LIN28/Let-7/PD-L1 Pathway As a Target for Cancer Immunotherapy Yanlian Chen1,2, Chen Xie1, Xiaohui Zheng1,3, Xin Nie1, Zining Wang2, Haiying Liu1, and Yong Zhao1,2

Published OnlineFirst January 16, 2019; DOI: 10.1158/2326-6066.CIR-18-0331

Research Article Cancer Immunology Research LIN28/let-7/PD-L1 Pathway as a Target for Cancer Immunotherapy Yanlian Chen1,2, Chen Xie1, Xiaohui Zheng1,3, Xin Nie1, Zining Wang2, Haiying Liu1, and Yong Zhao1,2

Abstract

The immunocheckpoint protein PD-1/PD-L1 is considered be a strategy to prevent immune evasion of cancer cells. We a promising target for cancer immunotherapeutics. However, found that treatment with a LIN28 inhibitor, the small com- the objective response rate using antibodies that block the pound C1632, increases let-7 and suppresses PD-L1 expres- interaction between PD-1 and PD-L1 was less than 40%, and sion, leading to reactivation of antitumor immunity in vitro the mechanism underlying regulation of PD-1/PD-L1 expres- and in vivo. In addition, C1632 also displayed the capacity to sion is poorly understood. In this study, we identified the inhibit cancer cell proliferation and tumor growth in mice. miRNA let-7 that posttranscriptionally suppresses PD-L1 Altogether, these findings identified LIN28/let-7 as a target for expression. LIN28, an RNA binding protein upregulated in PD-L1–mediated immunotherapeutics and reveal the poten- most cancer cells, inhibits the biogenesis of let-7, thus pro- tial of C1632 and its derivatives as promising oncotherapeutic moting PD-L1 expression. Therefore, inhibition of LIN28 may agents.

40%). This could reflect individual variation in expression Introduction of PD-L1 and the complexity of the tumor microenviron- fi Programmed death ligand-1 (PD-L1), also known as CD274 ment (7, 8). Speci cally, poor responders may express low fi or B7-H1, is a type I transmembrane protein that interacts PD-L1 or demonstrate less frequent in ltration of tumor tissue with the T-cell inhibitory receptor programmed cell death pro- by antibodies or T cells. Thus, there is great need for new tein-1 (PD-1). The PD-1/PD-L1 complex on the cell surface therapeutic modalities and a better understanding of how promotes T-cell tolerance and minimizes autoimmune- expression of PD-L1 is regulated. fi mediated inflammation under normal physiologic conditions LIN28 is an RNA binding protein rst discovered in Caenor- (1). Thus, PD-L1 is considered to be an immune-checkpoint habditis elegans (9). In mammals, the LIN28 family includes protein. However, PD-L1 is often overexpressed in cancer cells, LIN28A and LIN28B, two protein isoforms that share similar most likely as a mechanism to escape immune surveillance, structure and function (10). LIN28 regulates many biological avoid recognition of tumor-specificantigensbyTcells,and processes, including stem cell differentiation and cell prolifer- promote tumor progression. High expression of PD-L1 is asso- ation (11, 12). LIN28 is frequently dysregulated in tumor ciated with more aggressive and malignant tumor subtypes, cells, and upregulation of LIN28 correlates with advanced higher risk of cancer mortality, and poor patient prognosis disease and poor prognosis for many cancer subtypes (13). (2). Clinical trials that tested the ability of monoclonal anti- It is well established that LIN28A/LIN28B interferes with con- – bodies against PD-L1 to improve the outcome of immunother- version of pre let-7 transcripts to mature let-7 miRNAs, thereby – apy for kidney, lung, and melanoma skin cancer had promising suppressing let-7 mediated downstream effects (14, 15). results (3–6), despite a low objective response rate (less than The let-7 family includes 12 miRNA members, which are thought to play important roles as tumor suppressors (16). Let-7 binds to the 30-UTRs of key oncogenes, including RAS and 1Key Laboratory of Gene Engineering of the Ministry of Education, School of Life MYC, and inhibits their expression (17, 18). In addition, let-7 2 Sciences, Sun Yat-sen University, Guangzhou, P.R. China. State Key Laboratory also suppresses cancer cell proliferation by inhibiting expres- of Oncology in South China, Collaborative Innovation Center for Cancer sion of cell-cycle regulators such as E2F2, CDK6, and Medicine, Sun Yat-Sen University Cancer Center, Guangzhou, Guangdong, P.R. China. 3School of Pharmaceutical Sciences, Wenzhou Medical University, CCND2 (19, 20). Therefore, let-7 is frequently downregulated Wenzhou, P.R. China. in human cancers, and reduced expression of let-7 family members is highly associated with poor prognosis for cancer Note: Supplementary data for this article are available at Cancer Immunology – Research Online (http://cancerimmunolres.aacrjournals.org/). patients (21 23). Here, we investigate the possibility that dysregulation of Y. Chen and C. Xie contributed equally to this article. LIN28/let-7 plays a role in tumor cell–driven evasion of immune Corresponding Author: Yong Zhao, Sun Yat-sen University, 135 West Xingang, surveillance. We demonstrate that let-7 interacts with the 30-UTR Guangzhou 510027, P. R. China. Phone: 86-203-994-3401; Fax: 86-203-933- and suppresses expression of PD-L1 in multiple human cancer cell 2944; E-mail: [email protected] lines, and this mechanism is in turn blocked by LIN28. However, doi: 10.1158/2326-6066.CIR-18-0331 an inhibitor of LIN28, C1632, restores let-7–mediated down- 2019 American Association for Cancer Research. regulation of PD-L1, inhibits tumor cell growth, and prevents

www.aacrjournals.org OF1

Chen et al.

evasion of immune-mediated tumor surveillance in vitro and Bioinformatic analysis in vivo. These results reveal a targetable pathway that regulates RegRNA software (online software provided by National Chiao immune surveillance and cancer cell proliferation simultaneous- Tung University, Taiwan, China. http://regrna.mbc.nctu.edu.tw/ ly, and may provide a mechanism for improving the efﬁcacy of html/prediction.html) was used to predict binding sites for let-7 in antitumor immunotherapies. PD-L1 mRNA.

