Induction of Antiviral Cytotoxic T Cells by Plasmacytoid Dendritic Cells for Adoptive Immunotherapy of Posttransplant Diseases

Authors

  • C. Aspord,

    Corresponding author
    1. EFS Rhone-Alpes, R&D Laboratory, La Tronche F-38701, France
    2. Joseph Fourier University, Grenoble F-38041, France
    3. INSERM, U823, Immunobiology & Immunotherapy of Cancers, La Tronche F-38706, France
    Search for more papers by this author
  • D. Laurin,

    1. EFS Rhone-Alpes, R&D Laboratory, La Tronche F-38701, France
    2. Joseph Fourier University, Grenoble F-38041, France
    Search for more papers by this author
  • M.-J. Richard,

    1. Joseph Fourier University, Grenoble F-38041, France
    2. INSERM, U823, Immunobiology & Immunotherapy of Cancers, La Tronche F-38706, France
    3. CHU Grenoble, Michallon Hospital, Cancerology and Biotherapy, Grenoble F-38043, France
    Search for more papers by this author
  • H. Vie,

    1. Institut de Recherche Thérapeutique de l’Université de Nantes, UMR INSERM, U892, 8 quai Moncousu, BP 70721, 44007, Nantes Cedex 1, France
    Search for more papers by this author
  • L. Chaperot,

    1. EFS Rhone-Alpes, R&D Laboratory, La Tronche F-38701, France
    2. Joseph Fourier University, Grenoble F-38041, France
    3. INSERM, U823, Immunobiology & Immunotherapy of Cancers, La Tronche F-38706, France
    Search for more papers by this author
  • J. Plumas

    1. EFS Rhone-Alpes, R&D Laboratory, La Tronche F-38701, France
    2. Joseph Fourier University, Grenoble F-38041, France
    3. INSERM, U823, Immunobiology & Immunotherapy of Cancers, La Tronche F-38706, France
    4. University College London, Cancer Institute, 72 Huntley Street, WC1E 6BT, UK
    Search for more papers by this author

Caroline Aspord, carolineaspord@yahoo.com

Abstract

Virus-associated hematologic malignancies (EBV lymphoproliferative disease) and opportunistic infections (CMV) represent a major cause of hematopoietic stem cell and solid organ transplantation failure. Adoptive transfer of antigen-specific T lymphocytes appears to be a major and successful immunotherapeutic strategy, but improvements are needed to reliably produce high numbers of virus-specific T cells with appropriate requirements for adoptive immunotherapy that would allow extensive clinical use. Since plasmacytoid dendritic cells (pDCs) are crucial in launching antiviral responses, we investigated their capacity to elicit functional antiviral T-cell responses for adoptive cellular immunotherapy using a unique pDC line and antigens derived from Influenza, CMV and EBV viruses. Stimulation of peripheral blood mononuclear cells from HLA-A*0201+ donors by HLA-A0201 matched pDCs pulsed with viral-derived peptides triggered high levels of multi-specific and functional cytotoxic T-cell responses (up to 99% tetramer+ CD8 T cells) in vitro. Furthermore, the central/effector memory cytotoxic T cells elicited by the pDCs strongly display antiviral activity upon adoptive transfer into a humanized mouse model that mimics a virus-induced malignancy. We provide a simple and potent method to generate virus-specific CTL with the required properties for adoptive cellular immunotherapy of post-transplant diseases.

Abbreviations: 
APC

antigen-presenting cell

CsA

cyclosporin A

CMV

cytomegalovirus

CTL

cytotoxic T lymphocyte

EBV

Epstein–Barr virus

HSCT

hematopoietic stem cell transplantation

PBMC

peripheral blood mononuclear cells

pDC

plasmacytoid dendritic cell

RAP

rapamycin

SOT

solid organ transplantation

TLR

toll-like receptor

Introduction

Epstein–Barr virus (EBV) lymphoproliferative disease and opportunistic infections such as CMV represent a major cause of hematopoietic stem cell (HSCT) and solid organ (SOT) transplantation failure, morbidity and mortality. To treat these diseases, which are often refractory to conventional chemotherapy, radiotherapy or antiviral treatments, adoptive transfer of antigen-specific T lymphocytes appears to be an emerging and successful immunotherapeutic strategy (1–3).

Different approaches have been developed to expand in vitro high quantities of specific T cells that target virus-specific antigens (4). The generation of specific cytotoxic T cells (CTL) has been commonly achieved through repetitive stimulation of peripheral blood mononuclear cells (PBMC) with antigen-loaded antigen-presenting cells (APCs). For example, the adoptive transfer of CMV-specific T-cell lines produced by serial stimulation of PBMC with irradiated feeder cells or dendritic cells (DCs) in the presence of antigen and IL2 reconstituted a protective cellular immunity against CMV in allogeneic bone marrow transplantation recipients (5–8). The infusion of EBV-specific CTL lines generated by serial stimulation of autologous EBV-transformed B-cell lines (LCLs) with IL2 led to the control of EBV-associated lymphoproliferation following HSCT (9,10), SOT (11) and EBV-related Hodgkin's disease in immunocompetent hosts (12). Other strategies include the use of transduced or nucleofected DCs with vectors encoding viral antigens (13,14), selection of antigen-specific T cells from blood leukocytes by tetramer labeling (15) or IFNγ-secreting cell capture (16–19) followed by in vitro amplification before infusion. Genetic engineering has been used to redirect effector T-cell specificity or function (20) by either transduction with a high-affinity tumor-specific TcR (21) or expression of a chimeric antigen receptor combining the ScFv of a tumor antigen-specific antibody and a T-cell activation endodomain (22). All of these approaches are often lengthy, costly and sophisticated, requiring additional clinical grade products and technical improvements. In addition, T cells do not always display optimal proliferative, effector and long-lived potential in vivo. Trials have demonstrated the safety and efficacy of adoptive transfer of antigen-specific CTL, but there is a real need for a simple and fast procedure to reliably produce high numbers of virus-specific T cells with appropriate requirements for adoptive immunotherapy that would allow extensive clinical use.

