Characterization of broadly neutralizing antibody responses to HIV-1 in a cohort of long term non-progressors

Background Only a small fraction of HIV-1-infected patients develop broadly neutralizing antibodies (bNAbs), a process generally associated to chronic antigen stimulation. It has been described that rare aviremic HIV-1-infected patients can generate bNAbs but this issue remains controversial. To address this matter we have assessed bNAb responses in a large cohort of long-term non-progressors (LTNPs) with low or undetectable viremia. Methods Samples from the LTNP cohort of the Spanish AIDS Research Network (87 elite and 42 viremic controllers) and a control population of 176 viremic typical-progressors (TPs) were screened for bNAbs using Env-recombinant viruses. bNAb specificities were studied by ELISA using mutated gp120, neutralization assays with mutated viruses, and peptide competition. Epitope specificities were also elucidated from the serum pattern of neutralization against a panel of diverse HIV-1 isolates. Results Broadly neutralizing sera were found among 9.3% LTNPs, both elite (7%) and viremic controllers (14%). Within the broadly neutralizing sera, CD4 binding site antibodies were detected by ELISA in 4/12 LTNPs (33%), and 16/33 of TPs (48%). Anti-MPER antibodies were detected in 6/12 LTNPs (50%) and 14/33 TPs (42%) whereas glycan-dependent HIV-1 bNAbs were more frequent in LTNPs (11/12, 92%) as compared to TPs (12/33, 36%). A good concordance between standard serum mapping and neutralization-based mapping was observed. Conclusion LTNPs, both viremic and elite controllers, showed broad humoral immune responses against HIV-1, including activity against many major epitopes involved in bNAbs-mediated protection.

Introduction Production of broadly neutralizing antibodies (bNAbs) against HIV represents a relatively infrequent event in HIV-infected patients [1,2]. One major issue to induce such antibodies resides in the high variability of the viral envelope and structural mechanisms hiding crucial epitopes for neutralization. Besides, maturation leading to high affinity antibodies represents a major challenge for the immune system that can be impaired by the immunodeficiency associated with HIV infection. Affinity maturation of antibodies is critical to confer effective neutralization against HIV and this maturation capacity becomes altered along infection [3][4][5].
The study of the mechanism of action of bNAbs is essential to understand the mechanisms of antibody neutralization and escape by HIV-1. Several studies have suggested that the development of neutralizing antibodies is a consequence of viral replication [1,21]. On the other hand, it is generally accepted that bNAbs are not able to contribute the control of viremia due to continuous escape by HIV from immune pressure through mutation or glycosylation. However, it has been recently described a role of bNAbs in HIV control in one patient with EC phenotype raising the possibility of an active role of bNAbs in the control of autologous viruses [22].
We have shown that patients receiving antiretroviral treatment are capable of inducing a broad and potent humoral immune response against HIV despite having undetectable levels of viremia [23]. According to these results, it is possible that long-term nonprogressors (LTNPs), individuals with low levels of viremia who maintain stable CD4 T cell counts over 10 years of infection, develop neutralizing antibodies with a high affinity profile. In fact, in isolated LTNP patients the presence of bNAbs has been described [24][25][26]. We have explored the hypothesis that preserved B cell function in LTNPs could result in the production of a broad humoral response. To get a better understanding of this issue, we have assessed the presence of bNAbs in a large cohort of LTNP, including both viremic and elite controllers. Furthermore we have characterized the epitopes targeted by bNAbs found in LTNPs in comparison with those in HIV typical progressors (TPs).

HIV-1 infected subjects
This study has been approved by Research Ethics and Animal Welfare Committee of Instituto de Salud Carlos III (CEI PI 42_2011-v2).
