin (BCG) immunotherapy (GSE19423) (17) or MVAC (methotrexate, vinblastine, doxorubicin, and cisplatin) chemotherapy (GSE70691) (18).

Clinical data and patient information of urothelial carcinomas treated with immunotherapy Tumor samples and medical records from a cohort of 23 patients (20 males and 3 females; Age from 44 to 77 years; Median age 65 years) with urothelial carcinoma (cancers in bladder, renal pelvis, ureter or urethra that showed predominantly transitional-cell features on histologic testing) treated at Fudan University Shanghai Cancer Center were analyzed. The protocol of this study was approved by the Ethical and Scientific Committee (equivalent to Institutional Review Board (IRB)) at Fudan University Shanghai Cancer Center. Urothelial carcinoma samples were obtained from the primary or metastatic lesions of 23 patients before they underwent therapy with anti-PD1 treatment until disease progression according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1,15 (19). All patients provided written consent before enrolling in the study. Formalin-fixed paraffin-embedded (FFPE) tumor specimens with sufficient viable tumor content were required before the start of the study. One sample (patient 7) was excluded for immunohistochemistry (IHC) staining evaluation because the specimen was too small. A total of 22 specimens were used for FOXA1 IHC staining evaluation.

Immunohistochemistry (IHC) of urothelial carcinoma patient specimens Urothelial carcinoma FFPE samples were deparaffinized, rehydrated and subjected to heat-

mediated antigen retrieval. Sections were incubated with 3% H2O2 for 15 min at room temperature to quench endogenous peroxidase activity. After antigen retrieval using unmasking solution (Vector Labs), slides were blocked with normal goat serum for 1 h and incubated with primary antibody at 4°C overnight. IHC analysis of tumor samples were performed using primary antibodies for FOXA1 (dilution 1:500; Abcam, # ab170933). The sections were washed three times in 1× PBS and treated for 30 min with biotinylated goat-anti-rabbit IgG secondary antibodies (#BA-9200, Vector Labs). After washing three times in 1× PBS, the sections were incubated with streptavidin-conjugated HRP (#3999S, Cell Signaling Technology). After washing three times in 1× PBS for 5 min each, specific detection was developed with 3,3′3diaminobenzidine (#D4168-50SET, Sigma-Aldrich). For IHC staining score (IS) or intensity, 0 = <1% positive cells, 1 = 1-20% positive cells, 2 = 20-50% positive cells, 3 = >50% positive cells. FOXA1 expression levels with IS = 0 or 1 are considered as “low” and IS = 2 or 3 are considered as “high”.

Data availability

All data and softwares used in this study are accessible. The raw data of single cell RNA-seq from prostate cancer patient tissues reported in this paper is accessible with Gene Expression

Omnibus (GEO) accession number GSE141445: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE141445, token: yhmhquokxjqfzgj. The GEO accession number for the ChIP-seq and RNA-seq data from cell lines reported in this paper is GSE142221: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE142221; Secure

token: yderayksdtazlsr. Previously deposited transcriptomic data (GSE106977; GSE41998;

GSE34138; GSE21974; GSE53922; GSE19423; GSE70691; GSE130408; GSE124821;

GSE128421) that are used in this study are also available at GEO

(https://www.ncbi.nlm.nih.gov/geo/).

Supplementary references 1. Robinson MD, McCarthy DJ, and Smyth GK. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2010;26(1):139-40. 2. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545-50. 3. Mootha VK, Lindgren CM, Eriksson KF, Subramanian A, Sihag S, Lehar J, et al. PGC-1alpha- responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet. 2003;34(3):267-73. 4. Zuo D, Shogren KL, Zang J, Jewison DE, Waletzki BE, Miller AL, 2nd, et al. Inhibition of STAT3 blocks protein synthesis and tumor metastasis in osteosarcoma cells. J Exp Clin Cancer Res. 2018;37(1):244. 5. Meerbrey KL, Hu G, Kessler JD, Roarty K, Li MZ, Fang JE, et al. The pINDUCER lentiviral toolkit for inducible RNA interference in vitro and in vivo. Proc Natl Acad Sci U S A. 2011;108(9):3665-70. 6. Jayaprakash P, Ai M, Liu A, Budhani P, Bartkowiak T, Sheng J, et al. Targeted hypoxia reduction restores T cell infiltration and sensitizes prostate cancer to immunotherapy. J Clin Invest. 2018;128(11):5137-49. 7. Kimball AK, Oko LM, Bullock BL, Nemenoff RA, van Dyk LF, and Clambey ET. A Beginner's Guide to Analyzing and Visualizing Mass Cytometry Data. J Immunol. 2018;200(1):3-22. 8. Chalmers ZR, Connelly CF, Fabrizio D, Gay L, Ali SM, Ennis R, et al. Analysis of 100,000 human cancer genomes reveals the landscape of tumor mutational burden. Genome Med. 2017;9(1):34. 9. Hellmann MD, Ciuleanu TE, Pluzanski A, Lee JS, Otterson GA, Audigier-Valette C, et al. Nivolumab plus Ipilimumab in Lung Cancer with a High Tumor Mutational Burden. N Engl J Med. 2018;378(22):2093-104. 10. Goetz MP, Kalari KR, Suman VJ, Moyer AM, Yu J, Visscher DW, et al. Tumor Sequencing and Patient-Derived Xenografts in the Neoadjuvant Treatment of Breast Cancer. J Natl Cancer Inst. 2017;109(7). 11. Horak CE, Pusztai L, Xing G, Trifan OC, Saura C, Tseng LM, et al. Biomarker analysis of neoadjuvant doxorubicin/cyclophosphamide followed by ixabepilone or Paclitaxel in early-stage breast cancer. Clin Cancer Res. 2013;19(6):1587-95. 12. de Ronde JJ, Lips EH, Mulder L, Vincent AD, Wesseling J, Nieuwland M, et al. SERPINA6, BEX1, AGTR1, SLC26A3, and LAPTM4B are markers of resistance to neoadjuvant chemotherapy in HER2-negative breast cancer. Breast Cancer Res Treat. 2013;137(1):213-23. 13. Santonja A, Sanchez-Munoz A, Lluch A, Chica-Parrado MR, Albanell J, Chacon JI, et al. Triple negative breast cancer subtypes and pathologic complete response rate to neoadjuvant chemotherapy. Oncotarget. 2018;9(41):26406-16. 14. Stickeler E, Pils D, Klar M, Orlowsk-Volk M, Zur Hausen A, Jager M, et al. Basal-like molecular subtype and HER4 up-regulation and response to neoadjuvant chemotherapy in breast cancer. Oncol Rep. 2011;26(4):1037-45. 15. Hollern DP, Xu N, Thennavan A, Glodowski C, Garcia-Recio S, Mott KR, et al. B Cells and T Follicular Helper Cells Mediate Response to Checkpoint Inhibitors in High Mutation Burden Mouse Models of Breast Cancer. Cell. 2019;179(5):1191-206 e21. 16. Araki H, Pang X, Komatsu N, Soejima M, Miyata N, Takaki M, et al. Haptoglobin promoter polymorphism rs5472 as a prognostic biomarker for peptide vaccine efficacy in castration- resistant prostate cancer patients. Cancer Immunol Immunother. 2015;64(12):1565-73. 17. Kim YJ, Ha YS, Kim SK, Yoon HY, Lym MS, Kim MJ, et al. Gene signatures for the prediction of response to Bacillus Calmette-Guerin immunotherapy in primary pT1 bladder cancers. Clin Cancer Res. 2010;16(7):2131-7. 18. McConkey DJ, Choi W, Shen Y, Lee IL, Porten S, Matin SF, et al. A Prognostic Gene Expression Signature in the Molecular Classification of Chemotherapy-naive Urothelial Cancer is Predictive of Clinical Outcomes from Neoadjuvant Chemotherapy: A Phase 2 Trial of Dose- dense Methotrexate, Vinblastine, Doxorubicin, and Cisplatin with Bevacizumab in Urothelial Cancer. Eur Urol. 2016;69(5):855-62. 19. Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer. 2009;45(2):228-47.

Supplementary Figure 1 TCGA database A Melanoma pretreated with anti-PD1 or anti- B 100 CTLA-4 (Liu et al. Nature Medicine 2019) 3 P<0.0001

80 2

60 1

40 P = 0.0286 0 Overallsurvival (%)

IFN activity low (median survival: 15.06 months) IFN score 20 IFN activity high (median survival: 33.5 months) -1 0 0 20 Months 40 60 -2

Spearman rank of top genes C negatively coefficient with IFN activity -3 Prostate cancer (TCGA) Breast cancer (TCGA) Spearman’s Spearman’s Genes P-value Genes P-value rho rho ACACA -0.56348 1.74E-42 TOM1L2 -0.38571 1.31E-36 FOXA1 -0.53136 3.99E-37 SUMF1 -0.37199 5.59E-34 Melanoma C9ORF152 -0.53029 5.88E-37 FAM172A -0.36778 3.38E-33 Lung cancer RAB3IP -0.52284 8.48E-36 XBP1 -0.35779 2.19E-31 Kidney cancer SLC35E1 -0.52243 9.82E-36 FOXA1 -0.35233 2.00E-30 ProstateLimited cancer response Better response LEO1 -0.51939 2.85E-35 CASZ1 -0.3465 2.03E-29 to ICIs to ICIs MAML3 -0.51089 5.35E-34 EIF4B -0.34103 1.71E-28 Breast cancer (Lumi) ERGIC1 -0.50862 1.16E-33 KDM4B -0.33776 6.01E-28 GALNT7 -0.50333 6.78E-33 FAM47E -0.33671 8.95E-28 PPM1H -0.50109 1.42E-32 MTURN -0.33189 5.50E-27 D ACLY -0.4919 2.80E-31 THSD4 -0.33164 6.04E-27 Single RNA-seq from 13 prostate cancer patient tissues SHANK2 -0.49052 4.34E-31 NEDD4L -0.33157 6.20E-27 PPP1R9A -0.48824 8.92E-31 KIAA1324 -0.33145 6.49E-27 TTC3 -0.48706 1.29E-30 BCAM -0.33137 6.66E-27 CMTM4 -0.48546 2.14E-30 KIAA0556 -0.3294 1.39E-26 TMEM268 -0.48425 3.12E-30 PLA2G12A -0.32827 2.10E-26 KCTD3 -0.48253 5.32E-30 EEF2 -0.32789 2.41E-26 LRIG1 -0.47984 1.22E-29 CYB5D2 -0.32748 2.80E-26 25 DOP1B -0.47958 1.32E-29 ZNF396 -0.32673 3.69E-26 BEND4 -0.47624 3.65E-29 HCAR1 -0.32657 3.92E-26 SCARB2 -0.47571 4.29E-29 ERBB3 -0.32586 5.08E-26 TMEM192 -0.47546 4.63E-29 USP22 -0.32474 7.63E-26 ZNF24 -0.47532 4.82E-29 RUNDC1 -0.32359 1.16E-25 GSE1 -0.47468 5.86E-29 MLPH -0.32198 2.07E-25 TOM1L1 -0.47382 7.59E-29 KIF13B -0.32186 2.16E-25

GFPT1 -0.47254 1.12E-28 ZBTB42 -0.32076 3.20E-25 tSNE2 0 HIPK2 -0.47152 1.51E-28 ALDH6A1 -0.31868 6.72E-25 Luminal epithelial cells REPS2 -0.47125 1.64E-28 DRC3 -0.31866 6.77E-25 MBOAT2 -0.47096 1.79E-28 UNC119B -0.31678 1.32E-24 Basal/intermediate cells DSC2 -0.46895 3.25E-28 ARSD -0.31668 1.36E-24 IQGAP2 -0.46689 5.96E-28 SPRED2 -0.31587 1.81E-24 Endothelial cells GOLPH3 -0.46675 6.22E-28 ERBB4 -0.31474 2.70E-24 UAP1 -0.46672 6.28E-28 LONRF2 -0.31446 2.97E-24 Fibroblasts TAB3 -0.46669 6.33E-28 LMX1B -0.3141 3.37E-24 -25 GNPNAT1 -0.46557 8.80E-28 GLUD1 -0.31332 4.42E-24 Mast cells MMUT -0.46547 9.06E-28 IFT140 -0.31273 5.42E-24 FNBP1L -0.46511 1.01E-27 REEP6 -0.31113 9.45E-24 Myeloid cells ZBED4 -0.46376 1.49E-27 RERE -0.30985 1.47E-23 LARS -0.4633 1.70E-27 MEGF9 -0.30974 1.53E-23 T cells ZNF542P -0.46278 1.98E-27 PCDH1 -0.30938 1.72E-23 SLC46A1 -0.46071 3.61E-27 MOAP1 -0.30871 2.17E-23 SMARCC1 -0.46065 3.67E-27 ARSG -0.30825 2.53E-23 -20 0 20 40 TMTC4 -0.45951 5.09E-27 MAPT -0.308 2.76E-23 PLA2G12A -0.45926 5.47E-27 POLI -0.30773 3.03E-23 tSNE1 VPS54 -0.45911 5.72E-27 CACNA1D -0.30764 3.12E-23 STEAP2 -0.45624 1.29E-26 ATP8B1 -0.30749 3.29E-23 CANT1 -0.45598 1.40E-26 SHROOM3 -0.30686 4.07E-23 PHKB -0.45593 1.41E-26 THRB -0.30637 4.80E-23 NWD1 -0.45575 1.49E-26 IGF1R -0.30451 8.98E-23

