|Home | About | Journals | Submit | Contact Us | Français|
T-bet (TBX21) is a transcription factor required for the optimal development of Type-1 immune responses. While initially characterized for its intrinsic role in T cell functional polarization, endogenous T-bet may also be critical to the licensing of Type-1-biasing antigen presenting cells. Here, we investigated whether human dendritic cells (DC) genetically engineered to express high levels of T-bet (i.e. DC.Tbet) promote superior Type-1 T cell responses in vitro. We observed that DC.Tbet were selective activators of Type-1 effector T cells developed from the naïve pool of responder cells, while DC.Tbet and control DC promoted Type-1 responses equitably from the memory pool of responder cells. Naïve T cells primed by (SEB or tumor-associated protein-loaded) DC.Tbet exhibited an enhancement in Type-1-, and a concomitant reduction in TH2- and Treg-associated phenotype/function. Surprisingly, DC.Tbet were impaired in their production of IL-12 family member cytokines (IL-12p70, IL-23 and IL-27) when compared with control DC, and the capacity of DC.Tbet to preferentially prime Type-1 T cell responses was only minimally inhibited by cytokine (IL-12p70, IL-23, IFN-γ) neutralization or receptor (IL-12Rβ2, IL-27R) blockade during T cell priming. The results of transwell assays suggested the DC.Tbet-mediated effects are predominantly the result of direct DC-T cell contact or their close proximity, thereby implicating a novel, IL-12-independent mechanism by which DC.Tbet promote improved Type-1 functional polarization from naïve T cell responders. Given their superior Type-1 polarizing capacity, DC.Tbet may be suitable for use in vaccines designed to prevent/treat cancer or infectious disease.
Dendritic cells (DC) are professional antigen-presenting cells that capture, process, and present antigens to T cells in the form of peptides complexed with MHC molecules (1, 2). DC support the activation and functional maturation of TH1, TH2, TH17, and regulatory CD4+ T cells, as well as, CD8+ T cells, NK cells, and innate myeloid immune cells (3–6).
In a diverse array of infectious disease states and in the cancer setting, host protection is largely afforded via the generation of Type-1 immunity (7). Type-1 T cell induction is believed to require DC presentation of cognate antigen, in addition to costimulator molecules, such as B7 and TNF family member molecules, and polarizing cytokines, such as IL-12, IL-23, and IL-27 (8, 9). Functional polarization of Type-1 T cells can be augmented by IL-12 and IL-27, which act through STAT4 and STAT1, respectively, to promote IFN-γ and Type-1 associated accessory molecules (10, 11). However, IL-12/STAT-4-independent mechanisms of Type-1 T cell induction have also been reported (12, 13). In such cases, Type-1 polarization requires intrinsic expression of the T-bet transcription factor in T cells which is regulated in a TCR- and STAT1-dependent manner (14–16). Silencing of T-bet in T cells suppresses IFN-γ and STAT1 expression levels during antigen-specific T cell differentiation, resulting in the unbalanced development of IL-4-secreting TH2 cells (17, 18). Conversely, T-bet expression suppresses TH2 differentiation by interfering with the Type-2 transactivator GATA-3 (19, 20).
Intrinsic, low-level expression of T-bet in (at least a subset of) DC also appears crucial to the generation of Type-1 immunity (15, 21–23). Our results suggest that human DC, engineered using recombinant adenovirus to express high levels of T-bet protein in a high percentage of DC, selectively prime and expand Type-1 T cells from naïve precursors in vitro, while concomitantly restricting TH2 and Treg polarization profiles. Notably, DC.Tbet cells pulsed with tumor antigenderived protein or peptide epitopes proved to be superior activators of melanoma antigen-specific TH1 and TC1 effector cells in vitro, thereby supporting the potential utility of these APC in vaccines against infectious disease or cancer.
DC (> 98% CD11c+CD14−) were generated from normal donors with written consent under an IRB-approved protocol, as previously described (24). Where indicated, day 5 immature DC were activated for 24h by incubation with inflammatory stimuli including i.) IL-1β (25 ng/mL; Sigma-Aldrich, St. Louis, MO) + TNF-α (50 ng/mL; Sigma-Aldrich) + IFN-α (3000 units/mL; Intron A-IFNα2b; Schering-Plough Corp., Kenilworth, NJ) + IFN-γ (1000 units/mL; Endogen, Woburn, MA) + polyinosinic:polycytidylic acid (poly I:C; 20 µg/mL, Sigma-Aldrich) yielding αDC1 (25); ii.) LPS (250 ng/mL; Sigma-Aldrich) yielding DC.LPS; iii.) LPS (250 ng/mL) + IFN-γ (1000 units/mL) yielding DC.LPS/IFN or iv.) 1 nM bryostatin-1 (Sigma-Aldrich) yielding DC.BS1 (26).
Plastic non-adherent cells, enriched in T cells, were collected and stored at −80°C for 5–7 days during the DC culture period. After thawing, naïve or memory T cells were negatively-isolated using CD45RO or CD45RA MACS™ microbeads (Miltenyi Biotec, Auburn, CA), respectively, per the manufacturer’s protocols. Isolated cell populations were > 98% pure based on corollary flow cytometry analyses. In some cases, CD4+ or CD8+ naïve or memory T cell subsets could then be further isolated by positive-selection using specific MACS beads as indicated. In additional experiments, CD45ROneg and/or CD45RAneg cells were depleted of CD56+ cells, or they were separated into their CCR7+ vs. CCR7neg sub-populations using specific MACS beads (Miltenyi), as indicated.
