Molecular mechanisms for direct sensing of virus-like antigens by B cells

Abstract

Particulate antigens (Ags) such as viruses can often induce strong B-cell responses in vivo very effectively, but the molecular determinants of this complex process remain incompletely understood. In this review, we focus on recent mechanistic insights into the earliest steps in the initiation of primary B-cell responses to viruses, gained by exploiting a new generation of model particulate Ag, synthetic virus-like structures. We also review the characteristics of the resulting short- and long-term antibody (Ab) responses in mice. These studies reveal that a repeating pattern of epitope display on a virus-sized scaffold is a fundamental biophysical feature of viruses that triggers a qualitatively distinct mode of B-cell Ag receptor (BCR) signal transduction relative to soluble Ag display, and consequently serves as a standalone danger signal for Ag-specific B-cell activation. Quantitative variation in epitope density (ED) on such scaffolds modulates the degree and quality of B-cell activation both in vitro and in vivo. The presence of internal nucleic acid (iNA) in the interior of these virus-like structures can profoundly influence the resulting Ab responses for the lifespan of immunized animals. We conclude that the ED of viral surface Ags and the iNA genomes provide two essential signals that together are sufficient for B-cell activation and Ab production during antiviral responses. We place these findings in the context of the literature, discuss implications for rational vaccine design, and highlight unanswered questions to guide future research directions.

B cell responses to viral particulate antigens

Introduction

Edward Jenner’s work on vaccination against smallpox is widely regarded as the dawn of immunology as a scientific discipline (1, 2). It is well documented that foreign particulate antigens (Ags), such as viral particles, can elicit protective antibodies (Abs), and this forms the basis for over a dozen antiviral vaccines that have been licensed by the US Food and Drug Administration (3). However, how such particulate Ags activate and direct the immune system to produce protective and durable Ab responses remains incompletely understood. Mechanistic details of B-cell antigen receptor (BCR) signaling, B-cell activation, as well as Ag internalization, processing, and presentation to cognate T cells have each been defined with widely used tools to ligate the BCR (e.g. anti-IgM Fab’2) and well-established model Ags such as hapten–protein conjugates. However, how these different B-cell pathways are qualitatively and quantitatively modulated by encounter with a structurally complex Ag, such as a viral particle, is far less well understood. As a result, we lack a systematic understanding of the mechanistic basis of viral immunogenicity, and we only partially appreciate the molecular features essential for the design of an optimized vaccine to elicit a potent and protective Ab response to viruses.

Given the importance of B cells in antiviral immunity, here we review the current knowledge on B-cell responses to virus-like particulate Ags, highlighting in context recent work from our groups. We focus on the mechanisms by which fundamental biophysical features of viruses are sensed and interpreted by B cells, as well as the impact of this recognition on B-cell responses. In organizing this review, we follow the biological order of events that should occur in order to produce high-affinity Abs: from entry points that facilitate B-cell encounter with a particulate Ag, to recognition, internalization, and evolution of humoral responses in vivo. We consider the impact of fundamental viral features that constitute minimal determinants for B-cell recognition on the characteristics of resulting Abs and point out important questions that remain.

Entry points for direct B-cell encounter with particulate Ags

Uptake of cell-free particulate matter from the interstitial fluid to the lymphatics, and its subsequent accumulation in regional draining lymph nodes, has been directly observed and measured (4, 5). This has the potential to serve as a nonspecific yet efficient mechanism for the collection of viruses, regardless of their molecular features, from peripheral tissues to lymph nodes. Once such free viruses and virus-like particles reach lymph nodes, professional Ag-presenting cells, including subcapsular sinus macrophages and dendritic cells, can capture and present such particulate Ag directly to Ag-specific B cells for recognition (6–13). Early studies by Ochsenbein et al. showed that marginal zone macrophages in the spleen also help capture blood-borne viral Ags for presentation to B cells in a manner that depends on complement receptors, and this pathway is important to elicit the T-independent neutralizing-IgM response toward vesicular stomatitis virus (VSV), poliomyelitis virus, and recombinant vaccinia virus (14). Later in the course of an adaptive T-dependent immune response, follicular dendritic cells in the germinal center capture and present Ag—including intact virions—via the complement receptor (15, 16). This Ag capture and display is facilitated by low concentrations of virus-specific Ab that arise early in the immune response (reviewed in (17)), and can be further potentiated by the low protease activity in B-cell follicles to preserve the structural integrity of captured Ag (13, 18, 19). Collectively, these studies reveal several pathways for direct presentation of intact particulate Ags to B cells in vivo at both early and late stages of a physiological immune response.

