Selection and immune recognition of HIV-1 MPER mimotopes

Abstract

The membrane proximal external region (MPER) of HIV-1 gp41 is targeted by several neutralizing antibodies (NAbs) and is of interest for vaccine design. In this study, we identified novel MPER peptide mimotopes and evaluated their reactivity with HIV + plasma antibodies to characterize the diversity of the immune responses to MPER during natural infection. We utilized phage display technology to generate novel mimotopes that fit antigen-binding sites of MPER NAbs 4E10, 2F5 and Z13. Plasma antibodies from 10 HIV + patients were mapped by phage immunoprecipitation, to identify unique patient MPER binding profiles that were distinct from, and overlapping with, those of MPER NAbs. 4E10 mimotope binding profiles correlated with plasma neutralization of HIV-2/HIV-1 MPER chimeric virus, and with overall plasma neutralization breadth and potency. When administered as vaccines, 4E10 mimotopes elicited low titer NAb responses in mice. HIV mimotopes may be useful for detailed analysis of plasma antibody specificity.

1. Introduction

Neutralizing antibodies (NAbs) can prevent human immunodeficiency virus (HIV-1) infection by binding to the HIV-1 envelope (Env) and inhibiting structural changes required for Env fusion with the target cell (Sun et al., 2008). The dynamic process that occurs during cellular infection requires significant structural rearrangement of the Env spike, a gp120 and gp41 heterotrimer. After gp120 binding to cellular CD4 receptor and then coreceptor, a series of conformational changes occur in gp41 (Wilen et al., 2012). The gp41 fusion peptide inserts into the host cell membrane, extending gp41 into the pre-hairpin intermediate state. Next, the N- and C-heptad repeats collapse, bringing the gp41 fusion peptide and the membrane proximal external region (MPER), associated with the host and viral membranes respectively, in proximity for fusion (Khasnis et al., 2016). The stable, post-fusion gp41 six-helical bundle remains on the surface of the infected cell. The different conformations of gp41 have potentially distinct antigenic and immunogenic properties that require further characterization (Pancera et al., 2014).

The MPER of gp41 is a major neutralizing determinant of HIV-1 Env and of interest for vaccine design. Anti-gp41 antibodies arise within two weeks of HIV-1 infection but initially target the stable, post-fusion six-helical bundle and are non-neutralizing as cellular infection has already occurred (Frey et al., 2010). MPER-specific antibodies with either neutralizing or non-neutralizing functions can develop in a subset of individuals with different studies reporting frequencies between < 1% and 66%, depending on the cohort and detection method used (Molinos-Albert et al., 2017; Landais et al., 2016; Walker et al., 2010; Gray et al., 2007, 2009a, 2009b, 2011; Li et al., 2009; Tomaras et al., 2011; Binley et al., 2008). Low accessibility of the MPER epitope, proximity to the membrane and lipid cross-reactivity are thought to limit generation of MPER-specific antibodies during natural infection. In fact, NAbs targeting MPER are observed less frequently in HIV + plasma than NAbs with other Env specificities, such as gp120 V3 (Landais et al., 2016; Gray et al., 2009a). Consequently, fewer MPER-specific NAbs have been isolated as compared to other Env specificities, limiting our understanding of MPER complexity and structure of functional, pre-fusion epitopes.

HIV-1 vaccines have failed to elicit potent MPER-specific NAbs, potentially due to the inability of immunogens to mimic the neutralization-competent, pre-fusion structure of gp41. Recombinant MPER proteins and peptide formulations can elicit MPER-specific antibodies but they are non-neutralizing, indicating the requirement for neutralization exceeds amino acid sequence recognition (Molinos-Albert et al., 2017). Attempts have been made to stabilize the pre-fusion form of gp41 for use as an immunogen, however further definition of relevant structures that mimic NAb-recognized form of MPER may be required for success in this approach (Frey et al., 2008, 2010; Kim et al., 2007; Dennison et al., 2011).

Defining the prevalence and range of MPER-specific antibodies elicited during natural infection and identifying stable neutralization-competent structures of MPER may improve vaccine design. In this study, we utilize phage display technology to identify novel peptides, or mimotopes, recognized by MPER NAbs 4E10, 2F5 and Z13. Mimotopes are unique peptides recognized by a target antibody and can represent linear or conformational epitopes, as well as non-protein molecules such as lipid. MPER mimotopes identified in this study were used to map MPER epitope-specificities of plasma IgG from 10 chronically, subtype B HIV-1 infected individuals. HIV-infected individuals were found to have unique humoral immune responses to MPER, with IgG specificities that overlapped with, but were distinct from, 4E10 and 2F5 NAbs. Neutralization inhibition assays were used to down-select mimotopes that were capable of reducing plasma MPER-specific neutralization. Immunogenicity of selected MPER mimotopes was evaluated in conjunction with a gp145 Env protein; HIV-1 specific immune responses were elicited with MPER mimotope boosting incrementally improving the neutralizing activity of a gp145 Env subunit candidate vaccine.

