Priming with DNA Expressing Trimeric HIV V1V2 Alters the Immune Hierarchy Favoring the Development of V2-Specific Antibodies in Rhesus Macaques

Abstract

The RV144 vaccine trial revealed a correlation between reduced risk of HIV infection and the level of nonneutralizing-antibody (Ab) responses targeting specific epitopes in the second variable domain (V2) of the HIV gp120 envelope (Env) protein, suggesting this region as a target for vaccine development. To favor induction of V2-specific Abs, we developed a vaccine regimen that included priming with DNA expressing an HIV V1V2 trimeric scaffold immunogen followed by booster immunizations with a combination of DNA and protein in rhesus macaques. Priming vaccination with DNA expressing the HIV recombinant subtype CRF01_AE V1V2 scaffold induced higher and broader V2-specific Ab responses than vaccination with DNA expressing CRF01_AE gp145 Env. Abs recognizing the V2 peptide that was reported as a critical target in RV144 developed only after the priming immunization with V1V2 DNA. The V2-specific Abs showed several nonneutralizing Fc-mediated functions, including ADCP and C1q binding. Importantly, robust V2-specific Abs were maintained upon boosting with gp145 DNA and gp120 protein coimmunization. In conclusion, priming with DNA expressing the trimeric V1V2 scaffold alters the hierarchy of humoral immune responses to V2 region epitopes, providing a method for more efficient induction and maintenance of V2-specific Env Abs associated with reduced risk of HIV infection.IMPORTANCE The aim of this work was to design and test a vaccine regimen focusing the immune response on targets associated with infection prevention. We demonstrated that priming with a DNA vaccine expressing only the HIV Env V1V2 region induces Ab responses targeting the critical region in V2 associated with protection. This work shows that V1V2 scaffold DNA priming immunization provides a method to focus immune responses to the desired target region, in the absence of immune interference by other epitopes. This induced immune responses with improved recognition of epitopes important for protective immunity, namely, V2-specific humoral immune responses inversely correlating with HIV risk of infection in the RV144 trial.

Keywords: ADCC; ADCP; C1q; DNA vaccine; Env; HIV; NAb; V1V2; antibody; cyclic V2; gp145; linear peptide; prime-boost; rhesus macaque.

FIG 1

Development of DNA vaccine to target immune responses to V2 of HIV Env. (A) Design and expression of V1V2 scaffolds. The cartoons show the details of two different protein scaffolds (2F5K and 2J9C) containing the 80-aa region of V1V2_A244, the N-terminal tPA signal peptide, and the C-terminal FLAG-tag (left). The cartoon also depicts the FLAG-tagged V1V2-2J9C protein with the PDGFR transmembrane domain (V1V2-TM) used for in vitro studies (plasmid 447H) and the V1V2-2J9C DNA used for the vaccine studies, with the FLAG tag removed (plasmid 418H). HEK293T cells were transfected with V1V2_A244-2J9C-FLAG DNA and V1V2_A244-2F5K-FLAG DNA, and 2 days later, cells and supernatants were analyzed by Western blotting by loading 1/200 of each fraction, and the membranes were probed with anti-FLAG–HRP Ab. Lanes 1 and 6, Kaleidoscope prestained protein marker; lanes 2, 3, 7, and 8, V1V2_A244-2F5K; lanes 4, 5, 9, and 10, V1V2_A244-2J9C. Two independent DNA plasmid clones from each construct were tested (lanes 2 and 3, lanes 4 and 5, lanes 7 and 8, and lanes 9 and 10). Equal loading of the cell-associated fraction was controlled by probing the membrane with a β-actin Ab. (B) CM244 Env was expressed as cell-associated gp145 and as a soluble monomeric gp120 protein from the CMV promoter. The cartoon indicates the furin cleavage site and the TM region in gp145. HEK293T cells were transfected with two independent clones of each DNA. Two days later, cells and supernatants were analyzed by Western blotting by loading 1/200 of the fractions. The membrane was probed with a mixture of plasma from macaques vaccinated with HIV Env followed by anti-monkey IgG–HRP. The ECL detection method was used to visualize the expressed proteins. Lanes 1 and 6, Kaleidoscope prestained protein marker; lanes 2, 3, 7, and 8, gp120_CM244; lanes 4, 5, 9, and 10, gp145_CM244. (C) Expression levels of the endotoxin-free plasmid preparations used for macaque vaccination producing gp145 (plasmid 416H) or V1V2-2J9C (plasmid 418H) after 24 h upon transfection in HEK293T cells. Twofold serial dilutions (1:2 to 1:16) of total proteins were analyzed, and the membrane was probed with the V2-specific CH58 MAb. Lanes 1 and 7, Kaleidoscope prestained protein marker; 2 to 6, gp145; lanes 8 to 12, V1V2-2J9C. (D and E) Stable HEK293H cell lines were generated expressing V1V2_A244-2J9C anchored to the cell membrane via PDGFR-TM (D) and membrane-anchored trimeric gp145_CM244 (E). Histogram overlays show the binding of selected V2-specific MAbs (PG9, 697-30D, CH58, CH59, and PGT145) to V1V2_A244 (red histograms) (D) and membrane-anchored trimeric gp145 (black histograms) (E). Gray-shaded histograms show the binding of the anti-p24^Gag MAb 241-D, used as a negative control. The results show recognition by (i) PG9 (V2q-type MAb), supporting the quaternary structure of the trimeric V1V2 in a glycan-dependent fashion; (ii) 697-30D (V2i-type MAb), recognizing a discontinuous epitope overlapping the α₄β₇ integrin binding motif (LDI/V); and (iii) CH58 and CH59 (V2p-type MAbs), isolated from RV144 trial volunteers known to target linear and cyclic peptides spanning aa 170 to 176 of V2.

