Differential immune responses mediated by adenovirus- and lentivirus-transduced DCs in a HER-2/neu overexpressing tumor model

Abstract

Recent investigations have demonstrated that adenoviral and lentiviral vectors encoding HER-2 can be utilized in cancer immunotherapy. However, it is not known whether both viral systems elicit a similar immune response. Here, we compare the immune response in mice induced by dendritic cells (DCs) infected with either recombinant adenovirus or lentivirus encoding rat HER-2 (rHER-2). Both vaccine types yielded similar control of tumor growth, but we found clear differences in their immune responses 10 days after DC immunization. Adenovirus rHER-2-transduced DCs elicited locally and systemically high frequencies of CD4+ and CD8+ T cells, while lentivirus rHER-2-transduced DCs predominantly led to CD4+ T-cell infiltration at the tumor site. Splenocytes from mice immunized with lentivirus rHER-2-transduced DCs secreted higher levels of interferon (IFN)-γ, mainly by CD4+ T cells, following stimulation by RM-1-mHER-2 tumors. In contrast, the adenovirus vaccinated group exhibited CD4+ and CD8+ T cells that both contributed to IFN-γ production. Besides an established cellular immune response, the rHER-2/DC vaccine elicited a significant humoral response that was highest in the adenovirus group. DC subsets and regulatory T cells in the spleen were also differentially modulated in the two vaccine systems. Finally, adoptive transfer of splenocytes from both groups of immunized mice strongly inhibited in vivo tumor growth. Our results suggest that not only the target antigen but also the virus system may determine the nature and magnitude of antitumor immunity by DC vaccination.

INTRODUCTION

The treatment of tumors that overexpress HER-2/neu (HER-2) remains a challenge in many therapeutic contexts in the clinic. HER-2 is a transmembrane protein receptor and a potent oncoprotein implicated in cancer initiation and progression. It has been shown that HER-2 is immunogenic as observed by the presence of antibodies and cellular immune responses.^1,2 However, tumor resistance to Trastuzumab, an anti-HER-2 monoclonal antibody currently used in the clinic, shows that more effective treatments are needed for HER-2-expressing tumors.³

Gene transfer vectors are promising tools for the prevention and treatment of cancer.^4,5 Viral vectors provide an efficient way to modify and incorporate exogenous genes, or to perturb the expression of endogenous genes. Adenovirus vectors (AdVs) and lentivirus vectors (LVs) are among the most popular systems for gene delivery. LVs can stably transduce non-dividing cells with long-term expression of the transgene through proviral integration into the target cell genome.^6,7 Another advantage of the use of LVs is the possibility of modifying the virus’ envelope protein, permitting a broader tropism. Moreover, LVs are not inherently immunogenic, and do not elicit immune or inflammatory responses in hosts unless given at very high doses.⁸ We have previously shown that injected LVs can circulate in mice in vivo without eliciting a response against the virus itself.⁹ Unlike LVs, AdVs do not integrate into the host genome at an appreciable frequency, eliminating the possibility of insertional mutagenesis, and thus are intrinsically safer from an oncogenic perspective. However, AdVs have been shown to induce strong auxiliary immune responses in its hosts.

Direct infection with AdVs or LVs that engineer tumor antigen expression is one way to employ these vectors as cancer vaccines. However, dendritic cells (DCs) are considered the most clinically effective vaccine platforms. Genetically modified DCs have been shown to be more efficient in the induction of a specific antitumor memory T-cell responses relative to peptide loaded DCs.¹⁰ Viral transduction is a useful platform to load DCs with tumor antigens that has advantages over direct virus injections including the ability to deliver repeated vaccinations because virus-specific neutralizing antibodies are avoided and better activation of both tumoricidal CD4+ T cells and NK cells occurs.^11-13

Although it is known that AdVs and LVs can effectively introduce genes into DCs, there is a paucity of data directly comparing the immunogenicity and systemic consequences of employing these two major vector systems in side-by-side experiments. In this study, we used a murine tumor model to compare the immune response induced by DCs transduced with AdVs and LVs encoding a xenoantigenic form of rHER-2, in order to more effectively break the tolerance to self-antigens.^11,14 Such a head-to-head comparison points to mechanistic differences and can lead to protocol designs that may favor certain responses over others.

RESULTS

Adenovirus and lentivirus infections do not affect the differentiation and maturation of DCs in vitro

In this study, we engineered both AdVs and LVs encoding the full-length rHER-2 complementary DNA. Transgene expression was driven by the murine cytomegalovirus promoter in the AdV vector (Figure 1), while in the LV, expression was regulated by the elongation factor-1α promoter (Figure 1). Corresponding vectors engineering expression of enhanced green fluorescent protein (enGFP) were used as controls.

Figure 1.

Schema of AdV and LV constructs encoding rHER-2 and enGFP complementary DNAs (cDNAs). The full-length cDNA for rat HER-2/neu was used to construct the recombinant adenovirus and LVs. For safety, a point mutation was introduced into HER-2.1 inactivating the kinase activity of neu. The adenovirus consists of murine cytomegalovirus (MCMV) promoter and a SV40 polyA signal. The enGFP adenovector also has a MSIINFEKL peptide. Lentivirus contains LTR, long terminal repeat; ψ, packaging signal; RRE, rev-responsive element; cPPT, central polypurine tract; 3′SIN, self-inactivating LTR.