APCs used to amplify virus-specific T cells are usually in vitro-made DCs, monocytes or lymphoblastoid cell lines (LCLs). However, the unique cells specialized in launching antiviral responses are the plasmacytoid dendritic cells (pDCs) (23,24). pDCs play a crucial role in immunity to viruses due to their ability to detect the presence of single stranded RNA and CpG DNA through their Toll-like receptor 7 (TLR7) and 9 (TLR9) and to subsequently produce large quantities of type I IFN. It has been shown that pDCs can cross-present viral antigens (25,26), induce virus-specific adaptive immune responses in vitro (27–30) and are essential to the induction of CTL in vivo following viral infection (31). Until now, the potential of pDC has not been worked out in immunotherapies in humans.

We have developed a new immunotherapeutic strategy using a unique human pDC-cell line (GEN) established from leukaemic HLA-A*0201+ pDC (32) and exploiting its ability to promote immune responses toward specific antigens. We have already demonstrated the potential of irradiated peptide-loaded pDCs to induce tumor-specific responses in HLA-A*0201 matched settings and their efficiency as antitumor vaccines in an active immunotherapeutic strategy for melanoma (33). Here we investigate their potential to elicit antiviral responses and their efficiency as an adoptive immunotherapy strategy using antigens derived from CMV and EBV viruses which are often responsible for transplantation relapse or hematologic malignancies. We demonstrated in vitro the tremendous efficacy of the pDC line to promptly induce highly functional virus-specific T cells from PBMC, and in vivo the therapeutic potential of the transferred CTL to strongly block tumor growth in a humanized mouse model that mimics a virus-induced malignancy. We propose a potent and promising new method for adoptive cellular immunotherapy of virus-induced posttransplant diseases.

Materials and Methods

Cell lines and reagents

Cultures were performed in RPMI-1640 Glutamax supplemented with 1% nonessential amino-acids, 1 mM sodium pyruvate (Sigma, Lyon, France), 100 μg/mL gentamycin and 10%FCS (Invitrogen, Villebon sur Yvette, France). T2 and K562 lines were purchased from ATCC (LGC, Molsheim, France). CMV-or IE-transduced EBV-cell lines GRE-pp65 (HLA-A0201+), COL-pp65 (HLA-A0201-) and GRE-IE were generated by C. Retiere (EFS Nantes, France) (34). Anti-human CD45, CD3, CD8 Abs were purchased from Beckman (Roissy, France), CD45RA, CCR7, CD27, HLA-A2 Abs from BD (Port de Claix, France).

Peptides and tetramers

We used the following viral-derived HLA-A*0201-restricted peptides (NeoMPS) and corresponding iTag HLA-A*0201 tetramers (Beckman): FluM158–66 (GILGFVFTL) from the influenza matrix protein, CMVpp65495–503 (NLVPMVATV) from CMV internal matrix protein, EBV BMLF1280–288 (GLCTLVAML) from lytic cycle protein and EBV LMP2426–434 (CLGGLLTMV) from latent membrane protein.

PBMC and pDC line

PBMC were obtained from HLA-A*0201+ healthy donors by Ficoll-Hypaque density gradient centrifugation (Eurobio, Courtaboeuf, France). The human pDC line GEN2.2, that shares most of the phenotypic and functional features of primary pDC, was obtained from leukaemic pDC and cultured as previously described (32). This study was approved by the French Blood Agency Institutional Review Board. All donors signed informed consent forms.

Specific T-cell response induction

GEN cells were washed three times with serum-free RPMI and resuspended at 1 × 106 cells/ml. β2-microglobulin (0.1 μg/mL) (Sigma) and single or pooled peptide(s) (1–10 μM final) were added. After 3 h of incubation at 37°C, peptide-loaded GEN were washed twice, irradiated 30 Gy and cocultured with HLA-A*0201-matched PBMC at a 1:10 ratio in RPMI 10%FCS for 7 days. Cultures were weekly restimulated with peptide-loaded GEN and 200 U/mL IL2 Proleukine (Chiron, San Diego, CA, USA). Viability and cell count was assessed by Trypan-blue exclusion and numeration. Specific CD8 T-cell responses were measured by tetramer labeling of PBMC initially and at different steps of the culture. Cells were resuspended in HBSS 2% FCS, stained with CD45 FITC, iTAg HLA-A*0201 tetramer PE, CD3 PC7, CD8 APC antibodies and submitted to flow cytometry analysis using a 4-colors FACSCalibur and CellQuest software (BD). Differentiation phenotype of tetramer+ CD8 T cells was assessed by labelling the cells with CD45RA, CCR7 and CD27 antibodies and analyzing using a 6-colors FACSCanto-II and DIVA software (BD).

