|Home | About | Journals | Submit | Contact Us | Français|
The membrane-proximal external region (MPER) of HIV-1, located at the C terminus of the gp41 ectodomain, is conserved and crucial for viral fusion. Three broadly neutralizing monoclonal antibodies (bnMAbs), 2F5, 4E10, and Z13e1, are directed against linear epitopes mapped to the MPER, making this conserved region an important potential vaccine target. However, no MPER antibodies have been definitively shown to provide protection against HIV challenge. Here, we show that both MAbs 2F5 and 4E10 can provide complete protection against mucosal simian-human immunodeficiency virus (SHIV) challenge in macaques. MAb 2F5 or 4E10 was administered intravenously at 50 mg/kg to groups of six male Indian rhesus macaques 1 day prior to and again 1 day following intrarectal challenge with SHIVBa-L. In both groups, five out of six animals showed complete protection and sterilizing immunity, while for one animal in each group a low level of viral replication following challenge could not be ruled out. The study confirms the protective potential of 2F5 and 4E10 and supports emphasis on HIV immunogen design based on the MPER region of gp41.
Eliciting broadly neutralizing antibodies is an important goal of HIV vaccine design efforts, and the study of broadly neutralizing monoclonal antibodies (bnMAbs) can assist in that goal. Human bnMAbs against both gp120 and gp41 of the HIV-1 envelope spike have been described. Three bnMAbs to gp41, 2F5, 4E10, and Z13e1, have been identified and shown to recognize neighboring linear epitopes on the membrane proximal external (MPER) region of gp41 (3, 24, 25, 37, 47). In a comprehensive cross-clade neutralization study by Binley et al., 2F5 neutralized 67% and 4E10 neutralized 100% of a diverse panel of 90 primary isolates (2). Similar broad neutralization was seen against sexually transmitted isolates cloned from acutely infected patients (22). More recently, a comprehensive study showed that 2F5 neutralized 97 isolates from a 162-virus panel (60%) and that 4E10 neutralized 159 isolates (98%) (41). Although less potent, the monoclonal antibody Z13, isolated from an antibody phage display library derived from a bone marrow donor whose serum was broadly neutralizing (47), has cross-clade neutralizing activity. Z13e1 is an affinity-enhanced variant of the earlier-characterized MAb Z13 that is directed against an access-restricted epitope between and overlapping the epitopes of 2F5 and 4E10. Both MAbs 2F5 and 4E10 were originally obtained as IgG3 antibodies in hybridomas derived from peripheral blood mononuclear blood lymphocytes (PBMCs) of HIV-1-seropositive nonsymptomatic patients and were later class switched to IgG1 to enable large-scale manufacturing and to prolong in vivo half-life (3, 6, 32).
Despite the interest in the MPER as a vaccine target, there is limited information on the ability of MPER antibodies to act antivirally in vivo either in established infection or prophylactically. A study using the huPBL-SCID mouse model showed limited impact from 2F5 when the antibody was administered in established infection (31). Passive administration of 2G12, 2F5, and 4E10 to a cohort of acutely and chronically infected HIV-1 patients provided little direct evidence of 2F5 or 4E10 antiviral activity, whereas the emergence of escape variants indicated unequivocally the ability of 2G12 to act antivirally (18, 39). Indirect evidence did, however, suggest that the MPER MAbs may have affected virus replication, as indicated by viral rebound suppression in a patient known to have a 2G12-resistant virus prior to passive immunization (39). Another study of 10 individuals passively administered 2G12, 2F5, and 4E10 before and after cessation of combination antiretroviral therapy (ART) showed similarly that 2G12 treatment could delay viral rebound, but antiviral activity by 2F5 and 4E10 was not clearly demonstrated (21). In prophylaxis, an early 2F5 passive transfer study with chimpanzees suggested that the antibody could delay or lower the magnitude of primary viremia following HIV-1 challenge (7). A study using gene transfer of 2F5 in a humanized SCID mouse model suggested that continuous plasma levels of approximately 1 μg/ml of 2F5 may significantly reduce viral loads in LAI- and MN-challenged mice (34). Protection studies of rhesus macaques using simian-human immunodeficiency virus SHIV89.6PD challenge did not provide definitive direct evidence for MPER antibody-mediated protection. One of three animals was protected against intravenous (i.v.) challenge when 2F5 was administered in a cocktail with HIVIG and 2G12 (19), but all three animals treated with 2F5 alone at high concentration became infected. In a vaginal challenge study with SHIV89.6PD (20), four of five animals were protected with a cocktail of HIVIG, 2F5, and 2G12, but a 2F5/2G12 combination protected only two of five animals. Further protection studies have used MPER MAbs in combination with other MAbs, leaving the individual contributions of these antibodies uncertain (1, 8).