Human T-bet (hT-bet) was PCR cloned from peripheral blood lymphocytes using the following primers: hT-bet: Fwd 5’-GTCGACGACGGCTACGGGAAGGTG-3’, Rev 5’-GGATCCTTAGTCGGTGTCCTCCAACC-3’. The product was then digested with the restriction enzymes SalI and BamHI and the 1.7Kb fragment containing full-length hT-bet was ligated into the adenoviral-Cre-Lox (Ad.lox) vector. After sequence validation, recombinant adenoviruses were generated, as previously described (27). The mock (empty) adenoviral vector Ad.ψ5 and/or Ad.EGFP (encoding the enhanced green fluorescence protein) were used as negative controls, as indicated. Adenoviral vectors encoding full-length human IL-12p70 or MART-1 protein (Ad.MART1) have been previously described (27, 28). All adenoviruses were expanded, purified and provided by the University of Pittsburgh’s Vector Core Facility (Shared Resource).
Day 5 immature DC were infected with adenoviruses at an MOI of 300 at 37°C for 48h as previously described (27). 293-T human kidney epithelial cells (American Type Culture Collection, Manassas, VA) were infected with Ad.MART1 or Ad.ψ5 at an MOI of 20 for 48h before being used to generate freeze-thaw cell lysates.
Antibodies reactive against T-bet (Santa Cruz Biotechnology, Santa Cruz, CA), CCR7, CD3, CD4, CD8, CD25, CD45RA, CD45RO, CD54, CD70, CTLA-4, IFN-γ, IL-17A, MHC I, MHC II (BD Biosciences, San Jose, CA), CD80, CD86, B7-H1, CXCR3, Granzyme-B, Foxp3, IL-4 (eBioscience, San Diego, CA), CD212, Jagged-1 (Jag-1), TGF-βRII (R&D Systems, Minneapolis, MN), IFN-γ, IL-4, IL-10 (Miltenyi Biotec), CD11c, GITR, GITR-L (BioLegend, San Diego, CA), DLL4 (Novus Biologicals, Littleton, CO), MART-1 (Vector Laboratories, Burlingame, CA) or β-actin (Invitrogen, Carlsbad, CA) were used in flow cytometry, immunofluorescence microscopy and western blot experiments, as indicated. Anti-HLA-A2 mAbs BB7.2 and MA2.1 (American Type Culture Collection, ATCC, Manassas, VA) and anti-HLA-DR4 mAb 359-13F10 (the kind gift of Dr. Janice Blum (Indiana University) were used to determine the HLA phenotype of normal donors and melanoma patients based on flow cytometry analysis of PBMC. Neutralizing/blocking anti-hIL-12p70 polyclonal Ab (pAb; R&D Systems), anti-IL12Rβ2 pAb (R&D Systems), anti-hIL-23 pAb (R&D Systems), anti-IL-27R pAb (TCCR/WSX-1; R&D Systems), anti-hIFN-γ pAb (R&D Systems), anti-hIFN-γR1 monoclonal Ab (mAb; R&D Systems), anti-CD70 (Ancell, Bayport, MN) and rhNotch-1/Fc chimera (R&D Systems) were used at final concentrations of 10 µg/ml, per manufacturers’ recommendations.
Cell surface and intracellular staining of cells was performed and monitored by flow cytometry, as previously described (24). For immunofluorescence microscopy, 1 × 105 DC were cytospun and fixed onto slides. Cells were permeabilized and stained with T-bet primary antibody (Santa Cruz) and conjugated with goat anti-mouse Alexa Fluor 488 secondary antibody (Invitrogen). The counterstains used included Hoechst nuclear dye (Sigma), and f-actin-binding rhodamine phalloidin (Invitrogen). Fluorescence images were then captured using an Olympus BX51 microscope (Olympus America, Melville, NY).
For mRNA analysis, DC were harvested on day 2 (48h post-transduction) and MACS™ isolated naïve or memory CD4+ T cells were harvested on day 3 after initial priming by DC. RNA was isolated with Trizol (Invitrogen). Reverse transcription was performed using MuLV reverse transcriptase (Applied Biosystems, Carlsbad, CA) and random hexamers (Applied Biosystems). Semi-quantitative PCR was used to amplify cDNA for expression of gene-specific products. Specific primers were used for IL-12p35, IL-12/23p40, IL-23p19, IL-27p28, EBI-3, IL-15, IL-18, IL-10, TGF-β, IFN-α, and IFN-γ as previously described (24). Additional primer sequences included: T-bet: Fwd 5’-CCACCAGCCACTACAGGATG-3’ and Rev 5’-GGACGCCCCCTTGTTGTTT-3’; GATA-3: Fwd 5’-GTGCTTTTTAACATCGACGGTC-3’ and Rev 5’- AGGGGCTGAGATTCCAGGG-3’; Foxp3: Fwd 5’-GCACCTTCCCAAATCCCAGT-3’ and Rev 5’-TAGGGTTGGAACACCTGCTG-3’; and RORγt: Fwd 5’-AAATCTGTGGGGACAAGTCG-3’ and Rev 5’-TGAGGGTATCTGCTCCTTGG-3’. β-actin primers were used as an internal positive control (24).