IFNγ secretion and CD107 expression by tetramer+ CD8 T cells

T cells were first labeled with iTAg HLA-A*0201 tetramer PE for 30 min at RT, washed and restimulated with peptide-pulsed T2 cells (10:1 ratio) for 5 h 30 min. For IFNγ secretion, 1 μL/mL brefeldin-A (BD) was added for the last 3 h. Cells were then surface-labeled with anti-CD3 PC7 and anti-CD8 APC antibodies and submitted to IFNγ intracellular staining (BD). For CD107 detection, anti-CD107a and CD107b FITC antibodies (10 μL/1 × 106 cells) (BD) were added at the beginning of the restimulation in presence of Golgi-STOP (0.67 μL/mL) for the last 4 h. Cells were then labeled with anti-CD3 PC7 and anti-CD8 APC antibodies. IFNγ and CD107 staining were analyzed on the tetramer+ CD8 T cells, tetramer CD8 T cells and CD4 T cells. In some experiments, CsA (100 ng/mL), FK-506 (10 ng/mL) or RAP (10 ng/mL) (SIGMA) were added during the restimulation.

Cytotoxicity assay

Cytotoxic activity was measured by performing a standard 51Cr release assay. Effector T cells were sorted from the coculture using an EasySep human T-cell enrichment kit (StemCell). Target cells were labeled with 50 μCi for 1 h at 37°C, washed three times and plated with effector T cells at the indicated E:T ratio in round bottom 96-well plates. In some experiments, CsA (100 ng/mL), FK-506 (10 ng/mL) or RAP (10 ng/mL) were added during the test. After 4 h radioactivity was measured on 30 μL supernatants on a microplate scintillation counter TopCount-NXT (PerkinElmer, Courtaboeuf, France). The mean of triplicate measurements was expressed as a percentage of specific lysis using the formula: (sample release–spontaneous release)/(maximal release–spontaneous release) ×100.

In vivofunctional assays in humanized mice

NOD-SCIDβ2m−/– immunodeficient mice (NOD.Cg-PrkdcSCIDβ2mTm1Unc/J) were purchased from Jackson ImmunoResearch Laboratories (BarHarbor, ME, USA) and bred at the Plateforme de Haute Technologie Animale (LaTronche, France). HuPBL mice were constructed by transplanting intraperitoneally 50 × 106 PBMC from healthy donors into sublethally irradiated NOD-SCIDβ2m−/– mice (120–150cGy). Mice were further vaccinated intra-peritoneally with 5 × 106 irradiated peptide-pulsed GEN cells once a week. Response to vaccination was analyzed in blood, peritoneal lavage, spleen and lymph nodes. Organs were digested 30 min at 37°C with 2 mg/mL collagenase D (Roche, Meylan, France). Resultant cell suspensions were washed with HBSS 2% FCS, stained using anti-human CD45 FITC, iTAg HLA-A*0201 tetramer PE, CD3 PC7, CD8 APC antibodies and submitted to flow cytometry analysis. Ex vivo cytotoxic activity of the in vivo primed T cells was analyzed using a standard 51Cr release assay as previously described after selection of the human CD45+ cells (Miltenyi) from different organs. For adoptive therapy experiments, sublethally irradiated NOD-SCIDβ2m−/– mice were implanted subcutaneously into the flank with 2.5 × 106 pp65/IE-transduced HLA-A*0201+ or HLA-A*0201 EBV cell lines and treated 5 days later with three weekly peritumoral injections of equal numbers of anti-CMVpp65 or anti-Flu purified CD8 T cells (1–5 × 106 cells), generated in vitro from HLA-A0201+ PBMC after one or two rounds of stimulation(s) with the peptide-loaded pDC line. Tumor size was monitored every 2–3 days and tumor volume calculated using the formula: (short diameter)2xlong diameter/2. Specific T cells were analyzed in tumor and spleen-cell suspensions by tetramer labeling and flow cytometry analysis. This study was carried out in strict accordance with the European Union guidelines for handling of laboratory animals (86/609/CEE) and the French National Chart guidelines. The protocol was approved by the Regional Committee for Animal Ethic Rhone-Alpes.

Statistical analysis

The statistics were performed by using Mann–Whitney nonparametric U-test, paired and unpaired t-test and ANOVA using Prism software.

Results

Human HLA-A*0201+ pDCs induce multivirus-specific T-cell responses from HLA-A*0201-matched healthy donors’ PBMC with a strong efficiencyin vitro

To assess whether the HLA-A*0201+ pDC line can induce specific T-cell responses against Flu, CMV and EBV viruses, PBMC from HLA-A*0201+ healthy donors were stimulated with the irradiated peptide-loaded pDC line. We obtained a massive amplification of specific T cells after only 7 days of culture, as detected by tetramer labeling (Figure 1A). Such responses were reproduced with cells from 20 to 26 healthy donors and with various virus-derived antigens (Figure 1B). This induction was further enhanced by serial stimulations with the peptide-loaded pDC line combined with IL-2. In some cases, we obtained up to 99% tetramer+ CD8 T cells after 40 days (Figures 1C and S1). The systematic fold increase of the absolute numbers of Flu, CMV and EBV-specific T cells reached a mean of 113 in 7 days (range 10–797) and 1272 in 13 days (range 57–9393) (Figure 1D). By starting cultures from up to 25 000 specific T cells depending on donors and antigens, we were able to obtain up to 4 × 106 specific T cells in 7 days and 6.2 × 107 specific T cells in 13 days for the three viruses (Figure S2). Importantly, CD4 T cells were not amplified in the culture (Figures 1D and S2). We then evaluated whether it would be possible to easily generate multivirus-specific T cells to avoid multicultures. We showed that by loading the pDC line with a mixture of several peptides, multivirus-specific T-cell lines could be obtained in a single culture with the same efficacy as with single peptide loading (Figures 1E and F). Thus, strong virus-specific T-cell responses can be elicited easily in vitro from healthy donors’ cells by culture with the pDC line loaded with viral peptides.