Direct recognition of intact virions by Ag-specific B cells: significance for the B-cell antiviral response

In vivo encounter of B cells with particulate Ags in undigested form implies that the native structure of the Ag, including complex conformational Ag epitopes and the biophysical organization of the supramolecular structure of a virion, can be directly presented to Ag-specific B cells for recognition by cognate BCRs. The capture and internalization of intact virions by B cells further enable B cells to present physically linked Ags to elicit T-cell help, although this is influenced by particle size and complexity (20–22). This expands the supply of co-stimulation a B cell can access and is critical for entry to and competition within the germinal center (23). Even more importantly, direct B-cell encounter with intact viral structures facilitates the generation of neutralizing Abs (nAbs). Lessons from licensed vaccine trials suggest that nAb titer is the single best correlate of immune protection in the majority of cases (3). An nAb must bind viral surface Ags mostly in the context of a native virus, and serves directly as an inhibitor of viral entry by blocking interaction with and productive infection of host cells by the virus (24). To produce these types of nAbs, B cells must recognize entry-productive conformations of viral surface Ags through their Ag receptor (25) rather than the denatured structure of a digested peptide/linear epitope derived from the virus. More fundamentally—as we review below—accumulating evidence suggests that direct B-cell recognition of intact virions profoundly impacts B-cell signaling and activation, and this is critical to generate nAbs and mount antiviral defense.

The matter of epitope density and related Ag features

Direct in vivo encounter of B cells with particulate Ags in their native state implies that the fundamental biochemical and biophysical features of a viral particle can be directly presented to Ag-specific B cells for recognition. Two universal and essential attributes shared by all animal viruses are the display of viral-specific proteins, with certain spatial density, typically in a patterned array on the surface of individual virions, and the encapsulation of viral genetic material inside the virion. Naturally occurring virions exhibit diverse spatial densities of viral surface proteins, ranging from 200 to 30 000 molecules per μm², which represents a variation over 2 orders of magnitude (3). Direct presentation of particulate Ags to B cells implies that Ag-specific B cells will encounter this variation of epitope density (ED).

The concept of Ag valency and avidity (when combined with epitope affinity) is the subject of intense investigation to identify optimal Ag characteristics for maximal immunogenicity (26–31). Prior work has shown how powerful valency and avidity are at driving robust germinal center and extrafollicular B-cell responses, facilitating recruitment of rare and low-affinity B-cell precursors in the face of interclonal competition, and thereby promoting clonal diversity (32–35). Importantly, the concept of ED is related to but distinct from valency, and we propose that this is a defining biophysical property of viruses recognized by B cells; in the context of naturally occurring viruses, we define ED as the number of a specific viral surface epitope molecules per unit area of virion surface (3). Therefore, ED is related to both the total number of a specified viral surface Ag and the physical size of the virion structure. For virions of similar size, the comparison of ED can be reduced to the total number of a viral-specific Ag displayed on a single virion surface. However, it is important to distinguish ED (Fig. 1A and B) from the valence of an Ag (Fig. 1C). Valence of an epitope specifically refers to the number of epitopes in one unit of Ag that are available for BCR binding, independent of the physical size of the Ag or epitope spacing. This term valence was originally used in chemistry to denote the degree of the combining capacity of an element with other monovalent elements, such as hydrogen. ED, on the other hand, is not only dependent upon the valence of an Ag, but also upon the physical size of the virion structure. The physical size of a virion impacts the average spacing in between virion epitopes (Fig. 1A), a biophysical feature that is critical for recognition by BCRs, which exhibit nanoscale organization on the B cell surface (36, 37). For example, many decades ago Dintzis et al. first showed that the spacing in between epitopes, rather than valence per se, can determine the immunogenicity of a T-independent Ag (27); although 12–16 properly spaced DNP hapten epitopes per polymer gave efficient responses in their study, increasing valence to 25 DNP groups per polymer was not immunogenic, suggesting that average spacing between haptens modulates the B-cell response. More recently, newer strategies such as DNA origami have suggested that epitope spacing and Ag size have functional significance (38–40); B-cell calcium signaling was shown to be sensitive to the physical spacing of repeating epitopes on DNA origami nanoparticles in vitro (38). Moreover, it is important to note that valency and spacing thresholds for B-cell responses in vivo may well depend on context, such as B-cell precursor frequency and clonal competition, epitope affinity, along with co-stimulatory signals generated by nucleic acid (NA) recognition and T-cell help, and naïve or recall context. Lastly, the concept of ED was also used in studies that involved Ags immobilized on a planar lipid bilayer (29) (Fig. 1D), a very useful in vitro model for studying B-cell responses toward Ags presented on a cell surface (41). However, a virus-like particulate Ag as a whole can be internalized through BCR-mediated endocytic pathways (Fig. 1E) (see section below: "BCR-mediated particle internalization"), in contrast to the extraction of the same Ags on a planar lipid bilayer (41–43). We postulate that B-cell sensitivity to ED on virus-like structures can generate uniquely immunogenic responses upon recognition of these biophysical features.