To estimate the profile of cytokines produced by DC after cognate DC-T cell interaction, 6 × 104 DC were co-cultured with J558 CD40-Ligand (i.e. CD40L contributes signals normally provided by newly-activated T cells; ref. 29) expressing fibroblasts for 24h at a DC:J558 ratio of 1:2 in 96-well flat-bottom plates in 200 µl of AIM-V culture medium. Supernatants were collected and stored at −80°C prior to analysis using commercial ELISAs for human IL-12p70, IL-23, TNF-α and IL-10 (all BD Biosciences, except for IL-23 ELISA from BenderMedSystems, Burlingame, CA). Additional studies included DC stimulation for 24h using agonists to TLR2 (HKLM; Invivogen, San Diego, CA) TLR3 (poly I:C; Sigma-Aldrich, St. Louis, MO), TLR4 (LPS, Sigma-Aldrich), TLR5 (flagellin; Invivogen), TLR7 (imiquimod; Invivogen), as well as a trimeric form of CD40 ligand (the kind gift of Dr. Andrea Gambotto, University of Pittsburgh), as indicated.
The superantigen staphylococcus enterotoxin B (SEB) model for priming autologous T cells was used in these studies (6, 30). Briefly, DC.Tbet (ectopic T-bet expressing DC) or control DC were pulsed with SEB (Sigma-Aldrich) at 0.1–10 ng/ml (with a standard dose of 1 ng/ml selected for standard use based on preliminary studies; Supplemental Fig.1) in AIM-V media (Invitrogen) for 3h at 37°C prior washing and the addition of 104 DC to 96-well round bottom plates. Sorted CD45ROneg (naïve) or CD45RAneg (memory) T cells were labeled with 0.5 µM of CFSE (Invitrogen) in PBS for 15 minutes at 37°C, before being washed twice, with 105 T cells [resuspended in TcMEM: Iscove’s Modified Dulbecco Medium supplemented with 10% heat-inactivated human AB serum, L-glutamine, penicillin/streptomycin and non-essential amino acids; all reagents from Invitrogen with the exception of serum (Sigma-Aldrich)] added to wells containing DC along with 100 U/ml of rhIL-2 (Peprotech). Responder T cells were evaluated for CSFE dilution by flow cytometry on day 3 of co-cultures.
T cells were plated with SEB pulsed-DC.Tbet or control DC at an E:T ratio of 1:10 in TcMEM. Supernatant of DC-T cell co-cultures were collected on day 3 and analyzed for hIFN-γ production using a commercial ELISA (BD Biosciences). Additionally, on day 3, CD4+ T cells were MACS™ isolated from DC co-cultures. Total RNA was isolated for RT-PCR analysis or T cells were co-stained with mAbs to CD4, CD212 (IL-12Rβ2) and T-bet for flow cytometric analysis. In additional studies, T cells cultured with SEB pulsed-DC.Tbet or control DC on day 0 were restimulated on day 5 with identically-prepared DC and supplemented with 20 U/ml of rhIL-2 (Peprotech) and 5 ng/ml of rhIL-7 (Sigma-Aldrich) every other day. On day 12 or 14 of co-culture, T cells were collected and assayed for cytokine (IFN-γ, IL-4, IL-17A and IL-10), cell surface (CXCR3), and intracellular (Foxp3 and Granzyme-B) protein expression by flow cytometry. To evaluate intracellular cytokine expression, T cells were stimulated with PMA (1 µg/ml) and Ionomycin (50 ng/ml) for 4h, with 2 µM monensin (all from Sigma-Aldrich) added over the final 2h of culture. Where indicated, cell culture supernatants were analyzed for secreted levels of hIFN-γ, hIL-4, hIL-10 (all from BD Biosciences) and hIL-17A (BioLegend) using commercial ELISAs.
Briefly, DC.Tbet or control DC were plated with T cells at a DC:T cell ratio of 1:10 in triplicate in 96-flat bottom plates, in the presence or absence of neutralizing/blocking Abs or recombinant fusion protein. On day 3, cell-free supernatants were collected and evaluated using a human IFN-γ-specific ELISA. Alternatively, T cells were restimulated on day 5 with SEB-pulsed DC and supplemented with rhIL-2, rhIL-7, and neutralizing antibodies, with T cells harvested on day 14 for analysis of intracellular IFN-γ production by flow cytometry.
DC.Tbet or control DC (5 × 105) were plated in the bottom chamber of a 24-well transwell plate in 400 µl of TcMEM. Twenty-four hours later, 1 × 106 naïve T cells along with 1 × 105 SEB pulsed-immature DC or 3 × 105 anti-CD3/CD28 microbeads (Invitrogen) were placed in the upper chamber of the transwell plate bringing total volume to 600 µl of TcMEM. Cell supernatants were collected from the upper chamber on day 3 for performance of IFN-γ ELISA analyses.