Materials and Methods RNA extraction and real-time quantitative RT-PCR Total RNA was extracted using RNAiso Plus (Takara) according Cell culture and transfection to the manufacturer's instructions. cDNA was synthesized with Human 293T, Hela, MCF-7, U2OS, A549, MNNG/HOS, PrimeScript II first-strand cDNA Synthesis Kit (Takara), followed SKLU-1, and mouse TUBO and 4T1 cells were obtained from by amplification with RealStar Power SYBR Mixture (GenStar). Chinese Typical Culture Center from 2014 to 2015. MG63, miRNAs were quantified using the stem–loop method, with U6 SAOS2, and THP-1 cells were kindly provided by Zhou Son- snRNA as internal control. Specific primer sequences were gyang's lab at 2016. All the cells were verified by standardized designed as described previously (25). qPCR was performed with short tandem repeat analysis. In our lab, each new aliquot was a LightCycler 480 Real-Time PCR system (Roche). Data were passaged for less than 2 months after resuscitation, and then analyzed using the comparative Ct (2DDCt) method (26). new seed stocks were thawed. Mycoplasma was regularly exam- All experiments were performed in triplicate. qPCR primers for ined during cell culturing, and no contamination occurred mRNA detection were as follows: GAPDH-forward 50-CCCA- during this study. Cells were grown in Dulbecco's modified TGTTCGTCATGGGTGT-30; GAPDH-reverse 50-TGGTCATGAG- Eagles medium (Life Technologies) supplemented with 10% 0 0 TCCTTCCACGATA-3 ; LIN28A-forward 5 -CAACCAGCAGTTTG- fetal bovine serum, 1% penicillin/streptomycin at 37 Cin5% CAGGTGG-30; LIN28A-reverse 50-GCGGTCATGGACA GGA- CO . For gene overexpression, cells were transiently transfected 2 AGCC-30; LIN28B-forward 50-ATATTCCAGTCGATGTATTTG- with plasmids using lipofectamine 3000 (Life Technologies). TACA-30; LIN28B-reverse 50-TGGGTCTTCTTTCACTTCCTAA-30; For knockdown experiments, siRNA or negative control (siRNA PD-L1-forward 50-ACC TACTGGCATTTGCTGAAC-30; PD-L1- with scrambled sequence; GenePharma) were transfected into reverse primer 50-TCCTCCATTTCCC AATAGACA-30. cells using RNAi Max (Life Technologies) according to the manufacturer's protocol. Dual-luciferase reporter assays 293T cells (2 104) were plated in 100 mL medium in 96-well Plasmids plates. The cells were transfected with 100 ng reporter plasmid Let-7 sensor plasmid psiCHECK2-let-7 4 (#20930) and DNA and 300 ng miRNA expression plasmid DNA using PEI pBABE-hLin28B (#26358) were purchased from Addgene. The (Polyscience Inc.) and incubated for 48 hours. Transfected cells Lin28A gene was obtained by amplification of its cDNA from were harvested and assayed using the dual-luciferase reporter human 293T cells. The PCR product was then inserted into pCDH- assay kit (Promega) according to the manufacturer's instructions. CMV-MCS-EF1-copGFP-T2A-Puro vector (cat. #CD511B-1, SBI) Transfection and assays were performed in triplicate. using NovoRec PCR Cloning Kit (Novoprotein Scientific Inc.). Let- LIN28 inhibitor synthesis 7 expression vectors were kindly provided by Songshan Jiang's As shown in Scheme 1, CL285032 (N-methyl-N-[3-(3- lab (24). To deplete PD-L1 in TUBO mouse cells, lentivirus vectors methyl[1,2,4] triazolo[4,3-b]-pyridazin-6-yl)phenyl]acetamide) expressing shRNA were generated: pLVX-shRNA2-T2A-Puro and was synthesized according to a literature-reported procedure (27). PMD2.G/PSPAX2 were purchased from Oligobio Co., Ltd and Initial raw materials 1 (N-(3-Acetylphenyl)-N-methylacetamide), Addgene, respectively; shRNA sequence: GAGGTAATTTGGATA- 2 (3-methyl-4H-1,2,4-triazol-4-amine hydrochloride), and 4 AATAGTttcaagagaACTGTTTGTCCAGATTACCTC. (Bredereck reagent) were purchased from J&K Chemical Compa- ny. CL285032 was obtained by flash chromatography using Analysis of RNA-seq data MeOH/CH2Cl2 (0–15%) as a light-yellow solid: 1H NMR Annotated and preprocessed level 3 RNA-seq data were down- (CDCl3, d, ppm): 8.42 (d, J ¼ 9.6 Hz, 1H), 8.11 (br s, 2H), loaded from The Cancer Genome Atlas (TCGA) database (http:// 7.98 (d, J ¼ 9.6 Hz, 1H), 7.66 (t, J ¼ 7.8 Hz, 1H), 7.57 (d, J ¼ 6.8 cancergenome.nih.gov/). Hz, 1H), 3.24 (s, 3H), 2.78 (s, 3H), 1.85 (s, 3H); LC–MS (ESI):

Scheme 1. The synthesis of CL285032 (27).

OF2 Cancer Immunol Res; 7(3) March 2019 Cancer Immunology Research

Targeting LIN28/let-7 for PD-L1–Mediated Immunotherapy

2 282.0; Elemental analysis (calcd, found for C15H15N5O): C the formula V ¼ length width /2 was used to calculate tumor (64.04, 63.79), H (5.37, 5.42), N (24.90, 25.01), O (5.69, 5.78). volume.