2. Materials and methods

Nab epitope selection by biopanning:

Biopanning experiments were performed in 96-well microtiter Immunol B polystyrene plates using two phage display random peptide libraries, Ph.D.12 and Ph.D.c7c (New England Biolabs, Ipswich, MA), following manufacture’s instructions. Nabs 4E10, 2F5 or Z13 (150 μl at 20 μg/ml in 0.1 M NaHCO₃) were incubated overnight at 4 °C with gentle agitation in a humidified container. Unbound Nab was removed; wells were blocked with 300 μl of 10% nonfat milk at 4 °C for 1 h. Wells were washed six times with 300 μl of 0.1% PBST (PBS containing 0.1% Tween-20), phage library was added (1 × 10¹¹ M13/well in 0.1% PBST), and incubated at room temperature with gentle agitation for 5 min. Unbound phage were removed by washing ten times with 300 μl 0.1% PBST then bound M13 were eluted with 100 μl of 0.2 M Glycine-HCl (pH 2.2). Eluted phage were amplified and titered on E. coli then used as input for additional rounds of biopanning. For biopanning rounds two and three, 0.5% PBST (PBS containing 0.5% Tween-20) was used to wash plates. After three rounds of biopanning, individual phage were expanded as clones and sequenced. Selection of 4E10-binding epitopes was also performed under competition with HIV-1 (17.5 ng/ml 873) or sCD4-bound HIV-1 (17.5 ng/ml 873 + 20 μg/ml sCD4), by incubated with target Nab for 30 min before direct addition of M13.

M13 Binding ELISA:

Microtiter plates were coated with Nab (1 μg/well), incubated overnight at 4 °C, then washed three times with 0.05% PBST. Plates were blocked with 10% non-fat milk for 1 h at 4 °C, then washed twice. M13 was titered in 2-fold serial dilutions starting at 1 × 10¹¹ M13/ml in blocking buffer and added to the plate. Plates were incubated at room temperature for 1 h, then washed three times. HRP-labeled anti-M13 Ab (Pharmacia, Whitehouse Station, NJ) diluted to 1:5000 in blocking buffer was added, plates were incubated for 30 min at room temperature, then washed three times. TMB (Seracare, Milford, MA) was added, incubated for 15 min at 37 °C, then the reaction was stopped by adding of 1 M phosphoric acid. Plates were read on a spectrophotometer at 450 nm.

Peptide Competition ELISA:

The M13 Binding ELISA was modified to include peptide competition, including MPER peptide (LELDKWASLWNWFNITNWLWYIK(amide)) and MPER scrambled peptide (LSINEAFKWLDWWTLNDLWYIWK(amide)). Assay plates were coated with specific Nab and blocked as described for the M13 Binding ELISA. A single concentration of M13 (7.5 × 10⁸ M13/ml) and titered peptide (5-fold serial dilutions starting at 100 μg/ml) were added concurrently to each well. Samples were incubated at room temperature for 1 h; the assay continued as described for the M13 Binding ELISA.

HIV-1 Capture Competition Assay:

Reacti-Bind Protein A/G Coated 96-well plates (Thermo Fisher Scientific, Waltham, MA) were wash three times with 0.05% PBST, then 1 μg Nab diluted in SuperBlock Blocking Buffer (Thermo Fisher Scientific, Waltham, MA) was added. Plates were incubated at room temperature for 1 h, then washed three times. M13 was titered in 2-fold serial dilutions starting at 4 × 10¹¹ M13/ml in RPMI-1640 (Quality Biologicals, Inc, Gaithersburg, MD); HIV-1 pseudoviruses were diluted in RPMI-1640 to 35 ng/ml p24 protein. Equal volumes of M13 and HIV were added concurrently to each well; plates were incubated at 37 °C for 1 h, then washed six times with RPMI-1640. Disruption buffer (Advanced Bioscience Laboratories, Rockville, MD) was added to each well and bound pseudoviruses were quantitated by p24 antigen capture ELISA (Advanced Bioscience Laboratories, Rockville, MD), as per manufactures protocol. Assays measuring the effect of sCD4 binding on HIV-1 capture were performed using HIV-1 stocks that were pre-incubated with 10 μg/ml sCD4 at 37 °C for 1 h before titration and addition.

Immunoprecipitation and immunoblot:

Protein A beads (Invitrogen, Carlsbad, CA) were washed by resuspending 50 μl beads in 0.05% PBST, then liquid was removed after 1 min on a magnetic rack. The beads were resuspended in 0.02% PBST (PBS containing 0.02% Tween-20) containing 1.5 μl of filtered undiluted plasma, then mixed by rotation at room temperature for 30 min. Liquid was removed and the beads were resuspended in 10% nonfat milk containing phage at 1 × 10¹⁰ M13/ml, then were mixed by rotation at room temperature for 30 min. The beads were washed seven times with 1.5% PBST (PBS containing 1.5% Tween-20), then were resuspended in 20 μl elution buffer (0.2 M Glycine-HCl, pH2.2) and incubated at room temperature for 5 min. Liquid was collected, samples were lysed and subjected to gel electrophoresis using a SDS, reducing gel (Invitrogen, Carlsbad, CA) followed by transfer to a PVDF 0.2 μm membrane (Invitrogen, Carlsbad, CA). M13 capsid protein was detected on the membrane using murine primary M13 g8p Ab (AbCam, Cambridge, UK) and secondary IRdye conjugated goat anti-mouse IgG Ab (LI-COR Biosciences, Lincoln, NE). The membrane was scanned using an Odyssey infrared imaging system (LI-COR Biosciences, Lincoln, NE) and bands were normalized and quantified using protein standards.

