Verteporfin Inhibits PD-L1 Through Autophagy and the STAT1–IRF1

Published OnlineFirst April 7, 2020; DOI: 10.1158/2326-6066.CIR-19-0159

CANCER IMMUNOLOGY RESEARCH | RESEARCH ARTICLE

Verteporﬁn Inhibits PD-L1 through Autophagy and the STAT1–IRF1–TRIM28 Signaling Axis, Exerting Antitumor Efﬁcacy A C Jiyong Liang1, Lulu Wang2, Chao Wang1,3, Jianfeng Shen2, Bojin Su1, Anantha L. Marisetty4, Dexing Fang4, Cynthia Kassab4, Kang Jin Jeong1, Wei Zhao1, Yiling Lu1, Abhinav K. Jain5, Zhicheng Zhou1, Han Liang6, Shao-Cong Sun7, Changming Lu8, Zhi-Xiang Xu9, Qinghua Yu1, Shan Shao1, XiaoHua Chen1, Meng Gao1, Francois X. Claret1, Zhiyong Ding1, Jian Chen10, Pingsheng Chen11, Michelle C. Barton5, Guang Peng2, Gordon B. Mills12, and Amy B. Heimberger4

ABSTRACT ◥ Programmed cell death 1 ligand 1 (PD-L1) is a key driver of STAT1–IRF1–TRIM28 signaling cascade, but did not affect the tumor-mediated immune suppression, and targeting it with anti- proinflammatory CIITA-MHC II cascade. Within the tumor bodies can induce therapeutic responses. Given the costs and microenvironment, verteporfin inhibited PD-L1 expression, associated toxicity of PD-L1 blockade, alternative therapeutic which associated with enhanced T-lymphocyte infiltration. Inhi- strategies are needed. Using reverse-phase protein arrays to assess bition of chromatin-associated enzyme PARP1 induced PD-L1 drugs in use or likely to enter trials, we performed a candidate drug expression in high endothelial venules (HEV) in tumors and, screen for inhibitors of PD-L1 expression and identified verte- when combined with verteporfin, enhanced therapeutic efficacy. porfin as a possible small-molecule inhibitor. Verteporfin sup- Thus, verteporfin effectively targets PD-L1 through transcription- pressed basal and IFN-induced PD-L1 expression in vitro and al and posttranslational mechanisms, representing an alternative in vivo through Golgi-related autophagy and disruption of the therapeutic strategy for targeting PD-L1.

1Systems Biology, The University of Texas MD Anderson Cancer Center, Houston, Introduction Texas. 2Clinical Cancer Prevention, The University of Texas MD Anderson Cancer Discovered first as a B7 family ligand of the programmed cell death 1 3 Center, Houston, Texas. Department of Obstetrics & Gynecology, Obstetrics & (PD1) encoded by the CD274 gene (1), PD-L1 is a checkpoint molecule Gynecology Hospital of Fu Dan University, Shanghai, China. 4Neurosurgery, The that initiates an inhibitory pathway, suppressing CTLs upon PD-1 University of Texas MD Anderson Cancer Center, Houston, Texas. 5Genes and Development Graduate Program, Department of Epigenetics and Molecular engagement. This mechanism that normally serves to prevent over Carcinogenesis, and Center for Cancer Epigenetics, The University of Texas MD activation of immune responses is frequently coopted by cancer cells to Anderson Cancer Center, Houston, Texas. 6Bioinformatics and Computational evade immune surveillance (2). Subsequent studies have provided Biology, The University of Texas MD Anderson Cancer Center, Houston, Texas. compelling evidence that the PD-L1/PD1 cascade is a highly effective 7 Immunology, The University of Texas MD Anderson Cancer Center, Houston, therapeutic target for immune checkpoint blockade therapy that yields Texas. 8The Institute of Oral Health, School of Dentistry, University of Alabama durable anticancer efficacy and prolongs patient survival (3). at Birmingham, Birmingham, Alabama. 9Division of Hematology and Oncology, Comprehensive Cancer Center, University of Alabama at Birmingham, Birming- PD-L1 is expressed in a variety of cancer types in either a consti- ham, Alabama. 10Department of General Surgery, Second Affiliated Hospital, tutive (or intrinsic) or IFN-induced manner, and the mechanisms Zhejiang University School of Medicine, Zhejiang Province, Hangzhou, China. controlling the expression of PD-L1 mRNA and protein have been the 11Department of Pathology, Medical School of Southeast University, Nanjing, subject of numerous studies. The STAT1–IRF1 axis plays a central role 12 Jiangsu, China. Cell, Developmental & Cancer Biology, Oregon Health and in mediating IFN-induced PD-L1 transcription in both cancer and Sciences University, Portland, Oregon. noncancer cells (4). In addition to transcription regulation, epigenetic Note: Supplementary data for this article are available at Cancer Immunology mechanisms, 30 UTR variations, and miRNAs have roles in fine-tuning Research Online (http://cancerimmunolres.aacrjournals.org/). the relative expression of PD-L1 in context-dependent settings (5). J. Liang, L. Wang, and C. Wang contributed equally to this article. Other mechanisms control PD-L1 degradation, ubiquitination, autop- Current address for J. Shen: Department of Ophthalmology, Ninth People's hagy, glycosylation, and recycling from plasma membrane and intra- Hospital, Shanghai Jiao Tong University School of Medicine and Shanghai Key cellular compartments (6, 7). In contrast to IFN-dependent PD-L1 Laboratory of Orbital Diseases and Ocular Oncology, Shanghai, China; and transcription, the mechanisms underlying intrinsic PD-L1 expression current address for Z.-X. Xu, School of Life Sciences, Henan University, in cancer are poorly understood. Kaifeng, China. Toxicity and cost are two key issues, among others, that limit the use Corresponding Authors: Amy B. Heimberger, The University of Texas MD Ander- of antibody approaches to immune checkpoints (8, 9). The complexity son Cancer Center, 1400 Holcombe Boulevard, Unit 442, Houston, TX 77030. of the mechanisms controlling PD-L1 expression poses another major Phone: 713-792-2400; Fax: 713-794-4950; E-mail: [email protected]; Jiyong Liang, [email protected]; and Gordon B. Mills, Oregon Health and challenge for the development of small-molecule PD-L1 inhibitors. Sciences University, Portland, OR 97239. E-mail: [email protected] Targeting one of the mechanisms controlling either intrinsic or IFN- induced PD-L1 expression may not be sufficient because this pathway Cancer Immunol Res 2020;8:952–65 controls the expression of >7,000 genes. Many of these genes are doi: 10.1158/2326-6066.CIR-19-0159 essential for immune responses, especially those along the CIITA- 2020 American Association for Cancer Research. MHC cascade, which are crucial for cancer immunogenicity (10).

