Incorporation of Glycosylphosphatidylinositol-Anchored Granulocyte- MacrophageColony-Stimulating Factor or CD40 Ligand Enhances Immunogenicity of Chimeric Simian Immunodeficiency Virus-Like Particles

For enzyme-linked immunospot (ELISPOT) assays and enzyme-linked immunosorbent assays (ELISAs), the following peptide stimulants were used: (i) two peptides derived from SIVmac239 Env, Env amino acids (aa) 211 to 230 (CNTSVIQESCDKHYWDAIRF) and Env aa 231 to 250 (RYCAPPGYALLRCNDTNYSG), as Env MHC I peptide stimulants (at a final concentration of 1μ g/ml); (ii) RQIINTWHKVGKNVYL Env (aa 435 to 450) as an MHC II peptide (final concentration of 2.5 μg/ml); and (iii) SIVmac239 Env peptide pools, in which peptides are 15 amino acids in length with 11-amino-acid overlaps between sequential peptides. All peptides were obtained from the NIH AIDS Research and Reference Reagent program.

For in vitro cultures, soluble recombinant murine GM-CSF, IL-4, and CD40L were purchased from Peprotech (Rocky Hill, NJ) and used at 50 ng/ml (500 U/ml). The final concentration of VLPs in culture supernatants ranged from 1 to 2μ g/ml.

Plasmids containing cDNAs encoding GM-CSF with a GPI-anchoring domain of CD59 and LFA3 were described previously (36). These plasmids were digested with HindIII and ApaI (for CD59 GPI) or XbaI (for LFA3 GPI) to obtain DNA fragments of GM-CSF fused to the GPI-anchoring domain. After filling in with the Klenow polymerase fragment to obtain blunt ends, GM-CSF DNA fragments were ligated into the pSP72 vector with the T7 promoter, which was subsequently used to transform Escherichia coli DH5α cells. pSP72 constructs isolated from bacterial clones were screened by restriction enzymes, and the correctness of the GM-CSF DNA constructs was confirmed by DNA sequencing and protein expression in HeLa T4 cells using the T7 recombinant vaccinia virus expression system, as previously described (24). SmaI and XbaI DNA fragments of GM-CSF pSP72 plasmids were cloned into the baculovirus expression vector pc/pS1 and used to produce recombinant baculoviruses (rBV). GM-CSF containing pc/pS1 plasmids were transfected into Sf9 insect cells using the Baculo-Gold transfection kit (BD Pharmingen) by following the manufacturer's manual. Plaques of rBV were screened by their ability to express GM-CSF. The cDNA encoding mouse CD40L (obtained from Mark Feinberg, Emory University) was PCR amplified using the following primers: F-CD40L-SmaI, 5′-CCTTCCCGGGACCATGATAGAAACATACAGC-3′, and R-CD40L-XbaI, 5′-CTG CAG TCT AGA TCA GCG CAC TGT TCA G-3′ (underlining indicates restriction enzyme recognition sites). The SmaI- and XbaI-digested CD40L-encoding DNA fragment was cloned into the pSP72 vector, and CD40L pSP72 constructs were screened by restriction enzymes (SmaI and XbaI) and confirmed by DNA sequencing. Similarly, the DNA segment for CD40L from the pSP72 construct was cloned into the pc/pS1 rBV shuttle vector, and an rBV expressing CD40L was generated using a Baculo-Gold transfection kit. The virus titer was determined with a Fast Plax titration kit according to the manufacturer's instructions (Novagen, Madison, WI).

Sf9 insect cells were infected with rBV expressing SIV Gag (multiplicity of infection [MOI], 2) as a negative control, GM-CSF (MOI, 2) or CD40L (MOI, 2) and cultured in suspension. One million cells were harvested and stained for flow cytometry. For GM-CSF, a rat anti-GM-CSF (A2/F17-107) monoclonal antibody was mixed with rBV-infected cells at a final concentration of 5 μg/ml and incubated at 4°C for 30 min. As a secondary antibody, fluorescein isothiocyanate (FITC)-conjugated goat anti-rat immunoglobulin G (IgG) (Zymed) was incubated at a dilution of 50 in phosphate-buffered saline (PBS) with 2% fetal bovine serum (FBS) for 30 min at 4°C. For CD40L, phycoerythrin (PE)-conjugated hamster anti-mouse CD40L (BD-PharMingen) was used at a dilution of 200 in PBS with 2% FBS. After staining, cells were fixed in PBS with 1% paraformaldehyde and analyzed with a FACSCalibur instrument (Becton Dickinson) and WINMDI 2.8 software (Scripps Research Institute Cytometry Software).