Samples (129) from the cohort of LTNPs from the RIS (median RNA copies/ml: 104, median CD4+: 734 cells/μl and asymptomatic HIV infection over 10 year after seroconversion) were kindly provided by the HIV BioBank integrated in the Spanish AIDS Research Network (RIS) [27]. The HIV BioBank, integrated in the Spanish AIDS Research Network, is partially funded by the RD12/0017/0037 project as part of the Plan Nacional R + D + I and cofinanced by ISCIII-Subdirección General de Evaluación y el Fondo Europeo de Desarrollo Regional (FEDER) and Fundación para la investigación y prevención del SIDA en España (FIPSE). Samples were processed following current procedures and frozen immediately after their reception. All patients participating in the study gave their informed consent and protocols were approved by institutional ethical committees. A population of 176 untreated TPs (median RNA copies/ml: 10,241, median CD4+: 567 cells/μl) from Hospital Clinic, Barcelona, was analyzed as control [23]. The overall rate of CD4 cell decline in TPs was 50-100 cells/μL per year. Patients in the present study signed informed consent. All subjects on this study were antiretroviral naïve at the time of sampling.
LTNPs were classified as elite and viremic controllers (Table 1). Among LTNPs, 87 of them were elite controllers with persistent viral load below 50 RNA copies/ml and a median CD4 + T cell count of 773 cells/μl and 42 were viremic controllers. Viremic controllers had a median number of viral RNA copies/ml in plasma of 3,450, a median CD4 + T cell count of 655 cells/μl and viral load was always below 10000 RNA copies/ml.
Infectious supernatants were generated by calcium phosphate transfection in HEK 293 T cells with 5 μg of the plasmids [31]. Titrated recombinant viruses were preincubated with the dilutions of sera (1/200-1/2000) for 30 minutes at 37˚C before the infection of the U87.CD4.

Epitope mapping
i. Neutralization assays. For the characterization of neutralizing antibodies, neutralization assays were performed using single-round infection HIV-1 Env pseudoviruses and TZM-bl target cells as previously described [33,34]. To determine the serum concentration producing 50% reduction in RLU value, serial dilutions were made and the neutralization dose-response curves were fitted by non-linear regression using a four-parameter hill slope equation.
For the assessment of CD4bs-directed neutralization, antibody-mediated neutralization was blocked with specific protein probes in a competition assay [7]. Briefly, 25 μg/ml of RSC3 (Resurfaced Stabilized Core 3) or 25 μg/ml of the mutant RSC3 Δ371I/P363N was incubated with sera serially diluted 4-fold starting from 1:10 for 30 min. JR-FL (subtype B), RW020 (subtype A) or ZA012 (subtype C) pseudovirus was added 30 min before the addition of TZM-bl cells and the infection proceeded for 48 h. CD4bs-directed activity was calculated as a 30% reduction in the ID50 values of the sera in the presence of RSC3 compared to RSC3 Δ371I/ P363N [35].
For the mapping of bNAbs directed to glycan structures in the variable region (V1V2 and V3), neutralization assays in TZM-bl were performed using JR-CSF virus with the N160K mutation and the N332A mutation respectively [36,37]. The mutation of N160 to lysine, an Nglycosylation site in the V2 loop, abolishes the neutralization mediated by PG9 and PG16 and N332A mutation removes a glycosylation site at the base of V3 loop essential for the formation of epitopes recognized by 2G12 and PGT bNAbs. A sample is considered positive if there is a decrease in ID50 greater than or equal to 50% for the mutant compared to the wild-type virus [38]. For the mapping of bNAbs to the membrane-proximal external region (MPER) of gp41, sera were tested for neutralizing activity against a chimeric HIV-2 virus containing the HIV-1 MPER region of gp41 (71312-C1) and the parenteral HIV-2 7312A clone [39].
The serum neutralizing antibodies were also mapped to the MPER region by a method of selective peptide inhibition of neutralization [40]. gp41-specific overlapping peptides (MPR.03: KKKNEQELLELDKWASLWNWFDITNWLWYIRKKK, 2F5.01: NEQELLELDKWASLWN, 4E10.22: CNWFDITNWLWYIRKKK and Z13e1.01: WASLWNWFDITNKKKK) were added to the serum for 30 min at a concentration of 25 μg/ml prior to the addition of 7312A-C1 virus (HIV-2/YU2 MPER chimera). A scrambled MPER region peptide (MPR.Scr.02: KKKRIYWLWNTIDFWNWLSAWKDLELLEQENKKK) was included as negative control. In these assays the reduction of the neutralization activity caused by MPER derived peptides is compared to that due to the mock peptide (defined as an equivalent volume of DMEM medium).
ii. ELISA analyses. ELISA assays were performed as previously described [7]. Plates were coated with the antigen in PBS at 2 μg/ml and incubated overnight at 4˚C. The ELISA plates were coated with the following probes: • YU2 gp120wt and YU2 gp120 D368R protein. The mutation at position 368 reduces or knocks out binding of most CD4bs Ab.