AACRJournals.org | 952

Verteporﬁn Inhibits PD-L1

Thus, it is imperative to preserve the CIITA-MHC cascade and IRF3 (#11904), STAT1 (D1K9Y-#14994), YAP1 (#15117), TRIM28 immune responses while still targeting PD-L1 expression. (#4123), and protein A-HRP (#12991) were obtained from Cell In mammalian cells, the endoplasmic reticulum (ER)-Golgi net- Signaling Technology; p62/SQSTM1 (610832) from BD Biosciences, work plays an important role in autophagy and is a major source of HLA-Ds (sc-53302), and ERK2 from Santa Cruz Biotechnology; membrane structures contributing to formation of autophago- COG3 (11130-1-AP) from ProteinTech group; and COG7 somes (11, 12). PD-L1 is one of a large number of glycosylated proteins, (EPR9942-ab168362) from Abcam. including most of the MHC antigens, that are processed through the Golgi apparatus (13). The conserved oligomeric Golgi proteins (COG) Western blotting in the Golgi apparatus play an important role in posttranslational Cell lysis was performed by lysing 5 106 cells with NP-40 lysis processing and transport of secreted peptides and proteins that are buffer containing 50 mmol/L Tris (pH 7.4), 150 mmol/L NaCl, 0.5% glycosylated and targeted to the plasma membranes (14). The oligo- NP-40, protease inhibitors (1873580001, Millipore Sigma), and phos- meric Golgi complex includes 8 members (COG1-8), with the central phatase inhibitors (4906837001, Millipore Sigma) on ice for 15 COG1 surrounded by a COG2-4 forming lobe A and COG5-8 forming minutes followed by microcentrifugation at 14,000 g for 15 minutes lobe B (15–19). PD-L1 associates with 7 of the 8 COGs (20), whereas at 4C to remove debris. Immunoblotting was performed using 50 mg the MHC molecules associate with the coatomers (20, 21), indicating of protein lysates resolved in SDS-PAGE, transferred to polyvinylidene distinct routes of processing. It is suggested that an essential role for the difluoride membranes (88520, Thermo Fisher Scientific), probed with COGs is maintaining posttranslational homeostasis of either protein primary and secondary antibodies (A9917 and AP307P, Sigma), glycosylation or glycosylated proteins (22). However, it remains followed by signal detection using the ECL reagents (GERPN2019) unclear whether PD-L1 expression can be targeted selectively because from Sigma as instructed. Equal protein loading was verified by the central mechanisms controlling PD-L1 expression, for example, blotting ERK2 (tumor lines), LDHB, b-actin, mTOR, or pan AKT. IRF1-dependent transcription and posttranslational processing of PD-L1 through the ER–Golgi network, remain insufficiently under- Immunoprecipitation stood. Indeed, despite the essential role of IRF1 in PD-L1 mRNA Immnoprecipitation was performed per the manufacturer's proto- expression, little is known regarding how this mechanism is regulated col (Cell Signaling Technology). Briefly, 1 mg of protein lysate was at the PD-L1 gene promoter and whether PD-L1 transactivation can precleared with protein A magnetic beads (#73778) and was used for be dissected from other IRF1 downstream genes. As such, the goal of the COG3 and IRF1 immunoprecipitation at 1:50 antibody to lysate this study was to identify potential therapeutic agents that could inhibit dilution (v/v). Following overnight incubation at 4C, the immuno- PD-L1 while maintaining antitumor immune responses. complexes were incubated with magnetic protein A beads for 20 minutes at room temperature and then washed five times followed Materials and Methods by Western blotting. Cell culture Chromatin immunoprecipitation assay All cell lines obtained from the ATCC were curated by and redis- Chromatin immunoprecipitation (ChIP)-PCR assay was performed tributed through The University of Texas MD Anderson Cancer as described previously (25). Briefly, cells were crosslinked with 1% Center (MD Anderson, Houston, TX) Characterized Cell Line Core formaldehyde for 10 minutes and the reaction was stopped by adding from 2015 to 2019. CRF-SB, EFE184, ETN1, KLE, HOC7, UNP251, 125 mmol/L glycine. Cells were then lysed and chromatin DNA was OVCAR3, OVCAR8, and MCF7 cells were maintained in RPMI1640 fragmented using a Bioruptor Sonicator (Diagenode). The samples medium supplemented with 5% FBS and incubated at 37Cina were immunoprecipitated with 4 mg of the IRF1 (#8478, Cell Signaling humidified incubator in an atmosphere of 95% air/5% CO2. Cell line Technology) antibody overnight at 4 C and a concentration-matched identities were reauthenticated by short tandem repeat (STR) DNA rabbit IgG isotype control (#3900, Cell Signaling Technology) was used fingerprinting using the AmpF STR identifier kit according to the for mock precipitation. The protein–DNA cross-links were then manufacturer's instructions (Applied Biosystems, catalog no. reversed and the DNA purified and analyzed by quantitative real- 4322288). The STR profiles were compared with known ATCC time PCR on the ABI ViiA 7 Real-Time PCR system using the Power fingerprints; with the Cell Line Integrated Molecular Authentication SYBR Green PCR Master Mix (4368577, Thermo Fisher Scientific). database, version 0.1.200808 (23); and with the MD Anderson fin- The primers used for qPCR are as follows. Pair 1: -666 forward- gerprint database. The STR profiles matched known DNA fingerprints TCCTTAGGGTGGCAGAATATCAG, -605 reverse-CCCATCCC- or were unique. The ID8 mouse ovarian surface epithelial cells, GAGCTAC-ATCTTT; Pair 2: -103 forward-TGAAAGCTTCCGCC- obtained from Vahid Afshar-Kharghan (MD Anderson, Houston, GATTT, -53 reverse-TGCCGGG-CGTTGGA; Pair 3: þ120 forward- TX), were maintained in DMEM (high-glucose, Cellgro) supple- TGCCCACGGCCCAGTAT, þ179 reverse-GTAGAGACCCTCCG- mented with 4% FBS, 100 U/mL penicillin, 100 mg/mL streptomy- TCCTAAAGTG; Pair4: þ146 forward-GCTCGCTGGGCACTT- cin, 5 mg/mL insulin, 5 mg/mL transferrin, and 5 ng/mL sodium TAGG; þ205 reverse-TCCTCTCTCCATCCCAAAGAAA. selenite. All lines were tested for Mycoplasma every 6 months. MDA-MB-231 demonstrates reliable cell surface expression of PD-L1 IHC for flow cytometery (24). Tissues were first fixed in in 10% neutral buffered formalin for 24 hours, preserved in 70% ethanol, processed, and embedded in Antibodies paraffin. Tissue blocks were then cut into 5-mm sections on positively PD-L1 (E1L3N-#13684), CIITA (#3793), IDH1 (#8137), mTOR charged slides, deparaffinized, and rehydrated. Mouse specific PD-L1 (7C10-#2983), Ub (P4D1-#3936), p62, AKT (40D4-2920), LDHB, LC3 (#64988, 1:200) and CD8a (#98941, 1:200) antibodies from Cell (D11-#3868), ATG5 (2630), GM130 (D6B1-#12480), HSP60, VPS34 Signaling Technology, and CD3e (sc-20047, 1:100) from Santa Cruz (D9A5-4263), b-actin (8H10D10-#3700), ATG16L1 (D6D5-8089), Biotechnology were used for IHC using SignalStain reagents (#8112, BECN1 (#3738), IFIH1 (#5321), IRF1 (D5E4-#8478), IRF2 (#4943), #14746, and #8114 or #8125) per Cell Signaling Technology protocol.

AACRJournals.org Cancer Immunol Res; 8(7) July 2020 953

Liang et al.

Quantitative data were obtained with ImageJ analysis of positive cells slides were blotted sequentially with Re-Blot (Chemicon), I-Block, per random area. and a biotin blocking system (Dako), probed with primary antibodies, and incubated with biotin-conjugated secondary antibodies. Flow cytometry The signals were then amplified using a Catalyzed Signal Ampli- Cell surface PD-L1 was stained with PE-conjugated PD-L1 antibody fication kit (DakoCytomation) according to the manufacturer's (clone MIH1; catalog no. 12-5983-42) from Thermo Fisher Scientific instructions. The processed slides were scanned and quantitated for 30 minutes at 4C. Control staining was performed using man- using the Microvigene software (VigeneTech Inc.) and the quan- ufacturer recommended mouse IgG1 isotype and treatment-matched titative values of five consistently expressed proteins (p38, JNK, samples. Data were acquired using the FACSCelesta Cell Analyzer (BD ERK, mTor, GSK) as internal controls. To screen for the effects on Biosciences) and analyzed using the Flowjo software (Flowjo, LLC). PD-L1 protein levels, a total of 40 drugs or drug combinations targeting cancer pathways that are entering or likely to enter clinical RNAi trials were surveyed across established cancer cell lines and RPPA ON-TARGETplus siRNAs targeting human COG2 (L-019487-01- datasets (Supplementary Table S1). 0005), COG3 (L-013499-02-0010), ATG5 (LQ-004374-00-0005), YAP/TAZ (LQ-012200-00-0005/L-009608-00-0010), IRF3 (L- Mouse studies and in vivo tumor models 009608-00-0010), IRF1 (LQ-011704-00-0005), and IRF2 (L-011705- All animal work and protocols were supervised and approved by the 02-0010) were from Dharmacon. RNAi silencing was performed using Institutional Animal Care and Use Committee of MD Anderson Lipofectamine RNAiMAX Transfection Reagent (Thermo Fisher Sci- (Houston, TX). ID8 cells (1 106 in 100 mL PBS) were intraperito- entific) according to the manufacturer's protocol. Briefly, 20 nmol/L neally injected into C57BL/6 mice (female, 6–8 weeks old, CRL/NCI). siRNA was transfected and control cells were transfected with After transplantation, cells were allowed to grow for 1 week, and then 20 nmol/L nontargeting siRNA (Dharmacon). mice with established tumors were randomly sorted into different treatment groups with 5 mice/group. ID8-bearing mice were then Quantitative real-time PCR assay treated with isotype control IgG or PD-L1 antibody (200 mg/mouse in Total RNA was isolated using the RNeasy Plus Mini Kit (Qiagen, 100 mL PBS, B7-H1, clone 10F.9G2, BioXCell; i.p. injection) every three #74136) according the manufacturer's instructions. Briely, 5 106 days, verteporfin 60 mg/kg daily in 8% DMSO (100 mL; i.p.), BMN 673 cells were lysed and homogenized. RNA concentration was deter- 0.33 mg/kg daily in 2.64% ethanol (100 mL; oral gavage), or the mined using a NanoDrop2000c Spectrophotometer (Thermo Fisher combination of BMN 673 with either PD-L1 antibody or verteporfin. Scientific) and 2 mg RNA was used for first strand cDNA synthesis Tumor progression was monitored once a week using a Xenogen IVIS using the High Capacity cDNA Reverse Transcription kit (Invitrogen, Spectrum in vivo bioluminescence imaging system (PerkinElmer). #4368814). Real-time PCR was performed in a total volume of 20 mL Tumor volume was determined on the basis of the total flux (photons per reaction in duplicates in twin.tec real-time PCR Plates (Eppendorf, per second). Mice reaching a humane endpoint or weighing more than #0030132718), covered with Masterclear Real-Time PCR Film 35 g as a result of tumor growth and/or ascites were euthanized. (Eppendorf, #0030132947). PCR probes CD274 (Hs01125301_m1) Lewis lung carcinoma (LLC) cells (5 105 in 100 mL PBS) were and ACTB (Hs01060665_g1) from TaqMan Gene Expression Assays inoculated subcutaneously into C57BL/6 mice or nude (NU/J, The (Invitrogen) were used as instructed. Each reaction contained 40 ng of Jackson Laboratory; 8–12 weeks) in both flanks. Five days after cDNA, 10 mL of TaqMan Fast Universal PCR Master Mix (2), no inoculation, mice were randomly divided into 4 groups (n ¼ 6/group) AmpErase UNG (Invitrogen, #4366073). TaqMan reactions were run and treated with vehicle, BMN 673 0.33 mg/kg daily (oral gavage), on Mastercycler ep realplex (Eppendorf) with the following thermal verteporfin 30 mg/kg daily (i.p. injection), or the combination of cycling protocol: 95C for 2 minutes followed by 40 cycles of 95C for BMN 673 with verteporfin for 16 days. Tumors size were measured 15 seconds, 55C for 15 seconds, and 68C for 20 seconds. CD274 gene every 2 days by digital calipers to determine tumor volume using the DDCt 2 expression was quantified using the comparative Ct (2 ) method formula [length/2] [width ]. with Ct values normalized to the housekeeping gene (ACTB). Statistical analysis Cell fractionation All statistical analyses were done in GraphPad Prism 8 software. Subcellular fractionation was performed using the FractionPREP Correlations between TRIM28 and CIITA-MHC II gene expression cell fractionation kit (K270) from Biovision according to the manu- was analyzed using the linear regression test. On the basis of pilot facturer's instructions. studies of the anti–PD-1 experiment, 5 mice per group was sufficient to identify the expected effects with 90% power. Overall survival of various Cell viability assay treatment groups was analyzed using the Cox regression model, and the Cell viability was assayed using the CellTiter-Blue Cell Viability Assay log-rank test was used to determine the P values. Otherwise, unpaired t kit (G8081) from Promega according to the manufacturer's protocol. tests were used to generate two-tailed P values and P < 0.05 was considered statistically siginificant. The Cancer Genome Atlas (TCGA) Reverse-phase protein microarray analysis and high- data were analysed using cBioPortal (cbioportal.org) and PanCancer throughput drug screen Atlas datasets for each cancer types were used. Reverse-phase protein microarray analysis (RPPA) assays were performed in the MD Anderson CCSG core as described at http://www.mdanderson.org/education-and-research/resources-for- Results professionals/scientific-resources/core-facilities-and-services/functional- Identification of verteporfin as a potent inhibitor of PD-L1 proteomics-rppa-core/index.html. Briefly, serially diluted lysates were expression spotted onto FAST slides (Schleicher & Schuell BioSciences) using a Using downregulation of PD-L1 as a read-out by RPPAs to assess robotic GeneTAC arrayer (Genomic Solutions, Inc.). After printing, the drugs in use or likely to enter trials with well-characterized molecular

