MHC class I and class II Presentation of Tumor Antigen in Retrovirally and Adenovirally Transduced Dendritic Cells

Abstract

The unique antigen-presenting capabilities of dendritic cells (DCs) make them an attractive means with which to initiate an antitumor immune response. Using DCs transduced with tumor antigens for immunotherapy has several theoretical advantages over peptide-pulsed DCs including the possibility that transduced DCs are capable of presenting epitopes on both class I and class II MHC molecules. To test this theory, we inserted the human tumor antigen gp100 into mouse DCs transgenic for HLA-DRb1^*0401 using either adenoviral vector or a VSV-G pseudotyped retroviral vector. DCs transduced with tumor antigen were able to be recognized by both a murine CD8⁺ T-cell clone and a murine CD4⁺ T-cell line in a cytokine release assay, thereby demonstrating presentation of both MHC class I and class II gp100 epitopes. This study describes the simultaneous presentation of a tumor-associated antigen to both CD4⁺ and CD8⁺ T cells and lends support to the use of gene-modified DCs as a means to initiate both CD4⁺ and CD8⁺ antitumor responses.

Tumor-antigen-reactive T cells have been detected in the circulation of patients with cancer.¹ However, despite this, tumors continue to progress in most patients, potentially because tumor-reactive T cells are in a nonresponsive state and low in number and may need to be activated to achieve an antitumor response. Many different methods are being attempted for the activation of T cells including vaccines composed of peptides from tumor-associated antigens² and viral³ or DNA vectors⁴ encoding these antigens. An alternative strategy involves the use of dendritic cells (DCs) that constitute a particularly attractive adjuvant for T-cell stimulation because they play a pivotal role in the initiation of immunity.⁵ DCs are able to stimulate both CD8⁺ and CD4⁺ subsets of T cells, both of which are important in generating a robust antitumor response.⁵

DCs loaded with tumor antigens have been shown to inhibit tumor growth in several mouse models.^6,7 Methods used to load tumor antigen in these studies include pulsing DCs with defined peptide epitopes or with tumor lysates. These methods primarily result in antigen presentation on MHC class I or class II molecules that stimulate either CD8⁺ or CD4⁺, respectively. However, because both CD8⁺ and CD4⁺ T cells play crucial roles in antitumor responses it is therefore advantageous to develop a method whereby DCs can simultaneously present MHC class I and II tumor antigen epitopes.

Genetic modification of DCs has been shown to result in presentation of a number of MHC class I and II restricted epitopes to T cells including those from beta-galactosidase,⁸ but direct presentation by DCs of tumor-associated antigen via both MHC class I and II to tumor-specific T cells is less well characterized.

We have previously developed methods for gene-modifying DCs,⁸ and in this study we used these methods to insert the human tumor antigen gp100 into DCs. Human gp100 has well-defined class I and class II epitopes restricted to H-2D^b and HLA-DRβ1^*0401.^9,10 Using T-cell lines specific for these two epitopes, we were able to demonstrate that retrovirally and adenovirally transduced DCs are able to simultaneously present epitopes to CD4⁺ and CD8⁺ cells.

Materials and methods

Animals and peptides

Murine class II-deficient, DR4-IE transgenic mice¹¹ ex-press chimeric class II molecules consisting of the antigen-binding domains from HLA-DRα and HLA-DRβ1^*0401 molecules and the remaining domains from the IE^d-a2 and IE^d-b2 chains^II. Animals were obtained from Taconic (Germantown, NY) and maintained in accordance with institutional standards. The following peptides were synthesized using standard fluorenylmethoxycarbonyl (F-moc) chemistry. Mouse gp100_25-33: EGSRNQDWL (restricted by H-2Db), human gp100_25-33: KVPRNQDWL (restricted by H-2Db), and human gp100_44-59: RQLYPEWTEAQRL (restricted by HLA-DRβ1^*0401). The molecular masses of peptides were verified by laser desorption mass spectrometry (Biosynthesis, Lewisville, TX). Purity was also determined in this manner and found to be >99%.