• The antigenically resurfaced glycoprotein RSC3 containing the CD4bs and RSC3 Δ371I/ P363N mutant, which affects the CD4 binding loop and reduces b12 and VRC01 binding.
• The RSC3 G367R probe [7,35] that creates a steric clash for mAb b12 binding but it causes little interference with VRC01 binding.
Sera that showed a loss of reactivity to the CD4bs mutants (YU2 D368R and RSC3 Δ371I/ P363N) and reacted to YU2, RSC3 and RSC3 G367R were classified as containing CD4bs antibodies as described previously [35,41]. In those samples with low reactivity to YU2 gp120wt or/and RSC3 (endpoint titer below or equal to 2500) ELISA with RSC3 G367R probe was not performed.
iii. Neutralization-based serum delineation analysis. Serum specificities were delineated using a neutralization fingerprint algorithm, as described previously [42]. Briefly, a reference set of monoclonal antibody-virus neutralization data was obtained for a set of 21 diverse HIV-1 strains (subtypes A, B and C) against a set of representative monoclonal antibodies, divided into epitope-specific antibody clusters (VRC01, b12, CD4, HJ16, 8ANC195, PG9, PGT128, 2G12, 2F5 and 10E8-like). For each serum, the pattern of neutralization of the same set of 21 strains was compared to the neutralization patterns (fingerprints) of the set of reference antibodies, and the relative contribution of antibodies from each cluster to the neutralization by a given serum was estimated.

Results
In order to better understand the breadth and the spectrum of neutralization against HIV-1 in LTNP patients, we evaluated the neutralizing capacity of sera samples from the cohort of LTNPs of the Spanish AIDS Research Network (RIS). The neutralizing activity in LTNPs was compared with that in the control group of TPs. In these neutralization assays, neutralizing activity of these sera were tested against a mini-panel of recombinant viruses with different subtypes (clades and tier categorizations are given in parentheses): VI191 (A, tier 2), NL4-3 (B, tier 1A), 92BR025 (C, tier 1B) and CM244 (AE, tier 2) (S1 Fig) [10,[43][44][45]. A VSV-pseudotyped HIV-1 was included as a control for nonspecific neutralizing activity and none of the sera showed neutralizing activity against VSV-pseudotyped HIV-1. We considered that a serum sample displayed broadly neutralizing activity (bNA) when it was capable of neutralizing all the recombinant viruses tested across with a titre !200, with no neutralization of the VSV-pseudotyped control.
LTNP serum samples capable to neutralize all the viruses in the mini-panel with an ID50!200 were selected and these samples were screened against an extended panel of 10 viruses more with various subtypes (Fig 1).
Although the number of broad neutralizer individuals in the viremic population was higher, broad neutralizer patients were also found among the elite controllers. Six of the LTNPs with bNA had viral loads under 50 copies/ml indicating that patients with undetectable viral loads are capable to develop a broadly neutralizing response. These six individuals represent 7% of the elite controllers´population (6/87) while in the case of the viremic LTNPs the ratio of individuals with bNA was 14% (6/42) ( Table 1).
The epitopes for the binding of the neutralizing antibodies contained in LTNPs sera and in TPs sera from patients with a broad neutralization profile were mapped.
