Strategies For Eliciting Multiple Lineages of Broadly Neutralizing Antibodies to HIV by Vaccination

Abstract

A prophylactic vaccine would be a powerful tool in the fight against HIV. Passive immunization of animals with broadly neutralizing antibodies (bnAbs) affords protection against viral challenge, and recent data from the Antibody Mediated Prevention clinical trials support the concept of bnAbs providing protection against HIV in humans, albeit only at broad and potent neutralizing antibody titers. Moreover, it is now clear that a successful vaccine will also need to induce bnAbs against multiple neutralizing epitopes on the HIV envelope (Env) glycoprotein. Here, we review recent clinical trials evaluating bnAb-based vaccines, and discuss key issues in the development of an HIV vaccine capable of targeting multiple Env neutralizing epitopes.

Introduction

Vaccines that induce broadly neutralizing antibodies (bnAbs) are a leading objective for HIV vaccine research. The recently published results of the Antibody Mediated Prevention (AMP) trials provided proof-of-concept that a bnAb-inducing vaccine could also protect people from infection [1]. However, the AMP trials also demonstrated an HIV vaccine that elicits only one specificity of bnAb would likely not be clinically effective. For an HIV vaccine to be successful, it will need to elicit bnAbs against several distinct epitopes on the envelope glycoprotein (Env). In this review, we discuss results of clinical trials of bnAbs for passive immunization and prophylaxis, and review strategies for design and delivery of bnAb-based vaccines.

HIV broadly neutralizing antibodies for prevention

BnAbs prevent HIV infection by binding to conserved epitopes on Env, thereby interfering with the viral machinery to enter target cells. Seven neutralizing sites of bnAbs have been identified on the Env gp120-gp41 trimer complex: sites on the first and second variable regions (V1/V2-glycan), sites on the third variable region (V3-glycan), the CD4 binding site (CD4bs), the gp120-gp41 interface, the membrane-proximal external region (MPER), the Env silent face, and the fusion domain [2–4] (Figure 1A). An eighth bnAb type, Fab-Dimerized Glycan-reactive (FDG) antibodies, represents a new class of bnAbs that bind only to glycans and can target multiple glycan sites [5]. Prototypical bnAbs for each neutralizing epitope, some of which are being tested in passive immunization studies for their capacity to protect people from infection, are shown in Figure 1B.

Figure 1:

Neutralizing epitopes on the HIV Envelope glycoprotein. (A) Neutralizing epitopes are mapped on a representative HIV Env trimer and shown in top view (left), side view (middle), and bottom view (right). BnAb epitopes are indicated by color: green, V3-glycan; red, V1/V2-glycan; blue, CD4bs; orange, fusion domain; magenta, Env silent face; cyan, gp41-gp120 interface. The position of the MPER region is shown in the side view and the first amino acid of the MPER peptide is marked with an asterisk in the bottom view. (B) Examples of bnAbs for each neutralizing epitope are listed and color-coded according to their epitope. An additional bnAb category, Fab-Dimerized Glycan-reactive (FDG) bnAbs have been recently described that bind to multiple Env glycans.

Can bnAbs protect humans from HIV infection?

BnAb infusions of humanized mice and non-human primates protect against HIV viral challenge [6–10]. A meta-analysis of rhesus macaque studies of bnAb administration followed by SHIV (chimeric simian-human immunodeficiency virus) challenge demonstrated that serum neutralization titer has a significant effect on infection risk and that viral protection by a single bnAb likely will require relatively high levels of neutralization at the time of exposure [11]. Administration of bnAbs to HIV-infected people transiently reduces viremia [12], confirming that bnAbs exert anti-viral effects in humans as well. The critical follow-up question has been do bnAbs protect people from HIV infection and, if so, what serum titer of bnAb is needed for protection?