Figure 1.

Figure 1.

HLA-A*0201+ pDCs induce multivirus-specific T-cell responses from HLA-A*0201+ healthy donors’ PBMC with a high efficiency in vitro. PBMC from HLA-A*0201+ healthy donors were stimulated by the irradiated peptide-loaded HLA-A*0201+ pDC line and weekly restimulated in the presence of IL2. Specificity of the T cells was determined at different time points by tetramer labeling and flow cytometry analysis. (A) Representative dot plots of tetramer+ T cells initially (d0) and after 7 days of culture (d7) with the irradiated pDC line loaded with Flu M1-, CMVpp65-, EBV BMLF1- or LMP2-derived HLA-A*0201-restricted peptide, respectively (gated on CD8+ T cells). (B) Percentages of Flu-, CMV- and EBV-specific T cells initially and after 7 days of culture of PBMC with the pDC line loaded with the corresponding peptide (n = 26, 20, 21 and 20 healthy donors, respectively, bars at mean). (C) Follow-up example of the percentage of CMVpp65 tetramer+ CD8 T cells during the culture of PBMC weekly stimulated with CMVpp65-loaded pDC line. Flu tetramer was used as negative control. (D) Amplification of the absolute numbers of tetramer+ specific T cells (left panel) or CD4 T cells (right panel) obtained after one or two stimulations with the pDC line loaded with Flu M1-, CMVpp65- or EBV BMLF1-derived peptides (ratio of the absolute number of specific T cells at dX by the absolute number of specific T cells at d0, data from n = 26 independent cultures). (E, F) PBMC from HLA-A*0201+ healthy donors were stimulated by the irradiated HLA-A*0201+ pDC line loaded either with a single viral peptide or with a mix of the four viral peptides. (E) Representative dot plots of tetramer+ T cells initially (d0) and after 7 days of culture (d7). (F) Percentages of Flu-, CMV- and EBV-specific T cells initially and after 7 days of culture of PBMC with the pDC line loaded either with a single or a mix of four peptides (n = 7 healthy donors, bars at mean).

Figure 1.

Figure 1.

HLA-A*0201+ pDCs induce multivirus-specific T-cell responses from HLA-A*0201+ healthy donors’ PBMC with a high efficiency in vitro. PBMC from HLA-A*0201+ healthy donors were stimulated by the irradiated peptide-loaded HLA-A*0201+ pDC line and weekly restimulated in the presence of IL2. Specificity of the T cells was determined at different time points by tetramer labeling and flow cytometry analysis. (A) Representative dot plots of tetramer+ T cells initially (d0) and after 7 days of culture (d7) with the irradiated pDC line loaded with Flu M1-, CMVpp65-, EBV BMLF1- or LMP2-derived HLA-A*0201-restricted peptide, respectively (gated on CD8+ T cells). (B) Percentages of Flu-, CMV- and EBV-specific T cells initially and after 7 days of culture of PBMC with the pDC line loaded with the corresponding peptide (n = 26, 20, 21 and 20 healthy donors, respectively, bars at mean). (C) Follow-up example of the percentage of CMVpp65 tetramer+ CD8 T cells during the culture of PBMC weekly stimulated with CMVpp65-loaded pDC line. Flu tetramer was used as negative control. (D) Amplification of the absolute numbers of tetramer+ specific T cells (left panel) or CD4 T cells (right panel) obtained after one or two stimulations with the pDC line loaded with Flu M1-, CMVpp65- or EBV BMLF1-derived peptides (ratio of the absolute number of specific T cells at dX by the absolute number of specific T cells at d0, data from n = 26 independent cultures). (E, F) PBMC from HLA-A*0201+ healthy donors were stimulated by the irradiated HLA-A*0201+ pDC line loaded either with a single viral peptide or with a mix of the four viral peptides. (E) Representative dot plots of tetramer+ T cells initially (d0) and after 7 days of culture (d7). (F) Percentages of Flu-, CMV- and EBV-specific T cells initially and after 7 days of culture of PBMC with the pDC line loaded either with a single or a mix of four peptides (n = 7 healthy donors, bars at mean).

The virus-specific T cells elicited by the peptide-loaded pDC display a central/effector memory phenotype

As the stage of differentiation of adoptively transferred T cells is important for the subsequent therapeutic efficiency, we investigated the detailed phenotype of the virus-specific T cells elicited by the pDC line. We analyzed the naïve (N), central memory (CM), effector memory (EM) and effector memory CD45RA (EMRA) populations within tetramer+ CD8 T cells using CCR7, CD45RA and CD27 markers (Figure 2A). We found that upon one (d7) or two (d13) stimulations with the peptide-loaded pDC line, the central memory (CD45RA- CCR7+) and effector memory (CD45RA- CCR7-) subsets were predominant (Figure 2B). Indeed the CM and EM subsets represented 38–68% and 28–52% of tetramer+ T cells, respectively. At d13, CD27 was found to be expressed by 39–79% of virus-specific T cells (Figure 2C). Thus, the pDC line elicits CM and EM Flu-, CMV- and EBV-specific T cells.

Figure 2.

