Introduction

Dendritic cells (DCs) are the most potent antigen-presenting cells. In their steady state, DCs in the skin are in an immature state and have great phagocytic capacity. After antigen take-up, they undergo changes (maturation) including down-regulation of antigen take-up and processing capacity, and migrate to regional lymph nodes (LN) where they present the antigen to naïve T cells, thereby generating potent immunity (Wright-Browne et al. 1997). It has been reported that efficient anti-tumor immunity is generated by immature DCs because DCs phagocytose both necrotic and apoptotic tumor cells and mature in vivo (Chen et al. 2001). These results suggest that DCs can be used efficiently in cancer immunotherapy.

DC-based immunotherapies have already been tried in mice as well as humans and positive results were obtained against many tumors, especially malignant melanoma, B cell lymphoma, colorectal cancer, and prostate cancer (Nestle et al. 1998; Timmerman et al. 2002; Morse et al. 1999; Ragde et al. 2004). Although promising results have been obtained in many studies, several important issues still need to be clarified, such as the source, injection sites, optimal number of DCs, method of maturation, and frequency of vaccination (Cranmer et al. 2004). For example, the impact of DC immunization via different routes on either immune responses or anti-tumor efficacy is diverse. DCs administered intranodally were superior for the induction of anti-tumor immunity when compared to intravenously (i.v.) or subcutaneously (s.c.)-injected DCs in the lung metastases model (Lambert et al. 2001), whereas s.c.-administered DCs induced a higher immune response when compared with i.v.-injected DCs in murine tumor models (Eggert et al. 1999; Okada et al. 2001).

In this study, we have assessed the therapeutic efficacy of increasing the dosage of immature DCs against a pre-established solid tumor model (melanoma) in mice. In addition, we analyzed the relationship between the therapeutic and immunological results.

Materials and methods

Animals

C57BL/6 female mice (6–8 weeks old) were purchased from Polas International (Seoul, Korea) and housed in pathogen-free units at the Yonsei Medical Research Center.

Media and cytokines

Complete medium (CM) consisted of RPMI 1640 (Gibco, BRL, Grand Island, NY) supplemented with 10% (v/v) heat-inactivated FBS, 100 IU penicillin/ml, 100 μg streptomycin/ml, 0.1 mM non-essential amino acids (Sigma, Saint Louis, MO), 1 mM sodium pyruvate, 10 mM HEPES, and 50 μM 2-mercaptoethanol. Recombinant mouse GM-CSF and IL-4 were purchased from Endogen (Woburn, MA).

Preparation of bone marrow-derived DCs

Bone marrow cells were obtained from the tibias and femurs of C57BL/6 mice (6–8 weeks old) and depleted of erythrocytes with red blood cell lysis buffer (Sigma). Cells were then treated with a mixture of antibodies (Abs) [anti-CD4 (GK1.5), anti-CD8 (53-6.72), anti-B220 (RA3-3A1/6.1), anti-I-Ab,d,q & I-Ed,k (M5/114.15.2), anti-erythrocytes, neutrophils & B cells (J11d.2, TIB-183); all obtained from ATCC] for 30 min at 4oC, followed by a rabbit complement (Cedarlane, Ont., Canada) for 1 h at 37oC. Cells were then layered onto lympholyte M (density: 1.0875 ± 0.001 g/ml, Cedarlane) gradients, centrifuged, and cells at the low-density interface were collected. Cells were washed three times with CM and incubated in CM supplemented with 10 ng GM-CSF/ml and 10 ng IL-4/ml in 24 well plates at 7–10 × 105 cells/well. On day two (D2) of culture, floating cells were gently removed, and fresh medium containing GM-CSF and IL-4 was added. On day four (D4), floating cells were harvested and fresh medium containing GM-CSF and IL-4 was added. The harvested cells were plated in new 24 well plates in CM supplemented with GM-CSF and IL-4 at 5 × 105/ml. On day six (D6), non-adherent and loosely adherent DCs were collected.

Flow cytometry

To identify expression of surface molecules on D6 DCs, cells (6–10 × 105) were washed twice with 0.4% (w/v) BSA/PBS and stained for 30 min at 4oC with monoclonal Abs against CD11c, CD80, CD86, and MHC class II (I-Ab,d,q & I-Ed,k). Rat IgG2a were used as isotype controls. All Abs were purchased from Pharmingen (San Diego, CA). After two washes with 0.4% BSA/PBS, secondary staining was performed using FITC-conjugated F(ab′)2 goat anti-rat Ig (Biosource, Camarillo, CA) and FITC-conjugated anti-hamster IgG. After 30 min, cells were washed twice with 0.4% BSA/PBS and resuspended in 400 μl of 0.4% BSA/PBS. Propidium iodide (Sigma) was added to exclude dead cells from analysis. Flow cytometric analysis was performed on a FACSCalibur (Becton Dickinson, Mountain View, CA).