954 Cancer Immunol Res; 8(7) July 2020 CANCER IMMUNOLOGY RESEARCH

Verteporﬁn Inhibits PD-L1

targets, we performed a candidate drug screen (n ¼ 40) for inhibition as a quality control mechanism. However, depletion of COG3 but not of PD-L1 expression discarding low-potency, narrow-spectrum, and COG2 attenuated verteporfin-induced loss of PD-L1 (Fig. 2F), sug- upregulating candidates, which identified a single drug candidate, gesting a crucial role for COG3, but not COG2, in verteporfin-induced verteporfin. Verteporfin suppressed PD-L1 expression effectively in all PD-L1 removal. six cell lines (T-cell leukemia; B-cell leukemia; ovarian; endometrium n ¼ 3; Fig. 1A; Supplementary Table S1; Supplementary Fig. S1). In an Verteporfin inhibited IRF1-dependent PD-L1 transcription additional panel of eight human cancer cell lines (ovarian, n ¼ 5; Supporting a role for autophagy in mediating verteporfin-induced osteoblastoma, n ¼ 1; and lung cancers, n ¼ 2) and two murine cancer loss of PD-L1, RNAi-mediated knockdown of ATG5, a gene that is cell lines (ovarian and lung), verteporfin abolished basal PD-L1 protein essential for canonical autophagy (29), mitigated the effect of verte- expression, including differential glycosylated states as reflected by the porfin on PD-L1 downregulation (Fig. 3A). Despite a role for autop- double bands on Western blots (7), regardless of genetic background, hagy in downregulation of PD-L1 protein expression, verteporfin was lineage specificity, and basal (intrinsic) PD-L1 expression (Fig. 1A–D). sufficient to decrease PD-L1 expression in ATG5-depleted cells at Cell fractionation revealed that verteporfin decreased membrane- increased concentrations (1 mmol/L vs. 0.5 mmol/L). At higher drug associated PD-L1 (functionally relevant PD-L1) in EFE184 cells concentrations, verteporfin abolished IFN-induced PD-L1 expression (endometrial cancer; Fig. 1E) and flow cytometry showed that verte- in the presence of chloroquine (Fig. 3B), suggesting that chloroquine- porfin reduced PD-L1 expression on both the surface of cancer cells sensitive autophagy was dispensable for loss of PD-L1 in this setting. (Fig. 1F) and on antigen-presenting cells (Supplementary Fig. S1D). Because verteporfin treatment also resulted in significant downregula- Verteporfin suppressed both IFN-induced PD-L1 protein expression tion of PD-L1 mRNA expression (Fig. 1G), we explored mechanisms (Supplementary Fig. S1B–S1D) and mRNA expression (Fig. 1G). controlling PD-L1 gene expression. However, in contrast to the marked loss of PD-L1 protein, verteporfin Verteporfin inhibits YAP/TAZ function and YAP1 inhibits IRF3 had little effect on intrinsic PD-L1 mRNA expression in the absence of signaling (30). Thus, RNAi was performed to interrogate the roles of IFNg (Fig. 1H). Thus, verteporfin engaged at least two independent the YAP1, TAZ (WWTR1), IFIH1, and IRF3 cascade in association mechanisms to downregulate PD-L1 expression. with PD-L1 expression. Knockdown of these genes had minimal effects on both the basal expression of PD-L1 or IFN-induced PD-L1 expres- Verteporfin activated Golgi-related PD-L1 autophagy sion, with depletion of YAP1 and TAZ (WWTR1) leading to increased As verteporfin inhibits autophagy (26), we evaluated whether rather than decreased PD-L1 in the presence of IFNg (Supplementary autophagy was required for verteporfin-induced loss of PD-L1. Fig. S2). Thus, the effect of verteporfin on PD-L1 expression is unlikely Although verteporfin exhibited a modest dose-dependent growth- to be attributed to the YAP/TAZ signaling cascade. inhibitory effect on EFE184 cells, cotreatment with chloroquine, which Given the ability of verteporfin to abolish IFN-induced PD-L1 inhibits the final step of autophagy, led to an approximately 16-fold expression, we next sought to determine whether verteporfin acted increase in verteporfin-induced growth suppression and a marked loss to disrupt PD-L1 gene transcription. Both the signal transducer and of cell viability (Fig. 2A); these data were consistent with a crucial role activator of transcription 1 (STAT1) and the main transcription factor of autophagy in maintenance of viability of cells treated with verte- IRF1, downstream of the IFNg receptor, are essential for IFN-induced porfin. Transmission electron microscopy (TEM) revealed a marked PD-L1 gene expression (31). Indeed, IFN increased IRF1 expression in increase in autophagosomes in the cells treated with verteporfin, and a panel of cancer cell lines, and IRF1 protein expression directly altered Golgi apparatus with swollen and disrupted structures correlated with PD-L1 levels (Supplementary Fig. S3). IRF1 depletion (Fig. 2B). The morphologically disrupted Golgi networks were in abolished IFN-induced expression of PD-L1 (Fig. 3C), whereas IRF2 proximity to autophagosomes. Consistent with the TEM data, verte- knockdown had no effect despite a possible role of IRF2 in either porfin treatment led to progressive loss of high molecular weight opposing or complimenting that of IRF1 (32). Thus, we examined the ubiquitinated proteins in a dose-dependent manner, increased LC3 effect of verteporfin on PD-L1 mRNA in the context of IRF1 depletion. lipidation (LC3 1 and II), and loss of the selective autophagy substrate Verteporfin markedly decreased PD-L1 mRNA expression, and IRF1 and adaptor p62/SQSTM1 (Fig. 2C and D), indicating active autop- depletion had a stronger effect, whereas verteporfindidnotcause hagy consistent with a prior report (27). decreases in PD-L1 mRNA expression in the setting of IRF1 In addition to loss of p62 and PD-L1, verteporfin induced a marked depletion (Fig. 3D), suggesting that verteporfin acted via inhibition decrease in the Golgi proteins COG3, COG7, and the Golgi matrix of IRF1-dependent PD-L1 transcription. However, IFN-induced protein GM130 (Fig. 2D). Chloroquine treatment abrogated verte- STAT1 and IRF1, including their nuclear localization, were not porfin-induced loss of PD-L1, p62/SQSTM1, and Golgi proteins affected by verteporfin(Fig. 3E), suggesting intact IFN-STAT1 (Fig. 2D). These data suggested that verteporfin induced autop- signaling upstream of IRF1. hagy-mediated degradation of the Golgi apparatus, which was likely To determine where in the STAT1 pathway verteporfin might be a consequence of verteporfin-induced organelle damage. Indeed, we acting, we performed IRF1 immunoprecipitation and found that found COG3 physically associated with VPS34 (a crucial membrane whereas IFN-g induced a marked increase in IRF1–STAT1 association, component of the autophagy machinery), the autophagy adaptor verteporfin treatment led to complete disruption of the IRF1–STAT1 molecule p62, and GM130 (Fig. 2E). Next, we interrogated whether complex (Fig. 3F). We then performed ChIP assays to determine the the lobe A subunits might have a role in autophagy-dependent PD-L1 effect of verteporfin on IFN-induced IRF1-binding to PD-L1 gene removal in cancer cells. RNAi-mediated gene silencing was performed promoter regions (Fig. 3G). Indeed, IRF1-binding could not be for COG2 and COG3. The COG2 RNAi led to cross depletion of detected in unstimulated cells, but IFNg induced marked increases COG3, consistent with the finding that both COG3 and COG2 are in the binding of IRF1 to the PD-L1 promoter (Fig. 3H). Strikingly, required for stabilizing lobe A complexes (28). Nonetheless, RNAi of despite inhibiting PD-L1 expression, verteporfin treatment led to a either COG2 or COG3 increased intrinsic PD-L1 expression, with modest increase in IRF1 binding in the absence of IFN, and a marked COG2 depletion having a stronger effect (Fig. 2F), consistent with a increase was detected in the presence of IFN (Fig. 3F). The paradoxical role for the oligomeric Golgi complex in regulating PD-L1 expression increase in IRF1 binding to the PD-L1 promoter after verteporfin

AACRJournals.org Cancer Immunol Res; 8(7) July 2020 955

Liang et al.