SIV VLPs were produced using a modification of a previously described method (57). For Gag VLPs, Sf9 insect cells were infected with rBV expressing SIVmac239 Gag at a MOI of 2 and incubated at 27°C for 72 h. SIV VLPs containing SIV Env and Gag were produced from Sf9 cells coinfected with rBVs expressing SIV Gag and SIV Env (SIVmac239) at MOI ratios of 1:4. Chimeric SIV VLPs were produced from Sf9 cells coinfected with rBV expressing SIV Gag, SIV Env, and CD40L or GM-CSF at MOI ratios of 1:4:4. Three days postinfection, the culture supernatants were collected and centrifuged at 1,500 × g for 20 min and filtered through a 0.45-μm-pore size filter, and the VLPs were pelleted at 100,000 × g for 1 h at 4°C in a Beckman SW28 rotor. The pellets were resuspended in PBS at 4°C overnight and VLPs were further purified through a 20-35-60% discontinuous sucrose gradient at 28,000 rpm for 1 h at 4°C. The VLP bands were collected, washed with PBS, pelleted, and resuspended overnight in PBS. To quantitate the yield of purified VLPs, the protein concentration of each sample was estimated with the Bio-Rad protein assay (Bio-Rad Laboratories, Inc., Hercules, CA). For protein analysis, all samples were normalized to 1μ g/ml and loaded onto sodium dodecyl sulfate (SDS) polyacrylamide gels at the same concentration (5 or 10 μg), and SIV Gag and Env proteins, GM-CSF, and CD40L were probed using monkey anti-SIVmac239 serum (kindly provided by Silvija Staprans, Emory Vaccine Center), rabbit anti-mouse GM-CSF (Peprotech), and goat anti-mouse CD40L (Peprotech).

To analyze the quality and purity of VLP preparations, purified VLPs (5 μg) were applied to a Formvar carbon-coated grid at room temperature for 1 min. Excess VLP suspension was blotted with filter paper, and the grid was immediately stained with 1% uranyl acetate for 30 s. Excess stain was removed by filter paper, and the samples were dried and examined using a transmission electron microscope.

To determine whether the targeting molecules (GM-CSF, CD40L) are efficiently incorporated into the same VLPs as the viral Env protein, samples from each VLP group were immunoprecipitated with 1:50 or 1:100 diluted anti-GM-CSF or anti-CD40L antibodies, subjected to SDS-polyacrylamide gel electrophoresis (PAGE), and probed with anti-SIV antibody (1:5,000 in PBS-Tweenwith 1% skim milk). Other samples of the chimeric VLP group were immunoprecipitated with anti-SIV serum (1:100 or 1:500) and then probed with anti-GM-CSF or anti-CD40L antibodies (1:1,000 and 1:7,000) after being separated on SDS-PAGE gels.

To estimate the percentage of Env incorporation in VLPs, we used a sandwich ELISA; 96-well Nunc Maxisorb flat-bottom plates were coated overnight with a 1:1,000 dilution of SIVmac251 gp120 monoclonal antibody (KK46) (NIH AIDS Research and Reference Reagent Program). The VLPs were pretreated with 0.1% radioimmunoprecipitation assay buffer (1 M Tris buffer [pH 8.0], 5 M NaCl, 10% Triton, 10% sodium deoxycholate, 10% SDS) and added at a concentration of 400 ng/well. Goat anti-SIV gp120 (1:4,000) and rabbit anti-goat IgG horseradish peroxidase (HRP) (1:4,000) diluted in PBS plus 0.05% Tween 20 supplementedwith 2% bovine serum albumin were used as primary and secondary antibodies. Purified SIVmac239 gp130 was used to construct the standard curve to detect the Env concentration and was provided from the NIH AIDS Research and Reference Reagent Program (National Institutes of Health, Rockville, MD).