Differentiation stage phenotype of virus-specific T cells elicited by stimulation with the peptide-loaded pDC line. Differentiation stages of the virus-specific T cells was assessed at different time points of the culture of HLA-A*0201+ PBMC with the viral peptide-loaded pDC line. (A) Principle of the differentiation stages determination. CD45RA, CCR7 and CD27 markers expression was determined on tetramer+ T cells to define naïve (N), central memory (CM), effector memory (EM) and effector memory RA (EMRA) populations upon weekly stimulation of HLA-A*0201+ PBMC with the pDC line loaded with Flu M1-, CMVpp65-, EBV BMLF1- and EBV LMP2-derived peptides (gated on tetramer+ T cells). (B) Evolution of the percentages of the different N, CM, EM and EMRA populations within Flu-, CMV- and EBV-specific T cells before (d0) and after one (d7) or two (d13) stimulations with the viral peptide-loaded pDC (mean of 4–5 experiments with different donor cells for each antigen). (C) Evolution of CD27 expression on tetramer+ T cells before (d0) and after one (d7) or two (d13) stimulations with the viral peptide-loaded pDC (mean of 4–5 experiments with different donor cells for each antigen).

The virus-specific T cells induced by HLA-A*0201-matched pDC exhibited in vitro functional HLA-A*0201 and antigen-restricted activity

We further examined the functionality of the virus-specific T cells induced by the pDC line by analyzing their capacity to secrete IFNγ and express CD107 upon restimulation and their cytotoxic activity. When cocultured with peptide-loaded T2 cells, specific T cells secreted specifically IFNγ and expressed CD107 in the presence of the relevant but not the control peptide (Figure 3A). We significantly obtained higher IFNγ+ tetramer+ and CD107+ tetramer+ CD8 T cells upon specific restimulation compared to control conditions (p < 0.03 and p < 0.003, respectively; Figure 3B). Importantly, we did not observe any response induction from nonspecific tetramer negative CD8 T cells and CD4 T cells upon T2 or GEN restimulation as attested by the poor IFNγ production and CD107 expression by these subsets (Table 1), suggesting that alloreactive T cells are not elicited in GEN-PBMC HLA-A*0201 restricted cocultures. We also observed the absence of activation of the tetramer+ T cells toward GEN or T2 cells loaded with an irrelevant peptide (Table 1), implying that the virus-specific CTL are not cross-reactive toward other HLA class I molecules. In addition, virus-specific T cells exhibited a strong cytotoxicity toward T2 cells loaded with the relevant but not with control peptide (Figure 3C). A higher specific lysis was obtained for Flu-, CMV- and EBV-specific T cells compared to control conditions (p < 0.007; Figure 3D). Importantly, using unselected T cells as effectors, we did not observe any killing of the GEN cells loaded with control peptide (Figure S3), suggesting the absence of cross-reactivity of the virus-specific CTL and absence of other cytotoxic alloreactive cells. Importantly, the effector T cells function in the setting of immunosuppression as demonstrated by similar CD107 expression upon restimulation and cytotoxicity in the presence of cyclosporine A, tacrolimus or sirolimus (Figures 4 and S4). Thus, the pDC line elicits fully functional virus-specific T cells without bystander allogeneic responses.

Figure 3.

Figure 3.

The virus-specific T cells elicited by the pDC line exhibited in vitro functional activity. (A,B) Virus-specific T cells induced by the pDC line secrete IFNγ and express CD107 on the surface upon specific restimulation. Cells from the PBMC/peptide-loaded pDC culture (d7, d13) were submitted to tetramer labelling and restimulated with T2 cells pulsed with the relevant or a control peptide. IFNγ production was assessed by intracellular staining and CD107 expression by adding anti-CD107a+b antibodies during the restimulation. (A) Representative dot plots of IFNγ secretion and CD107 expression by tetramer+ CD8 T cells elicited by Flu M1-, CMVpp65- and EBV BMLF1-derived peptide-loaded pDC and restimulated with peptide-loaded T2 cells (gated on tetramer+ T cells). (B) Percentages of IFNγ+ tetramer+ T cells (upper panels) and CD107+ tetramer+ T cells (lower panels) obtained upon restimulation of Flu, CMV and EBV virus-specific T cells with T2 cells loaded with the corresponding or a control peptide from n = 4, 6 and 17 experiments, respectively. (C, D) Virus-specific T cells induced by the pDC line are cytotoxic. T cells were selected from the PBMC/peptide-loaded pDC cultures and submitted to a 51Cr release assay using peptide-loaded T2 cells as targets. T cells killed T2 cells loaded with the relevant but not the control peptide. (C) One representative CTL assay for each virus-specific T cells. (D) Percentages of specific killing obtained with Flu M1, CMVpp65 and EBV BMLF1-specific effectors T cells at E:T ratio 60:1 (mean ± SEM, n = 6, 3 and 14 experiments, respectively).

Figure 3.

Figure 3.