A B C − − − H226 H358 1.2 VP + + + −VP +VP − + − + PD-L1 55KD VP 1 PD-L1 Actin 0.8 ES2 HOC1 UPN251

PD-L1 0.6 Actin D 0.4 ID8 LLC 0.2 IFNγ −−++ −−++ − − − − 0 VP + + + + PD-L1

β-Actin

E F −IFNγ +IFNγ Cytosol Membrane Isotype = 0.94 PD-L1 Isotype = 1.08 PD-L1

IFNγ −−++ −−++ VP − + − + − + − +

−VP − PD-L1 VP

CD74 Isotype = 3.77 PD-L1 Isotype = 3.14 PD-L1

IDH1

mTOR +VP +VP

G +IFNγ H 1.5 2 −IFNγ ) t ) 1.5 1 t

ΔΔ C 1

** ΔΔ C

0.5 L1 mRNA - 0.5 (fold, PD (fold, 0 0 PD-L1 mRNA level mRNA PD-L1 Ctl VP 00.51 VP (μmol/L)

Figure 1. Verteporfin decreased intrinsic and IFN-induced PD-L1 expression. A, Summarized data of PD-L1 protein expression (fold change relative to untreated baseline) as determined by Western blot analysis in human ovarian(ES2,HOC1,andUPN251),osteoblastoma(U2OS),andlung(H226andH358)cancercellstreatedwith (þ) or without () verteporfin(1mmol/L, 24 hours), as shown in B and C. Western blots for PD-L1 expression in human ovarian cancer cells (B) and lung cancer cells (C). D, Murine cancer cells treated with (þ) or without () verteporfinandIFNg. E, Western blots of PD-L1, a loading control (mTOR), membrane fractionation (CD74), and cytosolic fraction (IDH1) of EFE184 cells (endometrial carcinoma) treated with (þ) or without ()IFNg (10 ng/mL) and verteporfin (1 mmol/L, 24 hours). Data represent two or more independent experiments. F, Flow cytometry analysis of PD-L1 expression on the surface of MDA-MB-231 (breast carcinoma) cells treated with (þ) or without () verteporfin(2mmol/L) for 24 hours. Relative PD-L1 mRNA expression in EFE184 cells treated with or without verteporfin for 24 hours either in the presence (G)orabsence(H)ofIFNg (5 ng/mL). Data represent mean SD of three independent experiments. , P < 0.001 (two-tailed t test). VP, verteporfin.

956 Cancer Immunol Res; 8(7) July 2020 CANCER IMMUNOLOGY RESEARCH

Verteporﬁn Inhibits PD-L1

Figure 2. A CQ (μmol/L) VP (1X, 62.5 nmol/L) Autophagy was required for cell survival, 0 5 10 20 fi and vertepor n-induced loss of PD-L1. A, Cell 0 1 2 density after treatment with either verteporfin (0–1,000 nmol/L) or chloroquine (CQ, 0– 4816 20 mmol/L) alone or the combination of the two drugs for 6 days. B, Autophagosomes as revealed by TEM in EFE184 cells treated with B C D CQ −−+ + (þ) or without () verteporfin(1mmol/L) for VP 0 .03 .1 .3 1 3 (μmol/L) VP − + − + 24 hours. Red arrows, Golgi; yellow arrows, KD autophagosomes. Data represent mean SD PD-L1 130 from 30 cells. , P < 0.001 (two-tailed t test). 90 C, Western blots of ubiquitin (Ub) and p62/ 1 0.2 2.5 2.6 SQSTM1 (p62) in EFE184 cells treated with ver- 55 p62 −VP teporfin(0–3 mmol/L). AKT and LDHB were Ub 35 I blotted for loading control. D, Western blot LC3 24 II analyses of protein expression in EFE184 cells treated with (þ) or without () verteporfin GM130 (0.5 mmol/L) and chloroquine (10 mmol/L) for 17 96 hours. Quantification data (densitometry COG3 values normalized to lane 1) are provided below 1 0.3 1.3 0.7 respective blots. E, Western blots for proteins p62 COG7 present in COG3 immunoprecipitates from 1 0.3 1.1 0.6 EFE184 cells treated with (þ) or without () AKT HSP60 verteporfin(1mmol/L) and chloroquine +VP 35 (10 mmol/L) for 24 hours. F, Western blots for LDHB protein expression in EFE184 cells treated with F −VP +VP (þ) or without () verteporfin for 24 hours after siNT + −−+ −− transfection of nontargeting (siNT), COG2 siCOG2 − + −−+ − (siCOG2), or COG3 (siCOG3) siRNA for 48 hours. siCOG3 −−+ −−+ Densitometry data (fold change) of PD-L1 E IP COG3 expression are presented below the blot. 40 CQ −−+ + HSP60, AKT, and LDHB were blotted for loading ** PD-L1 30 VP − + − + NR-IgG Ab only control. Western blotting data represent two or more independent experiments. CQ, chloro- 20 COG3 3 quine; VP, verteporfin. 2 10 GM130 1

Autophagosomes 0 0 −VP +VP VPS34 123456

p62 GM130

p62

LDHB

treatment appeared to suggest futile recruitment of IRF1, probably as a dependent tumor immunogenicity. We next sought to understand the result of disruption of the IRF1–STAT1 protein complex. Indeed, underlying mechanism behind this. TRIM28 is a cofactor acting either additional ChIP assays showed that verteporfin decreased the recruit- to activate or repress target gene transcription (33). Although previ- ment of STAT1 to the PD-L1 promoter in either the absence or ously reported in association with IRF1 (34), the role of TRIM28 in presence of IFN (Supplementary Fig. S4). Taken together, our data regulating IRF1 signaling remains unclear. We found in parallel to suggested that verteporfin disrupts IFN-induced IRF1–STAT1 inter- STAT1, that the TRIM28–IRF1 interaction was markedly increased in action, leading to suppression of PD-L1 transcription with nonpro- response to IFN and that the interaction was abrogated by verteporfin ductive trapping of IRF1 to the PD-L1 promoter. (Fig. 3F). Strikingly, shRNA-mediated TRIM28 depletion led to marked increases in CIITA and HLA-D (Fig. 4C), suggesting TRIM28 Differential effects of verteporfin on IRF1-dependent normally acted to inhibit the CIITA-MHC II cascade in a feed-forward transcription manner that is responsive to IFN. In addition to loss of PD-L1 expression, IRF1 knockdown also led Indeed, analyses of TCGA data revealed that TRIM28 expression to a marked decrease in the CIITA transcription factor and its inversely correlated with CIITA-MHC II gene expression and CD74 in downstream CD74 protein (Fig. 4A). The MHC II cascade branch the majority of human cancer types that span immunologic reactivity of IRF1 signaling was not inhibited in verteporfin-treated cells in the (Supplementary Fig. S5A and S5B). Gene expression profiling revealed presence of IFN (Fig. 4B), suggesting that verteporfin blocked the two distinct tumor clusters in lung cancer, conforming to high and low IFN–IRF1–PD-L1 axis specifically but had little effect on IRF1-CIITA. CIITA signature classes of gene expression, respectively (Supplemen- Such pathway specification might allow the PD-L1 immune check- tary Fig. S5C), which in turn exhibited a very strong (P ¼ 9.27 10 16) point pathway to be blocked without collateral suppression of CIITA- inverse association with TRIM28 mRNA expression (Supplementary

AACRJournals.org Cancer Immunol Res; 8(7) July 2020 957

Liang et al.

A −IFNγ + IFNγ B C −− − −IFN −−− −−− VP + + +IFN siATG5 +++ +++ CQ −−−++ siNT+ −− + −− VP 00.51 0 0.5 1 0 0.5 1 0 0.5 1 (μmol/L) IFNγ − ++++ siIRF1 − + −−+ − siIRF2 −−+ −−+ ATG5 PD-L1 PD-L1 ERK2 p62 IRF1 PD-L1 IRF2 20 STAT1 10 IRF3 0 123456789101112 LDHB E Cytosol Nuclear ** D VP − + − + − + − + 45 * IFN −−++ −−++ 40 N.S. ) siNC t 35 siIRF1 N.S. IRF1 ΔΔ C 30 STAT1 25 IRF3 20 15 H2B 10 N.S.

PD-L1 mRNA(fold, LDH 5 * 0 hnRNPA1 Control VP (0.5) VP (1.0) IFN IFN IFN +VP(0.5) +VP(1.0) F IP: IRF1 Input IRF1-de H 1.8 Ctl VP −−+ −−+ −−+ NR IgG 1.6 IFN − ++ − ++ − ++ VP γ 55 1.4 IFN IRF1 1.2 IFNγ + VP 1 STAT1 100 0.8 TRIM28 70 0.6 % Input 0.4 G Exon 1 Intron 1 0.2 0 R1 R2 R3/R4 R1 R2 R3 R4 −0.2

Figure 3. IRF1 was essential for IFN-induced PD-L1, and verteporfin disrupted IRF1 complex. A, Western blots for protein expression in EFE184 cells treated with (þ) or without () verteporfin for 24 hours after transfection of nontargeting (siNT) or ATG5 (siATG5) siRNA for 48 hours. B, Western analysis of protein expression in EFE184 cells treated with (þ) or without () verteporfin(2mmol/L), chloroquine (CQ; 20 mmol/L), and IFNg (5 ng/mL) for 24 hours. C, Western blots for protein expression in EFE184 cells treated with (þ) or without ()IFNg (5 ng/mL) for 24 hours after transfection of nontargeting (siNT), IRF1 (siIRF1), or IRF2 (siIRF2) siRNA for 48 hours. D, qPCR for PD-L1 mRNA expression in EFE184 cells treated with (þ) or without () verteporfin(mmol/L) and IFNg for 24 hours after transfection with nontargeting (siNT) or IRF1 (siIRF1) siRNA for 48 hours. Data represent mean SD (n ¼ 6). , P < 0.05; , P < 0.001; N.S., not significant (two-tailed t test). E, Western blots for proteins after subcellular fractionation in EFE184 cells treated with (þ) or without () verteporfin(1mmol/L) and IFNg (5 ng/mL) for 24 hours. Western blotting data represent two or more independent experiments. F, Western blots for proteins present in IRF1 immunoprecipitates (IP: IRF1) and mock immunoprecipitation with normal rabbit IgG (NR-IgG) from EFE184 cells treated with (þ) or without ()IFNg and verteporfin for 24 hours. Blots for input and IRF1-immunodepleted (IRF1-de) lysates are shown on the right. PD-L1 gene (CD274) promoter regions (R14; G) and CHIP assays for IRF1 (H) and STAT1 (Supplementary Fig. S4) recruitment in EFE184 cells treated with or without IFNg and verteporfin for 24 hours. VP, verteporfin.