In our previous studies, we successfully used the SHIV/macaque model to demonstrate neutralizing antibody protection against mucosal challenge, and we have begun to explore how that protection is achieved (12, 30). Here, we conducted a protection study with the two broadly neutralizing MPER-directed antibodies 2F5 and 4E10. We show that the antibodies can prevent viral infection and thereby support the MPER as a vaccine target.
All protocols for male Indian rhesus macaques were reviewed and approved by the Institutional Animal Care and Use Committees of the Scripps Research Institute and the Wisconsin National Primate Research Center. The animals were housed in accordance with the American Association for Accreditation of Laboratory Animal Care Standards. At the start of all experiments, all animals were experimentally naïve and were negative for antibodies against HIV-1, simian immunodeficiency virus (SIV), and type D retrovirus. For intrarectal inoculations, animals were sedated with ketamine hydrochloride (10 mg/kg intramuscularly [i.m.]) and placed within a biosafety cabinet in ventral recumbency with the hindquarters elevated. The tail was elevated dorsally, and 2,000 50% tissue culture infective doses (TCID50) SHIVBa-L contained in 1 ml of sterile tissue culture medium were administered atraumatically into the rectum. Following virus administration, the tail was lowered, ensuring complete delivery, and the hindquarters remained elevated for 15 min before the anesthesia was reversed.
SHIVBa-L retains the R5 phenotype of HIV-1Ba-L. The virus used in this study was produced as described by Pal et al. (27, 29) by transfection of HeLa-Tat cells and expansion in PM-1 cells. To enhance infectivity for macaques, the virus was passaged serially in vivo through four pigtail macaques. The infectious titer increased from 500 TCID50 per ml for unpassaged virus to 2,050 TCID50 per ml of in vivo-passaged virus, demonstrating an adaption of SHIVBa-L for macaque cells. After animal-to-animal passage, the virus was expanded on PM-1 cells to make stock and was harvested when the SIV p27 concentration was determined to be 100 ng/ml. SHIVBa-L stock was propagated in phytohemagglutinin (PHA)-activated rhesus macaque peripheral blood mononuclear cells (PBMCs) and obtained through Advanced Bioscience Labs, Inc., Kensington, MD (27, 29). An infectious titer (TCID50) of 4.1 × 104 per ml was determined in TZM-bl cells, and the SIV p27 concentration was determined to be 139 ng/ml (Zeptometrix Corp., Buffalo, NY).
Recombinant IgG1 2F5 and 4E10 were obtained from Polymun Scientific, Vienna, Austria (3, 6, 32). The isotype control antibody DEN3, an anti-dengue virus NS1 human IgG1 antibody, was expressed in Chinese hamster ovary (CHO-K1) cells in glutamine-free custom-formulated Glasgow minimum essential medium (GMEM selection medium; MediaTech Cellgro, MD). For large-scale tissue culture, the medium was supplemented with 3.5% Ultra Low bovine IgG fetal bovine serum (Invitrogen, Carlsbad, CA) and grown in 10-layer Cellstacks and Cell Cubes (Corning, Corning, NY). The antibody was purified using protein A affinity matrix chromatography (GE Healthcare, Chalfont St. Giles, United Kingdom) and dialyzed against phosphate-buffered saline (PBS). Care was taken to minimize endotoxin contamination, which was monitored using a quantitative chromagenic Limulus amoebecyte lysate assay (Cambrex Corp., East Rutherford, NJ) performed according to the manufacturer's recommendations. Antibody used for the passive transfer experiments contained <1 IU per mg of endotoxin.