To estimate the percentage of incorporation of growth factors into the chimeric VLPs, we used a direct ELISA. VLPs were used to coat each well (5 μg per well) of a 96-well Nunc Maxisorb flat-bottom plate. For the quantitative determination of GM-CSF, rabbit anti-mouse GM-CSF (1:2,000) and goat anti-rabbit IgG coupled to HRP (1:2,000) were used. For the quantitation of CD40L, goat anti-mouse CD40L (1:3,000) and rabbit anti-goat IgG HRP (1:5,000) were used. For the standard curves, we used soluble recombinant murine GM-CSF and CD40L (Peprotech, Inc, Rocky Hill, NJ). O-Phenylenediamine (OPD) substrate tablets (Zymed, San Francisco, CA) dissolved in citrate buffer, pH 5.0, were used to develop color in all aforementioned assays. Optical density was read at 450 nm.

Bone marrow cells were prepared as described elsewhere (23). Bone marrow cells (1 × 10⁶) were labeled with CFSE (carboxyfluorescein diacetate, succinimidyl ester) (Molecular Probes, Eugene, OR) at a final concentration of 1 μM, and CFSE was quenched by further incubation in serum-containing medium and extensively washed in RPMI medium. To determine the effect of chimeric VLPs on cell proliferation, the CFSE-labeled bone marrow cells were cultured in RPMI medium in the presence of 1 μg/μl VLPs. Following incubation at 37°C in 5% CO₂ for 4 days, cells were harvested and analyzed by fluorescence-activated cell sorting. As a separate experiment to identify the phenotypes of cells, bone marrow cultures expanded in the presence of VLPs or recombinant GM-CSF (rGM-CSF) plus rIL-4 were stained with phycoerythrin (PE)-conjugated anti-CD11c and allophycocyanin (APC)-conjugated anti-CD11b for 20 min at 4°C, fixed in PBS with 1.5% paraformaldehyde, and analyzed with a FACSCalibur instrument (Becton Dickinson). Viable cells were counted by light microscopy after staining with trypan blue.

To measure antibody production, 2 × 10⁶ spleen cells were cultured in triplicate in 48-well round-bottom plates in 500 μl medium with or without VLPs or growth factors. VLPs were added at 1- or 2-μg/ml final concentrations. After incubation at 37°C in 5% CO₂ for 3 to 5 days, 100 μl of supernatant was collected at days 3, 4, and 5 for measurements of IgG, IgG1, IgG2a, IgG2b, IgG3, IgM, and IgA on ELISA plates coated with 4 μg/ml of Ig (heavy plus light chain) as described previously (22). To analyze the phenotypes of activated cells, spleen cells were cultured as described above and collected on day 4. One million cells were stained with PE-conjugated anti-CD69, peridinin chlorophyll protein (PerCP)-conjugated anti-B220, APC-conjugated anti-CD8, or FITC-conjugated anti-CD4 (eBioscience) for 20 min at 4°C. After staining, cells were washed and fixed with 1% paraformaldehyde and analyzed with a FACSCalibur instrument.

Female BALB/c mice (6 to 8 weeks of age, six mice per group) (Charles River Laboratory, Wilmington, MA) were immunized subcutaneously (s.c.) with VLPs or chimeric VLPs to assess immune responses. All immunizations were performed with one VLP preparation that met all the quality control requirements (Western blot, electron microscopy, coimmunoprecipitation, and quantitative ELISA of proteins incorporated). All animals received a priming immunization (s.c.) of VLPs (50 μg/dose) followed by two s.c. boosters with the same dose at weeks 4 and 8. At 2 weeks after each immunization, the animals were bled from the retro-orbital plexus, and the sera were used for the detection of SIV Env-specific antibodies with ELISAs and neutralization assays.

All sera were individually collected, and SIV Env-specific-antibody levels for IgG, IgG1, IgG2a, IgG2b, and IgG3 were quantitatively determined by ELISAs as previously described (22). The substrate OPD (Zymed, San Francisco, CA) dissolved in citrate buffer, pH 5.0, was used to develop color. Optical density was read at 450 nm, and antibody concentrations were determined based on standard curves of each subtype antibodies. Neutralization activity was determined using SMAGI cell assays as described previously (22). Briefly, preimmune and immune sera were heat inactivated at 56°C for 30 min, serially diluted, incubated with SIVmac1A11 virus (100 PFU) for 1 h at 37°C, and then added to the SMAGI cells.β -Galactosidase (β-Gal)-expressing blue foci indicating infectious spots were counted. Neutralization titers were expressed as reciprocal values of dilution factors giving 50% reduction ofβ -Gal-stained infected-cell foci compared to control wells without serum samples.