293T human kidney epithelial cells (ATCC) were infected with Ad.MART1 at an MOI of 20 for 48h, at which time freeze-thaw lysates were generated as previously described (31). 293T cells infected with Ad.ψ5 (MOI = 20) were used to generated a negative control lysate. Expression of MART-1 mRNA/protein in transduced 293T cells was determined using RT-PCR and immunohistochemistry, and MART-1 protein in lysates confirmed by western blot, as previously described (28). Total lysate protein content was estimated by OD280nm (1.2 mg = 1.0 AUFS280nm) and lysates stored at −80°C until being used to load DC for T cell induction and recognition assays.
MACS™-isolated, naive CD4+ T cells were isolated from HLA-DR4+ (based on monocyte staining with anti-HLA-DR4 mAb 359-13F10 as monitored by flow cytometry) normal donors as outlined above, and stimulated twice on a weekly basis with control DC (DC.null) pulsed with the MART-151–73 peptide (10 mM). On day 14 of culture, T cells were harvested and assessed for their ability to recognize (in IFN-γ ELISPOT assays) autologous DC.null, DCψ5 or DC.Tbet cells pre-pulsed for 48h with freeze-thaw lysates (50 µg/ml) generated from Ad.MART-1- vs. Ad.ψ5-infected 293T cells.
Peripheral blood mononuclear cells were isolated from healthy, normal donors with written-consent under an IRB-approved protocol. For CD8+ T cell responses, DC were generated from HLA-A2+ normal donors (i.e. lymphocytes staining with both the anti-HLA-A2 mAbs BB7.2 and MA2.1 as monitored by flow cytometry), as outlined above, and DC.Tbet or control DC were pulsed with HLA-A2-restricted peptide epitopes (EphA2883–891, gp100209–217; 10 µM each; refs. 25, 32) for 3h at 37°C prior to culturing with MACS™-isolated naïve CD8+ T cells at a 10:1 T cell:DC ratio in presence of 5 ng/ml of rhIL-7. CD8+ T-cell cultures were expanded by a second stimulation on day 7 with identically-prepared DC or with peptide-pulsed autologous, irradiated PBMC. Restimulated cultures were supplemented with 20 U/ml of rhIL-2 and 5 ng/ml of rhIL-7, with cytokines replenished every other day. On day 14, the frequency of peptide-specific CD8+ T cells was analyzed in IFN-γ ELISPOT assays (Mabtech, Stockholm, Sweden) using HLA-A2+ T2 cells as antigen presenting cells that were performed as previously described (32). The HLA-A2-presented HIV-nef190–198 peptide (32) served as a (negative) specificity control in these assays. For CD4+ T cell responses, DC were generated from HLA-DR4+ normal donors (based on monocyte staining with anti-HLA-DR4 mAb 359-13F10 as monitored by flow cytometry) as outlined above. DC.Tbet or control DC were pulsed with 50 µg/ml of freeze-thaw lysate generated from 293T cells infected with Ad.MART1 vs. Ad.ψ5 for 24h, 37°C and used to stimulate autologous MACS™-isolated, naive CD4+ T cells, as outlined above for CD8+ T cell responses. On day 7 of cultures, responder CD4+ T cells were restimulated with identically-prepared, Ag-loaded DC and cultures supplemented with rIL-2 and rIL-7 as noted above. On day 14, the frequency of MART-1-specific CD4+ T cells was analyzed in IFN-γ and IL-5 ELISPOT assays (Mabtech) using autologous control DC pulsed with the MART-151–73 vs. the HIV-nef192–204 (negative) control HLA-DR4-presented peptide epitopes as target cells (33).
A two-tailed Student’s t test was used for data analysis. Null hypothesis was rejected and differences were assumed to be significant at a value of p < 0.05.
Human DC were generated from peripheral blood monocytes and transduced with recombinant adenovirus encoding human T-bet (DC.Tbet) or control Ad.ψ5 (DC.ψ5) for 48h. DC were also generated using known Type-1 polarizing culture conditions, yielding DC1. Harvested DC were analyzed for T-bet mRNA (via RT-PCR; Fig. 1A) and protein expression (via Western Blot and flow cytometry; Fig. 1B, C). As shown in Fig. 1A–C, T-bet expression in untreated immature DC (DC.null) and DC.ψ5 was very low (at both the transcript and protein levels), with expression levels augmented in DC.null cells by 24h culture in the presence of inflammatory stimuli (24–27). However, in marked contrast to the < 1 % frequency of T-bet+ DC developed using non-viral culture methods, DC.Tbet were 63 ± 18% T-bet+ over 15 independent experiments as determined by intracellular staining, as exemplified in Fig. 1C. Notably, immunofluorescence microscopy revealed that T-bet protein was expressed predominantly in the nucleus of DC.Tbet cells (Fig. 1D).
Given previous reports that intrinsic (low-level) expression of T-bet in DC is crucial to the ability of these APC to promote Type-1 T cell responses (15, 21–23), we hypothesized that DC.Tbet cells might be enhanced in this capacity. We used a superantigen (staphylococcus enterotoxin B; SEB) model to investigate DC.Tbet-induced “specific” responses from naïve vs. memory T cell populations in vitro. Briefly, DC.Tbet or control DC were pulsed with 1 ng/ml of SEB prior to co-culture with autologous naïve (MACS-isolated CD45ROneg cells) or memory (MACS-isolated CD45RAneg) bulk (CD4+ and CD8+) T cells at a DC:T cell ratio of 1:10 for 72h. These condtions were chosen based on dilutional analyses (Supplemental Fig.1) in which optimal IFN-γ was observed from responder T cells within both the control DC- and DC.Tbet-stimulated cohorts at an SEB dose of 1 ng/ml.