Owing to the transient nature of AdV-induced transgene expression, bone marrow-derived DCs were infected on day 5 of in vitro culture and injected into recipient animals on day 8. However, LV transductions were performed on day 3 to maximize transduction efficiency and also injected on day 8. Transduced DCs will herein be referred to as rHER-2/DCs and enGFP/DCs, respectively. We determined a similar level of rHER-2 expression on DCs empirically using an multiplicity of infection (MOI) of 100 for AdVs and MOI of 15–20 for LVs. Analysis of rHER-2 expression on the day of vaccination illustrates an infection rate of 50.6% (±4.8%) and 46.9% (±7.4%) for AdV and LV, respectively, using these conditions. Representative data are shown in Figure 2a. Similar results were obtained for enGFP expression; 52.6% (±7.3%) of AdV-transduced DCs were positive for enGFP and 51.8% (±6.4%) of LV-transduced DCs were positive for enGFP expression (Figure 2a).

Figure 2.

DC surface phenotype and rHER-2 expression following AdV- and LV-mediated rHER-2 gene transfer. DCs were generated from murine bone marrow in the presence of granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin (IL)-4, and matured on day 7 by tumor necrosis factor (TNF)-α stimulation for 24h. Cells were infected on day 3 with LV/rHER-2 (MOI of 15-20) or on day 5 with AdV/rHER-2 (MOI of 100). Infections with vectors encoding enGFP (that is, LV/enGFP and AdV/enGFP) were used as controls. (a) Expression of rHER-2 and enGFP on day 8 following AdV and LV infection; (b) Immature or (c) TNF-α-matured DCs stained with PE anti-CD11c, FITC anti-CD80, FITC anti-CD86 and FITC anti-major histocompatibility complex (MHC)-II antibodies or isotype controls on day 8 of in vitro culture and analyzed by flow cytometry. Results are representative of three independent experiments. AdV (solid black line), LV (dashed black line), non-transduced (filled gray) and isotype control (dashed gray line).

Next, we analyzed whether the AdV and LV transductions would affect the in vitro differentiation and maturation of the DCs. DCs were analyzed before (day 7) and after (day 8) maturation with tumor necrosis factor-α. rHER-2/DCs were typical as noted by their morphology, dendrites (data not shown), and the expression of differential marker CD11c (Figure 2b). In addition, we observed that neither infection induced the maturation of DCs. rHER-2/DCs maintained an immature phenotype similar to the non-transduced DCs (NT/DCs) with high expression of CD11c, low expression of major histocompatibility complex II, and low expression of costimulatory molecules CD80 and CD86 (Figure 2b). Maturation was achieved by tumor necrosis factor-α addition; the DCs demonstrated upregulation of major histocompatibility complex II, CD80 and CD86 on their cell surface (Figure 2c). A similar effect was observed on the enGFP/DCs (data not shown).

Vaccination with both adenovirus- and lentivirus-transduced rHER-2/DCs induces a strong and similar inhibition of tumor growth in mice

To investigate possible differences in the gross induction of antitumor immune responses between rHER-2/DCs transduced by AdV and LV, we employed an aggressive RM-1 tumor model overexpressing murine HER-2 (RM-1-mHER-2) as previously described by our group.¹³ We chose a small dose of genetically modified DCs to evaluate the intensity of immune response induced by AdV and LV. Groups of C57BL/6 mice were injected subcutaneously with RM-1-mHER-2 tumor cells and the tumors were allowed to grow to a moderate size (~20mm³) for 4 days before the vaccination. At day 4 post-tumor implantation, the mice were injected in their footpad with 2×10⁵ rHER-2/DCs transduced either with AdV or LV. NT/DCs and enGFP/DCs were used as controls.

As can be seen in Figure 3, treatment with both the AdV- and LV-rHER-2/DCs markedly reduced the tumor growth for 2 weeks when compared with the NT/DC- and enGFP/DC-treated groups in three separate experiments. In contrast, tumors from the enGFP/DC-treated groups progressed from day 8 post-tumor injection; by day 14 all mice in these groups developed large (~1400 mm³) and/or ulcerated tumors. Under the same conditions, vaccination with NT/DCs produced similar uncontrolled tumor growth. Interestingly, no significant differences in tumor volume were observed between mice treated with AdV- or LV-transduced DCs.

Figure 3.

Efficacy of AdV- and LV-rHER-2/DC vaccines in tumor-bearing mice. C57BL/6 mice were inoculated with RM-1-mHER-2 tumor cells (2×10⁵) subcutaneously (s.c.) on day 0. After 4 days, when the tumors were visible, the mice were immunized with AdV- or LV-rHER-2/DCs and enGFP/DCs or NT/DCs. DCs were injected into the footpad at the dose of 2×10⁵ cells per mouse. The arrow indicates the time point that the mice were immunized. The data represent the mean (±s.e.m.) of 15 measurements from three separate independent experiments, *P<0.04.

