Translational Science

Small-Molecule Inhibition of PD-1 TranscriptionIs an Effective Alternative to Antibody Blockadein Cancer TherapyAlison Taylor1,2, David Rothstein3, and Christopher E. Rudd2,4,5

Abstract

The impact of PD-1 immune checkpoint therapy promptsexploration of other strategies to downregulate PD-1 for cancertherapy. We previously showed that the serine/threonine kinase,glycogen synthase kinase, GSK-3a/b, is a central regulator of PD-1transcription in CD8þ T cells. Here, we show that the use of small-molecule inhibitors of GSK-3a/b (GSK-3i) to reduce pcdc1 (PD-1)transcription and expression was as effective as anti–PD-1 andPD-L1–blocking antibodies in the control of B16 melanoma,or EL4 lymphoma, in primary tumor and metastatic settings.Furthermore, the conditional genetic deletion of GSK-3a/breduced PD-1 expression on CD8þ T cells and limited B16pulmonary metastasis to the same degree as PD-1 gene defi-ciency. In each model, GSK-3i inhibited PD-1 expression ontumor-infiltrating lymphocytes, while increasing Tbx21 (T-bet)

transcription, and the expression of CD107aþ (LAMP1) andgranzyme B (GZMB) on CD8þ T cells. Finally, the adoptivetransfer of T cells treated ex vivo with a GSK-3 inhibitor delayedthe onset of EL4 lymphoma growth to a similar extent as anti–PD-1 pretreatment. Overall, our findings show how GSK-3inhibitors that downregulate PD-1 expression can enhanceCD8þ T-cell function in cancer therapy to a similar degree asPD-1–blocking antibodies.

Significance: These findings show how GSK-3 inhibitors thatdownregulate PD-1 expression can enhanceCD8þ T-cell function incancer therapy to a similar degree as PD-1 blocking antibodies,offering a next-generation approach in the design of immunother-apeutic approaches for cancermanagement. Cancer Res; 78(3); 706–17.�2017 AACR.

IntroductionThe coreceptor programmed cell death 1 (PD-1; PDCD1) is a

member of the B7 gene family that negatively regulates T-cellfunction (1–3). PD-1 is expressed in response to T-cell activa-tion and contributes to the exhaustion of CD8þ T cells duringchronic infections (4, 5). The coreceptor binds to ligands,programmed cell death ligand 1 and 2 (PD-L1/L2), on lym-phoid and nonlymphoid cells (6–8). Immune checkpointblockade with anti–PD-1 or anti–PD-L1 has proven successfulin the treatment of human cancers, either alone or in combi-nation with anti–CTLA-4 (9, 10). PD-1 expression on tumor-infiltrating CD8þ T cells correlates with impaired effector cellfunction (2, 11), while PD-L1 expression on tumors can facil-itate escape from the host immune system (3) and can serve as a

prognostic factor (12). Recent evidence indicates that recoveryof responses from anti–PD-1 blockade depends on the relatedcoreceptor (13–15).

The nature of the intracellular signaling pathways that regulatePD-1 expression on T cells has been the subject of much interest.Pdcd1 expression can be positively and negatively regulated bydifferent transcription factors such as nuclear factor of activated Tcells (NFAT), Forkhead box protein O1 (FoxO1), Notch, activatorprotein 1 (AP1), and B-lymphocyte maturation protein 1(Blimp1; refs. 16–19). Despite this, the identity of the upstreamsignaling event(s) that control PD-1 expression has been unclear.We and others previously showed that T cells are activated byprotein tyrosine kinases p56lck and ZAP-70 (20, 21). p56lck bindsto coreceptors CD4 and CD8 (22–24) and phosphorylatesimmune receptor activation motifs (ITAM) needed for ZAP-70recruitment to the TcR–CD3 complex (20, 23, 25). In contrast,glycogen synthase kinase 3 (GSK-3) is a serine/threonine kinasethat is active in resting T cells and becomes inactivated with T-cellactivation (26, 27). Differentially regulated isoforms of GSK-3aand b differ in their N- and C-terminal sequences and caninfluence pathways initiated by diverse stimuli. The inactivationof GSK-3 can be mediated by several upstream kinases, includingprotein kinase B (PKB/AKT). In CD4þ T cells, GSK-3 promotes theexit of NFAT from the nucleus (28, 29). TCR and CD28 ligationphosphorylates and inactivates GSK-3 (30–32), while expressionof active GSK-3b inhibits the proliferation of T cells (30). Engage-ment of PKB/AKT and GSK-3 in T cells operates independently ofguanine nucleotide exchange factor VAV-1 (31). Clinical trialsusing GSK-3 inhibitors have been undertaken in the treatment oftype II diabetes and various neurologic disorders (27, 33–35).

Recently, we reported that the inactivation of GSK-3a/bwith siRNAs and small-molecule inhibitors (SMI) specifically

1Leeds Institute of Cancer & Pathology Wellcome Trust Brenner Building, StJames's University Hospital, Leeds United Kingdom. 2Cell Signaling Section,Department of Pathology, Tennis Court Road, University of Cambridge, Cam-bridge, United Kingdom. 3Thomas E. Starzl Transplantation Institute, Universityof Pittsburgh School of Medicine, Pittsburgh, Pennsylvania. 4Department ofImmunology-Cell Therapy, Centre de Recherche Hôpital Maisonneuve-Rose-mont, Montreal, Quebec, Canada. 5Department of Medicine, University ofMontr�eal, Montreal, Quebec, Canada.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Corresponding Author: Christopher E. Rudd, Centre de Recherche HopitalMaisonneuve-Rosemont, Universit�e de Montr�eal, 5345, Boulevard de l'Assomp-tion, Montreal, Quebec H1T 4B3, Canada. Phone: 514-252-3400 x5878; E-mail:christopher.e.rudd@umontreal.ca

doi: 10.1158/0008-5472.CAN-17-0491

�2017 American Association for Cancer Research.

CancerResearch

Cancer Res; 78(3) February 1, 2018706

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491

downregulate PD-1 expression for enhanced CD8þ cytolyticT cell (CTL) function and clearance of viral infections (36).The approach has introduced the possibility that SMIs of GSK-3may be effective in the downregulation of PD-1 in the treatmentof cancer. Here, we show that SMIs of GSK-3 are as effective asanti–PD-1 in the control of B16 melanoma and EL4 lymphomagrowth in mice. Our findings demonstrate, for the first time, thesuccessful application of a GSK-3 inhibitor for the downregula-tion of PD-1 on T cells in cancer immunotherapy.

Materials and MethodsMice and cells

C57/Bl6mice were used alongside OT-1 Tg and Rag2 knockoutmice. Spleen cells were treated with a hypotonic buffer with 0.15mol/L NH4CL, 10 mmol/L KHCO3, and 0.1 mmol/L EDTA, pH7.2 to eliminate red blood cells before suspension in RPMI1640medium supplemented with 10% FCS, 50 mmol/L b-mercap-toethanol, sodium pyruvate, 2 mmol/L L-glutamine, and 100U/mL penicillin and streptomycin (Gibco). T cells were isolatedfrom tumor-infiltrating cells, spleen, and lymph node samples byuse of T-cell purification columns (R&D Systems). In some cases,whole lymphocyte samples were used for flow cytometry todetermine PD-1 expression in other cell types. Cells includedB16 F10 melanoma and EL4 lymphoma cells (obtained from theATCC). Each cell line was grown to achieve adequate numbers forfreezing, followed by repeated thawing for use in the describedexperiments. The length and time between thawing and use inexperiments was on average 3 to 4 weeks. The cell lines wereauthenticated by means of cell surface staining and flow cytome-try for characteristic markers and by their growth properties asdescribed in the literature. Cell cultures were occasionally testedfor mycoplasma (last tested in 2011). The research was regulatedunder the Animals (Scientific Procedures) Act 1986 AmendmentRegulations 2012 following ethical review by the University ofCambridge Animal Welfare and Ethical Review Body (AWERB)Home Office UK PPL No. 70/7544.