The quantity of SIV viral RNA (vRNA) genomic copy equivalents per ml in EDTA-anticoagulated plasma was determined using two methods. The first was a quantitative reverse transcription-PCR (QRT-PCR) assay as previously described (10). Briefly, vRNA was isolated from plasma using a GuSCN-based procedure as described previously (5). QRT-PCR was performed using the SuperScript III Platinum one-step quantitative RT-PCR system (Invitrogen, Carlsbad, CA). Reaction mixes did not contain bovine serum albumin (BSA). Reactions were run with a Roche LightCycler 2.0 instrument and software. The vRNA copy number was determined using LightCycler 4.0 software (Roche Molecular Diagnostics, Indianapolis, IN) to interpolate sample crossing points onto an internal standard curve prepared from 10-fold serial dilutions of a synthetic RNA transcript representing a conserved region of SIV gag.
In the second method, available selected samples of sufficient volume were further tested in a highly sensitive plasma viral load assay designed to permit reliable measurement of SIV genomic RNA down to 10 copy equivalents per ml of plasma, as fully described by Cline et al. (5).
2F5 and 4E10 antibody concentrations in macaque sera were determined in enzyme-linked immunosorbent assays (ELISAs) by three different methods: (i) by binding to M41xt, a construct comprised of the short peptide, GELDKWASLC, and a fusion protein of the ectodomain of gp41JR-FL linked to the C terminus of the maltose binding protein and more fully described elsewhere (46); (ii) by binding to epitope-specific peptides synthesized from the HIV-1 MN Env sequence by AnaSpec, San Jose, CA (HIV-1 MN Env-165, catalog no. 28760, for 2F5 serum ELISAs and HIV-1MN Env-168, catalog no. 28763, for 4E10 serum ELISAs; and (iii) by binding to recombinant gp41 (Vybion, Ithaca, NY). For each ELISA, serially diluted serum was used to determine the concentration by comparison to the corresponding antibody standard curve using nonlinear regression.
The presence of serum antibodies against HIVBa-L gp120 was evaluated by ELISA. Briefly, purified HIVBa-L gp120 was immobilized on enzyme immunoassay (EIA) plates (Costar; Corning, NY) at 2 μg/ml in phosphate-buffered saline overnight at 4°C. All subsequent incubations were conducted at room temperature. Plates were washed four times and blocked with 4% nonfat dry milk for 1 h. All wash steps were carried out with PBS-0.05% Tween. Serum samples were diluted to 1:20 and serially diluted in a 3-fold titration in 1% BSA-PBS-0.02% Tween, transferred to assay plates, and incubated for 2 h. Plates were washed as described above, followed by the addition of the secondary antibody, goat anti-human IgG F(ab′)2 (alkaline phosphatase conjugate; Jackson Immuno Research) and incubated for 1 h. Plates were developed with an alkaline phosphatase substrate and read at 405 nm.
Neutralization of antibodies was assessed by two different methods. Neutralization of the primary isolate SHIVBa-L, was performed using phytohemagglutinin (PHA)-activated peripheral blood mononuclear cells (PBMCs) from a single rhesus macaque (no. 355) as target cells. Cells from this animal replicate SHIV efficiently. Neutralization assessment was carried out as described previously (30). Neutralization titers of animal sera were also determined and reported by Monogram Biosciences, South San Francisco, CA, after preparation of an HIV-1 envelope-pseudotyped luciferase SHIVBa-L capable of single-round replication. The pseudovirus-based neutralization assay was performed as previously described (33).