Immunotherapy with AdV-rHER-2/DCs rather than LV-rHER-2/DCs results in a higher infiltration of CD4+ and CD8+ T cells

Given that AdV- and LV-rHER-2/DCs induced a similar inhibition of tumor development, we sought to determine whether this correlated with similar expansion of CD4+ and CD8+ T-cell subsets. To address this question, splenocytes obtained from vaccinated mice at day 10 post-DC immunization were stained with anti-CD4 and anti-CD8 antibodies and analyzed by flow cytometry. As expected, the frequencies of CD4+ and CD8+ T cells in the spleen were higher in the rHER-2/DC-treated group than in the NT/DC- and enGFP/DC-treated groups (Figure 4a). However, the most pronounced difference between rHER-2/DC- and enGFP/DC-treated groups was observed in the AdV-infected DCs, in which the differences were 10.7% for CD4+ and 6.5% for CD8+ splenocytes. The increase in the CD4+ and CD8+ T-cell frequencies between LV-rHER-2/DCs and enGFP/DCs were 6.7 and 3.6%, respectively. No differences were noted among the control groups. Interestingly, mice treated with rHER-2/DCs from the AdV transductions had significantly higher expansion of CD4+ and CD8+ T cells in the spleen than the mice that were vaccinated with LV-rHER-2/DCs. The percentages of CD4+ and CD8+ splenocytes in the mice immunized with AdV-rHER-2/DCs vaccine were 25.7% (±1.4%) and 15.3% (±0.8%) respectively, while for LV-rHER-2/DCs the levels were 20.9% (±0.9%) for CD4+ cells and 12.9% (±0.5%) for CD8+ T cells (Figure 4a).

Figure 4.

Profile of cellular immune responses induced by AdV- and LV-rHER-2/DCs. Tumor-bearing mice vaccinated with either AdV- or LV-rHER-2/DCs, or with the controls DCs (AdV- or LV-enGFP/DCs, and NT/DCs) were killed at day 10 after vaccination. (a) Splenocytes were obtained and immunostained with FITC anti-CD4 and PE anti-CD8 antibodies. The frequency of CD4+ and CD8+ T-cell subsets were determined by flow cytometry. The results represent a mean (±s.e.m.) of 15 samples per group from three independent experiments (CD4, *P<0.04; CD8, *P<0.04). (b, c) Tumor tissues were also obtained and processed for immunohistochemical analyses. Tissue sections were incubated with anti-CD4 and anti-CD8 antibodies followed by Alexa Fluor-488 anti-IgG secondary antibody and 4,6-diamidino-2-phenylindole (DAPI). Three to five areas of each sample (total of five samples per group) were analyzed using the ImageJ software. The results are shown as mean (±s.e.m.). (CD4, *P<0.02, ***P<0.001; CD8, *P<0.02). (d) DCs present in the spleen were stained with anti-CD11c, anti-CD11b, anti-B220, anti-Gr-1 and anti-CD8 antibodies and phenotypically characterized as myeloid (CD11c+CD11b+B220−), plasmacytoid (CD11c+B220+Gr-1+), and lymphoid (CD11c+CD11b−B220−CD8+). For staining of regulatory T cells, splenocytes were incubated with anti-CD4, anti-CD25 and anti-Foxp3 antibodies. Splenocytes were also stained with an anti-NK1.1 antibody. The results are shown as absolute number of cells per spleen. *P<0.0199, **P<0.0076.

To further assess whether the observed differences in the spleens between animals administered AdV- and LV-rHER-2/DCs was also reflected at the tumor site, we performed immunohistochemical analysis of the tumors 10 days after DC injection. Sections of tumors were prepared and stained with anti-CD4 and anti-CD8 antibodies and the presence of positively stained cells was evaluated by microscopy. Tumors examined from mice vaccinated with rHER-2/DCs, either transduced with AdVs or LVs, were strongly infiltrated by CD4+ T cells, while the control groups exhibited a small number of positive cells (Figure 4b). LV-rHER-2/DC therapy induced sixfold (P<0.0010) more CD4+ T-cell infiltration than the control group, whereas AdV-rHER-2/DCs had a threefold increase (P<0.0195) (Figure 4c). The difference in this measurement was not significant between AdV- and LV-DCs. However, the accumulation of CD8+ T cells within the tumor was approximately twofold (P<0.0290) higher in mice that received the AdV-rHER-2/DCs than in animals treated with the DCs transduced with LVs (Figure 4c). An increased number of CD8+ T cells was also found in the AdV-rHER-2/DCs vaccinated mice in comparison with their counterpart, enGFP/DC control group (Figure 4c). In contrast, LV-rHER-2/DCs did not induce a significant infiltration of CD8+ T cells at the tumor site.

The role of DCs driving CD4+ and CD8+ T cells is well known. Owing to our initial finding that AdV- and LV-rHER-2/DCs led to a differential balance of CD4+ and CD8+ T cells, we decided to investigate the DC subsets in the spleen at 10 days after vaccination. DCs were characterized as myeloid (CD11c+CD11b+B220−), plasmacytoid (CD11c+Gr-1+B220+) and lymphoid (CD11c+CD11b−B220−CD8+). Although the absolute number of DC subsets was similar between both therapeutic-DC vaccines, there were differences in the type of DC subsets between the groups when compared with their respective controls. For instance, mice treated with AdV-rHER-2/DCs had significantly more myeloid (~1.6-fold, P<0.0076) and plasmacytoid (~2.6-fold, P<0.0101) DCs than the AdV-enGFP/DC control group, whereas myeloid (~1.9-fold, P<0.0199) DCs were the most significant DC population in the LV-rHER-2/DC group (Figure 4d).

Regulatory T cells with the CD4+CD25+Foxp3+ phenotype were also evaluated after vaccination. The LV-rHER-2/DC group displayed an ~twofold (P<0.0065) decrease in CD4+CD25+ Treg cell numbers compared to the LV-enGFP/DCs group (Figure 4d). Although the AdV-rHER-2/DC vaccinated mice had significantly lower frequency of CD4+CD25+ Treg cells in comparison with the AdV-enGFP/DCs group (P<0.0074, data not shown), the absolute number of cells was similar between both groups (Figure 4d). Furthermore, we did not find any differences in the frequency or absolute number of Tregs at that time point between AdV and LV-DCs immunized mice. Finally, there were no differences observed in the NK1.1+ cell population present in the spleens of mice treated with rHER-2/DCs and that of control DCs (Figure 4d).