Immunogen preparation:

Antigens, M13 phage and gp145 protein, were encapsulated in liposome prior to immunization. Liposomes composed of dimyristoyl phosphatidylcholine, dimyristoyl phosphatidylglycerol and cholesterol in molar ratios of 1.8:0.2:1.5 were prepared by dispersion of lyophilized mixtures of lipids at a phospholipid concentration of 50 mM in Dulbecco’s PBS with 0.4 g/ml lipid A, either with or without antigen. Liposomes were washed twice in sterile saline to remove the unencapsulated antigen.

Animal Immunizations:

BALB/C mice (five mice per group) were immunized by the intramuscular route on weeks 0, 3, 6 and 8 with 50 μl of the following immunogens at 5 × 10¹¹ phage or 10 μg gp145 protein each per dose: M13 4E10 12D4, M13 4E10 12B7, M13 4E10 (c7c4, 12B6, 12B7, 12D1, 12D4), gp145, M13 4E10 (c7c4, 12B6, 12B7, 12D1, 12D4) + gp145, M13 no insert or unencapsulated liposomes. Blood was collected at two-week intervals starting two weeks prior to the first immunization ending on week 10 when the animals were euthanized to collect spleens and lymph nodes.

HIV-1 Neutralization Assays:

TZMbl (Emmer et al., 2016) and PBMC (Matyas et al., 2009; Edmonds et al., 2010) neutralization assays were performed as previously described for serially diluted monoclonal antibody, or HIV + or vaccinee sera, with inhibition of HIV infection detected as a reduction in firefly or renilla luciferase RLU, respectively. For both assay platforms, the reciprocal titers were derived through the performance of two independent assays for which the IC₅₀ or ID₅₀ values were within 5-fold range. PSVs were produced with a pSG3Δenv DNA plasmid encoding the HIV backbone genes and a plasmid encoding the env gene, including CO6980.v0.c22 (subtype C), 20,635–4 (subtype C), 93MW965 (subtype C), GS015 (subtype C), BZ167 (subtype B), MN (subtype B), SF162 (subtype B), 93UG065 (subtype D), TH023 (CRF01_AE), 55,815 (CRF02_AG), and MuLV (murine leukemia virus; nonspecific control). HIV-2 and HIV-2/MPER chimera viruses (kindly provided by Dr. George Shaw) were used to quantify MPER-specific neutralizing activity in sera. Using the HIV-1/MPER chimera virus, the TZMbl neutralization assay was modified to quantify neutralizing titer of HIV + sera in the presence of M13-displayed MPER epitopes. Sera was diluted to 2 × ID₅₀, or serially diluted and mixed with an equal volume of M13 at 5 × 10¹⁰ M13/ml and incubated for 30 min before the addition of virus. Subsequent steps follow the standard TZMbl neutralization assay method.

Intracellular cytokine staining flow cytometry assay:

Cells from spleens or lymph nodes were stimulated with 5 μg/ml acute C gp145 (HIV-1 C06980, Advanced Bioscience Laboratories, Kensington, MD), gp140 (HIV-1 IIIB, Advanced Bioscience Laboratories, Kensington, MD), yeast-derived gp41 (Meridian Biosciencees, Cincinnati, OH) or 10 μg/ml cathepsin degraded, yeast-derived gp41 or ConA as the positive control for 22 h at 37 °C. Brefeldin A (1 mg/ml, Sigma-Aldrich, St. Louis, MO) and monensin (0.07 mg/ml, Becton Dickinson, Franklin Lakes, NJ) were added and cells were incubated for 20 h. Cells were analyzed on a LSR II flow cytometer (Becton Dickinson, Franklin Lakes, NJ) and 500,000 events were collected using FACSDiva software (Becton Dickinson, Franklin Lakes, NJ). Dead cells were excluded using a viability marker and B-cells were excluded. The CD3⁺ CD4⁺ and the CD3⁺ CD8⁺ T-cells were gated and analyzed for the expression of, IL-2, TNF-a, IFN-g and CD107a. The data were analyzed using FlowJo software (Tree Star, Ashland, OR).

Antigen-specific serum IgG ELISA:

Antigen specific IgG titers were determined by binding ELISA titrations using gp145 and gp41 proteins as targets. Antigens were diluted to 0.25 μg/ml in PBS (pH 7.4), 100 μl/well was added to 96-well microtiter Immunol 2 polystyrene plates. Plates were incubated overnight at 4 °C then washed three times with 300 μl 0.1% PBST. Serum was titered in 2-fold serial dilutions starting at 1:50 dilution in serum diluent (0.1% PBST containing 5% non-fat milk), and 100 μl/well each dilution was added to the plate. Plates were incubated at 37 °C for 1 h then washed three times with wash buffer. HRP-labeled anti-mouse IgG antibody diluted to 1:16,000 in serum diluent was added, 100 μl/well. Plates were incubated for 1 h at 37 °C then washed three times with wash buffer. TMB (100 μl/well; KPL, Gaithersburg, MD) was added, incubated for 30 min at 37 °C and the reaction stopped by adding 100 μl/well of 1 M phosphoric acid. Plates were read on a spectrophotometer at 410 nm, 570 nm reference filter. Antigen binding titer was determined by calculating the concentration at which binding was detectable above three times background. Two independent assays were performed and the results were averaged.