Drug treatment Statistical analysis Cancer cells (1 105) were seeded in 6-well plates. C1632 was Statistical analyses were performed with GraphPad (GraphPad added to the culture medium at the indicated dose and incubated Prism) using a two-tailed Student t test. Values are presented as for 3 days. IFNg was added to a final concentration of 10 ng/mL mean standard deviation (SD). P values < 0.05 were considered 24 hours before cell harvest, after which population doubling statistically significant (, P < 0.05; , P < 0.01; , P < 0.001). In was calculated, and surface PD-L1 protein was analyzed by flow RNA-seq data analysis, the Pearson correlation coefficient, r, and P cytometry mean fluorescence intensity (MFI). values were calculated using GraphPad. P < 0.05 was considered to be statistically significant. Cell viability assay Cancer cells (5 103) were seeded in 96-well plates. DMSO or C1632 was added to the culture medium at the indicated dose and Results incubated for 3 days. Cell viability was measured with a kit (CCK- 0 Let-7 binds to the coding region and 3 -UTR of PD-L1 and 8) provided by Biotool. suppresses translation To test the idea that let-7 might regulate expression of PD-L1, Apoptosis assay transcriptome data from 1,189 breast carcinoma tissue biopsies Cell apoptosis was determined by flow cytometry and immu- were obtained from TCGA database and analyzed. The results noblot assay. Cells were detected with the Annexin V–FITC revealed a moderate but statistically significant negative correla- Apoptosis Detection Kit (KeyGen Biotech) and quantified by flow tion between the expression of PD-L1 and let-7a (r ¼0.083, P < cytometry according to the manufacturer's instructions. For 0.01) or let-7e (r ¼0.118, P < 0.001; Fig. 1A). Similar analysis of immunoblot, blots were probed with a primary antibody against transcriptome data from 1,045 lungs (Supplementary Fig. S1A) Caspase-3 (Cell Signaling Technology). b-Actin (Proteintech) was and 309 cervical carcinoma tissue biopsies (Supplementary Fig. used as a control. S1B) in TCGA database also demonstrated statistically significant negative correlations between abundance of let-7 and PD-L1 Tumor-infiltrating lymphocyte analysis and flow cytometry transcripts. These data suggest that let-7 might negatively regulate Twenty-four hours after the last injection of C1632, mice were expression of PD-L1. anesthetized, and tumors were excised. Freshly isolated primary Analysis of the PD-L1 mRNA using RegRNA revealed the tumor tissue was digested in collagenase (Sigma) and DNase presence of three putative let-7 binding sites, one in the PD-L1 (Sigma) at 37 C for 1 hour. Lymphocytes were enriched on a 0 coding region (target site 1) and two in the 3 -UTR (target sites 2 Ficoll (Solarbio) gradient for flow cytometry analysis. T cells were and 3; Fig. 1B). These predicted let-7 binding sites were cloned into stained using CD3-FITC (eBioscience), CD8-PerCP (eBioscience), a dual-luciferase reporter construct and assayed for ability to CD69-APC (eBioscience), and CD279-PE (eBioscience). Tumor interact with let-7 and alter reporter gene activity. Let-7 signifi- cells were isolated and analyzed by flow cytometry on a BD cantly downregulated the activity of all three luciferase reporter FACSAriaII cytometer. Cells were stained with CD45-APC gene constructs and that the strongest effects were observed when (eBioscience) to distinguish immune cells from tumor cells. let-7a, let-7c, and let-7e bound to target site 1 (Fig. 1C). The Armenian Hamster IgG-FITC (eBioscience), Rat IgG2a-PerCP specificity and functional effect of the binding of let-7 to putative (eBioscience), Armenian Hamster IgG-APC (eBioscience), and target sites in PD-L1 was further tested by mutating three let-7 Rat IgG2b-PE (eBioscience) antibodies were used as isotype binding sites in PD-L1 and comparing the expression of wild-type controls. Data were analyzed using FlowJo V7.6 (Tree Star). (WT) and mutant PD-L1 in the absence (EV) and presence of let- 7a, c, or e. Let-7a, c, or e suppressed expression of WT PD-L1 by Determination of IFNg and TNFa in mice serum approximately 30%, but did not suppress expression of PD-L1 Twenty-four hours after the last injection of C1632, whole with mutant variants of let-7 target sites (Fig. 1D). This confirms blood samples were collected as previously described (28). Serum that let-7 negatively regulates expression of PD-L1, and this effect was isolated, and relative concentrations of IFNg and TNFa were is mediated by let-7 binding to specific sequences in PD-L1. measured by an ELISA kit following the protocol provided by the manufacturer (Multiscience). Let-7 suppresses PD-L1 expression posttranscriptionally in Animal treatment protocol human cancer cells Wild-type BALB/c mice and BALB/c nu/nu mice (4–6-week-old, The following experiments investigated let-7 regulation of PD- female) were purchased from the Laboratory Animal Center of L1 expression in MCF-7, U2OS, and Hela cancer cells using a let-7 Sun Yat-sen University (SYSU). Experiments were conducted "sponge" as a means to inhibit endogenous let-7 in these cells. The under guidelines approved by the Institutional Animal Care and "sponge" has multiple RNA sequences complementary to the Use Committee of SYSU. TUBO- or PD-L1–depleted TUBO cells heptameric seeds of let-7 members (29). Here, the let-7 sponge (5 105)in50mL PBS were injected into the mammary fat was cotransfected into 293T cells along with a luciferase reporter pad under #3 mammary gland in both wild-type and immune- plasmid carrying the let-7 complementary sequence fused to the deficient BALB/c mice. Tumor-bearing mice were randomly divid- luciferase gene (30). In cells expressing the let-7 sponge, luciferase ed into two groups and intravenously treated with PBS or a LIN28 activity was 155% higher (2.55-fold) than control cells expressing inhibitor at a dose of 40 mg/kg for 12 days (6 injections in total). GFP (Supplementary Fig. S2A), confirming the specificity and Tumor growth was assessed every 2 days with a digital caliper, and efficiency of let-7 sponge.

www.aacrjournals.org Cancer Immunol Res; 7(3) March 2019 OF3

Chen et al.

Figure 1. Let-7 binds to the coding region and 30-UTR of PD-L1 and suppresses translation of PD-L1 mRNA. A, Inverse correlation between let-7 and PD-L1 expression detected by transcriptome analysis of 1,189 TCGA breast cancer samples. B, Predicted target sequences for let-7 in human PD-L1 mRNA. C, The luciferase reporter gene was fused to let-7 target sites 1, 2, or 3 to prepare the luciferase reporter gene constructs. Luciferase activity was quantiﬁed in the presence of the indicated let-7 miRNAs. D, Putative let-7 target sites in PD-L1 were mutated, and wild-type (PD-L1 WT) and mutant PD-L1 (PD-L1 mut) were fused to the luciferase reporter gene. Luciferase activity was quantiﬁed in the presence of the indicated let-7 miRNA. The results show luciferase activity normalized to control cells transfected with empty vector (EV). Error bars, SD from mean for three independent experiments. P values were calculated using a two-tailed Student t test (, P < 0.05; ,P< 0.01; ,P< 0.001).