Splenocytes from LV-rHER-2/DC-vaccinated mice strongly respond to RM-1-mHER-2 tumor cells with increased secretion of interferon (IFN)-γ by CD4+ T cells

Having observed significant differences in the frequency of T-cell infiltration in the spleen and tumor site between animals treated with AdV- and LV-transduced DCs, we were interested in determining the functionality of these splenocytes in vitro. To test this, we evaluated spontaneous cytokine production by total splenocytes from vaccinated mice. As expected, rHER-2/DC therapy induced significantly higher IFN-γ expression from splenocytes in comparison with the control groups (Figure 5a). This finding is consistent with the high frequency of CD4+ and CD8+ T cells in the spleen that probably induced tumor growth inhibition. Comparing both AdV- and LV-rHER-2/DC treatments, the spontaneous secretion of IFN-γ was statistically higher in the splenocyte culture from mice receiving AdV-DCs. Indeed, the cytokine concentration in the supernatant from cells derived from mice receiving AdV-rHER-2/DCs was 15.9 ng ml⁻¹ (±2.3 ng ml⁻¹), compared with 9.5 ng ml⁻¹ (±0.8 ng ml⁻¹) for LV-rHER-2/DCs (Figure 5a).

Figure 5.

Differences in IFN-γ production by CD4+ and CD8+ splenocytes from mice immunized with AdV- and LV-rHER-2/DCs. Splenocytes were obtained 10 days after vaccination and cultured with and without RM-1-mHER-2 tumor cells at a splenocyte-to-tumor cell ratio of 10:1, 20:1, and 40:1 for 48h. Cell culture supernatants were collected and cytokines measured by ELISA. (a) IFN-γ *P<0.02, **P<0.01; (b) transforming growth factor (TGF)-β *P<0.03; Results are representative of 15 samples (mean±s.e.m.) of three independent experiments. (c) Splenocytes co-cultured with RM-1-mHER-2 tumor cells (1:40) were also immunostained with anti-CD4, anti-CD8 and anti-IFN-γ antibodies and analyzed by flow cytometry. Data represent a mean (±s.e.m.) of four individual mice per group. *P<0.0206, **P<0.0067, ***P<0.0005.

Next, to determine the antitumor specificity of IFN-γ-secreting cells, the splenocytes were cultured in the presence of RM-1-mHER-2 tumor cells at the ratio of 1:10 (tumor cells:splenocytes). Splenocytes from the control groups did not react to the tumor cells, maintaining the same level of IFN-γ as the non-stimulated cells. Surprisingly, IFN-γ production was apparently downregulated when the cells from rHER-2/DC-treated mice were co-cultured with the tumor cells. The concentration of IFN-γ dropped from 15.9 to 6.6 ng ml⁻¹ (P<0.0001) for the AdV group and from 9.5 to 5.5 ng ml⁻¹ (P<0.0026) for the LV group (Figure 5a). In order to see if this effect was reversible, we measured the change in cytokine secretion after increasing the ratio of splenocytes per tumor cell. The IFN-γ secretion from the rHER-2/DC group increased significantly in the co-cultures of 1:20 and 1:40 ratios. Splenocytes from the LV-rHER-2/DC mice group strongly responded to tumor cells with 2.5-fold (1:20, P<0.0172) and 4-fold (1:40, P<0.0119) increases in IFN-γ secretion, in comparison with the spontaneous level, whereas a 1.5-fold (1:20, P<0.0042) and 2-fold (1:40, P<0.0048) increase was detected in the AdV-rHER-2/DC splenocyte co-culture (Figure 5a).

We also examined the levels of immunosuppressive cytokines interleukin-10 and transforming growth factor-β in the co-culture of splenocytes and tumor cells. We did not find any differences in interleukin-10 secretion between the rHER-2/DC- and enGFP/DC-treated mice (data not shown). No significant differences were found in the secretion of transforming growth factor-β in the co-culture 1:10 ratio among the animal groups (Figure 5b). However, the splenocytes from rHER-2/DCs mice co-cultured at the 1:40 ratio produced significantly less transforming growth factor-β than the enGFP/DC group (AdV P<0.0363, LV P<0.0123) (Figure 5b).

Next, we explored whether IFN-γ was produced mainly by CD4+ or CD8+ T cells in each vaccination group. Unfractionated splenocytes from vaccinated mice were co-cultured with RM-1-mHER-2 tumor cells and stained with anti-CD4, anti-CD8, and anti-IFN-γ antibodies. Not surprisingly, the number of CD4+IFN-γ+ and CD8+IFN-γ+ cells in the rHER-2 groups increased significantly after re-stimulation with tumor cells (AdV: CD4+IFN-γ+, P<0.0140 and CD8+IFN-γ+, P<0.0215; LV: CD4+IFN-γ+, P<0.0123), with the exception of the CD8+IFN-γ+ cells from the LV group, which did not change in this condition (Figure 5c). In addition, the IFN-γ-producing CD4+ and CD8+ T cells were higher in the rHER-2 groups when compared with enGFP/DCs groups (AdV: CD4+IFN-γ+, P<0.0260 and CD8+IFN-γ+, P<0.0005; LV: CD4+IFN-γ+, P<0.0067 and CD8+IFN-γ+, P<0.0206). In this study, IFN-γ appeared to be predominantly produced by CD4+ T cells. Notably, when we compared both vector systems, CD4+IFN-γ+ T cells were more abundant in the LV- rHER-2/DC group (P<0.0323), while CD8+IFN-γ+ T cells were similar in abundance between the treatments (Figure 5c).