3. Results

Selection of MPER mimotopes.

Phage display biopanning was used to identify novel mimotopes that bound specifically to MPER NAbs 4E10, 2F5 and Z13. Two M13-displayed, random peptide libraries, expressing linear dodecapeptides (PhD 12) or disulfide-constrained heptapeptides (PhD c7c) fused to the bacteriophage minor coat protein p3, were utilized for mimotope selection. Solid phage biopanning was performed by direct binding of phage to the target NAb 4E10, 2F5 or Z13. Phage recovery increased over three progressively more stringent rounds of biopanning (Supplementary Table 1), with between 4 and 23 unique peptide sequences identified for each condition (Supplementary Table 2). 2F5 and Z13 mimotopes identified three amino acid motifs DKW and NxxDxT with 91% and 98% of the peptides including those sequences, respectively. 4E10 mimotopes identified a six amino acid motif, N/SWFDxS/TxxL, however less MPER homology was observed for the 4E10 mimotopes (amino acid logos are shown in Fig. 1). 25% of the linear 4E10 mimotopes selected by direct binding contained only two MPER homologous amino acids and may represent structural mimics of the epitope. Conserved motifs, other than the core MPER epitopes of the respective target NAbs, were not identified; BLAST analysis did not identify alternative viral or cellular sequences with which the mimotope sequences might correspond.

Fig. 1.

Consensus mimotope sequences. Mimotope sequences, identified using direct binding selection with Z13, 2F5 or 4E10 NAb targets, were aligned to generate sequence logos to identify conserved amino acids. Conserved mimotope sequences corresponded to the known HIV-1 gp41 MPER sequence, shown above the logos. The height of each stack and each letter indicates the conservation and relative frequency of each amino acid at that position respectively.

Competitive biopanning in the presence of HIV-1 or sCD4-bound HIV-1 was evaluated for the 4E10 NAb in an attempt to recover mimotopes with increased binding capacity. A subtype B HIV-1 pseudovirus (PSV), 90US_873, sensitive to neutralization by MPER NAbs, was used for competition with phage for the target NAb. Amino acids in the HIV-1 4E10 motif, N/SWFxxS/T, were identified with increased frequency in mimotopes selected under competition with HIV-1 (80% MPER homology) and sCD4-bound HIV-1 (93% MPER homology), than mimotopes selected by direct binding (69% MPER homology; data not shown).

Characterization of MPER mimotope binding to target NAbs.

All MPER mimotopes were evaluated for binding to the target NAb by ELISA. Mimotopes with the lowest M13 binding titers, indicating higher binding capacity, or with novel sequences were selected for further characterization (Table 1). A greater number of 4E10 mimotopes were included to reflect the higher sequence diversity observed for this target NAb. Disulfide-constrained mimotopes bound to 4E10 with significantly lower M13 titer than linear mimotopes (8.3-fold lower; Mann-Whitney U test, p = 0.007). Linear 2F5 and Z13 mimotopes bound with lower M13 titer than disulfide-constrained mimotopes (10- and 2.1-fold lower respectively). On average, the lowest M13 binding titers were observed for Z13 NAb. No binding was observed to the glycan-specific, HIV-1 gp120 NAb, 2G12 used as an assay control (> 2.5E+9 M13/ml); NAb binding was also not observed to the control M13 that did not express the MPER peptide. (Table 1)

Table 1.

Mimotopes sequences, NAb binding titers and MPER-peptide competition binding titers of selected 4E10, 2F5 and Z13 NAb-specific mimotopes.