The recently published results of the AMP trials [HIV Vaccine Trials Network (HVTN) 704/HIV Prevention Trials Network (HPTN) 085 and HVTN 703/HPTN 081] offer important insights [1]. In two parallel trials, at-risk cisgender men and transgender people (HVTN 704) or at-risk women (HVTN 703) were infused every eight weeks with a high or low dose of the CD4bs bnAb VRC01 or placebo for a total of ten infusions. VRC01 infusion did not provide a significant reduction in overall HIV transmission, but the frequency of transmission of VRC01-sensitive viruses was significantly lower in the treatment groups than in the placebo groups, with VRC01 infusion providing 75% protective efficacy against sensitive viruses [1]. This result provided proof-of-concept that passive immunization with bnAbs can protect people against viral acquisition. The lack of overall protection against HIV infection, however, highlights that any strategy for bnAb-based HIV prophylaxis will require consistently high titers of neutralizing antibodies against a wide breadth of HIV isolates.

The results of the AMP trials suggest that a mixture of bnAbs targeting different neutralizing epitopes will expand protection against various HIV isolates, and results of ex vivo neutralization studies using combinations of bnAbs targeting different neutralizing epitopes support this idea [13,14]. Clinical trials are underway to test whether infusion of two or three bnAbs targeting different epitopes indeed provide better protection (Table 1). An ongoing phase I clinical trial (NCT04173819) is evaluating the safety, tolerability, and pharmacokinetics of the combination of a CD4bs and a V3-glycan bnAb. Similar trials are testing different combinations of two or three antibodies that target different bnAb epitopes (NCT04212091 and NCT03928821). An alternative to using bnAb combinations is to engineer single antibodies bearing two or three different antigen binding sites, each of which binds a different bnAb epitope. Two phase I trials are underway to test the safety of bi- or tri-specific antibody approaches (NCT03875209 and NCT03705169).

Table 1:

Representative clinical trials of passive infusion of multiple bnAbs

Clinical trial	Concept	Trial Phase
NCT03571204	Infusion of 3BNC117 (CD4bs) and 10-1074 (V3-glycan) bnAbs	Phase 1
NCT04173819	Infusion of 3BNC117 (CD4bs) and 10-1074 (V3-glycan) bnAbs, i.v. vs. s.c.	Phase 1/2
NCT04212091	Infusion of PGT121 (V3-glycan) and VRC07 (CD4bs) bnAbs	Phase 1
NCT03928821	Infusion of combinations (two or three mAb) of PGT121 (V3-glycan), PGDM1400 (V1/V2-glycan, 10-1074 (V3-glycan), andVRC07 (CD4bs)	Phase 1
NCT03875209	Bi-specific antibody injection 10E8.4/iMab (MPER + CD4)	Phase 1
NCT03705169	Tri-specific antibody SAR441236 (CD4bs + V1/V2-glycan + MPER)	Phase 1

Can vaccines elicit bnAbs?

To date, bnAbs have only been induced by retroviral infection in HIV-infected individuals or SHIV-infected rhesus macaques [15,16]. Unfortunately, vaccination has not reliably induced potent bnAbs in animal models or human clinical trials. However, preclinical and clinical studies have informed our understanding of the immunological roadblocks in generating bnAbs through vaccination.

BnAbs exhibit unusual properties, including long heavy-chain complementarity-determining region (CDR) 3 loops and/or autoreactivity, which normally trigger immune tolerance mechanisms that prevent the development or activation of B cells bearing these receptors [17,18]. In addition, mature bnAbs contain a large number of somatic mutations and, in some cases, insertions or deletions in the immunoglobulin heavy and light chain genes, suggesting that prolonged periods of germinal center occupancy are needed for bnAb development [17]. Thus, the two major challenges for HIV vaccine development are: 1) design of “germline-targeting” immunogens that can activate rare and low-affinity bnAb precursor B cells, and 2) selection of bnAb precursor B cells to enter and stay in germinal centers in order to accumulate the mutations required for potency and breadth.

