Targeted delivery of CD40L promotes restricted activation of antigen-presenting cells and induction of cancer cell death

To determine the binding characteristics of anti-EpCAM:CD40L, the EpCAM-negative cell line HEK293 and the EpCAM-transfectant cell line HEK293.EpCAM were incubated with anti-EpCAM:CD40L and analyzed for CD40L by flow cytometry. Incubation of EpCAM-negative HEK293 cells with anti-EpCAM:CD40L did not trigger an increase in fluorescence upon anti-CD40L antibody staining (Figure 2A; left panel). By contrast, incubation of HEK293 cells engineered to express EpCAM (HEK.EpCAM) with anti-EpCAM:CD40L resulted in a significant increase in surface CD40L staining (Figure 2A; right panel). As with HEK293, anti-EpCAM:CD40L did not bind HEK293 cells that only expressed the intracellular domain of EpCAM (HEK293.EpICD; Figure 2A; middle panel). Further, incubation of the EpCAM⁺ colon carcinoma cell line DLD-1 with anti-EpCAM:CD40L showed a strong anti-CD40L fluorescence signal (Figure 2B; left panel) that was completely inhibited by pre-incubation with an epitope-competing EpCAM-blocking antibody (Figure 2B, middle panel). Similar antibody fragment-selective binding characteristics were obtained for the anti-CD20:CD40L fusion protein, with selective binding to CD20⁺ B cells (Figure 2C; left panel) that could be inhibited with the parental mAb rituximab (Figure 2C; middle panel). Furthermore, anti-CD20:CD40L failed to bind to CD20^- Jurkat T cells (Figure 2C; right panel). Taken together, these data demonstrate that the scFv:CD40L fusion proteins have a high binding affinity for their respective target antigens EpCAM and CD20.

To verify that N-terminal fusion of an scFv antibody fragment to sCD40L did not negatively affect the biological activity of the trimeric sCD40L, the capacity of scFv:CD40L fusion proteins to induce DC-maturation was assessed using anti-EpCAM:CD40L. Of note, non-targeted sCD40L has previously been shown to induce DC-maturation at relatively high concentrations ranging from 1 to 3 μg/ml [20]. Therefore, monocyte-derived immature dendritic cells (iDCs) were treated with high concentrations of the scFv:CD40L proteins (1 μg/ml) and analyzed for induction of DC maturation markers (Additional file 1: Figure S1A for protocol). Monocyte-derived iDCs had a typical DC morphology and expressed low cell surface levels of CD80 and CD86, intermediate levels of CD40 and HLA-DR, and lacked expression of CD83, a marker for mature DCs (mDC) (Figure 3A). Treatment of iDC with anti-EpCAM:CD40L dose-dependently increased the percentage of CD80^high/CD86^high DCs (Figure 3B) and significantly up-regulated HLA-DR expression (Figure 3C). Importantly, the HLA-DR^high but not the HLA-DR^low DCs also expressed the mDC markers CD83 and CCR7 (Figure 3D; histograms and MFI for all markers in Figure 3E and Additional file 1: Figure S1B, respectively) and retained a CD14-negative phenotype (Additional file 1: Figure S1C). These changes were in line with the phenotypic changes observed after treatment of iDC with lipopolysacharide (LPS), a well-established strong trigger of DC maturation (Additional file 1: Figure S1D). In line with this maturation marker profile, anti-EpCAM:CD40L induced significant production of the pro-inflammatory DC cytokine IL-12 after 3 days, whereas untreated iDC produced negligible levels of IL-12 (Additional file 1: Figure S1E).

Next, the anti-EpCAM:CD40L-matured DCs were analyzed for their capacity to induce T cell proliferation in allogeneic PBMCs. Hereto, allogeneic PBMCs were labeled with CFSE and mixed with DCs at the indicated ratios, whereupon T-cell proliferation in mixed cultures of iDC/PBMC and anti-EpCAM:CD40L matured DCs/PBMCs was analyzed after 7 days. In these mixed cultures, anti-EpCAM:CD40L matured DCs proved more potent at inducing T cell proliferation than iDC (exemplary proliferation plots in Figure 3F, quantified in Figure 3G), with a significant increase in the percentage of proliferating T cells (Figure 3H). Taken together, these data demonstrated that the soluble CD40L domain in scFv:CD40L fusion proteins remains biologically active and, in correspondence with its mode of action as a soluble molecule, requires high concentrations over prolonged periods of time to trigger efficient CD40 responses.