Mimotope	Selection Condition	Sequence	Binding Titer (M13/ml)		Peptide Comp (IC50 μg/ml)
			Target bNAb	2G12	MPER	Scramble
4E10
PhD c7c-1	direct binding	SIFDWPR	1.7E+06	> 2.5E+9	26.7	> 100
PhD c7c-2	direct binding	ERMLFEW	8.6E+06	> 2.5E+9	44.9	> 100
PhD c7c-3	direct binding	TLDIFRN	8.8E+06	> 2.5E+9	7.9	> 100
PhD c7c-4	direct binding	YFFDRSS	2.2E+07	> 2.5E+9	3.1	> 100
PhD 12–1	direct binding	NWFNLTQTLMPR	6.5E+07	> 2.5E+9	11.3	> 100
PhD 12–2	direct binding	SVSVGMKPSPRP	4.3E+07	> 2.5E+9	7.2	> 100
PhD 12–3	direct binding	NWFDRSHTLFHS	2.0E+08	> 2.5E+9	2.3	> 100
PhD 12–4	direct binding	NWFDGTTTLWHR	4.3E+07	> 2.5E+9	4.0	> 100
PhD 12–6	direct binding	DMRSIFHDNPFN	2.7E+07	> 2.5E+9	2.6	> 100
PhD 12–7	direct binding	GYWSDYWGMTTH	2.3E+08	> 2.5E+9	> 100	> 100
PhD 12–8	comp. w/sCD4:HIV	QSYNWFDHTRWI	1.7E+08	> 2.5E+9	4.7	> 100
PhD 12–9	comp. w/sCD4:HIV	NFWELSKYLHMA	1.6E+07	> 2.5E+9	15.5	> 100
PhD 12–10	comp. w/sCD4:HIV	TNFFELTTKLHR	1.4E+07	> 2.5E+9	76.6	> 100
PhD 12–11	comp. w/sCD4:HIV	LPNWFNLSSNLM	5.1E+07	> 2.5E+9	7.1	> 100
2F5
PhD c7c-1	direct binding	MELDKWA	2.9E+08	> 2.5E+9	0.5	> 100
PhD c7c-2	direct binding	DKWASFD	4.6E+07	> 2.5E+9	0.2	> 100
PhD 12–1	direct binding	HNSQNLDRWAII	8.2E+06	> 2.5E+9	0.6	> 100
PhD 12–2	direct binding	SEMDDKWAKMKI	3.6E+07	> 2.5E+9	0.6	> 100
PhD 12–3	direct binding	SEELDKWYNTLT	6.9E+06	> 2.5E+9	1.0	> 100
Z13
PhD c7c-1	direct binding	ALCLGPR	9.5E+06	> 2.5E+9	> 100	> 100
PhD c7c-2	direct binding	PFGFELW	8.4E+06	> 2.5E+9	> 100	> 100
PhD 12–1	direct binding	SHTLNFLDLTST	3.5E+06	> 2.5E+9	> 100	> 100
PhD 12–2	direct binding	QHWNYFDLSEQQ	1.3E+06	> 2.5E+9	> 100	> 100
PhD 12–3	direct binding	GPILANYSDITN	8.0E+06	> 2.5E+9	> 100	> 100
Control M13	NA	No insert	> 2.5E+9	> 2.5E+9	> 100	> 100

Mimotope binding assays were then modified to include HIV-1 MPER peptide competition for target NAb binding. Titered MPER peptide was used to compete with mimotopes for binding to target NAbs; 50% inhibitory concentration (IC₅₀) of peptide was determined. Lower IC₅₀ values indicate greater mimotope sensitivity to binding competition. Mimotope binding to 4E10 and 2F5 NAbs was inhibited by MPER peptide, but binding to Z13 was not (Table 1, right). A range of peptide inhibition was observed for 4E10 mimotopes; IC₅₀ values were significantly lower for 2F5 mimotopes than for 4E10 (30-fold lower; Mann-Whitney U test, p = 0.0002). No difference was observed between linear and disulfide-constrained mimotopes; peptide inhibition did not correlate with M13 binding titer to target NAb (data not shown). No inhibition was observed with the MPER scrambled peptide, indicating the observed competition was sequence-specific.

To further explore the ability of 4E10 and 2F5 mimotopes to compete with HIV-1 MPER antigens for target NAb binding, viral capture competition assays were performed with mimotopes, viral particles and target NAbs. For this analysis, MPER NAb neutralization-sensitive (green symbols) and neutralization-resistant (red symbols) PSVs were used; additionally a vesicular stomatitis virus (VSV)-pseudotyped virus (black symbol) was utilized as a non-specific control. Firstly, virion capture by target NAbs, without mimotope competition, was determined. The relationship between the ELISA input p24 concentration and the captured p24 concentration remaining after multiple rounds of washing, is shown (Fig. 2A and B). Neutralization sensitive viruses were captured by both 2F5 and 4E10. Neutralization resistant viruses and VSV were not captured by 2F5, but were captured at low levels by 4E10 indicating a potential interaction with non-Env viral components. For the tested Envs, virus capture by 2F5 or 4E10 correlated directly with MPER NAb-mediated neutralization of the PSVs.

Fig. 2.

HIV capture and mimotope competition assays. Binding competition assays were performed to evaluate the ability of MPER mimotopes to inhibit binding of HIV-1 viral particles to 2F5 and 4E10 NAb. Capture of HIV-1 by A) 4E10 and B) 2F5 NAbs was evaluated using MPER neutralization-sensitive and neutralization-resistant PSVs from subtype A and B; VSV PV was used as negative control. PSV MPER neutralization sensitivity is shown on the bottom right. Serially diluted HIV-1 stocks were bound to target NAb on microplates and captured HIV-1 was quantified by p24 ELISA. The dashed line indicates the reliable detection limit for the p24 ELISA used. C) Serially diluted mimotopes were then used to complete with HIV-1 PVs for binding to the target NAbs. The M13 concentration as which 50% of HIV-1 PV binding inhibition was observed, is reported.

Mimotopes were then used to inhibit NAb capture of HIV-1 and VSV (for 4E10 only; used as a measure of potential non-Env mediated lipid-specific reactivity), and the 50% inhibitory phage concentration (IC₅₀, M13/ml) for each mimotope clone and PSV pair was determined (Fig. 2C). Inhibition of HIV-1 binding was evaluated with all PSVs for 4E10 mimotopes and for neutralization-sensitive PSVs for 2F5 mimotopes. 4E10 mimotopes inhibited 4E10 NAb binding to all PSVs and VSV and no statistical difference was observed between neutralization sensitive and resistant PSVs. 4E10 mimotope binding titer to the 4E10 target NAb correlated significantly with mimotope inhibition of HIV PSV capture (spearman correlation, r = 0.7327, p < 0.0001; data not shown), but not VSV. 2F5 mimotopes inhibited HIV binding either weakly or not at all. Experiments were repeated using HIV-1 pre-incubated with sCD4 and similar trends were observed, however significantly higher mimotope IC₅₀ values were determined using sCD4-bound HIV (Mann Whitney U test, p < 0.0001; data not shown).