Higher titer of anti-mHER-2 antibody production following AdV- rHER-2/DC vaccination

To further investigate whether AdV- and LV-rHER-2/DCs elicit humoral antigen-specific immunity, sera from immunized mice were tested on cell-based enzyme-linked immunosorbent assay (ELISA) 10 days after the injection of DCs. As a negative control, sera from naive mice were used. RM-1 tumor cells not overexpressing mHER-2 (that is, RM-1-NT) were also used as a control. Corroborating the tumor growth profile, mice that received the rHER-2/DC vaccine developed a significantly higher titer of anti-mHER-2 antibodies (total immunoglobulin G (IgG)) than the control groups (Figure 6a) at 10 days after immunization. Interestingly, the AdV-rHER-2/DCs vaccine generated the highest antibody response titer by ELISA.

Figure 6.

Induction of specific anti-mHER-2 antibodies post- AdV- and LV-rHER-2/DC vaccination. Serum was obtained from tumor-bearing mice at day 3 and day 10 post-vaccination. Pre-immunization serum was used as negative control. Specificity was controlled by incubating the serum with parent RM-1 tumor cells. (a) Titration curves of anti-mHER-2 antibody (day 14) against RM-1-mHER-2 tumor cells in a cell-based ELISA. Serial dilution results are shown as absorbance mean at 450 nm minus the background value developed by incubation with RM-1-NT (non-overexpressing HER-2 tumor cells). Each dilution was tested in duplicate; a total of four individual sera were analyzed per group. *P<0.0313, **P<0.0025. (b) Frequency of antibody–cell binding. Serum was incubated with 0.5×10⁶ RM-1-mHER-2 tumor cells at 1:1000 dilution. After incubation, the cells were incubated with PE anti-mouse IgG antibody and the positive cells analyzed by flow cytometry. The data are shown as mean (±s.e.m.) of individual positive cell frequencies. Ten samples for each group from two different experiments were assayed. **P<0.006.

Alternatively, we analyzed the binding of serum antibody to RM-1-mHER-2 tumor cells by flow cytometry. Validating the data obtained from the cell-based ELISA, we found that the sera from rHER-2/DC-vaccinated mice reacted strongly with the tumor cells in a flow cytometry-based assay, independent of the virus system used to transduce the DCs. The frequency of antibody-RM-1-mHER-2 cell binding increased significantly from day 3 to day 10 in the rHER-2/DC-treated mice (Figure 6b). In contrast, the levels of anti-mHER-2 antibody in the sera of mice immunized with NT/DCs and enGFP/DCs did not change significantly from day 3 to day 10, although the antibody levels were higher in comparison with pre-immunization levels (Figure 6b). Moreover, the specificity of antitumor antibodies present in the sera of immunized mice was confirmed by incubation with RM-1-NT cells, in which minimal cross-reaction was detected (Figure 6b). Sera from AdV-rHER-2/DC-immunized mice recognized a 10-fold (P<0.0061) higher frequency of tumor cells than their respective controls, while a 3-fold (P<0.0018) increase was detected between the LV-rHER-2/DC and enGFP/DC groups.

Splenocytes from AdV- and LV-rHER-2/DC-immunized mice elicit antitumor immunity following adoptive transfer into naive tumor-bearing mice

Having found differences in the magnitude of antitumor immunity generated by vaccination with AdV- and LV-rHER-2/DCs, we next aimed to determine whether tumor-reactive splenocytes from vaccinated mice would provide antitumor immunity in non-DC-vaccinated tumor-bearing mice. Mice at 4 days after tumor implantation received 7.5×10⁶ splenocytes intravenously that were obtained from immunized tumor-bearing donors at 8 days post-vaccination. The frequencies of CD4+ and CD8+ T cells in the donor splenocytes were 23.6 and 14.5% for AdV-rHER-2/DCs, 20.3 and 15.1% for LV-rHER-2/DCs, 12.2 and 9.3% for AdV- enGFP/DCs, 11.2 and 7.5% for LV-enGFP/DCs, and 9.2 and 6.9% for NT/DCs, respectively (Figure 7a).

Figure 7.

Adoptive transfer of splenocytes from AdV- and LV-rHER-2/DC-vaccinated mice induce similar antitumor immunity in tumor-bearing mice. Splenocytes (7.5×10⁶ cells) obtained 10 days post-vaccination were transferred intravenously to tumor-bearing mice 4 days after tumor injection. (a) Splenocytes were stained with FITC anti-CD4 and PE anti-CD8 antibodies and the frequency of positive cells determined by flow cytometry. Representative results are shown. (b) Tumor volume was measured every 2 days and the results shown as mean (±s.e.m.) of 5–7 mice per group. *P<0.01.

Adoptive transfer of rHER-2-programmed splenocytes into tumor-bearing mice led to statistically significant tumor inhibition. At 14 days post-splenocyte transfer, the tumor volume of the AdV- and LV-rHER-2 group was fivefold smaller than the control groups (Figure 7b). We did not detect any significant differences between mice treated with splenocytes derived from AdV- or LV-rHER-2/DC-immunized mice.