To evaluate whether target antigen-selective immobilization of scFv:CD40L would augment CD40 signaling, EpCAM^-CD40^- HT1080 cells and CD40-transfected HT1080 cells (HT1080.CD40) were co-cultured with EpCAM⁺/CD40^- DLD-1. Subsequently, induction of CD40 signaling was analyzed by measuring IL-8 production. Treatment of HT1080.CD40 with anti-EpCAM:CD40L triggered IL-8 at higher doses of 100–1000 ng/ml, in line with the activity profile of sCD40L (Figure 4A; left panel). However, in co-cultures of HT1080.CD40 with EpCAM⁺ DLD-1 cells, the treatment with anti-EpCAM:CD40L resulted in IL-8 secretion at significantly lower doses, corresponding to a ~20-fold down-shift in the ED50 value (Figure 4A; middle panel). In mixed cultures of parental HT1080 with DLD-1 cells, no IL-8 secretion was detected (Additional file 1: Figure S1F).

To subsequently assess the potentiating effect of targeted delivery of CD40L on DC maturation, monocyte-derived iDC were co-cultured with HEK293.EpCAM and overnight production of IL-12 was determined as read-out for iDC maturation. Monocultures of HEK293.EpCAM and iDCs treated with increasing concentrations of anti-EpCAM:CD40L did not produce IL-12 even at 1 μg/mL (Figure 4B: left panel). By contrast, in iDC/HEK.EpCAM co-cultures, treatment with anti-EpCAM:CD40L dose-dependently induced IL-12 production (Figure 4B: middle panel). IL-12 production was dependent on EpCAM-specific binding since pre-incubation with the parental EpCAM blocking antibody largely blocked IL-12 production (Figure 4B: right panel). In control mixed cultures of iDC with HEK.YFP and HEK.EpICD only marginal levels of IL-12 were detectable (Figure 4B: middle panel). Furthermore, in co-culture experiments of iDCs with a panel of EpCAM⁺ cancer cell lines, treatment with anti-EpCAM:CD40L induced significant IL-12 production that was inhibited by the EpCAM blocking antibody (Figure 4C). By contrast, no IL-12 was detected upon treatment of co-cultures of iDCs and EpCAM^- cancer cell lines with anti-EpCAM:CD40L (Figure 4C), nor when EpCAM⁺ cells were treated with an scFv:CD40L fusion protein of irrelevant specificity (Additional file 1: Figure S1G).

In line with this IL-12 production, treatment of co-cultures of iDC and HEK.EpCAM-YFP with anti-EpCAM:CD40L also triggered phagocytic uptake of high numbers of HEK.EpCAM cells (Figure 4D), whereas no such phagocytic uptake was detected in co-cultures with HEK or HEK.EpICD cells (Figure 4D). Next, we assessed whether these findings extended to DC-induced T cell proliferation. To this end, we co-cultured a panel of five EpCAM⁺ cancer cell lines with iDCs and allogeneic T cells in the presence or absence of anti-EpCAM:CD40L using a modified procedure of the proliferation assay described above. Importantly, in 4 out of 5 cells lines, treatment with anti-EpCAM:CD40L was associated with increased T cell proliferation (Figure 4E).