To determine whether the CD4bs antibodies contained in sera were responsible for broad neutralization, neutralization competition assays using RSC3 glycoprotein containing the CD4bs were performed (Fig 2). RSC3 addition inhibited neutralization mediated by two samples of sera from LTNPs (449326 and 597473). For sample 449326 RSC3 addition inhibited  For the mapping of V1V2 and V3 glycan-dependent HIV-1 NAbs (Fig 3 and S2 Fig), neutralization assays in TZM-bl using JRCSF.N160K and JRCSF.N332A respectively were performed. A decrease of 2 fold or more in serum neutralization against JRCSF.N160K or JRCSF. N332A compared to the neutralization of wild-type JRCSF indicates the presence of neutralizing antibodies directed to glycans in V1V2 or V3, respectively. From the results obtained it can be deduced that glycan-dependent HIV-1 NAbs are significantly more abundant in the samples from LTNPs (11/12) than in those from TPs (12/33) (p value = 0.0017, Fisher's exact test).
Sera were analyzed to detect antibodies directed against MPER region of gp41 (Fig 4 and S3  Fig) assessing neutralizing activity against 7312A and 7312A-C1 viruses and anti-MPER antibodies were detected in both groups of patients. To confirm the presence of MPER neutralizing antibodies in the sera, the specificity of the antibodies against MPER was inhibited with soluble peptides containing different fragments of this region (Table 4). Four serum samples In summary, we found that most of the bNAbs from LTNPs map known neutralization epitopes and that in some subjects the neutralization breadth is mediated by antibodies with different specificities (Figs 5 and 6). A summary of serum neutralization specificities found with standard mapping is shown (Fig 5).
It has been shown that similarity in neutralization fingerprint correlates with similarity in epitope [42]. Therefore, epitope specificities of HIV-1-neutralizing antibodies in serum were Sera with V1V2 and V3 glycandependent HIV-1 NAbs are detected by neutralization assays with the parental JRCSF and the corresponding N160K and N332A mutants (V1V2 and V3 respectively). Sera with reactivity to glycan-dependent motives in V2 and V3 are indicated in orange and blue respectively. Percent serum neutralizing activity sensitive to the indicated residue replacement was calculated using the equation [1 − (ID50 mutant/ID50 wild type)] x 100. A sample is considered positive if there is a decrease in ID50 greater than or equal to 50% for the mutant relative to the wild-type virus. Monoclonal antibodies 2G12 and PG9 have been used as controls and are indicated by a black box. SEMs of two independent assays are shown. In this figure only samples with neutralizing antibodies directed to glycans in V1V2 and/or V3 are shown. elucidated from the serum pattern of neutralization against a panel of 21 HIV-1 isolates. The patterns of neutralization of the sera were compared with a reference set of 10 epitope-specific neutralization fingerprints (one for each epitope-specific antibody cluster). The predicted prevalence of the different clusters is shown as a heat map for each serum (Fig 6A). These data exhibited a high concordance with the ones obtained using the experimental assays (Fig 5), with at least one of the top two neutralization-based specificities identified by standard mapping in 92% of the sera (Fig 6B). Epitopes involved in bNAbs-mediated protection have been characterized by these different approaches and activity against all the analyzed epitopes was displayed in LTNPs and TPs (Fig 6C).

Discussion
It has been suggested that in HIV-infected individuals high levels of viral replication and the time since infection correlates with the induction of bNAbs [1,21,46]. However, in this work we have detected broadly neutralizing antibodies against HIV-1 in a cohort of LTNPs with low Characterization of bNAbs to HIV-1 in LTNPs or undetectable levels of viremia. Actually, LTNP were classified in two sub-groups according to viremia levels: elite controllers (persistent undetectable viremia) and viremic controllers (VL<10.000 copies/ml). Although we have found a higher percentage of individuals with bNAbs in viremic controllers (14% vs 7%), bNAbs were also found in long-term elite controllers, suggesting that other factors besides persistently detected viremia could drive the development of bNAbs. However, we cannot exclude that this broad humoral immune response could be due to hidden viral replication in other tissues such as gut-associated lymphoid tissue (GALT) or tonsils.
LTNPs and TPs developed antibodies against all kinds of epitopes analyzed. Therefore, bNAbs in LTNPs are mapped to specific known neutralization epitopes. These data have been obtained using two different approaches, neutralization-based and standard serum mapping. The concordance between these two methods was high. These two approaches have been previously compared in a population of 21 sera [42]. The data obtained in the present study of 19 additional sera validate the previous reports of concordance between neutralization fingerprinting and standard serum mapping, further underlining the utility of the neutralizationbased serum-epitope predictions.