Supplementary Figure 1. Inverse correlation between FOXA1 mRNA expression and IFN signaling and APM genes in prostate and breast cancer patient samples. (Related to Figure 1) A. Overall survival of patients with IFN activity-low or -high in the cohort of melanoma pretreated with anti-PD1 or anti-CTLA-4 (Liu et al. Nature Medicine 2019). Statistical significance was determined by Log-rank (Mantel-Cox) test (n = 117). (See details in Methods for the calculation of IFN gene expression score.) B. Comparison of IFN score in prostate cancer and luminal breast cancer with melanoma, lung and kidney cancers from TCGA database. Data shown as means ± SD; Statistical significance was determined by one-way analysis of variance (ANOVA). C. Spearman rank of top genes negatively correlated to IFN score in the TCGA cohorts of prostate and breast cancers. Statistical significance was determined by Spearman's rho correlation test. D. T-SNE projections of ten stages from single cell RNA sequencing data. T-SNE plot was performed for the entire dataset (n = 53,090 cells) and cells are colored and labelled by clusters. The luminal epithelial cells (n = 23,674) were used for the heatmap shown in Figure 1B. Clinic information of patients is shown in Supplementary Table 1 and Supplementary Table 2. Supplementary Figure 2

A Bulk RNA-seq of primary PCa Bulk RNA-seq of metastatic Bulk RNA-seq of bone metastatic B Bulk RNA-seq of BCa Bulk RNA-seq of BCa (TCGA, n = 491) PCa (SU2C, n = 270 ) PCa (Mayo Clinic cohort, n = 54) (TCGA, n = 960) (METABRIC, n = 1904) FOXA1 FOXA1 FOXA1 FOXA1 FOXA1 CXCL1 SERPING1 CXCL2 ACACB ACACB BST2 GBP1 SERPING1 BIRC3 CXCL1 LGALS9 IFI16 BST2 CXCL2 ACACB HERC5 CXCL2 DDX60 RASGRP3 IFITM1 BIRC3 DHX58 SERPING1 SLC15A3 HERC5 BIRC3 ISG20 GBP1 IFI27 BIRC3 LGALS9 IFI16 RASGRP3 HERC5 ISG20 SLC15A3 MX1 IFI44 IFI16 RASGRP3 SERPING1 XAF1 SLC15A3 IFI27 IFI44 GBP1 GBP1 IFI44L SLC15A3 IFI16 LGALS9 CXCL1 IFI44 LGALS9 BST2 IFI44 IFI44L IFIH1 IFI44L ISG20 ISG20 XAF1 IFIT3 XAF1 DHX58 CXCL2 IFITM1 IFIH1

IFIH1 - 6 OAS1

BIRC3 e IRF7 MX1

6 SP110 MX1 . DHX58 ISG15 PARP12 PARP12 = 1.5e−1= 2 < 2.2e−16< = 2.4e−4= = 1 IFI44L IFI27 ISG20 IFI27 <2.2e−16 P P

P XAF1 P IFIT3 ISG15 P ISG15 LGALS9 STAT1 ISG15 STAT1 MX1 OAS2 RASGRP3 IFIT3 IFIT3 OAS1 DDX60 SP110 OAS2 OAS2 Interferon (IFN) response Interferon (IFN) response signature OAS2 IRF7 IFITM1 IFITM1 IFI27 PARP12 SAMD9 OAS1 IFIH1 OAS1 RASGRP3 PARP12 IRF7 ISG15 OAS1 IRF7 SAMD9 BST2 MX1 HERC5 BST2 SERPING1 SAMD9 IFIT3 OAS2 DDX60 SLC15A3 SP110 IFI44 STAT1 SP110 DDX60 IFI44L DHX58 DDX60 STAT1 DHX58 IFITM1 IFIH1 ACACB XAF1 ACACB IRF7 HLA−A HLA−A B2M HLA-A HLA−J HLA−B PSMB9 HLA−B TAPBP HLA−C PSMB8 HLA-H HLA−C HLA−G HLA−E HLA-F HLA−E HLA−E HLA−E HLA−F HLA-E HLA−F B2M HLA−G HLA-B HLA−G HLA−A TAP1 HLA−H PSMB8 HLA−H TAP2 HLA−J HLA-C HLA−J

< 2.2e−16< HLA−C < 2.2e−16< 8.2e−12=

HLA−C 3.2e−4= B2M B2M B2M <2.2e−16 P PSMB9 P P P P PSMB9 HLA−B HLA-G PSMB9 PSMB8 machinery (APM) machinery PSMB8 TAP1 PSMB8 Antigen presentation HLA−H TAP1 TAP1 TAP2 TAP1 HLA−F TAP2 HLA-J TAP2 PSMB9 HLA−A TAPBP TAPBP TAPBP HLA−B -2 -2 0 2 -2 0 2 -2 0 2 0 2 -2 0 2

C Prostate cancer Prostate cancer Bone metastatic prostate cancer Breast cancer Breast Cancer (TCGA, n = 491) (SU2C, n = 270) (Mayo Clinic cohort, n = 54) (TCGA, n = 960) (METABRIC, n = 1904) r = − 0.40, P < 2.2e−16 r = − 0.33, P = 1.6e-6 r = − 0.40, P = 2.4e−4 r = − 0.30, P = 1.5e−12 r = −0.31, P < 2.2e−16 1.00 1.00 1.00 1.00 1.00 0.80 0.75 0.75 0.75 0.75 0.60 0.50 0.50 0.50 0.50 0.40 0.25 0.25 0.25

IFN score IFN 0.25 0.20 0.00 0.00 0.00 0.00 0.00 10 14 15 0.0 2.5 5.0 7.5 10 11 12 13 0 50 100 150 200 5 15 -4 -2 0 FOXA1, log2(RSEM read count+1) FOXA1, log2(FPKM+1) FOXA1 (FPKM) FOXA1, log2(RSEM read count+1) FOXA1, mRNA Expression z−Score

Breast cancer Prostate cancer Breast Cancer Prostate cancer (TCGA, n = 960) D Prostate cancer (Mayo Clinic cohort, n = 54) (METABRIC, n = 1904) (SU2C, n = 270) r = − 0.32, P < 2.2e−16 (TCGA, n = 491) r = − 0.47, P = 3.2e−4 r = −0.35, P < 2.2e−16 r = − 0.40, P = 8.2e−12 1.00 r = − 0.44, P < 2.2e−16 1.00 1.00 1.00 1.00 0.80 0.75 0.75 0.75 0.75 0.60 0.50 0.50 0.50 0.50 0.40

APM score APM 0.25 0.25 0.25 0.20 0.25 0.00 0.00 0.00 10 11 12 13 14 15 0.0 2.5 5.0 7.5 0.00 0.00 0 50 100 150 200 10 15 FOXA1, log2(RSEM read count+1) FOXA1, log2(FPKM+1) 5 -4 -2 0 FOXA1 (FPKM) FOXA1, log2(RSEM read count+1) FOXA1, mRNA Expression z−Score

Supplementary Figure 2. Inverse correlation between FOXA1 mRNA expression and IFN signaling and APM genes in prostate and breast cancer patient samples. (Related to Figure 1) A. Heatmaps show the inverse correlation of FOXA1 expression with IFN response signature genes and APM genes in bulk tissue RNA-seq data of PCa samples. Statistical significance was determined by Pearson correlation test. B. Heatmaps show the inverse correlation of FOXA1 expression with IFN response signature genes and APM genes in bulk tissue RNA-seq data of BCa samples. Statistical significance was determined by Pearson correlation test. C. The correlation of FOXA1 level with IFN score (see details in Methods for the calculation of gene expression score) in TCGA, SU2C, Mayo Clinic cohorts of prostate cancer and TCGA and METABRIC cohorts of breast cancer. Statistical significance was determined by Pearson correlation test. D. The correlation of FOXA1 level with APM score (see details in Methods for the calculation of gene expression score) in TCGA, SU2C, Mayo Clinic cohorts of prostate cancer and TCGA and METABRIC cohorts of breast cancer. Statistical significance was determined by Pearson correlation test. Supplementary Figure 3

Prostate cancer Breast cancer A (TCGA data, n = 491) (TCGA, n = 960) B FOXA1 mRNA IFNE 0.8% 2.6% Prostate cancer IFNA1 1% 2.5% IFNA2 0.8% 2.9% (TCGA data, n = 491) IFNA8 0.8% 2.6% r = 0.03, P = 0.51 IFNA13 1% 2.8% 3.0 IFNA6 0.8% 2.7% 2.5% IFNA5 1% 2.6% 2.5 IFNA14 0.8% 2.6% IFNA17 1% 2.6% 0.6% 2.0 IFNA10 2.5% 1% IFNA16 2.5% IFNA4 0.6% 2.3% 1.5 IFNA7 0.6% 2.6% IFNA21 0.6% 2.1% 1.0 IFNW1 0.6% 2.1% (Mutations (Mutations MB) per IFNB1 0.6% 1.9% Tumor mutation burden IRF2 1.4% 2.2% 0.5 JAK2 1.6% 1.2% IL31RA 5% 2.1% 0.0 ISG15 0.4% 1.4% PRKCD 0.4% 1% -5.0 -2.5 0.0 2.5 5.0 IFITM1 0.2% 1.2% IFITM2 0.2% 1.3% FOXA1 mRNA Expression IFRD2 0.4% 0.6% (log RNA Seq V2 RSEM) IFIT2 5% 1.2% IFIT1 5% 1.3% IFIT1B 4% 1.3% Breast cancer 0.9% IFIT3 5% (TCGA data, n = 960) IRF1 1% 1% IRF4 1.4% 0.7% r = − 0.0098, P = 0.78 IL10RA 1% 2.2% 20 0.9% PIAS4 0.6% 1.5% EBI3 0.8% 0.8% IRF8 5% 2.7% 15 2% IFI27 0.7% 0.4% IFI6 0.4% IL5RA 2% 2.5% IL2RG 0.4% 0.4% 10 MX1 10% 2.1% JAK1 2.7% 2.5% IL12B 1% 1.1% SP110 0.6% 1.2% 5 (Mutations (Mutations MB) per IRGM 0.4% 0.8% Tumor mutation burden IFNGR1 0.4% 1.8% 0.6% IL20RA 1.5% 0 IL22RA2 0.6% 1.8% CREB1 0.4% 1.3% -6 -4 -2 0 2 IFIH1 2.7% 0.5% IL13RA1 0.4% 0.9% FOXA1 mRNA Expression IRF9 0.4% 1.4% (log RNA Seq V2 RSEM) LEPR 1.6% 2.2% PSME1 0.6% 1% IFI35 1.6% 1.2% IFI44 1.6% 1.8% IFI44L 1.8% 1.2% IRF6 0.8% 8% 2.3% TYK2 0.6% 9% ADAR 2.7% 1.2% CNTFR 0.6% 0.8% Genomicalterations ofsignaling IFN pathway genes CREB3 1.2% 1.7% 0.4% CSF3R 1.8% CXCL10 0.4% 1.1% F3 0.6% 9% IFI16 2.9% 1.4% IFI30 0.8% 2.5% IFNG 0.8% 1.1% IFNGR2 0.4% 1.1% IFRD1 1.2% 1.4% IL10RB 0.2% 1.4% IL11RA 0.8% 1.7% IL12A 1.8% 0.8% IL20RB 1.8% 0.2% IL21R 1.6% 4% IL2RB 0.2% 0.8% IL4R 1.8% 4% IL6 2.3% 1.1% IL6R 2.3% 9% IL7R 0.8% 2.2% IRF2BP1 0.6% 1.3% IRF2BP2 1.6% 11% IRF3 0.6% 2.1% 1.3% IRF5 1.2% 1.4% OAS1 0.8% 0.1% PIAS1 0.6% 0.8% PIK3CA 5% 37% 2.3% PYHIN1 9% SOCS1 1% 4% SOCS3 0.4% 4% STAT1 0.2% 1.2% STAT2 0.6% 0.8% STAT3 2.9% 2.3% STAT4 1% 2.1% TBX21 1% 2.9%

Inframe Mutation Missense Mutation Truncating Mutation

Amplification Deep Deletion No alterations

Supplementary Figure 3. FOXA1 expression is not associated with genomic alterations in IFN signaling pathway genes and tumor mutation burden (TMB). (Related to Figure 1) A. Heatmaps show that there was no obvious correlation between FOXA1 expression and the genomic alterations in the IFN signaling pathway genes in the TCGA cohorts of prostate and breast cancer. Samples are ranked based on FOXA1 transcript levels. B. The correlation between FOXA1 level and TMB in TCGA cohorts of prostate cancer and breast cancer. Statistical significance was determined by Pearson correlation test. Supplementary Figure 4

