sirnasilencing of pd-l1 and pd-l2 on dendritic cells ......responses, the process of alloreactive...

IMMUNOBIOLOGY

siRNA silencing of PD-L1 and PD-L2 on dendritic cells augments expansion andfunction of minor histocompatibility antigen–specific CD8� T cellsWillemijn Hobo,1 Frans Maas,1 Niken Adisty,1 Theo de Witte,2 Nicolaas Schaap,3 Robbert van der Voort,1 and Harry Dolstra1

1Department of Laboratory Medicine–Laboratory of Hematology, and Departments of 2Tumor Immunology and 3Hematology, Radboud University NijmegenMedical Center, Nijmegen Center for Molecular Life Sciences, Nijmegen, The Netherlands

Tumor relapse after human leukocyteantigen–matched allogeneic stem celltransplantation (SCT) remains a seriousproblem, despite the long-term presenceof minor histocompatibility antigen(MiHA)–specific memory T cells. Den-dritic cell (DC)–based vaccination boost-ing MiHA-specific T-cell immunity is anappealing strategy to prevent or counter-act tumor recurrence, but improvement isnecessary to increase the clinical benefit.Here, we investigated whether knock-down of programmed death ligand 1 (PD-L1) and PD-L2 on monocyte-derived DCs

results in improved T-cell activation. Elec-troporation of single siRNA sequencesinto immature DCs resulted in efficient,specific, and long-lasting knockdown ofPD-L1 and PD-L2 expression. PD-L knock-down DCs strongly augmentedinterferon-� and interleukin-2 productionby stimulated T cells in an allogeneicmixed lymphocyte reaction, whereas noeffect was observed on T-cell prolifera-tion. Moreover, we demonstrated thatPD-L gene silencing, especially com-bined PD-L1 and PD-L2 knockdown, re-sulted in improved proliferation and cyto-

kine production of keyhole limpethemocyanin–specific CD4� T cells. Mostimportantly, PD-L knockdown DCsshowed superior potential to expandMiHA-specific CD8� effector and memoryT cells from leukemia patients early afterdonor lymphocyte infusion and later dur-ing relapse. These data demonstrate thatPD-L siRNA electroporated DCs are highlyeffective in enhancing T-cell proliferationand cytokine production, and are there-fore attractive cells for improving theefficacy of DC vaccines in cancer pa-tients. (Blood. 2010;116(22):4501-4511)

Introduction

Alloreactive CD8� T cells play a crucial role in graft-versus-tumor(GVT) responses after allogeneic stem cell transplantation (al-loSCT) and donor lymphocyte infusion (DLI).1,2 In human leuko-cyte antigen (HLA)–matched alloSCT, these alloreactive CD8�

T-cell responses are directed against minor histocompatibilityantigens (MiHAs), which are HLA-bound peptides derived frompolymorphic genes that differ between donor and recipient.2,3 Uponinfusion into the preconditioned patient, donor-derived MiHA-specific CD8� T cells clonally expand and become effector cells,which subsequently have the capacity to target MiHA-expressingmalignant cells.4 After the primary immune response, most MiHA-reactive CD8� T cells die through apoptosis, and only a small poolof long-lived memory cells that is able to respond quickly uponreencounter of the antigen survives.5 Importantly, we and othershave shown that after DLI, the emergence of MiHA-specific CD8�

effector T cells coincides with a decrease in the number ofmalignant cells.6-8 Moreover, years after the initial response,MiHA-specific CD8� memory T cells can still be detected, playingan important role in protective immunity in transplanted patients.However, these alloreactive CD8� memory T cells do not alwaysrespond efficiently to recurring tumor cells and this failure maycontribute to the occurrence of tumor relapses.6 This lack ofresponsiveness may result from inefficient memory T-cell activa-tion due to absence of effective antigen presentation, impairedcostimulation and/or enhanced signaling by coinhibitory receptors.Interestingly, we recently found that the programmed death (PD)–1/PD–ligand (PD-L) axis is involved in inhibiting the function ofMiHA-specific CD8� T cells upon their engagement with PD-L1

expressing myeloid leukemia cells and dendritic cells (DCs; Nordeet al, manuscript in preparation).

PD-1 is a coinhibitory receptor that is inducibly expressed byT and B cells upon activation. Engagement of this receptor witheither of its ligands, PD-L1 (B7-H1; CD274) and PD-L2 (B7-DC;CD273), results in phosphorylation of the cytoplasmic tail andnegatively regulates T-cell receptor (TCR)–induced cytokine pro-duction and proliferation via the phosphatidylinositol-3 kinasepathway.9,10 Whereas PD-L2 expression is mainly restricted toactivated lymphocytes and antigen-presenting cells (APCs), suchas DCs and macrophages, PD-L1 is also expressed by manynonlymphoid tissues and tumor cells.9,11 In the past few years, acrucial role for the PD-1/PD-L1 pathway has been demonstrated inchronic viral infections, such as HIV and hepatitis C virus(HCV).12-15 It has been shown that HIV- and HCV-specific T cellshave increased PD-1 expression, resulting in functional exhaustionof these cells. Moreover, PD-1/PD-L interactions have beenimplicated in the functional impairment of tumor antigen–reactiveCD8� T cells in solid tumors and chronic myeloid leukemia.16-18

Therefore, tumor cells may exploit the same pathway to escapeeradication by MiHA-specific CD8� memory T cells after alloSCT.

DC vaccination therapy would be an appealing strategy torevive functionally inactive MiHA-specific CD8� memory T cellsand, by this means, restore an effective immune response againstrecurring or persistent MiHA-expressing tumor cells. DCs havebeen widely exploited as adjuvants in vaccination therapy becauseof their capability to effectively initiate and reactivate T cell–basedimmune responses. In GVT and graft-versus-host disease (GVHD)

Submitted April 9, 2010; accepted July 25, 2010. Prepublished online as BloodFirst Edition paper, August 3, 2010; DOI 10.1182/blood-2010-04-278739.

The publication costs of this article were defrayed in part by page charge

payment. Therefore, and solely to indicate this fact, this article is herebymarked ‘‘advertisement’’ in accordance with 18 USC section 1734.

© 2010 by The American Society of Hematology

4501BLOOD, 25 NOVEMBER 2010 � VOLUME 116, NUMBER 22

For personal use only.on June 9, 2017. by guest www.bloodjournal.orgFrom

http://www.bloodjournal.org/

http://www.bloodjournal.org/site/subscriptions/ToS.xhtml

responses, the process of alloreactive T-cell stimulation andactivation is also orchestrated by MiHA-presenting DCs.19 Cur-rently, the most commonly used cell source for DC generation isperipheral blood monocytes, because high numbers of monocytescan be easily obtained and differentiated into mature DCs(mDCs).20,21 Using interleukin (IL)-4 and granulocyte-macrophagecolony-stimulating factor (GM-CSF), immature DCs (iDCs) areobtained after 5-7 days of culture, followed by 2 days of maturationin the presence of a cytokine cocktail containing IL-1�, IL-6, tumornecrosis factor � (TNF-�), and prostaglandin E2 (PGE2). Uponmaturation, the expression of peptide-HLA complexes is increased,and many accessory molecules involved in cell adhesion, cellmigration, and costimulation are acquired. For instance, CCR7plays an important role in DC migration to draining lymph nodes,whereas adhesion molecules CD54 and CD58 are involved inT-cell activation.22 Furthermore, mature DCs have a high expres-sion of costimulatory molecules CD80, CD86, and inducible T-cellcostimulator ligand (ICOS-L). However, coinhibitory molecules,such as PD-L1, PD-L2, herpes virus entry mediator (HVEM), andB7-H3, are also highly up-regulated. Expression levels and interac-tions of these co-stimulatory and -inhibitory ligands, with theircounterreceptors on T cells, ultimately determine the turn of thebalance toward antigen (Ag)–specific T-cell activation orinhibition.22-24

Here, we investigated whether the knockdown of PD-1 ligandsin monocyte-derived DCs results in improved T-cell function. Weidentified small-interfering RNAs (siRNAs) that efficiently si-lenced PD-L1 and PD-L2 expression after electroporation of iDCs.We demonstrated that PD-L gene silencing resulted in significantimprovement of interferon (IFN)-� and IL-2 production, andproliferation of keyhole limpet hemocyanin (KLH)-experiencedT cells. Moreover, PD-L knockdown DCs showed superior poten-tial to expand PD-1–expressing MiHA-specific CD8� effector andmemory T cells from leukemia patients shortly after DLI and, later,during relapse. Together, these findings indicate that PD-L–silenced DCs are attractive cells for improving the efficacy of DCvaccines in cancer patients.