Figure 1.

Epitope density (ED), valency, and synthetic virus-like structures (SVLS). (A) A virus-sized particulate Ag that displays a protein epitope at a relatively low density on its surface. This low ED is also manifested as the relatively large spacing between individual epitopes. (B) A virus-sized particulate Ag that displays a protein epitope at a relatively high density on its surface. This high ED is also manifested as the relatively close distance between individual epitopes. (C) Ag valence refers to the number of epitopes attached to a single structure, such as a cylindrical-shaped protein molecule. (D) Protein Ags displayed on the surface of a lipid bilayer as a model for the study of B-cell recognition of membrane-bound Ags. (E) Direct endocytosis of virus-like particulate Ags by Ag-specific B cells, in which viral features such as ED and internal nucleic acids are further sensed and integrated for downstream signaling. (F) Schematic of SVLS built around neutral liposomes that quantitatively incorporate two essential features of biological viruses: conformation-preserved and orientation-specific display of epitope at programmed surface density, and internal nucleic acids (iNAs). T-cell epitope peptides can also be encapsulated within the interior of these structures for recruitment of additional T-cell help to promote B-cell responses.

A novel platform of model Ags to dissect B-cell response to fundamental features of virus-like structures

To study the mechanisms of B-cell responses to virus-like particulate Ags, we have developed a liposome-based platform of synthetic virus-like structures (SVLS) (Fig. 1F) that incorporates and facilitates independent titration of the essential features of virions discussed above (44, 45). We used this platform to generate a library of SVLS that display an Ag of choice across a range of programmable ED by taking advantage of maleimide chemistry to conjugate a desired number of Ag to the liposomal surface at a defined orientation. This platform enables us to dissect the biological effect of ED in isolation as well as in combination with internal nucleic acids (iNAs) for the evaluation of synergistic effects. Importantly, SVLS incorporate ED and iNAs simultaneously in a single synthetic structure similar to that of a virion without the leakage of iNAs, which is essential for B-cell intrinsic endosomal TLR activation. These iNAs are oligomeric NA molecules with fully native phosphodiester backbones that are incorporated on demand to mimic natural viral genomes. Moreover, by incorporating affinity variants of an Ag, an additional parameter of high relevance to naïve B-cell responses can be layered into the library as a variable titrated independently from ED. The entire system of SVLS is prepared using highly purified materials, including neutral lipids, proteins, and NAs. This is critical for dissecting the molecular basis of virion immunogenicity because any impurities, present even in small amounts, could act as potential adjuvants for Ab responses in vivo or B-cell responses in vitro, complicating data interpretation. This family of virus-like particulate Ags enables us to interrogate ED, epitope affinity, and iNAs individually and collectively for their effects on B cells both in vitro and in vivo, in ways that were not possible before.

Recent studies we have conducted using this new generation of model particulate Ags have revealed that such a patterned display of epitopes potently directs the earliest steps of B-cell activation, differentiation, and class-switch recombination even in the absence of cognate T-cell help or any other adjuvants (46–48). Such structures even license B cells to overcome self-tolerance (46). Moreover, internalization of an intact viral particle upon BCR-mediated endocytosis (see the section below: "BCR-mediated particle internalization") enables Ag-specific B cells to interact with contents inside virion particles, which include viral NA genomes and other structural proteins that can serve as T-cell epitopes. Internalized viral NAs can be detected by endosomal NA-sensing Toll-like receptors (TLRs) and can promote both B-cell expansion and differentiation. Importantly, the dual engagement of BCRs and NA-sensing TLRs can synergistically enhance Ab responses, but systematic and quantitative exploration of this process remains to be pursued (49, 50).