Spleens were collected from individual mice at 2 weeks after the final immunization, and a single-cell suspension was prepared and used for ELISPOT assays and cytokine ELISAs as described previously (22). Spleen cells or mesenteric lymph node cells (0.2 × 10⁶/200μ l complete RPMI medium) were prepared from immunized mice at 2 weeks after the last immunization and stimulated in vitro with Env peptide pools at a final concentration of 1 μg/ml in complete RPMI medium. After 72 h, the cells were centrifuged and the supernatant was collected and stored at −80°C until assayed. The ELISA reagents for IL-6, IL-10, IL-12, and tumor necrosis factor alpha were purchased from eBioscience (San Diego, CA), and those for gamma interferon (IFN-γ) and IL-4 were purchased from BD-PharMingen. Cytokine levels were determined according to the manufacturer's instructions. For ELISPOT assay, freshly isolated splenocytes (0.5 to 1.0 × 10⁶/200 μl complete RPMI) from immunized mice were cultured for 36 h in the presence of peptide stimulants in complete RPMI medium, as previously described (22). All ELISPOT reagents were purchased from BD-PharMingen.

Results are expressed as means ± standard errors of the means (SEM). Statistical comparisons were performed by a two-tailed paired test, and a P value of <0.05 was considered statistically significant.

We previously produced SIV VLPs by expression of the Gag and Env proteins in insect cells using rBVs as expression vectors (22, 57). Here, we designed membrane-anchored forms of GM-CSF to enable its incorporation into VLPs containing SIV Env and Gag as an approach to induce enhanced immune responses against SIV antigens. Since GM-CSF is a secreted protein, we used GPI-anchored forms of GM-CSF constructs (36) as shown in Fig. 1A. To anchor GM-CSF to VLPs, we used the GPI signal sequences from CD59 and LFA3. Since CD40L is normally expressed in a membrane-bound form, there was no need to attach any additional membrane-anchoring sequences. To express GM-CSF and CD40L in insect cells, we cloned these constructs into a baculovirus shuttle vector and generated rBVs. We then confirmed the cell surface expression of GPI-anchored GM-CSF fusion proteins and CD40L by infecting insect cells with these rBVs and analyzing them by flow cytometry. Both CD59- and LFA3-GPI anchored GM-CSFs (GM-CSF_CD59, GM-CSF_LFA3) as well as CD40L were found to be expressed on the cell surface (data not shown).

Chimeric SIV VLPs containing membrane-anchored GM-CSF or CD40L were produced by coinfecting insect cells with rBVs expressing SIV Env, Gag, and GM-CSF_CD59, GM-CSF_LFA3, or CD40L. We harvested VLPs from the culture supernatants and purified them using sucrose gradient ultracentrifugation. The purified VLP preparations were routinely tested for integrity and homogeneity by electron microscopy; the particles were about 90 to 100 nm in diameter (Fig. 2A). SIV Env proteins were found to be present at similar levels among various SIV VLP preparations (Fig. 2B), which were estimated to be 1.5 ± 0.2% of total VLP proteins by quantitative ELISA. We measured the incorporation of immunostimulatory molecules by Western blot analysis of purified VLPs using antibodies specific to GM-CSF or CD40L (Fig. 2C and D). GM-CSF_CD59 and GM-CSF_LFA3 were found to be incorporated into SIV VLPs at similar levels. Also, CD40L was incorporated efficiently into SIV VLPs. The levels of GM-CSF and CD40L were quantitated by ELISA and determined to be approximately 0.1% and 0.14% of total VLP proteins, respectively.

To determine whether viral proteins and cytokines are directly incorporated into the same VLP structures, we used coimmunoprecipitation assays. Chimeric SIV VLPs containing GM-CSF were immunoprecipitated with anti-GM-CSF antibody, and the proteins were probed with anti-SIV antibody after separation by SDS-PAGE. SIV Env and Gag proteins were found to be coprecipitated by these antibodies, indicating that GM-CSF and SIV antigens are present in the same VLP structures (Fig. 2E). Similarly, when chimeric SIV VLPs were first immunoprecipitated with anti-SIV antibody and the blots were subsequently probed with anti-GM-CSF, the growth factor was found to be coprecipitated (Fig. 2F). Analogous experiments confirmed the incorporation of Env, Gag, and CD40L into the same VLP structures (data not shown).

