Interleukin-15-induced CD56(+) myeloid dendritic cells combine potent tumor antigen presentation with direct tumoricidal potential

Abstract

Dendritic cells (DCs) are the quintessential antigen-presenting cells of the human immune system and play a prime role in coordinating innate and adaptive immune responses, explaining the strong and still growing interest in their application for cancer immunotherapy. Much current research in the field of DC-based immunotherapy focuses on optimizing the culture conditions for in vitro DC generation in order to assure that DCs with the best possible immunogenic qualities are being used for immunotherapy. In this context, monocyte-derived DCs that are alternatively induced by interleukin-15 (IL-15 DCs) have attracted recent attention due to their superior immunostimulatory characteristics. In this study, we show that IL-15 DCs, in addition to potent tumor antigen-presenting function, possess tumoricidal potential and thus qualify for the designation of killer DCs. Notwithstanding marked expression of the natural killer (NK) cell marker CD56 on a subset of IL-15 DCs, we found no evidence of a further phenotypic overlap between IL-15 DCs and NK cells. Allostimulation and antigen presentation assays confirmed that IL-15 DCs should be regarded as bona fide myeloid DCs not only from the phenotypic but also from the functional point of view. Concerning their cytotoxic activity, we demonstrate that IL-15 DCs are able to induce apoptotic cell death of the human K562 tumor cell line, while sparing tumor antigen-specific T cells. The cytotoxicity of IL-15 DCs is predominantly mediated by granzyme B and, to a small extent, by tumor necrosis factor-α (TNF-α)-related apoptosis-inducing ligand (TRAIL) but is independent of perforin, Fas ligand and TNF-α. In conclusion, our data provide evidence of a previously unappreciated role for IL-15 in the differentiation of human monocytes towards killer DCs. The observation that IL-15 DCs have killer DC capacity lends further support to their implementation in DC-based immunotherapy protocols.

Figure 1. Phenotypic characteristics of CD56⁺ IL-15 DCs.

(A) CD14⁺ monocytes were cultured for 24–36 hr in the presence of GM-CSF and IL-15 (IL-15 DCs) and analyzed by flow cytometry for expression of CD11c/CD56 (left). The percentage between parentheses indicates the mean (± SEM) percentage of CD56⁺ cells among the total IL-15 DC population (n = 17). These CD56⁺ cells were then immunomagnetically separated, cultured for another 16–20 hr in the presence of DC maturation cocktail and analyzed for co-expression of CD11c/BDCA-1 (middle) and CD56/CD7 (right). Quadrant gates were set using corresponding isotype controls. (B) Matured CD56⁺ IL-15 DCs were further analyzed by flow cytometry for expression of the indicated NK cell-associated (CD56, CD7, CD16, CD69, NKG2D, NKp46), NKDC-associated (CD11c, B220, NKR-P1A) and DC-related surface antigens. Histogram overlays show expression of the indicated markers (solid line histograms) compared to their respective isotype controls (filled grey histograms). All plots are representative of at least 4 independent experiments.

Figure 2. Allostimulatory capacity of CD56⁺ and CD56⁻ IL-15 DCs in a MLR.

CD56⁺ and CD56⁻ IL-15 DCs were co-cultured with allogeneic, CFSE-labeled lymphocytes at a 1∶10 ratio for 5 days. CFSE-labeled cells stimulated with IL-4 DCs (1∶10 stimulator/responder ratio) or PHA/IL-2 were used as positive controls. Histograms show the degree of CFSE dilution, indicative of T cell proliferation, among gated CD3⁺CD4⁺ T cells in the absence (filled grey histograms) and presence (solid line histograms) of the indicated stimulators. Numbers above the bracketed lines indicate the background-subtracted percentages of proliferated (i.e. CFSE^low) T cells within the CD3⁺CD4⁺ gate. Data shown are representative of three donors.

Figure 3. Differential ability of CD56⁺ and CD56⁻ IL-15 DCs to stimulate a WT1-specific CTL clone.