DISCUSSION

Overexpression of HER-2 is a common occurrence in a variety of human tumors and often correlates with poor prognosis. Considerable efforts have been made to find a treatment for HER-2-positive tumors using gene therapy. The selection of the virus platform to implement DC vaccination has traditionally been a dilemma. We questioned whether suboptimal results from some AdV- and LV-based antitumor immunotherapy experiments stem from the choice of the virus gene delivery system, where a particular tumor-associated antigen may work better in one virus system than another. To our knowledge, there are no direct reports comparing AdV- and LV-transduced DCs expressing the same antigen. Here, we performed a comparative analysis of DCs overexpressing rHER-2 after AdV and LV infections. We observed that the infection either by AdV or LV did not induce the maturation of DCs. Contrary to some reports, we did not observe maturation of immature DCs following AdV infections.^15-17 Various other reports support our finding that the AdV itself does not induce DC maturation.^18-20 These contradictions may be related to differences in virus preparations and infection protocols, such as the number of virus particles per cell (MOI) used or even different lots of fetal calf serum used in the activating cultures. Importantly, in our model, AdV and LV infection of immature DCs did not impair maturation following tumor necrosis factor-α stimulation.

LV vectors are well known for stable and long-lasting expression of transgenes. When we evaluated tumor volumes in animals treated with AdV- and LV-rHER-2/DCs, we found a similar and effective inhibition of tumor growth in both groups. Although AdV transductions result in a non-stable expression of transgenes in comparison with LV transductions, it may not be as relevant in the context of ex vivo-directed DC transductions. Fully differentiated mature DCs do not divide and have a short lifespan that varies from 2 to 9 days once they enter the lymph nodes.²¹ We presume that neither AdV nor LV infection would alter the longevity of DCs. Moreover, the maturation status and rHER-2 levels on the cell surface were similar between cells infected with both virus systems, suggesting that both cells had the same potential with respect to the expression of co-stimulatory molecules required to sufficiently activate T cells in vivo.

The outcome of tumor growth was similar between animals receiving AdV- and LV-rHER-2/DCs. That said, the magnitude of the immune response mechanism was different. AdV-rHER-2/DC-treated mice mounted a local and systemic immune response with high frequencies of CD4+ and CD8+ T cells while spontaneous IFN-γ secretion was higher than animals treated with LV-rHER-2/DCs. A previous study using AdVs engineering expression of rHER-2 showed that tumor protection was completely dependent on anti-HER-2 antibody production with participation of CD4+ T cells in the early phase of immunization, but it does not depend on CD8+ T cells.²² Our results are in agreement with the idea that the generation of anti-HER-2 antibody is important for the development of an efficient antitumor response. Both AdV- and LV-rHER-2/DCs were able to elicit a humoral immune response in comparison with control groups. Contrary to this previous report, in the AdV-rHER-2/DC model the antitumor immunity appears to involve both CD4+ and CD8+ T cells, whereas for LV-rHER-2/DCs, responses were more related to CD4+ T cells as observed by T-cell infiltration at the tumor site. Other studies have also shown that tumor-specific CD4+ T cells have more potential to eradicate tumors than do CD8+ T cells.^23,24 On the other hand, our findings demonstrate that splenocytes from LV- rHER-2/DC-immunized mice responded more robustly to RM-1-mHER-2 tumor cells with enhanced secretion of IFN-γ by CD4+ T cells in contrast to splenocytes derived from mice that received the AdV-rHER-2/DCs. Moreover, the IFN-γ levels induced by AdV vaccines appeared to be dependent on both CD4+ and CD8+ T cells.

Another difference between AdV- and LV-rHER-2/DC vaccines was observed in our study regarding the subsets of DCs found in the spleen. It is known that different subsets of DCs promote a distinct pattern of cytokines and T-cell stimulation.²⁵ Here, myeloid and plasmacytoid DCs were found to be predominant in the AdV-rHER-2/DC group while the LV counterpart was mainly represented by the plasmacytoid DC subset. Despite the involvement of plasmacytoid DCs with infection immune responses,^26,27 data in the literature also support the direct or indirect antitumor effect of this specific DC subset.²⁸ A synergistic effect of myeloid and plasmacytoid DCs inducing specific antitumor immune response has also been discussed before.^29,30 The combination of both subsets enhanced the frequency of antigen-specific CD8+ T cells and antitumor immunity. Consistent with these findings, our results with AdV showed an antitumor effect that suggest a main role of CD8+ T cells or a balance of CD4+/CD8+ T cells.

We suggest that the differences observed may be related to an altered repertoire of immunodominant epitopes presented by the transduced DCs. One vaccine may narrow the immune response toward a limited number of immunodominant epitopes while the other goes against a subdominant repertoire. This is supported by a recent publication from Dai et al.³¹ involving directly administered boosting immunizations using LV-gag and AdV-gag. Here, the authors suggested that the LVs may direct DCs to load and present antigens more efficiently in this context.

We also hypothesize that the enhanced frequency of CD4+ and CD8+ T cells in the AdV-DCs immunization may be due to the additional anti-viral response that would be seen with this vaccine schema, which would not be present with the LV-DCs. This is supported by findings that AdV-transduced DCs are also loaded with viral capsid antigens³² and can stimulate an anti-viral response. Regardless of the differences in anti-vector responses, the overall antitumor activity was the same between both vectors. Taken together, our results indicate that AdV- and LV-mediated xenoantigen rHER-2 can be expressed at similarly high levels in DCs, which in turn, can activate antigen-specific cellular and humoral immune responses. Although the transient expression of foreign antigens may not be a concern in cancer gene therapy, the virus system seems to have an impact on determining the nature of the immune responses. Future studies may seek to capitalize on this to direct particular responses.