As shown in Fig. 2A, we noted that IFN-γ production from activated naïve, but not memory, bulk (CD4+ and CD8+) T cells was significantly up-regulated when primed by SEB-pulsed DC.Tbet vs. SEB-pulsed control DC (p = 0.004). Macroscopically, DC.Tbet activated cultures developed from naïve, bulk T cell precursors contained very large cellular clusters (Fig. 2B), suggestive of differentially-enhanced T cell proliferation within such cultures. However, a repeated series of assays implementing bulk CD45ROneg vs. CD45RAneg T cells that were pre-labeled with 0.5 µM CFSE prior to co-culture with control DC or DC.Tbet, revealed no significant changes in the frequencies of daughter cell generations (CD45ROneg T cells; Figs. 2C and 2D) or T cell yields on days 3 or 7 of culture (Fig. 2C), although the enhanced ability of daughter T cells to produce IFN-γ in DC.Tbet (+ CD45ROneg bulk T cell) co-cultures was readily apparent (Fig. 2D). This latter increase was evident in both the percentage of IFN-γ+CSFElow+ events (Fig. 2D) and the approximate doubling in MFI levels for IFN-γ expression in responder T cells (157 for DC.Tbet cultures vs. 73 or 71 for DC.null or DC.ψ5 cultures, respectively; data not shown). These data strongly suggest that DC.Tbet enhance Type-1 responses from bulk, CD45ROneg T cells via differential polarizing, rather than proliferative, signals.
Since our initial bulk CD45ROneg responder cell population also contained a sub-population of approximately 15–18% CD4negCD56+ NK/NK T cells (Fig. 3A), which could serve as a direct source of IFN-γ and/or act as an intermediary for DC-induced Type-1 T cell function (34), we MACS™-isolated CD45ROnegCD56neg T cells (Fig. 3A) and repeated our in vitro stimulation assays using autologous SEB-pulsed DC.Tbet vs. control DC as APC. As shown in Fig. 3B, depletion of CD56+ cells from CD45ROneg, bulk T cell responders did not inhibit the ability of DC.Tbet to promote superior IFN-γ production. This was further corroborated for CD4+ T responder cells positively-isolated from the CD45ROneg and CD45RAneg bulk populations of cells (Fig. 3C), where SEB-pulsed, autologous DC.Tbet elicited superior IFN-γ production only from CD4+CD45ROneg responder T cells (Fig. 3D).
Having discounted the importance of contaminant NK cells as a source of IFN-γ resulting from DC.Tbet priming of bulk, CD45ROneg responder cells, we next considered differential responsiveness of the various T cell functional subsets to DC.Tbet-based stimulation. Since T cell functional subsets may be discriminated into naïve (CD45ROnegCCR7+CD62L+), effector (TE; CD45ROneg, CCR7neg, CD62Ldim+ or CD45RAneg, CCR7neg, CD62Ldim+), central memory (TCM; CD45RAneg, CCR7+, CD62L+) and effector memory (TEM; CD45RAneg, CCR7neg, CD62Ldim+) subpopulations based on their composite phenotypes (35), we performed CCR7-MACS selection after first isolating CD45ROneg and CD45RAneg cells (Fig. 4A). These populations, enriched in the various T cell functional subsets, were then stimulated with autologous SEB-pulsed DC.Tbet vs. control DC for 72h and culture supernatants analyzed for IFN-γ levels (Fig. 4B). Despite the lack of absolute purity for each of the T cell functional subsets, only naïve T cells that were highly-enriched (approximately 90% pure) for the CCR7+CD62L+ phenotype exhibited differential responsiveness to DC.Tbet (vs. control DC) based on a dramatic upregulation in their production of IFN-γ (Fig. 4B).
To further assess responder T cell polarization status, we isolated CD3+ T cells from DC-bulk T cell co-cultures after 72h, and analyzed these cells for their comparative expression of mRNAs encoding transactivator proteins (i.e. T-bet (TH1), GATA-3 (TH2), RORγt (TH17), and Foxp3 (Treg)) linked to T cell function (Fig. 5A). We observed that naïve T cells stimulated with DC.Tbet cells were enriched (approximately 5-fold as assessed by densitometry analysis of gel bands; data not shown)) in T-bet, and reduced in GATA3 (approximately 4-fold), RORγt (less than 2-fold) and Foxp3 (approximately 5-fold) transcripts when compared to T cells stimulated with control DC (Fig. 5A). Furthermore, as T-bet directly binds to the IL-12Rβ2 promoter and enhances its expression in T helper subsets (8), we performed flow cytometry analyses on CD4+ T cells harvested from DC.Tbet-driven cultures established with naïve T cell responders. These analyses revealed that responder T cells were enriched in cells bearing the IL-12Rβ2+T-bet+ phenotype in DC.Tbet (vs. control DC)-driven cultures (Supplemental Fig. 2A). Corollary studies revealed that DC.Tbet differentially (vs. control DC) induced naïve (Supplemental Fig. 2B), but not memory (Supplemental Fig. 2C) Tc1 cell responses, based on CD8+ T cell expression of IFN-γ and Granzyme-B. Notably, responder CD4+ and CD8+ T cell expression of the CXCR3 chemokine receptor (associated with Type-1 T cell recruitment into (inflammatory) tumor sites; ref. 36) was also increased approximately 2-fold (based on MFI levels) if these T cells had been activated by DC.Tbet vs. control DC (Supplemental Fig. 2A and data not shown).