958 Cancer Immunol Res; 8(7) July 2020 CANCER IMMUNOLOGY RESEARCH

Verteporﬁn Inhibits PD-L1

A B VP −-+-+− C sh-CTL sh-TRIM28 − − IFN --++ VP -+− −-+−-+−-+ −IFN +IFN −− − − HLA-D IFN --++ --++ siNT +-−− -+−− - - 1 1 3 3.5 siIRF1 −-+--−−−+- − PD-L1 −--− +-−− -+ CD74 siIRF2 1 0.3 46 21 1 0.5 46 15 CIITA 1 1.5 14 29 TRIM28 CD74 PD-L1 1 0 7 3 ERK2 CIITA LDHA

Skin melanoma D E LUAD HLA-D TRIM28 high TRIM28 high 1 1.3 31 32 2.9 2.8 113 101 (n = 109) (n = 92) β-Actin TRIM28 low TRIM28 low (n = 178) (n = 138) Surviving Surviving

P = 0.0065 P = 0.0022

Survival (months) Survival (months) F Melanoma G LUAD H Breast CIITA high CIITA high CIITA high n n (n = 85) ( = 78) ( = 259) CIITA low CIITA low CIITA low (n = 202) (n = 152) (n = 704) Surviving Surviving Surviving

P = 0.0049 P = 0.0015 P = 0.0104

Survival (months) Survival (months) Survival (months)

I Ovary J Stomach K GBM

CIITA high CIITA high CIITA high n (n = 98) ( = 65) (n = 48) CIITA low CIITA low (n = 271) CIITA low (n = 117) (n = 88) Surviving Surviving Surviving P = 0.565 P = 0.048 P = 0.0167

Survival (months) Survival (months) Survival (months)

Figure 4. TRIM28 suppressed the CIITA-MHC II cascade downstream of interferon signaling. A, Western blots for interferon target proteins in EFE184 cells treated with (þ)or without ()IFNg (5 ng/mL) for 24 hours after transfection of nontargeting (siNT), IRF1 (siIRF1), or IRF2 (siIRF2) siRNA for 48 hours. B, Western blots for proteins in EFE184 cells treated with (þ)orwithout()IFNg (5 ng/mL) and verteporfin(VP;0.5mmol/L) for 24 hours. Quantification data (densitometry values normalized to lane 1) are provided below respective blots. Data represent two or more independent experiments. C, Western blot analyses of protein expression in control (sh-CTL) and TRIM28 knockdown (sh-TRIM28) MCF7 cells treated with (þ)orwithout() verteporfin(1mmol/L) and IFNg (5 ng/mL) for 24 hours. Quantification data (densitometry values normalized to lane 1) are provided below respective blots. Kaplan–Meier estimates of overall patient survival of melanoma (P ¼ 0.0065, n ¼ 287; D) and lung cancer (P ¼ 0.0022, n ¼ 230; E) expressing high and low expression of TRIM28 mRNA and melanoma (P ¼ 0.0049, n ¼ 287; F), lung adenocarcinoma (P ¼ 0.0015, n ¼ 230; G), breast cancer (P ¼ 0.0104, n ¼ 963; H), ovarian cancer (P ¼ 0.048, n ¼ 182; I), stomach cancer (P ¼ 0.0167, n ¼ 369; J), and glioblastoma multiforme (P ¼ 0.565, n ¼ 136; GBM; K) expressing high and low expression of CIITA mRNA from the TCGA database. P values < 0.05 were considered statistically significant as determined by log-rank tests. GBM, glioblastoma multiforme; LUAD, lung adenocarcinoma.

Fig. S5D–S5F). Thus, the role of TRIM28 in suppressing the IRF1– are considered immunologically reactive, with a demonstrated beneﬁt CIITA pathway appeared to be generalizable. to immune checkpoint inhibition (35). To determine whether the Importantly, higher TRIM28 expression was associated with poorer statistical association between TRIM28 and patient outcomes might patient outcomes in melanoma and lung cancer (Fig. 4D and E), which simply reﬂect a noncausal correlation, we analyzed the effect of a

AACRJournals.org Cancer Immunol Res; 8(7) July 2020 959

Liang et al.

downstream event, that is, the expression of CIITA. We found that the significantly poorer overall and disease-free survival in patients mRNA levels of CIITA were indeed positively associated with more with lung cancer and with decreased mRNA expression of favorable patient outcomes, with high statistical power across a broad CIITA-MHC and PD-L1 genes (Fig. 5B–G). spectrum of immunologically reactive human cancers albeit not PARP inhibitors have activity across a spectrum of cancers and necessarily in immunologically nonreactive tumors such as GBM are approved for first-line treatment of BRCA-mutant ovarian (Fig. 4F–K), suggesting that the effects of TRIM28 and CIITA on cancers (37). Notably, murine ID8 ovarian carcinoma cells exhib- patient outcomes might be mechanistically linked in specific cancer ited little sensitivity to verteporfin in cell culture and no synergism lineages. with the BMN 673 PARP inhibitor (Fig. 6A). Given these properties, we tested the in vivo therapeutic effects of verteporfin Verteporfin mediated PD-L1 blockade in syngeneic tumor monotherapy or in combination with BMN 673 in established microenvironments immune competent mice bearing ID8 cells intraperitoneally in We next hypothesized that therapeutic interventions that acti- direct comparison with PD-L1 antibody (Fig. 6B). Monotherapy vate the CIITA-MHC immunogenicity cascade, which usually also with either verteporfin or PD-L1 had a modest effect on the survival drive PD-L1 expression, would synergize with verteporfin. Because of ID8-burdened mice (Fig. 6B), consistent with previous studies PARP inhibitors synergize with immune checkpoint blockade showing these cells to have modest intrinsic immunogenici- (36), we sought to determine whether PARP inhibition might affect ty (38, 39). In contrast, the combination of verteporfin and BMN the CIITA-MHC cascade. We found that PARP1 copy number 673 improved outcome compared with either monotherapy that was increases in over 85% of lung cancers (adenocarcinomas), which equivalent to the combination of PD-L1 and BMN 673 (Fig. 6B). correlated with PARP1 mRNA expression (Fig. 5A; Pearson r ¼ Independent of BRCA mutations, PARP inhibition induces immune 0.511). Higher PARP1/2 mRNA expression was associated with responses in cancer models including that of lung cancer, which has

A B C n 2% 11% 3%84% PARP1 high (n = 68) PARP1 high ( = 68) PARP1 low (n = 162) PARP1 low (n = 162)

P = 0.0311 P = 0.0276 Survival Disease free PARP1 mRNA PARP1 (V2 RCSEM, log)

sDel Dip Gain Amp Months disease free Survival (months) PARP1 copy # (GISTIC) D E F CIITA HLA-DMA HLA-DPB1 mRNA (V2 RSEM, log)

P = 8.703 × 10−5 P = 7.35 × 10−7 P = 5.31 × 10 −7 PARP1/2: High Low High Low High Low (n = 122) (n = 108) (n = 122) (n = 108) (n = 122) (n = 108) G CD274 mRNA (V2 RSEM, log)

P = 4.47 × 10−7

Figure 5. PARP1 expression was associated with PD-L1 expression and the CIITA cascade. Correlation of PARP1 mRNA expression with putative PARP1 gene copy number (A), disease-free survival (B), and overall survival (C) of patients with lung adenocarcinoma (n ¼ 230, TCGA database). P < 0.05 as determined by log-rank tests. Enrichment analysis for the association of CIITA (D), MHC (E and F), and PD-L1 (G) expression with high and low expression of PARP1/2 mRNA expression in lung adenocarcinoma (TCGA database). P values < 0.05 were considered statistically signiﬁcant as determined by two-tailed unpaired t tests.

960 Cancer Immunol Res; 8(7) July 2020 CANCER IMMUNOLOGY RESEARCH

Verteporﬁn Inhibits PD-L1

A 1.2 ID8 B

1 100 Control Anti−PD-L1 0.8 BMN 673 VP * 0.6 50 VP + BMN 673 * Anti−PD-L1 + BMN 673 0.4 Relative viability VP

0.2 BMN 673 Percentage survival BMN 673 + VP 0 0204060 0 ABCDEF Survival (days) [Drug]

1,500 C57/BL6 C BMN 673 D Control 2.0 VP VP VP (1.0 μmol/L) + BMN 673 )

3 BMN 673* 1.5 1,000 VP + BMN 673**

1.0

OD570 500

0.5 Tumor size (mm

0.0 0 0 0.01 0.04 0.11 0.33 1.0 6 9 12 15 17 19 21 Concentration (μmol/L) Days Control E F 3,500 Nude 16 CTL VP ) 3,000 14 VP 3 BMN 673 BMN 673 2,500 12 VP + BMN 673 VP + BMN 673 10 2,000 8 1,500 6 1,000 Fold change 4 Tumor size (mm Tumor 500 2 0 0 PD-L1 CD8 8 1114182123 Days

Figure 6. Verteporfin and PARPi inhibited syngeneic ovarian cancer and lung carcinoma cell growth. A, Dose–response curves for relative viability of mouse ovarian cancer cells (ID8) treated with or without VP, BMN 673, or the combination of both for 72 hours in standard cell culture. A–F ¼ 0, 0.012, 0.04, 0.12, 0.37, and 1.1 mmol/L, respectively. Data represent mean SD (n ¼ 9). B, Kaplan–Meier survival estimates of C57/BL6 mice intraperitoneally implanted with ID8 cells (N ¼ 5). Established tumors were treated with vehicle (control), PD-L1, BMN 673, verteporfin, or the combination of BMN 673 with either PD-L1 or verteporfin. , P < 0.05 (log-rank test). C, Dose–response curves of verteporfin in combination with BMN 673 for relative viability of LLC cells. D, Growth curves of LLC in C57BL/6 mice treated with or without verteporfin (30 mg/kg), BMN 673, or the combination of the two drugs. , P < 0.05; , P < 0.01 (two-tailed t test). N ¼ 8 tumors per group. The experiment was repeated with similar results. E, IHC staining (summarized data) of PD-L1 and CD8A in LLC tumors treated with or without verteporfin, BMN 673, or the combination of the two drugs. Data represent average SD of 13 areas of 300 nuclear events. , P < 0.05; , P < 0.01 (two-tailed t test). The control was used as the baseline and set at 1. Supplementary Figure S6 shows representative IHC images. F, Growth curves of LLC in nude mice treated with or without verteporfin (30 mg/kg), BMN 673, or the combination of the two drugs. VP, verteporfin.