MATERIALS AND METHODS

Cell lines and mice

RM-1 prostate tumor cells, syngeneic to the C57BL/6 strain, were previously engineered to overexpress a kinase-truncated form of murine HER-2/neu (mHER-2).¹³ Cells were cultured in Dulbecco’s modified Eagle’s medium (Sigma Aldrich, St Louis, MO, USA) media supplemented with 10% fetal bovine serum, 2mm l-glutamine, and 1% penicillin–streptomycin (Invitrogen, Carlsbad, CA, USA) and maintained at 37 °C at 5% CO₂; passages were performed every 2 days.

C57BL/6 mice were bred and housed under specific pathogen-free conditions at the University Health Network Animal Resource Centre. Animal experimentation followed protocols approved by the UHN Animal Care Committee (ACC).

Vector construction and virus production

A kinase-inactivated rat HER-2 AdV was previously constructed and generously provided by Dr Y Wan.¹¹ Importantly, a single amino acid replacement (lysine to alanine) at codon 758 was performed to inactivate the kinase domain. Construction of the rat HER-2 LV was performed by subcloning of the rat HER-2 complementary DNA into the pCCL LV transfer vector backbone kindly provided by Dr L Naldini.³³ Briefly, the enGFP and delta Low-affinity Nerve Growth Factor Receptor dual-promoter system cassette in the pCCL vector was removed by digestion and the elongation factor-1α promoter subcloned into ClaI and AscI restriction enzyme sites. Subsequently, the full-length rHER-2 was amplified from the pJ4 Neu N plasmid by PCR using the KOD Hot Start DNA polymerase (EMD4Biosciences, San Diego, CA, USA). The rHER-2 complementary DNA fragment of approximately 3.7 kb in size was subsequently subcloned downstream the elongation factor-1α promoter into AscI and SalI restriction enzyme sites to yield the plasmid pCCL.SIN.cPPT.EFIa.rHER2-neu.Wpre. The final construction was verified by bidirectional DNA sequencing.

Recombinant rHER and enGFP AdV (denoted as AdV-rHER-2 and AdV-enGFP) were generated by the co-transfection of the rat HER Ad5 shuttle vector with the rescue vector pBHG10 into HEK 293 cells as previously described.¹¹ All recombinant Ad vectors were propagated using 293 cells and purified using CsCl gradient centrifugation as described previously.³⁴ Recombinant rHER-2 and enGFP LVs (denoted as LV-rHER-2 and LV-enGFP) were produced by transient co-transfection of HEK 293T cells with three plasmids: pMD.G (VSV-g envelope), pCMVΔ8.91 (packaging) and either rHER-2 or enGFP lenti-transfer vectors.^6,35 Transfections were performed using branched polyethylenimine as previously described.³⁶ Viral supernatants were harvested 48 and 72 h after transfection and concentrated by ultracentrifugation at 50 000 g. The final concentrated virus was stored at −80 °C until use. Functional viral titers were determined by serially diluted transductions of HEK 293 T cells, followed by rHER-2 expression analysis using flow cytometry.

DCs generation and infection

To generate DCs, murine bone marrow cells were cultured in RPMI-1640 medium (Sigma Aldrich) supplemented with 10% fetal bovine serum, 2 mm l-glutamine, 50mm 2-mercaptoethanol, 5% non-essential amino acids (Gibco, Carlsbad, CA, USA), 100Uml⁻¹ penicillin, and 100 μg ml⁻¹ streptomycin (complete medium). Briefly, bone marrow was flushed from the femurs and tibias of C57BL/6 mice and bone marrow cells were plated in Petri dishes in complete medium, which was further supplemented with 50 ng ml⁻¹ recombinant murine granulocyte-macrophage colony-stimulating factor (Peprotech, Rocky Hill, NJ, USA). DCs were transduced on day 3 with LVs at MOI of 15–20 and on day 5 with AdVs at MOI of 100. On day 7 of culture, DC maturation was induced by the addition of 100 ng ml⁻¹ tumor necrosis factor-α (Peprotech) for 24 h. The following day, non-adherent and loosely adherent cells were harvested by gentle pipetting, washed, and resuspended in phosphate-buffered saline (PBS) for injection. Mice were immunized with matured DCs on day 8 of culture.

Tumor inoculation and immunization

C57BL/6 mice were injected subcutaneously on the dorsal region with 2×10⁵ RM-1-mHER-2 tumor cells in PBS. On day 4 when the tumors reached approximately 20 mm³ in volume, the mice were immunized in the footpad with 2×10⁵ rHER-2/DCs transduced by either AdV or LV. NT/DCs or enGFP/DCs were used as controls. Mice were observed every other day and the length (l), width (w) and height (h) of each tumor was measured using a digital caliper. Tumor volume was calculated using the formula l×w×h. Mice were euthanized when tumors ulcerated or surpassed 1500 mm³ in volume (as required by the UHN ACC).

Adoptive transfer of splenocytes

Eight days after immunization with DCs, spleens were harvested and homogenized to single-cell suspensions. Unfractionated splenocytes were depleted of red blood cells and resuspended in PBS for injection. Four days after tumor implantation, tumor-bearing mice received 7.5×10⁶ splenocytes intravenously. Tumor growth was monitored every 2 days with calipers and measurements were expressed as tumor volume as above.