The virus-specific T cells elicited by the pDC line exhibited in vitro functional activity. (A,B) Virus-specific T cells induced by the pDC line secrete IFNγ and express CD107 on the surface upon specific restimulation. Cells from the PBMC/peptide-loaded pDC culture (d7, d13) were submitted to tetramer labelling and restimulated with T2 cells pulsed with the relevant or a control peptide. IFNγ production was assessed by intracellular staining and CD107 expression by adding anti-CD107a+b antibodies during the restimulation. (A) Representative dot plots of IFNγ secretion and CD107 expression by tetramer+ CD8 T cells elicited by Flu M1-, CMVpp65- and EBV BMLF1-derived peptide-loaded pDC and restimulated with peptide-loaded T2 cells (gated on tetramer+ T cells). (B) Percentages of IFNγ+ tetramer+ T cells (upper panels) and CD107+ tetramer+ T cells (lower panels) obtained upon restimulation of Flu, CMV and EBV virus-specific T cells with T2 cells loaded with the corresponding or a control peptide from n = 4, 6 and 17 experiments, respectively. (C, D) Virus-specific T cells induced by the pDC line are cytotoxic. T cells were selected from the PBMC/peptide-loaded pDC cultures and submitted to a 51Cr release assay using peptide-loaded T2 cells as targets. T cells killed T2 cells loaded with the relevant but not the control peptide. (C) One representative CTL assay for each virus-specific T cells. (D) Percentages of specific killing obtained with Flu M1, CMVpp65 and EBV BMLF1-specific effectors T cells at E:T ratio 60:1 (mean ± SEM, n = 6, 3 and 14 experiments, respectively).

Table 1.  IFNγ secretion and CD107 expression assessed on tetramer + CD8 T cells or on nonspecific tetramer negative CD8 or CD4 T cells upon restimulation with T2 or GEN cells loaded with the peptide used to stimulate antiviral T cells or a control peptide
Culture (d7)Stimulation% IFNγ+ within% CD107+ within
Tetramer + CD8 T cellsTetramer − CD8 T cellsCD4 T cellsTetramer + CD8 T cellsTetramer − CD8 T cellsCD4 T cells
PBMC + GEN Flu3.1 ± 2.91.4 ± 1.70.12 ± 0.0213 ± 6.82 ± 1.54.5 ± 1
T2 control3.1 ± 2.40.6 ± 0.50.14 ± 0.0213.6 ± 62.4 ± 1.75.2 ± 1.3
T2 Flu30.7 ± 13.80.9 ± 0.90.14 ± 0.0171.4 ± 5.12.8 ± 2.15.3 ± 0.8
GEN control4.8 ± 4.22 ± 1.20.7 ± 0.316.1 ± 9.26.7 ± 3.59 ± 1.7
GEN Flu38.8 ± 19.61.5 ± 1.30.7 ± 0.266.7 ± 9.85.5 ± 3.58.1 ± 1.1
PBMC + GEN CMVT2 control7.9 ± 4.30.2 ± 0.20.08 ± 0.0347.5 ± 11.20.9 ± 0.92 ± 1.2
T2 CMV13.9 ± 5.50.4 ± 0.50.07 ± 0.0468.6 ± 3.71.1 ± 1.12.2 ± 1.1
PBMC + GEN BMLF12 ± 00.05 ± 00.4 ± 012.9 ± 00.2 ± 01 ± 0
T2 control2.6 ± 1.50.1 ± 0.050.2 ± 0.0912.8 ± 4.71.1 ± 0.42 ± 0.8
T2 BMLF121.9 ± 70.3 ± 0.20.2 ± 0.164 ± 13.11.3 ± 0.61.9 ± 0.8
GEN control3 ± 00.1 ± 00.5 ± 013.2 ± 01.1 ± 02.1 ± 0
GEN BMLF110 ± 00.1 ± 00.3 ± 068.6 ± 00.8 ± 01.6 ± 0
PBMC + GEN LMP23.7 ± 0.10.3 ± 0.30.1 ± 0.0526.9 ± 0.20.9 ± 0.53.7 ± 0.4
T2 control13.8 ± 5.20.1 ± 0.070.1 ± 0.0738.1 ± 15.61.1 ± 0.52.1 ± 0.8
T2 LMP228.6 ± 9.90.1 ± 0.080.2 ± 0.163.1 ± 13.81.2 ± 0.62.1 ± 0.8
GEN control2.7 ± 00.7 ± 00.5 ± 025.6 ± 02 ± 05.5 ± 0
GEN LMP225.8 ± 00.6 ± 00.4 ± 061.9 ± 02.5 ± 05.5 ± 0

The peptide-loaded pDC line elicits strong and widespread functional virus-specific T-cell responses in vivo in humanized mice

We next investigated the ability of the peptide-loaded pDC line to induce viral-specific T cells in a whole organism. We used a humanized mouse model constructed by xenotransplanting human HLA-A*0201+ PBMC into immunodeficient NOD-SCIDβ2m−/- mice (HuPBL SCID model) and further vaccinated these mice with the pDC line loaded with Flu M1-, CMVpp65-, EBV BMLF1- and EBV LMP2-derived peptides. Notably, a single injection of the irradiated peptide-loaded pDC line induced strong antigen-specific T-cell responses toward the four viral antigens. Human tetramer+ CD8 T cells were found at the site of immunisation (peritoneal lavage) but also in the circulation (blood) and lymphoid organs (spleen, draining mesenteric lymph nodes) (Figure 5A). Seven days after a single pDC vaccination, within all HuPBL mice reconstituted with PBMC containing initially means of 0.11% (FluM1), 0.12% (CMVpp65) and 0.22% (EBV) tetramer+ T cells, virus-specific T cells reached means of 0.5– 1.9% for FluM1, 1.1–5.9% for CMVpp65, and 1.6–4.9% for EBV depending on the organs (Figure S5). Furthermore, specific T cells elicited in vivo by the pDC vaccine were functional, as demonstrated by their ex vivo cytotoxic activity toward peptide-pulsed T2 cells and importantly virus-expressing tumor cells in an HLA-A*0201 and antigen-restricted manner (Figure 5B). Importantly, such T cells were unable to kill the HLA-A*0201 negative COLpp65 cells, suggesting that the pDC line does not elicit cross-reactive nor alloreactive T cells. Thus, the peptide-pulsed pDCs elicits strong and widespread functional virus-specific T-cell responses in vivo.