Antibody-dependent cell-mediated viral inhibition (ADCVI) antibody activity was measured using methods similar to those described previously (9, 12). Briefly, target cells (CEM.NKR-CCR5 cells infected with virus for 48 h) were incubated with MAb and with fresh human PBMC effector cells (effector/target ratio = 10:1). Seven days later, p27 from the supernatant was determined by ELISA (Zeptometrix Corporation, Buffalo, NY). The percent virus inhibition at each concentration of test MAb was determined in comparison to a negative control MAb (DEN3).
Frozen peripheral blood mononuclear cells (PBMCs) isolated from EDTA-anticoagulated blood were used in enzyme-linked immunospot (ELISPOT) assays for detection of CD4+ and CD8+ responses to pools of 10 overlapping 15-mer peptides from SIVmac239 Gag, Nef, Vif, and Pol proteins to test for the presence of SIV-specific cellular immune responses. These proteins were chosen because of the homology between SHIVBa-L and SIVmac239, the genes of which were used to construct SHIVBa-L (27). We detected gamma interferon (IFN-γ)-positive cellular immune responses using a nonhuman primate IFN-γ ELISPOT assay kit (MabTech) according to the manufacturer's instructions. Stimulating peptides at a 1 μM final concentration were added to 1.0 × 105 PBMCs per well and incubated for 16 to 18 h at 37°C in a 5% CO2 incubator. All tests were performed in duplicate. Wells were imaged with an AID ELISPOT reader (AID, Strassberg, Germany) and counted by AID EliSpot Reader version 3.1.1 (AID, Strassberg, Germany). Spots were counted by an automated system with set parameters for size, intensity, and gradient. Background levels (mean for two to six wells/plate without peptide) were subtracted from the level for each well on the plate. A response was considered positive if the mean number of spot-forming cells (SFCs) of duplicate sample wells exceeded the background plus two standard deviations and was >50 SFCs per 1 × 106 cells.
We tested Gag, Nef, and Vif SIVmac239 proteins for cellular responses by intracellular cytokine staining (ICS). The stimulation was performed as described recently by Hansen et al. (11). Briefly, 500,000 Ficoll purified peripheral blood mononuclear cells (PBMCs) were incubated for 1.5 h at 37°C in 200 μl of R10 (RPMI 1640 containing 10% fetal calf serum and antibiotics) with anti-CD28, anti-CD49d, and synthetic peptides (pools of 15-mers) of Gag amino acids (aa) 1 to 291, Gag aa 285 to 510, Nef aa 1 to 263, and Vif aa 1 to 214. The final concentration of the peptides was 0.1 μM. Then, 10 μg of brefeldin A per ml was added to prevent protein transport from the Golgi apparatus, and the cells were incubated for a further 5 h at 37°C. Cells were washed and stained for surface expression of CD4 and CD8 markers and fixed overnight in 1% paraformaldehyde at 4°C. The following day, cells were permeabilized in buffer containing 0.1% saponin and stained for expression of the cytokines gamma interferon (IFN-γ), interleukin-2 (IL-2), and TNF-α before being fixed in 1% paraformaldehyde for 2 h at 4°C. Events were collected on a BD LSR II flow cytometer (Becton Dickinson, San Jose, CA) with FASCDiva 6.0 software and analyzed with FlowJo v.8.7.3. (Treestar, Ashland, OR). A positive result is defined as being at least 2-fold higher than that for the negative control, preferably with the mean fluorescence intensity at least in the second decade above the negative peak.
The frequencies of CD4+ and CD8+ T cells and memory and naïve CD4 or CD8 T-cell subsets within the lymphocyte population were determined by fluorescence-activated cell sorter (FACS) analysis using anti-human CD4-allophycocyanin (APC) (clone SK3), anti-human CD3-Alexa700 (clone SP34-2), anti-human CD8-Pacific Blue (clone RPA-T8), anti-human CD95-fluorescein isothiocyanate (FITC) (clone DX2), and anti-human CD28-phycoerythrin (PE) (clone CD28.2). All the antibodies were purchased from Becton Dickinson, San Jose CA. Data were acquired on a custom-made BD LSR II flow cytometer (Becton Dickinson, San Jose CA) equipped with 488-nm blue, 635-nm red, and 405-nm violet lasers. FCS files were analyzed by FlowJo software version 8.7.3 (Tree Star, Inc., Ashland, OR).