GM-CSF is a potent activator of hematopoietic progenitor cells and induces their differentiation and expansion into myeloid DC populations when supplemented with rIL-4 (25, 53). In order to determine whether GM-CSF incorporated into VLPs maintained this activity, we tested whether these VLPs could induce proliferation of bone marrow cells. We found that after cells were cultured for 4 days in the presence of 1 μg/ml SIV VLPs, GM-CSF_CD59 and GM-CSF_LFA3 anchored to SIV VLPs increased the overall numbers of bone marrow cells four- and threefold, respectively, compared to the medium control (Fig. 3A). In contrast, SIV VLPs did not induce a significant increase in number of viable cells compared to the medium control. We then labeled bone marrow cells with 1 μM CFSE and incubated them for 4 days in the presence of 1μ g/ml SIV VLPs or GM-CSF-anchored SIV VLPs. As controls, we set up analogous cultures with spleen cells. As expected, we observed massive expansion of bone marrow cells in the presence of GM-CSF incorporated into VLPs compared to the SIV VLP control (Fig. 3B). In contrast, neither VLP induced proliferation of spleen cells (data not shown). We then analyzed the expanded bone marrow cell cultures using flow cytometry for the presence of dendritic cells. We observed that bone marrow cells cultured in the presence of either GM-CSF_CD59 or GM-CSF_LFA3 anchored on SIV VLPs contained significantly higher numbers of CD11c⁺ CD11b⁺ myeloid DCs than cultures treated with control SIV VLPs (Fig. 3C). As expected, CD40L incorporated into VLPs had no effect on stimulation of proliferation of bone marrow cells (data not shown).

We also examined bone marrow and spleen cell cultures incubated in the presence of chimeric GM-CSF or CD40L VLPs, SIV VLPs, and controls such as recombinant soluble GM-CSF by light microscopy. We observed distinct morphological changes in bone marrow cultures, which were more pronounced after incubation with chimeric VLPs containing GM-CSF and were similar to those observed with rGM-CSF plus IL-4, indicating cell activation and differentiation into DCs (Fig. 4b, c, and d). SIV VLPs and CD40L SIV VLPs induced only minimal changes in the cell morphology of bone marrow cells (Fig. 4a). In contrast to the bone marrow cultures, we observed a very characteristic circular clustering of spleen cells when they were cultured with VLPs containing CD40L (Fig. 4h) and random amorphous clustering in the presence of rGM-CSF and IL-4 (Fig. 4f) or VLPs containing GM-CSF (Fig. 4g). These results suggest that although the target cell populations and the mechanisms of action of GM-CSF or CD40L are different, both molecules retain biological activities when incorporated into SIV VLPs.

Both GM-CSF and CD40L play a critical role in the activation of B cells. To determine whether GM-CSF- or CD40L-bearing chimeric VLPs are capable of activating B cells, we cultured splenocytes for 4 days in the presence of GM-CSF or CD40L anchored to VLPs or of soluble GM-CSF or CD40L. We then analyzed B-cell activation by checking the expression of the activation marker, CD69 (Fig. 5A). We found increased numbers of activated CD69⁺ B220⁺ cells when splenocytes were cultured with chimeric VLPs but not with soluble GM-CSF, CD40L, or SIV VLP (Fig. 5A). GM-CSF_LFA3 anchored to VLPs doubled the number of CD69⁺ B220⁺ cells (P = 0.0052) compared to the SIV VLP control. The addition of VLPs containing CD40L in the culture tripled the number of CD69⁺ B220⁺ cells compared to the addition of SIV VLPs (P= 0.0152) and significantly enhanced the numbers of double-positive CD69⁺ B220⁺ cells compared to the addition of GM-CSF VLPs (P = 0.0019 for VLPs containing GM-CSF_CD59 and P = 0.05 for GM-CSF_LFA3 VLPs). However, no significant increases in numbers of CD4⁺ CD69⁺ and CD8⁺ CD69⁺ T cells were observed (data not shown).