Notably, this standalone B-cell activation by viral structures suggests that the immune system can rapidly initiate an effective antiviral IgG response via clonal expansion of virus-specific B cells without the substantial time delay required to search for and recruit cognate T-cell help; this is especially critical for a naïve host in which both T- and B-cell responders will be rare. Moreover, early secretion of Ab can positively feed back to enhance humoral immune responses (17) and produce multiple levels of effector functions for antiviral defense, all of which can take place without cognate T-cell help (51–53).

Molecular basis for potent B-cell response to particulate Ag

Although studies of in vivo responses to multivalent Ags are extensive, we have a limited understanding of early events in BCR signal transduction initiated by complex multivalent Ags, in contrast to those “rules” established for soluble Ags (see reviews (54, 55)). In vitro studies by Batista, Pierce, Tolar, and others evaluated B-cell responses to Ag displayed on large surfaces (e.g. cell membrane, planar lipid bilayer, or even coated on plastic) (42, 56–63). These studies revealed unique biology relative to soluble monovalent Ag. Imaging approaches revealed that B-cell encounter with Ag on a cell membrane produces a micron-scale, synapse-like structure that recruits the co-receptor CD19 to enhance BCR sensitivity to such Ag (59, 61). Conversely, inhibitory B-cell co-receptors such as FcγRIIB and CD22 are gradually excluded from such synapses (56, 60). This reorganization of membrane co-receptors is postulated to enhance B-cell response to membrane Ag; even low-affinity Ag is potent when immobilized (56, 64). Engagement of such co-receptors by Ag context has a profound impact; Pierce and colleagues suggested that co-engagement of FcγRIIB with the BCR was sufficient to suppress immune synapse organization by membrane Ag (60). Similarly, forcible engagement of CD22 can suppress B-cell signaling and activation in response to a particulate model Ag (65). Conversely, soluble Ag tethered to complement can engage the CD21/CD19 complex to potentiate BCR signaling, reducing the B-cell activation threshold by many orders of magnitude and enabling very low-affinity Ag to trigger responses (66).

What, though, are the mechanisms for the potent and standalone response of B cells to ED on a virus-like particle? We created a library of SVLS with internal phosphate-buffered saline (PBS) that displays two different HEL affinity mutants (HELD = HEL2X 8 × 10⁷ M⁻¹ and HELT = HEL3X, 1.5 × 10⁶ M⁻¹ affinity for the well-studied Hy10 BCR) (67) with a broad spectrum of ED that mimicked naturally occurring viruses. All these structures have diameters around 120 nm (48), which is very close to that of HIV, influenza virus, and SARS-CoV-2 (68–70). Using primary mouse Hy10 BCR Tg naïve B cells, we revealed exquisite sensitivity to ED, and complete insensitivity to affinity changes over 50-fold down to physiologic naïve µM affinity epitopes—in marked contrast to soluble Ag (47).

We compared soluble and particulate formats of HEL affinity variants and discovered a dramatic potency gap; HELD/2X on liposomes at optimal density (250 ED approx.) gave a 3 orders of magnitude boost in potency as assessed by upregulation of markers of B-cell activation, while the gap in EC50 for HELT/3X approached 5000×. This was reminiscent of analogous comparisons made by Batista et al. for membrane-immobilized Ag (56). This potency gap implies that naïve B cells can readily detect virus-like epitope display while ignoring self-like soluble display of identical epitopes and suggests that B cells can discriminate biophysical structures of foreign and self Ags. To our surprise, when we analyzed more proximal signaling events such as Erk phosphorylation and cytosolic Ca²⁺ (comparing equipotent doses of soluble and particulate format Ag), we discovered remarkable and unique digital signaling behavior triggered by particles such that titration of particle number altered the fraction of responding cells but not the magnitude of responses within individual cells. Such population-level all-or-none behavior by B cells implies an amplification step in the signal transduction cascade and indeed this is precisely what we identified for particulate Ag; we found minimal tyrosine phosphorylation of proximal BCR signaling components like CD79, Syk, and identified BCR signal amplification at the level of the second messenger phosphatidylinositol 3,4,5-triphosphate (PIP3) in response to SVLS but not soluble Ag (47). Nevertheless, robust downstream signal amplification by SVLS was entirely dependent on the proximal BCR pathway (Syk, Btk, PI3K) (47). Notably, this finding emphasizes that mere cross-linking by high valency Ag (which would trigger enhanced signaling at the most proximal point in the cascade) does not account fully for the remarkable potency we identified for such particles.