Prostate cancer (TCGA) Breast cancer (TCGA) A B 5 100 2 3 P = 0.0001 P = 0.0005 P = 0.0026 P = 0.0016

2

50 0

0 1

0 -2 0 mRNA (Z score) mRNA (Z score) IFN score IFN IFN score IFN

-5 -1 FOXA1 FOXA1 -50 -4

-2

-10 -100 -6 -3

C Prostate cancer Prostate cancer (Singh Prostate Data, Breast cancer Breast cancer (Curtis Breast Data, (TCGA data) Singh et. al. Cancer Cell, 2002) (TCGA Data) Curtis et. al. Nature, 2012) 4.5 P < 0.0001 600 P < 0.0001 4.0 600 P < 0.0001 P < 0.0001 3.5 500 7.0 500 3.0 2.5 6.5 6.0 2.0 400 400 5.5 1.5 5.0 1.0 300 4.5 centered intensity)centered

300 levelmRNA mRNA levelmRNA 0.5 4.0 mRNA levelmRNA 0.0 200 3.5 centered intensity)centered 200 -0.5 levelmRNA 3.0

FOXA1 2.5 FOXA1

-1.0 (Transcript per million) 100

FOXA1 2.0

(Transcript per million) 100 -1.5 1.5 -2.0 0 FOXA1

- (Log2 median 1.0 0 -2.5 0.5 -3.0 -100 0.0 -100 - (Log2 median -0.5

-3.5

D Prostate cancer (TCGA data) Breast cancer (TCGA Data)

3 6 P < 0.0001 P = 0.049 2.5 5

2 4

1.5 3 mRNA levelmRNA mRNA levelmRNA 1 2

0.5 1 CD274 CD274 CD274 CD274 (Transcript per million) (Transcript per million) 0 0

-0.5 -1

Supplementary Figure 4. FOXA1 mRNA is overexpressed and CD274 mRNA is downregulated in prostate and breast cancer patients. (Related to Figure 1) A. Relative FOXA1 mRNA level and IFN score in the cases of FOXA1 amplification and non-amplification in prostate cancer from cohorts of TCGA. Data shown as means ± SD; Statistical comparison was done using Mann-Whitney U test. B. Relative FOXA1 mRNA level and IFN score in the cases of FOXA1 amplification and non-amplification in breast cancer from cohorts of TCGA. Data shown as means ± SD; Statistical comparison was done using Mann-Whitney U test. C. Comparison of FOXA1 mRNA level between normal tissues and tumor tissues in the indicated cohorts of prostate or breast cancer. Data shown as means ± SD; Statistical comparison was done using Mann-Whitney U test. D. Comparison of CD274 (PD-L1) mRNA level between cancer specimens and normal tissues in the TCGA cohorts of prostate or breast cancers. Data shown as means ± SD; Statistical comparison was done using Mann-Whitney U test. Supplementary Figure 5

C A B 293T DU145 (PCa) MDA-MB-231 (BCa) IFNα (50 ng/L) - + + + + + ------IFNα (50 ng/L) - - + + - IFNγ (50 ng/L) ------+ + + + + IFNα (50 ng/L) - + +

231 Lenti-Vector + - + - - - Lenti-Vector + - + SFB-FOXA1 - - - - Lenti-FOXA1 - + - + Lenti-FOXA1 - + - + MB - SFB-FOXA1 FOXA1 FOXA1 RT4 LNCaP 293T VCaP DU145 MCF7 MDA P-STAT1 P-STAT1 P-STAT1 FOXA1 STAT1 STAT1 STAT1 P-STAT2 P-STAT2 ERK2 P-STAT2 STAT2 STAT2 STAT2 ERK2 ERK2 ERK2

LNCaP (PCa) MCF7 (BCa) 1 139 316 480 575 851 D Input IP: Flag Input IP: Flag F STAT2 NTD CCD DBD LD CTD E STAT2 (DBD-CTD) DBD LD CTD IFNα (50 ng/L) - - + + IFNα (50 ng/L) - - + + V5-FOXA1 V5-FOXA1 + + + + STAT2 (LD-CTD) LD CTD siControl - + - siControl + - + - GFP-IRF9 + + + + + + HA-STAT1 + + + + + + + Input siFOXA1 - + - + Flag-STAT2 + + + + + + Flag-STAT1 + + IP: HA siFOXA1 - + - + IFNα + + + + + + IFNγ + + + + + + FOXA1 FOXA1 CTD) Flag-STAT2 Flag-STAT1 CTD) - - CTD) CTD) - P-STAT1 P-STAT1 - STAT1 HA-STAT1 STAT1 STAT1 STAT2 P-STAT2 P-STAT2 GFP-IRF9 STAT2 (DBD STAT2 STAT2 (LD STAT2 (DBD STAT2 STAT2 V5-FOXA1 STAT2 (LD - - - - - STAT2 - V5-FOXA1 HA HA HA HA HA ERK2 ERK2 HA V5-FOXA1 + + + + + + KDa V5-FOXA1 50

100 IB: HA 50 294 472 269 1 168 141 37 G H FOXA1 binding region I FOXA1: FKHD 486)

- STAT2-DBD 312 486 FOXA1-N (1-140): STAT2-DBD 312 426 FOXA1-M (141-294): FKHD STAT2-DBD 312 366 FOXA1-C (295-472): J GST Input IP: Flag DBD (312DBD GST Pulldown Input - IB: FOXA1 pulldown N M C N M C ------SFB-FOXA1-M(141-294) + - + - IB: FOXA1 Input SFB-FOXA1-M(141-247) - + - +

STAT2 + + - IFNα + + 50 KDa FOXA1 FOXA1 FOXA1 FOXA1 FOXA1 FOXA1 37 KDa Coomassie P-STAT2 GST GST 25 KDa blue staining SFB - SFB - SFB - SFB - SFB - SFB - GST + + + STAT2 IB: FOXA1 pulldown GST-STAT2-DBD: IB: Flag P-STAT1 IB: FOXA1 Input

50 KDa 50 KDa STAT1 37 KDa 37 KDa ERK2 25 KDa 20 KDa Coomassie blue staining Flag Coomassie blue staining of GST-STAT2-DBD

K ISRE-luc reporter M FOXA1-luc reporter

150 luc N

luc P < 0.0001 15 - Chromatin-free nuclear extract O P < 0.0001 H3 α 100 10 Incubation with SBR R261G WT △ △ - Biotin-ISRE (DNA) + Flag-FOXA1 (protein) - 50 P = 0.2056 5 AR Vector FOXA1 Vector FOXA1 luciferase activity FOXA1 FOXA1 luciferase activity Relative ISRE Biotin pulldown with streptavidin beads 0 - Relative FOXA1 0 Biotin-ISRE probe: + + + + + + + IFNα (100 ng/L) - + + + + + FOXA1-luc (100 ng) + + + + + IFNα: - + + + + + + ISRE-luc (10 ng) + + + + + + pRL-TK (10 ng) + + + + + Unbound supernatant pRL-TK (10 ng) + + + + + + Flag IP Shift ISRE probe

EMSA WT αH3 SBR Flag-FOXA1: - - IB: Biotin IFNα: - + + + + +△+ + +△ + + Biotin-ISRE (DNA): + + + + + + + + + + + Free ISRE probe

STAT2 L P-STAT2 Biotin-FRE probe: + + + + + FOXA1 Biotin pulldown STAT2 STAT2 Shift probe P-STAT1

EMSA FOXA1 FlagIP: STAT1 IB: Biotin Free probe FOXA1 AR Whole Whole cell lysis FOXA1 STAT2 Input

IB: Flag HDAC2 ERK2 Supplementary Figure 5. Effects of FOXA1 on expression of total and phosphorylated STAT1 and STAT2 proteins and STAT2 activity. (Related to Figure 2) A. Western blot analysis of expression of FOXA1 protein in the indicated cell lines. ERK2 was used as a loading control. B. Western blot analysis of the indicated proteins in 293T cells transfected with SFB-tagged FOXA1 in different doses in the presence or absence of IFNα or IFNγ. C. Western blot analysis of the indicated proteins in FOXA1-negative DU145 PCa and MDA-MB-231 BCa cells infected with lentivirus expressing vector or FOXA1 and treated with or without IFNα. D. Western blot analysis of the indicated proteins in FOXA1-high LNCaP PCa or MCF7 BCa cells transfected with control or FOXA1-specific siRNAs and treated with or without IFNα. E. Co-IP analysis of the effect of FOXA1 on the formation of STAT1-STAT2-IRF9 and STAT1-STAT1 complexes in 293T cells. F. Co-IP analysis of interaction of FOXA1 with STAT2 full length or its N-terminal truncation mutants in 293T cells. NTD, N- terminal domain; CCD, coiled-coil domain; DBD, DNA binding domain; LD, linker domain; CTD, C-terminal domain. G. GST pulldown examining the interaction of FOXA1 with STAT2 DBD in vitro. STAT2 DBD protein was purified from E.coli and FOXA1 protein was produced by in vitro transcription/translation. H. GST pulldown analysis of the interaction of GST-tagged STAT2 DBD and its C-terminal truncation mutants with FOXA1 protein produced by in vitro transcription/translation. I. GST pulldown analysis of the interaction of GST-tagged STAT2 DBD with different fragments of FOXA1 protein produced by in vitro transcription/translation. FKHD, forkhead domain; N, N-terminal; M, middle portion; C, C-terminal. J. Co-IP analysis of the interaction of SFB-tagged FOXA1 truncation mutants FOXA1(141-294) and FOXA1(141-247) with STAT1 and STAT2 in 293T cells treated with IFNα. K. Luciferase reporter assay performed to assess the effect of FOXA1 WT and truncation mutants on IFN stimulation response element luciferase reporter (ISRE-luc) activity in 293T cells treated with IFNα. Data shown as means ± SD (n = 3); Statistical significance was determined by one-way ANOVA with Bonferroni correction for multiple tests. L. EMSA assay performed to analyze the binding of FOXA1 WT and indicated mutants to the FOXA1 response element (FRE) present in the KLK3 (PSA) gene enhancer. M. Luciferase reporter assay performed to determine the effect of FOXA1 WT and indicated mutants on the activity of luciferase reporter containing FOXA1 response element (FRE) derived from the KLK3 gene enhancer in 293T cells. Data shown as means ± SD (n = 3). N. Biotin-labeled DNA pulldown assay followed by co-IP assay (see Supplementary Methods). ISRE, interferon stimulation response element. O. EMSA assay. Biotin-labeled ISRE was incubated with nuclear extract of vehicle- or IFNα-treated DU145 cells expressing FOXA1 WT, FOAX1 mutants or AR prior to PAGE and/or Western blots. Supplementary Figure 6 A B Correlation of the replicates of STAT2 ChIP-seq data 8 8 8

Rep1 8

r=0.943 P=0.000 r=0.918 P=0.000 - r=0.973 P=0.000

r=0.929 P= 0.000 Rep1 α Rep1 - 6 - 6 6 6 α Rep1 siCon + - + ------siFOXA1 - + - + + + + + 4 4 4 4 IFNα - - + + + + + +

siCon 2 2 2 2 siFOXA1

FOXA1 0 0 siCon+IFN 0 0

0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 siFOXA1+IFN 0 2 4 6 8 siCon-Rep2 siFOXA1-Rep2 siCon+IFNα-Rep2 siFOXA1+IFNα-Rep2 P-STAT2 WT ∆H3 -

STAT2 R261G 8 8 r=0.959 P=0.000 1 r=0.948 P=0.000 8 r=0.953 P=0.000 8 r=0.943 P=0.000 ∆SBR 6 6 Rep1 Rep1 Rep1 6 - - P-STAT1 - 6 α α α Rep1

4 4 - 4

4 α 2 +IFN +IFN STAT1 +IFN 2 2 2 0 0 0 +IFN 0 siFOXA1+FOXA1 siFOXA1+FOXA1 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 siFOXA1+FOXA1-WT siFOXA1+FOXA - siFOXA1+FOXA1∆H3 siFOXA1+FOXA1∆SBR ERK2 siFOXA1+FOXA1-R261G siFOXA1+FOXA1 +IFNα-Rep2 +IFNα-Rep2 +IFNα-Rep2 +IFNα-Rep2 LNCaP (PCa) C NSA-peaks IFNα - - + + + + + + siCon + siFOXA1 + siCon + siFOXA1 + siFOXA1 + siFOXA1 + siFOXA1 + siFOXA1 + Vector Vector Vector Vector FOXA1-WT FOXA1-R261G FOXA1 αH3 FOXA1 SBR R1 R2 R1 R2 R1 R2 R1 R2 R1 R2 R1 R2 R1 R2 R1 R2 5 △ △ 4 3 2 1 Binding intensity