We first evaluated MAbs 2F5 and 4E10 in PBMC and pseudovirus neutralization assay formats against SHIVs that have previously been characterized in macaques. As protection in vivo is generally correlated with neutralization in vitro, we sought a SHIV that was reasonably neutralization sensitive to both MAbs. A summary of the neutralization data is shown in Table Table1.1. The R5 SHIVBa-L (27, 29) was chosen over the X4 SHIV89.6P due to its greater sensitivity to neutralization by 2F5 and 4E10 and because an R5 virus is more representative of those involved in human transmission than an X4 virus.
A total of 16 male Indian rhesus macaques were intrarectally challenged with 2,000 50% tissue culture infectious (TCID50) of SHIVBa-L. One day prior to challenge, each animal was given an intravenous dose of 50 mg/kg of either 2F5, 4E10, or the isotype control IgG (anti-dengue virus NS1, DEN3). Experimental animals were divided into treatment groups of six animals for administration of 2F5 and 4E10. The isotype control group consisted of two animals, plus two additional untreated animals that were challenged prior to the beginning of the protection study to confirm viral infectivity. We decided to use a relatively high antibody dose to provide the best opportunity of observing protection based on the observation that many multiples of the neutralizing titer of the antibody (90% inhibitory concentration [IC90]) measured in a PBMC assay have been described previously as required for protection (26, 30). Based on earlier pharmacokinetic observations with 2F5 and 4E10, we expected the serum antibody half-life to be relatively short (4 to 5 days) (14, 19), so a second intravenous antibody treatment was given at day +1 following viral challenge in order to ensure that serum antibody concentrations were maintained long enough to favor a positive outcome. Blood draws were conducted at regular intervals following challenge to monitor viral infection, serum levels of passively administered antibody, and serum neutralizing activity.
The outcome of the protection study is depicted in Fig. Fig.1.1. The two isotype control-treated animals became infected, with peak viremia of approximately 107 virus copies per ml between days 14 and 21. The two control animals (96068 and 2025) which were challenged prior to the beginning of the protection study to confirm viral infectivity, but not treated with antibody, were also infected, with onset and peak viremias similar to those for the isotype control-treated animals.
In contrast, all 2F5- and 4E10-treated animals appeared to be protected against the establishment of infection. In the 4E10 antibody treatment group, no viremia was detected in six of six animals. Viral loads were first measured at days 1, 7, 10, 14, 17, and 21 and weekly thereafter for 6 weeks after challenge in quantification method with a detection threshold of 150 copies per ml. All samples were negative, except for one animal (99053) that showed a viremia of 2.7 × 102 copies per ml on day 14. However, a second, highly sensitive (detection threshold of 10 to 15 copies per ml) viral load measurement method did not confirm the presence of viral RNA in the same sample, suggesting a contamination with the first measurement. In addition, the highly sensitive viremia detection method confirmed the absence of detectable viral replication for all the samples with sufficient available volume to be retested (Table (Table22).
In the 2F5 treatment group, all animals but one showed the absence of viral replication in the assay with a detection limit of 150 copies per ml, with a low-level viremia of 4.0 × 102 copies per ml on day 35 measured in animal 93060. Sufficient sample volume from day 35 was not available to retest the plasma in the highly sensitive assay. An additional evaluation of plasma samples for which sufficient volume was available confirmed the absence of viral replication in all other 2F5-treated animals (Table (Table2).2). In summary, all animals in the 4E10 group and all animals but one with a potential viral blip in the 2F5 group appeared to be completely protected.
We investigated anti-Env antibody responses in an HIV-1Ba-L gp120 ELISA for evidence of any low-level infection in any of the treated animals. Serum samples from infected animals were included as a positive control (Fig. (Fig.2).2). We evaluated day 21, day 42, and month 6 sera of all 2F5 and 4E10-treated animals and detected no gp120-specific responses.