Since splenic B cells were activated by chimeric SIV VLPs containing either GM-CSF or CD40L, we then determined whether these VLPs could induce B cells to produce antibodies. Briefly, we cultured spleen cells for 5 days in the presence of VLPs, soluble GM-CSF, or CD40L. At days 4 and 5, we collected culture supernatants and analyzed them for Ig levels by ELISA. GM-CSF_CD59 or GM-CSF_LFA3 VLPs or CD40L VLPs stimulated the production of Ig subclasses compared to the SIV VLP controls (Fig. 5B and C). Only marginal differences in the IgG1 and IgG2a subclasses were observed when GM-CSF VLPs and CD40L VLPs were compared (Fig. 5B and C). The GM-CSF_CD59 VLPs enhanced production of IgG1 (Fig. 5B), whereas CD40L VLPs showed enhanced IgG2a levels at day 5 (Fig. 5C). The most pronounced effect of CD40L VLPs was observed in the total IgM levels, which were at least twofold higher than those induced by other VLPs (Fig. 5D). Neither soluble GM-CSF nor CD40L activated splenic B cells to secrete Igs. These results demonstrate that GM-CSF or CD40L incorporated into VLPs can activate splenic B cells to secrete antibodies specific to Env antigen on VLPs.

To investigate whether GM-CSF or CD40L incorporated into SIV VLPs can enhance humoral immune responses to the SIV Env protein, groups of mice were immunized s.c. with SIV VLPs (Env plus Gag), chimeric SIV VLPs (Env plus Gag plus GM-CSF or CD40L), or control Gag VLPs (Env-negative VLPs). We also included an additional control of SIV VLPs administered with soluble rGM-CSF. We used 10 ng soluble rGM-CSF because the VLP immunization dose of 50 μg contained approximately 10 ng of GPI-anchored GM-CSF. We measured serum levels of SIV Env-specific IgG at 2 weeks after each immunization (Fig. 6A) and the isotypes IgG1, IgG2a, IgG2b, and IgG3 after the last immunization by ELISA (Fig. 6B). The IgG levels induced by the chimeric GM-CSF_CD59 and GM-CSF_LFA3 VLPs were found to be 2.3-fold (P = 0.0104) and 2.8-fold (P = 0.0001) higher than those induced by SIV VLPs, 2 weeks after the last immunization. In contrast, CD40L incorporated into SIV VLPs did not induce a comparable enhancement of serum antibody responses. Compared with the groups that received a mixture of SIV VLPs and soluble GM-CSF, the chimeric GM-CSF SIV VLPs induced significantly higher levels of antibody responses (2.5- to 3.0-fold) (P = 0.03 for GM-CSF_CD59 and P = 0.0012 for GM-CSF_LFA3). Sera from mice that received the negative-control Gag VLPs exhibited minimal background titers (data not shown). In addition, the data show that incorporation of GM-CSF into VLPs did not alter the Th1-versus-Th2 profiles of the antibody responses; the SIV Env-specific IgG2a/IgG1 ratios were similar in all groups (Fig. 6B).

We were surprised that CD40L-containing VLPs did not induce strong IgG responses following immunization, because these VLPs had shown profound effects on in vitro cultures of splenocytes (Fig. 4 and 5). One possibility is that CD40L-containing VLPs fail to induce class switching from IgM to IgG. Therefore, we analyzed sera from immunized mice for SIV Env-specific IgM titers. We found that while the IgM levels in mice immunized with GM-CSF VLPs decreased after the second booster, they remained the same in CD40L VLP-immunized mice following the primary and secondary boosters (Fig. 6C). Taken together, these data suggest that chimeric CD40L VLPs fail to induce IgM-to-IgG class switching in Env-specific B cells.

Next, we assessed the neutralizing activity of the induced antibodies by determining the ability of serum to neutralize live SIV 1A11 virus. Sera from the SIV Gag VLP-immunized control mice showed very low neutralizing activity (titer < 10), similar to levels found in unimmunized mice (data not shown). Sera from mice immunized with SIV VLPs, coimmunized with SIV VLPs and soluble rGM-CSF, or immunized with chimeric CD40L VLP exhibited neutralizing antibody endpoint titers of 80. In contrast, sera from mice immunized with chimeric GM-CSF VLPs demonstrated a significant increase in neutralization activity, with an endpoint titer of 320 (Fig. 7). Based upon these results, we conclude that chimeric VLPs containing a membrane-anchored form of GM-CSFs are significantly more effective in inducing neutralizing antibodies than VLPs containing only the Gag and Env proteins, and that anchoring the adjuvant molecule to the VLP is important for the enhancement of this response.