Production and retroviral transduction of DC

DCs were produced from bone marrow of DR4-IE mice as follows. Bone marrow was expelled from the long bones of DR4-IE mice. Erythrocytes were depleted in ACK lysing buffer (Biofluids, Rockville, MD). Cells expressing B220, I-Ab (MHC class II), CD8, and CD4 were removed using hybridoma supernatant from TIB-146, TIB-229, TIB-150, TIB-207 (ATCC, Manassas, VA) (0.2 mL of each supernatant per 10⁷ cells) followed by incubation with rabbit complement (10 mL per 10⁸ cells) (Cedarlane, Ontario, Canada). The remaining cells were plated at a concentration of 7 × 10⁵ cells/mL in six well plates, and cultured for 6 days in DC medium (RPMI supplemented with 5% heat-inactivated FCS (both from Gibco BRL), 2 mM glutamine (Biofluids), 100 U/mL penicillin and 100 μ g/mL streptomycin (Biofluids), and 5 × 10 ^-5 M 2-mercaptoethanol (Sigma, St. Louis, MO). Recombinant murine GM-CSF was added to a final concentration of 20 ng/mL (Peprotech, Rocky Hill, NJ) and recombinant murine IL-4 (Peprotech) added to a final concentration of 100 ng/mL. Cytokines were replenished on days 2, 4, and 6. On day 6, nonadherent cells were collected and replated at 1 × 10⁶ cells/mL in a six-well plate. On day 7, the nonadherent cells were collected and washed three times in PBS before use. The retroviral plasmids pCLNCGFP and pCLNChgp100 contain the gene encoding GFP or gp100 driven by a CMV promoter embedded within a Moloney Murine Leukemia Virus backbone. The pCLNC plasmid and a plasmid encoding the VSV-G envelope protein were cotransfected into 293 cells constitutively expressing the gag and pol proteins.¹² Serial supernatant harvests were performed on days 2, 3, 4, and 5 posttransfection of the 293GP cells, passed through a 0.45-μ m filter (Millipore, Bedford, MA), and frozen. Supernatants were thawed on the day of transduction and concentrated in an ultracentrifuge (50,000 g, 1.5 hours, 4°C). Because retroviral transduction requires proliferating cells, bone marrow cells were exposed to retroviral supernatants during the proliferative phase of culture. One day after bone marrow harvest, the concentrated viral supernatant was added to the DCs along with 8 μ g/mL polybrene (Sigma, St. Louis, MO) and 10 mM HEPES (Biofluids). Cells were then spun on a tabletop centrifuge for 1 hour (1000 ×g,25-30°C).

Adenoviral transduction of DCs

The recombinant adenoviral vectors Ad.GFP and Ad.hgp100 were a gift from Bruce Roberts (Genzyme, Framingham, MA) and are described previously.¹³ Because adenoviral transduction is transient, bone marrow cells were exposed to recombinant adenovirus following proliferation and differentiation into DCs. After replating the DCs on day 6, the adenovirus was added at a multiplicity of infection of 500:1 and allowed to incubate 24 hours at 37°C, 5% CO₂.

Generation of murine CD8⁺ and murine CD4⁺ T-cell lines

Generation of the CD8⁺ T-cell clone specific for the H-2D^b restricted epitope of gp100 (gp100_25-33) used in this experiment has been described in detail elsewhere.⁹ Briefly, CD8⁺ T cells were derived from splenocytes of mice immunized by gene gun with human gp100 DNA. This line was thereafter restimulated with 1 μ M mouse gp100_25-33 peptide. These cells are able to recognize both human and murine gp100_25-33 epitopes presented on D^b molecules, despite a three amino acid difference between the murine and human epitope. All T cells were used between 5 and 10 days after restimulation. Generation of the CD4⁺ T-cell line used in this experiment has been described in detail elsewhere.¹⁰ Briefly, they were derived from DR4-IE transgenic mice that were originally immunized with hgp100 protein. Cells were subsequently stimulated with irradiated syngeneic DR4-IE splenocytes pulsed with hgp100_44-59.