The possibility of low-level infection was also investigated by looking for specific CD4+ and CD8+ T-cell responses in PBMCs isolated from all study animals at month 6 using SIVmac239 Gag, Nef, Vif, and Pol peptide pools. These proteins were chosen because of the high homology between SHIVBa-L and SIVmac239, the genes of which were used to construct SHIVBa-L (27). No cellular immune responses were detected against any of the peptide pools in any animal but one. The 4E10-treated animal 01050 displayed both CD4+ and CD8+ T-cell responses against the Gag A-G peptide pool (Fig. (Fig.33).
The 2F5 and 4E10 antibody concentrations in the sera of the macaques were measured using three different ELISA formats: (i) by binding to the M41xt construct (46), (ii) by binding to epitope-specific peptides HIV-1 MN Env-165 for 2F5 and HIV-1MN Env-168 for 4E10 (AnaSpec, San Jose, CA), and (iii) by binding to recombinant HIV-1 gp41 (Vybion, Ithaca, NY). In all formats, a dilution series of serum was compared to the appropriate antibody standard curve and the concentration determined using a nonlinear regression curve fit analysis. The serum concentrations derived from the three formats were generally in good agreement. Table Table33 and Table Table44 summarize the serum antibody concentrations in 2F5- and 4E10-treated animals. The data shown in Fig. Fig.44 represent the mean serum antibody concentrations for each group of animals generated by averaging values from the three ELISA formats for each animal and then averaging over all animals in each group. Following the first i.v. antibody transfer at 50 mg/kg, the mean concentrations of 2F5 and 4E10 at the time of challenge were 742 μg/ml and 866 μg/ml, respectively.
The antibodies were transferred in two doses: first at 1 day before viral challenge (day −1) and again at 1 day following viral challenge (day +1). Antibody kinetics is usually described by two exponential functions reflecting the distribution and elimination phases. However, in our study, we did not have enough early data points (no data points between day 1 and day 7), during the distribution phase, to derive both rates. Therefore, elimination half-lives were calculated from day 7 onwards (Fig. (Fig.4,4, solid lines), resulting in values of 4.6 days for 2F5 and 4.1 days for 4E10 (GraphPad Prism 5 for Mac). We do not expect any significant error from this single-exponential fitting, as the distribution of the administered antibodies is significantly faster than the elimination (14). A model with one single rate constant that took into account the data points from days 0 and +1 as well as the decay following the first antibody dose and the peak resulting from the second dose resulted in virtually the same half-lives (4.4 and 4.1 days for 2F5 and 4E10, respectively) is shown in Fig. Fig.44 as dashed lines.
Serum samples were assayed from all animals beginning on the day of the first antibody i.v. transfer and continuing for the following 3 weeks in a SHIVBa-L pseudovirus single-round assay established by Monogram Biosciences (33). The 50% (IC50) and 90% (IC90) serum neutralizing titers for all 2F5- and 4E10-treated macaques are shown in Table Table55 and Table Table6,6, respectively. In both the 2F5 and 4E10 groups, IC50s were similar and on the order of 1:2,000 on the day of viral challenge, and IC90 values were on the order of 1:200.
Studies have indicated that 2F5 and 4E10 bind somewhat weakly to infected cells (35, 47) and therefore may not efficiently mediate activities such as antibody-dependent cell-mediated cytotoxicity (ADCC). Because FcR-mediated antibody effector functions are important for protection by MAb b12 (12), we wished to investigate whether 2F5 and 4E10 could inhibit virus yield from infected cells in the presence of Fc receptor-bearing effector cells in vitro. In the antibody-dependent cell-mediated viral inhibition (ADCVI) assay, an infected cell culture from which supernatant containing excess free virus is removed is incubated with antibody and effector cells, and then viral infection is estimated 7 days later (9). In Fig. Fig.55 we show that 4E10 is poor at mediating ADCVI, while 2F5 displays significant activity only at high concentrations, being overall notably less effective than b12.