Flow cytometry analysis

For cell surface staining, murine bone marrow-derived DCs were incubated with phycoerythrin (PE) anti-CD11c (BD Pharmingen, Franklin Lakes, NJ, USA), fluorescein isothiocyanate (FITC) anti-CD80 (BD Pharmingen), FITC anti-CD86 (BD Pharmingen) and FITC anti-major histocompatibility complex-II (BD Pharmingen) antibodies. Antibodies labelled with PE were used for the cells transduced with virus-encoding enGFP. rHER-2 expression was detected by anti-rHER-2 primary antibody (Calbiochem, San Diego, CA, USA) followed by the secondary FITC anti-IgG antibody staining (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Nonspecific staining was blocked with anti-CD16/CD32 antibody (BD Pharmingen). The phenotypic characterization of splenocytes was performed by incubating the cells with FITC or antigen-presenting cell anti-CD4 (eBioscience, San Diego, CA, USA), PE anti-CD8 (BD Pharmingen), antigen-presenting cell anti-CD11c (BD Pharmingen), FITC anti-CD11b (BD Pharmingen), PE anti-B220 (BD Pharmingen) and FITC anti-NK1.1 (BD Pharmingen) antibodies. Intracellular staining of IFN-γ was performed using the BD Cytofix/Cytoperm Kit (BD Biosciences, Franklin Lakes, NJ, USA). Briefly, splenocytes were co-cultured with or without RM-1-mHER-2 at a ratio of 1:40 (splenocytes:tumor cells) for 48 h and monensin (GolgiStop, BD Biosciences) was added in the last 4 h. Cell surface staining was performed with antigen-presenting cell anti-CD4 and PE anti-CD8 antibodies. After fixation and permeabilization, cells were incubated with FITC anti-IFN-γ (BD Pharmingen) antibody. Staining of regulatory T cells was done using the Foxp3 eBioscience protocol. After CD4 and CD25 staining, splenocytes were incubated with fixation/permeabilization (eBioscience) working solution overnight at 4 °C. Subsequently, cells were stained with antigen-presenting cell anti-Foxp3 (eBioscience) antibody. The labelled cells were analyzed using a FACSCalibur flow cytometer (BD Biosciences).

Anti-mHER-2 antibody detection

Serum samples were collected pre- and post-immunization and assayed for the presence of mHER-2 antibody. A cell-based ELISA assay was developed for antibody titration. Briefly, 4×10⁴ RM-1-mHER-2 tumor cells or RM-1-NT (non-overexpressing HER-2) were seeded in a 96-well plate (Costar, Lowell, MA, USA) and incubated overnight at 37 °C. Cells were fixed with 10% formalin buffer (Fisher Scientific, Waltham, MA, USA) for 15 min at room temperature and then blocked with 3% bovine serum albumin in PBS for 2 h at 37 °C. Serial dilutions of serum were prepared in duplicate and added into wells for 2 h. Subsequently, the plates were washed with PBS/0.05%Tween-20 solution and the bounded antibody detected by addition of a horseradish-peroxidase goat anti-mouse IgG antibody (Southern Biotech, Birmingham, AL, USA) diluted 1:8000 in 1.5% bovine serum albumin/PBS. Tetramethylbenzidine (Sigma Aldrich) was added as substrate and the enzymatic reaction stopped with 2n hydrochloric acid. Serial dilution was plotted versus optical density at 450 nm.

The frequency of specific antibody/cell binding was also analyzed by flow cytometry. For detection of antibody-cell binding, 5×10⁵ RM-1-mHER-2 tumor cells were incubated with 100 μl of serum at a dilution of 1:1000 for 1 h at 4 °C. Next, the cells were washed and incubated with PE goat anti-mouse IgG secondary antibody (BD Pharmingen) for 1 h. Stained cells were then analyzed by flow cytometry and the frequency of positive cells was expressed as percentages. RM-1 tumor cells not overexpressing mHER-2 (RM-1-NT) were used as a control.

ELISA for cytokines

Splenocytes were obtained at sacrifice and co-cultured for 48 h with RM-1-mHER-2 tumor cells at splenocyte-to-tumor cell ratios of 10:1, 20:1 and 40:1. Cells cultured alone were used to measure the spontaneous secretion of cytokines. Measurement of IFN-γ (BD Biosciences), interleukin-10 (BD Biosciences), and transforming growth factor-β (R&D Systems, Minneapolis, MN, USA) in the cell supernatant was accomplished by sandwich ELISAs according to the manufacturer’s instructions.

Immunohistochemistry

Tumor tissue were embedded in tissue freezing medium, snap frozen in liquid nitrogen and subsequently cut into 5 μm thick sections. Frozen tissue sections were fixed in acetone for 15 min and then incubated with PBS+1% bovine serum albumin+0.2% gelatine-blocking solution for 1 h. After washing, the sections were stained with 1:50 rat anti-mouse CD4 (eBioscience) and rat anti-mouse CD8α primary antibody (eBioscience) for 1 h, followed by incubation with Alexa 488 goat anti- rat IgG (Molecular Probes, Eugene, OR, USA) at a dilution of 1:200 for 1 h. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (Sigma Aldrich). Background control staining was performed by incubation of the secondary antibody alone. Images were acquired using an Olympus (Tokyo, Japan) BX50 microscope at the Advanced Optical Microscopy Facility (AOMF, University Health Network). The evaluation of stained cells was performed using the ImageJ software (http://rsb.info.nih.gov/ij/).

Statistical analyses

Data are presented as means ± s.e.m. Statistical comparisons were obtained by Student’s t-tests by using Instat (GraphPad, San Diego, CA, USA). P-values<0.05 were considered statistically significant.