IFN-γ release assay

After the DC harvest on day 7, 1 × 10⁵ viable T cells and 1 × 10⁵ DCs were incubated in duplicate for 12 hours in 96-well plates. The murine IFN-γ concentration in the supernatant was determined by ELISA, using commercially available reagents according to the manufacturer’s instructions (Endogen, Rockford, IL). Peptide-pulsed DCs were also used as targets in some assays, in which case nontransduced DCs were pulsed with 1 μ M peptide for MHC class I-restricted epitopes or 10 μ M peptide for MHC class II-restricted epitopes.

Results

DCs derived from DR4-IE transgenic mice can be efficiently transduced using adenovirus or VSV-G pseudotyped retrovirus

To determine the transduction efficiency of HLA-DRβ1^*0401 transgenic DCs with retroviral and adenoviral vectors, GFP-transduced DCs derived from DR4-IE mice were analyzed by flow cytometry. Both adenovirally and retrovirally transduced cells expressed GFP (Fig 1, a and b). However, GFP was expressed in a greater percentage of retrovirally transduced DCs (∼50%) compared to adenovirally transduced DCs (∼20%). Expression levels were also higher in DCs following transduction with retrovirus when compared to adenovirus. Retrovirally and adenovirally transduced cells were also strongly positive for HLA-DR expression (Fig 1, c and d). Both types of GFP-transduced cells also displayed cell surface markers similar to nontransduced DCs. Specifically, the DCs were strongly positive for CD11c, B7.1, and B7.2 expression (data not shown).

Figure 1.

Transduced DCs express GFP and MHC class II. DCs were transduced with either retroviral CLNC-GFP (a,c) or adenoviral Ad-GFP (b,d) and analyzed by flow cytometry. GFP expression was clearly seen in both retrovirally (a) and adenovirally (b) GFP-transduced DCs (thin line) compared to gp100-transduced DCs (thick line). However, lower levels of GFP were detected in adenovirally-transduced cells compared to retrovirally-transduced cells. High levels of MHC class II were also demonstrated in retrovirus-transduced (c) and adenovirus-transduced (d) DCs when stained with antibody specific for DRb1^*0401 (thin line) compared to isotype control (thick line).

A murine CD8⁺ T-cell clone specific for hgp100_25-33 specifically recognizes hgp100-transduced DCs

To determine if the transduced DCs were able to process and present a class I epitope encoded by the transgene, a murine CD8⁺ T-cell clone specific to the H-2D^b epitope of gp100 (hgp100_25-33) was used in a cytokine release assay with the transduced DCs. This clone was able to specifically recognize the gp100-transduced DCs above GFP-transduced controls for both adenovirally and retrovirally transduced cells (Fig 2). Recognition of adenovirally transduced DCs was lower than retrovirally transduced DCs, which may be due to lower expression levels of gp100 in adenovirally transduced cells. Numbers given are representative of several experiments. This indicates that the transduced DCs efficiently presented transgene-encoded epitopes along the presented endogenous MHC class I pathway.

Figure 2.

Transduced DCs present antigen on MHC class I. DCs transduced with adenovirus or retrovirus encoding GFP or human gp100 were cocultured with a mouse T-cell line specific for human gp100_25-33 (restricted by H-2D^b). In this representative experiment, specific recognition of gp100-transduced DCs was evident from the higher levels of IFN-γ secretion in response to gp100-transduced cells when compared to GFP-transduced cells. DCs transduced with gp100 using retrovirus were recognized more strongly than those transduced with adenovirus. Interestingly, T cells secreted IFN-γ in response to DCs pulsed with gp100 peptide at a lower level than that seen in response to retrovirally transduced DC.