12 10 8 6 4 2

Binding signal 0 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 D Down-peaks IFNα - - + + + + + + siCon + siFOXA1 + siCon + siFOXA1 + siFOXA1 + siFOXA1 + siFOXA1 + siFOXA1 + Vector Vector Vector Vector FOXA1-WT FOXA1-R261G FOXA1 αH3 FOXA1 SBR R1 R2 R1 R2 R1 R2 R1 R2 R1 R2 R1 R2 R1 R2 R1 R2 20 △ △ 15 10 5 2.5 Binding intensity

20 30 25 15 Binding signal 10 5 0 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5

E Up-peaks NSA-peaks Down-peaks

30% 30% 30%

25% 25% 25%

20% 20% 20%

15% 15% 15% gene associations 10% 10% 10%

Region - 5% 5% 5%

0% 0% 0%

Distance to TSS (kb) Distance to TSS (kb) Distance to TSS (kb)

F GO-BP of Up-peak- GO-BP of NSA-peak- GO-BP of Down-peak- associated genes associated genes associated genes 60 60 value) value) P

P 40 No overlaps of

( 40 (

10 pathways were found 20 10 20 Log Log - 0 - 0 1 2 3 4 5 1 2 3 4 5 Supplementary Figure 6. Effect of FOXA1 expression on STAT2 occupancy on chromatin in LNCaP cells. (Related to Figure 3) A. Western blot analysis of indicated proteins in indicated LNCaP cells. LNCaP cells stably expressing vector, FOXA1 WT or mutants were transfected with control (siCon) or FOXA1-specific siRNA (siFOXA1) followed by treatment with vehicle or IFNα. Independent sets of cells were also used for STAT2 ChIP-seq. B. Analysis of correlation between replicates of STAT2 ChIP-seq data in each indicated cellular condition. C. The STAT2 binding intensity (top) and binding signal heatmap (bottom) of NSA-peaks (identified by STAT2 ChIP-seq as shown in Figure 3A) in the indicated cellular conditions. D. The STAT2 binding intensity (top) and binding signal heatmap (bottom) of Down-peaks (identified by STAT2 ChIP-seq as shown in Figure 3A) in the indicated cellular conditions. E. Distributions of STAT2 ChIP-seq peaks (Up, NSA or Down) in the genome relative to the transcription start site (TSS). F. GO-BP pathways analysis of genes associated with STAT2 ChIP-seq Up-peaks, NSA-peaks and Down-peaks. Supplementary Figure 7

chr1:1012448 chr1:78648449 chr6:30488299 chr6:32853063 chr8:23679937 chr14:75982261 A -1014796 (hg38) -78651650 (hg38) -30491000 (hg38) -32855013 (hg38) -23682982 (hg38) -75985693 (hg38) IFNα 200 20 30 35 20 90 + - Vector siCon

200 20 30 35 20 90 - Vector siFOXA1 +

200 20 30 35 20 90 + + Vector siCon

200 20 30 35 20 90

+ Vector siFOXA1 + seq -

200 20 30 35 20 90 ChIP - WT + siFOXA1 + FOXA1 STAT2

200 20 30 35 20 90

+ -- R261G siFOXA1 + FOXA1 200 20 30 35 20 90 H3 α △ + siFOXA1 + FOXA1 200 20 30 35 20 90 SBR △ + siFOXA1 + FOXA1 200 20 30 35 20 90 IgG

+ Vector siFOXA1 +

200 20 30 35 20 90 + - seq - Vector siCon 90 ChIP 200 20 30 35 20 + + Vector FOXA1 siCon

ISG15 IFI44 HLA-E PSMB9 NKX3.1 TGFB3 STAT2 target genes FOXA1 target genes

B MCF7 siCon ISG15 IFI44 HLA-E PSMB9 + - + - 0.12 0.20 0.020 0.015 siFOXA1 - + - + IgG IgG IgG IgG P = 0.0135 P = 0.0272 P = 0.0325 P = 0.0227 IFNα - - + + 0.10 Anti-STAT2 Anti-STAT2 Anti-STAT2 0.15 Anti-STAT2 0.015 FOXA1 0.08 0.010 P-STAT2 0.06 0.10 0.010 ns % of Input of % % of Input of %

% of Input of % 0.005 STAT2 0.04 Input of % ns 0.05 0.005 0.02 ns P-STAT1 ns 0.00 0.00 0.000 0.000 STAT1 IFNα 1- 2- +3 +4 IFNα 1- 2- +3 +4 IFNα 1- 2- +3 +4 IFNα 1- 2- +3 +4 siCon + - + - siCon + - + - siCon + - + - siCon + - + - ERK2 siFOXA1 - + - + siFOXA1 - + - + siFOXA1 - + - + siFOXA1 - + - +

Supplementary Figure 7. Effect of FOXA1 expression on STAT2 occupancy on chromatin in LNCaP cells. (Related to Figure 3) A. UCSC tracks (integration of two replicates) show ChIP-seq signals of IgG, STAT2 and FOXA1 at genomic loci of STAT2 target genes ISG15, IFI44, HLA-E, PSMB9 and of FOXA1 target genes NKX3.1 and TGFB3. B. STAT2 ChIP-qPCR validation of STAT2 occupancy at genomic loci of Stat2 target genes ISG15, IFI44, HLA-E and PSMB9 in MCF7 cells transfected with control (siCon) or FOXA1-specific siRNAs (siFOXA1) and treated with or without IFNα. Data shown as means ± SD (n = 3). Statistical significance was determined by one-way ANOVA with Bonferroni correction for multiple tests. Supplementary Figure 8

A FOXA1 ChIP-seq in LNCaP D IFNα (-) IFNα (+) STAT2 binding peaks overlapped with FOXA1 peaks Rep.1 Rep.2 Rep.1 Rep.2 14 Total Sample ChIP-seq MEME-ChIP top hit motif STAT2 STAT2 STAT2 8 peaks Up-peaks Down-peaks NSA-peak

0.0 (n = 192) (n = 45) (n = 557)

Binding intensity Number Ratio Number Ratio Number Ratio FOXA1 motif (E-value<1e-300) LNCaP cell line FOXA1 50,838 10 5% 22 48% 128 22% No potential STAT2 motif was found FOXA1 motif (E-value=5.1e-046) Primary prostate 1.0 cancer patient tissue FOXA1 12,379 1 1% 12 26% 81 14%

Binding intensity GSM4569882 0.5 No potential STAT2 motif was found

0.0 FOXA1 motif (E-value<1e-300) -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 -1.5 0 1.5 Primary prostate cancer patient tissue FOXA1 47,396 4 2% 13 28% 64 11% GSM4569883 B No potential STAT2 motif was found FOXA1 motif (E-value=6.0e-189) Primary prostate STAT2 Up-peaks after cancer patient tissue FOXA1 40,180 1 1% 14 31% 81 14% FOXA1 ChIP-seq in LNCaP 182 FOXA1 KD in LNCaP GSM4569884 No potential STAT2 motif was found Top hit: FOXA1 motif (E-value<1e-300) 10 Top hit: STAT2 motif (E-value<1e-300) FOXA1 motif (E-value=1.1e-126) Primary prostate cancer patient tissue FOXA1 20,138 3 1% 15 33% 74 13% (No potential STAT2 GSM4569885 (No potential FOXA1 No potential STAT2 motif was found motif was found) 50,678 motif was found) FOXA1 motif (E-value=9.8e-050) Primary prostate cancer patient tissue FOXA1 14,616 1 1% 13 28% 84 15% 128 GSM4569886 No potential STAT2 motif was found 429 22 FOXA1 motif (E-value=1.2e-204) 23 Primary prostate cancer patient tissue 40,370 4 2% 13 28% 74 13% STAT2 Down-peaks after FOXA1 STAT2 NSA-peaks after GSM4569887 FOXA1 KD in LNCaP FOXA1 KD in LNCaP No potential STAT2 motif was found FOXA1 motif (E-value=1.3e-218) (No potential STAT2 and (No potential STAT2 and Primary prostate FOXA1 motif was found) FOXA1 motif was found) cancer patient tissue FOXA1 25,100 1 1% 12 26% 86 15% GSM4569888 No potential STAT2 motif was found

FOXA1 motif (E-value=1.7e-049) Primary prostate cancer patient tissue FOXA1 13,918 1 1% 14 31% 82 14% GSM4569889 No potential STAT2 motif was found FOXA1 motif (E-value<1e-300) C LNCaP Prostate cancer Primary prostate (TCGA, n = 491) 0.3 15 cancer patient tissue FOXA1 64,140 5 2% 15 33% 78 14% P < 0.0001 P < 0.0001 GSM4569890 14 No potential STAT2 motif was found FOXA1 motif (E-value<1e-300) 0.2 13 Primary prostate

PKM) cancer patient tissue FOXA1 55,558 3 1% 12 26% 65 11% R GSM4569891 12 No potential STAT2 motif was found 0.1 FOXA1 motif (E-value=1.8e-079) Log2 ( 11 Primary prostate cancer patient tissue 12,639 0 0% 15 33% 61 10% Relative mRNA Relative mRNA level FOXA1 10 GSM4569892 0.0 No potential STAT2 motif was found ZN502 motif (E-value=1.2e-087) Primary prostate

FOXA1 STAT2 cancer patient tissue FOXA1 2,050 0 0% 15 33% 80 16% GSM4569893 No potential STAT2 motif was found FOXA1 motif (E-value<1e-300) Primary prostate cancer patient tissue FOXA1 66,960 11 5% 12 26% 63 11% GSM4569893 No potential STAT2 motif was found

E IFN-independent STAT2 binding sites IFN-dependent STAT2 binding sites

Other motifs ISRE

IFN-independent STAT2 DNA binding IFN-dependent STAT2 DNA binding (Mediated by other factors) (Induced by IFN stimulation) Supplementary Figure 8. Impact of IFNα treatment on FOXA1 occupancy on chromatin and genome-wide overlap between FOXA1 binding sites with STAT2 binding sites. (Related Figure 3) A. The FOXA1 binding intensity (top) and binding signal heatmap (bottom) of FOXA1 binding sites in LNCaP cells treated with or without IFNα (100 ng/L). B. Venn diagram shows the overlap of total FOXA1 binding peaks with Up-peaks, Down-peaks and NSA-peaks of STAT2 induced by FOXA1 KD in LNCaP treated with IFNα (100 ng/L). The top hit of the DNA motif defined by MEME-ChIP analysis in each group is also shown. C. The relative expression level of FOXA1 and STAT2 mRNA in LNCaP cells (Left, normalized to GAPDH mRNA expression level) and RNA-seq reads of FOXA1 and STAT2 mRNA in primary PCa tissues of the TCGA cohort (Right). Data shown as means ± SD (n = 3, left; n = 491, right). Statistical comparison was done using Mann-Whitney U test. D. Detailed information regarding the overlap of total FOXA1 binding peaks identified in LNCaP or 13 primary prostate cancer tissues with the Up-peaks, Down-peaks and NSA-peaks of STAT2 induced by FOXA1 KD in IFNα-treated LNCaP cells. Ratio of overlap = Numbers of overlapped peaks / Total STAT2 peaks of each cluster × 100%. E. A hypothetical model deciphering FOXA1 overexpression-mediated suppression of IFN-dependent but not IFN-independent STAT2 DNA binding activity. Supplementary Figure 9

Cytoplasm Nucleus Type Ⅰ/ Ⅲ IFN Type Ⅱ IFN A B C D promoter (ISRE-luc) promoter (GAS-luc) P < 0.0001 P = 0.0049 25 P < 0.0001 15 P = 0.0115

H3 P < 0.0001 P = 0.0171 H3 - WT - H247Q - R261G - F266L α - WT - H247Q - R261G - F266L 20 P < 0.0001 P = 0.0069 α - WT △ - H247Q - R261G 10 - WT - H247Q - R261G △ 15

V5-FOXA1: Relative 10 5 FOXA1 FOXA1 FOXA1 FOXA1 FOXA1 FOXA1 FOXA1 FOXA1 5 Vector Vector FOXA1 FOXA1 FOXA1 FOXA1 Vector Vector FOXA1 FOXA1 FOXA1 FOXA1

+ + + + + + + + siCon luciferase activity 0 0 Flag-STAT2: siCon + + ------+ + ------siFOXA1 - - IFNα (50 ng/L) ------IFNα: + + + + + + + + siFOXA1 - - + + + + + + + + + + + + + + + + + + + + - + + + + + IFNα - + - + - + - + - + - + IFNα - + - + - + - + - + - + IFNγ (50 ng/L) ------+ + + + + V5-FOXA1 FOXA1 FOXA1 ISRE-luc (10 ng) + + + + + + ------GAS-luc (10 ng) ------+ + + + + + P-STAT1 P-STAT1 pRL-TK (10 ng) + + + + + + + + + + + +