Our preliminary analyses support large decreases in GATA-3 and Foxp3 (and little-to-no change in ROR-γt) mRNA expression levels in naïve, bulk T cells primed using DC.Tbet vs. control DC (Fig. 5A). To corroborate these findings at the protein level, we assessed the polarization state of responding CD4+ T cells by analyzing their cytokine production profiles. We confirmed reductions in the levels of IL-4 and IL-10 produced by naïve (but not memory) T cells stimulated with autologous SEB-pulsed DC.Tbet (vs. control DC; p < 0.05) as analyzed in ELISA and intracellular staining protocols (bulk cells analyzed in Fig. 5B and CD3+ T cells assessed in Fig. 5C, respectively). We also noted that the frequency of responder CD4+Foxp3+ T cells was reduced after activation of naïve (but not memory T cells) with SEB-pulsed DC.Tbet vs. control DC (Fig.5D). In slight contradiction to the RT-PCR data reported for ROR-γt in Fig. 5A, we noted a modest increase in IL-17A protein production from naïve T cells primed using DC.Tbet vs. control DC (Figs. 5B, 5C).
The ability of DC.Tbet to selectively augment Type-1 responses from naïve T cells initially suggested the likely involvement of DC-produced IL-12 family members such as IL-12p70, IL-23 and IL-27 (8, 10, 11). We found that although DC.Tbet expressed reduced levels of IL-27p28 mRNA, transcript levels for all other IL-12- and IL-23-associated mRNAs, as well as, a number of alternate DC-associated cytokines were unchanged in DC.Tbet vs. control DC (Fig. 6A). Strikingly, despite DC.Tbet exhibiting an essentially control DC cytokine mRNA profile, these APC were profoundly suppressed (vs. control DC) in their capacity to secrete any cytokine evaluated (i.e. IL-12p70, IL-23, TNF-α and IL-10) either spontaneously, or in response to, CD40 ligation or TLR stimulation (Fig. 6B and Supplemental Fig. 3). Consistent with the lack of expression of IFN-γ mRNA in any DC population analyzed in Fig. 6A, IFN-γ was not produced at detectable levels by any of the DC cohorts (i.e. < 4.7 pg/ml as determined by specific ELISA; data not shown). Additional analyses suggest that the inability of DC.Tbet to produce these cytokines was not the result of reduced DC vitality or enhanced sensitivity of these APC to apoptosis vs. control DC (Supplemental Fig. 4).
Despite low levels of cytokine production by DC.Tbet, we evaluated whether IL-12 family member cytokines (or IFN-γ itself) were involved in the priming of Type-1 polarized T cell responses by SEB-pulsed DC.Tbet vs. control DC. In vitro stimulations of naïve, bulk T cells were recapitulated in the absence or presence of neutralizing/blocking antibodies reactive against IL-12p70, IL-23, IL12Rβ2, IL27R and/or IFN-γ (Fig. 7A–C). IFN-γ production by T cells primed by all control DC populations (including DC/IFN, DC/IFN + LPS and αDC1) was clearly dependent on IL-12p70 and/or IFN-γ itself, as well as, a functional IL-12Rβ2-dependent signaling pathway. However, this was not the case for naïve, bulk T cells activated using DC.Tbet cells. Indeed, antagonism of these cytokines/cytokine receptors did not significantly affect the ability of DC.Tbet to prime Type-1 T cell responses in vitro (Fig. 7A–C).
Since RT-PCR and ELISA analyses suggested the coordinate silencing of cytokine secretion by DC.Tbet, this implicated the likely dominant involvement of cell membrane interactions rather than soluble mediators in the differential ability of DC.Tbet to drive Type-1 T cell responses in vitro. We confirmed this hypothesis by co-culturing CD45ROneg, bulk T cells with anti-CD3/CD28 mAb-coated beads or with SEB-pulsed control DC in the upper chambers of transwell plates, with DC.Tbet or control DC placed in the lower chambers. After 72h of culture, supernatants harvested from the various T cell cultures were all found to contain comparable levels of IFN-γ (Fig. 7D), suggesting that physical separation of DC.Tbet from responder T cells mitigates their capacity to promote superior Type-1 immunity in vitro.
