CD137L dendritic cells induce potent response against cancer-associated viruses and polarize human CD8⁺ T cells to Tc1 phenotype

Abstract

Therapeutic tumor vaccination based on dendritic cells (DC) is safe; however, its efficacy is low. Among the reasons for only a subset of patients benefitting from DC-based immunotherapy is an insufficient potency of in vitro generated classical DCs (cDCs), made by treating monocytes with GM-CSF + IL-4 + maturation factors. Recent studies demonstrated that CD137L (4-1BBL, TNFSF9) signaling differentiates human monocytes to a highly potent novel type of DC (CD137L-DCs) which have an inflammatory phenotype and are closely related to in vivo DCs. Here, we show that CD137L-DCs induce potent CD8⁺ T-cell responses against Epstein–Barr virus (EBV) and Hepatitis B virus (HBV), and that T cells primed by CD137L-DCs more effectively lyse EBV⁺ and HBV⁺ target cells. The chemokine profile of CD137L-DCs identifies them as inflammatory DCs, and they polarize CD8⁺ T cells to a Tc1 phenotype. Expression of exhaustion markers is reduced on T cells activated by CD137L-DCs. Furthermore, these T cells are metabolically more active and have a higher capacity to utilize glucose. CD137L-induced monocyte to DC differentiation leads to the formation of AIM2 inflammasome, with IL-1beta contributing to CD137L-DCs possessing a stronger T cell activation ability. CD137L-DCs are effective in crosspresentation. PGE2 as a maturation factor is required for enhancing migration of CD137L-DCs but does not significantly reduce their potency. This study shows that CD137L-DCs have a superior ability to activate T cells and to induce potent Tc1 responses against the cancer-causing viruses EBV and HBV which suggest CD137L-DCs as promising candidates for DC-based tumor immunotherapy.

Electronic supplementary material

The online version of this article (10.1007/s00262-018-2144-x) contains supplementary material, which is available to authorized users.

Introduction

Dendritic cells (DCs) are sentinels of the vertebrate immune system with a specialized ability to capture and present antigens [1]. DCs thus play a critical role in orchestrating T-cell responses against pathogens and transformed cells. Immunotherapies with antigen-loaded DCs have been tested for treatment of a variety of cancers [2, 3]. Clinical studies with DC vaccines have so far relied on monocyte-derived DCs, generated in vitro using GM-CSF and IL-4. Although proven to be safe, DC-based immunotherapy has failed to meet expectations, since it yields limited clinical success as measured by patient survival [4].

Virus-associated malignancies account for approximately 15% of cancers worldwide with Epstein–Barr virus (EBV) and Hepatitis B virus (HBV) being the most common cancer-associated viruses. EBV is associated with Burkitt lymphoma, classic Hodgkin lymphoma, diffuse large B-cell lymphoma, nasopharyngeal carcinoma, and gastric carcinoma, while 53% of hepatocellular carcinoma cases are positive for HBV. Apart from contributing to virus-mediated carcinogenesis, viral proteins offer well-defined non-self antigens that can be specifically targeted via immunotherapy. Adoptive cell therapies and vaccines targeting EBV and HBV as anti-cancer immunotherapies have induced immune responses with clear survival benefits [5–8].

Previous studies demonstrated that CD137 ligand (CD137L) signaling differentiates human monocytes to highly potent dendritic cells (CD137L-DCs) without the requirement of exogenous cytokines [9, 10]. CD137L-DCs have an inflammatory phenotype and are closely related to in vivo DCs [11]. These findings are in line with data identifying reverse CD137L signaling as a powerful inducer of myeloid cell differentiation and activation [12].

CD137 expression is induced on vascular endothelial cells by TNF, and facilitates the extravasation of monocytes at sites of inflammation [13–15], where they then can differentiate to CD137L-DC. CD137L is expressed on a subset of hematopoietic progenitor cells in man and mouse, and CD137L signaling induces myeloid differentiation towards the monocytic lineage [16–18].

In this study, we show that CD137L-DCs induce potent CD8⁺ T-cell responses against EBV and HBV, and that T cells primed by CD137L-DCs more effectively lyse EBV⁺ and HBV⁺ target cells than T cells primed by GM-CSF + IL-4 derived classical DCs (cDCs). Chemokines released by CD137L-DCs polarize CD8⁺ T cells to a Tc1 phenotype, which are metabolically more active, have a higher ability to utilize glucose, and express lower levels of exhaustion markers. CD137L signaling induces the AIM2 inflammasome in differentiating monocytes, and IL-1 contributes to the higher T-cell activation capacity of CD137L-DCs.