Antibodies and reagentsThe following antibodies were used in experiments: anti-CD3

(2C11), anti–PD-1 (CD279, J43) and anti–CTLA-4 (9H10; Bio XCell); PD-L1 (E1L3N; Cell Signaling Technology), anti-granzymeB and anti–T-bet (Abcam plc); anti–GSK-3a/b, CD279 (cloneEH12.2H7) coupled FITC and mouse IgG1 FITC control (BioLe-gend); conjugated antibodies anti-CD8a (clone, 53-6.7), anti-CD4 (clone, RM4–5), CD44, CD62L, CD25, CD69 (eBioscience).Carboxyfluorescein succinimidyl ester (CFSE) antibodies along-side PE Annexin V Apoptosis Detection Kit with 7-AAD (BioLe-gend) was used for viability and proliferation assays. GSK3inhibitors SB415286 3-(3-chloro-4-hydroxyphenylamino)-4-(2-nitrophenyl)-1H-pyrrole-2,5-dione and AZ1080 (Abcam plc),OVA257-264 peptide (Bachem Ag).

Flow cytometryFlow cytometry of antibody staining of surface receptors was

conducted by suspending 106 cells in 100 mL PBS and addingantibody (1:100) for 2 hours at 4�C. Cells were thenwashed twicein PBS and in some cases suspended in 100 mL PBSwith secondaryantibody for a further 1 hour at 4�C.Cell stainingwas analyzed ona BD FACSCalibur flow cytometer and by FlowJo software. Forintracellular staining, cells were fixed in 4% paraformaldehyde,

permeabilized with 0.3% saponin (Sigma-Aldrich) and stainedwith thedesired antibody in saponin containingPBS for 2hours at4�C, followed by a secondary Ab incubation where primaryantibodies were not conjugated.

qPCRSingle-strand cDNA was synthesized with an RT-PCR Kit

(Qiagen) according to the manufacturer's instructions. Reversetranscription was performed using the RNA PCR Core Kit(Applied Biosystems). qPCR used SYBR Green Technology(Roche) on cDNA generated from the reverse transcription ofpurified RNA. After preamplification (95�C for 2 minutes), thePCRs were amplified for 40 cycles (95�C for 15 seconds and60�C for 60 seconds) in a sequence detection system (PE Prism7000; PerkinElmer Applied Biosystems). The exponentialphase, linear phase, and plateau phase of PCR amplificationwere carefully monitored to ensure a measurement of real-timetranscription (33). mRNA expression was normalized againstGAPDH expression using the standard curve method. PD-1-FW,50-CCGCCTTCTGTAATGGTTTGA-30; PD-1-RV, 50-GGGCAGC-TGTATGATCTGGAA-30; Tbet-FW, 50-GATCGTCCTGCAGTCTC-TCC-30; Tbet-RW, 50-AACTGTGTTCCCGAGGTGTC-30; GAPDH-FW, 50-CAACAGCAACTCCCACTCTTC-30; GAPDH- RW, 50-GGTCCAGGGTT TCTTACTCCTT-30.

Melanoma lung tumor establishment in wild-type miceB16 melanoma cells [(2 � 105 taken from the log phase of

in vitro growth) OVA-peptide pulsed or nonpulsed] were trans-ferred intravenously into syngeneic C57BL/6Jmice 10 to 12weeksold. The lungs were removed 14 days after the transfer, andvisible metastatic colonies on the lungs were counted. In somecases, live imaging usedB16 cells taggedwith luciferase.Micewereinjected intraperitoneally with luciferin (2 mg/mouse), anaesthe-tized with isoflurane, and scanned with an IVIS Lumina (CaliperLife Sciences).

Adoptive transfer of in vitro generated CTLsIn vitro generated T cells were injected into mice with 7-day

established EL4 tumors. For this,OT-1CTLswere generated in vitroas described previously (36). Primary mouse T cells were isolatedusing T-cell purification columns (R&D Systems). OVA-specificCD8þ cytolytic T cells were generated by incubating OT-I sple-nocytes with SIINFEKL peptide of OVA (OVA257-264) at 10 ng/mLfor 5 to 7 days. Isolated T cells (105 cells) were injected intrave-nously into mice with established EL4 tumors that had beenintradermally injected into mice 7 days before cell transfer.

Intradermal tumor establishmentEL4 or B16 tumor cells were taken from the log phase of in

vitro growth (�70% confluency). In some cases where stated,cells were pulsed with OVA peptide for 1 hour at 37�C). Theywere then washed and injected into mice (typically 3� 106 cellsfor EL4 and 2 � 105 for B16 cells). Tumors were clearly visibleafter 1 week and grew progressively in an encapsulated fashion.Induced tumors were measured on a daily basis using a verniercaliper. Tumors, spleens and lymph nodes were harvested asindicated, either on day 10 or when the tumor reached amaximum diameter of 12 mm. PCR and flow cytometry wereperformed.

GSK-3 Downregulation of PD-1 in Cancer Therapy

www.aacrjournals.org Cancer Res; 78(3) February 1, 2018 707

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491

Isolation of tumor-infiltrating lymphocytesSolid tumors or nodules from lungs were harvested from

mice at the time indicated. Tissue was disrupted using a bladeand then incubated in HBSS solution containing 200 U/mL ofcollagenase at 37�C for 2 hours. Tissue was then passed througha strainer and cells collected and layered onto Ficoll beforecentrifugation. Tumor-infiltrating cells were then collectedfrom the lymphocyte layer.

Statistical analysisThe mean and SE of each treatment group were calculated for

all experiments. The number of samples is indicated in thefigure legends. Unpaired Student t tests or ANOVA tests wereperformed using the InStat 3.0 software (GraphPad). In certaininstances, statistics were done using two-way ANOVA, or by

nonparametric Mann–Whitney at each time point. �, P < 0.05;��, P < 0.01; ���, P < 0.001.

ResultsGSK-3 inhibits the growth of intravenous- and intradermal-injected tumors

To assess whether the downregulation of PD-1 by GSK-3inhibition (GSK-3i) was effective in limiting tumor growth,B16 tumor cells were injected intravenously into C57/b6 micewith the GSK-3 inhibitor SB415286 and/or anti–PD-1 (Fig. 1A).The optimal dose of SB415286 and anti–PD-1 established in thismodel was 200 mg and 100 mg/injection/mouse, respectively. TheSMI or antibody was administered every 2 days following theinjection of tumor, followed by a harvest of lungs on day 14 and

Figure 1.