Two MHC class I alleles (Mamu-B*08 and Mamu-B*17) have been associated with elite control of SIV replication. We therefore evaluated all experimental animals by MHC genotyping using PCR with sequence-specific primers (PCR-SSP) to test macaque samples against a panel of nine class I alleles shown to be important in SIV epitope presentation (4, 15, 17, 28, 42). The results shown in Table Table77 reflect that no animals in either antibody treatment group express the Mamu-B*08 or Mamu-B*17 allele. The Mamu-A*01 allele, an allele that occurs frequently in many colonies and that has been associated with moderate reduction of SIVmac239 replication (4, 23, 28, 44), appears in four macaques in the 2F5 group and two macaques in the 4E10 group. Mamu-A*08, also expressed at relatively high frequency, is seen in four animals treated with 2F5 and two animals treated with 4E10. This allele has so far been linked to a single epitope presentation derived from SHIVHXBC2, and its role in SIV pathogenesis is unclear (40). The Mamu-B*01 allele is present in two macaques in the 2F5 group and one macaque in the 4E10 group. This allele remains on the panel based on early reporting of SIV-derived epitopes (38, 43), but subsequent studies have shown that Mamu-B*01 does not bind SIV-derived epitopes and has no effect on SIV disease progression (17). Overall, even with the presence of the Mamu-A*01, Mamu-A*08, and Mamu-B*01 alleles (which are not associated with elite control of SIV replication) there is no apparent correlation with the allelic profiles of the animals in this study that would account for any unusual ability to resist infection.
To assess the possibility that antibody administration followed by SHIV viral challenge may have initiated a long-term protective immune response, rechallenges were conducted with 2,000 TCID50 of SHIVBa-L at 6 months after the first SHIV challenge. Figure Figure66 shows that all animals in each antibody treatment group were infected except 2F5-treated animal 93060. To ensure that 2F5-treated animal 93060 was uninfected due to any virus-specific immune response, we evaluated PBMCs taken just prior to the second rechallenge and performed an ELISPOT assay using peptide pools for SIVmac239 Gag, Nef, Vif, and Pol. No immune responses were detected, confirming the ICS data presented in Fig. Fig.3.3. Six months later a second rechallenge was conducted with 93060, and the animal became infected with a peak viremia of 2.83 × 107 vRNA copies per ml measured in plasma taken on day 14 following the second viral rechallenge (Fig. (Fig.6B6B).
The results of this study suggest that gp41 MPER antibodies 2F5 and 4E10 can offer full protection against mucosal SHIV challenge. The serum neutralizing titers at the time of challenge for both MAbs were approximately 1:2,000 and 1:200 measured as IC50 and IC90 values in a pseudovirus assay. Based on the PBMC IC90 for 2F5 and 4E10 (Table (Table1)1) and measured serum concentrations, the neutralizing titers (IC90) at challenge were estimated as approximately 1:26 for 2F5 and 1:37 for 4E10 in a PBMC-based assay. Sterilizing immunity has been associated for other antibodies, isolates, and challenge routes with neutralizing titers of very roughly 2 orders of magnitude times the antibody IC90 in a PBMC assay (26, 30, 36), with notable exceptions (13, 20). Therefore, sterilizing immunity was observed here at a somewhat lower neutralizing titer on the day of challenge. However, it should be noted that in the present study, to maximize our chances of observing protection and given concerns about shorter half-lives of MPER antibodies, we gave a second bolus of passive antibody 1 day after challenge. Pharmacokinetics modeling suggests that the peak concentrations attained on day 1 were 1,282 μg/ml and 1,614 μg/ml for 2F5 and 4E10, respectively, corresponding to PBMC IC90 titers of about 1:64 for 2F5 and 1:49 for 4E10. Serum neutralizing titers were maintained at >1:1,000 (pseudovirus assay IC50) through day 7, still corresponding to an IC90 titer of roughly 1:50 in a PBMC assay. A titration of protection against MAb dose will be required to fully understand the minimal antibody neutralizing titers affording protection, as has been described for MAb b12 (30) and a monospecific polyclonal antibody preparation (26).
