The HHD transgenic mice kindly provided by F.A. Lemonnier, Institut Pasteur, Paris, France, express a transgenic monochain histocompatibility class I molecule in which the C-terminus of the human β₂-microglobulin (β₂m) is covalently linked to the N-terminus of a chimeric heavy chain (HLA-A2.1 α1-α2, H-2D^bα3-transmembrane and intracytoplasmic domains). HHD mice are homozygous for the transgene, H-2D^b−/− and β₂m^−/− double knockout.

Epitope peptides identified as relevant for HIV-1 vaccination (12, 13) were synthesized by Schafer-N, Copenhagen, Denmark. Peptide-binding affinities to purified HLA-A*0201 molecules were measured using an in vitro biochemical binding assay (14). Briefly, pre-oxidized MHC-I molecules when diluted into a reaction buffer containing human β₂m and the appropriate test peptide, fold rapidly and efficiently only in the presence of the test peptide.

Bone marrow cells from HHD transgenic mice were kept overnight in culture medium [RPMI 1640 with 10% (v/v) FCS, 1% (v/v) penicillin, streptomycin, 1 ‰ (v/v) 2-mercapta-ethanol] and the non-adhering cells were cultured with GM-CSF, 10 ng/ml, and IL-4, 20 ng/ml (PreproTech Inc., Rocky Hill, NJ, USA). At day 3, fresh culture medium with GM-CSF and IL-4 was supplied. At day 6, the cells were harvested and cultured overnight with GM-CSF, IL-4 and 1 μg/ml of LPS (Sigma, St. Louis, MO, USA). At day 7, the mature DCs were harvested and used for vaccination.

Mature DCs were either loaded with a single peptide for individual peptide immunogenic ranking or loaded with peptides in four different combinations at a concentration of 10 μg/ml/peptide (Fig. 1). Vaccination protocols were performed as follows: (i) all seven peptides were loaded onto the DCs simultaneously, washed thoroughly, resuspended in phosphate-buffered saline (PBS) and injected intracutaneously (i.c.) in one flank of the mouse. (ii) The same number of DCs was aliquoted in seven vials and pulsed separately with one peptide for each vial, then washed thoroughly and mixed in one syringe followed by i.c. injection in one flank of the mouse. (iii) Peptides were loaded separately onto DCs as described in group B, but just prior to injection, epitopes ranked 1–3 according to their individual immunogenicity (Table 1) were pooled and administered in the left flank of the mouse and epitopes ranked 4–7 (Table 1) were pooled and administered in the right flank of the mouse. (iv) Epitopes ranked 1–3 (Table 1) and 4–7 (Table 1) were loaded onto DCs in two separate vials followed by i.c. injection in left and right flank of the mouse, respectively. A CD4⁺ T helper peptide PADRE (15) was included with each CTL epitope peptide to provide necessary CD4 T-cell help. Each mouse received 3 × 10⁶ DCs for individual peptide response or alternatively 1.3 × 10⁶ pulsed DCs for each peptide resulting in 9 × 10⁶ DCs in total for every mouse in the competition experiments.

HHD transgenic mice, 8–12 weeks old, were injected i.c. in the flank with matured DCs pulsed with the relevant HLA-A*0201-binding HIV-1 CTL peptide and a T-helper peptide PADRE as described above (15). DCs were resuspended in 100 μl PBS for each injection. At day 11 after vaccination, mice were sacrificed and the splenocytes were recovered and re-stimulated with the vaccination-peptide as described previously (13). After 5 days, the CTLs were used for flow cytometric analysis or harvested and depleted of NK and B cells using anti-CD49b antibodies and anti-CD19 antibodies (Mylteneyi Biotec, Gladbach, Germany), respectively. Cells were depleted on a column, MACS LS, according to the manufacturer’s protocol. The non-depleted cells were then used directly for cytolytic ⁵¹Cr-release assays and IFN-γ ELISpot assays.

Splenocytes cultured for 5 days were stimulated for 2 h with 2 μg/ml peptide antigen and subsequently incubated for 5 h with protein transport inhibitor (BD GolgiStop; BD Pharmingen, San Diego, CA, USA) at 37 °C and 5% CO₂. Thereafter, cells were stored overnight in 4 °C. Due to the low frequency of CD8⁺ T cells in this HHD⁺ transgenic mouse model, CD8⁺ T cells were negatively isolated using a CD8a⁺ T cell isolation kit (Miltenyi Biotec, Teterow, Germany) according to the manufacturer’s protocol, followed by surface staining using anti-CD3-PerCP (BD Pharmingen) and anti-CD8-FITC (BD Pharmingen). Cells were washed and subsequently fixed and permeabilized using the Fix and Perm kit (Caltag Laboratories, Der Grub, Austria) and stained intracellularly with anti-IFN-γ-Allophycocyanin (APC) (BD Pharmingen) and anti-tumor necrosis factor-α (TNF-α)-Phycoerythrin (PE) (BD Pharmingen). Cells were washed and resuspended in staining buffer prior to immediate flow analysis. Data were acquired on a BD LSRII instrument using FACSdiva software (BD Biosciences) and analyzed with FACSdiva or FlowJo software (TreeStar, Ashland, OR, USA).