been exploited for therapeutic intervention (40, 41). Similar to ID8 combination, were not effective in immunodeficient nude mice cells, murine LLC cells exhibited limited sensitivity to verteporfin relative to control (P > 0.1 one-way ANOVA; Fig. 6F), emphasizing and no synergism with the BMN 673 PARP inhibitor in vitro the role of immune responses in the drug action. The difference in (Fig. 6C). Monotherapy with either verteporfinorBMN673had tumorsizeonday21oftheverteporfin-treated mice was not a modest effect on tumor growth in a syngeneic setting, whereas the statistically significant. combination of verteporfin and BMN 673 produced a significant reduction in tumor volume (Fig. 6D). In LLC tumors, verteporfin PARP inhibition induced PD-L1 expression on high endothelial treatment led to marked decreases in PD-L1 expression and venules þ increases in CD8 T cells, especially in the vereporfin and BMN 673 While T-cell infiltration was consistent with the overall outcome of treatment group (Fig. 6E; Supplementary Fig. S6A and S6B). In PD-L1 suppression, we observed that PD-L1 expression was highly contrast, verteporfin and BMN 673, either administrated alone or in heterogeneous in the LLC tumor microenvironment. The strongest

AACRJournals.org Cancer Immunol Res; 8(7) July 2020 961

Liang et al.

A B Figure 7. VP −− PD-L1 expression on high endothelial venules BMN 673 BMN 673 − ++++++ and the effects of PARPi and verteporfin. A PD-L1 PD-L1 and B, Western blots for CIITA and PD-L1 protein expression in SVEC4.10 (murine lymphatic endothelial) cells treated with BMN CIITA CIITA 673 (0–100 nmol/L for 96 hours; A)orBMN 673 (100 nmol/L for 96 hours) and vertepor- β-Actin Actin fin(0–2 mmol/L for 48 hours; B). Data represent two or more independent experiments. C and D, Immunofluorescent staining D for MECA-79, a selective marker for HEVs, C Control BMN 673 25 MECA-79 ** and PD-L1 in syngeneic LLC tumor treated MECA-79 ** without () or with (þ)BMN 673. Scale bars, PD-L1 20 20 mm. E, Schematic of the mechanism underlying the synergy between PARP inhibition and verteporfin-mediated PD-L1 block- 15 ade. PARP inhibition promoted HEV neogenesis and PD-L1 expression, including on HEVs, whereas verteporfin abrogated PD-L1 PD-L1 10 expression through Golgi-related autophagy and suppression of IRF1-dependent transcription. The combination of PARP inhibition Cells per field 5 and PD-L1 blockade produced a therapeutic synergy. VP, verteporfin.

Merge 0 CTL BMN-79

PD-L1 signal was detected in the tumor vasculature in BMN 673– drug doses (i.e., ≥1 mmol/L; Fig. 7B). Exvivoanalysis of tumors treated tumors (Supplementary Fig. S6A). Although PARP inhibition stained for the selective HEV marker MECA-79 formed poorly did not affect tumor cell expression of PD-L1 (Supplementary developed vascular-like structures in control tumors but highly Table S1). We then sought to determine whether tumor-associated organized structures were detected in BMN-673–treated tumors endothelial cells were the sources of PD-L1 expression in the BMN with strong luminal MECA-79 staining (Fig. 7C). Strikingly, despite 673-treated tumors. Although PD-L1 expression appeared to be a weak coexpression of PD-L1 and MECA-79 in control tumors, associated with LYVE1-positive lymphatic vasculature (Supplemen- strong PD-L1 expression was detected in MECA-79–positive cells in tary Fig. S6B), closer examination only revealed a relatively weak BMN 673–treated tumors (Fig. 7D), suggesting that PARP inhibi- expression of PD-L1 in LYVE1-positive cells. tion induced PD-L1 expression in HEVs. PD-L1 expression on the Because high endothelial venules (HEV), which are specialized HEVs thus inhibited effector and proliferating T cells as they exited post-capillary structures, are the only known routes of tumor from the peripheral circulation into the tumor microenvironment infiltration of lymphocytes (42, 43), we tested the hypothesis that (Fig. 7E). PARP inhibition drives PD-L1 expression in HEVs. PARP inhibition with BMN 673 on lymphatic endothelial cells led to increased proteinexpressionofCIITAandPD-L1(Fig. 7A). Whereas PARPi- Discussion induced PD-L1 expression was suppressed by verteporfin, CIITA High-throughput screening identified verteporfin as a lead candi- expression was not significantly altered even at supraphysiological date for inhibiting PD-L1, suggesting the potential repurposing of the

962 Cancer Immunol Res; 8(7) July 2020 CANCER IMMUNOLOGY RESEARCH

Verteporﬁn Inhibits PD-L1

FDA-approved ophthalmological drug verteporfin for oncologic ther- in the tumors (40). We found a strong inverse correlation between the apeutic benefit. Verteporfin is a photosensitizer commonly prescribed mRNA expression of PARP1/2 and the proinflammatory CIITA- for retinopathy but also has antitumor properties in various experi- MHC cascade genes in multiple human cancers. Thus, PARP inhibi- mental settings (44, 45). Verteporfin possesses intrinsic antitumor tion was likely to activate the CIITA-MHC cascade, immunogenicity, activities through targeting YAP1 and other mechanisms (44, 45). and T-cell priming through multiple pathways, which consequently Given the patient safety profile and relative low cost ($1700 per elicited immune responses and adaptive PD-L1 expression. injection, wholesale), verteporfin could be considered as an alternative Despite the strong induction of both HEV neogenesis and PD-L1 in approach to PD-L1 blockade. PD-L1 blockade has therapeutic benefit response to PARP inhibition, different mechanisms might be involved in non–small cell lung carcinoma in a randomized phase III study (46). in these processes. In addition to DNA repair, PARP1 regulates Notably, the patients with lung cancer with the highest expression of gene expression through multiple mechanisms (58) and can act as PD-L1 derived the greatest benefit from PD-L1. PARP inhibitors, such a cofactor of multiple transcriptional regulators including NFkB and as BMN 673, upregulated PD-L1 on HEVs and the combination of TNFa (59, 60). Indeed, the expression of a number of NFkB- PARP inhibitors and PD-L1 blockade increased the therapeutic effects dependent genes are impaired in PARP1 / cells as well as in response in vivo. Our preclinical data may provide a strong rationale for the to PARP inhibition (61). Whereas the lymphotoxin–b-LTBR pathway combinatorial use of verteporfin and a PARP inhibitor in patients with is essential for HEV neogenesis, the signaling of the closely related lung cancer, especially given the high frequency of PARP1 gain in lung cytokine TNFa has the opposite role in experimental melanoma (62). cancer. In many other cancers such as GBM, PD-L1 is low (47) and the Future studies will be directed at ascertaining the applicability of lack of immune effector responses in the tumor microenvironment modulating HEV PD-L1 expression across cancer lineages and its may not necessarily benefit from this strategy. relevance as a biomarker for response to PD-L1 strategies. Our study revealed two unique mechanisms of action of verteporfin Enrichment markers for response to PD-L1 strategies have not had on PD-L1 expression; an autophagy-dependent mechanism mediated high predictive value for therapeutic response, which is likely the by the oligomeric Golgi complex of the COG proteins (specifically scenario for responses to verteporfin as it may also have direct COG2 and COG3) for the removal of intrinsically expressed PD-L1 antitumor activities through targeting YAP1 (44, 45). At the dosage and suppression of IFNg-induced PD-L1 expression. This selective range that abrogated PD-L1 expression used in our study verteporfin downregulation of PD-L1 with verteporfin maintained the proinflam- had little effect on either YAP1 expression or its phosphorylation, and matory IRF1–CIITA–MHC II signaling cascade. IRF1 is a key regu- exhibited minimal cell-autonomous cytotoxicity on cancer cells. Ver- latory factor for IFNg, TLR, and TNF (48–50); thus, IRF1 plays an teporfin in combination with PARP inhibition produced a strong important role in inflammatory responses. Indeed, IRF1 knockout in synergistic effect in syngeneic models of both ovarian and lung cancers mice leads to a multifaceted defect in immune surveillance (51). but not in cell culture, suggesting a therapeutic efficacy independent of Depending on the specific target gene, IRF1-mediated transactivation cell-autonomous anticancer activity. Nevertheless, verteporfin may can occur with or without its cofactor STAT1 (52). In the case of PD-L1 prove uniquely effective in cancers that rely on oncogenic signals from expression, both IRF1 and STAT1 are essential (4). We showed that YAP1. In addition, the striking coexpression pattern of MECA-79 and verteporfin disrupted the IRF1–STAT1 interaction, thereby suppres- PD-L1 suggested that HEV staining could be considered as a potential sing IFNg-induced PD-L1 transcription, but it increased IRF1 binding biomarker for PD-L1 blockade, especially with a combination of PARP to the PD-L1 promoter in response to IFNg. It remains unclear as to inhibitors and verteporfin within the context of a clinical trial. whether or not this promoter-trapping effect is required for suppression of PD-L1 transcription in addition to loss of STAT1 binding. Disclosure of Potential Conflicts of Interest Supporting this concept, IRF1 is involved in target gene silencing (48). H. Liang is a consultant/advisory board member for and has ownership interest Our study also unveiled a previously unrecognized role for TRIM28 in (including patents) in Precision Scientific. G.B. Mills is a scientific advisory board inhibiting the IRF1–CIITA–MHC II axis but not PD-L1 expression. member for AstraZeneca, Chrysalis, ImmunoMET, Ionis, Lilly, PDX Pharma, fi Signalchem Lifesciences, Symphogen, and Tarveda and has ownership interest Decommissioning of this role of TRIM28 upon vertepor n treatment (including patents) in ImmunoMET, Signalchem Lifesciences, Tarveda, Catena provided another layer of selectivity, allowing precision targeting of the Pharmaceuticals, and Spindletop Ventures. No potential conflicts of interest were IRF1–PD-L1 axis. More studies are needed to determine whether disclosed by the other authors. TRIM28 contributes to the development of resistance to adaptive immune surveillance in human cancer and whether TRIM28 repre- Disclaimer sents a therapeutic target for reversing the immunoediting process The funding agencies had no role in the data analysis, interpretation of the results, associated with suppression of the CIITA-MHC genes (53). or writing of this article. We also found that the PARP inhibitors induced PD-L1 expression on specialized tumor-associated endothelial structures known as the Authors’ Contributions HEVs, the “gateway” for tumor-infiltrating lymphocytes (43). The Conception and design: J. Liang, L. Wang, C. Wang, Y. Lu, J. Chen, G.B. Mills, presence of HEVs correlate with more favorable patient outcomes in A.B. Heimberger Development of methodology: J. Liang, L. Wang, C. Kassab, K.J. Jeong, Y. Lu multiple cancer types (54). As shown in other tumor-associated Acquisition of data (provided animals, acquired and managed patients, provided endothelial cells (55, 56), HEV expression of PD-L1 was probably facilities, etc.): J. Liang, L. Wang, J. Shen, B. Su, D. Fang, C. Kassab, Y. Lu, Z. Zhou, modulating the effector and proliferative function of the T cells as they C. Lu, Z.-X. Xu, Q. Yu, S. Shao, X. Chen, M. Gao move from circulation into the tumor microenvironment. This mech- Analysis and interpretation of data (e.g., statistical analysis, biostatistics, anism would be in addition to the previously documented role computational analysis): J. Liang, L. Wang, J. Shen, C. Kassab, K.J. Jeong, of PARP inhibition, the accumulation of double-stranded DNA breaks W. Zhao, Y. Lu, A.K. Jain, H. Liang, S.-C. Sun, X. Chen, F.X. Claret, Z. Ding, P. Chen Writing, review, and/or revision of the manuscript: J. Liang, L. Wang, D. Fang, in cancer cells with increases in mutation loads or the presence of C. Kassab, K.J. Jeong, S.-C. Sun, M.C. Barton, G. Peng, G.B. Mills, A.B. Heimberger damaged chromosomal DNA in the cytoplasm triggering the cGAS- Administrative, technical, or material support (i.e., reporting or organizing data, innate immune response pathway (57). STING pathway activation in constructing databases): L. Wang, Y. Lu, M. Gao, J. Chen, P. Chen, G.B. Mills, this context also likely drives the robust T-cell infiltration we observed A.B. Heimberger