GSK-3 inhibition, anti–PD-1, and combination reduce pulmonary metastasis of B16 melanoma to the same extent. A, Schematic representation oftreatment regime (top left). Histogram showing the number of lung spots per animal with or without the described treatment (right), photograph belowshows one example of each group (n ¼ 6 mice/condition). B, Flow cytometric profiles of PD-1 expression for T cells isolated from spleen, draininglymph nodes, and TILs (data representative of 6 samples). C, qPCR of PD-1 and T-bet transcription of splenic T cells. D, Percent of CD8þ cellsexpressing granzyme B of tumor-infiltrating cells as determined by flow cytometry (n ¼ 6). ��, P < 0.001. E, Dose response of SB415286 on B16melanoma. C57/bl6 mice were injected intravenously with luciferase-tagged B16 melanoma cells and treated with doses of SB41586 as indicated.Luminescent images show B16 metastasis at day 14 and flow cytometric profiles to the right of each image show the level of PD-1 expressionon different cell subsets, taken from the spleen at day 14 (n ¼ 3).

Taylor et al.

Cancer Res; 78(3) February 1, 2018 Cancer Research708

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491

an assessment of numbers of B16 nodules. GSK-3 inhibitorSB415286 reduced the number of B16 spots from a mean of145 to 60 (i.e., >55% inhibition). This effect was comparable withanti–PD-1, which showed a mean of 70 spots (i.e., >50% inhi-bition). Furthermore, the combination of SB415286 and anti–PD-1 had the same effect as SB415286 and anti–PD-1 individu-ally (n¼ 6). Flow cytometry confirmed that GSK-3i reduced PD-1expression on T cells from the tumor [i.e., tumor infiltrating(TIL), spleen and draining lymph nodes (Fig. 1B)]. In contrast,no effect on the expression of other receptors, such as CD3, CD8,CD44, CD62L, CD25, and CD69, was observed (SupplementaryFig. S1A), similar to previous results involving GSK3i in viralinfection (36). GSK-3i reduced pcdc1 (PD-1) transcription in Tcells from isolated spleen of tumor-bearing mice (i.e., 3.7–0.8),concurrent with an increase in Tbx21 (T-bet) transcription (i.e.,1.6–4.2; Fig. 1C). Concurrent with reduced tumor growth,SB415286 treatment increased the percent of CD8þ TILs expres-sing CD107aþ (Lamp1) and granzyme B (GZMB; i.e., mean % of14–23), indicative of an increased presence of CD8þ killer T cellsin the tumor mass (Fig. 1D). These data showed that the down-regulation of PD-1 with a small-molecule inhibitor of GSK-3 canbe as effective as anti–PD-1 in the control of B16 pulmonarymetastasis in mice.

We also examined the effect of SB415286 doses on the growthof B16 cells tagged with luciferase (Fig. 1E). At day 14, mice wereinjected intraperitoneally with luciferin and scanned by IVISLumina imaging. SB416286 reduced the luciferase signal at 100and 200 mg/mouse, the dose of 200 mg being more effective(left). In terms of T-cell subsets, both doses reduced PD-1expression on the surface of CD8þ cells (right). In contrast,PD-1 expression was reduced to a lesser extent on CD3�

NKp46þ natural killer (NK) cells. No reduction on NK cellswas observed at 100 mg, despite the ability of this dose of drug toreduce tumor burden. Furthermore, we found that PD-1 expres-sion on CD4þ T cells, and CD4þ CD25þ FoxP3þ PD-1þ regu-latory T cells (Treg) taken from tumors was unaffected. Thesedata showed that GSK-3i preferentially downregulated PD-1 onCD8þ T cells, while having a lessor effect on NK cells and noobvious effect on CD4þ T cells in the B16 tumor model. In acomparison of injection frequency, more frequent injections of200 or 400 mg was optimal (Supplementary Fig. S1B). Four to 6injections of each dose seemed optimal for tumor rejection andPD-1 downregulation in vivo. A dose of 200 mg was as effective as400 ug in these setting.

To assess the role of GSK-3 genetically, we next compared B16tumor growth in GSK-3a/b conditional knockout mice relative towild-type and PD-1–deficient mice (Pdcd1�/�). GSK-3a/b con-ditional knockout (GSK-3a/b�/�) mice were generated from aGSK-3 alpha flox/flox/beta flox/flox Lck Creþ parental mice(parental line kindly provided by Dr. Jim Woodgett, Universityof Toronto, Toronto, Canada). T cells from conditional knockoutGSK-3a/b�/�mice showed the expected reduction in PD-1 expres-sion on CD8þ T cells when assessed frommice injected with B16cells and assayed on day 10 (Fig. 2A, left). In contrast, no effect onPD-1 expression on CD4þ T cells was noted, neither in terms ofcell number or mean fluorescent intensity (MFI; Fig. 2A, right).Second, B16pulmonarymetastasis wasmarkedly reduced inGSK-3a/b�/�mice (Fig. 2B). Intriguingly, the tumor growth inGSK-3a/b�/�micewas reduced to a similar extent as seen inPdcd1�/�mice(i.e., from 110 spots to <10 in both sets of mice; Fig. 2C).Furthermore, the injection of anti–PD-1 in GSK-3a/b�/� mice

had no further effect on the number of nodules in lungs (Fig. 2B,top). Conversely, the injection of SB415286, or another GSK-3inhibitor, AZ1080 in Pdcd1�/� mice had no additional effect inreducing the number of nodules (Fig. 2C, top). Overall, the datashowed that GSK-3 inhibitor preferentially downregulated PD-1expression on CD8þ T cells and that the loss or inhibition ofSB415286 and PD-1 in mice had the same effect in limiting B16pulmonary metastasis. This observation, combined with thefinding that GSK-3 inhibitors had no further inhibition of tu-mor growth in Pdcd1�/�mice, and that anti–PD-1 had no furthereffect on tumor growth in GSK-3a/b�/� mice supported thenotion that GSK-3 inhibition operated to limit tumor growthprimarily via the downregulation of anti–PD-1.

To further exclude that SB415286 had a direct effect ontumor cells, the SMI was injected into Rag2�/� mice (missingB and T cells) with the B16 tumor (Fig. 2D). Under thiscondition, SB415286 had no effect on limiting tumor growth.In addition, CFSE-labeled B16 cells were cocultured ofSB415286 in vitro over 5 days and assessed for differences.SB415286 had no obvious effect on the growth of B16 cells(Supplementary Fig. S2). Taken together, these data were mostconsistent with the interpretation that protective effects of GSK-3i were due primarily to an effect on CD8þ T cells.

We next also assessed whether the SMI SB415286 could affectPD-1 expression on human T cells (Supplementary Fig. S3).Human CD4þ CD8þ T cells isolated from human peripheralblood were stimulated with anti-CD3/CD28 for 72 hours priorto resting overnight and then incubated with SB415286 forvarious times and stained by flow cytometry for PD-1 expres-sion. The SMI reduced the percent of cells expressing PD-1 by>55% and the MFI for expression by >65% for cells whenassayed at 48 hours (left top and bottom panels). Lesser inhi-bition was also apparent at >96 hours. See examples in rightpanels at 10 minutes, 24, and 48 hours (right). As a control, theexpression of CD3, CD4, and CD8 was unaffected. These dataconfirm that GSK-3i can be used to downregulate PD-1 onCD3þCD8þ human T-cells.