To assess whether these effects could be extended to leukemic cancer cells, we next investigated the anti-CD20:CD40L fusion protein. Similar to the effects observed for anti-EpCAM:CD40L, monocultures of iDCs treated with increasing concentrations of anti-CD20:CD40L only produced marginal levels of IL-12 (Figure 5A). By contrast, in co-cultures of iDC with the CD20⁺ cell line BJAB, treatment with anti-CD20:CD40L dose-dependently induced IL-12 production (Figure 5A). IL-12 production was dependent on CD20-binding since pre-incubation with the parental CD20 blocking antibody rituximab fully abrogated the increased IL-12 production (Figure 5A). In control mixed cultures of iDC with CD20^- Jurkat stimulated with anti-CD20:CD40L, only marginal levels of IL-12 were detected (Figure 5B). These findings were extended to the B cell leukemic lines Mino, Jeko and Daudi with BJAB serving as positive control (Figure 5C). Treatment of iDC:B cell co-cultures with either 30 or 300 ng/mL anti-CD20:CD40L induced dose-dependent production of IL-12 across all cell lines. As before, rituximab fully abrogated the induction of IL-12, indicating a CD20-dependent effect. Of note, treatment of B cell lines or iDC alone induced no or only marginal production of IL-12 in these experiments (Figure 5C).

In leukemic cells, signaling via CD40 has been linked to both pro-survival and anti-cancer effects [21, 22]. However, most of these studies have been performed using varying CD40-targeted agents (e.g. crosslinked CD40L or antibodies), complicating interpretation. Therefore, to assess the direct effects of anti-CD20:CD40L on leukemic B cells, BJAB and Raji were treated for 3 days with increasing concentrations of anti-CD20:CD40L and changes in cell surface expression of B cell maturation markers were assessed. In these experiments, the optimal dose for non-saturating induction of signaling was determined at 250 ng/mL (not shown). In both cell lines, treatment with anti-CD20:CD40L induced an increase in cell surface CCR7 (Additional file 1: Figure S1H and S1I), whereas no significant changes were observed in the cell surface levels of HLA-DR, CD80 or CD86 (Additional file 1: Figure S1H and S1I). Interestingly, treatment with anti-CD20:CD40L induced an upregulation of CD83 on BJAB cells and a downregulation of CD83 on Raji cells, indicating cell type-specific effects of CD40 signaling (Additional file 1: Figure S1H and S1I). As expected, treatment of leukemic B cell lines with control fusion protein anti-EpCAM:CD40L had no effect in this setting even when applied at concentrations up to 1 μg/mL (not shown). Finally, we did not observe proliferation or cell death as a result of treatment with anti-CD20:CD40L within this time frame (not shown).

We previously demonstrated a synergistic anti-cancer effect by an anti-CD20 fusion protein that relied on simultaneous induction of 1. death receptor-mediated apoptosis by FasL and 2. cell death via crosslinking of CD20 by the Rituximab scFv [19]. Therefore, we assessed whether targeted delivery of CD40L to CD20 by means of anti-CD20:CD40L could likewise induce CD20-mediated cell death. Treatment of leukemic B cell lines BJAB, Raji and Z138 induced a significant loss of viable cells after 5d, similar to that observed for treatment with rituximab (Figure 6A-C). Treatment of cells with anti-EpCAM:CD40L had no apparent effects at concentrations up to 1 μg/mL (Figure 6A-C). Of note, induction of cell death by rituximab occurred only in the presence of a secondary crosslinker, whereas anti-CD20:CD40L-induced cell death did not require oligomerization. These findings are in agreement with the data on anti-EpCAM:CD40L and suggest a model where binding of anti-CD20:CD40L to the cell surface induces CD40-clustering and concomitant crosslinking of CD20.

Expression vectors encoding the Flag-tagged CD40 ligand are derived from a variant of pCR3 (Invitrogen), which was provided by P. Schneider (University of Lausanne, Epalinges, Switzerland) and encodes an Ig signal peptide followed by a Flag tag. To obtain scFvRit:CD40L (anti-CD20:CD40L), amplicons of CD40L encoding amino acids 116–261 including a stop codon were first inserted in frame 5’ to the Flag tag by standard cloning techniques. Next, the leader encoding fragment of the Flag-ligand encoding plasmid was replaced by an amplicon encoding the leader sequence plus the scFvRit cDNA as previously reported [17]. scFvC54 (anti-EpCAM) was generated analogously using PCR from vector pEE14-scFvC54:TRAIL and inserted 5’ to the Flag tag as described above.