AACRJournals.org Cancer Immunol Res; 8(7) July 2020 963

Liang et al.

Study supervision: J. Chen, M.C. Barton, G. Peng, G.B. Mills, A.B. Heimberger Brockman Foundation (to A.B. Heimberger), and the Shanghai Pujiang Program Other (dosing the animals for the study): A.L. Marisetty (17PJ1401400; to C. Wang). This research was performed in the Flow Cytometry Other (provided RPPA analysis on this study): Y. Lu, Q. Yu & Cellular Imaging Facility, which is supported in part by the NIH through MD Other (provided the plasmid DNA materials and expertise in data interpretation): Anderson's Cancer Center Support Grant CA016672. A.K. Jain The costs of publication of this article were defrayed in part by the payment of page Acknowledgments charges. This article must therefore be hereby marked advertisement in accordance The authors thank Kenneth Dunner Jr at the High Resolution Electron with 18 U.S.C. Section 1734 solely to indicate this fact. Microscopy Facility for assistance in performing electron microscopy; David M. Wildrick, PhD, for editorial assistance; and Audria Patrick for assisting in manuscript preparation. This project was supported by the Gynecologic SPORE Received March 6, 2019; revised October 7, 2019; accepted April 2, 2020; (5P50CA098258, NIH/NCI; to G.B. Mills), the Provost Retention Fund and the published ﬁrst April 7, 2020.

References 1. Dong H, Zhu G, Tamada K, Chen L. B7-H1, a third member of the B7 family, co- 21. Paulsson KM, Kleijmeer MJ, Griffith J, Jevon M, Chen S, Anderson PO, et al. stimulates T-cell proliferation and interleukin-10 secretion. Nat Med 1999;5: Association of tapasin and COPI provides a mechanism for the retrograde 1365–9. transport of major histocompatibility complex (MHC) class I molecules from the 2. Drake CG, Jaffee E, Pardoll DM. Mechanisms of immune evasion by tumors. Golgi complex to the endoplasmic reticulum. J Biol Chem 2002;277:18266–71. Adv Immunol 2006;90:51–81. 22. Foulquier F, Vasile E, Schollen E, Callewaert N, Raemaekers T, Quelhas D, et al. 3. Sharma P, Allison JP. The future of immune checkpoint therapy. Science 2015; Conserved oligomeric Golgi complex subunit 1 deficiency reveals a previously 348:56–61. uncharacterized congenital disorder of glycosylation type II. Proc Natl Acad Sci 4. Garcia-Diaz A, Shin DS, Moreno BH, Saco J, Escuin-Ordinas H, Rodriguez GA, U S A 2006;103:3764–9. et al. Interferon receptor signaling pathways regulating PD-L1 and PD-L2 23. Romano P, Manniello A, Aresu O, Armendo M, Cesaro M, Parodi B. Cell expression. Cell Rep 2017;19:1189–201. Line Data Base: structure and recent improvements towards molecular 5. Kataoka K, Shiraishi Y, Takeda Y, Sakata S, Matsumoto M, Nagano S, et al. authentication of human cell lines. Nucleic Acids Res 2009;37(Database Aberrant PD-L1 expression through 30-UTR disruption in multiple cancers. issue):D925–32. Nature 2016;534:402–6. 24. Guan H, Lan Y, Wan Y, Wang Q, Wang C, Xu L, et al. PD-L1 mediated the 6. Lim SO, Li CW, Xia W, Cha JH, Chan LC, Wu Y, et al. Deubiquitination and differentiation of tumor-infiltrating CD19(þ) B lymphocytes and T cells in stabilization of PD-L1 by CSN5. Cancer Cell 2016;30:925–39. Invasive breast cancer. Oncoimmunology 2016;5:e1075112. 7. Li CW, Lim SO, Xia W, Lee HH, Chan LC, Kuo CW, et al. Glycosylation and 25. Li Y, Wen H, Xi Y, Tanaka K, Wang H, Peng D, et al. AF9 YEATS domain links stabilization of programmed death ligand-1 suppresses T-cell activity. histone acetylation to DOT1L-mediated H3K79 methylation. Cell 2014;159: Nat Commun 2016;7:12632. 558–71. 8. Andrews A. Treating with checkpoint inhibitors-figure $1 million per patient. 26. Donohue E, Tovey A, Vogl AW, Arns S, Sternberg E, Young RN, et al. Inhibition Am Health Drug Benefits 2015;8:9. of autophagosome formation by the benzoporphyrin derivative verteporfin. 9. Naidoo J, Page DB, Li BT, Connell LC, Schindler K, Lacouture ME, et al. J Biol Chem 2011;286:7290–300. Toxicities of the anti-PD-1 and anti-PD-L1 immune checkpoint antibodies. 27. Tanida I, Minematsu-Ikeguchi N, Ueno T, Kominami E. Lysosomal turnover, Ann Oncol 2015;26:2375–91. but not a cellular level, of endogenous LC3 is a marker for autophagy. Autophagy 10. Der SD, Zhou A, Williams BR, Silverman RH. Identification of genes differen- 2005;1:84–91. tially regulated by interferon alpha, beta, or gamma using oligonucleotide arrays. 28. Blackburn JB, Lupashin VV. Creating knockouts of conserved oligomeric golgi Proc Natl Acad Sci U S A 1998;95:15623–8. complex subunits using CRISPR-mediated gene editing paired with a selection 11. Geng J, Klionsky DJ. The Golgi as a potential membrane source for autophagy. strategy based on glycosylation defects associated with impaired COG complex Autophagy 2010;6:950–1. function. Methods Mol Biol 2016;1496:145–61. 12. Ge L, Wilz L, Schekman R. Biogenesis of autophagosomal precursors for LC3 29. Codogno P, Mehrpour M, Proikas-Cezanne T. Canonical and non-canonical lipidation from the ER-Golgi intermediate compartment. Autophagy 2015;11: autophagy: variations on a common theme of self-eating? Nat Rev Mol Cell Biol 2372–4. 2011;13:7–12. 13. Hsu JM, Li CW, Lai YJ, Hung MC. Posttranslational modifications of PD-L1 and 30. Zhang Q, Meng F, Chen S, Plouffe SW, Wu S, Liu S, et al. Hippo signalling their applications in cancer therapy. Cancer Res 2018;78:6349–53. governs cytosolic nucleic acid sensing through YAP/TAZ-mediated TBK1 14. Blackburn JB, D'Souza Z, Lupashin VV. Maintaining order: COG complex blockade. Nat Cell Biol 2017;19:362–74. controls golgi trafficking, processing, and sorting. FEBS Lett 2019;593: 31. Lee SJ, Jang BC, Lee SW, Yang YI, Suh SI, Park YM, et al. Interferon regulatory 2466–87. factor-1 is prerequisite to the constitutive expression and IFN-gamma-induced 15. Bailey Blackburn J, Pokrovskaya I, Fisher P, Ungar D, Lupashin VV. COG upregulation of B7-H1 (CD274). FEBS Lett 2006;580:755–62. complex complexities: detailed characterization of a complete set of HEK293T 32. Harada H, Fujita T, Miyamoto M, Kimura Y, Maruyama M, Furia A, et al. cells lacking individual COG subunits. Front Cell Dev Biol 2016;4:23. Structurally similar but functionally distinct factors, IRF-1 and IRF-2, bind to 16. Whyte JR, Munro S.The Sec34/35 Golgi transport complex is related to the the same regulatory elements of IFN and IFN-inducible genes. Cell 1989;58: exocyst, defining a family of complexes involved in multiple steps of membrane 729–39. traffic. Dev Cell 2001;1:527–37. 33. Li J, Xi Y, Li W, McCarthy RL, Stratton SA, Zou W, et al. TRIM28 interacts with 17. Fotso P, Koryakina Y, Pavliv O, Tsiomenko AB, Lupashin VV. Cog1p plays a EZH2 and SWI/SNF to activate genes that promote mammosphere formation. central role in the organization of the yeast conserved oligomeric Golgi complex. Oncogene 2017;36:2991–3001. J Biol Chem 2005;280:27613–23. 34. Narayan V, Pion E, Landre V, Muller P, Ball KL. Docking-dependent ubiqui- 18. Ungar D, Oka T, Brittle EE, Vasile E, Lupashin VV, Chatterton JE, et al. tination of the interferon regulatory factor-1 tumor suppressor protein by the Characterization of a mammalian Golgi-localized protein complex, COG, ubiquitin ligase CHIP. J Biol Chem 2011;286:607–19. that is required for normal Golgi morphology and function. J Cell Biol 2002; 35. Brahmer JR, Tykodi SS, Chow LQ, Hwu WJ, Topalian SL, Hwu P, et al. Safety and 157:405–15. activity of anti-PD-L1 antibody in patients with advanced cancer. N Engl J Med 19. Ungar D, Oka T, Vasile E, Krieger M, Hughson FM. Subunit architecture of the 2012;366:2455–65. conserved oligomeric Golgi complex. J Biol Chem 2005;280:32729–35. 36. Higuchi T, Flies DB, Marjon NA, Mantia-Smaldone G, Ronner L, Gimotty 20. Huttlin EL, Ting L, Bruckner RJ, Gebreab F, Gygi MP, Szpyt J, et al. The PA, et al. CTLA-4 blockade synergizes therapeutically with PARP inhibi- BioPlex network: a systematic exploration of the human interactome. Cell tion in BRCA1-deficient ovarian cancer. Cancer Immunol Res 2015;3: 2015;162:425–40. 1257–68.