Cellular PD-L1 protein increased by 1.5-fold in 293T cells L1 to normal surface expression (Fig. 3G). Conversely, overex- expressing the let-7 sponge (Supplementary Fig. S2B), demon- pression of LIN28A was able to rescue decreased PD-L1 by let-7c strating let-7–mediated suppression of PD-L1. Accordingly, (Fig. 3H). Altogether, these results indicated that LIN28 regulates expression of the let-7 sponge increased PD-L1 protein on cell PD-L1 expression in a let-7–dependent manner. surface in MCF-7, U2OS, and Hela cells (Fig. 2A and B and To exclude the possibility that the LIN28 protein directly Supplementary Fig. S2C). Overexpression of let-7 signiﬁcantly upregulates transcription of PD-L1, the conserved upstream decreased PD-L1 mRNA (Fig. 2C and D) and protein on cell region of PD-L1, including the putative primary PD-L1 promoter, surface (Fig. 2E and F and Supplementary Fig. S2D). These or fragments of this region were fused to luciferase reporter genes, results demonstrate that let-7 negatively regulates expression of and the resulting reporter gene constructs were transfected into PD-L1 at the level of translation (i.e., posttranscriptionally). It 293T cells with or without cotransfection of LIN28. The results is worth noting that different let-7 family members regulate PD- showed similar luciferase activity, in the presence or absence of L1 in different cell lines, for example, let-7a, -7c, and -7e are LIN28 (Supplementary Fig. S3C and S3D), indicating that LIN28 most active in MCF-7 cells, whereas let-7d and let-7i are most does not directly promote transcription of PD-L1. Our results active in U2OS cells. further suggested that the regulatory effect of LIN28 on PD-L1 is mediated by let-7. LIN28 downregulates let-7 and promotes expression of PD-L1 LIN28 binds to let-7 pre-miRNA and blocks its matura- Inhibition of LIN28 with C1632 decreases PD-L1 and tion (15, 31). Consistent with this, knockdown of LIN28A or inhibits cell growth LIN28B by siRNA increased let-7 (Fig. 3A) and decreased PD-L1 on Malignant tumors frequently exhibit high expression of the cell surface of U2OS, MCF-7, and Hela cells (Fig. 3B and C and LIN28, and therefore low let-7 and high PD-L1 expression, Supplementary Fig. S3A), whereas overexpression of LIN28A/B suggesting that LIN28 and/or let-7 might be a therapeutic target signiﬁcantly decreased let-7 (Fig. 3D) and increased cell-surface to decrease PD-L1. Because different let-7 family members PD-L1 (Fig. 3E and F and Supplementary Fig. S3B). However, in regulate PD-L1 in different cancer cells and all members LIN28A-overexpressed cells, overexpression of let-7c reduced PD- are regulated by LIN28, LIN28 has a greater potential as a

OF4 Cancer Immunol Res; 7(3) March 2019 Cancer Immunology Research

Targeting LIN28/let-7 for PD-L1–Mediated Immunotherapy

Figure 2. Effect of let-7 on PD-L1 expression in human cancer cells. Cultured cells were transfected with a let-7 sponge or constructs encoding individual let-7 family members. A–B, 48 hours after transfection, cell-surface expression of PD-L1 in MCF-7 cells (A) or U2OS cells (B) transfected with the let-7 sponge or control (GFP) was quantified by flow cytometry MFI. Representative histograms are shown on the left for corresponding quantitative data on the right. C–D, In MCF-7 cells (C) or U2OS cells (D) expressing let-7 miRNAs as indicated (let-7a, 7b, 7c, 7d, 7e, 7f,and7i) or control (EV), PD-L1 mRNA level was quantified by qPCR. E–F, Cell-surface expression of PD-L1 in MCF-7 cells (E) or U2OS cells (F) expressing let-7 miRNAs as indicated or control (EV) was analyzed by flow cytometry MFI. For all experiments, cells were stimulated with IFNg for 24 hours prior to analysis. Error bars, SD from mean for three independent experiments. P values were calculated using a two-tailed Student t test (, P < 0.05; ,P< 0.01; ,P< 0.001).

targetable candidate. A small-molecule inhibitor of LIN28, dose, C1632 inhibited cell proliferation without affecting cell N-Methyl-N-[3-(3-methyl[1,2,4]triazolo[4,3-b]pyridazin-6-yl) morphology (Fig. 5C and D). Both flow cytometry and immu- phenyl]acetamide (C1632), was identifiedandreportedto noblot assay revealed that C1632 treatment, even at a high dose, block LIN28's activity to interact with let-7, thus rescuing the did not significantly increase cell apoptosis (Fig. 5E–G), sug- maturation of let-7 and increasing let-7 in cancer cells (32). gesting limited cytotoxicity. The inhibitory effect of C1632 on Our results showed that C1632 decreased LIN28 mRNA in cell viability and proliferation was also tested in MCF-7 (Sup- U2OS,MCF-7,andHelacells(Fig.4A–C) and increased let-7 in plementary Fig. S6A and S6B), Hela (Supplementary Fig. S6C these human cancer cells (Fig. 4D–F) and mouse TUBO and S6D), and A549 cells (Supplementary Fig. S6E and S6F). and 4T1 cancer cells (Supplementary Fig. S4A and S4B). As The results showed that although a low concentration of C1632 expected, cell-surface expression of PD-L1 decreased in human has no or limited effect on cell viability and proliferation, (Fig. 4G–I) and mouse cancer cells (Supplementary Fig. S4C C1632 at high concentration significantly decreases cell viabil- and S4D) treated with C1632. Experiments repeated in multi- ity and suppresses cell proliferation. Consistent with previous ple human cancer cell lines including MNNG/HOS, MG63, findings that low expression of PD-L1 decreases rate of cancer SAOS2, A549, and SKLU-1 confirmed the inhibition on PD-L1 cell proliferation (33), depletion of PD-L1 by shRNA in TUBO expression by C1632 (Supplementary Fig. S5A–S5E). In addi- cells (Supplementary Fig. S7A) showed moderate but consis- tion, overexpression of LIN28 was able to rescue the inhibitory tent inhibition of cell growth (Supplementary Fig. S7B). It has effect of C1632 on PD-L1 in U2OS cells (Supplementary Fig. also been reported that increased let-7 can suppress the prolif- S5F). These results support the idea that C1632 has the poten- eration of cancer cells in a PD-L1–independent manner (19). tial as an anticancer treatment that suppresses expression of PD- We thus proposed that C1632 treatment decreases PD-L1 by L1 by decreasing LIN28 expression and blocking interaction increasing let-7 and could inhibit cancer cell growth. between LIN28 and let-7. The effect of C1632 on cancer cell growth was also investigated. C1632 suppresses growth of syngeneic mammary fat pad C1632 decreased the viability of U2OS cells at high dose (68 mg/ tumors in mice L), but not at low dose (17 mg/L; Fig. 5A and B). Low dose of The above results prompted us to examine the effect of C1632 C1632 had a limited effect on cell growth, whereas at a high on syngeneic mammary fat pad tumors in immunocompetent

www.aacrjournals.org Cancer Immunol Res; 7(3) March 2019 OF5

Chen et al.