We showed genetically that this signal amplification at the level of PIP3 by SVLS is due to evasion of LYN-dependent inhibitory signaling downstream of ITIM-containing co-receptors (47) (Fig. 2). Interestingly, PIP3 hydrolysis by inositol phosphatases PTEN and SHIP1 is critical to suppress B-cell responses to self Ag (71–73). Indeed, we found that SVLS stimulation partly overcomes B-cell anergy (47). This makes teleological sense because B-cell anergy is meant to store protective specificities that can be unleashed for host defense against self-mimicking pathogens.

Figure 2.

Differences in proximal signaling between soluble Ags and the same Ag but conjugated on SVLS. (A) For soluble Ag, upon BCR activation, LYN can immediately phosphorylate the ITIM in CD22 that inhibits BCR activation. (B) For the same Ag conjugated on SVLS, upon BCR activation, phosphorylation of the ITIM in CD22 does not occur. (C) Erk phosphorylation after 20 minutes BCR stimulation in LYN-sufficient (MD4 Lyn^+/+) compared with LYN-deficient (MD4 Lyn^−/−) B cells by various agents. Control SVLS: “naked” liposome without Ag, sHELD: soluble HEL double mutant (HELD) R73E D101R, pHELD: SVLS displaying HELD at high (HD) and low ED (LD) (Adapted from (47)).

Moreover, we discovered that ED alone can trigger robust NF-kB activation via the canonical BCR signaling pathway (Btk-dependent) and independent of MyD88 or IRAK1/4 (47). In contrast, soluble Ag can only trigger weak NF-kB activation. As a consequence, a co-stimulatory second signal is dispensable for B-cell growth, survival, and proliferation in the presence of SVLS, but is absolutely essential for B-cell proliferation in response to the same epitope in soluble form. These key differences between soluble and SVLS (71–73) Ags in both proximal and distal BCR signaling (summarized in Table 1) are not merely quantitative, but qualitative, even though the epitope sensed by monoclonal BCR Tg naïve B cells is identical.

Table 1.

Major differences in the B-cell response to soluble and particulate Ags.

	Soluble Ag	SVLS (no iNA)
1, Sensitivity to affinity of the epitope	Yes	No
2. Lyn-dependent inhibitory tone	Robust inhibition	No effect
3. Feature of signal amplification	Analog	Digital
4. Extent of NF-κB activation	Weak	Robust activation without MyD88 or IRAK1/4
5. Dependence on co-stimulatory signal for p65 nuclear translocation	Yes	No
6. Second signal needed for proliferation?	Yes	No
7. Potency of B-cell activation	Weak	Strong

These differences are not merely quantitative (orders of magnitude higher potency for particulate Ag response), but also qualitative; particulate Ag evade Lyn-dependent inhibitory tone, trigger digital all-or-none B-cell signaling, maximal NF-κB activation, and are consequently independent of T-cell help for survival and proliferation.

BCR-mediated particle internalization

A unique and important feature of particulate Ags is that particles can be internalized through BCR-mediated endocytosis (43). This is crucial for virus detection because the viral genome packaged within virion particles can be internalized and, upon processing, exposed to endosomal NA-sensing TLRs in B cells. This B-cell intrinsic TLR signaling can dramatically influence B-cell activation and downstream events (49, 74–76).

B cells lack the nonspecific capacity of macrophages to internalize and degrade particulate Ag (77). But in the presence of cognate Ag interactions with BCRs, B cells can internalize different particulate Ags with a range of different sizes (7, 64, 78–80). Moreover, a threshold of ED may exist that allows this internalization to occur (75). However, we observed that as low as a few Ag molecules per virus-like structure are sufficient to trigger potent and long-lasting Ag-specific IgG responses, provided that there is a strong TLR agonist packaged inside the structure (48, 81), suggesting that this ED threshold for internalization could be extremely low. This important experimental observation in vivo also argues strongly against the assumption that high-density display of Ags on viruses is required for strong Ab responses (28). Moreover, these two viral features in SVLS, ED and iNAs, are integrated by individual Ag-specific B cells to elicit Ag-specific IgG, as evidenced by the compensatory signaling between the BCR and TLR in the elicitation of Ag-specific IgG by SVLS (48). Therefore, for our discussion of virion recognition by Ag-specific B cells, the natural features of a virion, including multivalent viral surface epitopes and the internal presence of NAs, both have to be considered so that the model will be relevant for understanding the molecular basis of virion immunogenicity.