Mature CD56⁺ and CD56⁻ IL-15 DCs were electroporated (EP) with WT1 RNA and co-cultured with an HLA-A*0201-restricted WT1_126–134-specific CTL clone. Negative controls included: T cells cultured without DCs (T cells only), WT1 RNA-electroporated DCs cultured without T cells (DCs only), and T cells cultured with non-antigen-loaded DCs (non/mock EP). (A) After overnight incubation, IFN-γ concentrations (pg/mL) in the culture supernatants were measured by ELISA. Bars represent mean (± SEM) IFN-γ concentrations of triplicate wells of three independent experiments (**, P = 0.003). (B) Antigenic responses were quantified in parallel by IFN-γ ELISpot. Illustrative single-well images from an IFN-γ ELISpot plate of one representative experiment are shown. Values in parentheses indicate the mean (± SEM) number of IFN-γ spot-forming cells (SFCs) of triplicate ELISpot wells.

Figure 4. Lysis of K562 cells but not of a WT1-specific CTL clone by IL-15 DCs.

PKH67-labeled target cells were mixed at varying E:T ratios with mature CD56⁺ and CD56⁻ IL-15 DCs or, where indicated, with conventionally generated IL-4 DCs and then subjected to PI/Annexin-V staining after overnight incubation. Target cell viability was defined as the percentage of PI⁻/Annexin-V⁻ cells within the PKH67⁺CD11c⁻ gate. (A) Viability profiles of gated K562 tumor cells cultured alone (control) or with either CD56⁻ or CD56⁺ IL-15 DCs at an E:T ratio of 50:1. One representative experiment out of five is shown. Percentages of viable K562 cells are displayed in the lower left quadrants and expressed as mean (± SEM) of 5 independent experiments. (B) Graph depicting the specific lysis of K562 tumor cells by CD56⁺ IL-15 DCs (solid black line, ▪; n = 5), CD56⁻ IL-15 DCs (solid grey line, □; n = 5) and IL-4 DCs (dashed grey line, ○; n = 3) at the indicated E:T ratios. Results are expressed as mean (± SEM) percentages of specific lysis. Asterisks refer to a statistically significant difference in cytotoxic activity at the indicated E:T ratio between CD56⁺ and CD56⁻ IL-15 DCs. (C) Bar graphs showing the viability of a WT1_126–134-specific CTL clone after overnight culture in the absence or presence of either CD56⁻ (□) or CD56⁺ (▪) IL-15 DCs at an E:T ratio of 50:1. Data are presented as mean (± SEM) percentages of viable T cells from three experiments.

Figure 5. Expression of lytic molecules by IL-15 DCs.

(A) Matured CD56⁺ IL-15 DCs were analyzed by flow cytometry for cell surface expression of TNF-α, FasL and TRAIL (solid line histograms). Filled grey histograms represent isotype controls. Data are from one experiment representative of three. (B) Both CD56⁺ (solid line) and CD56⁻ (dashed line) IL-15 DCs were stained for intracellular expression of TRAIL, granzyme B and perforin. Filled grey histograms represent isotype controls. One representative experiment out of 6 (for TRAIL) and 9 (for granzyme B and perforin) is shown. (C) Supernatants from overnight washout cultures of CD56⁺ (▪) and CD56⁻ (□) IL-15 DCs were analyzed for granzyme B release by ELISA. Bars represent mean (± SEM) granzyme B concentations (pg/mL) from 7 experiments.*, P<0.05.

Figure 6. Inhibition of CD56⁺ IL-15 DC-mediated cytotoxicity by neutralizing anti-TRAIL mAbs and concanamycin A.

Matured CD56⁺ IL-15 DCs were co-cultured with PKH67-labeled K562 target cells at an E:T ratio of 50:1 in the presence of either anti-TRAIL blocking mAb (left) or the granule exocytosis inhibitor concanamycin A (right). Parallel experiments were performed using TRAIL isotype-matched control mAb and medium control devoid of concanamycin A, respectively. Lysis of target cells was determined after overnight incubation using a flow cytometry-based cytotoxicity assay, as described above. Results are expressed as mean (± SEM) percentages of specific target cell lysis. Data are from 5 (for TRAIL) and 10 (for concanamycin) independent experiments. *, P<0.05; **, P<0.01.