Figure 3. Effect of LIN28 on let-7–mediated regulation of PD-L1. A–C, LIN28 was depleted by siRNA in U2OS cells (A, B) or MCF-7 cells (C). Forty-eight hours after transfection with siRNA, let-7 (let-7a, -7c, -7e) transcripts and PD-L1 protein on cell surface were measured by qPCR (A)andﬂow cytometry MFI (B, C), respectively. D–F, LIN28A, LIN28B, or GFP were overexpressed in U2OS cells (D, E) or in MCF-7 cells (F). The abundance of let-7 (let-7a, -7c, -7e) and PD-L1 protein on cell surface were determined by qPCR (D)andﬂow cytometry MFI (E, F), respectively. G, Overexpression of let-7c rescues LIN28A caused increase of PD-L1 in U2OS cells. H, Overexpression of LIN28A partially rescues the let-7c–mediated decrease of PD-L1 in U2OS cells. For all experiments, cells were stimulated with IFNg for 24 hours prior to analysis. Error bars, SD from mean of three independent experiments. P values were calculated using a two-tailed Student t test (, P < 0.05; ,P< 0.01; ,P< 0.001).

þ mice, in order to determine whether C1632 could enhance CD8 T cells were more abundant (Fig. 6D; mean percentage of endogenous antitumor activity in vivo. Recipient mice were control group ¼ 33.43 2.36; C1632-treated group ¼ 39.95 þ þ injected with 5 105 TUBO cells prior to start of treatment. Five 4.82, P ¼ 0.026, n ¼ 5) and CD8 PD1 T cells were less abundant days later, tumor-bearing mice were randomly divided into two (Fig. 6E; mean percentage of control group ¼ 52.10 6.21; groups, and treated with C1632 or PBS (control) for 12 days. C1632-treated group ¼ 29.77 16.66, P ¼ 0.023, n ¼ 5) and þ During treatment, no significant body weight loss was observed in tumor-infiltrating activated CD69 T cells were more abundant control- or C1632-treated mice (Fig. 6A). Tumor size was mea- (Fig. 6F; mean percentage of control group ¼ 9.10 1.05; C1632- sured every other day during treatment. The results showed that treated group ¼ 12.54 2.61, P ¼ 0.026, n ¼ 5) in C1632-treated average tumor size was significantly smaller in immunocompe- mice than in control mice. tent mice treated with C1632 (Fig. 6B). As expected, tumors from To further understand the inhibitory effect of C1632 on C1632-treated mice had a lower expression of PD-L1 (Fig. 6C) tumor growth, the above experiments were repeated using than tumors from control mice. In addition, tumor-infiltrating PD-L1–deficient TUBO cells (Supplementary Fig. S7A). We

OF6 Cancer Immunol Res; 7(3) March 2019 Cancer Immunology Research

Targeting LIN28/let-7 for PD-L1–Mediated Immunotherapy

Figure 4. C1632 rescues let-7 expression and suppresses PD-L1 in human cancer cells. Human cancer cells were treated with DMSO or C1632 (17 and 68 mg/L) for 3 days. Cells were stimulated with IFNg for 24 hours prior to analysis. Cells were collected for flow cytometry analysis, and RNA was isolated for qPCR. A–C, LIN28 mRNA was measured after drug treatment in U2OS (A), MCF-7 (B), and Hela cells (C).D–F, Impact of C1632 on expression of let-7 family members in U2OS (D), MCF-7 (E), and Hela cells (F). G–I, Quantification of flow cytometry MFI showing PD-L1 expression changes following drug treatment in U2OS (G), MCF-7 (H), and Hela cells (I). Error bars, SD from the mean. P values were calculated using the Student t test (, P < 0.05; ,P< 0.01; ,P< 0.001; ns: not significant). observed that depletion of PD-L1 suppressed tumor growth expression of PD-L1 on cell surface (Supplementary Fig. S8D and with average tumor size decreasing from 1,432 to 856 mm3, and S8E). Consistently, IFNg and TNFa, indicators of activated immu- C1632 treatment further reduced average tumor size to 501 mm3 nity that are primarily secreted by T cells and APCs, respective- (Fig. 6B). These results suggested that in addition to its role on ly (35, 36), are significantly increased in response to C1632 inhibition of PD-L1 in cancer cells, C1632 may have another treatment (Fig. 6G and H). function that enhances the tumor suppression. To further confirm that the immune system is required for It has been previously found that the blockade of PD-L1 in C1632 induced tumor suppression, TUBO cells with or without APCs is able to suppress tumor growth by enhancing T-cell PD-L1 were injected into immune-deficient (BALB/c nu/nu) mice. activation (34). We thus speculated that C1632 may also inhibit We observed that there is no difference in tumor size regardless of PD-L1 in APCs, thereby promoting T cell–mediated antitumor PD-L1 or C1632 treatment (Fig. 6I), further demonstrating the immunity. Indeed, our experiments showed that expression requirement of T cell–mediated immunity in tumor suppression. of PD-L1 is regulated by the LIN28/let-7 axis in THP-1 macro- Altogether, in vivo syngeneic tumor experiments suggested that phages (Supplementary Fig. S8A–S8C) and that C1632 treatment C1632 inhibits tumor growth by stimulating T cell–mediated significantly increases the abundance of let-7 and decreases the antitumor activity.

www.aacrjournals.org Cancer Immunol Res; 7(3) March 2019 OF7

Chen et al.