To determine whether DC.Tbet were capable of promoting enhanced TC1 immunity against tumor antigens (such as EphA2 (32) and gp100 (25)), naïve CD8+ T cells were isolated from HLA-A2+ normal donors and then co-cultured with autologous DC.Tbet or control DC pulsed with an equimolar mixture of the EphA2883–891 and gp100209–217(2M) peptides. T cells were restimulated after 7 days of culture and then assessed for populational frequencies of peptide-specific, IFN-γ-producing CD8+ T cells on day 14. We observed that T cell cultures primed using DC.Tbet (vs. control DC) contained significant increases in their frequencies of Type-1 CD8+ T cells reactive against both the EphA2 and gp100 peptides, but not a negative control HIV-nef peptide epitope (Fig. 8A). Notably, we observed elevated antigen-specific responses for the DC.Tbet-primed cohort of CD8+ T cells regardless of whether peptide-pulsed autologous DC.Tbet or PBMC were used as APC in the restimulation phase of this experiment (Fig. 8A). This supports the likelihood that the dominant impact of DC.Tbet on specific TC1 responses occurs during the priming phase.
To address whether DC.Tbet were similarly capable of promoting improved TH1 responses against a tumor antigen, we initially showed that these APC were fully competent to uptake and process exogenous recombinant MART-1 protein (in the form of a freeze-thaw lysate of 293T previously transduced with a recombinant adenovirus encoding hMART-1; Supplemental Fig. 5) and then present the derivative HLA-DR4-presented MART-151–73 epitope (37) to a peptide-specific CD4+ T cell line (Supplemental Fig. 6). To determine whether MART-1 protein-pulsed DC.Tbet cells were competent to preferentially prime Type-1 responses from naïve CD4+ T cell responders, DC.Tbet and control DC were loaded with 293T.MART1 lysate for 24h and then used to prime and boost (on day 7 of culture) autologous, naive CD4+ T cells isolated from normal HLA-DR4+ donors. As shown in Fig. 8B, CD4+ T cells analyzed on day 14 of culture displayed superior levels of reactivity against the MART-151–73 peptide epitope in IFN-γ (and reduced specific responses in IL-5) ELISPOT assays using autologous DC.null as APC if they had been developed using MART-1+ lysate-pulsed DC.Tbet vs. control DC (p < 0.05).
The transcription factor T-bet was originally identified as a master regulator of TH1 development, but has since been found to differentially regulate genes in CD8+ effector T cells, B cells, and natural killer (NK) and NKT cells (38–40). In particular, Glimcher et al. have shown that endogenous expression of T-bet in DC is necessary for optimal induction of Type-1 T cell responses (22, 23, 40). A major finding in the current studies is that ectopic (over)expression of T-bet (as a result of recombinant adenoviral T-bet cDNA delivery) to “license” DC to preferentially support the in vitro development of Type-1 (over Type-2 and Treg) polarized responses from naïve (CD45ROnegCCR7+CD62L+), but not memory, T cell precursors. Preferential enhancement in Type-1 T cell development was reflected at the level of differential transactivator molecule mRNA expressed (with T-bet increased and GATA-3, as well as, Foxp3 being decreased) and cytokines secreted (with IFN-γ increased, and IL-4, as well as, IL-10 being decreased). Furthermore, levels of cell surface (CXCR3, IL-12Rβ2) and effector (Granzyme-B, IFN-γ) molecules associated with Type-1 functionality were increased in naïve T cells after specific activation with DC.Tbet vs. control DC. Although, ROR-γt mRNA transcripts appeared unaffected or, in some cases, somewhat reduced in naïve T cells primed with DC.Tbet vs. control DC, we found that the level of IL-17A secreted by these responder T cells tended to be modestly increased (p < 0.05 vs. control DC-stimulated T cells). This may not be too surprising due to the mutual functional exclusivity between Foxp3+ Treg (suppressed after DC.Tbet stimulation) and TH17 T cells (potential compensatory enhancement), as previously reported by others (41). Furthermore, we did not detect TH17 cells co-producing both IFN-γ and IL-17A (Fig. 5C), suggesting that IFN-γ analyzed in our studies is stringently associated with bona fide Type-1 T cell responses.
A second major finding in our work relates to the IL-12 cytokine family-independent mechanism(s) involved in DC.Tbet activation of Type-1 CD4+ and CD8+ effector cells from naïve T cells. Indeed, we noted that i.) production of IL-12p70, IL-23, and IL-27, as well as, all other cytokines evaluated, were suppressed in DC.Tbet vs. control DC and ii.) neutralizing antibodies against IL-12p70, IL-23p19, IL-12Rβ2 and IL-27R all failed to attenuate DC.Tbet-mediated induction of Type-1 responses from naïve T cells. It remains formally possible that the absence of cytokine (i.e. IL-23 and IL-27)-mediated signaling into T cells could reinforce their Type 1 functional polarization, as others have previously shown that i.) IL-27 mediates the differentiation of naïve T cells into IL-10 producing Tr1 cells (42) and ii.) signals mediated via the IL-23R are crucial for the development of TH17 responses (43).
Results obtained in transwell assays support the critical importance of DC.Tbet-T cell interaction or proximity in order for Type-1 polarizing signals to be conveyed during the T cell priming event. Yet a survey of DC surface molecules for expression levels revealed no striking differences between DC.Tbet and control DC.ψ5 (or DC.EGFP) for MHC molecules, integrins, co-stimulatory/inhibitory molecules or modulatory receptors (Supplemental Figs. 7, 8. 9A). CD70 and NOTCH ligands delta like-4 (DLL4) and Jagged-1 (Jag-1) which have been previously shown to contribute to the functional polarization of responder T cells by DC (12, 13, 44), were not expressed (or expressed poorly) by DC.Tbet (Supplemental Fig. 9A), and appeared functionally irrelevant in our model system since the inclusion of specific blocking reagents had no perceptible impact on the ability of DC.Tbet to support enhanced Type-1 responses from CD45ROneg, bulk T cells (Supplemental Fig. 9B).