Methods

Patient and donor material

Peripheral blood mononuclear cells (PBMCs) were isolated from healthydonor buffy coats (Sanquin Blood Bank). Blood samples were collectedafter written informed consent was obtained in ongoing clinical SCT andDC vaccination protocols, approved by the Radboud University NijmegenMedical Center (RUNMC) Institutional Review Boards. PBMC containingKLH-specific T cells were obtained from multiple myeloma (MM) patientsvaccinated with autologous DC, pulsed with KLH as an adjuvant andimmune monitoring tool, and various tumor-associated antigens in anongoing clinical trial. These patients received 3 DC vaccines intravenouslyand intradermally with biweekly intervals. PBMCs obtained 2 weeks afterthe second vaccination were used for functional studies, and DCs were

cultured from autologous apheresis material collected before the start ofvaccination therapy. To examine the stimulatory capacity of PD-L knock-down DCs on MiHA-specific T-cell responses, we used PBMCs obtainedfrom 3 leukemia patients who developed a MiHA-specific CD8� T-cellresponse after DLI. Patient UPN543 suffered from an acute myeloidleukemia (AML) and developed a strong LRH-1–specific CD8� T-cellresponse upon preemptive DLI. Then, the patient remained in completeclinical and cytogenetic remission until 3 years after DLI, whereuponextramedullary relapses developed without leukemic involvement in thebone marrow.25,26 Patient UPN640 had a blast crisis (BC) chronic myeloidleukemia (CML) and was treated with preemptive DLI after alloSCT. Thisresulted in a HA-1–specific CD8� effector T-cell response. The third patient(UPN389) suffered from an accelerated-phase (AP) chronic myeloidleukemia (CML) and was successfully treated with therapeutic DLI afteralloSCT. After DLI, a LRH-1–specific CD8� T-cell response was observed,which coincided with remission of CML-AP.6,26 However, despite thelong-term presence of LRH-1–specific CD8� memory T cells, the patientrelapsed 4 years after DLI. Because no apheresis material of the correspond-ing donors was available, DCs were cultured from apheresis material ofallogeneic HLA-B7�LRH-1� or HLA-A2�HA-1�donors. All cells of healthydonors and patients were obtained after written informed consent.

Cell culture

LRH-1–specific CD8� cytotoxic T lymphocyte (CTL) culture RP1 wasisolated from CML-AP patient UPN389 and recognized the 9-mer epitope,TPNQRQNVC, in the context of HLA-B*0702. CTL RP1 was cultured asdescribed previously.6,25 Recipient and donor Epstein-Barr virus (EBV)–transformed lymphoblastoid cell lines (LCLs) of patient UPN389 werecultured in Iscove modified Dulbecco medium (IMDM; Invitrogen) supple-mented with 10% fetal calf serum (FCS; Integro).

Generation of monocyte-derived DCs

Monocytes were isolated from PBMCs via plastic adherence in tissue-culture flasks (Greiner Bio-One). Immature DCs were generated byculturing monocytes in X-VIVO-15 medium (Lonza) supplemented with2% human serum (HS; PAA Laboratories), 500 U/mL IL-4 (Immunotools),and 800 U/mL GM-CSF (Immunotools). After 2-3 days, half of the mediumwas replaced with fresh X-VIVO-15/2% HS medium containing 1000 U/mL IL-4 and 1600 U/mL GM-CSF. Maturation of DCs was induced at day6-7 by culturing 0.5 � 106 immature (electroporated) DCs/mL in 6-wellplates (Costar; Corning) in X-VIVO-15/2% HS containing 500 U/mL IL-4,800 U/mL GM-CSF, 5 ng/mL IL-1ß, 15 ng/mL IL-6, 20 ng/mL TNF-� (allImmunotools), and 1 �g/mL PGE2 (Pharmacia & Upjohn). At day 8,mature DCs were harvested and used for T-cell stimulation, unless statedotherwise.

Stealth siRNA duplexes

Three different PD-L1 and PD-L2 Stealth siRNA duplexes and 2 recom-mended Stealth negative control siRNA duplexes, for either low or mediumGC content, were obtained from Invitrogen. The PD-L1 siRNA and PD-L2siRNA target sequences are listed in Table 1. All siRNAs were dissolved ordiluted to a concentration of 25�M in diethyl pyrocarbonate (DEPC) water(Invitrogen) and subsequently stored at �20°C.

Table 1. Target sequences of Stealth PD-L siRNAs

siRNA Sense sequence (5�3 3�) Antisense sequence (5�3 3�)

PD-L1 no. 1 GAGGAAGACCUGAAGGUUCAGCAUA UAUGCUGAACCUUCAGGUCUUCCUC

PD-L1 no. 2 CCUACUGGCAUUUGCUGAACGCAUU AAUGCGUUCAGCAAAUGCCAGUAGG

PD-L1 no. 3 UGAUACACAUUUGGAGGAGACGUAA UUACGUCUCCUCCAAAUGUGUAUCA

PD-L2 no. 1 GCCUGGAAUUGCAGCUUCACCAGAU AUCUGGUGAAGCUGCAAUUCCAGGC

PD-L2 no. 2 GGGACUACAAGUACCUGACUCUGAA UUCAGAGUCAGGUACUUGUAGUCCC

PD-L2 no. 3 CCCUCCUGCAUCAUUGCUUUCAUUU AAAUGAAAGCAAUGAUGCAGGAGGG

4502 HOBO et al BLOOD, 25 NOVEMBER 2010 � VOLUME 116, NUMBER 22




DC electroporation

Immature DCs were harvested at days 6-7, then washed once with Hanksbalanced salt solution (HBSS; Lonza) and twice with phenol red freeOptimem buffer (Gibco Invitrogen). Cells were resuspended in Optimembuffer and � 7.5 � 106 cells were transferred to a 0.4-cm gene-pulsercuvette (Bio-Rad). Subsequently, 0.25 nmol of a single siRNA or, in thecase of double knockdown, 0.125 nmol per siRNA, was added. DCs wereelectroporated at 300 V, 150 �F using a Bio-Rad Genepulser II. Directlyafter electroporation, cells were resuspended in prewarmed phenol red freeX-VIVO-15/7% HS and incubated at 37°C. After 1 hour, phenol red freemedium was replaced with conventional maturation medium. Cells werecultured for 1 to 5 days.