FIG 2

Induction of V1V2-specific Abs by priming vaccination with DNA expressing secreted V1V2_A244 scaffold. (A) Schematic representation of the vaccination regimens, including two DNA primes (months 0 and 1) and three DNA-protein coimmunization boosts (months 2, 3, and 6) in two groups of macaques (n = 4). As the prime, the V1V2 DNA group received V1V2_A244-2J9C DNA and the gp145 DNA group received gp145_CM244 DNA. For boosts, both groups received gp145_CM244 DNA, and the V1V2 DNA group continued to receive V1V2_A244-2J9C DNA. The animals also received gp120_CM244 protein adjuvanted with GLA-SE as part of the boost. Ab responses were analyzed after the prime and after the boost, as indicated. (B) Endpoint titers of the vaccine-induced plasma Ab to gp120_CM244 after the prime, as measured by ELISA. The P value is from a parametric t test. (C) Immunoprecipitation, performed under nondenaturing conditions with HEK293-produced FLAG-tagged V1V2_A244-2J9C protein and macaque plasma samples collected after administration of two doses of each of the DNA primes. The antigen-Ab complexes that had bound to protein A-Sepharose beads were analyzed on 12% denaturing gels. Protein A-Sepharose-bound (50% of sample) and unbound (25% of sample) materials were loaded onto gels, and the membranes were probed with anti-FLAG–HRP Ab. Bands labeled “bound” reveal the presence of V1V2-2J9C-FLAG in antigen-Ab complexes, whereas the bands labeled “unbound” show the absence of Abs in plasma specific for V1V2-2J9C-FLAG. Lanes 1 to 4, V1V2 DNA group; lanes 5 to 8, gp145 DNA group. (D to F) Histogram overlays show flow cytometry analysis of plasma Ab binding to membrane-anchored Env. Binding of plasma from the V1V2 DNA group (top) and the gp145 DNA group (bottom) to the stable HEK293H cell line expressing membrane-anchored V1V2-TM (D), gp145_CM244 (E) and gp145_CH505 (F) is shown. Data from individual animals from the V1V2 DNA (red histograms) and gp145 DNA (dark gray histograms) vaccine groups are shown. Binding of prebleed plasma from the corresponding animals is also shown (light gray). Macaque Ab binding was measured using a PE-conjugated anti-human IgG MAb cross-reactive with macaque IgG.