ScFvRit:CD40L (anti-CD20:CD40L) and scFvC54:CD40L (anti-EpCAM:CD40L) were produced in HEK293 cells by transient transfection of the expression vector. In brief, cells were electroporated (50 × 10⁶ cells/ml, 4-mm cuvette, 250 V, 1800 μF, maximum resistance) with 30 μg of DNA in 1 ml of culture medium containing 10% FCS using an Easyject Plus electroporator (PeqLab). Electroporated cells were recovered overnight in RPMI 1640 with 10% FCS (PAA Laboratories), and the next day the medium was replaced by low serum RPMI 1640 (0.5% FCS). After 3–4 days supernatants were collected and clarified by centrifugation. Fusion proteins were purified by affinity chromatography on anti-Flag M2 agarose beads (Sigma-Aldrich) and eluted with TBS containing 100 μg/ml Flag peptide (Sigma-Aldrich). The fractions containing the recombinant proteins were finally dialyzed against PBS, purity assessed by gel electrophoresis (Figure 1B) and stored at -20°C for further analysis.

OvCAR-3, HT-29, DLD-1, HCT-8, Sw480, Sw620, Caco-2, HT1080, HEK293, A375M, 786-O, BJAB, Jurkat, Daudi, Jeko and Mino were obtained from the American Tissue Culture Collection (ATCC) and characterized by short tandem repeat profiling, karyotyping and isoenzyme analysis. HT1080.CD40 cells were generated as described before. HEK293.YFP, HEK293.EpICD-YFP and HEK.EpCAM-YFP were a kind gift from Dr. O. Gires (University of München) [34]. Cells were cultured in RPMI 1640 medium supplemented with 10% fetal calf serum and 500 μg/ml Geneticin.

Monocytes and PBMCs were isolated from rests of blood buffy coats of fully anonymized donors obtained from the Institute of Clinical Transfusion Medicine and Hämotherapy of the University Hospital Würzburg and required no special written informed consent. PBMCs were isolated by density gradient centrifugation with lymphocyte separation medium (PAA Laboratories) [35]. Monocytes were separated from PBMCs by MACS magnetic bead separation with anti-CD14-coated beads (Miltenyi Biotec). Monocytes were seeded in 6-well plates (3 × 10⁶ cells) with RPMI 1640, 10% FCS, and penicillin/streptomycin. For DC differentiation 100 ng/ml GM-CSF and 20 ng/ml IL-4 (ImmunoTools) were added every second day. On day 7, cells were stimulated as indicated. On day 10, functional analyses were performed as indicated.

To determine cell surface expression of the indicated markers by flow cytometry, cells were incubated on ice with anti-CD14-FITC, anti-CD80-PE, anti-CD83-FITC, anti-CD86-APC, anti-CCR7-APC, anti-HLA-DR-PE and/or indicated combinations thereof or the appropriate isotype control Abs (all from Becton Dickinson, Amsterdam, The Netherlands). After three washes with ice-cold PBS, cell-associated immunofluorescence was determined by FACS analysis. For binding studies with the scFv:CD40L constructs, indicated cell lines were incubated on ice for 1 h with scFv:CD40L fusion proteins, washed three times with ice-cold PBS and stained with anti-CD40L-PE antibody (Becton Dickinson, Amsterdam, The Netherlands) on ice for another 1 h. Subsequently, cells were washed three more times with ice-cold PBS and used for flow cytometry. For blocking studies, cells were pre-incubated with control antibodies (Rituximab or MOC31-Fc) for 1 h on ice prior to incubation with fusion protein.

For immunofluorescence, cells were washed and stained with the indicated PE-conjugated antibodies (anti-CD40-PE, anti-CD80-PE, anti-CD83-PE, anti-CD86-PE and anti-HLA-DR-PE all from Becton Dickinson, Amsterdam, The Netherlands) on ice. After three washes with ice-cold PBS, cell-associated immunofluorescence was determined by fluorescent microscopy. All fluorescent microscopy was performed on an EVOS fl Digital Fluorescence Microscope.