Two approaches have been used in the design of bnAb germline or unmutated common ancestor (UCA)-targeting immunogens: 1) well-folded, native Envs based on transmitted/founder viruses isolated from HIV-infected people who eventually generated bnAbs, and 2) proteins engineered to bind UCAs and mimic Env neutralizing epitopes. A key consideration in the development of a germline/UCA targeting immunogen is the affinity of the immunogen for B cell receptors (BCRs) expressed by naïve bnAb precursor B cells [19]. Transgenic mice, in which a subset of naïve B cells express the computationally inferred UCA precursor BCR for a certain bnAb lineage, are powerful tools for testing the capacity of novel immunogens to activate B cells with bnAb potential in vivo [20–23]. After immunization, serial analysis of the breadth of neutralization induced in plasma, the frequency of bnAb precursors induced in blood and lymphoid tissues, and the mutations induced in bnAb B cell lineage members provides insight as to how well a immunogen formulation is performing and predicts how well a vaccine-induced bnAb type or class will protect against viral strains. It is also important to determine if a candidate germline-targeting immunogen can activate not only inferred germline precursors in mice, but also authentic naïve B cells in humans. For example, B cells that bind the N332-GT2 and eOD-GT8 immunogens have been found in the naïve human B cell repertoire [22,23]. Isolating and characterizing reactive naïve human B cells will be critical for evaluating other germline targeting immunogens, although this process can be very challenging, since naïve germline precursors, particularly those for long HCDR3-bearing bnAbs, are rare and of low affinity [19].

Success with various germline or UCA targeting immunogens in pre-clinical studies has led to their evaluation in human clinical trials (Table 2). The eOD-GT8 60mer nanoparticle was recently tested in a phase I clinical trial (NCT03547245, IAVI G001), with early encouraging results [24]. In this trial, 48 HIV negative adults received two doses of eOD-GT8 or placebo two months apart. Most of the immunized individuals (97%) who received the eOD-GT8 60mer exhibited activated B cells with features of germline precursors of the VRC01 lineage.

Table 2:

Immunogens that target B cell germline precursors for bnAb lineages

Immunogen	Epitope	BnAb GL or UCA targeted	Clinical Trial
CH505 TF gp120	CD4bs	CH103	HVTN 115 (NCT03220724)
CH848 10.17DT NP	V3-glycan	DH270.6	HVTN 3XX
BG505 MD39.3 Trimer	V3-glycan	PGT121	HVTN 302
eOD-GT8	CD4bs	VRC01	IAVI G001 (NCT03547245)
BG505 SOSIP.GT1.1 gp140 trimer	CD4bs	VRC01	IAVI C101 (NCT04224701)
CH505 TF SOSIP Trimer (Clade A BG505 backbone)	CD4bs	CH103	HVTN 300
426c Core NP	CD4bs	VRC01	HVTN 301
MPER-656 peptide-liposome	MPER	2F5, 10E8, 4E10, DH511	HVTN 133 (NCT03934541)

Design of bnAb boosting immunogens

The results of germline/UCA targeting immunogens in animal studies and early clinical trials have raised optimism that bnAb precursors are present and functional in the naïve B cell pool. However, germline/UCA targeting immunogens alone have yet to induce the degree of somatic mutation in bnAb lineages needed for neutralization breadth. Thus, there is considerable focus on the use of sequential boosting immunogens that mature bnAb lineages in germinal centers [17].