SB415286 and anti–PD-1 had similar effects on the growthof more established B16 pulmonary metastases in mice(Fig. 3A). B16 tumor cells were injected intravenously and leftfor 7 days before beginning treatment. Lungs were harvested onday 19 and assessed for B16 nodules. Because of the extendedtime, nontreated animals showed larger nodules than seen inthe previous experiments. Despite this, injection of mice witheither SB415286 or anti–PD-1 greatly reduced the number ofnodules from 135 to 5–20. Both reagents also reduced the sizeof the remaining nodules from a mean of 3 to 0.05 mm indiameter. Furthermore, the combined injection of SB415286and anti–PD-1 reduced the number of spots to the same extentas each individual treatment (i.e., 5–10). As a control, qPCRmeasurements confirmed the reduction in pcdc1 transcription,with a concurrent increase in Tbx21 transcription in splenic Tcells (Fig. 3B). This reduction in PD-1 expression was confirmedby flow cytometry where we observed that SB415286 or com-bined therapy reduced PD-1 expression in T cells from spleenand TILs (Fig. 3C). Furthermore, this reduction in expressioncorrelated with an increase in Lamp1- and GZMB-expressingCD8þ TILs (Fig. 3D).

We next investigated whether GSK-3 inactivation was also aseffective as anti–PD-1 checkpoint blockade in the control of thegrowth of solid B16 tumors. For this, B16 cells were injected

GSK-3 Downregulation of PD-1 in Cancer Therapy

www.aacrjournals.org Cancer Res; 78(3) February 1, 2018 709

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491

intradermally, followed by intraperitoneal injections of eitherSB415286 or anti–PD-1 (Fig. 4A). We found that SB415286,anti–PD-1, and the combination slowed tumor growth suchthat 10–mm–sized tumors were not seen until days 19 to 22,rather than day 14 in untreated mice. Furthermore, SB415286or anti–PD-1 increased overall survival to the same generalextent, as seen in the Kaplan–Meier survival plot with a 40% to50% survival at day 30 (Fig. 4B). As a control, qPCR of isolatedspleen T cells showed a marked decrease in pdcd1 transcription,while increasing the transcription of Tbx21 (Fig. 4C). Flowcytometry also showed a decrease in PD-1 expression in T cellsfrom the spleen and extracted TILs (n ¼ 5; Fig. 4D). ReducedPdcd1 and increased Tbx21 transcription was also observed inTILs (Fig. 4E). Finally, both SB415286 and combination ther-apy increased the numbers of CD107aþGZMBþCD8þ cellsindicative of the increased presence of more effective killer T

cells (Fig. 4F). There was no significant difference between thenumber of CD107aþGZMBþCD8þ cells with SB415286 versusSB415286 and anti–PD-1.

We also assessed the effect of GSK-3 inactivation on immunecell rejection of another tumor model, EL4 lymphoma cells(Fig. 5). Priming of OT-1 OVA-specific T cells with SIINFEKLpeptide of OVA (OVA257-264) produces a specific CTL responseagainst tumor targets (37). EL4 cells were pretreated with 0, 2,5, and 10 mg of OVA peptide, and washed, prior to injection.EL4 cells not exposed to peptide were injected into the left flankof OT-1 Tg mice, while those exposed to peptide were injectedinto the right flank. SB415286 was injected intraperitoneally onday 0, and tumor growth was then monitored over 10 daysbefore harvesting (Fig. 5A, top). Tumor size was reducedespecially with 5 and 10 mg OVA peptide; however, tumorswere still evident at all peptide concentrations (bottom). In

Figure 2.

Pulmonary metastasis of B16 melanoma is reduced to the same extent in GSK-3a/b�/� and Pdcd�/� mice. A, CD8þ T-cells from GSK-3a/b�/� mice showa decrease in PD-1 expression relative to CD8þ T cells from wild-type control mice CD8þ T cells (left), CD4þ T cells (right). B, GSK-3a/b�/� enhancedclearance of lung melanoma. Top inset, expanded abscissa range showing a lack of an effect of anti–PD-1 on numbers of nodules in GSK-3a/b�/� mice (n ¼ 5mice/condition). � , P < 0.05; ��, P < 0.001; ns, no significant difference relative to controls. C, Pdcd�/� mice show enhanced clearance of lung B16melanoma. Top right inset, expanded abscissa range showing a lack of an effect of GSK-3 inhibitors SB415286 and AZ1080 on numbers of B16 nodulesin Pdcd�/� mice (n ¼ 5 mice/condition). Panel shows a representative experiment. � , P < 0.05; �� , P < 0.001; ns, no significant difference relative tocontrols. D, Inhibition of GSK-3 does not affect pulmonary metastasis of B16 melanoma in Rag2�/� mice. Left, photo of lungs harvested from Rag2�/� mice14 days after injection of B16 cells in the presence or absence of the injection of 6 doses of 200 mg SB415286; right, histogram shows number of spotsfrom mice in the left (n ¼ 3).

Taylor et al.

Cancer Res; 78(3) February 1, 2018 Cancer Research710

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491

contrast, in 4 of 7 experiments, SB415286 treatment resultedin a complete loss of tumor mass at all peptide concentra-tions in >80% of mice. This remarkable tumor clearance wasobserved in mice of different ages, 4 to 6 weeks (SupplementaryFig. S4A), 6 to 10 weeks (Supplementary Fig. S4B), and 6months (Supplementary Fig. S4C). As a control, real-time PCRof splenic T cells confirmed SB415286 in vivo injection inhib-ited pcdc1 transcription, while increasing Tbx21 transcription(Fig. 5B). Flow cytometry staining of cells with anti–PD-1 alsoconfirmed the reduction in coreceptor expression. In experi-ments where the tumors were not eliminated by SB415286 atall peptide concentrations, the drug and anti–PD-1 had similareffects (Fig. 5C). At 5 mg/mL OVA, untreated mice carriedtumors of 5 mm by day 14, while SB415286, anti–PD-1, orthe combination, delayed the appearance of this sized tumoruntil days 20 to 22. Similarly, in the case of mouse survival, at

2 mg/mL OVA peptide, the SMI SB415286, anti–PD-1 andcombination therapy increased survival from 17 to 24–26 daysas seen by the Kaplan–Meier survival plot (bottom). At 5 mg/mLOVA, SB415286, anti–PD-1, and combination therapy increas-ed survival from 21 to 30–31 days. At 10 mg/mL OVA peptide,SB415286 and anti–PD-1 completely protected against death,compared with day 25 for untreated mice. As an additionalcontrol, flow cytometric analysis of intracellular/surface stainedcells showed that SB415286 treatment did not affect theexpression of other markers (i.e., other than PD-1) such asCD44, CD4, FasL, FoxP3, CD152 (CTLA-4), and CD25 (Sup-plementary Fig. S5).

In another approach, we assessed PD-1 blockade usinga blocking antibody to the PD-1 ligand (PD-L1; Fig. 5D).SB415286, anti–PD-L1, and anti-PD-1 delayed the onset of EL-4 tumor growth at 5mgOVApeptide fromanonset of growth from

Figure 3.