Materials and methods

Recombinant proteins, antibodies, and reagents

Anti-human CD137L (Ultra LEAF™ purified, clone 5F4) and anti-CD3 antibody (clone OKT3) were purchased from Biolegend (MN, USA). GM-CSF, IL-4, IL-2, TNFα, IL-1β, IL-6, and IFN-γ were from ImmunoTools (Friesoythe, Germany); human IL-1Ra, prostaglandin E2 (PGE2), and Z-VAD from Sigma–Aldrich (MO, USA); Resiquimod (R848) from Invivogen (CA, USA), human Fc receptor blocker from Miltenyi (Bergisch Gladbach, Germany), antibodies against human PD1 (clone J105), CTLA-4 (L3B10), TIM3 (clone F38-2E2), CD80 (clone 24D10), CD86 (clone IT2.2), and HLA-DR (clone L243) from eBioscience (CA, USA); anti-human CCR7 antibody (clone G043H7) from Biolegend (USA); Immunocult™ Human CD3/CD28/CD2 T cell activator from STEMCELL Technologies (BC, Canada); and L7™ hSPC Passaging solution from Lonza (Basel, Switzerland).

Peptide antigens, TCR sequences, and MHC I multimers

The human melanoma-associated antigen Melan-A/MART-1 epitope employed was as described [19]. The EBV antigen EBNA3B, the HBV HBsAg epitope, and the corresponding TCR sequences have been described [20, 21]. All peptides were synthesized by 1st BASE, Singapore, and were > 95% pure. The TCR sequences were synthesized and cloned in pVAX1 vectors by GenScript (IL, USA). MHC-I dextramers were purchased from Immudex (Copenhagen, Denmark). The LMP1 peptide was a gift from Dr. Paul Macary, National University of Singapore. Table 1 describes the peptides used in the study.

Table 1.

Antigen peptides

Antigen	Amino acid sequence	HLA restriction
MART-1 (Short)	ELAGIGILTV	HLA-A2
MART-1 (Long)	GHGHSYTTEELAGIGILTVILGVL	HLA-A2
LMP2a (EBV)	CLGGLLTMV	HLA-A2
EBNA3B (EBV)	AVFDRKSDAK	HLA-A11
LMP1 (EBV)	YLLEMLWRL	HLA-A2
HBs Ag (HBV)	FLLTRILTI	HLA-A2

Isolation of primary cells and cell culture

Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll-Paque (GE Healthcare, IL, USA) density gradient centrifugation. Monocytes were isolated from PBMCs by negative selection using Human Monocyte Enrichment Kit (STEMCELL Technologies). Pan T or CD8⁺ T cells (naïve/total) cells were isolated from PBMCs using the appropriate Human T-cell Isolation Kit or Human CD8⁺ T cell Isolation Kit (STEMCELL Technologies). The MART-1 TCR-expressing T-cell clone was a kind gift from Dr. John R. Wunderlich, National Cancer Institute, Bethesda, USA, and was maintained in AIMV media supplement with 2% human AB serum (AIMV 2ab) and 100 IU/ml IL-2. HLA-A2-restricted HepG2 cells stably transfected with the luciferase gene were cultured in RPMI-1640 supplemented with 10% FBS, 50 µg/ml streptomycin, and 50 IU/ml penicillin (R10PS) supplemented with 1% non-essential amino acids and 2 µg/ml puromycin.

DC generation and maturation

To generate CD137L-DCs, freshly isolated monocytes were plated on 6-well polystyrene plates coated with 5 µg/ml of anti-CD137L antibody (clone 5F4) in R10PS media for 7 days. DCs were matured for the last 18 h with R848 (1 µg/ml) + IFN-γ (50 ng/ml). In addition, PGE2 (1 µg/ml) was added to the maturation cocktail wherever mentioned. Classical DCs (cDCs) were generated by culturing monocytes in R10PS media in the presence of GM-CSF (80 ng/ml) and IL-4 (100 ng/ml) for 7 days. For the last 18 h, cells were matured with TNF (10 ng/ml), IL-1β (10 ng/ml), IL-6 (10 ng/ml), and PGE2 (1 µg/ml) [22]. CD137L-DCs were harvested by incubation with L7™ hSPC passaging solution for 15 min at 37 °C, followed by gentle scraping. cDCs were harvested by gentle pipetting.

ELISPOT assay

The ELISPOT assay was performed using the Human IFN-γ T cell ELISPOT single-color enzymatic kit from Cellular Technologies (OH, USA) at a DC:T cell ratio of 1:10. DCs were pulsed with 10 µg/ml of antigen peptide(s). After 12–18 h incubation at 37 °C, wells were washed, developed, and read on an Immunospot S6 Versa ELISPOT reader (Cellular Technologies,). Co-cultures with un-pulsed DCs served as controls for unspecific T-cell activation.