Antigen pulsing and intratumoral injection of DCs

Six days after harvesting, the DCs were incubated with B16F10 melanoma lysates (freeze–thawed three times, centrifuged at 3,000 g for 10 min, supernatant collected) at 150 μg/ml for 18 h. After an 18 h incubation, DCs were harvested, washed twice, and resuspended in PBS (30 μl) for injection. After the first intratumoral injection, another two intratumoral injections of DCs were performed at 2 day-intervals.

Experimental design

Malignant melanoma was made by inoculating 5 × 105 B16F10 cells in 100 μl PBS subcutaneously into the backs of C57BL/6 mice. Six days after the inoculation, when the tumor size was about 0.5 cm across the major axis, 20 mice were divided into four groups: (1) treated with PBS (Group I), (2) treated with 1 × 105 DCs (Group II), (3) treated with 7 × 105 DCs (Group III), and (4) treated with 21 × 105 DCs (Group IV). In the survival experiment, 40 mice were divided into the same four groups.

Measurement of tumor size and survival rate

After injection of the DCs, the lengths of the short and long axes of the tumors were measured with calipers (Mitutoyo, Japan) at two- to three-day intervals, and tumor volumes (V) were calculated by the formula: V = 1/2 × A 2 × B, where A is the length of the short axis and B that of the long axis. Survival rates were calculated by dividing the number of surviving mice by the total number of mice.

Delayed type hypersensitivity assay

Seven days after the third DC injection, each mouse received 500 μg of tumor lysates (suspended in 40 μl of PBS) subcutaneously injected into the right footpad area. As a negative control, an identical amount of PBS was also injected into the left footpad of the same mouse. After 48 h, the extent of swelling was measured using calipers (Mitutoyo). Results were reported as the differences in thickness (in millimeters) between both footpads.

ELISA

Splenocytes were obtained from each experimental group 1 week after the last DC injection, and commercial erythrocyte lysis buffer was added. After 5 min, the reaction was stopped with CM, and the cells were washed twice with CM. Triplicate samples of 1.5 × 105 cells in CM were seeded in 96 well round-bottom plates (Corning, USA). The splenocytes (responder, R) were stimulated with B16F10 cells (stimulator, S) treated with 1% (w/v) paraformaldehyde at a S:R ratio of 1:100. Splenocytes from group I mice were used as a negative control. The cells were incubated at 37oC for 4 days. The amount of secreted IFN-γ, IL-12p70, and IL-4 was measured by ELISA, which was performed as described in the OptEIA (Pharmingen) catalog. OD450 was measured on an ELISA Reader (Molecular Devices, Sunnyvale, CA). Data are represented as the average of duplicate samples ±SE.

IFN-γ ELISPOT assay

Erythrocyte-depleted splenocytes were obtained as described above. Ninety-six well PVDF plates were coated with anti-mouse IFN-γ and incubated overnight at 4°C. After washing, plates were blocked with CM for 2 h at room temperature. After removal, splenocytes were plated in the wells at 1 × 105/100 μl, and B16F10 cells treated with 1% (w/v) paraformaldehyde were added at 1 × 104/100 μl. After 24 h of incubation at 37°C, the cells were removed. Spot-forming colonies were performed according to the ELISPOT kit protocol (Pharmingen) and were counted with an ELISPOT reader system (AID, Germany). To identify CD8+ T cells, CD4+ T cells were depleted with anti-CD4 Ab (H129.19).

Statistical analysis

Statistical analyses were performed using SPSS (SPSS Inc., USA) software. To compare tumor sizes in the main experiment, an ANOVA was applied with the Sheffé method. The Kruskal–Wallis test, a non-parametric equivalent to ANOVA, was used to compare DTH, ELISA, and ELISPOT assay results. P-values <0.05 were considered statistically significant.

Results and discussion

Changes in DC surface molecules

Erythrocyte-, lymphocyte-, and neutrophil-depleted bone marrow cells of C57BL/6 mice were incubated with rmGM-CSF and rmIL-4 for 6 days. Upon FACS analysis, high expression levels of costimulatory and MHC class II molecules were observed, which indicated that the DCs were generated from bone marrow cells (Fig. 1). The expression of CD40, CD80, CD86, and MHC class II molecules did not increase further after pulsing the DCs with B16F10 melanoma lysates (data not shown).