GSK-3 inhibition and anti–PD-1 enhance clearance of established B16 tumors. A, Schematic representation of treatment regime (top left). Histogramshowing the number of lung spots per animal with or without the described treatment (right). Photograph below shows one example of each group (n ¼ 6mice/condition). B, qPCR of PD-1 and T-bet transcription of splenic T cells from animals used in A (n ¼ 5). � , P < 0.05; ��, P < 0.001; ns, no significantdifference relative to controls. C, Flow cytometric profiles of PD-1 expression on T cells isolated from spleen and TILs (data representative of 5 samples).D, Percent of CD8þ cells expressing granzyme B of tumor-infiltrating cells as determined by flow cytometry (n ¼ 5).

GSK-3 Downregulation of PD-1 in Cancer Therapy

www.aacrjournals.org Cancer Res; 78(3) February 1, 2018 711

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491

day 10 to 20–21 (top left). At 10 mg OVA, SB415286, anti–PD-L1,and anti–PD-1 completely eradicated the presence of tumors(top right). Again, as a control with 5 mg peptide, SB415286markedly reduced in pdcd1 transcription while increasingT-bet expression. These studies showed SB415286, anti–PD-L1and anti–PD-1 were remarkably similar in delaying the onset oftumor growth.

We then compared the effect of GSK-3 inactivation versusanti–PD-1 on EL4 solid tumors in the absence of OVA peptide(Fig. 6). This required coinjections of SB415286 and/or anti–PD-1 as depicted (Fig. 6A) and an assessment of tumor growthover a longer period of 40 days. EL4 tumors grew to 10 mm byday 5 in untreated mice, which was prolonged to day 12 inSB41528 and anti–PD-1–treated mice (bottom). The tumorcompletely regressed in 1 to 3 mice by day 18 with anti–PD-1,or the combination of SB41528 and anti–PD-1. The coopera-tive effect of combination therapy was also reflected in survival

(Fig. 6B). Although untreated mice died by 11 days, 50% ofmice on combined therapy remained alive by 40 days asshown in the Kaplan–Meier survival plot. Flow cytometry ofT cells isolated from the spleens, draining lymph nodes, andTILs demonstrated lower expression levels of PD-1 (Fig. 6C).This was confirmed with PCR demonstrating a reduction inPdcd1 and an increase in Tbx21 (Fig. 6D).

To determine the duration of the effect of SB415286, PD-1expression was monitored in mice coinjected with EL-4 tumorsand a single injection of the drug (Supplementary Fig. S6). Micewere sacrificed at various times, and PD-1 expression on spleenCD3þCD8þ T cells was assessed by flow cytometry. The effects onPD-1 expression onCD3þCD8þ T cells was sustained until 7 to 10days (i.e., 57% suppression at day 5; 42% at day 7) with expres-sion returning to control levels from day 10 to 14 (i.e., 22%suppression at day 10 and 7% at day 14). These data indicate thatthe effects of SB415286 were sustained for over 7 to 10 days.

Figure 4.

GSK-3 inhibition and anti–PD-1 inhibit growth of solid B16 tumors. A, Schematic representation of treatment regime (top); tumor growth curves (bottom).Number in bottom right corner represents how many mice (out of 3) from each treated condition were tumor free at the end of the study. B, Survival curves ofmice with and without treatment as shown (n ¼ 6; number of mice: 24). Panel shows a representative experiment. C, qPCR of splenic T cells. D, Flowcytometric profiles for T cells isolated from draining lymph nodes and TILs (data representative of 5 samples). E, qPCR of TILs. F, Percent of CD8þ

cells expressing CD107a/granzyme B of tumor-infiltrating cells as determined by flow cytometry. �� , P < 0.001; ns, no significant difference relativeto controls.

Taylor et al.

Cancer Res; 78(3) February 1, 2018 Cancer Research712

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491

Finally, we tested whether pretreatment of T cells ex vivo withSB415286 provided protection (Fig. 7). EL4 cells were intra-dermally injected into mice for 7 days followed by the transferof OT-1 CTLs that had been cultured in vitro for 7 days in theabsence or presence of OVA peptide plus the SMI SB415286or anti–PD-1. The transfer of cells without any SB415286 oranti–PD-1 pretreatment delayed the onset of growth at allpeptide concentrations (Fig. 7A). However, pretreatment withSB415286, anti–PD-1, and in combination further slowedtumor growth. At 5 mg Ova, each treatment delayed the onsetof tumor growth until days 19 to 21. Furthermore, tumorsattained a diameter of 8 mm by day 16 and at days 25 to 27with SB415286, anti–PD-1 or combined therapy. In thisinstance, the effect of SB415286 was the same as anti–PD-1.

Precultured T cells (prior to adoptive transfer) showed theexpected decrease in pcdc1 transcription while increasing thetranscription of Tbx21 (Fig. 7B). Flow cytometry prior to trans-fer confirmed the reduction in PD-1 expression at different OVAconcentrations (Fig. 7C). These data showed that the efficacyof GSK-3 inhibition was the result of a direct effect on T cellsand that the pretreatment of T cells ex vivowith either anti–PD-1or SB415286 provided protection against tumor growth.

DiscussionImmune checkpoint blockade with anti–PD-1 or anti–PD-L1

has proven to be a highly promising treatment of humancancers, either alone or in combination with other reagents

Figure 5.

GSK-3 inhibition and anti–PD-1/PL1 attenuate growth to a similar extent of solid EL4-OVA tumors. A, GSK-3 inhibition enhances clearance of solid EL4-OVAtumors. Top, schematic representation of treatment regime for EL4 solid tumor model with different concentrations of OVA peptide. Non–OVA-pulsedEL4 cells were injected into the left flank and EL4-OVA cells into the right flank; middle, photograph showing tumor growth after 10 days of OVA-pulsed EL4 tumor cells in OT-1 Tg mice with or without SB415286 injection (top); bottom, tumor growth curves (n ¼ 6 mice/condition). Panel shows arepresentative experiment. B, Real-time PCR of splenic T cells confirmed SB415286 inhibition of PD-1 and T-bet transcription. Flow cytometric profile ofanti–PD-1 staining of spleen T cells. C, SB415286 and anti–PD-1 alone and in combination attenuate B16 cell growth (top). Bottom, SB415286 and anti–PD-1alone and in combination increased mouse survival to the same degree. Two micrograms OVA peptide (bottom left); 5 mg OVA peptide (bottommiddle); 10 mg/mL (bottom right). Number in bottom right corner represents how many mice (out of 6) were tumor free at the end of the study. D, SB415286,anti–PD-1, and anti–PD-L1 (PD-1 ligand) show a similar effect on the inhibition of EL4 tumor growth (top). Left, effects in the presence of 5 mg OVApeptide; right, effects in the presence of 10 mg OVA peptide. Number in bottom right corner represents how many mice (out of 3) from each treatedcondition were tumor free at the end of the study. Bottom, qPCR of PD-1 and T-bet transcription.

GSK-3 Downregulation of PD-1 in Cancer Therapy

www.aacrjournals.org Cancer Res; 78(3) February 1, 2018 713

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491

such as anti–CTLA-4 (2, 9, 10). However, only a minority ofpatients is responsive to this therapy, and there is a need to findalternate ways to complement present approaches. We previ-ously showed that the kinase GSK-3a/b is a central regulator ofPD-1 expression and that SMIs of GSK-3 (GSK3i) are effectivein promoting viral clearance (36). In this article, we have shownthat GSK-3i inhibition of pcdc1 (PD-1) transcription with asmall-molecule inhibitor (i.e., SB415286) is as effective asanti–PD-1 and PD-L1–blocking antibodies in the control of

B16 and EL-4 tumor growth. Our findings identify a potentialalternate approach using small-molecule inhibition of PD-1expression in cancer therapy.