Viral features and germinal centers

Viral features such as multivalent display of a protein, and the presence of internal NAs can have a significant impact on the germinal center. Once they enter the germinal center, B cells will undergo repeated rounds of selection for higher-affinity BCRs while self-reactivity is counter-selected or “redeemed” (82–85). In the germinal center, a particulate Ag can be displayed on the surface of follicular dendritic cells, and thus viral features remain accessible for B-cell recognition and response (13, 19). Pulling forces required for antigen capture are important for affinity discrimination by germinal center B cells (42, 86, 87). Conversely, direct presentation of complex multivalent virus-like Ag may subvert this process, permitting persistence of low-affinity clones in the germinal center. Using HIV-1 immunogens on different structural scaffolds to change the valency of epitopes, Kato et al. showed that high valency of Ags is correlated with more efficient recruitment of low-affinity B cells into germinal centers in mice (30). Using computationally designed self-assembling protein nanomaterials (88, 89), Ols et al. further showed in nonhuman primates that multivalent display of protein Ags on nanoparticles can improve recruitment of low-affinity B cells into the immune response, and thus increase B-cell clonetype diversity and the neutralization breadth of resulting Abs (35). These studies imply that higher ED on virus-like Ags may have similar effects: more efficient B-cell activation and broader recruitment of B cells into germinal centers. However, it should be noted that in most of the data reported in the latter two studies, animal immunizations were done with engineered adjuvants, and the impact of adjuvants on germinal center B-cell recruitment has not been explicitly discussed.

In the context of particulate Ags, using bacteriophage Qβ virus-like particles as model particulate Ags, Hou et al. showed that TLR9 ligand contained within a virus-like particle can substantially enhance the germinal center Ab response (49). However, TLR ligands (when supplied as an admixture with Ag instead of virus-like iNAs) can also disfavor germinal center responses (90) and promote rapid, shorter-lived, extrafollicular responses to facilitate viral clearance in the context of an acute infection (91, 92). How iNAs specifically regulate the balance and efficiency of extrafollicular and germinal center responses remains unclear, but may depend on their abundance and biophysical format as well as adjuvant context, and needs to be further explored for vaccine design.

Long-term Ab response to immunization with virus-like Ags

Although it has been well established that natural viral infections can induce long-term Ab responses (93), the fundamental molecular components that seed these long-term Ab responses have been less clear (94). This question is highly relevant because it is known that protein Ags in the presence of adjuvants can mount robust Ab responses in the absence of TLR signaling (95), that is, NAs are dispensible for a robust Ab response. On the other hand, how do NAs inside virion structures modulate a long-term Ab response? These questions are challenging to address using existing viruses because virion structures are highly organized, and proteins and NAs are usually integrated together in a small submicron-sized particle. Our recent development of SVLS offered a unique approach; the modular nature of this system allows us to examine Ab response triggered by individual components with clarity (44, 45, 48, 81). We showed recently using two unrelated protein Ags and immunization in mice of two different genetic backgrounds (81) that protein Ags alone arrayed on the surface of immune-inert lipid bilayer vesicles filled with PBS are sufficient to seed a long-term Ab response in the absence of any other adjuvants or TLR ligands, although at a low Ab concentration (Fig. 3). The presence of iNAs can dramatically enhance the magnitude of the long-term Ab response, suggesting that efficient TLR activation in the context of a virus-like Ag can substantially augment the production of long-lived plasma cells (and/or their capacity to secrete IgG). These data suggest that virus-like Ag display (ED) alone may be sufficient to seed a long-term Ab response, and TLR activation by the iNAs modulates the magnitude of this response.

Figure 3.

Duration of the IgG responses in B6 or BALB/c mice induced by a single injection of SVLS at sub-microgram doses without additional adjuvants. (a)(b) and (c)(d) Concentrations of Ag-specific IgG in mouse sera collected on different days after a single injection of various agents listed in each inset, where (a) and (b) are for anti-RBD IgG in wtB6 (a) and BALB/c mice (b), respectively, and (c) and (d) are for anti-HEL IgG in wtB6 (c) and BALB/c mice (d), respectively. All concentrations were measured using ELISA based on standard curves obtained from reference monoclonal Abs. Throughout Fig. 3, N = 4 for each time point (Adapted from (81)).