IP:Flag STAT1 STAT1 Flag-STAT2 P-STAT1 P-STAT2 P-STAT2 STAT1 STAT2 STAT2 V5-FOXA1 P-STAT2 ISG15 ISG15 STAT2 Flag-STAT2 MHC-Ⅰ MHC-Ⅰ FOXA1 Input ERK2 ERK2 ERK2 PARP1 LNCaP VCaP

HSP70 FOXA1 mutants FOXA1 mutants

E Prostate cancer (TCGA) Breast cancer (TCGA) P = 0.12 P = 0.22 P = 0.89 P = 0.68 0.20 0.30 0.45 0.15 0.40 0.20 0.40 0.10 0.10 0.30 0.05 0.35

0.00 0.30 0.00 score APM APM score APM IFN score IFN IFN score IFN 0.20 −0.05 0.25 −0.10 −0.10 0.10 H SFB-tagged Mutant Mutant WT Mutant WT 268) 268) 290) WT Mutant WT 268)

(n=27) (n=461) (n=27) (n=461) (n=30) (n=30) (n=964) 290) 268) 247) (n=964) - - - (168 - (227 - (247 - (1 (1 (1 (247 - WT ------FOXA1 FOXA1 FOXA1 FOXA1 FOXA1 FOXA1 FOXA1 FOXA1 F G IFN STAT2 α

EV -

Winged Helix Domain + EV NLS 1 168NLS 247 268 290 472 FOXA1-WT FOXA1-WT + Flag IP: Flag + FOXA1-(1-290) FOXA1-(1-290) + FOXA1-(1-268) FOXA1-(1-268) + FOXA1-(1-247) FOXA1-(1-247) + FOXA1-(247-268) FOXA1-(247-268) P < 0.0001 + STAT2 FOXA1-(247-290) FOXA1-(247-290) + FOXA1-(227-268) FOXA1-(227-268) + FOXA1-(168-268) FOXA1-(168-268) Input

SBR 0 0 10 20 30 40 10 20 30 40 Flag Relative ISRE-luc luciferase activity

Supplementary Figure 9. Effects of PCa-derived FOXA1 mutants on the expression of IFN signature and APM genes. (Related to Figures 1-3) A. Co-IP-based analysis of the interaction of ectopically expressed FOXA1-WT and PCa-associated mutants (FOXA1- H247Q, FOXA1- R261G and FOXA1-F266L) with STAT2 in 293T cells. B. Western blot analysis of the effect of FOXA1 KD and restored expression of FOXA1 WT or PCa-derived mutants on IFNα-induced expression of IFN response genes in LNCaP PCa cell. ERK2, a loading control. C. Western blot analysis of the effect of FOXA1 KD and restored expression of FOXA1 WT or PCa-derived mutants on IFNα-induced expression of IFN response genes in VCaP PCa cell. ERK2, a loading control. D. Effects of FOXA1-WT and PCa-derived mutants FOXA1-H247Q, FOXA1-R261G and FOXA1-F266L on the activity of ISRE-luc and GAS-luc reporters in 293T cells treated with vehicle, IFNα or IFNγ. Data shown as means ± SD (n = 3); Statistical significance was determined by one-way ANOVA with Bonferroni correction for multiple tests. E. Analysis of the association of FOXA1 mutations with the expression of IFN response genes (IFN score) and expression of antigen presentation machinery genes (APM score) in the TCGA cohorts of prostate and breast cancer. Data shown as means ± SD; Statistical comparison was done using Mann-Whitney U test. F. Diagram shows the FOXA1 truncation expression constructs. NLS, nuclear localization signal; SBR, STAT2 binding region. G. Luciferase reporter assay shows the inhibitory effect of the indicated FOXA1 WT or truncations on ISRE-luc reporter gene activity in DU145 cells. Data shown as means ± SD (n = 3). Statistical significance was determined by one-way ANOVA with Bonferroni correction for multiple tests. H. Co-IP analysis of the interaction of SFB-tagged FOXA1 WT or truncations with STAT2 in 293T cells treated with IFNα. Supplementary Figure 10 A Correlation of the replicates of RNA-seq data

12 12 12 Rep2 12 12 r=0.996 p=0.000 r=0.998 p=0.000 r=0.994 p=0.000 - r=0.998 p=0.000 r=0.998 p=0.000 α 10 Rep2 10 10 10

10 Rep2 - -

8 α 8 8 8 8 Rep2 Rep2 - - 6 6 6 6 α 6 4 4 4 4 4 2 +IFN siCon 2 2 2 2 siFOXA1 0 0 0 0 0 siCon+IFN 0 2 4 6 8 10 12 0 2 4 6 8 10 12 0 2 4 6 8 10 12 0 2 4 6 8 10 12 siFOXA1+siSTAT2 0 2 4 6 8 10 12 siCon-Rep1 siFOXA1-Rep1 siCon+IFNα-Rep1 siFOXA1+IFN siFOXA1+IFNα-Rep1 siFOXA1+siSTAT2 +IFNα-Rep1 12 12 12 12 Rep3 12

r=0.995 p=0.000 r=0.998 p=0.000 r=0.998 p=0.000 - r=0.997 p=0.000 r=0.996 p=0.000

10 10 10 α 10 10 Rep3 Rep3 - - Rep3 Rep3 8 8 8 α 8 8 - -

6 6 6 6 α 6 4 4 4 4 4 siCon 2 2 2 2 +IFN 2 0 siFOXA1 0 0 0 0 siFOXA1+siSTAT2

0 2 4 6 8 10 12 0 2 4 6 8 10 12 siCon+IFN 0 2 4 6 8 10 12 0 2 4 6 8 10 12 0 2 4 6 8 10 12 siCon-Rep1 siFOXA1-Rep1 siCon+IFNα-Rep1 siFOXA1+IFN siFOXA1+IFNα-Rep1 siFOXA1+siSTAT2 +IFNα-Rep1 12 12 12 12 12 Rep3

r=0.998 p=0.000 r=0.997 p=0.000 r=0.994 p=0.000 - r=0.997 p=0.000 r=0.995 p=0.000

10 Rep3 10 10 10 α 10 - Rep3

8 α 8 - 8 8 8 Rep3 Rep3 - - 6 6 6 6 6 α 4 4 4 4 4 2 siCon 2 2 2 2 +IFN

0 siFOXA1 0 0 0 0 0 2 4 6 8 10 12 siCon+IFN 0 2 4 6 8 10 12 0 2 4 6 8 10 12 0 2 4 6 8 10 12 0 2 4 6 8 10 12 siFOXA1+siSTAT2 siCon-Rep2 siFOXA1-Rep2 siCon+IFNα-Rep2 siFOXA1+IFN siFOXA1+IFNα-Rep2 siFOXA1+siSTAT2 +IFNα-Rep2

RNA-seq chr1:1013236-1014801 (hg38) IFNα (-) IFNα (+) 380 C B siCON siFOXA1 siCON siFOXA1 siCON R1 R2 R3 R1 R2 R3 R1 R2 R3 R1 R2 R3 380 OAS2 XAF 1 GBP4 SFN CXCL11 MX2 OAS1 siFOXA1 IFI44L BST2 IFIT1 IFIT3 IFITM1 CCL5 380 RSAD2 OASL SIGLEC1 IFI6 IFI27 BATF2 TNFSF10 IFIT2 OAS3 siCON + IFNα SAMD9 GPER1 CXCL10 DEGS2 IRF9 380 PSMB9 ISG15 GBP3 PLSCR1 SAMD9L INAFM1 LGALS9 IFI35 CD274 siFOXA1 + IFNα ESPN IFI44 UBTD1 UBE2L6 SLC9A3 CMPK2 380 HNRNPUL2-BSCL2 DDX60L DDX60 IFITM3 UBA7 IRF7 SPSB1 siFOXA1 +siSTAT2 + IFNα APOL6 PARP14 HLA-DRA GBP2 CRIP2 ISG20 HLA-E HSPB1 LAMP3 IFIT5 RENBP OGFR RARRES3 MED16 TMEM54 NKPD1 CEL ISG15 NME3 GALK1 MT1X LAMA5 TRIM14 DDX58 GBP1 TLR3 MOV10 MT2A chr1:78649831-78664073 (hg38) DHX58 HERC6 25 FAM46A PSMB8 APOL1 HERC5 APOL3 TNFRSF14 SAMD10 siCON HELZ2 IFIH1 USP18 LGALS3BP ATP5F1D CLDN3 25 LRP3 DVL1 NDUFB7 H1FX IGFBP2 BCL3 TRIM22 PSMB10 UNC93B1 siFOXA1 RNF208 TSPAN4 LRP1 RHPN1 TRABD HES4 25 CORO1B GBX2 SP110 EIF2AK2 PADI2 INTS1 PARP9 PIM3 STAT1 siCON + IFNα GMPR BIRC3 PKP3 LY6E WDR24 25 STUB1 SLC9A3R2 CCND3 TIGD5 HIST1H2BK APOL2 MY D88 CIC HSPG2 siFOXA1 + IFNα IFITM2 BTBD2 OPLAH HLA-F RPS28P7 TRIM21 25 D TX3L NAPA TDRD7 PML inducible genes upregulated by FOXA1 KD (n = 172) = (n KD inducible FOXA1 upregulated genes by PTMA

- SECTM128 GTPBP1 siFOXA1 +siSTAT2 + IFNα HLA-B TAP1 SCRIB GAK SAMHD1L LAP3 PARP10 TRIM25 TAPBP PNRC1 IFNα HLA-A SNN PARP12 APC2 FKBP8 TRIM51 IFI44 EHD4 POLD1 FAM89B HLA-C CNFN SLC25A10 SAC3D1 THEM6 POLRMT SSBP3 chr6:30489509-30494194 (hg38) TRIM5 SAMHD1 FZR1 PLCH2 30 siCON

30 siFOXA1

30 siCON + IFNα

30 siFOXA1 + IFNα

30 siFOXA1 +siSTAT2 + IFNα

HLA-E

5 chr6:32854192-32859851 (hg38) siCON 5 siFOXA1 5 siCON + IFNα 5 siFOXA1 + IFNα 5 siFOXA1 +siSTAT2 + IFNα

PSMB9 Supplementary Figure 10. Impact of IFNα treatment on STAT2 target gene expression. (Related Figure 4) A. Analysis of correlation between the replicates for RNA-seq data in each indicated cellular condition. B. Heatmaps of RNA-seq data showing a group of IFNα-inducible genes (n = 172) significantly upregulated (P < 0.0001) by FOXA1 KD in LNCaP cells. Statistical comparison was done using one-way analysis of variance (ANOVA). C. UCSC tracks (integration of three replicates) show RNA-seq signals at genomic loci of STAT2 target genes ISG15, IFI44, HLA-E and PSMB9. Supplementary Figure 11

A RNA-seq in LNCaP FOXA1 ChIP-seq in LNCaP RNA-seq D IFNα (-) IFNα (+) FOXA1 KD-affected genes FOXA1 binding site-associated genes siFOXA1 (n = 2613) (n = 14,053) siCON siFOXA1 siCON siFOXA1 + siSTAT2 R1 R2 R3 R1 R2 R3 R1 R2 R3 R1 R2 R3 R1 R2 R3

FOXA1 target genes (n = 1421) (n =(n 650)

B FOXA1 upregulated genes FOXA1 downregulated genes (n = 650) (n = 771) upregulated FOXA1 genes 60 GO-BP 50 GO-BP 40 value) value) 40 30 P P ( (

10 20

10 20 10 Log Log - - 0 0

1 2 3 4 5 1 2 3 4 5 (n =(n 771) FOXA1 downregulated downregulated FOXA1 genes

RNA-seq C chr14: 350 siCON

350 siFOXA1 350 siCON + IFNα 350 siFOXA1 + IFNα

350 siFOXA1 +siSTAT2 + IFNα

FOXA1

chr12: 30 siCON 30 siFOXA1 30 siCON + IFNα 30 siFOXA1 + IFNα 30 siFOXA1 +siSTAT2 + IFNα

STAT2 Supplementary Figure 11. Impact of IFNα treatment on FOXA1 target gene expression. (Related Figure 4) A. Venn diagram shows the overlap between the genes affected by FOXA1 KD in LNCaP identified by RNA-seq and FOXA1 target genes identified by FOXA1 ChIP-seq in LNCaP cells. B. GO-BP analysis shows the top 5 pathways of FOXA1-up- or down-regulated. C. UCSC track profiles of RNA-seq signals at the FOXA1 and STAT2 gene loci in LNCaP cells transfected with siRNAs and/or treated with or without IFNα. D. Heatmaps of RNA-seq data showing the expression of FOXA1 up- and down-regulated genes in LNCaP cells transfected with indicated siRNAs and/or treated with or without IFNα. Supplementary Figure 12