Anti-MART1 response and cross presentation assay

CD137L-DCs, generated from HLA-A2⁺ donors and pulsed with a short MART-1 peptide, were co-cultured with autologous naïve CD8⁺ T cells at a DC:T cell ratio 2:5. After 24 h, 10 IU/ml of IL-2 was added to the co-culture. Cells were maintained in the co-culture for 12 days with periodic media addition before they were stained for CD3, CD8 and MART-1 dextramer and analyzed by flow cytometry.

To test the crosspresentation ability of CD137L-DCs, DCs, generated from HLA-A2⁺ donors, and pulsed with short or long MART-1 peptide for 1 h were co-cultured in R10PS media with HLA-A2⁺ T cells, stably expressing the corresponding MART-1 TCR at DC:T cell ratios of 1:5 and 1:10. Cell proliferation was measured on day 5 by ³H-thymindine incorporation. IFN-γ levels were determined in supernatants, after 18 h of co-culture, by ELISA.

Generation of TCR-redirected T cells

The plasmid (pVAX1) expressing a TCR for LMP2a/EBNA3B/HBsAg was subjected to in vitro transcription using the mMESSAGE mMACHINE T7 Ultra Kit (Life Technologies, CA, USA). 20 µg TCR mRNA were then electroporated into corresponding HLA-matched T cells (10⁷) using the P3 Primary Cell 4D-Nucleofeactor kit (Lonza). Prior to electroporation, PBMCs were activated with anti-CD3 antibody (clone OKT3, 50 ng/ml) + IL-2 (600 IU/ml) for 7 days in AIMV 2ab media. On day 6, cells were supplemented with AIMV 2ab medium containing 100 IU/ml IL-2. On day 7, cells were harvested and electroporated with TCR mRNA. Following electroporation, the T cells were maintained in AIMV 2ab medium supplemented with 100 IU/ml IL-2 for 4 h at 37 °C and 5% CO₂ and then used for downstream assays.

DC–T cell co-culture

CD137L-DCs and cDCs were harvested and co-cultured with T cells at indicated DC:T cell ratios for 18–72 h. Thereafter, the T cells and/or supernatants were harvested for further assays. TCR-redirected T cells were co-cultured at a DC:T cell ratio of 1:10 for 24 h.

Cytotoxicity assay

For killing assay, TCR-electroporated T cells were generated and stimulated with antigen-pulsed DCs for 18 h. 24 h prior to T-cell harvesting, HepG2 cells were harvested and 20,000 cells/well were seeded in a 96-well flat bottom plate. After, 18–24 h cells were pulsed with 1 µg/ml of antigen peptide for the last 2 h in case of LMP2a. To initiate the killing assay, medium from the antigen-pulsed HepG2 (targets) wells was removed and harvested TCR-electroporated T cells (effectors), resuspended in AIMV 2ab medium, were added at. The plate was incubated at 37 °C and 5% CO₂ for 12–16 h. Target cell lysis was measured by quantifying the luciferase expression level in the remaining cells by the Steady-Glo Luciferase Assay Kit (Promega, WI, USA). Luminescence was measured using the VICTOR³ multi-label plate reader (PerkinElmer, MA, USA). HepG2 cells without any effector cells were used as a measure for maximum luminescence. For LMP2a, HepG2 cells not pulsed with LMP2a peptide served as control. For HBV, HepG2-env cells expressing the HBs antigen were used as target cells, while HepG2-core cells not expressing the HBs antigen were used as a control.

DC migration assay

The migration assay was performed using a 24 well transwell chamber with 5 µM pore size (Corning, NY, USA). DCs were suspended in RPMI-1640 supplemented with 2% FBS. The bottom chambers of the transwell plate were filled with 600 µl of medium supplemented with or without CCL19. In the top chamber, 8 × 10⁴ DCs were added. The plate was incubated at 37 °C for 2 h. Migration was determined by determining the number of migrated cells in the bottom chamber using flow cytometry and counting beads. Percentage of migrated = (Number of migrated cells/Total number of cells) × 100.

Quantitative real-time PCR

Total RNA was isolated using the RNeasy Mini Kit (Qiagen, Hilden, Germany). cDNA was synthesized using the GoScript™ reverse transcription kit (Promega). Quantitative real-time PCR was performed for the target gene(s) with 40–50 ng cDNA/reaction using GoTaq^® qPCR master mix (Promega) on ABI 7500 real-time PCR system (Applied Biosystems, CA, USA). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an endogenous control. Relative expression was computed by the ΔΔCt method. Primers used are listed in supplementary Table 1. 1st Base, Singapore synthesized all primers.

Nanostring assay

LMP2a TCR-redirected T cells were activated with LPM2a antigen-pulsed matured cDCs and CD137L-DCs (not-matured and matured) for 18 h. CD8⁺ T cells were sorted and lysed. A Nanostring assay using the Human Immunology Panel V2 was performed. Data were analyzed using the nSolver Analysis Software 3.0.