Our findings showed that GSK-3 inhibition with SMI treat-ment operates primarily via a reduction in PD-1 expression onthe CD8þ T cells. As in the case of drug inhibition of PD-1transcription, GSK-3a/b�/� T cells showed a reduction in PD-1expression, while B16 pulmonary metastasis was reduced to asimilar extent in Pdcd�/� and GSK-3a/b�/� mice. In each model,

Figure 6.

GSK-3 inhibition and combined anti–PD-1 therapy slows solid EL4 tumorgrowth. A, Schematic representationof treatment regime for EL4 solidtumor model without OVA-presentation (top). Tumor growthcurves (bottom; n ¼ 6 mice/condition). Number in bottom rightcorner represents howmanymice (outof 6) were tumor free at the end of thestudy. B, Survival curves. C, Flowcytometric profiles of PD-1 expression;data representative of 5 samples.D, qPCR of PD-1 and T-bettranscription (n ¼ 5).

Taylor et al.

Cancer Res; 78(3) February 1, 2018 Cancer Research714

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491

GSK-3i inhibited Pdcd1 transcription and PD-1 expression onTILs, while increasing T-bet transcription and the presence ofCD8þ TILs expressing LAMP1 and GZMB. Despite this, it is alsoimportant to note that GSK-3 is likely to affect other aspects of T-cell function in a PD-1–independent fashion. We showed thatthe enzyme upregulates T-bet expression (36), which regulatesthe expression of numerous other genes, such as GZMB andIFNg1 (38). GSK-3i may eventually be found to alter theexpression of other receptors and mediators and provide apotential advantage over anti–PD-1 blockade. However, in thecontext of the models examined to date, the downregulatoryeffect on PD-1 plays a central role in generating antitumorimmunity.

The role of the immune system in providing protective immu-nity viaGSK-3was also seen in conditionally deletedGSK-3a/b�/�

mice. Intriguingly, GSK-3 inactivation by gene ablation, or expo-sure to SMIs, preferentially reduced PD-1 expression on CD8þ Tcells. Expression on CD4þ T cells and CD4þ CD25þ FoxP3þ PD-1þ Tregs from tumors was unaffected. Although NK cells play keyroles in eliminating B16 tumors (39–42), PD-1 expression wasless affected on this subset than on CD8þ T cells. Furthermore, at

the lower SMI dose of 100 mg, reduced PD-1 expression onNKp46þ NK cells was not observed, despite the protective effectof the reagent on B16 tumor growth. At a higher dose, somereduction in PD-1 expression was observed, but to a lesser extentthan seen on CD8þ T cells. Furthermore, SB415286 had no effecton B16 pulmonary metastasis in Rag2�/�mice, which express NKcells. Therefore, althoughwedo exclude an effect onNK-mediatedkilling of tumors (42), the antitumor effect of GS3i in our modelwas primarily due to an effect on CD8þ T cells. This latterobservation also argued against a direct effect of the SMI ontumor growth. This lack of a direct effect of SMIs on tumors wassupported by the absence of an effect of SB415286 on the in vitrogrowth of B16 cells.

In addition to effects onB16 cells, GSK-3 inhibition, anti–PD-1,and anti–PD-L1had similar effects onEL4 lymphoma solid tumorgrowth. In most mice, SB415286 treatment eradicated tumors,while in other mice, depending on the concentration of OVApeptide, the SMI delayed the onset of growth by 20 to 22 days.GSK-3 inhibition, anti–PD-1, or anti–PD-L1 delayed tumorgrowth to the same extent. Furthermore, the combination ofSB415286 and anti–PD-1 had the same effect as monotherapy.

Figure 7.

Ex vivo pretreatment of T cells with SB415286, anti–PD-1, or combination followed by adoptive therapy delayed the onset EL4-OVA tumor growth. A, Regimeof adoptive therapy (top). Tumor growth curves (n ¼ 3; number of mice: >20; bottom). Panel shows a representative experiment. Number in bottomright corner represents how many mice (out of 3) from each treated condition were tumor free at the end of the study. B, Real-time PCR of splenicT cells prior to adoptive transfer confirmed SB415286 inhibition of PD-1 and T-bet transcription (top). C, Flow cytometric profiles of anti–PD-1 staining ofspleen T cells treated with different concentrations of OVA peptide prior to adoptive transfer.

GSK-3 Downregulation of PD-1 in Cancer Therapy

www.aacrjournals.org Cancer Res; 78(3) February 1, 2018 715

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491

Occasionally, we observed some cooperation between anti–PD-1and GSK-3i in the EL4 model, and therefore, we cannot excludethat future work will show cooperativity dependent on the tumormodel.

Finally, we found that the ex vivo pretreatment of CTLsfollowed by adoptive cell transfer was effective in the controlof tumor growth. As in the other models, SB415286 and anti–PD-1 delayed the onset of growth to the same extent. Thisexperiment confirmed that the protective effect of GSK-3 ontumor growth was due to an effect on the function of T cells.Combination therapy did not provide an obvious improve-ment, supporting the notion that SB415286 was acting toinhibit tumor growth primarily by downregulating the expres-sion of PD-1. GSK-3 SMIs may therefore have application to celltherapy, potentially to improve chimeric antigen receptor(CAR) therapy (43).

Overall, there are potential advantages and disadvantagesto the use of the GSK-3 SMI versus anti–PD-1 antibody ther-apies. Anti–PD-1 therapy is expensive and associated withadverse effects, such as fatigue, rash, and possible autoimmunecomplications, such as colitis. Although we cannot excludesome inflammatory effects with the use of GSK-3 inhibitors,we have seen no evidence of autoimmunity with SMIs or in theGSK-3a/b�/� mice over 2 years. Furthermore, the use of manyGSK-3 inhibitors is no longer restricted by patent coverage (27),and SMI inhibition offers the advantage of more accuratedosing, lower cost, and the potential of oral administration.Importantly, PD-1 expression on both murine and human Tcells was downregulated by GSK-3i.

The potential disadvantage of GSK-3 inactivation is a pos-sible direct effect on the function of other host cells or thetumor itself. However, lithium chloride, an inhibitor of GSK-3,has been used for decades for the treatment of bipolar diseasewithout a reported increase in tumor frequency. The dose ofSB415286 inhibitor (200 mg/20 g mouse) in our study wasroughly comparable with the dose of another inhibitor tide-glusib used in a phase II oral study (800 mg in an 80 kg patient)to treat progressive supranuclear palsy (35). Furthermore, weshowed that the effects of GSK-3 SMIs are durable such that asingle dose injection of SB415286 downregulated PD-1 for 10to 14 days. Although we failed to see any effect of SB415286 onthe growth of B16 melanoma cells, GSK-3 inhibition has beenreported to directly inhibit the growth of multiple myeloma,neuroblastoma, hepatoma, and prostate tumors (44–48). It istherefore possible that, in some instances, GSK-3 inhibitorsmight directly inhibit the growth of some tumors. Despite these

possibilities, the major effect of GSK-3i in our studies wasamplify the ability of the immune system to react againsttumor growth as shown by the effect of ex vivo–treated T cellsin adoptive cell therapy as well as by the elimination of tumorsin mice where GSK-3a/b conditionally deleted in the T cells.Certain tumors can also impair proximal TCR signaling eventsas a form of immune avoidance (49, 50). The inhibition ofGSK-3 could potentially circumvent this impairment given thatGSK-3 operates downstream of proximal signal mediators, suchas p56lck. Further work is needed to uncover the full range ofdownstream effects that may be regulated by GSK-3 regulationin antitumor immunity.