Cancer cells were seeded (2 × 10⁴ cells/well) in 96-well tissue cultures plates in 100 μl of RPMI 1640 medium with 10% FCS and grown overnight. On the next day, medium was exchanged to minimize the background of constitutive cytokine production, and cells were stimulated for 16 h with the indicated concentrations of fusion protein ± blocking antibody in the presence or absence of indicated HT1080 transfectants or Dendritic cells. Supernatants were analyzed for production of IL-8 or IL-12 using the BD human IL-8 ELISA or IL-12/IL-23 Duoset ELISA (BD Biosciences) according to the manufacturer’s instructions.

Dendritic cells (DCs) matured as indicated (or relevant controls) were harvested, washed three times and resuspended at a density of 1×10⁷ cells/mL. DCs were subsequently incubated with 50 μg/mL Mitomycin C solution in PBS (freshly prepared and filter sterilized at 0.5 mg/mL) in the dark for 20 min. at 37°C. After incubation, Mitomycin C was rapidly diluted by addition of an excess of warm culture medium and cells were washed three more times by centrifugation. Cells were resuspended at a density of 10×10⁶ cells/mL and plated at the indicated densities/ratios in round-bottom 96-well plates in the presence of 1×10⁵ T cells/well. DC-T cell co-cultures were briefly centrifuged (1 min, 100 g) to promote cell-cell contact, incubated for 6 days at 37°C in a humidified 5% CO₂ incubator and T cell proliferation assessed by CFSE dilution on a flow cytometer. For CFSE-labeling, the Vybrant^® CFDA SE Cell Tracer Kit (Invitrogen) was used according to the manufacturer’s instructions. In all experiments, T cells autologous to the Dendritic cell population were used as control and never displayed >5% proliferation. For co-cultures of dendritic cells and tumor cells this procedure was slightly modified. First, the ratios used in co-culture of immature dendritic cells and tumor cells were optimized per cell line to prevent cancer cells from overgrowing. Second, the percentage of viable dendritic cells within the mixed population after 3 day maturation was assessed using CD86/CD83/PI staining on flow cytometry and cell counts adjusted accordingly to a PBMC/DC ratio of 32. Of note, mitomycin C treatment did not fully abrogate cancer cell growth during the 6 day MLR, but slowly proliferating cancer cells did not affect T cell viability within this time frame.

HEK293.YFP, HEK293.EpICD-YFP or HEK293.EpCAM-YFP cells were harvested and kept in suspension. Cells were added at a 1:1 ratio to cultures of iDC in the presence or absence of anti-EpCAM:CD40L for 2 h, washed three times with ice cold PBS and stained with anti-CD80 and anti-CD86 antibodies as described in the section on flow cytometry. Cells were gated exclusively on the single cell population by FSC/SSC analysis to exclude DC:HEK cell doublets. DCs were identified based on co-expression of CD80 and CD86 expression and the increase in YFP signal assessed in the FL1 channel by means of the indicated histograms. Percentages were determined based on the amount of YFP⁺ cells within the total CD80/CD86 population.

Primary B cells were cultured for up to 5d as indicated in 6-well plates (1×10⁶ cells/well; 3 mL/well) in the presence or absence of the indicated concentrations of anti-CD20:CD40L and harvested for analysis. To determine specific maturation, binding of anti-CD20:CD40L was blocked by pre- and subsequently co-incubating cells with 10 μg/mL of parental antibody rituximab. Where indicated, B cells were also treated with (Fab-fragment crosslinked) Rituximab, an irrelevant IgG1 control, or anti-EpCAM:CD40L, all at a final concentration of 1 μg/mL. B cell maturation was assessed by flow cytometry as described above. Cell death in B cell lines was evaluated by loss of mitochondrial membrane potential (ΔΨ) using DioC6 according to the manufacturer’s instructions (Molecular Probes).