TRAMP-C2 A B 700

600 Vehicle Poly(I:C) ) IFNα - - - - + + + + 3 500 Vector Vector ns

FOXA1-WT 0.0009 =

FOXA1 ns FOXA1-WT ns P = 0.0045 = 400 ns

FOXA1 αH3 FOXA1 αH3 P 300 FOXA1△SBR Stat1 FOXA1△SBR △ △ 200

Stat2 Tumor volume (mm 100

Isg15 0

7 1 102 133 164 185 206 227 248 269 28 30 32 34 Days 10 11 12 13 Erk2 Vehicle Poly(I:C)

Vector + Vehicle FOXA1-WT + Vehicle FOXA1 αH3 + Vehicle FOXA1 SBR + Vehicle C 30 30 30 30 △ △ 20 20 20 20

10 10 10 10

0 0 0 0 TSNE 2 TSNE 2 TSNE 2 -10 -10 TSNE 2 -10 -10

-20 -20 -20 -20

-30 -30 -30 -30 -20 0 20 -20 0 20 -20 0 20 -20 0 20 TSNE 1 TSNE 1 TSNE 1 TSNE 1 Vector + Poly(I:C) FOXA1-WT + Poly(I:C) FOXA1 αH3 + Poly(I:C) FOXA1 SBR + Poly(I:C) 30 30 30 30 △ △ 20 20 20 20

10 10 10 10

0 0 0 0 TSNE 2 TSNE 2 TSNE 2 -10 -10 TSNE 2 -10 -10

-20 -20 -20 -20

-30 -30 -30 -30 -20 0 20 -20 0 20 -20 0 20 -20 0 20 TSNE 1 TSNE 1 TSNE 1 TSNE 1 Colored by phenotype: CD45- CD45+CD3+CD4-CD8+ T cells CD45+CD3+CD8-CD4+ T cells CD45+CD3-NK1.1+ NK cells CD45+CD11b+Gr-1+ CD45+CD11b+Gr-1-

D Vehicle Poly(I:C) 35 35 Vector Vector 30 FOXA1-WT 30 FOXA1-WT FOXA1 αH3 FOXA1 αH3 25 FOXA1 SBR 25 FOXA1 SBR △ ns △ 20 △ ns 20 △ ns ns ns 15 ns 15 % of events

% of events ns ns ns ns ns ns 10 = 0.0167 10 = 0.0125 ns P

P ns = 0.0451 ns ns ns = 0.0339 ns = 0.0019 = 0.0027

ns 5 P 5 ns ns P P ns P ns ns 0 0 TotalCD8-CD8 Cyto-CD8CTL-CD8 CD4 NK cells MDSCsMDSCs TotalCD8-CD8 Cyto-CD8CTL-CD8 CD4 NK cells MDSCs + + + + + + + (CD45+CD3e+ (CD45+CD3e+ (CD45+CD3e+ (CD45+Nk1.1+ (CD45+ (CD45 CD3e (CD45+CD3e+ (CD45 CD3e (CD45 Nk1.1 (CD45 + - + - - + + CD8+CD4-) CD8+Granzyme B+) CD4+CD8-) CD3e-) CD11b+Gr-1+) CD8 CD4 ) CD8+Granzyme B+) CD4 CD8 ) CD3e ) CD11b Gr-1 )

Supplementary Figure 12. FOXA1 overexpression suppresses PCa immune response in mice. (Related Figure 5) A. Western blot analysis of indicated proteins in TRAMP-C2 murine PCa cells transfected with indicated expression vectors and treated with or without IFNα. Erk2, a loading control. B. Growth of TRAMP-C2-Vector and TRAMP-C2-FOXA1-WT, TRAMP-C2-FOXA1ΔαH3 and TRAMP-C2- FOXA1ΔSBR tumors treated with or without Poly(I:C). The data are presented as the mean ± SD (n = 10). Statistical significance was determined by two-way analysis of variance (ANOVA). C. CyTOF analysis of TRAMP-C2 murine PCa cells. PhenoGraph-defined cellular distribution and clustering, as defined by tSNE1 and tSNE2, colored by cellular phenotypes in TRAMP-C2 tumors stably expressing vector, FOXA1-WT, FOXA1ΔαH3 and FOXA1ΔSBR from mice at two days after the last administration of vehicle or Poly(I:C). D. CyTOF analysis of TRAMP-C2 murine PCa cells. Data were derived from normalized viable single cells and analyzed with the PhenoGraph algorithm and are shown in the bar graphs as means ± SD (n = 3). Statistical significance was determined by one-way ANOVA with Bonferroni correction for multiple tests. CTL, cytotoxic T lymphocyte; NK, natural killer cells; MDSC, myeloid-derived suppressor cells. Supplementary Figure 13

A Vehicle Poly(I:C) FOXA1 CD8 DAPI Merge FOXA1 CD8 DAPI Merge Vector Vector WT WT - - FOXA1 FOXA1 H H α α △ △ 3 3 FOXA1 FOXA1 SBR SBR △ △ FOXA1 FOXA1

ns 80

60 P = 0.0069 P = 0.0027 cells / field

+ 40 ns ns

CD8 ns 20

0 Vector1 FOXA12 -WT FOXA13 αH3 FOXA14 SBR Vector5 FOXA16 -WT FOXA17 αH3 FOXA18 SBR Vehicle Poly(I:C) △ △ △ △ B 0.20

0.15 Vecotor ns FOXA1-WT = 0.0164 =

= 0.0096 = FOXA1 αH3 P 0.10 P ns ns FOXA1△SBR

= 0.0025 = ns = 0.0017 =

ns ns 0.0113 = = 0.0105 = P P

P △ P mRNA level 0.05 ns ns ns ns (relative to Gapdh) ns ns

= 0.0136 = ns = 0.0101 = P

ns ns ns P 0.00 IFN stimulated genes: Isg15Isg15 Ifi44 H2-k1H2-k1 Psmb9Psmb9 Isg15 Ifi44 H2-k1H2-k1 Psmb9Psmb9 Vehicle Poly(I:C)

Supplementary Figure 13. FOXA1 inhibits immune response in mice. (Related to Figure 5) A. Immunofluorescence chemistry analysis of expression of the transfected FOXA1-WT, FOXA1∆αH3, FOXA1ΔSBR and CD8 in TRAMP-C2 tumors from mice at 2 day after the last administration of vehicle or poly(I:C). CD8+ cells were quantified in 10 randomly selected fields from tumor section; Scale bar,10μm. The data are presented as the mean ± SD from 3 tumors (n = 30 fields). Statistical significance was determined by one-way ANOVA with Bonferroni correction for multiple tests. B. RT-qPCR analysis of STAT2 target genes Isg15, Ifi44, H2-k1 and Psmb9 in TRAMP-C2 tumors transfected empty vector, FOXA1-WT, FOXA1ΔαH3 or FOXA1ΔSBR at 2 day after the last administration of vehicle or poly(I:C). The data are presented as the mean± SD (n = 9); Statistical significance was determined by one-way ANOVA with Bonferroni correction for multiple tests. Supplementary Figure 14

Winged Helix Domain FOXA1 TAD 100 % TAD 472aa A 168 268 FOXA2 TAD 95 % TAD 457aa 157 257 FOXA3 TAD 97 % TAD 350aa 115 215

Essential region to bind STAT2 (SBR) FOXA1-human 168 HAKPPYSYISLITMAIQQAPSKMLTLSEIYQWIMDLFPYYRQNQQRWQNSIRHSLSFNDCFVKVARSPDKPGKGSYWTLHPDSGNMFENGCYLRRQKRFKCEKQP 273 Foxa1-mouse 168 HAKPPYSYISLITMAIQQAPSKMLTLSEIYQWIMDLFPYYRQNQQRWQNSIRHSLSFNDCFVKVARSPDKPGKGSYWTLHPDSGNMFENGCYLRRQKRFKCEKQP 273 FOXA2-human 157 HAKPPYSYISLITMAIQQSPNKMLTLSEIYQWIMDLFPFYRQNQQRWQNSIRHSLSFNDCFLKVPRSPDKPGKGSFWTLHPDSGNMFENGCYLRRQKRFKCEKQL 261 Foxa2-mouse 157 HAKPPYSYISLITMAIQQSPNKMLTLSEIYQWIMDLFPFYRQNQQRWQNSIRHSLSFNDCFLKVPRSPDKPGKGSFWTLHPDSGNMFENGCYLRRQKRFKCEKQL 261 FOXA3-human 115 HAKPPYSYISLITMAIQQAPGKMLTLSEIYQWIMDLFPYYRENQQRWQNSIRHSLSFNDCFVKVARSPDKPGKGSYWALHPSSGNMFENGCYLRRQKRFKLEEKV 219 Foxa3-mouse 117 HAKPPYSYISLITMAIQQAPGKMLTLSEIYQWIMDLFPYYRENQQRWQNSIRHSLSFNDCFVKVARSPDKPGKGSYWALHPSSGNMFENGCYLRRQKRFKLEEKA 221

Input IP: Flag B 30 C D luc

- FoxA1 20 FoxA2 Flag - FOXA1 Flag - FOXA2 Flag - FOXA3 Flag - FOXA1 Flag - FOXA2 Flag - FOXA3 Flag - FOXA1 Flag - FOXA2 Flag - FOXA3 Flag - FOXA1 Flag - FOXA2 Flag - FOXA3 IFNα + + + + + + + + + + + + 10 FoxA3 luciferase activity Relative ISRE STAT2 ErK2 0 Flag IFNα (50 ng/L) - + + + + + + + + + + + + Vector ------Flag-FOXA1 ------Flag-FOXA2 ------Flag-FOXA3 ------

Flag

P-STAT2

STAT2 ERK2

Supplementary Figure 14. Assessment of the IFN inhibitory effect of FOXA1, FOXA2 and FOXA3. (Related to Figures 5) A. Diagram shows the domain structure and similarity of Winged Helix domain in the indicated FOX-proteins. B. Luciferase reporter assay shows the inhibitory effect of different dose of FOXA1, FOXA2 and FOXA3 on ISRE-luc reporter gene activity in DU145 cells. Data shown as means ± SD (n = 3). C. Co-IP analysis of the interaction of Flag-tagged FOXA1, FOXA2 and FOXA3 with STAT2 in 293T cells treated with IFNα. D. Western blot analysis of expression of FOXA1, FOXA2, FOXA3 proteins in the indicated cell lines. ERK2 was used as a loading control. Supplementary Figure 15

IgG Anti-PD1/CTLA4 B IN-shFoxa1#1 IN-shFoxa1#1 IN-shFoxa1#1 IN-shFoxa1#1 A MyC-CaP (Dox -) MyC-CaP (Dox +) MyC-CaP (Dox -) MyC-CaP (Dox +) MyC-CaPIN-shFoxa1#2 Dox - + - + IFNα - - + + Foxa1

P-Stat2 Foxa1

Stat2

Isg15

Erk2 DAPI

IgG Anti-PD1/CTLA4 Staining setup C MyC-CaPIN-shFoxa1#1 MyC-CaPIN-shFoxa1#1 MyC-CaPIN-shFoxa1#1 MyC-CaPIN-shFoxa1#1 No staining control Foxa1 single staining control (Dox -) (Dox +) (Dox -) (Dox +) SSC

Foxa1+ (38.4%) Foxa1+ (13.2%) Foxa1+ (42.7%) Foxa1+ (11.3%)

Foxa1 staining (Alexa fluor 488)

MyC-CaPIN-shFoxa1#1 MyC-CaPIN-shFoxa1#1 MyC-CaPIN-shFoxa1#1 MyC-CaPIN-shFoxa1#1 No staining control CD8 single staining control (Dox -) (Dox +) (Dox -) (Dox +) SSC

CD8+ (1.1%) CD8+ (1.9%) CD8+ (3.9%) CD8+ (10.2%)

CD8 staining (Alexa fluor 647)

MyC-CaPIN-shFoxa1#1 MyC-CaPIN-shFoxa1#1 MyC-CaPIN-shFoxa1#1 MyC-CaPIN-shFoxa1#1 CD8 single staining control (Dox -) (Dox +) (Dox -) (Dox +) staining CD8+Gzmb+ CD8+Gzmb+ CD8+Gzmb+ CD8+Gzmb+ (0.13%) (0.16%) (1.2%) (3.3%) Gzmb (Alexa (Alexa fluor 594)

CD8 staining (Alexa fluor 647)