Glucose utilization and cell energy phenotype assay

T-cell metabolism was measured in a Seahorse XFe24 analyzer on V7-PET XF24 cell culture microplates (Seahorse Bioscience, USA). 5 × 10⁵ CD8⁺ T cells, activated with allogeneic DCs for 48 h, were harvested via sorting and seeded into the wells coated with Cell-Tak (Corning, USA) as per the manufacturer’s instructions. For the glucose utilization assay, the assay medium was XF basal medium supplemented with 10 mM glucose (Invitrogen, CA, USA) and 20 µl of Immunocult™ human T-cell activator was used. For the phenotype assay, the XF basal medium was supplemented with 1 mM pyruvate, 2 mM glutamine, and 10 mM glucose. 1 µM Oligomycin and 2 µM carbonilcyanide p-triflouromethoxyphenylhydrazone (FCCP) (Sigma-Aldrich) were used.

Flow cytometry

Flow cytometry was performed using the BD LSR Fortessa cell analyzer or BD Fortessa X-20 cell analyzer (BD Bioscience, CA, USA). Data were analyzed using the FlowJo data acquisition and analysis software. Fc receptor blocker was used during sample preparation. For staining with MHC I dextramers, T cells were first stained for CD3 and/or CD8.

ELISA

IFN-γ levels in cell culture supernatants were determined by human Duoset ELISA (R&D Systems, MN, USA). IL-1β and TNF levels were determined by human Ready-SET-Go!® ELISA kits (eBioscience). All measurements were performed in either duplicates or triplicates. Human cytokine/chemokine multiplex assay was performed using 40-plex panel for Luminex™ platform (Thermo Fischer Scientific, CA, USA). The assay was outsourced to Immunomonitoring Platform at SIgN, Singapore.

Statistical analysis

Statistical significance was determined by a two-tailed unpaired Student’s t test or ANNOVA as specified.

Results

CD137L-DCs induce potent CD8⁺ T-cell responses against EBV and HBV

The potency of CD137L-DCs to induce anti-EBV and anti-HBV responses was tested by ELISPOT assays for the EBV antigens LMP1 and LMP2a and by activation of TCR-redirected T cells for the EBV antigens LMP2a, EBNA3B, and the HBV envelope antigen HBsAg. Matured cDCs, currently being used for DC therapy, were used as a positive control. The peptide-pulsed DCs were co-cultured with autologous CD8⁺ T cells, and after 18 h, antigen-specific IFN-γ secreting T cells were identified by ELISPOT assay. All DCs were able to induce an EBV-specific CD8⁺ T-cell response (Fig. 1a). Maturation greatly enhanced the potency of CD137L-DCs, and matured CD137L-DCs induced significantly higher numbers of LMP1- and LMP2a-specific CD8⁺ T cells than matured cDCs (Fig. 1a).

Fig. 1.

CD137L-DCs induce potent T-cell responses against EBV and HBV. cDCs and CD137L-DCs were generated from monocytes of HLA-A2⁺ or HLA-A11⁺ donors and matured. a Autologous ELISPOT assay for HLA-A2-restricted LMP1 or LMP2a peptide. Depicted are means ± standard deviations of the number of spot forming cells (SFC) per 10⁶ CD8⁺ T cells of triplicate measurements. **p < 0.01 and ***p < 0.001 using one-way ANNOVA. b–d Activation of TCR-redirected T cells by autologous peptide-pulsed (1 µg/ml) DCs. TCR-redirected T cells without any DC stimulation (No APC) served as a control. Bar charts depict means ± standard deviations of duplicate measurements. ***p < 0.001 and ****p < 0.0001 using one-way ANNOVA. ns not significant, Mat matured. Results are representative of three independent experiments

TCR-redirected T cells have been used for treatment of variety of cancers, and provide an effective tool for comparatively evaluating the potency of CD137L-DCs. TCR-redirected T cells were generated as described above. Cells from HLA-A2⁺ donor(s) were electroporated with the mRNA of a HLA-A2-restricted TCR specific for LMP2a/HBsAg, and cells from HLA-A11⁺ donor(s) were electroporated with the mRNA of a HLA-A11-restricted TCR specific for EBNA3B. 24 h after electroporation more than 85% CD8⁺ T cells were positive for the TCR (Supplementary Fig. 1). Stimulation of TCR-redirected T cells with autologous peptide-pulsed DCs for 18 h showed that non-matured as well matured CD137L-DCs were significantly more potent than matured cDCs in activating TCR-redirected T cells for LMP2a, EBNA3B, and HBsAg as evidenced by the higher levels of IFN-γ and TNF-α in co-culture supernatants (Fig. 1b–d). TCR-redirected T cells without any DC stimulation (No APC) served as controls. In addition, CD137L-DCs effectively polarized naïve CD8⁺ T cells and cross-presented melanoma-associated antigen MART-1 (Supplementary Fig. 2).