To compare T-cell responses to VLPs, we measured cytokine production by ELISAs and ELISPOT assays as an indicator of cellular immune responses. Briefly, we isolated spleens from mice immunized with various VLPs and stimulated them with Env-specific MHC I- or MHC II-restricted peptides to quantitate Env-specific CD4 and CD8 cells secreting IL-2, IL-4, IFN-γ, IL-5, IL-6, IL-10, and IL-12 (Fig. 8). Splenocytes from GM-CSF- and CD40L-bearing chimeric VLP-immunized mice demonstrated a four- to eightfold increase of CD4 and CD8 T cells producing IL-4 compared to the SIV VLP group, with the highest numbers being observed in the GM-CSF_LFA3 group (Fig. 8A). GM-CSF SIV VLPs induce significantly higher levels of IFN-γ-secreting CD4 T cells than CD40L SIV VLP or control SIV VLPs (P = 0.0015 for GM-CSF_LFA3 and P = 0.05 for GM-CSF_CD59) (Fig. 8B). Both CD40L- and GM-CSF-bearing SIV VLPs induced significantly higher numbers of Env-specific, IFN-γ-secreting CD8 T cells than conventional SIV VLPs (P = 0.0016 for GM-CSF_LFA3, P= 0 for GM-CSF_CD59, and P = 0.0185 for CD40L) (Fig. 8B).

The number of CD8 T cells secreting IL-5 was dramatically increased in the GM-CSF_LFA3 VLPs and GM-CSF_CD59 VLPs groups compared to SIV VLP-immunized mice (Fig. 8C). For CD4 T-cell responses, only the GM-CSF_CD59 group was significantly higher than the SIV VLP and CD40L SIV groups. GM-CSF VLPs induced statistically significant differences in the levels of IL-6-secreting CD8 cells compared either to SIV VLP (P = 0.0008 for the GM-CSF_LFA3 group and P = 0.0003 for the GM-CSF_CD59 group) or to CD40L VLP (P = 0.0065 for GM-CSF_LFA3 and P = 0.0013 for GM-CSF_CD59) (Fig. 8D). In the case of IL-10, which is known to have a dual role as a regulatory cytokine and as a Th2-inducing cytokine (27), all VLP groups demonstrated a two- to sixfold increase in IL-10 production by CD8 upon stimulation with MHC class I-restricted peptides compared to the unimmunized group, but only the chimeric GM-CSF VLP groups exhibited a significant difference from the SIV VLP group (P = 0.0042 for GM-CSF_LFA3 and P = 0.0449 for GM-CSF_CD59) (Fig. 8E). Mice immunized with chimeric GM-CSF VLPs and CD40L VLPs showed similar increases in numbers of CD4 cells secreting IL-2 and IL-12. However, we observed significant differences in the CD8 cells secreting IL-2 between the GM-CSF VLPs and the SIV VLPs (P = 0.011 for GM-CSF_LFA3 and P = 0.0046 for GM-CSF_CD59 groups) (Fig. 8F). In contrast to the results obtained with GM-CSF VLPs, the CD40L VLP-immunized group did not show any effect on IL-6 or IL-2 production but exhibited a threefold increase in IL-12-producing CD8 T cells (P = 0.0008) (Fig. 8G).

Taken together, these data demonstrate that SIV VLPs are capable of inducing Th1- and Th2-type cytokine production. They also demonstrate a potent stimulatory effect of chimeric GM-CSF VLPs on the production of the Th2-type cytokines IL-4, IL-5, and IL-10 and the proinflammatory cytokine IL-6. In contrast, the CD40L VLP group was more potent in enhancing the secretion of IL-12, IFN-γ, and IL-10. Both GM-CSF and CD40L VLPs were highly effective in inducing IL-4-producing CD4 T cells, whereas CD40L VLPs were more efficient at inducing IFN-γ-producing CD8 T cells.