Only two of the sera with binding antibodies against the CD4bs detected by ELISA had detectable neutralization activity confirmed by RSC3 competition neutralization assays. One possible explanation is that if the virus is neutralized by antibodies directed against other epitope, there will not be an effect on neutralization reduction mediated by RSC3. The sera samples with no inhibition in neutralization by this glycoprotein could have few antibodies against CD4bs or antibodies with low affinity incapable of mediating neutralization, only detectable Characterization of bNAbs to HIV-1 in LTNPs

Fig 5. Summary of experimental serum mapping of HIV-1 sera from LTNPs and TPs with a broad neutralization
by ELISA techniques. Then, in these sera the neutralization may be due mainly to the presence of high-affinity antibodies against other domains. When we analyzed the sera samples with the standard serum mapping, we observed that V3 glycan-dependent HIV-1 NAbs were more abundant in LTNPs (11/12) than in TPs (12/33). This prevalence of the V3 glycan-dependent NAbs in LTNPs was also detected with the neutralization-based analysis. A previous study has also found high levels of 2G12-like antibodies in broadly neutralizing samples from LTNPs [47]. A question that arises from these results is whether these antibodies are contributing to the control of viremia in LTNP. The V3 region is known to be highly immunogenic and individuals develop antibodies directed against the C3-V4 region early in infection [48,49]. HIV overcomes the response through mutation but this variability is decreased in the low viral rates of replication. It could be possible that in some LTNPs bNAbs contribute to viral control. Actually it has been recently described that Abs from one EC patient can exhibit autologous neutralization and these antibodies are contributing to elite control in this individual [22]. To address this hypothesis autologous neutralization should be detected in LTNP patients but due to extremely low viral loads, viral isolation and cloning of the envelope is unfeasible in the majority of patients. For each serum the predicted relative prevalence of each reference antibody cluster is presented as a heat map where stronger neutralization signals by the antibody cluster are shown as darker colors (higher fractional numbers). Breadth was measured as the percentage of viruses neutralized with an ID50>50 and potency as the geometric mean ID50 titers of sera. Sera with breadth less than 20% are greyed out since the computational predictions in those cases are less reliable (B) Concordance between neutralization-based and standard serum mapping. Sera with breadth less than 20% were excluded. The agreement between the two methods was good. In 92% of the sera at least one of top two computational neutralization-based predictions was identified by standard mapping. (C) Frequency of the epitope specificities of HIV-1 neutralizing antibodies in the sera from LTNPs and TPs with broadly neutralizing activity. Sera with breadth less than 20% were excluded. https://doi.org/10.1371/journal.pone.0193773.g006

Characterization of bNAbs to HIV-1 in LTNPs
In some subjects neutralization breadth was mediated by more than one antibody epitope specificity which is in agreement with previous observations showing that broadly neutralizing activity of some HIV-1 infected individuals is due to antibodies that target more than one epitope [38,[50][51][52].
bNAbs require a large number of somatic mutations [7] that are related with preservation of T follicular helper cells (TFH). HIV infected individuals have several defects in the humoral immune system, including B cell abnormalities associated with HIV replication-induced immune cell activation and TFH priming [53][54][55]. Elite controllers with viral load consistently below 50 copies/ml could develop a robust response mediated by their well-preserved B cells generating high affinity antibodies. Therefore control of the viral load could be associated to an improved maturation of antibodies in the affinity for the antigen. TFH are involved in the development of these antibodies as B cell memory maturation and generation of high-affinity neutralizing antibodies is dependent on extensive signaling from TFH cells [56]. However, in productive HIV infection, high levels of HIV viremia drive the expansion of TFH cells which is associated with perturbation of the B cell compartment, resulting in deregulated antibody production [57,58]. One potential hypothesis could point to a better preserved B cell function in LTNPs including appropriate regulation of TFH resulting in a generation of bNAbs with high levels of somatic hypermutation despite lower levels of antigen.