RNA isolation and real-time quantitativereverse-transcription PCR

Total RNA was isolated from mature PD-L knockdown DCs using theQuick-RNA–miniprep (Zymo Research Corporation). cDNA synthesis andpolymerase chain reaction (PCR) amplification were performed as previ-ously described.27,28 The hydroxymethylbilane synthase (HMBS) housekeep-ing gene was used to normalize gene expression. One microliter of cDNAwas amplified in a 50-�L reaction mixture containing 1.25 U AmpliTaqGold (Applied Biosystems), 300nM gene-specific forward and reverseprimer, 100-300nM gene-specific Taqman probe (100nM for PD-L2 and300nM for PD-L1 and HMBS), 250�M of each deoxynucleotide triphos-phate (dNTP), 5mM MgCl2, and 1� Taqman PCR buffer (AppliedBiosystems). The following gene-specific primers and Taqman probes wereused: PD-L1; PD-L1 FW 5�-CATCTTATTATGCCTTGGTGTAGCA-3�,PD-L1 RV 5�-GGATTACGTCTCCTCCAAATGTG-3� and PD-L1 MGBprobe 5�-(TET)-ACATTCATCTTCCGTTTAAG-3�, PD-L2; PD-L2 FW5�-CAACTTGGCTGCTTCACATTTT-3�, PD-L2 RV 5�-TGTGGTGACAG-GTCTTTTTGTTGT-3� and PD-L2 probe 5�-(TET)-TTCATAGCCACAGT-GATAGCCCTAAGAAAACAACTCT-(TAMRA)-3�, HMBS; HMBS-FW5�-GGCAATGCGGCTGCAA-3�, HMBS-RV 5�-GGGTACCCACGC-GAATCAC-3�, and HMBS-probe 5�-(VIC)-CTCATCTTTGGGCTGTTT-TCTTCCGCC-(TAMRA)-3�. PCR amplification was performed using anABI Prism 7700 (Applied Biosystems), with the following PCR conditions:enzyme activation for 10 minutes at 95°C, followed by 45 cycles of 95°Cfor 15 seconds, and 60°C for 1 minute. PD-L1 and PD-L2 Ct values werenormalized for HMBS by calculating Ct Ctgene � CtHMBS per sample.Finally, PD-L1 and PD-L2 mRNA expression levels were quantifiedrelative to DCs electroporated without siRNA as follows:2�(Ct electroporated with siRNA � Ct electroporated without siRNA).

Flow cytometry

Phenotype and maturation state of DCs was analyzed by staining with thefollowing antibodies: anti-CD14 (clone TUK4, Dako), anti-CD80 (cloneMAB104), anti-CD83 (clone HB15a), anti-CD86 (clone HA5.2B7, all fromBeckman Coulter), anti-PD-L1 (clone MIH1), anti-PD-L2 (clone MIH18,both from Becton Dickinson), and isotype controls IgG1 FITC/PE dual-color control (Dako) and IgG2b phycoerythrin (PE; Beckman Coulter).LRH-1– and HA-1–specific T cells were detected by staining cell suspen-sions with PE-labeled LRH-1/HLA-B7 tetramers containing peptide TPN-QRQNVC and PE-labeled HA-1/HLA-A2 tetramers containing peptideVLHDDLLEA, respectively. Tetramers were kindly provided by Prof DrFred Falkenburg (Department of Hematology, Leiden University MedicalCenter, Leiden, The Netherlands). T-cell cultures were incubated with1.5-2 �g/mL tetramer for 15 minutes at room temperature. Subsequently,cells were labeled with the appropriate concentrations of anti-CD8 (ProIm-mune) and anti-CD3 (Beckman Coulter) for 30 minutes at 4°C. Afterwashing with phosphate-buffered saline (PBS)/0.5% bovine serum albumin(BSA; Sigma-Aldrich), cells were resuspended in washing buffer contain-ing 0.1% 7-amino-actinomycin D (7-AAD; Sigma-Aldrich). Cells wereanalyzed using the Coulter FC500 flow cytometer (Beckman Coulter).

Allogeneic mixed lymphocyte reactions (MLRs)

Allogeneic CD4� and CD8� T cells were isolated from nonadherent orcomplete PBMC fractions by direct magnetic labeling with the appropriateMicroBeads (Miltenyi Biotec), following the manufacturer’s instructions.Cells were resuspended in IMDM/10% HS to a concentration of0.5 � 106 cells/mL and stimulated with maturated PD-L knockdown DCs atT cell:DC ratios ranging from 1:0.1 to 1:0.0125. These cocultures wereperformed in 48-well plates (Cellstar; Greiner Bio-one) in a total volume of1 mL IMDM/10% HS. After 1-5 days of coculture, supernatant washarvested for cytokine analysis. At day 5, cells were resuspended andtransferred to 96-well round-bottomed plates (Costar; Corning). Subse-quently, 0.5�Ci [3H]-thymidine (PerkinElmer) was added to each well.After overnight incubation, [3H]-thymidine incorporation was measuredusing a 1205 Wallac Betaplate counter (PerkinElmer).

KLH-specific T-cell activation assays

Autologous monocyte-derived DCs were cultured from apheresis materialof MM patients as described above. At day 3, DCs were pulsed with orwithout 50 �g/mL KLH (Vacmune; Biosyn Corporation). Patient PBMCscontaining KLH-specific T cells were thawed, resuspended in IMDM/10%HS to a concentration of 0.5 � 106 cells/mL, and mixed with autologous2-day mature PD-L knockdown DCs at a T cell:DC ratio of 1:0.05. Thesecocultures were performed in 6-fold in 96-well round-bottomed plates(Costar; Corning) in a total volume of 200 �L. After 1 and 5 days ofcoculture, supernatant was harvested for cytokine analysis. At day 5,proliferation was analyzed using [3H]-thymidine, as described before.

MiHA-specific T-cell expansion assays

LRH-1–specific CTL RP1 was resuspended in IMDM/10% HS to aconcentration of 0.05 � 106 cells/mL and plated in 96-well round-bottomedplates (Costar; Corning). HLA-B7� PD-L knockdown DCs were loadedwith or without 10�M LRH-1 peptide TPNQRQNVC for 30 minutes atroom temperature and, subsequently, cocultured with CTL at a stimulationratio of 1:1 in a total volume of 200 �L. After 1 day, 100 �L of supernatantwas harvested, and fresh IMDM/10% HS containing 100 U/mL IL-2(Chiron) was added. At day 4, CTL proliferation was analyzed using[3H]-thymidine, as described before, and supernatant was analyzed forcytokine levels.

MiHA-specific CD8� memory T cells present in PBMCs from patientsUPN543, UPN640, and UPN389 were stimulated for 1-3 consecutiveweeks ex vivo with unloaded or MiHA peptide-loaded PD-L knockdownDCs. Mature allogeneic PD-L knockdown or control DCs, cultured fromapheresis material of a HLA-B7�LRH-1� or HLA-A2�HA-1� donor, wereloaded with or without 5�M MiHA peptide (LRH-1: TPNQRQNVC;HA-1: VLHDDLLEA) for 30 minutes at room temperature. PBMCs andDCs were subsequently cocultured at a ratio of 1:0.1 in 2 mL IMDM/10%HS in 24-well plates (Costar; Corning). After 5 days, CD4� T cells weredepleted from the cultures by direct magnetic labeling with CD4 IMagbeads (Becton Dickinson). CD4-depleted fractions were resuspended inIMDM/10% HS supplemented with 50 U/mL IL-2 and 5 ng/mL IL-15(Immunotools). At day 6 or 7, cells were harvested and counted, and thepercentage of LRH-1– or HA-1–specific CD8� T cells was determined byflow cytometry.

CD107a degranulation assay

PBMC cultures stimulated for 2 consecutive weeks with LRH-1 peptide-loaded PD-L knockdown DCs were used in a CD107a secretion assay todetermine functional recognition of the LRH-1 antigen by measuring CTLdegranulation. As target cells, we used the corresponding recipient EBV-LCLs and donor EBV-LCLs loaded with or without 5�M LRH-1 peptide.Briefly, 2 � 105 cells of each T-cell culture were stimulated with target cellsat an effector:target ratio of 1:1 in a total volume of 500 �L of IMDM/10%HS supplemented with 25 U/mL of IL-2 and 6.5 �L of anti-CD107a(Becton Dickinson). After overnight coculture, the percentage of CD107a�

cells was determined within the LRH-1-tetramer� T-cell population usingflow cytometry.