A murine CD4⁺ T-cell line specific for hgp100_44-59 specifically recognizes hgp100-transduced DCs

DCs identical to those used in the CD8⁺ assay were used in a cytokine release assay with CD4⁺ T cells recognizing the HLA-DRβ1^*0401-restricted hgp100_44-59 epitope. These CD4⁺ T cells were able to specifically recognize the gp100-transduced DCs above GFP-transduced controls for both adenovirally and retrovirally transduced DCs (Fig 3). This indicates that transduced DCs were also able to process and present a class II epitope from an endogenous antigen. Activating the transduced DCs with CD40L had little additional effect on either class I or class II presentation (data not shown).

Figure 3.

Transduced DCs present antigen on MHC class II. DCs were transduced with either adenovirus or retrovirus encoding the tumor antigen gp100 or GFP. CD4⁺T cells specific for human gp100_44-59 (restricted by HLA-DRβ1^*0401) were able to specifically recognize gp100-transduced DCs but not GFP-transduced DCs as seen by IFN-γ secretion following overnight coculture. Transduction of gp100 using retrovirus produced DCs that were recognized at a higher level than DCs transduced with adenovirus. T cells also secreted IFN-γ in response to DCs pulsed with human gp100_44-59.

Discussion

In previous work we developed methods of transducing DCs and demonstrated that injection of DCs transduced with a model antigen gene encoding beta-galactosidase into mice could inhibit the growth of established lung metastases expressing beta-galactosidase.⁸ Certain endogenous antigens have previously been shown to be efficiently presented on MHC class II molecules,¹⁴ and several groups have observed that immunizations with transduced DCs require CD4⁺ T-cell help for optimal antitumor effects;¹⁵ however, presentation of endogenous tumor antigen to CD4⁺ T cells by transduced DCs is less well documented.

Peptides bound to MHC class I molecules come predominantly from proteins degraded in the cytoplasm, whereas peptides bound to MHC class II molecules are mostly derived from exogenous or intravesicular sources.¹⁶ Recently a more complicated picture of antigen processing is emerging, in which specialized antigen-presenting cells in particular are able to sample proteins from either compartment for presentation on both class I and class II molecules.¹⁷

Here we demonstrate that adenovirally and retrovirally transduced DCs are capable of presenting class I and class II epitopes of the melanocyte differentiation antigen, gp100. Future studies will be important to generalize these findings to other antigens, as well as comparing these transduction methods using in vivo tumor models. Like other melanosome membrane glycoproteins, gp100 contains a melanosome transport signal,¹⁸ which has been shown to traffic antigens to the endocytic pathway for class II presentation.¹⁹ Because of this, class II presentation of all endogenous antigens cannot be presumed from the finding that transduced DCs are capable of presenting a gp100-derived class II epitope. However, endogenous antigens can be targeted to the class II compartment by incorporation of sequences from the MHC class II-associated invariant chain²⁰ or lysosomal-associated membrane protein (LAMP-1).²¹ It is also possible that MHC class II presentation in this study did not result from processing of endogenous antigen but rather was derived from exogenous antigen released from other transduced cells in the culture.

In addition to simultaneous presentation of MHC class I and II antigenic epitopes, there are other benefits of gene-modified DCs as vaccines compared to peptide-pulsed DCs including that all possible epitopes within the antigen could potentially be processed and presented. These epitopes would also be subjected to normal posttranslational processing, which has been shown to be important for presentation of some cysteinylated peptides, phosphopeptides, and glycopeptides to T cells.^22,23 Finally, because the protein continues to be expressed over time, transduced DCs may sustain epitope presentation longer than their peptide-pulsed counterparts.

The simultaneous presentation of class I and class II epitopes of a known tumor antigen in transduced DCs lends support to the use of gene-modified DCs as a means to simultaneously initiate a CD4⁺ and CD8⁺ antitumor response.