Figure 5.

Antiviral T cells elicited in vivo by the peptide-loaded pDC line in a humanized mouse model are functional. Immunodeficient NOD-SCIDβ2m−/- mice were reconstituted intraperitoneally with 50 × 106 human HLA-A*0201+ healthy donors’ PBMC and vaccinated by the same route with 5 × 106 irradiated peptide-loaded GEN cells. (A) Specific T-cell induction was analyzed at the injection site (peritoneal lavage), in the circulation (blood) and lymphoid organs (spleen, draining lymph nodes) by tetramer labeling of human T cells in cell suspensions. Representative dot plots of tetramer labelling of PBMC initially (upper panels) and after a single vaccination with viral peptide-loaded GEN cells in the different organs at day 8 (lower panels) (gated on CD8+ T cells). One mouse per group is shown. (B) Specific T cells induced in vivo by vaccination with the peptide-loaded pDC line are functional ex-vivo. Human T cells were purified from the indicated organs 9 days after vaccination with CMVpp65-loaded GEN cells. A 51Cr release assay was performed using T2 cells loaded with peptides and pp65- or IE-transduced HLA-A*0201+ GRE and HLA-A*0201 COL EBV lines as targets. T cells from the vaccination site and from the periphery are able to kill target cells in an HLA-A*0201 and antigen-restricted manner.

Virus-specific T cells elicited by the peptide-loaded pDC line display a strong antiviral activity upon adoptive transferin vivoin humanized mice

We further investigated the potential of this strategy in adoptive immunotherapy using humanized mice engrafted with human virus-expressing tumors. We asked whether the adoptive transfer of antigen-specific T cells generated in vitro by the pDC line could display antiviral activity in vivo. Immunodeficient NOD-SCIDβ2m−/– mice were inoculated subcutaneously with pp65-transduced HLA-A*0201+ lymphoblastoid cells and established tumors were treated 5 days later by peritumoral injections of CMVpp65-specific T cells generated in vitro from HLA-A*0201+ PBMC by the peptide-loaded pDC line. Treatment with anti-CMVpp65 T cells abrogated tumor development whereas control anti-Flu T cells did not (tumor size at day 20 = 18 mm3 compared to 150 mm3, p < 0.001) (Figures 6A and B). Tetramer labeling of cell suspensions recovered at day 30 revealed the presence of many specific T cells at the residual tumor site but also in the blood and lymphoid organs of the treated mice (Figure 6C), demonstrating their survival and patrolling potential in vivo. By contrast, the growth of HLA-A*0201+pp65 and HLA-A*0201pp65+ lymphoblastoid tumors was not inhibited by treatment with anti-CMVpp65 T cells compared to anti-Flu T cells (Figure 6D) demonstrating the HLA-A*0201- and antigen-restriction of this strategy. We found CMVpp65-specific T cells only in HLA-A*0201+pp65+ tumors (GRE-pp65) but not in HLA-A*0201+pp65 (GRE-IE) or HLA-A*0201pp65+ (COLpp65) tumors (Figure 6E). Thus the virus-specific T cells induced by the pDCs display specific and efficient therapeutic effects upon adoptive transfer in vivo.

Figure 6.

Specific T-cell elicited in vitro by the peptide-loaded pDC line inhibits tumor growth upon adoptive transfer in vivo in humanized mice. Specific T cells generated in vitro by peptide-pulsed GEN cells were adoptively transferred into tumor-bearing immunodeficient NOD-SCID β2m−/– mice. Tumor progression was followed up. (A–C) Mice were implanted with pp65-transduced HLA-A*0201+ EBV cells and treated 5 days later with three weekly injections of anti-CMVpp65 or control anti-Flu T cells. (A) Representative experiment of comparative tumor growth evolution. (B) Comparison of tumor size 20 days after implantation and after anti-CMVpp65 or anti-Flu T cell transfer (pool of four experiments, n = 11 and 9 mice, respectively). (C) Tetramer labeling of T cells within cell suspensions from tumor, blood and spleen of mice treated with anti-CMVpp65 T cells at day 30. (D) The therapeutic effect of the treatment is HLA-A*0201 and antigen-restricted. Growth evolution of HLA-A*0201+ pp65 and HLA-A*0201 pp65+ tumors implanted into mice treated with anti-CMVpp65 or anti-Flu T cells. Representative of two experiments, 5–6 mice per group, p = NS. (E) CMVpp65 tetramer+ T cells found at the tumor site following anti-CMVpp65 T cells transfer into GRE-pp65, GRE-IE or COL-pp65 tumor bearing mice (mean ± SEM, n = 6 mice).

Discussion

Adoptive T-cell therapy is an effective strategy to treat viral infections posttransplant in HSCT and SOT recipients. Current procedures to generate high quantities of antigen-specific T cells with appropriate functions are still limited for extensive clinical use (1). pDC are important antigen-presenting cells, particularly in the context of infectious diseases. We demonstrate here that an irradiated peptide-loaded pDC line can efficiently trigger highly functional HLA-A*0201-restricted multivirus-specific T-cell lines and prove its potential for adoptive immunotherapy of virus-induced malignancies.