The absolute number of mutations required to generate bnAbs from germline precursors represents a substantial hurdle in itself, but research in recent years has revealed additional challenges that must be overcome for a boosting immunogen to advance a bnAb lineage. Most bnAbs require somatic mutations in the framework regions of their immunoglobulin genes, which facilitates neutralization breadth by conferring the antibody with greater flexibility in antigen binding sites [25]. However, framework region mutations are less tolerated than mutations in antigen-contacting CDR loops, which may negatively impact B-cell fitness in germinal centers [25]. In addition, bnAbs are enriched for mutations in “cold” spots, sequences of immunoglobulin genes that are seldom subject to somatic mutation, and thus represent improbable events [26]. Lastly, approximately 40% of bnAbs require in-frame insertions or deletions, which occur only rarely during somatic hypermutation [27]. Thus, an effective boosting immunogen will need to bind BCRs on the mutated descendants of bnAb germline precursors, and then promote the selection of B cells expressing BCRs that were produced through rare events in the somatic hypermutation process. Nonetheless, the feasibility of the sequential immunization strategy has been demonstrated in mice, where immunization of bnAb precursor transgenic mice with germline targeting immunogens followed by boosts with heterologous immunogens results in greater neutralization breadth than repeated immunization with germline targeting immunogens alone [28,29].

An important consideration in the design of Env immunogens is the presence of immunodominant non-neutralizing epitopes that can take B cell lineages “off track.” These epitopes are normally occluded on the Env trimer but are exposed upon receptor CD4 binding or upon disassembly of Env trimers into monomers. To ensure that Env immunogens are maintained in a well-folded conformation following administration, multiple mutations have been integrated into immunogens that stabilize the Env trimer and minimize the exposure of undesired off-target epitopes [30–32].

Designing an HIV vaccine capable of inducing multiple specificities of bnAbs

Can germline targeting and sequential immunization approaches be adapted for a multi-bnAb vaccine? One might envision a regimen of immunogen cocktails, with a priming cocktail capable of activating bnAb germline precursors for various neutralizing epitopes, and subsequent boosting cocktails containing the next immunogens in each sequence that mature different bnAb lineages (Figure 2). With the induction of bnAbs against multiple epitopes would come protection against a broad spectrum of HIV strains (Figure 2).

Figure 2:

The sequential immunization approach for a single vs. multi-bnAb HIV vaccine. For a bnAb-inducing vaccine that targets a single neutralizing epitope (top), a germline (GL) targeting immunogen is administered for immunization #1 to activate bnAb precursors. Subsequent boosting immunizations rely on modified immunogens that favor the acquisition of mutations in immunoglobulin genes that increase neutralizing breadth. Induction of bnAbs against a single neutralizing epitope, however, will offer protection only against a subset of sensitive viruses. In contrast, a multi-bnAb vaccine (bottom) would rely on multiple immunogens that induce antibody lineages against different neutralizing epitopes. With multiple bnAbs induced, vaccine recipients will be protected against larger fraction of viruses.

An important consideration for a multi-bnAb vaccine will be the method for delivering several immunogens in one injection. Strategies to deliver multiple immunogens can be categorized into three approaches: 1) mixing of individual immunogens; 2) design of chimeric immunogens featuring targeting motifs for different neutralizing epitopes; and 3) display of different immunogens on the surface of a nanoparticle. The mixing of different immunogens to elicit antibodies against a variety of pathogen variants has been exploited for a variety of vaccines, including the pneumococcal vaccine, in which polysaccharides of different bacterial serotypes are combined to elicit an array of serotype-specific antibodies [33]. The efficacy of multivalent pneumococcal vaccines derives largely from the highly immunogenic nature of the bacterial polysaccharides, to the point that antibodies can be generated without T cell help. Thus, mixing of germline targeting or boosting immunogens for HIV may represent the simplest approach, but may lack the necessary immunogenicity to drive bnAb maturation in germinal centers. Moreover, a mixed HIV vaccine may suffer at the hands of pragmatism; the hypothetical immunization strategy shown in Figure 2, comprising a four-injection regimen with four bnAb-inducing immunogens per immunization, represents sixteen individual immunogens to undergo regulatory review and manufacture.