KNOCKDOWN OF PD-L ON DCs AUGMENTS T-CELL FUNCTION 4503BLOOD, 25 NOVEMBER 2010 � VOLUME 116, NUMBER 22




Cytokine analyses

IFN-� and granzyme B levels in culture supernatants were analyzed usingenzyme-linked immunosorbent assays (ELISAs; IFN-�: Pierce Endogen;granzyme B: MabTech). Release of IL-2, IL-4, IL-5, TNF-�, and IFN-� bythe T cells was simultaneously determined in pooled supernatant using aTh1/Th2 BD cytometric bead array (Becton Dickinson), following themanufacturer’s protocol, and measured using flow cytometry.

Statistical analysis

Statistical significance of differences between PD-L knockdown DCs andmedium GC siRNA control DCs were analyzed using one-way analysis ofvariance (ANOVA), followed by a Bonferroni posthoc test. P values� .05 were considered significant.

Results

Efficient, specific, and long-lasting siRNA–mediatedknockdown of PD-L1 and PD-L2 on DCs

To determine the efficacy and specificity of 3 different siRNAs insilencing PD-L1 and PD-L2 expression on monocyte-derived DCs,we studied the magnitude, off-target effects, and duration ofknockdown on mRNA and protein level. Immature DCs wereelectroporated with or without 0.25 nmol siRNA and subsequentlycultured in maturation medium containing IL-1�, IL-6, TNF-�, andPGE2. All 3 PD-L1 siRNAs appreciably reduced PD-L1 mRNA

Figure 1. Efficacy, specificity, and duration of siRNA-mediated PD-L1 and PD-L2 silencing on DCs. Immature DCs were electroporated with 0.25 nmol PD-L1, PD-L2, ornegative control (MED GC vs LOW GC) siRNA duplexes and subsequently cultured in maturation medium containing IL-1�, IL-6, TNF-�, and PGE2 for 2 days (A-D) or 1-5 days(E). (A) PD-L1 and PD-L2 mRNA expression was measured and subsequently normalized for HMBS expression using RT Q-PCR. PD-L mRNA expression in DCselectroporated without siRNA was set to 1. (B) PD-L1 and PD-L2 protein expression (black lines), compared with isotype control (gray histograms), was analyzed using flowcytometry. The numbers in the plots represent the mean fluorescence intensity (MFI). The data of 1 representative donor of 6 is shown. (C) Percentage of PD-L1� or PD-L2�

DCs was determined for 3-6 donors by flow cytometry. (D) Expression of maturation and costimulatory molecules by electroporated DCs was analyzed using flow cytometry.The bars represent DCs electroporated without siRNA (black), with MED GC control siRNA (white), PD-L1 siRNA 2 (dark gray), or PD-L2 siRNA 2 (light gray). Data areexpressed as mean � SEM of 3-6 donors. (E) Percentage of PD-L1� or PD-L2� DCs over time was determined by flow cytometry. The percentage of PD-L� DCselectroporated without siRNA was set to 100%. The lines represent DCs electroporated without siRNA (dotted line, F), with MED GC control siRNA (dotted line, E), PD-L1siRNA 2 (solid dark gray line), or PD-L2 siRNA 2 (solid light gray line). The data are representative of 2 independent experiments on 2 different donors.





levels at 2 days after electroporation, while the negative controlsiRNAs had no effect (Figure 1A). Similarly, all 3 PD-L2 siRNAsefficiently knocked down PD-L2 mRNA expression to the sameextent. To validate knockdown on the protein level, we analyzedPD-L1 and PD-L2 surface expression on siRNA-electroporatedDCs 2 days after maturation. PD-L1 siRNA number 2 induced themost pronounced reduction of PD-L1 cell-surface expression,resulting in an average of 21% PD-L1� DCs (n 6; Figure 1B-C).In the case of PD-L2, all 3 siRNAs were equally effective, resultingin � 10% PD-L2� DCs. Therefore, we decided to continue withPD-L1 and PD-L2 siRNA number 2 and the accompanyingmedium GC-negative control siRNA. To verify that our siRNAs ofinterest did not affect DC maturation or have other off-targeteffects, protein expression levels of various maturation and costimu-latory molecules were examined 2 days after electroporation. Noneof the molecules were affected by the siRNAs tested (Figure 1D).Finally, we determined the duration of the siRNA-mediatedknockdown effects by assessing PD-L1 and PD-L2 protein expres-sion over time upon DC electroporation. After 2 days, PD-L1 andPD-L2 were both maximally down-regulated by the respectivesiRNAs and expression remained low for at least 3 more days(Figure 1E). Collectively, these data show that siRNA electropora-tion of DCs results in efficient, specific, and long-lasting silencingof PD-L1 and PD-L2.

PD-L knockdown DCs show increased capacity to stimulatecytokine production in allogeneic MLR assays

To investigate whether knockdown of PD-L1 and/or PD-L2 usingthe selected siRNAs (ie, PD-L1 and PD-L2 target sequence no. 2 inTable 1) improved the stimulatory capacity of DCs, allogeneicT cells were cocultured for 5 days with mature PD-L knockdownDCs. Thereafter, T-cell proliferation and cytokine production wasanalyzed. A strong, dose-dependent increase in IFN-� productionwas observed upon stimulation with PD-L knockdown DC, butPD-L knockdown DCs were not capable of improving T-cellproliferation of allogeneic CD4� or CD8� T cells (Figure 2A-B).For CD4� T cells, IFN-� levels were enhanced up to 7-fold byPD-L1 knockdown DCs, whereas PD-L2 knockdown DCs onlymodestly improved IFN-� production (Figure 2B). Interestingly,stimulation with PD-L1/L2 double-knockdown DCs resulted in a10-fold higher IFN-� production, indicating a synergistic effectwhen PD-L1 and PD-L2 are knocked down simultaneously.Although CD8� T cells produced less IFN-� than CD4� T cells, astill considerable elevation of IFN-� levels was observed uponstimulation with PD-L knockdown DC. Especially, PD-L1 single-and PD-L1/L2 double-knockdown DCs improved IFN-� levels upto 3-fold, while the increase by PD-L2 single knockdown DCs wasagain modest.

Figure 2. PD-L knockdown DCs show enhanced stimulation of allogeneic T cells. Immature DCs were electroporated with 0.25 nmol of each single siRNA or 0.125 nmolof each siRNA in case of double knockdown. After 2 days of maturation, DCs were mixed with allogeneic T cells. (A) At day 5 of coculture, [3H]-thymidine was added to theculture, and the following day, T-cell proliferation capacity was determined by measuring [3H]-thymidine incorporation. Data are depicted as mean � SD and are representativefor 2 different donors. CPM counts per minute. (B) IFN-� levels were measured in the supernatant at day 5 of coculture using ELISA. Data are depicted as mean � SD andare representative for 3 different donors. (C) Cytokines produced by CD4� and CD8� T cells of the same donor stimulated with allogeneic DCs at a ratio of 1:0.1 were analyzedsimultaneously at days 1-5 of the coculture. Statistical analysis was performed using one-way ANOVA, followed by a Bonferroni post-hoc test. *P � .05, **P � .01, ***P � .001.





To get more insight into the levels and kinetics of variouscytokines produced by the different T-cell subsets upon stimulationwith PD-L knockdown DCs, coculture supernatant was harvestedover time and analyzed. Early after stimulation with DCs, alloge-neic T cells started producing IL-2, while production of IFN-�,TNF-�, and IL-5 initiated later (Figure 2C). No IL-4 was produced(data not shown). Overall, highest cytokine levels were observed incocultures with CD4� T cells. Furthermore, over time, PD-Lknockdown DCs evidently elevated IL-2, IFN-�, TNF-�, and IL-5production by CD4� as well as CD8� T cells. For all cytokines, thehighest levels were obtained after stimulation with the PD-L1/L2double-knockdown DCs.

Altogether, these results demonstrate that stimulation withPD-L knockdown DCs has no effect on T-cell proliferation, butprofoundly enhances the cytokine production capacity of stimu-lated T cells in a primary allogeneic MLR.