The ability of DCs to migrate to draining lymph nodes is essential for the induction of adaptive immune responses. CD137L-DCs, whether matured or not, displayed a significantly lower migratory activity than cDCs (Supplementary Fig. 3a). We, therefore, optimized the CD137L-DC maturation cocktail with PGE2 as an additional maturation factor to enhance migration of CD137L-DCs (Supplementary Fig. 3b). PGE2 may not impair DC potency as some cytokine and surface marker expression was reduced, while that of others was enhanced (Supplementary Figs. 3d–e, 4).

These data demonstrate the ability of CD137L-DCs to induce superior anti-EBV and anti-HBV CD8⁺ T-cell responses. Particularly matured CD137L-DCs were consistently more potent than matured cDCs across all the assays and antigen systems tested.

CD137L-DCs enhance antigen-specific killing by T cells

The ability of stimulated T cells to kill target cells in an antigen-specific manner is most important for the success of immunotherapy. Therefore, we tested the ability of CD137L-DC-activated T cells to lyse LMP2a⁺ (EBV) and HBsAg⁺ (HBV) target cells.

TCR-redirected T cells for LMP2a or HBsAg were generated for HLA-A2⁺ donors and stimulated with autologous antigen-pulsed, matured cDCs and CD137L-DCs (non-matured and matured) for 18 h. IFN-γ and TNF-α levels in co-culture supernatants of T cells with matured or non-matured CD137L-DCs were significantly higher than when matured cDCs or no APC were co-cultured, indicating an enhanced T-cell activation by CD137L-DCs (Fig. 2a, b, e, f).

Fig. 2.

CD137L-DCs enhance cytotoxicity of T cells. cDCs and CD137L-DCs were generated from HLA-A2⁺ donors and matured. On day 7, DCs were harvested and pulsed with 1 µg/ml of HLA-A2-restricted LMP2a or HBsAg peptide for 1 h. a, b, e, f Activation of TCR-redirected T cells by autologous peptide-pulsed DCs. c, d Killing of LMP2a peptide-pulsed c and un-pulsed d target cells by LMP2a TCR-redirected T cells stimulated with DCs. g, h Killing of HBsAg expressing target cells g and control cells h by HBsAg TCR-redirected T cells stimulated with DCs. Line and bar charts depict means ± standard deviations of triplicate measurements. ***p < 0.001 by 2 way ANNOVA. Mat matured. Results are representative of at least three independent experiments

When PGE2 was included in the maturation cocktail for CD137L-DCs, IFN-γ levels were reduced while TNF-α levels remained unchanged. However, IFN-γ and TNF-α levels in T-cell co-cultures with CD137L-DCs (whether matured or not) were in all cases significantly higher than in T-cell co-cultures with matured cDCs or T cells alone (No APC) (Fig. 2a, b, e, f).

T cells from these co-cultures were analyzed for their ability to lyse target cells. T cells stimulated by CD137L-DCs (non-matured or matured) had a ~ 20% higher cytolytic activity than T cells stimulated by matured cDCs, and an at least twofold higher cytolytic activity than T cells without any DC stimulation (No APC). Addition of PGE2 to R848 + IFN-γ as a maturation cocktail to enhance CD137L-DC migration did not affect the cytolytic activity of the T cells (Fig. 2c, d).

A similar trend as in the case of TCR-redirected T cells against LMP2a was observed for HBsAg where the target cell lysis by CD137L-DC-activated T cells was significantly higher than target cell lysis by cDC-activated T cells. Both matured as well as non-matured CD137L-DCs were more potent than cDCs in stimulating HBsAg-specific T cells, with maturation of CD137L-DCs further increasing their cytolytic activity (Fig. 2g, h).

These data demonstrate that CD137L-DCs significantly enhance the cytotoxic potential of T cells, and that PGE2 does not affect this critical T cell activity.

IL-1β secreted by CD137L-DCs contributes to superior T cells activation

IL-1β is a pyrogenic cytokine released by antigen presenting cells and plays an important role in promoting type 1, cell-mediated immune responses [23]. CD137L signaling in monocytes induced IL-1β secretion (Fig. 3a), which is caspase-dependent as it is significantly inhibited by the pan-caspase inhibitor Z-VAD (Fig. 3b). IL-1β was not only secreted at the beginning of monocyte to CD137L-DC differentiation, but also by differentiated CD137L-DCs. Maturation of CD137L-DCs further increased IL-1β levels profoundly, from 16.4 ± 0.6 to 137.4 ± 0.9 pg/ml. No IL-1β was detected in supernatants of cDCs (Fig. 3c). We could not measure IL-1β secretion by matured cDCs as IL-1β is a part of their maturation cocktail.