Conclusions and future perspective

B-cell recognition of and response to encounter with intact particulate Ags, such as viruses, is a fundamental feature of the immune system. B cells are sufficient to activate naïve CD4⁺ T cells without dendritic cells in response to virus-like particle or inactivated influenza virus vaccination (96), highlighting the important roles of B cells in coordinating antiviral host responses. The potent activity of virus-like particulate Ags in vivo has long been appreciated. Multivalent display of protein Ags, in particular, promotes recruitment of low-affinity B cells into the germinal center even in the face of low epitope affinity and low precursor frequency (32, 35). However, our understanding of how B cells recognize virus-like Ags at the molecular and cellular level remains incomplete. The two essential features of a virus-like Ag, ED together with iNA, influence every step of B-cell activation until terminal differentiation into plasma cells. While we understand substantial mechanistic details of these processes from studies using other related Ags (97, 98) or in other cells (99–104), we understand relatively little about how these signals are integrated in the context of virus-like particulate Ags in mature naïve B cells, especially in the absence of conventional adjuvants.

In particular, signaling by RNA-sensing TLR7 and DNA-sensing TLR9 appears to drive qualitatively different B-cell responses, evidenced by their opposing roles in autoimmune disease (105–108), with TLR7 promoting (109–111) while TLR9 negatively regulating (112, 113) disease pathogenesis. These observations suggest that the mechanisms of signal transduction and regulation (108, 114–116) differ profoundly between these two NA sensors. How these stimuli are integrated with ED at a biochemical level to impact extrafollicular and germinal center B-cell pathways needs to be defined. A systematic understanding of these interactions between B cells and essential viral features will shed light not only on the molecular basis of viral immunogenicity and potential strategies of viral immune evasion, but also on autoimmunity.

The conditions for memory B-cell activation are critically important for optimal vaccine design targeting many common viruses, notably influenza and SARS-CoV-2. Although memory B cells have lower thresholds for activation than naïve B cells (117–119), it remains to be explored in a systematic manner how they may respond differently to ED and iNAs compared to naïve B cells. Moreover, the biology of memory B cells is impacted by the level and isotype of BCR expression; unlike naïve B cells with high expression of IgD along with variable amounts of IgM, memory B cells express low levels of IgM alone or other isotypes with unique signaling properties (e.g. IgG, IgE, IgA) (120–123). Such differences, coupled with the distinct nanoscale organization of these BCR isotypes and co-receptors, may also impact sensitivity to ED and/or valency of Ag (59, 124, 125). Is lower ED or absence of iNAs sufficient to trigger recall responses? How does this compare to thresholds and optima for naïve B-cell engagement? Answers to these questions are important for guiding the design of influenza vaccines that heavily rely on memory B-cell engagement on the one hand, and HIV-1 vaccines designed to target germline BCR configurations in naïve B cells on the other (126). The final desired neutralizing-Ab response for the latter may further depend on activation of germinal center and/or memory B cells after the initial priming.

We still lack a conceptual framework for how different biophysical formats of Ag are interpreted by the immune system. Are valence and epitope spacing two independent variables for programming a B-cell response to virus-like Ag? How do particulate Ags evade Lyn-dependent inhibitory co-receptor engagement? What are the substrates of Lyn that enable B-cell discrimination between soluble and particulate Ags? Is there a lower affinity floor for B-cell responses to particulate Ags? How do other biophysical properties of viruses, such as particle size or membrane fluidity influence B-cell responses? To address these knowledge gaps in a manner that is germane to virus-like particulate Ags, the selection of a model Ag that adopts the essential features of a virion, and the proper characterization of the Ags are critical to ensure interpretable results. To understand the molecular basis of viral immunogenicity, in vivo immunization without the use of conventional adjuvants is important because the mechanisms of adjuvants are not well understood in many cases and may confound humoral immune responses triggered by virus-like stimuli. Despite significant progress, many gaps remain in our understanding of the molecular basis of virus immunogenicity, with profound implications for both fundamental B-cell biology and vaccine design.

Conflict of interest statement

W.C. has a patent pending on synthetic virus-like structures. No other conflicts were disclosed.

This work was supported by an NIH/NIAID grant (1R01AI155653-01A1) to W.C. and J.Z.