964 Cancer Immunol Res; 8(7) July 2020 CANCER IMMUNOLOGY RESEARCH

Verteporﬁn Inhibits PD-L1

37. Moore K, Colombo N, Scambia G, Kim BG, Oaknin A, Friedlander M, et al. regulatory factor-1 (IRF-1) in IL-27 gene expression. Eur J Immunol 2007;37: Maintenance olaparib in patients with newly diagnosed advanced ovarian 3499–508. cancer. N Engl J Med 2018;379:2495–505. 50. Yarilina A, Park-Min KH, Antoniv T, Hu X, Ivashkiv LB. TNF activates an IRF1- 38. Buckanovich RJ, Facciabene A, Kim S, Benencia F, Sasaroli D, Balint K, et al. dependent autocrine loop leading to sustained expression of chemokines and Endothelin B receptor mediates the endothelial barrier to T cell homing to STAT1-dependent type I interferon-response genes. Nat Immunol 2008;9:378–87. tumors and disables immune therapy. Nat Med 2008;14:28–36. 51. Matsuyama T, Kimura T, Kitagawa M, Pfeffer K, Kawakami T, Watanabe N, et al. 39. Duraiswamy J, Freeman GJ, Coukos G. Therapeutic PD-1 pathway blockade Targeted disruption of IRF-1 or IRF-2 results in abnormal type I IFN gene augments with other modalities of immunotherapy T-cell function to prevent induction and aberrant lymphocyte development. Cell 1993;75:83–97. immune decline in ovarian cancer. Cancer Res 2013;73:6900–12. 52. Ramana CV, Gil MP, Schreiber RD, Stark GR. Stat1-dependent and -indepen- 40. Shen J, Zhao W, Ju Z, Wang L, Peng Y, Labrie M, et al. PARPi triggers the dent pathways in IFN-gamma-dependent signaling. Trends Immunol 2002;23: STING-dependent immune response and enhances the therapeutic efficacy 96–101. of immune checkpoint blockade independent of BRCAness. Cancer Res 2019; 53. Chikuma S, Suita N, Okazaki IM, Shibayama S, Honjo T. TRIM28 prevents 79:311–9. autoinflammatory T cell development in vivo. Nat Immunol 2012;13: 41. Chabanon RM, Muirhead G, Krastev DB, Adam J, Morel D, Garrido M, et al. 596–603. PARP inhibition enhances tumor cell-intrinsic immunity in ERCC1-deficient 54. Wirsing AM, Ervik IK, Seppola M, Uhlin-Hansen L, Steigen SE, Hadler-Olsen E. non-small cell lung cancer. J Clin Invest 2019;129:1211–28. Presence of high-endothelial venules correlates with a favorable immune 42. Martinet L, Garrido I, Girard JP. Tumor high endothelial venules (HEVs) predict microenvironment in oral squamous cell carcinoma. Mod Pathol 2018;31: lymphocyte infiltration and favorable prognosis in breast cancer. Oncoimmu- 910–22. nology 2012;1:789–90. 55. Dieterich LC, Ikenberg K, Cetintas T, Kapaklikaya K, Hutmacher C, Detmar M. 43. Martinet L, Le Guellec S, Filleron T, Lamant L, Meyer N, Rochaix P, et al. High Tumor-associated lymphatic vessels upregulate PDL1 to inhibit T-cell activa- endothelial venules (HEVs) in human melanoma lesions: Major gateways for tion. Front Immunol 2017;8:66. tumor-infiltrating lymphocytes. Oncoimmunology 2012;1:829–39. 56. Hida K, Hida Y, Amin DN, Flint AF, Panigrahy D, Morton CC, et al. Tumor- 44. Liu-Chittenden Y, Huang B, Shim JS, Chen Q, Lee SJ, Anders RA, et al. Genetic associated endothelial cells with cytogenetic abnormalities. Cancer Res 2004;64: and pharmacological disruption of the TEAD-YAP complex suppresses the 8249–55. oncogenic activity of YAP. Genes Dev 2012;26:1300–5. 57. Chen Q, Sun L, Chen ZJ. Regulation and function of the cGAS-STING pathway 45. Zhang H, Ramakrishnan SK, Triner D, Centofanti B, Maitra D, Gyorffy B, et al. of cytosolic DNA sensing. Nat Immunol 2016;17:1142–9. Tumor-selective proteotoxicity of verteporfin inhibits colon cancer progression 58. Feng FY, de Bono JS, Rubin MA, Knudsen KE. Chromatin to clinic: the molecular independently of YAP1. Sci Signal 2015;8:ra98. rationale for PARP1 inhibitor function. Mol Cell 2015;58:925–34. 46. Rittmeyer A, Barlesi F, Waterkamp D, Park K, Ciardiello F, von Pawel J, et al. 59. Hassa PO, Hottiger MO. The functional role of poly(ADP-ribose)polymerase Atezolizumab versus docetaxel in patients with previously treated non-small-cell 1 as novel coactivator of NF-kappaB in inflammatory disorders. Cell Mol Life lung cancer (OAK): a phase 3, open-label, multicentre randomised controlled Sci 2002;59:1534–53. trial. Lancet 2017;389:255–65. 60. Amir RE, Haecker H, Karin M, Ciechanover A. Mechanism of processing of the 47. NduomEK,WeiJ,YaghiNK,HuangN,KongLY,GabrusiewiczK,etal.PD- NF-kappa B2 p100 precursor: identification of the specific polyubiquitin chain- L1 expression and prognostic impact in glioblastoma. Neuro-Oncol 2016;18: anchoring lysine residue and analysis of the role of NEDD8-modification on the 195–205. SCF(beta-TrCP) ubiquitin ligase. Oncogene 2004;23:2540–7. 48. Fujita T, Sakakibara J, Sudo Y, Miyamoto M, Kimura Y, Taniguchi T. 61. Petrilli V, Herceg Z, Hassa PO, Patel NS, Di Paola R, Cortes U, et al. Noncleavable Evidence for a nuclear factor(s), IRF-1, mediating induction and silencing poly(ADP-ribose) polymerase-1 regulates the inflammation response in mice. properties to human IFN-beta gene regulatory elements. EMBO J 1988;7: J Clin Invest 2004;114:1072–81. 3397–405. 62. Bertrand F, Rochotte J, Colacios C, Montfort A, Tilkin-Mariame AF, Touriol C, 49. Remoli ME, Gafa V, Giacomini E, Severa M, Lande R, Coccia EM. IFN-beta et al. Blocking tumor necrosis factor alpha enhances cd8 t-cell-dependent modulates the response to TLR stimulation in human DC: involvement of IFN immunity in experimental melanoma. Cancer Res 2015;75:2619–28.

AACRJournals.org Cancer Immunol Res; 8(7) July 2020 965

Verteporfin Inhibits PD-L1 through Autophagy and the STAT1−IRF1 −TRIM28 Signaling Axis, Exerting Antitumor Efficacy

Jiyong Liang, Lulu Wang, Chao Wang, et al.

Cancer Immunol Res 2020;8:952-965. Published OnlineFirst April 7, 2020.

Updated version Access the most recent version of this article at: doi:10.1158/2326-6066.CIR-19-0159

Supplementary Access the most recent supplemental material at: Material http://cancerimmunolres.aacrjournals.org/content/suppl/2020/04/07/2326-6066.CIR-19-0159.DC1

Cited articles This article cites 62 articles, 17 of which you can access for free at: http://cancerimmunolres.aacrjournals.org/content/8/7/952.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://cancerimmunolres.aacrjournals.org/content/8/7/952.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department Subscriptions at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerimmunolres.aacrjournals.org/content/8/7/952. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.