FIG 3

V1V2 DNA priming induced higher V2-specific Abs and gp120-specific Abs. (A) Plasma Ab binding (1:50 dilution) to cyclic V2 was measured by an SPR assay after the prime. Reactivity to cyclic V2 peptides from different clades (AE.92TH023 and C.1086) are reported as response units. Individual animals from the V1V2 (red symbols) and gp145 (black symbols) DNA-primed groups are shown. P values are from a parametric t test. (B and C) A Luminex assay shows Ab binding to cyclic V2 from different clades (AE.92TH023 and C.1086) (B) and to gp120 from different clades (A.MG505, B.YU2, and C.ZM53) (C) using serially diluted serum from the immunized macaques. Data are given as mean fluorescence intensity (MFI) for the V1V2 DNA group (red symbols) and the gp145 DNA group (black symbols) after the prime (top) and after the boost (middle [B] and bottom [C]). Values for prevaccination serum samples are shown (bottom [B]). The mean area-under-the-curve (AUC) values for the 2 vaccine groups (red and black lettering, respectively) after prime and after boost are shown. P values are from a parametric t test.

FIG 4

Distinct epitope recognition in V2 by Abs elicited by the two vaccine regimens. (A) ELISA analysis using linear peptides (20-mers overlapping by 14 aa) spanning the V2 region (amino acid numbering follows CM244 [GenBank accession number AAW57760]) was performed using a 1:50 plasma dilution after the DNA prime for each animal from both groups. (B) Alignment of V1V2 amino acid sequences of A244, ZM53, and ZM109. The five strands (A, B, C, C’, and D, and the helical turn K) composing the V1V2 beta-barrel are indicated by gray shading (10). Residue K169 (HXB2 numbering) was identified by sieve analysis in RV144 as mismatched in vaccine breakthrough infections (3). The peptide from aa 165 to 178 (LRDKKQKVHALFYK) is underlined and spans the residues shown to interact with Abs associated with reduced risk of infection from the RV144 trial (14). (C) Binding of plasma (1:50 dilution) from the immunized macaques to the wild-type V2 peptide and to its mutant (K169V) was measured by ELISA (HXB2 amino acid numbering). Individual plasma samples from V1V2 DNA animals were analyzed after the prime.

FIG 5

Distinct epitope recognition in V2 is maintained after the booster vaccination. (A) Ab binding to selected peptides was measured in a Luminex assay using serially diluted sera. Data obtained after the second prime (top) and after the last boost (bottom) from the V1V2 DNA group (red symbols) and gp145 DNA group (black symbols) are shown as MFI. The mean AUC values are given: red lettering indicates the V1V2 DNA group, and black lettering indicates the gp145 DNA group. P values are from a parametric t test. (B) ELISA analysis using linear peptides (20-mers overlapping by 14 aa) spanning the V2 region was performed as for Fig. 4A using a 1:50 plasma dilution of plasma collected after the DNA-protein booster vaccination from animals from both groups.

FIG 6

Distinct effector function of the Ab induced by the two vaccine regimens. (A) Neutralizing Abs (Nab) to tier 1A pseudovirus 92TH023.6 (left) and the mutant 92TH023.6N160K (right) were measured in serum from all the animals 2 weeks after the boost. Amino acid numbering is that for HXB2, and the location of N160 is indicated in Fig. 4B. The dashed line indicates the limit of detection of neutralization in the TZM-bl assay. (B) ADCC obtained with gp120_A244-coated CEM-NKR target cells was measured in sera from V1V2-primed animals (red symbols) and gp145-primed animals (black symbols) after the last booster vaccinations. (C) Phagocytosis of gp140_1086c-coated fluorescent beads by THP-1 cells triggered by antibodies induced after the prime in sera from V1V2 DNA group (red symbols) and the gp145 DNA group (black symbols). The fold differences were determined by dividing the ADCP score from immunized sera by the score from the respective prebled sera. P values are from a nonparametric t test. (D) C1q binding to V2-specific Abs was measured in the two vaccine groups after the prime and the boost, by ELISA using plates coated with cyclic V2_1086c. P values are from a nonparametric t test.