KLH-specific T-cell responses are strongly augmented by PD-Lknockdown DCs

To examine the stimulatory capacity of PD-L-silenced DCs insecondary responses by antigen-experienced T cells, we culturedPBMCs containing primed KLH-specific T cells together withautologous KLH-pulsed PD-L knockdown DCs. After 1-5 days,KLH-specific proliferation and cytokine production of DC-stimulated T cells was studied. Interestingly, the proliferativecapacity of primed KLH-specific T cells could be significantlyboosted with PD-L knockdown DCs (Figure 3A). Using PBMCsfrom MM patient 1, all 3 knockdown DC types increased KLH-specific T-cell proliferation, whereas in patient 2, significant

induction of proliferation was only seen upon stimulation withPD-L1/PD-L2 double-knockdown DCs. In accordance with theresults found in the allogeneic MLR setting, stimulation with PD-Lknockdown DCs resulted in increased cytokine production byKLH-specific T cells (Figure 3B). Stimulation with PD-L1 knock-down DCs resulted in 2-5� higher IFN-� levels than aftercoculture with control siRNA DCs. Despite the modest elevation ofIFN-� production after stimulation with PD-L2-silenced DCs, anevident synergistic augmentation (ie, 8-14-fold) of IFN-� levelswas observed when KLH-specific T cells were stimulated withPD-L1/L2 double-knockdown DCs. Already 1 day after stimula-tion with KLH-pulsed DCs, IL-2 production was markedly el-evated by all PD-L knockdown DC types (Figure 3C). Again, themost evident enhancement was seen for PD-L1 single and PD-L1/L2 dual knockdown DCs. Moreover, in PD-L1 single andPD-L1/L2 double-knockdown DC-stimulated cultures, IL-2 levelsremained considerably higher up to 5 days after stimulation of theKLH-specific T cells. These data demonstrate that KLH-pulsedPD-L knockdown DCs augment proliferation and strongly increasecytokine production by primed KLH-specific T cells of DC-vaccinated cancer patients.

MiHA-specific CTL expansion and function is enhanced byPD-L silenced DCs

Recently, we have observed that activated LRH-1–specific CD8�

T cells express PD-1, and that in vitro antibody blockade of PD-1signaling results in improved stimulation of LRH-1–specific CD8�

T cells by PD-L1–expressing DCs and myeloid leukemia cells(Norde et al, manuscript in preparation). To investigate the effect of

Figure 3. PD-L siRNA-silenced DCs show an increased capacity to stimulate KLH-specific T-cell proliferation and cytokine production in vaccinated cancerpatients. Autologous PD-L knockdown DCs pulsed with or without KLH were cocultured with PBMCs from vaccinated MM patients at a stimulation ratio of 0.1:1. (A) At day 5,[3H]-thymidine was added to the culture, and the following day, T-cell proliferation capacity was determined by measuring [3H]-thymidine incorporation. Representative data of2 of 5 patients are shown, and bars are depicted as mean � SD. CPM indicates counts per minute. (B) IFN-� levels were measured in the supernatant at day 5 of the coculture.Representative data of 2 of 5 patients are shown, and bars are depicted as mean � SD. (C) IL-2 levels were measured in coculture supernatant of days 1 and 5. Representativedata of 2 of 5 patients are shown. Statistical analysis was performed using one-way ANOVA, followed by a Bonferroni post-hoc test. *P � .05, **P � .01, ***P � .001.





PD-L silencing on MiHA-specific CD8� T-cell expansion andfunctionality initiated by DCs, we stimulated LRH-1–specific CTLclone RP1 with PD-L knockdown DCs loaded with or withoutLRH-1 peptide for 1-4 days. After DC stimulation, we measuredthe proliferative capacity as well as production of IFN-� andgranzyme B by LRH-1–specific CTL RP1. After 4 days ofcoculture, LRH-1–dependent proliferation of CTL RP1 was ob-served (Figure 4A). Single PD-L1 and dual PD-L1/PD-L2 knock-down significantly enhanced DC-mediated CTL proliferation,while the effect of PD-L2 knockdown was only modest. Further-more, we observed increased IFN-� production by LRH-1–specificCTL after stimulation with all 3 LRH-1 peptide-loaded PD-Lknockdown DCs (Figure 4B). Moreover, production of granzyme Bby CTL RP1 was augmented by PD-L1–silenced DCs upon peptiderecognition (Figure 4C). These results indicate that PD-L–silencedDCs are able to increase proliferation rate and degranulationcapacity of MiHA-specific CTLs.

PD-L knockdown DCs effectively boost ex vivo expansion andfunction of MiHA-specific CD8� effector and memory T cells

To investigate the effect of stimulation with MiHA-presenting,PD-L–silenced DCs on MiHA-specific CD8� effector and memoryT-cell proliferation, we used PBMCs from 3 different leukemiapatients collected early after SCT and DLI or during relapseddisease. First, we cultured PBMCs of patients UPN543 andUPN640, both obtained 7 months after DLI during the effectorresponse, with MiHA-presenting, PD-L–silenced DCs for 1 week.At start, PBMCs contained 0.12% LRH-1–specific and 0.14%

HA-1–specific CD8� T cells, respectively. In both patients, weobserved augmented percentages of MiHA-specific CD8� T cellsafter stimulation with MiHA-loaded, PD-L–silenced DCs, com-pared with control DCs (Figure 5A-B). Although the best effectswere observed upon stimulation with PD-L1/L2 double-knock-down DCs, all PD-L knockdown DC types evidently increased thenumber of LRH-1–specific CD8� T cells of patient UPN543 after1 week of culture ranging from 1.5- to 2.7-fold (Figure 5C). In thecase of patient UPN640, the effects were even more pronounced.One-week culture with PD-L knockdown DCs resulted in consider-ably augmented expansion of HA-1–specific CD8� T cells of up to12-fold for PD-L1/L2 double-knockdown DCs (Figure 5D).

Next, we investigated the effect of repetitive stimulations withMiHA-presenting PD-L–silenced DCs on MiHA-specific CD8�

memory T cells. For this, we used PBMCs from CML-AP patientUPN389 and AML patient UPN543 containing, respectively,0.06% and 0.09% LRH-1–specific CD8� T cells, collected at64 (UPN389) and 71 (UPN543) months post-DLI during relapseddisease. These PBMCs were stimulated for 2 or 3 consecutiveweeks with LRH-1 peptide-loaded DCs, and the number ofLRH-1–specific CD8� T cells was weekly determined by flowcytometry. After stimulation with LRH-1 peptide-loaded controlDCs, a specific increase in the percentage of LRH-1-tetramer�

CD8� T cells of 8% was observed over time in patientUPN389 (Figure 6A). Interestingly, PD-L1 single- and PD-L1/PD-L2 double-knockdown DCs improved LRH-1–specific CD8�

memory T-cell expansion up to 12% and 29%, respectively.Surprisingly, lower percentages of LRH-1-specific CD8� T cellswere detected upon stimulation with LRH-1 peptide-loaded PD-L2knockdown DCs. Because absolute T-cell numbers and the percent-age of CD8� T cells were higher in PD-L–silenced DC-stimulatedcultures compared with T-cell cultures stimulated with controlDCs, we also calculated the cumulative number of LRH-1–specificCD8� T cells over time (Figure 6B). In the first week, the foldexpansion upon stimulation with PD-L1 single and PD-L1/L2double-knockdown DCs was, respectively, 1.6 and 4.7� higherthan that of LRH-1 peptide-loaded control DCs. Ultimately,stimulation with PD-L1 single or PD-L1/L2 double-knockdownDCs resulted in a 2.8 and 23.3� higher number of LRH-1–specificCD8� T cells by the end of week 3.