A very low-level viremia was detected at day 35 in 2F5-treated animal 93060, after the antibody concentration had waned to undetectable levels (Fig. (Fig.1).1). It is likely that this “blip” may represent a contamination of the first viral load measurement procedure, but sufficient plasma was not available to retest the day 35 plasma sample in a more sensitive viral load assessment. However, day 21 viral RNA was measured to be below 15 copies per ml of plasma, and further evaluations revealed no antibody or T-cell responses in this animal even at 12 months postchallenge (Table (Table2).2). Interestingly, rechallenge of 93060 failed to initiate a productive infection (Fig. (Fig.6A).6A). However, a second rechallenge conducted at month 12 after the initial challenge (Fig. (Fig.6B)6B) led to a regular infection course in the animal, and the significance of the absence of infection following the first rechallenge is unclear. Nevertheless, the possibility of a low-level viral replication in animal 93060 cannot be excluded. The detection of CD4+ and CD8+ T-cell specific responses to Gag in the 4E10-treated animal 01050 (Fig. (Fig.3)3) does also suggest the occurrence of viral replication. As described in Hansen et al. (11), these responses may reflect an infection that was locally contained early on, in the present case thanks to neutralizing Ab and potentially to the de novo T-cell responses. However, the detection of specific T cells at month 6 only, in the absence of detection at day 42, rather suggests a persistent low level of viral replication, generating T-cell responses at a later point. No CD8 depletion experiment was performed here to verify the presence of persistent viral replication.
Because antibody was transferred in two doses, first on day −1 before viral challenge and again on day +1 following viral challenge, estimation of antibody half-life from observed serum concentrations required more complex modeling than when transferring a single dose (Fig. (Fig.5).5). We calculated the elimination half-life of 2F5 to be 4.6 days and that of 4E10 to be 4.1 days, which is in good agreement with previous studies with these antibodies in macaques (19) and in humans (14). The half-lives are at the bottom end of the range described for antibodies in clinical use, i.e., 4.3 to 21.8 days (16).
Another interesting observation here is that the antibodies 2F5 and 4E10, and particularly 4E10, are less effective than b12 at ADCVI in vitro (Fig. (Fig.5).5). Nevertheless, the concentrations of both MPER antibodies in sera during the protection experiments were, high and ADCVI mechanisms may contribute to protection for both antibodies. Further experimentation will be required to better understand the relative contributions of different mechanisms to MPER antibody protection in vivo.
The tremendous breadth of HIV-1 neutralization by 2F5 and 4E10 has traditionally made the MPER region of gp41 an interesting target for immunogen design. This study shows that anti-MPER antibodies can provide protection at moderate serum neutralizing titers. There are considerable challenges in eliciting antibodies to recognize a region close to the viral membrane, but if these are overcome, the relevant immunogen could contribute to HIV vaccine protection.
We thank Karen Saye-Francisco for the production of the isotype control antibody and quality control assistance and Jillian Sola for technical assistance at The Scripps Research Institute. We are grateful for the assistance provided by Gretta Borchardt and Caitlin MacNair with genotyping and viral load assessments at the Wisconsin National Primate Research Center. We thank Lars Hangartner for helpful discussions. We appreciate the assistance provided by Jeff Lifson and Michael Piatak and the AIDS and Cancer Virus Program, Science Applications International Corporation Frederick, National Cancer Institute-Frederick, in conducting the highly sensitive vRNA assays on macaque plasma.
Support for this work was provided by the International AIDS Vaccine Initiative (IAVI) through the Neutralizing Antibody Consortium and by NIH grant AI033292 (to D.R.B.). MHC genotyping by sequence-specific PCR was performed by the University of Wisconsin Genotyping Core with support of NIH grant 5R24RR16038-6 awarded to David I. Watkins.
Published ahead of print on 11 November 2009.