Disclosure of Potential Conflicts of InterestD. Rothstein and C.E. Rudd have ownership interest (including patents)

in Immune Ventures. No potential conflicts of interest were disclosed by theother authors.

Authors' ContributionsConception and design: A. Taylor, D. Rothstein, C.E. RuddDevelopment of methodology: A. Taylor, C.E. RuddAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): A. TaylorAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): A. Taylor, C.E. RuddWriting, review, and/or revision of the manuscript: A. Taylor, D. Rothstein,C.E. RuddAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): A. Taylor, C.E. RuddStudy supervision: C.E. Rudd

AcknowledgmentsWe thank Dr. Jim Woodgett, Lunenfeld-Tanenbaum Research Institute,

Mount Sinai Hospital, Toronto, for the heterozygotes of the GSK-3a/b condi-tional knockout mice. PD-1–deficient mice (Pdcd1-/) were a kind gift ofProf. T. Honjo, Kyoto University Faculty of Medicine, Japan.

C.E. Rudd was supported by Wellcome Trust 092627/Z/10/Z and aFoundation Award from the Centre de Recherche Hopital Maisonneuve-Rosemont. C.E. Rudd and A. Taylor were also supported by CRUK grantA20105.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received February 21, 2017; revised July 22, 2017; acceptedOctober 17, 2017;published OnlineFirst October 20, 2017.

References1. Schildberg FA, Klein SR, Freeman GJ, Sharpe AH. Coinhibitory

pathways in the B7-CD28 ligand-receptor family. Immunity 2016;44:955–72.

2. Baumeister SH, Freeman GJ, Dranoff G, Sharpe AH. Coinhibitorypathways in immunotherapy for cancer. Annu Rev Immunol 2016;34:539–73.

3. Iwai Y, Ishida M, Tanaka Y, Okazaki T, Honjo T, Minato N. Involvement ofPD-L1 on tumor cells in the escape from host immune system and tumorimmunotherapy by PD-L1 blockade. Proc Natl Acad Sci U S A 2002;99:12293–7.

4. Barber DL, Wherry EJ, Masopust D, Zhu B, Allison JP, Sharpe AH, et al.Restoring function in exhausted CD8 T cells during chronic viral infection.Nature 2006;439:682–7.

5. Day CL, Kaufmann DE, Kiepiela P, Brown JA, Moodley ES, ReddyS, et al. PD-1 expression on HIV-specific T cells is associatedwith T-cell exhaustion and disease progression. Nature 2006;443:350–4.

6. Freeman GJ, Long AJ, Iwai Y, Bourque K, Chernova T, Nishimura H, et al.Engagement of the PD-1 immunoinhibitory receptor by a novel B7 familymember leads to negative regulation of lymphocyte activation. J Exp Med2000;192:1027–34.

7. Okazaki T, Iwai Y, Honjo T. New regulatory co-receptors: inducible co-stimulator and PD-1. Curr Opin Immunol 2002;14:779–82.

8. Latchman Y, Wood CR, Chernova T, Chaudhary D, Borde M, Chernova I,et al. PD-L2 is a second ligand for PD-1 and inhibits T cell activation.Nat Immunol 2001;2:261–8.

Taylor et al.

Cancer Res; 78(3) February 1, 2018 Cancer Research716

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491

9. Topalian SL,Hodi FS, Brahmer JR,Gettinger SN, SmithDC,McDermottDF,et al. Safety, activity, and immune correlates of anti-PD-1 antibody incancer. N Engl J Med 2012;366:2443–54.

10. Wolchok JD, Kluger H, Callahan MK, Postow MA, Rizvi NA, Lesokhin A,et al. Nivolumab plus ipilimumab in advanced melanoma. N Engl J Med2013;369:122–33.

11. Ahmadzadeh M, Johnson LA, Heemskerk B, Wunderlich JR, Dudley ME,White DE, et al. Tumor antigen-specific CD8 T cells infiltrating the tumorexpress high levels of PD-1 and are functionally impaired. Blood2009;114:1537–44.

12. Ghebeh H, Mohammed S, Al-Omair A, Qattan A, Lehe C, Al-QudaihiG, et al. The B7-H1 (PD-L1) T lymphocyte-inhibitory molecule isexpressed in breast cancer patients with infiltrating ductal carcinoma:correlation with important high-risk prognostic factors. Neoplasia2006;8:190–8.

13. Krueger J, Rudd CE. Two strings in one bow: PD-1 negatively regulates viaco-receptor CD28 on T cells. Immunity 2017;46:529–31.

14. Kamphorst AO, Wieland A, Nasti T, Yang S, Zhang R, Barber DL, et al.Rescue of exhausted CD8 T cells by PD-1-targeted therapies is CD28-dependent. Science 2017;355:1423–7.

15. Hui E, Cheung J, Zhu J, Su X, Taylor MJ, Wallweber HA, et al. T cellcostimulatory receptor CD28 is a primary target for PD-1-mediated inhi-bition. Science 2017;355:1428–33.

16. StaronMM, Gray SM, Marshall HD, Parish IA, Chen JH, Perry CJ, et al. Thetranscription factor FoxO1 sustains expression of the inhibitory receptorPD-1 and survival of antiviral CD8(þ) T cells during chronic infection.Immunity 2014;41:802–14.

17. Oestreich KJ, Yoon H, Ahmed R, Boss JM. NFATc1 regulates PD-1 expres-sion upon T cell activation. J Immunol 2008;181:4832–9.

18. Mathieu M, Cotta-Grand N, Daudelin JF, Thebault P, Labrecque N. Notchsignaling regulates PD-1 expression during CD8(þ) T-cell activation.Immunol Cell Biol 2013;91:82–8.

19. Xiao G, Deng A, Liu H, Ge G, Liu X. Activator protein 1 suppressesantitumor T-cell function via the induction of programmed death 1. ProcNatl Acad Sci U S A 2012;109:15419–24.

20. Weiss A, LittmanDR. Signal transduction by lymphocyte antigen receptors.Cell 1994;76:263–74.

21. Rudd CE. Adaptors andmolecular scaffolds in immune cell signaling. Cell1999;96:5–8.

22. Rudd CE, Trevillyan JM, Dasgupta JD, Wong LL, Schlossman SF. The CD4receptor is complexed in detergent lysates to a protein-tyrosine kinase(pp58) from human T lymphocytes. Proc Natl Acad Sci U S A 1988;85:5190–4.

23. Barber EK, Dasgupta JD, Schlossman SF, Trevillyan JM, Rudd CE. TheCD4 and CD8 antigens are coupled to a protein-tyrosine kinase(p56lck) that phosphorylates the CD3 complex. Proc Natl Acad SciU S A 1989;86:3277–81.