A standard IFN-γ ELISpot assay was used as described previously (13). Cells, with titration of cells ranging from 300 000 to 18 750 each performed in triplicate, were stimulated with 10 μg of antigen/ml for 18 h of incubation at 37 °C, 5% CO₂. Spot-forming units were quantified using an automatic system ELR02 ELISPOT reader (Autoimmun Diagnostika GmbH, Strassberg, Germany) and analyzed with AiD3.1 S.R software.

A standard cytotoxic ⁵¹Cr-release assay was used as described previously (13). Briefly, HHD-EL4S3⁻Rob target cells (16) were mixed with peptide loaded splenocyte effector cells at effector:target cell ratios (E:T) of 100:1, 50:1, 25:1 and 12.5:1 in either triplicates or six replicates. ⁵¹Cr-release was measured using a microplate scintillation counter (Topcount, NXT; Packard, Boston, MA, USA). Spontaneous and total ⁵¹Cr-release was measured by adding culture medium or a detergent (Triton X-100) 2% v/v, respectively. The percentage of specific lysis was calculated as 100 × (experimental release − spontaneous release)/(total release − spontaneous release). Spontaneous and total release was in the range of 1200 and 9000 counts per minute, respectively.

To evaluate the individual immunogenicity of the minimal HIV-1-derived HLA-A*0201-binding CTL epitopes we immunized HLA-A*0201 tg mice with DCs pulsed with the individual CTL epitope. We used a standard specific cell killing assay (chromium release), confirmed by intracellular cytokine staining (ICS) and ELISpot assays (Table 1). We found that each of the seven CTL epitopes tested exhibited differential immunogenicity in vivo, with the two Gag epitopes Gag₄₃₃ and Gag₁₅₀^mod being the most immunogenic CTL epitopes. The immunogenicity was comparable with the HLA-A*0201 binding affinity where the two Gag peptides bind better than 100 nM and the remaining peptides bind in the intermediate 100 nM < IC₅₀ < 500 nM range (Table 1). For further in vivo competition experiments when immunizing multiple CTL epitopes simultaneously we ranked the seven CTL epitopes with respect to immunogenicity measured by functional lytic activity (Table 1).

We examined whether immune competition would occur and could be changed by presenting epitopes either separately or simultaneously on bone marrow-derived DCs prior to immunization in HLA-A*0201 transgenic mice. This was done by pulsing all seven peptides in one vial vs pulsing all seven peptides in seven individual vials of DCs as illustrated in Fig. 1A,B, respectively. These experiments were repeated three times. By simultaneous immunization, a hierarchy was established where Gag₁₅₀^mod and Gag₄₃₃ responses dominated (Fig. 2). Irrespective of whether the epitopes were pulsed separately before pooling, or pulsed together onto DCs, the resulting in vivo hierarchy was virtually unchanged as measured by cytolytic and ELISPOT assays. One representative experiment is shown in Fig. 2. These data were confirmed in another experiment using IFN-γ, TNF-α and IL-2 ICS (data not shown). In some experiments with individual pulsing or grouping according to the ranking shown in Table 1, up to three additional epitopes showed a weak response in some assays, as Vif₂₃ in Fig. 2B. However, these weaker responses were not consistent.

We then hypothesized whether we could overcome the immunodominance of Gag₄₃₃, Gag₁₅₀^mod and Vif₁₀₁, ranked 1–3 in Table 1, by homing DCs loaded with more dominant CTL epitopes and DCs loaded with less dominant CTL epitopes, to the same or different lymphoid compartments, respectively. We immunized dominant and less dominant CTL epitopes separately in the right and left flank of the mouse, respectively, as indicated in Fig. 1C,D. Administration of dominant and less dominant peptides in different lymphoid compartments did not increase the breadth or the magnitude of CTL responses in vivo (Fig. 2C,D). Again, the epitopes Gag₁₅₀^mod and Gag₄₃₃ dominated the responses in three individual experiments (Fig. 2).