We provide a simple and potent method to generate virus-specific CTL with the required properties for adoptive cellular immunotherapy. The irradiated peptide-pulsed pDC line can generate virus-specific T-cell responses displaying efficient functions. Interestingly, we demonstrated the possibility of inducing simultaneously multi-specific T cells toward several viral antigens, as we already showed for tumor antigens (33), which may permit adoptive immunotherapy to target multiple pathogens and multiple epitopes of the same pathogen to prevent immune evasion (35,36). The differentiation state of CD8 T cells is a crucial determinant of their efficacy upon adoptive transfer. It is known that the central memory component of the transferred T cells appeared as optimal for efficacy and long-term persistence of the T cells in vivo (37–40). Indeed, using the peptide-loaded pDC line, we generated CTL displaying a central memory phenotype CD45RA CCR7+ still expressing CD27 (41) that were able to inhibit virus-associated malignancies and persisted in vivo after their adoptive transfer in a humanized mouse model. Importantly, we further demonstrate the therapeutic relevance of this strategy by showing its antiviral activity in a humanized mouse model (42) that mimics a virus-induced malignancy in human settings. The ability of the virus-specific T cells to inhibit the growth of virus-expressing lymphoblastoid tumors demonstrated their potential to correctly migrate, specifically kill their targets and survive in vivo. Thus the specific T cells elicited by the pDC line display all the required qualities for adoptive immunotherapy: quick generation of high numbers of specific T cells, with optimal differentiation phenotype, strong functional capacities even in the setting of immunosuppression, the ability to home to tumors and long-lived potential in vivo (in the settings of the humanized mouse model). This strategy could also be used in an active way by vaccinating the patients with the irradiated peptide-loaded pDC line to elicit virus-specific CTL directly in vivo, as it has already been shown with peptide-pulsed autologous mDCs in the context of EBV+ nasopharyngeal carcinoma (43).

This strategy, even if limited to HLA-A*0201+ patients in the present setting that is less than 50% of the population, overcomes the limitations of current T-cell generation protocols with respect to time, complexity, technology and cost that have in the past restricted the broad application of these therapies (1). The generation of virus-specific CTL requires competent APCs and current procedures involve the production of APCs for each patient through long processes (4–6 weeks for EBV-LCLs) or with additional genetic modification, activators or adjuvants (for DCs). CTL expansion is then usually achieved through repeated stimulations in open cultures over several weeks. Our peptide-loaded pDC line can be produced and stored frozen as prevalidated batches ready to use for all patients, and allows the prompt production of broad spectrum antiviral CTL from the same sample of PBMCs and in closed systems. As an example, from only 20 mL of blood, we can generate 25 × 106 CMV-specific T cells in 13 days, which corresponds to doses used in current trials (1–5 × 107 cells/m2). The simple design of our potent strategy is therefore suitable for clinical use.

In the HSCT and SOT context, allogeneic responses represent a major cause of graft rejection. CTL therapy should therefore be specific. Our strategy allows the generation of potent antiviral CTL that are not cross-reactive to different HLA alleles, without bystander activation of alloreactive CD8 and CD4 T cells in vitro. The antigen- and HLA-A*0201-restriction of the killing by virus-specific CTL and the absence of killing of HLA-A*0201 negative tumor cells by total T cells of GEN-PBMC cocultures further underlined the specificity and absence of alloreactivity and cross-reactivity of the approach in vivo. A possible explanation for the selective HLA restricted- over allo-responses triggered by pDCs could be the differential mobilization of MHC-I and MHC-II molecules upon stimulation and the rapid mobilization of MHC-I molecules to the cell surface. Indeed, it has been suggested that pDCs are equipped with large “ready-made” intracellular stores of MHC-I molecules than can be rapidly mobilize to the cell surface to initiate antigen-specific CD8 T-cell responses (44). pDCs display low amounts of MHC-II molecules on their surface and did not upregulate MHC-II synthesis soon after activation (45,46). Importantly, pDCs can induce allogeneic T-cell hyporesponsiveness and subsequent prolonged graft survival (47). The absence of bystander allogeneic response upon induction of antiviral CTL with the pDC line allows consideration of our strategy as a safe procedure to generate CTL for adoptive therapy of posttransplant diseases.

Adoptive T-cell immunotherapy appears to be the most clinically effective strategy to treat viral and nonviral cancers and virus-induced malignancies. Overcoming limitations of current protocols could be promising for the future development of cellular immunotherapy. We propose a new method to simply but strongly activate and expand antigen-specific T cells with the required properties for adoptive therapy of virus-induced posttransplantation diseases. Its clinical applicability would, however, require confirming that such CTL generation can be achieved from PBMC of transplant patients; CTL will survive when adoptively transferred to patient on immunosuppression and minimal alloreactivity also occurs in the clinical setting. Our strategy represents, nevertheless, an advance in the challenging area of adoptive immunotherapy because pDCs have never been tested yet as inducers of antiviral CTL for adoptive cell therapy. Their efficiency could exceed that of current protocols and the straightforward design of the strategy renders it attractive for extensive clinical use.

Acknowledgments

We are grateful to C. Morand, I. Michaud, F. Bernard and their staff from EFS Rhone-Alpes for providing blood samples, F. Blanquet and R. Balouzat of the Plateforme de Haute Technologie Animale for expert animal care, F. Herodin and J.-F. Mayol from the CRSSA for animal irradiation. We thank G. Gallot for sending the transduced EBV cell lines, K. Peggs and C. Roddie for critical review of the manuscript. We thank the healthy volunteers who consented to participate in this study. This work was supported by grants from the Institut National du Cancer (ACI-63–04 and canceropole 2004–05) and Etablissement Français du Sang.

Disclosure

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

Ancillary