24. Veillette A, Bookman MA, Horak EM, Samelson LE, Bolen JB. Signaltransduction through the CD4 receptor involves the activation of theinternal membrane tyrosine-protein kinase p56lck. Nature 1989;338:257–9.

25. Chan AC, Iwashima M, Turck CW, Weiss A. ZAP-70: a 70 kd protein-tyrosine kinase that associates with the TCR zeta chain. Cell 1992;71:649–62.

26. Woodgett JR. Molecular cloning and expression of glycogen synthasekinase-3/factor A. EMBO J 1990;9:2431–8.

27. Cohen P, Frame S. The renaissance of GSK3. Nat Rev Mol Cell Biol2001;2:769–76.

28. Beals CR, Sheridan CM, Turck CW, Gardner P, Crabtree GR. Nuclear exportof NF-ATc enhanced by glycogen synthase kinase-3. Science 1997;275:1930–4.

29. Neal JW, Clipstone NA. Glycogen synthase kinase-3 inhibits the DNAbinding activity of NFATc. J Biol Chem 2001;276:3666–73.

30. Ohteki T, Parsons M, Zakarian A, Jones RG, Nguyen LT, Woodgett JR,et al. Negative regulation of T cell proliferation and interleukin 2

production by the serine threonine kinase GSK-3. J Exp Med 2000;192:99–104.

31. Wood JE, SchneiderH,RuddCE. TcR andTcR-CD28 engagement of proteinkinase B (PKB/AKT) and glycogen synthase kinase-3 (GSK-3) operatesindependently of guanine nucleotide exchange factor VAV-1. J Biol Chem2006;281:32385–94.

32. Appleman LJ, van Puijenbroek AA, Shu KM, Nadler LM, Boussiotis VA.CD28 costimulation mediates down-regulation of p27kip1 and cell cycleprogression by activation of the PI3K/PKB signaling pathway in primaryhuman T cells. J Immunol 2002;168:2729–36.

33. Jope RS, Roh MS. Glycogen synthase kinase-3 (GSK3) in psychiatricdiseases and therapeutic interventions.CurrDrugTargets 2006;7:1421–34.

34. Mazanetz MP, Fischer PM. Untangling tau hyperphosphorylation indrug design for neurodegenerative diseases. Nat Rev Drug Discov2007;6:464–79.

35. Tolosa E, Litvan I,HoglingerGU,BurnD, LeesA,AndresMV, et al. Aphase 2trial of the GSK-3 inhibitor tideglusib in progressive supranuclear palsy.Mov Disord 2014;29:470–8.

36. Taylor A, Harker JA, Chanthong K, Stevenson PG, Zuniga EI, Rudd CE.Glycogen synthase kinase 3 inactivation drives T-bet-mediated downre-gulation of co-receptor PD-1 to enhance CD8(þ) cytolytic T cell responses.Immunity 2016;44:274–86.

37. Boissonnas A, Fetler L, Zeelenberg IS, Hugues S, Amigorena S. In vivoimaging of cytotoxic T cell infiltration and elimination of a solid tumor. JExp Med 2007;204:345–56.

38. Lazarevic V, Glimcher LH. T-bet in disease. Nat Immunol 2011;12:597–606.

39. GrundyMA, Zhang T, Sentman CL. NK cells rapidly remove B16F10 tumorcells in a perforin and interferon-gamma independent manner in vivo.Cancer Immunol Immunother 2007;56:1153–61.

40. McMillan TJ, Rao J, Everett CA, Hart IR. Interferon-induced alterations inmetastatic capacity, class-1 antigen expression and natural killer cellsensitivity of melanoma cells. Int J Cancer 1987;40:659–63.

41. Takeda K, NakayamaM, Sakaki M, Hayakawa Y, Imawari M, Ogasawara K,et al. IFN-gamma production by lung NK cells is critical for the naturalresistance topulmonarymetastasis of B16melanoma inmice. J LeukocBiol2011;90:777–85.

42. Parameswaran R, Ramakrishnan P,Moreton SA, Xia Z,Hou Y, LeeDA, et al.Repression of GSK3 restores NK cell cytotoxicity in AML patients.Nat Commun 2016;7:11154.

43. Gill S, June CH. Going viral: chimeric antigen receptor T-cell therapy forhematological malignancies. Immunol Rev 2015;263:68–89.

44. Zhu Q, Yang J, Han S, Liu J, Holzbeierlein J, Thrasher JB, Li B. Suppressionof glycogen synthase kinase 3 activity reduces tumor growth of prostatecancer in vivo. Prostate 2011;71:835–45.

45. Klein PS, Melton DA. A molecular mechanism for the effect of lithium ondevelopment. Proc Natl Acad Sci U S A 1996;93:8455–9.

46. Piazza F, Manni S, Tubi LQ, Montini B, Pavan L, Colpo A, et al. GlycogenSynthase Kinase-3 regulates multiple myeloma cell growth and bortezo-mib-induced cell death. BMC Cancer 2010;10:526.

47. Dickey A, Schleicher S, Leahy K, Hu R, Hallahan D, Thotala DK. GSK-3beta inhibition promotes cell death, apoptosis, and in vivo tumorgrowth delay in neuroblastoma Neuro-2A cell line. J Neurooncol 2011;104:145–53.

48. Beurel E, Blivet-Van Eggelpoel MJ, Kornprobst M, Moritz S, Delelo R, PayeF, et al. Glycogen synthase kinase-3 inhibitors augment TRAIL-inducedapoptotic death in human hepatoma cells. Biochem Pharmacol 2009;77:54–65.

49. Finke JH, ZeaAH, Stanley J, LongoDL,MizoguchiH, TubbsRR, et al. Loss ofT-cell receptor zeta chain and p56lck in T-cells infiltrating human renal cellcarcinoma. Cancer Res 1993;53:5613–6.

50. Rayman P, Uzzo RG, Kolenko V, Bloom T, Cathcart MK, Molto L, et al.Tumor-induced dysfunction in interleukin-2 production and interleukin-2receptor signaling: amechanismof immune escape.Cancer J Sci Am2000;6Suppl 1:S81–87.

www.aacrjournals.org Cancer Res; 78(3) February 1, 2018 717

GSK-3 Downregulation of PD-1 in Cancer Therapy

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491

2018;78:706-717. Published OnlineFirst October 20, 2017.Cancer Res   Alison Taylor, David Rothstein and Christopher E. Rudd  Alternative to Antibody Blockade in Cancer TherapySmall-Molecule Inhibition of PD-1 Transcription Is an Effective

  Updated version

  10.1158/0008-5472.CAN-17-0491doi:

Access the most recent version of this article at:

  Material

Supplementary

  http://cancerres.aacrjournals.org/content/suppl/2017/10/20/0008-5472.CAN-17-0491.DC1

Access the most recent supplemental material at:

   

   

  Cited articles

  http://cancerres.aacrjournals.org/content/78/3/706.full#ref-list-1

This article cites 50 articles, 17 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/78/3/706.full#related-urls

This article has been cited by 3 HighWire-hosted articles. Access the articles at:

   

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

  Subscriptions

Reprints and

  .pubs@aacr.org

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/78/3/706To request permission to re-use all or part of this article, use this link

on September 21, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 20, 2017; DOI: 10.1158/0008-5472.CAN-17-0491