MPER mimotope epitope mapping of HIV-positive plasma IgG.

We next evaluated the ability of these MPER mimotopes to bind to antibodies elicited during the course of natural HIV-1 infection. This antibody mapping approach provides insight into the breadth of antibody recognition of HIV-1 MPER through use of multiple MPER mimotope variants, for 4E10 specifically. Plasma reactivity to MPER mimotopes was determined by mimotope immunoprecipitation using protein A-immobilized plasma IgG from ten chronically, subtype B HIV-infected individuals (Table 2). Mimotope binding to IgG was quantified by Western blot of the lysate to the M13 major capsid protein p8, present at 2700 copies per phage and therefore greatly amplifying detection of low titer mimotope-specific antibodies, and the binding units were quantified by densitometry. Normal human plasma (NHP) was used as a negative control and NHP spiked with 25 μg/ml 4E10, 2F5 or Z13 was used as positive controls for each cognate mimotope.

Table 2.

Normalized signal intensities of mimotope M13 p8 protein immunoprecipitated by plasma IgG from 10 subtype B HIV-infected individuals, a subtype B HIV + plasma pool and normal human plasma without or with MPER NAbs at 25 μg/ml; signal intensity quartiles are highlighted..

HIV + plasma IgG bound a minimum of 8 out of the 14 tested 4E10 mimotopes. Binding profiles overlapped with but were distinct from 4E10-spiked NHP. Individual 4E10 mimotopes were bound with varying signal intensities and were identified in 10%–90% of the evaluated individuals. Seven of the 10 plasma samples tested bound to one or both of the two 2F5 mimotopes tested; only weak binding was detected to the two Z13 mimotopes. A broadly reactive subtype B plasma pool bound to all 4E10 and 2F5 mimotopes, but did not to Z13 mimotopes. Minimal binding was detected to the control M13 lacking MPER peptide.

The neutralization specificity and potency of the HIV + plasma samples were then evaluated in the TZMbl neutralization assay using an HIV-2 env/HIV-1 MPER chimera C1C virus, shown to specifically detect MPER-directed neutralization (Yuste et al., 2006), and 11 PSVs from subtype A, B and C. MPER-specific neutralization was detected in 7 of the 10 plasma samples with ID₅₀, with values ranging from 30 to 959. Direct correlations were observed between plasma MPER-specific neutralization titer and plasma IgG 4E10 mimotope binding profiles for six 4E10 mimotopes (Fig. 3); however this trend was strongly driven by the most potently neutralizing plasma and, if excluded from analysis, the trends were no longer significant. The geometric mean of all 4E10, but not 2F5, mimotope IgG binding signal intensities correlated directly with plasma neutralization breadth (# of PSVs neutralized; spearman correlation, r = 0.7755, p = 0.0107) and potency (geometric mean ID₅₀ titer; r = 0.6970, p = 0.0306; Supplemental Fig. 1).

Fig. 3.

HIV + plasma MPER-specific neutralization and mimotope binding profile correlations. Correlations between HIV + plasma MPER-specific neutralization capacity, observed using the HIV-2/MPER chimera virus in the TZMbl NAb assay, and plasma IgG mimotope binding profiles, observed as M13 p8 signal intensity by Western blot after phage immunoprecipitation with plasma IgG; linear regression analysis is shown.

Select 4E10 and 2F5 mimotopes were then tested for their ability to inhibit 4E10 and 2F5 NAb and HIV + plasma-mediated neutralization of the HIV-2/HIV-1 MPER chimera C1C PSV. Single concentrations of mimotopes were incubated with titered NAb or plasma before the addition of HIV PSV; neutralization IC₅₀ and ID₅₀ values were determined in the presence or absence of mimotope competition (Fig. 4). No neutralization inhibition was observed with the non-specific M13 control. Moderate inhibition of 4E10 and 2F5 NAb neutralization was observed for 5 out of 6 4E10 mimotopes (Fig. 4A) and 2 out of 2 2F5 mimotopes (Fig. 4B) tested at a concentration 2.5 × 10¹⁰ M13/ml. 4E10 and 2F5 mimotopes capable of inhibiting target NAb HIV neutralization were also able to inhibit plasma neutralization by the subtype B HIV + plasma pool (Fig. 4C) and individual plasma, 50,780 (Fig. 4D). Less neutralization inhibition was observed with 2F5 mimotopes as compared to 4E10 mimotopes; this may be reflective of the lower HIV binding competition observed for 2F5 mimotopes as compared to 4E10 mimotopes (Fig. 2C).

Fig. 4.

Mimotope inhibition of HIV neutralization of the HIV-2/MPER chimera virus. Serially diluted NAbs, including A) 4E10 and B) 2F5 or HIV + plasma, including C) the subtype B HIV + plasma pool and D) individual HIV + plasma from patient 50,780, were tested for neutralization in the TZMbl NAb assay alone (black) or in the presence of mimotope (gray) at a single concentration. M13 control, without MPER peptide (dashed gray), was included as a negative control. Inhibition of neutralization in the presence of each inhibitor was determined and reported as an ID₅₀ or IC₅₀ values. Three independent experiments were preformed for each condition and standard error is shown. The dashed line indicated neutralization potency in the absence of inhibitor, with the dotted lines showing the 95% confidence around this value.