In addition, for AML patient UPN543, we observed a specificincrease in the percentage of LRH-1-tetramer� CD8� T cells to0.7% upon repetitive stimulations with LRH-1 peptide-loaded DCs(Figure 6C). Silencing of PD-1 ligands on these DCs resulted inimproved percentages of LRH-1–specific CD8� T cells of 2.5%and 10% for PD-L single and PD-L1/L2 double-knockdown DCs,compared with control DCs (0.7%). In absolute numbers, theadvantage of PD-L knockdown DCs over control DCs in augment-ing LRH-1–specific CD8� T-cell expansion is even more evident(Figure 6D). Eventually, this resulted in 13.3, 26.7, and 61� highernumbers of tetramer� CD8� T cells by PD-L1, PD-L2, andPD-L1/L2 double-knockdown DCs, respectively.

Finally, we studied the functional capacity of the expandedLRH-1–specific CD8� T cells of patient UPN389 after 2 weeks ofstimulation with LRH-1 peptide-loaded PD-L knockdown orcontrol DCs. These DC-stimulated CD8� T-cell cultures wereincubated overnight with different target cells in the presence ofanti-CD107a to determine the percentage of degranulating CD107a�

LRH-1–specific CD8� T cells. We observed that stimulation withrecipient EBV-LCL and LRH-1 peptide-loaded donor EBV-LCLresulted in 75% CD107a� LRH-1–specific CD8� T cells(Figure 6E), whereas stimulation with donor EBV-LCL or medium

Figure 4. Expansion and function of MiHA-specific CTL RP1 is augmented byPD-L knockdown DCs. Two days after electroporation with PD-L1 and/or PD-L2siRNA, mature HLA-B7� DCs were loaded with or without 10�M LRH-1 peptide andcocultured with MiHA-specific CTL RP1 at a stimulation ratio of 1:1. PD-L knockdownresulted in � 20% PD-L1� and/or � 1% PD-L2� DC. (A) After 4 days, proliferationwas determined following overnight [3H]-thymidine incorporation. Data of 1 of2 independent experiments are shown, and bars are depicted as mean � SD.CPM counts per minute. (B) IFN-� and (C) granzyme B levels were measured inCTL RP1 supernatant obtained at day 1 of coculture. Statistical analysis wasperformed using one-way ANOVA, followed by a Bonferroni post-hoc test. *P � .05,***P � .001.





resulted in � 25% nonspecific degranulation of the LRH-1–specific CD8� T cells. A slight difference in degranulation capacitywas seen between the LRH-1–specific CD8� T cells stimulatedwith PD-L knockdown DCs, in comparison to control DCs.Together, these results demonstrate that MiHA-specific CD8�

effector and memory T-cell expansion can be efficiently boostedusing PD-L knockdown DCs, and that these expanded MiHA-specific CD8� T cells have the capacity to degranulate uponrecognition of MiHA� target cells.

Discussion

Alloreactive CD8� T cells targeting MiHA on malignant cells ofthe recipient play a key role in GVT immunity after alloSCT andDLI. However, MiHA-specific CD8� T-cell responses inducedposttransplantation are in a substantial number of patients notsufficient to sustain complete remission of the malignant disease. Itis evident that distinct mechanisms are involved in dampeningantitumor T-cell responses, which may allow malignant cells toescape immune destruction. Among these mechanisms, T-cellinhibition or even exhaustion due to signaling of the PD-1/PD-Lpathway may contribute to abrogation of immune responses bylimiting the expansion and functionality of CD8� T cells.12-18

Recently, we showed that MiHA� leukemia can relapse withoutinducing MiHA-specific CD8� memory T-cell expansion, suggest-ing that these memory T cells are either inactive or are not activatedby MiHA-presenting APC.6,25 Furthermore, we found that likevirus-specific CD8� T cells in chronic viral infections and tumor-

infiltrating T cells in solid tumors, the PD-1/PD-L pathway isinvolved in the impairment of MiHA-specific CD8� T-cell function(Norde et al, manuscript in preparation).

DC-based vaccination is an attractive strategy to prevent or treatrecurrence of tumor cells by boosting MiHA-specific CD8� T-cellimmunity.26 Although results of DC vaccination therapy in cancerpatients are promising, the overall clinical outcome is not satisfac-tory yet.20 Previously, we have shown that mature DCs have thepotential to revive impaired MiHA-specific memory T cells fromrelapsed leukemia patients.26 However, mature DCs do not onlyexpress costimulatory molecules, but also highly express PD-L1and PD-L2, which may inhibit their stimulatory capacity uponencounter of PD-1� alloreactive CD8� T cells. Here, we investi-gated whether PD-L1 and PD-L2 knockdown in DCs results inimproved T-cell stimulation. We silenced PD-L1 and PD-L2expression on monocyte-derived DCs using Stealth Select siRNAduplexes, which have high specificity, longevity and stability, andlimited induction of nonspecific cellular stress-response pathways.Optimization experiments revealed that electroporation of iDCs,compared with mDCs, resulted in more efficient siRNA-mediatedknockdown of PD-L2 and, especially, PD-L1 (data not shown).Therefore, we decided to continue with iDC electroporationfollowed by 2 days of maturation using the conventional cytokinecocktail, containing IL-1�, IL-6, TNF-�, and PGE2. Testing of3 different siRNA sequences demonstrated that PD-L2 expressionlevels could be efficiently down-regulated by all 3 duplexes afterone electroporation, while duplex number 2, located on theboundary of exons 2 and 3, was the most efficient PD-L1 siRNAsequence. Our results contrast to findings reported by Breton et al,

Figure 5. MiHA-specific CD8� effector T-cell expansion can be enhanced by PD-L–silenced DCs. Two days after electroporation with PD-L1 and/or PD-L2 siRNA, matureHLA-B7�LRH-1� or HLA-A2�HA-1� DCs were loaded with or without 5�M MiHA peptide. PD-L knockdown resulted in � 17% PD-L1� and/or � 5% PD-L2� DCs. PatientPBMCs, containing MiHA-specific CD8� effector T cells, stimulated with PD-L knockdown DCs at a ratio of 1:0.1, were screened after 1 week for tetramer� CD8� T cells usingflow cytometry. The numbers in the FACS plots represent the percentage of (A) LRH-1– (UPN543) or (B) HA-1–specific (UPN640) CD8� T cells in the total CD3�CD8� T-cellpopulation. Total numbers of LRH-1– (C) or HA-1–specific (D) CD8� T cells obtained after 1 week of stimulation with PD-L knockdown DCs loaded with or without MiHA peptide.





Figure 6. PD-L knockdown DCs boost the proliferative capacity and function of MiHA-specific CD8� memory T cells. Two days after electroporation with PD-L1 and/orPD-L2 siRNA, mature HLA-B7�LRH-1� DCs were loaded with or without 5�M LRH-1 peptide. PD-L knockdown resulted in � 5% PD-L1� and/or � 2% PD-L2� DCs.(A) PBMCs of patient UPN389, consecutively stimulated with PD-L knockdown DCs at a ratio of 1:0.1, were weekly screened for tetramer� CD8� T cells using flow cytometry.The numbers in the FACS plots represent the percentage of LRH-1–specific CD8� T cells in the total CD3�CD8� T-cell population. Data are representative for 2 independentexperiments. (B) Cumulative numbers of LRH-1–specific CD8� T cells of patient UPN389 obtained after 3 consecutive stimulations with PD-L knockdown DCs loaded with orwithout LRH-1 peptide. (C) Percentage of LRH-1–specific CD8� T cells in PBMCs of patient UPN543. PBMCs were consecutively stimulated with PD-L knockdown DCs at aratio of 1:0.1 and weekly screened for tetramer� CD8� T cells using flow cytometry. (D) Cumulative numbers of LRH-1–specific CD8� T cells of patient UPN543 obtained after2 consecutive stimulations with PD-L knockdown DCs loaded with or without LRH-1 peptide. (E) Degranulation of LRH-1–specific CD8� T cells of patient UPN389, stimulatedfor 2 weeks with LRH-1 peptide-loaded PD-L knockdown DC, was measured by staining for CD107a during overnight stimulation with different target cells. PBMCs werecultured 1:1 with LRH-1� donor EBV-LCL, LRH-1 peptide-loaded donor EBV-LCL, LRH-1� recipient EBV-LCL, or in medium. ND indicates not determined.





who described that a 2-step electroporation method was necessaryto obtain sufficient PD-L1 knockdown on monocyte-derived DCs.29

The difference in efficacy could be related to the used PD-L1siRNA sequence or different electroporation conditions. Our se-lected PD-L1 and PD-L2 siRNAs showed specific effects for thetargeted proteins and did not affect DC maturation or expression ofcostimulatory molecules. Furthermore, the siRNA-mediated PD-Lknockdown lasted for at least 5 days. After identifying the mostefficient siRNAs and electroporation protocol, we determined thestimulatory capacity of PD-L knockdown DCs in primary andsecondary antigen-specific T-cell responses.