D Metastatic castration-resistant prostate cancer (CRPC) treated with personalized peptide vaccination (PPV) immunotherapy 100 FOXA1 mRNA IFN activity CD3E mRNA PRF1 mRNA GZMB mRNA 6.0 3.0 9 14 15 P < 0.0001 P < 0.0001 P = 0.0302 P = 0.0236 P = 0.0327 FOXA1 mRNA low 80 (median survival: 489 days) 14 FOXA1 mRNA high 5.5 2.5 12 8 60 (median survival: 321 days)

score 13 5.0 2.0 P = 0.0158 10 40 IFN 7 12

Overall survival (%) 4.5 1.5 20 8 Log2 (mRNA level)expression Log2 (mRNA Log2 (mRNA level)expression Log2 (mRNA Log2 (mRNA level)expression Log2 (mRNA Log2 (mRNA level)expression Log2 (mRNA 11 6 0 4.0 1.0 FOXA1 high FOXA1 low FOXA1 high FOXA1 low FOXA1 high FOXA1 low FOXA1 high FOXA1 low FOXA1 high FOXA1Llow 0 1000 2000 3000 4000 5000 L H L H L H L H H (n = 72) (n = 40) (n = 72) (n = 40) (n = 72) (n = 40) (n = 72) (n = 40) (n = 72) (n = 40) Days Supplementary Figure 15. Foxa1 inhibits immune response in mice and in patient of PCa. (Related to Figure 6) A. Western blot analysis of indicated proteins in MyC-CaP murine PCa cells stably expressing doxycycline-inducible lentiviral shFoxa1#2 (MyC-CaPIN-shFoxa1#2) and treated with or without doxycycline (Dox) or/and IFNα. Erk2, a loading control. B. Immunofluorescence chemistry analysis of expression of the transfected Foxa1 in MyC-CaPIN-shFoxa1#1 tumors from mice at 2 day after the last administration of IgG or Anti-PD1/CTLA4; Scale bar, 10 μm. C. Flow cytometry analysis of Foxa1, CD8 and Gzmb positive cells in MyC-CaPIN-shFoxa1#1 tumors from mice at two days after the last administration of IgG or anti-PD1/CTLA-4. Dox (-), without doxycycline treatment; Dox (+), with doxycycline treatment. D. Overall survival of patients with FOXA1-low or -high PCa treated with personalized peptide vaccination (PPV) immunotherapy. Analysis of microarray data (GSE53922) from this cohort of patients shows the inverse correlation of FOXA1 expression with the IFN activity and expression of CD3E, CD8A and GZMB genes. The mean value of FOXA1 expression level was used as the cutoff. Statistical significance was determined by Log-rank (Mantel-Cox) test for overall survival and the others by Mann-Whitney U test. Supplementary Figure 16

A Breast cancer treated with NAC (Mayo Clinic cohort, n = 128) Clinical molecular subtype pCR No-pCR

Lum A (n = 11) 0 (0%) 11 (100%)

Lum B (n = 37) 4 (10.8%) 33 (89.2%)

Lum unknown (n = 2) 0 (0%) 2 (100%)

ER+/HER2+ (n=16) 4 (25%) 12 (75%)

ER-/HER2+ (n = 20) 12 (60%) 8 (40%)

LAR (n = 6) 1 (16%) 5 (84%) TNBC Basal-like (n = 36) 23 (64%) 13 (36%)

B Lum A Lum B ER+/HER2+ ER−/HER2+ TNBC P-value of LAR correlation with FOXA1: 2.4e−06 No P-value available P = 0.00164 P = 0.221 P = 0.379 P-value of pCR vs No-pCR correlation with FOXA1: 0.188 10 10 10 10 10 LAR Basal-like 8 8 8 8 8 6 6 6 6 6 expression 4 4 4 4 4 2 2 2 2 2 FOXA1 0 0 0 0 0

Log2 -2 -2 -2 -2 -2 pCR No-pCR pCR No-pCR pCR No-pCR pCR No-pCR pCR No-pCR (n = 0) (n = 11) (n = 3) (n = 33) (n = 4) (n = 11) (n = 12) (n = 8) (n = 23) (n = 19) Breast cancer treated with NAC (Mayo Clinic cohort)

C TNBC treated with NAC (epirubicin, cyclophosphamide and docetaxel) (GEICAM/2006-03, NCT00432172; GSE106977) Before removing LAR 10 After removing LAR LAR 9 Basal-like

8 P = 0.0056 8 P = 0.0290 expression expression 7 7 FOXA1 FOXA1 6 6 Log2 Log2

5 5

4 4 pCR No-pCR pCR No pCR (n = 73) (n = 46) (npCR = 44) (n = 61) No pCR

D Breast cancer treated with NAC (Mayo Clinic cohort) E Breast cancer treated with NAC (Mayo Clinic cohort) 30 P = 0.4995 30 pCR r = -0.0815, P = 0.3997 No-pCR

25 25

20 20

15 15

10 10

5 5 Tumor mutation mutation perTumor burden (Mutations MB) Tumor mutation mutation perTumor burden (Mutations MB)

0 0 pCR (n = 40) No-pCR (n = 69) -2 0 2 4 6 8 Log2 FOXA1 expression level Note: some samples have missing mutation load values pCR (n = 40), No-pCR (n = 69); Note: some samples have missing mutation load values Supplementary Figure 16. Assessment of inverse correlation between FOXA1 expression and responsiveness to NAC in molecular subtypes of breast cancer. (Related to Figure 7) A. The number of cases of breast cancers in different molecular subtypes (from a cohort of patients of the Mayo Clinic) with pathological complete response (pCR) or No-pCR to neoadjuvant chemotherapy (NAC). B. Comparison of FOXA1 expression level in luminal A (Lum A), luminal B (Lum B), ER+/HER2+, ER-/HER2+, and TNBC (LAR or basal-like) tumors (from the Mayo Clinic cohort of patients) with pathological complete response (pCR) or No- pCR to NAC. Data shown as means ± SD; Statistical comparison was done using Mann-Whitney U test. C. Analysis of microarray data shows the association of FOXA1 expression in a cohort of TNBC tumors (GSE106977) with pCR or No-pCR to NAC. Right, before removing luminal AR (LAR); Left, after removing luminal AR (LAR). Data shown as means ± SD; Statistical comparison was done using Mann-Whitney U test. D. Analysis of the association of DNA mutation burden in the cohort of 126 BCa samples from Mayo Clinic with tumors response to NAC (pCR versus No-pCR). Data shown as means ± SD; Statistical significance was determined by Mann- Whitney U test. E. Analysis of the association of DNA mutation burden in the cohort of 126 BCa samples from Mayo Clinic with tumors response to FOXA1 expression levels. Statistical significance was determined by Pearson correlation test. Supplementary Figure 17

Cytoplasm Nucleus A Vehicle + - - + - - IFNα - + - - + - B siCon + - - + - - IFNγ - - + - - + siFOXA1 - + + - + + P-STAT1 IFNα - - - + + + STAT1 FOXA1

P-STAT2 P-STAT1 STAT2 IP:FOXA1 STAT1 FOXA1 P-STAT2 FOXA1 STAT2 P-STAT1 ISG15 STAT1 ERK2 P-STAT2 Input RT4 (bladder cancer) STAT2

PARP1

HSP70 RT4 (bladder cancer)

Supplementary Figure 17. The effect of FOXA1 on IFN signaling in bladder cancer. (Related to Figure 8) A. Co-IP analysis of FOXA1 interaction with STAT1 and STAT2 at the endogenous level in RT4 bladder cancer cells after treatment of IFNα or IFNγ. B. Western blot analysis of the indicated proteins in RT4 bladder cancer cells transfected with control siRNA (siCon) or FOXA1 specific siRNA (siFOXA1) and/or treated with or without IFNα. Supplementary Figure 18 Urothelial carcinomas treated FOXA1 IHC Urothelial carcinomas treated FOXA1 IHC with anti-PD1 (FOXA1 staining ) score (IS) with anti-PD1 (FOXA1 staining ) score (IS) Low magnification High magnification Low magnification High magnification

20-50% (IS = 2) 20-50% (IS = 2) Patient 14 Patient Patient 1 Patient

2 mm 200 µm 2 mm 200 µm

1-20% (IS = 1) 1-20% (IS = 1) Patient 15 Patient Patient 2 Patient

2 mm 200 µm 2 mm 200 µm

< 1% (IS = 0) Patient 16 Patient >50% (IS = 3) Patient 3 Patient

2 mm 200 µm 2 mm 200 µm

>50% (IS = 3) >50% (IS = 3) Patient 4 Patient Patient 17 Patient

2 mm 200 µm 2 mm 200 µm

< 1% (IS = 0) 20-50% (IS = 2) Patient 5 Patient Patient 18 Patient

2 mm 200 µm 2 mm 200 µm

>50% (IS = 3) Patient 19 Patient Patient 6 Patient 1-20% (IS = 1)

2 mm 200 µm 2 200 µm m m

1-20% (IS = 1) < 1% (IS = 0) Patient 20 Patient Patient 8 Patient

2 mm 200 µm 2 mm 200 µm

< 1% (IS = 0) 20-50% (IS = 2) Patient 21 Patient Patient 9 Patient

2 mm 200 µm 2 mm 200 µm

< 1% (IS = 0) Patient 22 Patient < 1% (IS = 0) Patient 10 Patient

2 mm 200 µm 2 mm 200 µm

20-50% (IS = 2) 1-20% (IS = 1) Patient 11 Patient Patient 23 Patient

2 mm 200 µm 2 mm 200 µm

1-20% (IS = 1) Patient 12 Patient

2 mm 200 µm

>50% (IS = 3) Patient 13 Patient

2 mm 200 µm Supplementary Figure 18. IHC analysis of FOXA1 protein expression in pre-treatment urothelial carcinoma specimens from a cohort of patients treated with anti-PD1 immunotherapy. (Related to Figure 8) FOXA1 IHC was performed using a FOXA1-specific antibody on pre-treatment urothelial carcinoma specimens from a cohort of 22 patients treated with anti-PD1 immunotherapy. Low and high magnification FOXA1 IHC images in each case and FOXA1 IHC scores are shown (see scoring details in Supplementary Methods and in Supplementary Table 3). Scale bars: 2 mm (low magnification); 200 µm (high magnification). Supplementary Figure 19

A Bladder cancer treated with Bacillus Calmette-Guerin (BCG) immunotherapy (GSE19423) FOXA1 mRNA low 100 (median survival: NA) FOXA1 mRNA IFN activity CD3E mRNA GZMB mRNA FOXA1 mRNA high (median survival: 70.93 months) P < 0.0001 60 P = 0.0197 P = 0.0422 11 P = 0.0311 80 14 8.5 40 13 10 60 20 8.0 P = 0.0659 12 0 9

40 IFN score -20 7.5 11 8 Overallsurvival (%) -40 20 10 -60 7

7.0 level)expression Log2 (mRNA Log2 (mRNA level)expression Log2 (mRNA Log2 (mRNA level)expression Log2 (mRNA FOXA1 low FOXA1 high FOXA1 low FOXA1 high FOXA1 low FOXA1 high FOXA1 low FOXA1 high L H 0 (n = 30) (n = 18) (n = 30) (n = 18) (n = 30) (n = 18) (n = 30) (n = 18) pCR pCR pCR 0 50 100 150 No pCR No pCR No pCR Months B Bladder cancer treated with MVAC (methotrexate, vinblastine, doxorubicin, and cisplatin) chemotherapy (GSE70691) FOXA1 mRNA low 100 (median survival: 144 months) FOXA1 mRNA IFN activity CD8A mRNA GZMB mRNA

FOXA1 mRNA high P < 0.0001 P < 0.0001 P = 0.0126 15 P = 0.0117 15 40 15 80 (median survival: 65.7 months) 20 60 P = 0.0408 0 10 10 10

40 IFN score -20

-40 20

Overallsurvival (%) 5 5 Log2 (mRNA level)expression Log2 (mRNA -60 level)expression Log2 (mRNA Log2 (mRNA level)expression Log2 (mRNA 5 FOXA1 low FOXA1 high FOXA1 low FOXA1 high FOXA1 low FOXA1 high FOXA1 low FOXA1 high 0 L H L H L H (n =L 20) (nH = 29) (n = 20) (n = 29) (n = 20) (n = 29) (n = 20) (n = 29) 0 100 200 300 400 Months

Supplementary Figure 19. FOXA1 confers to immunotherapy and chemotherapy resistance in bladder cancer. (Related to Figure 8) A. Overall survival of patients with FOXA1-low or -high bladder cancer treated with Bacillus Calmette-Guerin (BCG) immunotherapy. Analysis of microarray data (GSE19423) from this cohort of patients shows the inverse correlation of FOXA1 expression with the IFN activity and expression of CD3E, CD8A and GZMB T cell marker genes. The mean value of FOXA1 expression level was used as the cutoff. Statistical significance was determined by Log-rank (Mantel- Cox) test for overall survival and the others by Mann-Whitney U test. NA = Not available. B. Overall survival of patients with FOXA1-low or -high bladder cancer treated with methotrexate, vinblastine, doxorubicin, and cisplatin (MVAC) chemotherapy. Analysis of microarray data (GSE70691) from this cohort of patients shows the inverse correlation of FOXA1 expression with the IFN activity and expression of CD3E, CD8A and GZMB T cell marker genes. The mean value of FOXA1 expression level was used as the cutoff. Statistical significance was determined by Log-rank (Mantel-Cox) test for overall survival and the others by Mann-Whitney U test. Supplementary Table 1. Clinic information of prostate cancer samples used for single cell RNA-seq.