Immunogenicity of 4E10 mimotopes.

An immunogenicity study was then performed to determine if 4E10 mimotopes could elicit and boost HIV-specific immune responses. Five 4E10 mimotopes capable of inhibiting MPER-specific neutralization were evaluated alone, in combination, or in combination with an acute subtype C HIV-1 gp145 Env protein, previously shown to elicit neutralizing antibodies in rabbits (Wieczorek et al., 2015) and nonhuman primates (Bolton et al., in preparation). Female BALB/C mice, five animals per group, were vaccinated with 4E10 PhD 12 clone 11 or clone 7 alone, 5 4E10 mimotopes together (4E10 PhD mix: PhD 12 clones 6, 7, 8 and 11, and PhD c7c clone 4), gp145 alone, gp145 in combination with the 5 4E10 mimotopes, or non-specific control M13 (no responses observed; data not shown). The vaccines were formulated in liposomes and administered at weeks 0, 4 and 8; immune responses were evaluated pre-immunization and two weeks after the last immunization.

Binding antibody titers were determined by ELISA using gp145 and gp41 Env proteins. Gp145 alone or gp145 in combination with 4E10 PhD mix elicited equivalently high antibodies against both gp145 (GMT: 388,023 and 445,722, respectively) and gp41 (GMT: 27,858 and 24,251, respectively). No responses were observed for other groups (data not shown). Biacore was used to characterize IgG binding to MPER peptide by pooled serum; no binding was detected (data not shown).

TZMbl and PBMC target cell neutralization assays were used to assess mouse serum antibody neutralization (Fig. 5). Sera collected prior to immunization and 2 weeks after the final immunization were titered against 10 HIV-1 PSVs from multiple subtypes and murine leukemia pseudotyped virus (MuLV), a nonspecific control, in the TZMbl NAb assay. Sera were titered against 2 subtype B infectious molecular clones in the PBMC neutralization assay. Neutralization potency of the peak time point was determined using the pre-immune time point as baseline. No neutralization was observed for 4 Tier 2 HIV PSVs and MuLV (data not shown). Animals immunized with gp145/4E10 PhD mix had higher titers overall as compared to animals immunized with gp145 alone, for both the TZMbl (GMT gp145/4E10 PhD mix = 130, gp145 alone = 65) and PBMC (GMT gp145/4E10 PhD mix = 87, gp145 alone = 52) assays. In addition, titers were also higher when the animals received the gp145/4E10 PhD mix as compared to the 4E10 PhD mix alone. The combination of gp145 plus the mimotope mix yielded significantly higher neutralization titers than the 4E10 PhD mimotope mix alone for the 93MW965, and GS015 viruses (p-values = 0.0079 and 0.0079, respectively, Fig. 5A). Animals immunized with mimotopes alone (4E10 PhD 12–11, 4E10 PhD 12–7 and 4E10 PhD mix) also produced detectable, low titer neutralizing responses. All sera were screened against the HIV-2/MPER chimera in the TZMbl assay and no neutralization was observed. No other statistical differences were observed between groups.

Fig. 5.

Neutralizing antibody responses elicited by mouse immunization with mimotopes and Env gp145 protein. Neutralization potency of vaccinated mouse sera evaluated in the A) TZMbl neutralization assay using a multisubtype panel of PSVs or the B) PBMC neutralization assay using 2 subtype B infectious molecular clones. The dashed line shows the negative cut-off value for each assay.

Cellular immune responses were assessed by intracellular cytokine staining (ICS) assay for both CD4⁺ and CD8⁺ T cells from the spleen and lymph node. IL-2 responses were observed for all groups against HIV-1 Env antigens, including MPER peptide, gp41, gp140 and gp145 (Supplemental Fig. 2). TNFα, CD107a and INFγ responses were not detected (data not shown). The greatest response for both CD4 and CD8 T cells was observed in animals immunized with gp145 alone (Supplemental Figs. 2A and B); minimal responses were observed in the control M13 group. No significant differences were observed between groups.

4. Discussion

Elicitation of neutralizing antibodies is thought to be a critical component of HIV-1 vaccine development. Detailed understanding of MPER epitope structures targeted by NAbs may improve vaccine designs. The MPER is a highly conserved and dynamic linear stretch of 25 amino acids adjacent to the viral lipid bilayer and contains binding sites for several NAbs, including 2F5, Z13 and 4E10 (Apellaniz et al., 2015). MPER can acquire different conformations upon antibodies binding; in complex with 2F5, MPER peptide assumes a type 1 β-turn structure, while in complex with 4E10, an α-helical conformation forms (Bryson et al., 2001; Cardoso et al., 2005, 2007). The dynamic nature of MPER, along with other features such as auto-reactivity of MPER-specific antibodies, have made MPER an elusive vaccine target (Alam et al., 2007; Nieva et al., 2011; Finton et al., 2013; Chen et al., 2014), yet MPER NAbs can develop during the course of natural HIV-1 infection.