Figure 5. Effects of C1632 on cancer cell growth. U2OS cells were grown for 12 days in the presence of medium containing C1632 (17 and 68 mg/L) or DMSO control. A, Schematic figure showing treatment with C1632 or DMSO control; cells were passaged every 4 days. B, After 4 days, cell viability was determined by cell viability assay. C, Population doublings were calculated during 12-day treatment with C1632 or DMSO control. D, Cell morphology was observed by light microscopy on the eighth day of treatment with C1632 or DMSO control. E, Cell apoptosis was quantified by flow cytometry. Representative images show apoptotic U2OS cells after 12-day treatment with the indicated dose of C1632. F, Quantification of E. Error bars, SD from mean value. P values were calculated using the Student t test (, P < 0.01; ns, not significant). G, Immunoblot showing no increase of cleaved caspase-3 (marker of apoptosis) after 3-day treatment with C1632. Actin was used as an internal control.

OF8 Cancer Immunol Res; 7(3) March 2019 Cancer Immunology Research

Targeting LIN28/let-7 for PD-L1–Mediated Immunotherapy

Figure 6. Response of mammary fat pad tumor-bearing BALB/c mice to C1632 by tail-vein injection. Subcutaneous tumors were generated by injecting 5 105 TUBO cells with or without PD-L1 into mammary fat pad under #3 mammary gland of BALB/c mice. C1632 or PBS was delivered by intravenous tail-vein injection (n ¼ 5for the PD-L1þ group and n ¼ 3 for the PD-L1– group). Body weight and tumor size were measured every 2 days. Twenty-four hours after the last injection, all mice were sacrificed. A, Body weight of mice treated with PBS or C1632. P value was calculated using an unpaired two-sided Student t test (ns, not significant). B, Tumors were excised on day 13 from mice treated with PBS or C1632. C, Flow cytometry MFI analysis showing relative cell-surface expression of PD-L1 in tumor cells. D–F, Tumor-infiltrating T cells were isolated, and the percentage of CD8þ T cells (D), CD8þPD1þ T cells (E), and CD69þ T cells (F) was determined by flow cytometry. G–H, Analysis of IFNg (G) and TNFa (H) in blood serum from mice injected with PD-L1–depleted TUBO cells. I, TUBO cells with or without PD-L1 were injected into immune-deficient BALB/c mice. C1632 or PBS was delivered by intravenous tail-vein injection (n ¼ 3 per group). Tumor volume was measured every 2 days. Error bars, SD from the mean. P values were calculated by the Student t test (,P< 0.05; ,P< 0.01).

Discussion between PD-1 and PD-L1, especially when expression of PD-L1 is Previous studies show that PD-1/PD-L1 inhibits antitumor low (37, 38). Developing an alternative approach is thus desired. immune surveillance and that agents that disrupt PD-1/PD-L1 Although many factors have been identified to regulate the stimulate this tumor defense mechanism. For example, human- expression of PD-L1 in cancer cells (39–41), few of them can be ized antibodies that interfere with PD-1/PD-L1 significantly used as therapeutic targets due to lack of druggable molecules. In improved clinical outcomes for human cancer patients, including this report, we demonstrated that LIN28/let-7 regulates the post- patients with a wide range of malignancies. However, the transcriptional expression of PD-L1 and that C1632 is a small- response to antibody treatment was observed in fewer than molecule inhibitor of LIN28 that increases let-7, decreases PD-L1, 40% of patients, raising the possibility that antibody at the given and suppresses cell growth in a variety of human cancer cell types. dosage may not be sufficient to effectively block the interaction Results presented here also showed that C1632 suppresses

www.aacrjournals.org Cancer Immunol Res; 7(3) March 2019 OF9

Chen et al.

progression of syngeneic tumors in vivo in immunocompetent C1632, which was first developed as an anxiolytic agent (49), mice and stimulates T cell–mediated antitumor immune surveil- rescues maturation of let-7 by blocking the interaction between lance in these animals. LIN28 and let-7 (32). Although our results showed that C1632 is PD-1/PD-L1 signaling induces T-cell exhaustion and alters able to decrease the expression of PD-L1 in vitro and in vivo, it is far þ frequency of tumor-infiltrating CD8 T cells, whereas the block- from an ideal candidate for therapeutics. There is a need to ade of PD-1/PD-L1 rescues T-cell functions and reactivates anti- develop a molecule that can inhibit LIN28 and rescue let-7 with tumor immune surveillance. Our results demonstrated that much greater efficiency. Efforts to find derivatives of C1632 C1632 suppressed expression of PD-L1 in tumors, increased the or other inhibitors that block the interaction between LIN28 and þ percentage of tumor-infiltrating CD8 T cells, decreased the let-7 are warranted and would be facilitated by availability of the þ þ frequency of CD8 PD1 T cells, and increased the frequency of molecular structures of C1632 and LIN28/let-7 (50). Nonetheless, þ infiltrating activated CD69 T cells in tumor-bearing mice in vivo. this study establishes that LIN28/let-7–mediated regulation of Our results also showed that C1632 inhibited PD-L1 expression in PD-L1 is a promising anticancer target that could be exploited in THP-1 macrophages in vitro and increased TNFa and IFNg in mice developing novel anticancer therapeutics. blood, suggesting that activated APCs may further enhance the antitumor immunity. Taken together, these results suggest that Disclosure of Potential Conflicts of Interest C1632 inhibits immune evasion and stimulates T cell–mediated No potential conflicts of interest were disclosed. immune surveillance, and that it may be possible to exploit C1632 for therapeutic benefit toward human cancers. Because Authors' Contributions increased expression of LIN28 impairs T-cell development and Conception and design: Y. Chen, C. Xie, Y. Zhao leads to peripheral T-cell lymphoma (42), C1632 or other LIN28 Development of methodology: Y. Chen, C. Xie, X. Nie, Z. Wang, H. Liu, Y. Zhao Acquisition of data (provided animals, acquired and managed patients, inhibitors may be generally beneficial for T-cell maturation and provided facilities, etc.): Y. Chen, X. Zheng function. Analysis and interpretation of data (e.g., statistical analysis, biostatistics, Previous studies demonstrate frequent upregulation of LIN28 computational analysis): Y. Chen, C. Xie, X. Zheng, Y. Zhao in human tumors, and LIN28 has been proposed as a potential Writing, review, and/or revision of the manuscript: Y. Chen, C. Xie, Y. Zhao target for cancer therapeutics (42–44). LIN28 and let-7 influence a Administrative, technical, or material support (i.e., reporting or organizing wide range of cellular processes including proliferation, metab- data, constructing databases): Y. Chen Study supervision: Y. Chen, Y. Zhao olism, metastasis of cancer cells, and activation and maturation of – APCs (45 48). Our results also demonstrated the inhibitory effect Acknowledgments of C1632 on the viability and proliferation of cancer cells. This work was supported by the National Natural Science Foundation of Therefore, inhibition of LIN28 has dual functions for cancer China Grants (31571410, 31570827, 81771506, and 21701194), the National therapy: suppressing tumor growth by inhibiting cancer cell Key R&D Program of China (2018YFA0107000], and Guangzhou Municipal proliferation and stimulating antitumor immune responses. Sup- People's Livelihood Science and Technology Plan (201803010108 and pression of tumor growth may enhance the T-cell-to-tumor ratio 201604016111). and thus promotes antineoplastic immune activity, especially in tumors with reduced T-cell infiltration. Nevertheless, treatment The costs of publication of this article were defrayed in part by the payment of fi page charges. This article must therefore be hereby marked advertisement in with C1632 did not appear to signi cantly increase cell death accordance with 18 U.S.C. Section 1734 solely to indicate this fact. (apoptosis). It remains to be determined whether C1632 or other inhibitors of LIN28 are well tolerated at therapeutically effective Received May 17, 2018; revised October 15, 2018; accepted January 10, 2019; doses against cancer in humans. published first January 16, 2019.