PD-L-silenced DCs significantly increased cytokine production ofCD4� and CD8� T cells in allogeneic MLR assays; however, noimproved effect of PD-L knockdown DCs on the induction of prolifera-tive CD4� and CD8� T-cell responses was observed. In the early phaseof T-cell stimulation, we observed increased production of IL-2 and, inthe later phase, enhanced production of IFN-�, TNF-�, and IL-5, but notIL-4. These data indicate that PD-L–silenced DCs improve CD4� Th1and CD8� T-cell responses. Other studies reported increased T-cellproliferation and cytokine production during primary allogeneic MLRresponses after PD-L blockade using antibodies, especially upon initia-tion by weaker DCs, such as iDCs and IL-10–pretreated mDCs.30

However, more profound stimulatory effects by PD-L1 and PD-L2blockade were observed on memory and recently activated T cellsduring secondary immune responses.31 Therefore, we tested our PD-Lknockdown DCs also in an antigen-experienced setting using PBMCs ofMM patients containing KLH-specific T cells after vaccination withautologous KLH-pulsed DCs. Although effects vary to some extent, weobserved, in 5 different patients, significant improvement of KLH-specific T-cell proliferation after stimulation with KLH-pulsed, PD-L–silenced DCs. Furthermore, PD-L knockdown significantly enhancedIFN-� and IL-2 production by the KLH-specific T cells. Interestingly,the effects of PD-L2–silenced DCs were very modest, while simulta-neous knockdown of PD-L1 and PD-L2 on DCs resulted in a synergisticimprovement of T-cell cytokine production. A similar phenomenon wasobserved by Keir et al after stimulation of ovalbumin-specific CD4�

T cells with PD-L1/PD-L2�/� DCs in the presence of ovalbumin-peptide.32 This suggests that PD-L1 and PD-L2 can exert similar effectsvia overlapping pathways.

To investigate whether our PD-L knockdown DCs were capableof improving MiHA-specific CD8� T-cell responses, we usedLRH-1–specific CTL RP1 and PBMCs containing MiHA-specificCD8� effector or memory T cells, collected from 3 leukemiapatients early after DLI or years later during relapsed disease. Inthis MiHA-specific setting, LRH-1 peptide-loaded PD-L knock-down DCs improved the expansion rate of LRH-1–specific CTLRP1. In addition, IFN-� and granzyme B levels produced by thisCTL were significantly higher after stimulation with PD-L knock-down DCs. Most importantly, we demonstrated that LRH-1– andHA-1–specific CD8� T cells could be more efficiently expanded byMiHA peptide-loaded PD-L knockdown DCs during both effectorand memory T-cell responses. After the first stimulation with DCs,differences were still modest for some patient samples, but uponrestimulation, PD-L knockdown DCs profoundly improved the

proliferative capacity of the MiHA-specific CD8� T cells. Surpris-ingly, in patient UPN389, LRH-1–specific CD8� memory T-cellexpansion was lower after stimulation with PD-L2 knockdownDCs than with control DCs. We are very intrigued by thisobservation, especially since PD-L1/PD-L2 double-knockdownDCs are, evidently, the best stimulator cells. It could be that undercertain conditions PD-L2 acts as a costimulatory molecule by aPD-1–independent mechanism, as shown by others, and that bysiRNA-mediated silencing, we inhibited LRH-1–specific T-cellactivation.33,34 Another explanation could be that PD-L2 knock-down interferes with the postulated PD-L2 reverse signaling,resulting in DCs with lower stimulatory capacity.35,36 Finally, weshowed that the expanded LRH-1–specific CD8� T cells have thecapacity to degranulate upon recognition of MiHA-expressingtarget cells, demonstrating that these cells have competent cyto-toxic functions.

In conclusion, we demonstrated that efficient, specific, andlong-lasting silencing of PD-L1 and PD-L2 can be achieved bysingle siRNA electroporation of monocyte-derived DCs. We showedthat PD-L knockdown DCs profoundly boost the proliferativecapacity of antigen-experienced T-cells specific for KLH andMiHA. Moreover, both in primary allogeneic MLR and secondaryantigen-specific T-cell responses, PD-L knockdown DCs stronglyenhance cytokine production. Together, these findings indicate thatPD-L siRNA–electroporated DCs are attractive cells for improvingthe efficacy of MiHA-based DC vaccines to boost GVT immunityin SCT patients as well as antigen-loaded DCs in cancer patientsand chronic viral infections.

Acknowledgments

We thank Wieger Norde for helpful advice, Hanny Fredrix fortechnical support, and Prof Dr Fred Falkenburg and Michel Kester(Department of Hematology, Leiden University Medical Center,Leiden, The Netherlands) for providing LRH-1/HLA-B7 andHA-1/HLA-A2 tetramers.

This work was supported by a grant from the Dutch CancerSociety (KWF 2008-4018).

Authorship

Contribution: W.H. performed experiments, analyzed data, andwrote the manuscript; F.M. and N.A. assisted with the experiments;N.S. revised the manuscript and provided patient data and samples;T.W. and R.V. provided advice and revised the manuscript; andH.D. designed research and wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no compet-ing financial interests.

Correspondence: Harry Dolstra, Department of LaboratoryMedicine–Laboratory of Hematology, Radboud University Nijme-gen Medical Center, Geert Grooteplein 8, PO Box 9101, 6500 HBNijmegen, The Netherlands; e-mail: [email protected].

References

1. Riddell SR, Berger C, Murata M, Randolph S,Warren EH. The graft versus leukemia responseafter allogeneic hematopoietic stem cell trans-plantation. Blood Rev. 2003;17(3):153-162.

2. Bleakley M, Riddell SR. Molecules and mecha-nisms of the graft-versus-leukaemia effect. NatRev Cancer. 2004;4(5):371-380.

3. Falkenburg JHF, van de Corput L, Marijt EWA,Willemze R. Minor histocompatibility antigens inhuman stem cell transplantation. Exp Hematol.2003;31(9):743-751.

4. Williams MA, Bevan MJ. Effector and memoryCTL differentiation. Annu Rev Immunol. 2007;25171-25192.

5. Harty JT, Badovinac VP. Shaping and reshapingCD8� T-cell memory. Nat Rev Immunol. 2008;8(2):107-119.

6. de Rijke B, van Horssen-Zoetbrood A,Beekman JM, et al. A frameshift polymorphismin P2X5 elicits an allogeneic cytotoxic T lym-phocyte response associated with remission of





chronic myeloid leukemia. J Clin Invest. 2005;115(12):3506-3516.

7. Marijt WA, Heemskerk MH, Kloosterboer FM, etal. Hematopoiesis-restricted minor histocompat-ibility antigens HA-1- or HA-2-specific T cells caninduce complete remissions of relapsed leuke-mia. Proc Natl Acad Sci U S A. 2003;100(5):2742-2747.