Overall, our data appear to support a novel mechanism by which DC.Tbet preferential prime Type-1 T cell responses from naïve T cell precursors. This is manifest in enhanced DC- naïve T cell clustering at early phases of the induction process (i.e. day 3) via a process that was not correlated with T cell proliferation/expansion based on CFSE dilution analyses in vitro. These data could suggest that DC.Tbet- naïve T cell interactions may be uniquely prolonged due to the sustained interfacing of key MHC/TCR and costimulatory/integrin/adhesion molecules and/or to the abbreviated impact of co-inhibitory or intercellular repulsion molecules (45–48), resulting a reinforced commitment of newly-primed T cells towards a state of Type-1 functional polarization. It is also possible that DC.Tbet may be refractory to dissociating signals, such as those contributed via newly-activated T cell-expressed CTLA-4 (49). If such interactions underlie the observed selective priming of Type-1 immunity by DC.Tbet, this could explain the inability of DC.Tbet to affect superior Type-1 responses from the activated, memory T cell population, since memory T cells are known to exhibit a lower activation threshold requirement for both signal 1 (MHC/peptide) and signal 2 (costimulation) when compared with naïve T cells (50). We are currently pursuing a further characterization (genomic, proteomic) of changes occurring in DC.Tbet that may be implicated in the selective priming Type-1 responses from CD45ROnegCCR7+CD62L+ T cells.
Type-1 T cell responses appear most efficient in regulating disease development and progression in the cancer setting (7, 24, 25, 27, 28, 30, 34). Hence, the ability to predictably generate tumor-specific Type-1 immunity is a major target for cancer immunotherapy-based approaches. A means to accomplish this goal includes the use of vaccines that may selectively and predictably augment the development of Tc1 and TH1 effector T cell populations. While such vaccines have commonly integrated autologous DC as a “biologic adjuvant” (7, 28, 51) over the past decade, significant heterogeneity in DC subsets and variable states of maturation have yielded equivocal results in both preclinical tumor models and clinical trials applying DC-based modalities (51).
In this context, methods to condition or engineer DC1 that are particularly competent to expand and develop Type-1 T cell-mediated anti-tumor immunity may improve clinical efficacy of DC-based cancer vaccines. In this regard, (IL-12p70-independent) DC.Tbet promote at least equitable Type-1 T cell responses to (IL-12p70-dependent) αDC1, a current “gold standard” for clinically-applied DC1 (25). Given the apparent non-overlapping mechanism of Type-1 immune induction by DC.Tbet and IL-12p70, it might be envisioned that these 2 agents might act synergistically in promoting TC1 and TH1 responses. We are currently evaluating this possibility in vitro.
Our in vitro stimulation experiments using tumor peptide (i.e. EphA2 and gp100) or protein (recombinant MART-1)-pulsed DC.Tbet clearly support the improved capacity of this “vaccine” to promote specific TC1 and TH1 responses in vitro from naïve CD8+ and CD4+ T cells, respectively. Such Type-1 T cells would be predicted to be competent to both infiltrate tumor lesions (as associated increases in CXCR3 expression is observed) in vivo (36, 52) and to mediate robust anti-tumor activity within these sites (53). Furthermore, since DC.Tbet retain their capacity to uptake whole (tumor) proteins and to process and then prime tumor Ag-specific, Type-1 CD4+ and CD8+ T cell responses in vitro, they may also be envisioned as a therapeutic modality to be injected directly into tumor lesions in vivo (where they may acquire and then preferentially prime Type-1 anti-tumor T cell responses). Overall, the potent capacity of DC.Tbet to promote Ag-specific Type-1 T cell responses while coordinately minimizing Type-2/Treg functional responses suggests that (DC.Tbet-based) vaccines may yield enhanced therapeutic efficacy in vivo (54) in the settings of cancer and infectious disease.
The authors wish to thank Drs. James Finke, Rathindranath Baral and Laurie Glimcher for helpful comments provided during the performance of this work and the preparation of this report.
1This work was supported by National Institutes of Health P01 grant CA100327 (W.J.S.), the University of Pittsburgh Skin Cancer SPORE CA121973 (W.J.S.) and training grant 5T32CA082084-08 (M.W.L.).
Publisher's Disclaimer: This is an author-produced version of a manuscript accepted for publication in The Journal of Immunology (The JI). The American Association of Immunologists, Inc. (AAI), publisher of The JI, holds the copyright to this manuscript. This version of the manuscript has not yet been copyedited or subjected to editorial proofreading by The JI; hence, it may differ from the final version published in The JI (online and in print). AAI (The JI) is not liable for errors or omissions in this author-produced version of the manuscript or in any version derived from it by the U.S. National Institutes of Health or any other third party. The final, citable version of record can be found at www.jimmunol.org.