Broadly NAbs develop in 10–30% of HIV + individuals within the first 2–3 years infection. Different cohorts with different HIV-1 subtypes have reported varying frequencies of MPER reactive antibodies, between < 1% and 66% (Landais et al., 2016; Gray et al., 2009b, 2011; Molinos-Albert et al., 2014; van Gils et al., 2009). Genetically unrelated antibody lineages have been shown to target the MPER in natural infection, and when present, correlate with plasma neutralization breadth and potency (Jacob et al., 2015). Understanding the range of immune responses against MPER in natural infection will help inform the type of responses that may need to be elicited for an effective MPER-specific vaccine.

Several methods exist to evaluate the MPER specificity of plasma antibodies, including binding analysis with MPER peptide or gp41 protein, neutralization of HIV-2/HIV-1 MPER chimeric viruses (Yuste et al., 2006) and computational neutralization fingerprinting (Doria-Rose et al., 2017). The epitope mapping approach we utilized in this study, may be a complementary approach for analysis of binding antibody responses. In this study, we sought to develop HIV-1 MPER mimotopes to characterize the range of epitope recognition of MPER antibodies elicited during natural infection. We identified a variety of mimotopes that bind specifically to MPER NAbs 4E10, 2F5 and Z13. Mimotopes selected by 2F5 or Z13 NAbs shared conserved motifs, homologous to the MPER region, that are required for NAb binding. Mimotopes selected by 4E10 were more degenerate, in agreement with the range of epitope recognition previously observed for this Nab (Finton et al., 2013). The anti-viral breadth and potency of the 4E10 NAb may also be attributed to its ability to recognize MPER in these multiple conformations.

4E10 NAb has previously been shown to be highly polyspecific and autoreactive, features thought to limit production by elimination through B cell tolerance mechanisms (Alam et al., 2007; Finton et al., 2013; Haynes et al., 2005; Yang et al., 2013). In this study, we observed 4E10 mimotope-specific plasma IgG in 100% of subtype B HIV + individuals tested. Plasma IgG binding of MPER mimotopes indicates that the 2F5 and 4E10 NAb MPER epitopes were immunogenic during these subtype B HIV-1 infections, however antibodies that bound to the Z13 NAb epitope were not detected using these mimotopes. The Z13 NAb, which binds to a difference face of the MPER helix than the 4E10 NAb, was generated from an antibody phage display library and may not represent an epitope commonly targeted in vivo (Nelson et al., 2007; Zwick et al., 2001). Different patterns and relative binding intensities of 4E10-like specificities were observed in the polyclonal plasma, indicating potential qualitative differences in MPER-specific humoral immune responses between HIV + individuals.

HIV + plasma IgG bound equally to MPER homologous (e.g. 4E10 PhD 12–11) and non-homologous (e.g. 4E10 PhD 12–1) 4E10 mimotopes. During natural infection, MPER assumes multiple conformations from native state, fusion-intermediates to post-fusion forms. Antibodies are elicited against these different MPER conformations, however those directed against the post-fusion form of gp41 will not neutralize HIV-1 (Frey et al., 2010). We therefore evaluated plasma neutralization capacity using the HIV-2/MPER chimera virus to determine if there was a correlation with specific mimotope binding profiles. We observed positive correlations between MPER-specific neutralization potency and the intensity of IgG binding to several 4E10 mimotopes, however use of more potently neutralizing sera samples would have helped clarify this trend. 4E10 mimotope binding correlated directly with plasma breadth and potency of heterologous HIV neutralization, similar to previous reports (Jacob et al., 2015). Additionally, five 4E10 mimotopes were capable of reducing neutralization potency of 4E10 NAb and HIV + plasma.

Immunization with select mimotopes indicated that HIV-specific immune responses can be elicited, albeit at low titer. The low copy number of each MPER peptide, 5 copies per phage, may have limited the MPER-specific responses using phage alone. 4E10 mimotopes given in conjunction with the gp145 Env immunogen were capable of improving NAb responses over gp145 Env alone, however the differences between groups were not statistically significant. Also, addition of the M13 phage itself may have boosted the immune response to the gp145 Env immunogen, however a gp145 Env and control phage group were not included in this study for direct comparison. Limited mouse serum volumes did not allow us to map elicited IgG specificities using the 4E10 mimotopes. The utility of these mimotopes as immunogens could be further explored through generation in high copy number nanoparticles, as they may have potential to boost and broaden MPER-specific immune responses.

In this study, we developed and evaluated mimotopes for MPER-specific NAbs, however, NAb targeting different regions of HIV-1 Env may also prove useful. Mimotopes of the CD4 binding site NAb, VRC01, have also been reported (Chikaev et al., 2015) and we are continuing to expand our study to evaluate NAbs targeting V1V2, V3, and the CD4 binding site, as well as additional MPER Nabs (Schoen et al., 2017). We have greatest interest in utilizing these mimotopes to finely map epitope-specific responses elicited during HIV infection or via vaccination and are working to improve the throughput and sensitivity of this approach. Understanding the development of antibody responses to key neutralizing determinants of the HIV-1 Env will be crucial for developing successful vaccine regimens, and these mimotopes may prove to be useful tools for understanding the specificities and repertoire of antibodies required for protection.