8. Hoogendoorn M, Jedema I, Barge RM, et al.Characterization of graft-versus-leukemia re-sponses in patients treated for advanced chroniclymphocytic leukemia with donor lymphocyte infu-sions after in vitro T-cell–depleted allogeneicstem cell transplantation following reduced-inten-sity conditioning. Leukemia. 2007;21(12):2569-2574.

9. Keir ME, Butte MJ, Freeman GJ, Sharpe AH.PD-1 and its ligands in tolerance and immunity.Annu Rev Immunol. 2008;26677-26704.

10. Chemnitz JM, Parry RV, Nichols KE, June CH,Riley JL. SHP-1 and SHP-2 associate with immu-noreceptor tyrosine-based switch motif of pro-grammed death 1 upon primary human T cellstimulation, but only receptor ligation prevents Tcell activation. J Immunol. 2004;173(2):945-954.

11. Okazaki T, Honjo T. PD-1 and PD-1 ligands: fromdiscovery to clinical application. Int Immunol.2007;19(7):813-824.

12. Barber DL, Wherry EJ, Masopust D, et al. Restor-ing function in exhausted CD8 T cells duringchronic viral infection. Nature. 2006;439(7077):682-687.

13. Day CL, Kaufmann DE, Kiepiela P, et al. PD-1expression on HIV-specific T cells is associatedwith T-cell exhaustion and disease progression.Nature. 2006;443(7109):350-354.

14. Petrovas C, Casazza JP, Brenchley JM, et al.PD-1 is a regulator of virus-specific CD8� T cellsurvival in HIV infection. J Exp Med. 2006;203(10):2281-2292.

15. Larrubia JR, Benito-Martinez S, Miquel J, et al.Costimulatory molecule programmed death-1 inthe cytotoxic response during chronic hepatitis C.World J Gastroenterol. 2009;15(41):5129-5140.

16. Mumprecht S, Schurch C, Schwaller J,

Solenthaler M, Ochsenbein AF. Programmeddeath 1 signaling on chronic myeloid leukemia-specific T cells results in T-cell exhaustion anddisease progression. Blood. 2009;114(8):1528-1536.

17. Ahmadzadeh M, Johnson LA, Heemskerk B, etal. Tumor antigen-specific CD8 T cells infiltratingthe tumor express high levels of PD-1 and arefunctionally impaired. Blood. 2009;114(8):1537-1544.

18. Blank C, Kuball J, Voelkl S, et al. Blockade ofPD-L1 (B7-H1) augments human tumor-specific Tcell responses in vitro. Int J Cancer. 2006;119(2):317-327.

19. van Lochem E, van der Keur M, Mommaas AM,de Gast GC, Goulmy E. Functional expression ofminor histocompatibility antigens on human pe-ripheral blood dendritic cells and epidermal Lang-erhans cells. Transpl Immunol. 1996;4(2):151-157.

20. Tuyaerts S, Aerts JL, Corthals J, et al. Currentapproaches in dendritic cell generation and futureimplications for cancer immunotherapy. CancerImmunol Immunother. 2007;56(10):1513-1537.

21. Banchereau J, Palucka AK. Dendritic cells astherapeutic vaccines against cancer. Nat Rev Im-munol. 2005;5(4):296-306.

22. Banchereau J, Briere F, Caux C, et al. Immunobi-ology of dendritic cells. Annu Rev Immunol. 2000;18767-18811.

23. Greenwald RJ, Freeman GJ, Sharpe AH. The B7family revisited. Annu Rev Immunol. 2005;23515-23548.

24. Zang X, Allison JP. The B7 family and cancertherapy: costimulation and coinhibition. Clin Can-cer Res. 2007;13(18 Pt 1):5271-5279.

25. Norde WJ, Overes IM, Maas F, et al. Myeloid leu-kemic progenitor cells can be specifically targetedby minor histocompatibility antigen LRH-1-reac-tive cytotoxic T cells. Blood. 2009;113(10):2312-2323.

26. Overes IM, Fredrix H, Kester MG, et al. Efficientactivation of LRH-1-specific CD8� T-cell re-sponses from transplanted leukemia patients bystimulation with P2X5 mRNA-electroporated den-dritic cells. J Immunother. 2009;32(6):539-551.

27. Overes IM, de Rijke B, van Horssen-Zoetbrood A,et al. Expression of P2X5 in lymphoid malignan-cies results in LRH-1-specific cytotoxic T cell–mediated lysis. Br J Haematol. 2008;141(6):799-807.

28. Overes IM, Levenga TH, Vos JC, et al. Aberrantexpression of the hematopoietic-restricted minorhistocompatibility antigen LRH-1 on solid tumorsresults in efficient cytotoxic T cell–mediated lysis.Cancer Immunol Immunother. 2009;58(3):429-439.

29. Breton G, Yassine-Diab B, Cohn L, et al. siRNAknockdown of PD-L1 and PD-L2 in monocyte-derived dendritic cells only modestly improvesproliferative responses to Gag by CD8(�) T cellsfrom HIV-1-infected individuals. J Clin Immunol.2009;29(5):637-645.

30. Brown JA, Dorfman DM, Ma FR, et al. Blockadeof programmed death-1 ligands on dendritic cellsenhances T cell activation and cytokine produc-tion. J Immunol. 2003;170(3):1257-1266.

31. Cai G, Karni A, Oliveira EM, et al. PD-1 ligands,negative regulators for activation of naive,memory, and recently activated human CD4�T cells. Cell Immunol. 2004;230(2):89-98.

32. Keir ME, Liang SC, Guleria I, et al. Tissue expres-sion of PD-L1 mediates peripheral T cell toler-ance. J Exp Med. 2006;203(4):883-895.

33. Liu X, Gao JX, Wen J, et al. B7DC/PDL2 pro-motes tumor immunity by a PD-1-independentmechanism. J Exp Med. 2003;197(12):1721-1730.

34. Tseng SY, Otsuji M, Gorski K, et al. B7-DC, a newdendritic cell molecule with potent costimulatoryproperties for T cells. J Exp Med. 2001;193(7):839-846.

35. Nguyen LT, Radhakrishnan S, Ciric B, et al.Cross-linking the B7 family molecule B7-DC di-rectly activates immune functions of dendriticcells. J Exp Med. 2002;196(10):1393-1398.

36. Radhakrishnan S, Nguyen LT, Ciric B, van KeulenV, Pease LR. B7-DC/PD-L2 cross-linking inducesNF-kappaB-dependent protection of dendriticcells from cell death. J Immunol. 2007;178(3):1426-1432.





online August 3, 2010 originally publisheddoi:10.1182/blood-2010-04-278739

2010 116: 4501-4511

and Harry DolstraWillemijn Hobo, Frans Maas, Niken Adisty, Theo de Witte, Nicolaas Schaap, Robbert van der Voort

T cells+CD8specific−expansion and function of minor histocompatibility antigen

siRNA silencing of PD-L1 and PD-L2 on dendritic cells augments

http://www.bloodjournal.org/content/116/22/4501.full.htmlUpdated information and services can be found at:

(5489 articles)Immunobiology Articles on similar topics can be found in the following Blood collections

http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requestsInformation about reproducing this article in parts or in its entirety may be found online at:

http://www.bloodjournal.org/site/misc/rights.xhtml#reprintsInformation about ordering reprints may be found online at:

http://www.bloodjournal.org/site/subscriptions/index.xhtmlInformation about subscriptions and ASH membership may be found online at:

Copyright 2011 by The American Society of Hematology; all rights reserved.of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American Society


http://www.bloodjournal.org/content/116/22/4501.full.html

http://www.bloodjournal.org/cgi/collection/immunobiology

http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests

http://www.bloodjournal.org/site/misc/rights.xhtml#reprints

http://www.bloodjournal.org/site/subscriptions/index.xhtml





sirnasilencing of pd-l1 and pd-l2 on dendritic cells ......responses, the process of alloreactive...

Documents