References 1. Keir ME, Butte MJ, Freeman GJ, Sharpe AH. PD-1 and its ligands in 9. Horvitz HR, Sulston JE. Isolation and genetic characterization of cell- tolerance and immunity. Annu Rev Immunol 2008;26:677–704. lineage mutants of the nematode Caenorhabditis elegans. Genetics 2. Hamanishi J, Mandai M, Iwasaki M, Okazaki T, Tanaka Y, Yamaguchi K, 1980;96:435–54. et al. Programmed cell death 1 ligand 1 and tumor-inﬁltrating CD8þ T 10. Jiang S, Baltimore D. RNA-binding protein Lin28 in cancer and immunity. lymphocytes are prognostic factors of human ovarian cancer. Proc Natl Cancer Lett 2016;375:108–13. Acad Sci USA 2007;104:3360–5. 11. Shyh-Chang N, Daley GQ. Lin28: primal regulator of growth and metab- 3. Sznol M, Chen L. Antagonist antibodies to PD-1 and B7-H1 (PD-L1) in the olism in stem cells. Cell Stem Cell 2013;12:395–406. treatment of advanced human cancer. Clin Cancer Res 2013;19:1021–34. 12. Moss EG, Lee RC, Ambros V. The cold shock domain protein LIN-28 4. Dong H, Strome SE, Salomao DR, Tamura H, Hirano F, Flies DB, et al. controls developmental timing in C. elegans and is regulated by the lin-4 Tumor-associated B7-H1 promotes T-cell apoptosis: a potential mecha- RNA. Cell 1997;88:637–46. nism of immune evasion. Nat Med 2002;8:793–800. 13. Wang T, Wang G, Hao D, Liu X, Wang D, Ning N, et al. Aberrant regulation 5. Velcheti V, Schalper KA, Carvajal DE, Anagnostou VK, Syrigos KN, Sznol M, of the LIN28A/LIN28B and let-7 loop in human malignant tumors and its et al. Programmed death ligand-1 expression in non-small cell lung cancer. effects on the hallmarks of cancer. Mol Cancer 2015;14:125. Lab Invest 2014;94:107–16. 14. Newman MA, Thomson JM, Hammond SM. Lin-28 interaction with the 6. Pardoll DM. The blockade of immune checkpoints in cancer immuno- Let-7 precursor loop mediates regulated microRNA processing. RNA 2008; therapy. Nat Rev Cancer 2012;12:252–64. 14:1539–49. 7. Reck M, Rodriguez-Abreu D, Robinson AG, Hui R, Csoszi T, Fulop A, et al. 15. Viswanathan SR, Daley GQ, Gregory RI. Selective blockade of micro- Pembrolizumab versus chemotherapy for PD-L1-positive non-small-cell RNA processing by Lin-28. Science 2008;320:97–100. lung cancer. N Engl J Med 2016;375:1823–33. 16. Calin GA, Sevignani C, Dumitru CD, Hyslop T, Noch E, Yendamuri S, et al. 8. Herbst RS, Morgensztern D, Boshoff C. The biology and management of Human microRNA genes are frequently located at fragile sites and genomic non-small cell lung cancer. Nature 2018;553:446. regions involved in cancers. Proc Natl Acad Sci USA 2004;101:2999–3004.