Fig. 1
figure 1

Flow cytometric analyses of mouse bone marrow-derived DCs. On day six of culture, the surface phenotypes of bone marrow-derived DCs were examined by flow cytometry using CD40, CD80, CD86, MHC class II, and CD11c mAb. Similar results were obtained from three repeated experiments. The data shown are from one representative experiment

Full size image

Decreased tumor sizes and increased survival with intratumoral DC injection

DCs were first injected when the long axes of the tumors were about 0.5 cm (day 0), and tumor sizes were then measured in two- to three-day intervals (Fig. 2A). In total, three cycles of DCs were injected at two-day intervals. Reduced tumor growth was observed in groups III and IV after day 13 compared to the other groups (I and II). Twenty days after the first DC injection, the mean tumor sizes of groups I, II, III, and IV were 2,402 ± 293 mm3, 2,178 ± 414 mm3, 1,305 ± 174 mm3, and 986 ± 164 mm3 (approximately 41% tumor growth inhibition), respectively (mean ± standard error). The reduced growth observed in groups III and IV was significant when compared to groups I and II (P < 0.05), but no statistically significant difference was found between groups III and IV (P = 0.095). These results suggest that an increasing dosage of DCs reduces tumor growth, though a limitation in improving the therapeutic effect with increasing dosage may exist. However, the Pearson correlation coefficient was −0.90, indicating a definitive negative correlation between DC dosage and tumor size.

Fig. 2
figure 2

Tumor volume and survival rate changes after various treatments. Melanomas were generated by subcutaneously inoculating B16F10 melanoma cells (5 × 105 cells) into the backs of C57BL/6 mice. Six days after B16F10 inoculation, dendritic cells were injected intratumorally. (A) Mean tumor volumes were much reduced in group IV 20 days after inoculation. The data shown are mean values with their associated standard errors. (B) Survival rates of mice in each treatment group. Survival rates are presented as the percentage of animals surviving

Full size image

Survival rates were calculated as the number of surviving mice divided by the total number of mice (Fig. 2B). The survival rates of group I and group II rapidly decreased after day 20, and all mice died before day 33 and 38, respectively. The survival period of group III and group IV were 42 and 54 days, respectively. Treatment with larger numbers of DCs prolonged survival of the mice when compared with group I. These results indicate that anti-tumoral immunity was increased with a higher DC dosage.

Generation of tumor-specific immunity identified by in vivo and in vitro assays

Tumor lysates were injected into the footpads of mice to demonstrate whether tumor-specific immunity was generated or not, and the extent of footpad swelling was measured (Fig. 3A). The DTH response in groups III and IV was induced 15-fold and 26-fold higher than in group I, respectively. Although the tumor size and survival rate was similar to group I, the DTH response of group II was induced eight-fold more than in group I. Moreover, the DTH response in group IV was more severe than that in group III, and the Pearson correlation coefficient was +0.95, indicating there was a strong correlation between DC dosage and the DTH responses induced.

Fig. 3
figure 3

Induction of tumor-specific immunity after various treatments. (A) Delayed type hypersensitivity. A total of three cycles of Ag-pulsed DCs were injected intratumorally into C57BL/6 mice at two-day intervals. Six days after the last DC injection, 500 ug of Ag (in 40 μl PBS) and PBS were injected into the right and left footpads, respectively. Data presented are the mean values of three repeated experiments with error bars. (B) 1, IFN-γ; 2, IL-12p70; and 3, IL-4 secretions measured by ELISA. The cytokine contents of splenocyte supernatants after 4 days of stimulation were determined by ELISA. Data presented are the mean values of three repeated experiments with error bars. (C) IFN-γ-secreting cells measured by ELISPOT assays. CD4+ T cells were depleted from splenocytes with anti-CD4 Ab (H129.19). Data presented are the mean values of spots per 1 × 105 responders with error bars from three repeated experiments that were conducted in duplicate

Full size image

IFN-γ ELISA was also performed with the supernatants of splenocytes stimulated with fixed B16F10 melanoma cells. As shown Fig. 3B, the concentrations of IFN-γ were 1,596 ± 391 pg/ml, 3,751 ± 393 pg/ml, and 7,842 ± 1,112 pg/ml in Group II, III, and IV, respectively (mean ± standard error). No IFN-γ was detected in group I. The Pearson correlation coefficient was +0.98. The co-culture with B16F10 melanoma cells significantly induced IL-12p70 release from the spleen cells of mice treated with DCs (Groups III and IV, P < 0.03 compared with Group I). The spleen cells from DC-treated mice (Group IV) also showed increased production of IL-4 (105 ± 10 pg/ml) in co-culture with B16F10 cells, but the actual amount was much lower than IFN-γ (7,842 ± 1,112 pg/ml). These results show that DCs induce Th1-dominant immune response.

IFN-γ ELISPOT assays were performed to identify melanoma-specific T cells in each experimental group. After depleting CD4+ T cells with anti-CD4 Ab to analyze CD8+ T cells, splenocytes of each group were stimulated with fixed B16F10 melanoma cells for 24 h. The number of IFN-γ-secreting cells were much greater in group IV (Fig. 3C), and the Pearson correlation coefficient was +0.98. These findings indicate that anti-melanoma immunity was significantly enhanced with an increased dosage of DCs used for treatment.

In summary, using of more tumor antigen-loaded DCs yielded better therapeutic results in our experiments, and the results obtained by DTH, ELISA, and ELISPOT assays were consistent with our therapeutic outcome.