Fig. 3.

CD137L signaling induces an AIM2 inflammasome and IL-1β secretion which contributes to T-cell activation by CD137L-DCs. a IL-1β secretion by monocytes 18 h after induction of CD137L signaling measured by ELISA. b Effect of pan-caspase inhibitor Z-VAD (50 µM) and or the solvent DMSO on IL-1β secretion by monocytes 18 h after induction of CD137L signaling. Bar charts depict means ± standard deviations of duplicate measurements. ***p < 0.001 using Student’s t test. Mat matured. c CD137L-DCs were generated and matured. IL-1β levels were determined in DC culture supernatants by ELISA. Bar charts depict means ± standard deviations of duplicate measurements. ***p < 0.001 using Student’s t test. Mat, matured. d, e Change in AIM2 and IL-1β gene expression 4 h after induction of CD137L signaling followed by treatment with 5 µM ATP for 40 min. Cells cultured on uncoated plates without ATP treatment served as control. Data are represented as fold increase in gene expression relative to control and normalized to GAPDH. f, g Change in T-cell activation in presence of IL-1ra (10 µg/ml) or equal volume of solvent (H₂O) in CD137L-DC-T cell (ratio = 1:10) co-culture as measured by IFN-γ levels in supernatants by ELISA. Bar charts depict means ± standard deviations of duplicate measurements. *p < 0.05, ***p < 0.001 using student’s t test. Results are representative of three independent experiments

Caspase-dependent secretion of IL-1β is a marker for inflammasome induction. We initially identified the type of inflammasome in monocytic THP1 cells. Upon induction of CD137L signaling THP1 cells increased AIM2 gene expression when ATP was used as DAMP, while gene expression for NLRP3 and NLRC4 did not change significantly (Supplementary Fig. 5a). A corresponding increase in IL-1β gene expression and secretion was also observed (Supplementary Fig. 5b). The induction of AIM2 (Fig. 3d) and IL-1β (Fig. 3e) gene expression by CD137L signaling was confirmed in peripheral human monocytes.

IL-1β secretion by CD137L-DCs contributes to their ability to activate T cells. Addition of IL-1 receptor antagonist (IL-1Ra) to co-cultures of CD137L-DCs with allogeneic T cells significantly reduced secretion of IFN-γ at 24, 48 and 72 h. IL-1β is important for the T-cell-stimulating activity of non-matured (Fig. 3f) as well as matured (Fig. 3g) CD137L-DCs but had a bigger influence on non-matured CD137L-DCs.

These data demonstrate that CD137L signaling induces AIM2 inflammasome formation and IL-1β secretion by CD137L-DCs, which contribute to T-cell activation.

CD137L-DCs are pro-inflammatory DCs which polarize CD8⁺ T cells to a Tc1 phenotype and reduce exhaustion

Signals by costimulatory molecules are pivotal for T-cell activation and also for the direction of the T cell response. CD137L-DCs express high levels of CD137L (Fig. 4a) which has been demonstrated to be a powerful stimulator of T-cell activity, and to polarize T cells towards a cellular, type 1 response.

Fig. 4.

CD137L-DCs polarize CD8⁺ T cells to a Tc1 phenotype. a Expression of costimulatory molecules of matured (mat) cDC, CD137L-DCs, and mat CD137L-DCs determined by flow cytometry. Light grey histogram: unstained; dark grey histogram: marker-stained. Stated next to histograms are the percentages of positive cells and mean fluorescence intensity. b Chemokine levels in co-culture of LMP2a TCR-redirected T cells with autologous peptide (1 µg/ml) pulsed DCs. TCR-redirected T cells without any DC stimulation (No APC) served as controls. Heat map of chemokine levels from one experiment is shown. Data shown are representative of 2 independent experiments with 2 different donors. c Immune transcriptome of LMP2a TCR-redirected CD8⁺ T cells after activation with autologous peptide-pulsed (1 µg/ml) DCs determined via Nanostring assay. Data depicted are fold change v/s CD8⁺ T cells without DC stimulation (No APC) for 3 independent donors. d PD-1, CTLA-4, and TIM-3 levels on CD8⁺ T cells after co-culture with allogeneic DCs at DC: T cell ratio of 1:5 for 48 h. Bar charts depict means ± standard deviations of duplicate measurements. *p < 0.05, **p < 0.01 by 1 way ANNOVA. Mat matured. Results are representative of three independent experiments

Chemokines secreted by DCs also play an important role in shaping the T-cell response by facilitating the homing of circulating DCs and T cells to site of infection, inflammation, or lymphoid organs. For instance, DCs that polarize T cells to the Th1 or Th2 phenotype also secrete different chemokine profiles [24].