OF10 Cancer Immunol Res; 7(3) March 2019 Cancer Immunology Research

Targeting LIN28/let-7 for PD-L1–Mediated Immunotherapy

17. Johnson SM, Grosshans H, Shingara J, Byrom M, Jarvis R, Cheng A, et al. 34. Curiel TJ, Wei S, Dong H, Alvarez X, Cheng P, Mottram P, et al. Blockade of RAS is regulated by the let-7 microRNA family. Cell 2005;120:635–47. B7-H1 improves myeloid dendritic cell-mediated antitumor immunity. 18. Sampson VB, Rong NH, Han J, Yang Q, Aris V, Soteropoulos P, et al. Nat Med 2003;9:562–7. MicroRNA let-7a down-regulates MYC and reverts MYC-induced growth in 35. Boehm U, Klamp T, Groot M, Howard JC. Cellular responses to interferon- Burkitt lymphoma cells. Cancer Res 2007;67:9762–70. gamma. Annu Rev Immunol 1997;15:749–95. 19. Dong Q, Meng P, Wang T, Weiwei Q, Qin W, Wang F, et al. MicroRNA 36. Wherry EJ. T cell exhaustion. Nat Immunol 2011;12:492–9. let-7a inhibits proliferation of human prostate cancer cells in vitro and 37. Sunshine J, Taube JM. PD-1/PD-L1 inhibitors. Curr Opin Pharmacol 2015; in vivo by targeting E2F2 and CCND22010. 23:32–8. 20. Johnson CD, Esquela-Kerscher A, Stefani G, Byrom M, Kelnar K, Ovchar- 38. Tolba MF, Omar HA. Immunotherapy, an evolving approach for the enko D, et al. The let-7 microRNA represses cell proliferation pathways in management of triple negative breast cancer: Converting non-responders human cells. Cancer Res 2007;67:7713–22. to responders. Crit Rev Oncol Hematol 2018. 21. Boyerinas B, Park SM, Hau A, Murmann AE, Peter ME. The role of let-7 in 39. Casey SC, Tong L, Li Y, Do R, Walz S, Fitzgerald KN, et al. MYC regulates the cell differentiation and cancer. Endocr Relat Cancer 2010;17:F19–36. antitumor immune response through CD47 and PD-L1. Science 2016;352: 22. Shell S, Park SM, Radjabi AR, Schickel R, Kistner EO, Jewell DA, et al. Let-7 227–31. expression defines two differentiation stages of cancer. Proc Natl Acad Sci 40. Chen L, Gibbons DL, Goswami S, Cortez MA, Ahn YH, Byers LA, et al. USA 2007;104:11400–5. Metastasis is regulated via microRNA-200/ZEB1 axis control of tumour cell 23. Takamizawa J, Konishi H, Yanagisawa K, Tomida S, Osada H, Endoh H, PD-L1 expression and intratumoral immunosuppression. Nat Commun et al. Reduced expression of the let-7 microRNAs in human lung cancers in 2014;5:5241. association with shortened postoperative survival. Cancer Res 2004;64: 41. Gong AY, Zhou R, Hu G, Li X, Splinter PL, O'Hara SP, et al. MicroRNA-513 3753–6. regulates B7-H1 translation and is involved in IFN-gamma-induced B7-H1 24. Xie C, Chen W, Zhang M, Cai Q, Xu W, Li X, et al. MDM4 regulation by the expression in cholangiocytes. J Immunol 2009;182:1325–33. let-7 miRNA family in the DNA damage response of glioma cells. FEBS Lett 42. Beachy SH, Onozawa M, Chung YJ, Slape C, Bilke S, Francis P, et al. 2015;589:1958–65. Enforced expression of Lin28b leads to impaired T-cell development, 25. Wang Y, Zhou J, Chen Y, Wang C, Wu E, Fu L, et al. Quantification release of inflammatory cytokines, and peripheral T-cell lymphoma. of distinct let-7 microRNA family members by a modified stem-loop Blood 2012;120:1048–59. RT-qPCR. Mol Med Rep 2017. 43. Manier S, Powers JT, Sacco A, Glavey SV, Huynh D, Reagan MR, et al. The 26. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using LIN28B/let-7 axis is a novel therapeutic pathway in multiple myeloma. real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods Leukemia 2017;31:853–60. 2001;25:402–8. 44. Wang L, Yuan C, Lv K, Xie S, Fu P, Liu X, et al. Lin28 mediates radiation 27. Ghidu VP, Ilies MA, Cullen T, Pollet R, Abou-Gharbia M. A new and resistance of breast cancer cells via regulation of caspase, H2A.X and Let-7 efficient synthetic route for the anxiolytic agent CL285032. Bioorg Med signaling. PLoS One 2013;8:e67373. Chem Lett 2011;21:259–61. 45. Xu B, Zhang K, Huang Y. Lin28 modulates cell growth and associates with a 28. Cortez MA, Ivan C, Valdecanas D, Wang X, Peltier HJ, Ye Y, et al. PDL1 subset of cell cycle regulator mRNAs in mouse embryonic stem cells. RNA regulation by p53 via miR-34. J Natl Cancer Inst 2016;108. 2009;15:357–61. 29. Ebert MS, Neilson JR, Sharp PA. MicroRNA sponges: competitive inhibitors 46. Zhu H, Shyh-Chang N, Segre AV, Shinoda G, Shah SP, Einhorn WS, et al. of small RNAs in mammalian cells. Nature Methods 2007;4:721–6. The Lin28/let-7 axis regulates glucose metabolism. Cell 2011;147:81–94. 30. Iwasaki S, Kawamata T, Tomari Y. Drosophila argonaute1 and argonaute2 47. King CE, Cuatrecasas M, Castells A, Sepulveda AR, Lee JS, Rustgi AK. employ distinct mechanisms for translational repression. Molecular Cell LIN28B promotes colon cancer progression and metastasis. Cancer Res 2009;34:58–67. 2011;71:4260–8. 31. Sakurai M, Miki Y, Masuda M, Hata S, Shibahara Y, Hirakawa H, et al. 48. Zhang M, Liu F, Jia H, Zhang Q, Yin L, Liu W, et al. Inhibition of microRNA LIN28: a regulator of tumor-suppressing activity of let-7 microRNA in let-7i depresses maturation and functional state of dendritic cells in human breast cancer. J Steroid Biochem Mol Biol 2012;131:101–6. response to lipopolysaccharide stimulation via targeting suppressor of 32. Roos M, Pradere U, Ngondo RP, Behera A, Allegrini S, Civenni G, et al. cytokine signaling 1. J Immunol 2011;187:1674–83. A small-molecule inhibitor of Lin28. ACS Chem Biol 2016;11: 49. Albright JD, Moran DB, Wright WB Jr., Collins JB, Beer B, Lippa AS, et al. 2773–81. Synthesis and anxiolytic activity of 6-(substituted-phenyl)-1,2,4- 33. Clark CA, Gupta HB, Sareddy G, Pandeswara S, Lao S, Yuan B, et al. triazolo[4,3-b]pyridazines. J Med Chem 1981;24:592–600. Tumor-intrinsic PD-L1 signals regulate cell growth, pathogenesis, and 50. Loughlin FE, Gebert LF, Towbin H, Brunschweiger A, Hall J, Allain FH. autophagy in ovarian cancer and melanoma. Cancer Res 2016;76: Structural basis of pre-let-7 miRNA recognition by the zinc knuckles of 6964–74. pluripotency factor Lin28. Nat Struct Mol Biol 2011;19:84–9.

www.aacrjournals.org Cancer Immunol Res; 7(3) March 2019 OF11

LIN28/let-7/PD-L1 Pathway as a Target for Cancer Immunotherapy

Yanlian Chen, Chen Xie, Xiaohui Zheng, et al.

Cancer Immunol Res Published OnlineFirst January 16, 2019.

Updated version Access the most recent version of this article at: doi:10.1158/2326-6066.CIR-18-0331

Supplementary Access the most recent supplemental material at: Material http://cancerimmunolres.aacrjournals.org/content/suppl/2019/01/16/2326-6066.CIR-18-0331 .DC1

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 Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerimmunolres.aacrjournals.org/content/early/2019/02/15/2326-6066.CIR-18-0331. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.