A chimeric immunogen approach, in which targeting motifs for different neutralizing epitopes are incorporated onto a common backbone, can overcome the “too many immunogens” issue. Medina-Ramirez et al. demonstrated in transgenic mice that a two-epitope germline targeting protein was able to bind and activate germline precursors of two bnAb lineages, VRC01 and PGT121 [34]. Similarly, Saunders et al. demonstrated that a native Env-based CD4bs germline targeting immunogen was capable of eliciting V1/V2-glycan targeted neutralizing antibodies [35]. It is unclear, however, how many neutralizing epitopes can be incorporated into a single immunogen, and whether the correct sequential epitopes can be readily incorporated into the same molecular backbone.

Multimerizing Env immunogens on nanoparticles has become an important strategy to overcome the lack of Env immunogenicity, and it may have the added bonus of offering a platform for delivering different HIV immunogens on the same molecule. There is growing evidence that viral immunogens arrayed on a particle enhance B cell activation and induce higher antibody titers than soluble forms of the same immunogen [21,36–38]. Particle-based strategies include display on a variety of structures such as virus-like particles, liposomes, lipid nanocapsules, silica nanoparticles, and ferritin nanoparticles. The clinical trial of eOD-GT8 60mer, for example, used a multimerization strategy in which an average of 60 copies of the eOD-GT8 immunogen were arrayed on a bacterial lumazine synthase-based particle [39]. The HVTN 133 trial studies an MPER neutralizing epitope embedded in a liposome to mimic epitope presentation on a viral membrane, and in macaques initiated an MPER bnAb lineage [40]. Brouwer et al. engineered a self-assembling nanoparticle that is capable of displaying a variety of different Envs; in their original report, homogenous nanoparticles of Envs were tested, but platforms of this type may allow for the display of different immunogens on the same nanoparticle [41]. Indeed, recent studies have revealed the feasibility of displaying multiple immunogens on the same particle. Cohen et al. generated nanoparticles displaying Spike proteins from eight different coronavirus species as a strategy to elicit multiple neutralizing antibody lineages [42]. Most recently, a scaffolded nanoparticle quadrivalent influenza vaccine decorated with four different hemagglutinin molecules has been reported that induced broad influenza protection [43].

Multiple protein-based immunogens, however, may prove logistically impractical or prohibitively expensive for a multi-bnAb approach. Thus, an alternative strategy could employ nucleoside-modified mRNAs to encode multiple HIV immunogens. Modified mRNA vaccines offer many features that are amenable to a multi-bnAb, sequential immunization approach [44,45]. Modified mRNAs are simpler and faster to manufacture than traditional protein immunogens and, importantly, are potent inducers of T follicular helper cells and B cell germinal centers [46]. Moreover, multiple immunogens can be encoded as modified mRNAs and delivered in lipid nanoparticles: a modified mRNA vaccine encoding multiple influenza antigens protected mice against a panel of different influenza viruses [47]. Results from pre-clinical studies for modified mRNA-based, single bnAb HIV vaccines are ongoing [48,49], but studies with multiple co-administered modified mRNA-encoded immunogens are needed.

Conclusion

Considerable knowledge has been gained over the past three decades about the barriers to HIV vaccine-induced bnAb generation. The B cell germline precursors for bnAb lineages are rare, some are disfavored by immune tolerance mechanisms, and the path from germline precursors to bnAbs is torturous, requiring rare somatic hypermutations. Nonetheless, pre-clinical studies in animals suggest that optimal combination of immunogens may be able to promote bnAb development.

However, the results of the AMP trial were clear: inducing bnAbs can protect against susceptible HIV strains but the titers needed to protect will be relatively high, and it will require antibodies to multiple epitopes to protect against viral diversity. If a bnAb against a single neutralizing epitope can be guided through full maturation, the next challenge will be to design vaccines that elicit bnAbs against multiple neutralizing epitopes to maximize coverage against viral variants. Challenges for a multi-bnAb vaccine will likely involve the format for delivering multiple immunogens in a manner that elicits optimal responses from B cell germline precursors and intermediates for different bnAb lineages. With the development of chimeric immunogens, novel nanoparticle platforms, and modified mRNA-based vaccines, there is a growing toolbox for the delivery of a multi-bnAb HIV vaccine.