CD137L-DCs (non-matured or matured) express significantly higher levels of the pro-inflammatory chemokines CCL2 (MCP-1), CCL4 (MIP-1β), CCL5 (RANTES), CXCL1 (GRO), and CXCL10 (IP-10) than matured cDCs, while CCL22 (MDC), a chemokine preferentially expressed by Th2-polarizing DCs [24], was secreted at significantly higher levels by mat cDCs than by CD137L-DCs. Addition of PGE2 to the maturation cocktail of CD137L-DCs did not affect the chemokine secretion by matured CD137L-DCs with the exception of GRO where PGE2 increased secretion (Fig. 4b).

Not surprisingly, T cells stimulated by CD137L-DCs were more polarized towards a type 1, cellular immune response. We analyzed the immune transcriptome of LMP2a TCR-redirected CD8⁺ T cells which were stimulated by autologous LMP2a peptide-pulsed DCs. Genes associated with a Tc1 response such as CCL4, Granzyme B, IFN-γ, IRF4, CSF-1, and Tbet [25] were significantly stronger expressed in T cells stimulated by CD137L-DCs than in T cells stimulated by matured cDCs (Fig. 4c).

T-cell exhaustion is characterized by high levels of PD-1, CTLA-3 and TIM3 expression and low cytokine secretion [26]. Since T cells stimulated by CD137L-DCs secrete higher levels of cytokines, we hypothesized that they may express lower levels of exhaustion markers which could, indeed, be confirmed (Fig. 4d). These data demonstrate that CD137L-DCs are Th1-polarizing DCs that promote a Tc1 response with significantly less exhausted T cells.

T cells stimulated by CD137L-DCs have higher glucose utilization ability and are metabolically more active

T-cell proliferation and effector functions are energy-intensive processes and, therefore, highly dependent on the upregulation of glycolysis and oxidative phosphorylation [27, 28]. Activation of T cells by TCR and costimulatory molecules induces an increase in nutrient uptake and metabolic activity to meet these enhanced energy needs [29–31].

We, therefore, compared the rate of glycolysis and mitochondrial respiration under baseline and stressed condition of CD8⁺ T cells that were activated by cDCs or CD137L-DCs. CD8⁺ T cells that were stimulated by non-matured or matured CD137L-DCs had a higher glucose utilization than CD8⁺ T cells that were activated by matured cDCs or than T cells alone (No APC) as indicated by the higher extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) which are measures of the extend of glycolysis and oxidative phosphorylation, respectively. This was the case at basal level and after activation of the CD8⁺ T cells by a cocktail of anti-CD2, anti-CD3 and anti-CD28 antibodies (Fig. 5a, b).

Fig. 5.

T cells activated by CD137L-DCs are metabolically more active. CD8⁺ a, b Glucose utilization assay. a Extent of glycolysis as measured by the extracellular acidification rate (ECAR). b Extent of oxidative phosphorylation as measured by the oxygen consumption rate (OCR). c–f Energy phenotype assay c, d kinetic measurement. e, f Average of all time points. All measurements were performed in duplicates. Resting T cells without any DC stimulation (No APC) served as control. Results are representative of 2 independent experiments

To determine the energy phenotype, the metabolic inhibitors Oligomycin and FCCP were used to induce metabolic stress. ECAR and OCR levels were significantly enhanced in CD8⁺ T cells activated by CD137L-DCs (non-matured or matured) compared to CD8⁺ T cells activated by matured cDCs under metabolic stress (Fig. 5c–f). Thus, activation by CD137L-DCs enhances the metabolic activity and the glycolytic and oxidative phosphorylation capacity of CD8⁺ T cells.

Discussion

DC-based therapy has the potential to boost the immune system against tumor-associated antigens to induce a curative response as well as long-term protection from relapse. However, immune suppression by tumors, the scarcity of well-defined tumor antigens, and the lack of current DC vaccines to mount a sufficiently strong T-cell response are among the major factors for the so far limited success of DC-based cancer immunotherapy. The development of immune checkpoint inhibitors has provided a tool to reduce tumor-mediated immune suppression [32], and next-generation sequencing can identify cancer neoantigens as targets for immunotherapy. There are also improvements in the use of DC vaccines such as endowing DC with cytokine expression [33] or preconditioning the injection site to enhance DC migration [34–36]. However, a more potent alternative for current cDCs remains to be developed.

We have previously described a novel method for the development of highly potent inflammatory DCs (CD137L-DCs) which resemble inflammatory in vivo DCs found at sites of inflammation [9–11]. In this study, we report that CD137L-DCs induce stronger CD8⁺ T-cell responses against EBV and HBV than cDCs. Since both viruses are linked to several types of malignancies, CD137L-DCs could be the basis for an effective vaccine against EBV- and HBV-associated cancers, especially, since the capacity of DC to activate CD8⁺ T cells is pivotal for an anti-tumor effect [37].

