Systematic comparison of HIV-1 Envelope-specific IgG responses induced by different vaccination regimens: Can we steer IgG recognition towards regions of viral vulnerability?

Abstract

Immunogens and vaccination regimens can influence patterns of immune-epitope recognition, steering them towards or away from epitopes of potential viral vulnerability. HIV-1 envelope (Env)-specific antibodies targeting variable region 2 (V2) or 3 (V3) correlated with protection during the RV144 trial, however, it was suggested that the immunodominant V3 region might divert antibody responses away from other relevant sites. We mapped IgG responses against linear Env epitopes in five clinical HIV vaccine trials, revealing a specific pattern of Env targeting for each regimen. Notable V2 responses were only induced in trials administering CRF01_AE based immunogens, but targeting of V3 was seen in all trials, with the soluble, trimeric CN54gp140 protein eliciting robust V3 recognition. Strong V3 targeting was linked to greater overall response, increased number of total recognised antigenic regions, and where present, stronger V2 recognition. Hence, strong induction of V3-specific antibodies did not negatively impact the targeting of other linear epitopes in this study, suggesting that the induction of antibodies against V3 and other regions of potential viral vulnerability need not be necessarily mutually exclusive.

Keywords: CN54rgp140 vaccine; HIV; V2-antibodies; V3-antibodies; envelope-specific antibodies; immunogen sequence; linear peptide array; vaccine.

Figure 1

Vaccination schedules. Vaccination schedules and immunogens of the eight vaccination groups from five different HIV vaccine trials analysed herein. For each vaccination group the molecular forms as well as the HIV-1 clade of the Env immunogens, their delivery form, and time point of administration are stated. The immunogens included: gp120 monomers, gp140 soluble trimers, open trimeric structures, membrane anchored gp145, ‘native-like’ gp150 including the transmembrane region and native trimeric gp160. Immunogens were either administered as adjuvanted proteins or expressed in vivo using DNA, MVA, CP and/or Ad5. Most groups received prime-boost vaccine regimens including multiple sequence variants of the Env (RV144, TMV02, RV172), while some included only CN54 derived immunogens (UK003, X001). The colour coding of the surrounding frame for each vaccination group is kept consistently in this article. Viral Vectors: MVA, Modified Vaccinia Ankara; CP, Canary Pox; Ad5, Adenovirus 5; Proteins: CM244, CRF01_AE strain; 1-MN = B-strain, CN54 = C-strain.

Figure 2

Analysis of IgG epitope recognition along the HIV-1 Env protein in the 8 different vaccine groups. The frequency of responders (FOR; upper panel) and the mean fluorescence intensities (mean FI; lower panel) plotted against individual antigenic regions along the entire HIV-1 Env as included in the 10 full-length Env immunogen sequences comprising the array backbone. Each row of the respective heat maps displays the Env-specific IgG responses of one of 8 vaccination groups, tested four weeks after the last vaccination. IgG responses against individual antigenic regions were considered positive if the corresponding FI was above 3,500 after subtraction of the pre-vaccination value. The mean FI was calculated using the maximum FI values per position for each participant, only if peptide-specific IgG responses occurred in at least 25% of the vaccinees. IDRs 1–4 are indicated by red lines and are listed in Table 1 .

Figure 3

Statistical comparison of IgG responses targeting single peptide IDRs. Graphs (A-D) each depict the statistical comparison of the FI values of one representative peptide of the 4 IDRs. Each symbol indicates the maximum FI value of one single study participant. Values after subtraction of the baseline are shown. The cut-off for positive signals is indicated by a dotted line. P-values were calculated using a Mann-Whitey-U test. P-values of comparisons between groups with or without CN54gp140 protein boosts are shown. Triangular symbols indicate that the CN54gp140 protein was part of the study’s vaccination schedule, round symbols represent vaccinees that did not receive the CN54gp140 protein.

Figure 4

Targeting of IDR2a_V2 and IDR3c_V3 peptide variants by vaccine-induced antibodies. The mean magnitude of antibody responses against the respective peptide variant was calculated per group if positive responses occurred in >25% of vaccinees above background (3500 FI) after baseline subtraction. Mean FI values for each vaccination group (yellow to red) are illustrated as a heat map in the context of their frequency of occurrence in the HIV database (green coding on the left), and their clade representation (purple coding on the right). Green colour-coding symbolises the frequency of the respective peptide in the Los Alamos (www.hiv.lanl.gov) database representing the current global HIV epidemic, varying from grey (low) to green (high) according to the prevalence of the peptide. Red colour coding represents the magnitude of the IgG response towards each given peptide. The distribution of occurrences of a peptide variant within HIV-1 clades as a rounded fraction is depicted in purple. (A) Heat map of 18 peptide variants corresponding to the HXB163_TGMIDKMKEEYALFY V2 position. The CRF01_AE immunogen sequence is highlighted in grey. (B) 22 peptide variants were included for the V3 tip region (HXB304_RKSIRIGPGSTFYAT). Additional peptide variants were included in the peptide microarray to cover a broad range of variants and responses in these specific Env areas of interest.

Figure 5

Impact of V3 detection on V2 tip and total Env reactivity. Scatter plots depict the relationship between the strength of V3 IgG recognition and (A) V2 detection, (B) the total number of peptides detected, and (C) overall Env detection. Data were Z normalised (mean of 0, standard deviation of 1) to allow comparison of all trials. A Spearman correlation analysis was used to calculate the statistical relationship. (A) Correlation of the intensity of IDR3c_V3 IgG targeting with IDR2a_V2 recognition. For the scaled data, the r-value was: 0.475, the p-value was: 0.008 and the 95% CI calculated by bootstrapping was 0.1273 to 0.7187. V2_IDR2a and V3_IDR3c are represented by one single peptide each. Only data of study participants from studies with both responses (RV144 and TMV02) are plotted. (B) Correlation of the FI values of the whole IDR3_V3 with the total number of Env peptides detected excluding the IDR3_V3. The r-value was: 0.3024, the p-value was: 0.0088 and the 95% CI calculated by bootstrapping was 0.07250 to 0.5018. (C) Correlation of the FI values of the whole IDR3_V3 with the total Env detection strength bar the IDR3_V3. The observed r-value was: 0.3288, the p-value was: 0.0042 and the 95% CI calculated by bootstrapping was 0.1016 to 0.5234. In (B) and (C) the max FI values for the three IDRs belonging to V3 (IDR3a-c) were summed up per patient. The individual studies are differentiated by colour, with each symbol representing one participant. The line indicates the linear fit of the data. Participants whose regimen included CN54gp140 are symbolised by triangles, all others by dots.