Gleason pre-operative Clinical Clinical Clinical Neoadjuvant Tumor Sample ID WHO/ISUP pT pN cellularity score PSA (ng/mL) T N M ADT

Patient 1 4+5 5 60% 79.44 T4 N1 M0 T3b N1 NO

Patient 2 5+5 5 70% 18.51 T4 N0 M0 T3b N0 NO

Patient 3 4+5 5 80% 7.05 T3a N0 M0 T3b N0 NO

Patient 4 4+5 5 80% 22.57 T3a N0 M0 T3a N0 NO

Patient 5 4+3 3 80% 10.50 T2c N0 M0 T2c N0 NO

Patient 6 4+3 3 90% 336.00 T2c N1 M0 T3b N1 NO

Patient 7 4+5 5 10% 16.00 T2c N1 M1b T2c N1 NO

Patient 8 4+5 5 70% 16.00 T2c N1 M1b T2c N1 NO

Patient 9 4+5 5 90% 20.00 T2c N0 M0 T2c N0 YES

Patient 10 3+4 2 60% 81.09 T2c N0 M0 T2c N0 NO

Patient 11 4+5 5 80% 19.00 T3a N0 M1b T3a N0 NO

Patient 12 4+3 3 80% 51.00 T3a N0 M0 T3a N0 NO

Patient 13 4+5 5 50% 14.00 T2b N0 M1a T3a N0 NO

Supplementary Table 2. Single cell RNA-seq information of prostate cancer patients used for analysis.

Sample ID Average_nGene Average_nUMI Average_percent.mito Cell_Number Epithelia_Cell

Patient 1 2864.398856 14306.05616 0.082028703 1923 668

Patient 2 466.7349769 874.4571135 0.055694131 1947 335

Patient 3 2026.572198 8897.319504 0.104496127 3712 963

Patient 4 3034.626333 14956.16967 0.074302536 6000 5694

Patient 5 1360.559574 5324.886525 0.30085929 4230 1374

Patient 6 1543.450465 5390.983189 0.074496791 6127 3688

Patient 7 557.5352157 1272.217322 0.053934186 3152 951

Patient 8 1970.975065 8287.448083 0.104352916 5374 2852

Patient 9 1160.757277 4373.853596 0.097172687 2336 262

Patient 10 2399.793621 9504.188793 0.07756701 5800 2129

Patient 11 1651.909848 6503.126894 0.184457406 2640 798

Patient 12 1469.374933 6327.621578 0.144288349 3726 1014

Patient 13 2037.840928 8201.765801 0.217854816 6123 2946

Supplementary Table 3. FOXA1 IHC staining of urothelial carcinoma with anti-PD1 treatment (see FOXA1 staining data in Supplementary Figure 18).

Patient Cancer type Pathological Gender Age Disease Progression FOXA1 ID type (M = progression free staining score male; (1 = Yes, 0 survival (0 = “< 1%”, 1 F = = No) time = “1-20%”, 2 female) (Days) = “20-50%”, 3 = “>50%”) (Years) 1 bladder cancer urothelial M 54 1 111 2 carcinoma 2 bladder cancer urothelial M 67 0 927 1 carcinoma 3 bladder cancer urothelial M 44 0 196 0 carcinoma 4 bladder cancer urothelial M 51 1 112 3 carcinoma 5 bladder cancer urothelial M 66 0 296 0 carcinoma 6 ureteral cancer urothelial M 77 1 251 3 carcinoma 8 bladder cancer urothelial M 73 1 207 1 carcinoma 9 renal urothelial F 47 0 128 0 pelvis urothelial carcinoma carcinoma 10 renal urothelial M 64 0 50 0 pelvis urothelial carcinoma carcinoma 11 bladder cancer urothelial M 62 1 52 2 carcinoma 12 bladder cancer urothelial M 68 1 40 1 carcinoma 13 ureteral cancer urothelial F 71 1 70 3 carcinoma 14 ureteral cancer urothelial M 71 1 54 2 carcinoma 15 bladder cancer urothelial M 70 1 55 1 carcinoma 16 bladder cancer urothelial M 62 1 55 3 carcinoma 17 renal urothelial F 66 1 60 3 pelvis urothelial carcinoma carcinoma 18 bladder cancer urothelial M 66 1 23 2 carcinoma 19 ureteral cancer urothelial M 65 1 65 1 carcinoma 20 ureteral cancer urothelial M 68 0 51 0 carcinoma 21 bladder cancer urothelial M 48 0 36 2 carcinoma 22 bladder cancer urothelial M 63 1 21 0 carcinoma 23 renal urothelial M 56 0 219 1 pelvis urothelial carcinoma carcinoma

Supplementary Table 4. Primers for RT-qPCR and ChIP-qPCR.

Names Sequence (5’-3’) Used for GAPDH-F CCGGGAAACTGTGGCGTGATGG RT-qPCR GAPDH-R: AGGTGGAGGAGTGGGTGTCGCTGTT RT-qPCR FOXA1-F: AGATGGAAGGGCATGAAACCA RT-qPCR FOXA1-R: ATGTTGCCGCTCGTAGTCAT RT-qPCR STAT2-F GAGGAGAAGCAATGGGTCTTAG RT-qPCR STAT2-R GGTCCACAACCAACGAATAGA RT-qPCR IFI27-F: GTTCATCCTGGGCTCCATTG RT-qPCR IFI27-R: CCCCTGGCATGGTTCTCTT RT-qPCR HLA-E-F CCTACGACGGCAAGGA RT-qPCR HLA-E-R CCCTTCTCCAGGTATTTGTG RT-qPCR PSMB9-F GAGAGGACTTGTCTGCACATC RT-qPCR PSMB9-R GCATCCACATAACCATAGATAAAGG RT-qPCR IFI44-F TGGTACATGTGGCTTTGCTC RT-qPCR IFI44-R CCACCGAGATGTCAGAAAGAG RT-qPCR IFI44L-F TGCACTGAGGCAGATGCTGCG RT-qPCR IFI44L-R TCATTGCGGCACACCAGTACAG RT-qPCR IFI16-F ACTGAGTACAACAAAGCCATTTGA RT-qPCR IFI16-R TTGTGACATTGTCCTGTCCCCAC RT-qPCR PSMB9-F GAGAGGACTTGTCTGCACATC RT-qPCR PSMB9-R GCATCCACATAACCATAGATAAAGG RT-qPCR ISG15-F GCGCAGATCACCCAGAAGAT RT-qPCR ISG15-R TCCTCACCAGGATGCTCAGA RT-qPCR Gapdh-F: TGGAAAGCTGTGGCGTGAT RT-qPCR Gapdh-R: TGCTTCACCACCTTCTTGAT RT-qPCR Foxa1-F: TCCTTATGGCGCTACCTTGC RT-qPCR Foxa1-R: CCTTGGTAGTAGGCTGGCTC RT-qPCR Isg15-F: AAGCAGCCAGAAGCAGACTC RT-qPCR Isg15-R: CACCAATCTTCTGGGCAATC RT-qPCR Ifi44-F: AACTGACTGCTCGCAATAATGT RT-qPCR Ifi44-R: GTAACACAGCAATGCCTCTTGT RT-qPCR Psmb9-F: ACCGTGAGGACTTGTTAGCG RT-qPCR Psmb9-R: GTAAAGGGCTGTCGAATTAGCA RT-qPCR H2-k1-F: CCGCGGACGCTGGATA RT-qPCR H2-k1-R: GGCGATTCGCGACTTCTG RT-qPCR ISG15-F: TGCATCCTGTGAAGGATCTG ChIP-qPCR ISG15-R: TTTGGCTTCAGTTTCGGTTT ChIP-qPCR IFI44-F: TGAGAGAAGTTGGCATGCTG ChIP-qPCR IFI44-R: AGCTGAGGGTAGCTGCTCTGT ChIP-qPCR HLA-E-F: GCCCAGCCAGGACTAATTTC ChIP-qPCR HLA-E-R: CACCCATTGGGAATGAGAAC ChIP-qPCR PSMB9-F: CCGAGTCCTCCCCTACTGG ChIP-qPCR PSMB-R: GCAGGCCACTTTTGGAAGTA ChIP-qPCR

Supplementary Table 5. Antibodies used for CyTOF.

Antibodies Clone Tag Source Localization

Anti-Mouse CD45 30-F11 89Y Fluidigm, Cat#: 3089005B surface

Anti-Mouse CD3e 145-2C11 152Sm Fluidigm, Cat#: 3152004B surface

Anti-Mouse CD8a 53-6.7 153-Eu Fluidigm, Cat#: 3153012B surface

Anti-Mouse CD4 RM4-5 145-Nd Fluidigm, Cat#: 3145002B surface

Anti-Mouse CD11b M1/70 172-Yb Fluidigm Cat#: 3172012B surface

Anti-Mouse NK1.1 PK136 170Er Fluidigm, Cat#: 3170002B surface

Anti-Mouse Ly-6G 1A8 141Pr Fluidigm, Cat#: 3141008B surface

Anti-Mouse Ly-6C HK1.4 162Dy Fluidigm, Cat#: 3162014B surface

Anti-Mouse Foxp3 FJK-16s 158Gd Fluidigm, Cat#: 3158003A intracellular

Anti-Mouse Granzyme B GB11 173Yb Fluidigm, Cat#: 3173006B intracellular

Full unedited gel and blot images

Figure 2B, right Figure 2B, left Input IP: FOXA1 Input IP: FOXA1

P-STAT1 P-STAT1

STAT1 STAT1

P-STAT2 P-STAT2

STAT2 STAT2

FOXA1 FOXA1

PARP1 PARP1

HSP70 HSP70

Figure 2C Input IP: Flag Figure 2E Input GST-Pulldown

V5-FOXA1

Flag Flag-STAT2 Flag-STAT1

100 70 50 PARP1 37 25

HSP70 Figure 2F Input GST-Pulldown Figure 2G

Flag

Flag

130 100 70 50 37

25 20

Figure S5A Figure 6A

ERK2

FOXA1

Foxa1

P-Stat2

Stat2

Isg15

Erk2 Figure S5B, left Figure S5B, right

SFB-FOXA1 SFB-FOXA1

STAT1 P-STAT1 STAT1 P-STAT1

P-STAT2 STAT2 STAT2 P-STAT2

ERK2 ERK2

Figure S5C Figure S5D

P-STAT2 P-STAT2 FOXA1 FOXA1

P-STAT1 P-STAT1 P-STAT1 P-STAT1

P-STAT2 P-STAT2 STAT2 STAT2

STAT1 STAT1 STAT1 STAT1

STAT2 STAT2 FOXA1 FOXA1

ERK2 ERK2 Figure S5F Input IP: HA Figure S5EInput IP: Flag

V5-FOXA1

Flag-STAT2 HA

STAT1

GFP-IRF9 Figure S5G V5-FOXA1 Input GST-Pulldown

FOXA1 Flag-STAT1

HA-STAT1 10070 50 37 25 20 15 STAT2

V5-FOXA1

Figure 5I Figure S5H Input GST-Pulldown Input GST-Pulldown

FOXA1 Flag

100 70 50 37 25 130 100 70 50 37

25 20 Figure S5L Figure S5J

EMSA STAT1 IB: Biotin

P-STAT1

STAT2 P-STAT2

Flag ERK2

Figure S5O Figure S5N

STAT2

Biotin

Flag-FOXA1

STAT2 P-STAT2

P-STAT1 1 Flag-FOXA1

STAT2

Flag-FOXA1 STAT1

STAT2 Flag-AR

Flag-FOXA1 HDAC2

ERK2 Figure S6A Figure S7B

FOXA1 FOXA1

STAT2 P-STAT2

P-STAT2

STAT2 P-STAT1

P-STAT1 STAT1

STAT1

ERK2

ERK2

Figure S9B Figure S9A

FOXA1

FOXA1 IP: Flag

FOXA1 Input P-STAT2

P-STAT1

IP: Flag Flag-STAT2 STAT2

Input Flag-STAT2 STAT1

Input PARP1 ISG15

Input HSP70

MHCI

ERK2 Figure S9C Figure S9D

FOXA1

STAT1

P-STAT1 P-STAT1

STAT2 STAT1

P-STAT2

P-STAT2

SFB-FOXA1 STAT2

ERK2 ISG15

MHCI

ERK2

Figure S9H Figure S12A

Flag

STAT2

Flag

STAT2 Figure S14B Figure S14C

Flag

STAT2

P-STAT2

P-STAT1 Flag

STAT1

STAT2 ERK2 Figure S15A

Figure S14D - CaP - H660 NCI PC3 LNCaP TRAMP - C2 MyC

FOXA1 - CaP - H660 NCI PC3 LNCaP TRAMP - C2 MyC FOXA2 - CaP - H660 PC3 LNCaP MyC TRAMP - C2 NCI FOXA3

ERK2 Figure S17A Figure S17B

P-STAT1 FOXA1

STAT1 P-STAT1

P-STAT2 STAT1

STAT2

P-STAT2

FOXA1

STAT2

PARP1 ISG15

HSP70 ERK2