It is not only the stronger stimulation of CD8⁺ T-cell responses against EBV and HBV but also the predominant Th1/Tc1 polarization by CD137L-DCs that is promising. CD8⁺ T cells activated by CD137L-DCs have a characteristic Tc1 signature with a significant upregulation of CCL4, Granzyme B, IFN-γ, IRF8, IRF4, and Tbet. They secrete high levels of IFN-γ, and based on low expression of PD-1, CTLA-4 and TIM3, are less exhausted. The Th1/Tc1 polarization is likely driven, at least in part, by the pro-inflammatory chemokines secreted by CD137L-DCs.

Another factor may be the higher expression of costimulatory molecules, especially of CD137L on CD137L-DCs. The fact that CD137L-DCs, which have been generated by inducing CD137L signaling, express more CD137L points to a positive feed-back loop for CD137L expression. Since CD137 signaling enhances the CD8⁺ memory T-cell responses, e.g., against HIV and influenza [38, 39], it can be hypothesized that CD137L-DCs will also induce stronger CD8⁺ memory T-cell responses.

The induction of a Th1/Tc1 response by reverse CD137L signaling is thus paralleled by the promotion of a type 1, cellular immune response by CD137 forward signaling. CD137 is expressed by activated T cells and NK cells and powerfully enhances their activities, enabling rejection of tumors, transplants, and viruses [40], and that is also the reason for inclusion of the CD137 cytoplasmic domain in chimeric antigen receptors [41].

TCR-engineered T cells have been used for the treatment of cancers [8] and, therefore, apart from being employed as stand-alone DC vaccines, CD137L-DCs can be used to enhance the effector functions of TCR-engineered T cells prior to adoptive transfer. The tumor microenvironment is known to have reduced glucose and oxygen availability, which inhibits the function of cytotoxic T cells. In this regard, the enhanced glycolysis, mitochondrial respiration, and glucose utilization potential of CD137L-DC-stimulated CD8⁺ T cells could give them an advantage in effectively competing for the limited amount of glucose at the tumor site.

CD137L-driven monocyte to CD137L-DC differentiation may be prevented in some diseases such as chronic lymphocytic leukemia (CLL). Soluble CD137 which is generated by differential splicing [42, 43], and which inhibits CD137–CD137L interaction [44], and which is present at high concentration in sera of CLL patients [45] would not only block T-cell costimulation but also CD137L-driven monocyte to inflammatory DC differentiation.

Thus, CD137L-DCs induce a strong Tc1 response against cancer-associated antigens with enhanced T-cell fitness to perform effector functions, making them promising candidates for a therapeutic cancer vaccine.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

CD137L

CD137 ligand

CD137L-DC

CD137 ligand-generated dendritic cell

cDC

Classical DC

DC

Dendritic cell

EBNA

Epstein–Barr virus nuclear antigen

EBV

Epstein–Barr virus

ECAR

Extracellular acidification rate

HBsAG

HBV surface antigen

HBV

Hepatitis B virus

LMP-1

Latent membrane protein

NUS

National University of Singapore

OCR

Oxygen consumption rate

PGE2

Prostaglandin E2

Author Contributions

Conception and design: Bhushan Dharmadhikari, Zulkarnain Harfuddin, and Herbert Schwarz. Development of methodology: Bhushan Dharmadhikari, Emily Nickles, and Zulkarnain Harfuddin. Acquisition of data: Bhushan Dharmadhikari, Emily Nickles, Zulkarnain Harfuddin, and Nur Diana Binte Ishak. Analysis and interpretation of data: Bhushan Dharmadhikari, Emily Nickles, Zulkarnain Harfuddin, Nur Diana Binte Ishak, Antonio Bertoletti, and Herbert Schwarz. Writing, review, and/or revision of the manuscript: Bhushan Dharmadhikari, Emily Nickles, Zulkarnain Harfuddin, Qun Zeng, Antonio Bertoletti, and Herbert Schwarz. Administrative, technical, or material support: Bhushan Dharmadhikari, Emily Nickles, Zulkarnain Harfuddin, Nur Diana Binte Ishak, Qun Zeng, and Antonio Bertoletti.

Funding

This research was supported by a Grant (NMRC/BnB/018b/2015) from the National Medical Research Council, Singapore.

Compliance with ethical Standards

Ethical approval

This article does not contain any studies with human participants or animals performed by any of the authors. PBMC and monocytes were isolated from buffy coats of healthy volunteers, which were obtained from the Blood Donation Center of the National University Hospital, Singapore, after obtaining donors’ consent and approval by the NUS IRB (# B-15-320E). Cell lines used were newly purchased from ATCC (Manassas, VA, USA).