Nanoparticle Immunoadjuvant Complexes Augment Germinal Center Responses to Vaccination
Nicholas J. Tursi, Colby J. Agostino, Jinwei Huang, Toshitha Kannan, Niklas Laenger, Jennifer Londregan, Katlyn Lederer, Michaela Helble, Nicole Bedanova, Cory Livingston, Ebony N. Gary, Madison McCanna, Marta Tarquis Medina, Rumi Habib, Ignacio Rodriguez Relaño, Ivan Maillard
TL;DR
A new vaccine approach uses nanoparticles to boost germinal center responses and improve antibody production in mice.
Contribution
The study introduces nanoparticle immunoadjuvant complexes that enhance B cell maturation and antibody diversity.
Findings
GT8-IL-21-NICs increase serum antibody titers and GC B cell responses in mice.
Immunization with GT8-IL-21-NICs leads to increased antibody diversity and somatic hypermutation.
Transcriptomic analysis shows upregulation of selection-associated gene signatures in GC B cells.
Abstract
Vaccine approaches capable of eliciting enhanced germinal center (GC) responses would result in improved protective humoral immunity against infectious diseases. Here, we investigate whether a cytokine can be scaffolded onto a self‐assembling nanoparticle immunogen to enhance antigen‐specific GC responses and B cell maturation. To test this approach, we design chimeric nanoparticles bearing eOD‐GT8, a germline‐targeting HIV immunogen, and IL‐21, a canonical GC cytokine. DNA delivery of these nanoparticle immunoadjuvant complexes (GT8‐IL‐21‐NICs) drives improved serum antibody titers and antigen‐specific GC B cell responses in mice. Transcriptomic analysis of eOD‐GT8‐specific GC B cells demonstrates upregulation of selection‐associated gene signatures with GT8‐IL‐21‐NIC immunization. In mice harboring human bnAb precursor heavy and light chain genes, immunization with the GT8‐IL‐21‐NIC…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5- —NIH IPCAVD
- —NIH10.13039/100000001
- —INOVIO Pharmaceuticals10.13039/100016255
- —W.W. Smith Charitable Trust Distinguished Professorship in Cancer Research
- —The Jill and Mark Fishman Foundation
- —Commonwealth of Pennsylvania Health Research Formula Fund
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsImmunotherapy and Immune Responses · HIV Research and Treatment · RNA Interference and Gene Delivery
Introduction
1
Germinal center (GC) responses are crucial for the induction of humoral immunity, resulting in robust high affinity antibody responses. GCs are structures that form in B cell follicles of secondary lymphoid organs which feature organized collections of primarily B cells, T cells, and specialized stromal cells [1]. GC dynamics canonically involve GC B cell recognition and processing of antigen from a network of follicular dendritic cells (FDCs) that enable cognate interactions with T follicular helper (Tfh) cells in the light zone (LZ) [2]. Clones that receive Tfh help through costimulation and cytokine signals in the LZ are licensed to enter the dark zone (DZ) for rounds of clonal expansion and somatic hypermutation [3, 4]; the amount of Tfh help therefore dictates the amount of DZ activity among GC B cell clones [5]. Within the GC, cytokine and chemokine signaling mediators work in concert to promote GC initiation, maintenance, and exit [6]. Numerous studies have implicated IL‐21, IL‐6, IL‐4, and thymic stromal lymphopoietin among other cytokines as important regulators of GC dynamics, namely by influencing the differentiation and proliferation of both Tfh and GC B cells [3, 4, 7, 8, 9, 10]. Next‐generation vaccine approaches, such as germline targeting and sequential immunization, rely upon priming strong germinal center responses, which frequently leverage iterative structure‐guided design of immunogens to induce specific B cell lineages to generate broadly neutralizing antibodies (bnAbs) [11, 12, 13, 14, 15]. Germline targeting immunogens such as the eOD‐GT8‐60mer have demonstrated promise in the clinic to prime VRC01‐class bnAb‐precursor responses [16, 17]. However, due to low frequencies of many bnAb precursors [13, 18, 19, 20] and poor responder populations, vaccine approaches that can prime strong antigen‐specific GC responses, improve expansion of recruited clones, and improve response durability are growing in importance.
Multimeric nanoparticle immunogens have been shown to elicit potent and durable GC responses by increasing the breadth of humoral immunity through recruitment of lower affinity B cell clones compared to monomeric antigen [21, 22]. Generally, high antigen valency can promote potent B cell activation through B cell receptor (BCR) clustering and alternate antigen trafficking and uptake dynamics [23, 24, 25]. We have recently described DNA‐launched multimeric nanoparticle vaccines that assemble in vivo, are dose sparing, and induce robust humoral immune responses relative to traditional DNA vaccines [26, 27, 28, 29, 30]. Additionally, we have previously demonstrated that nucleic acid‐launched nanoparticles efficiently deposit onto FDCs [26]. Further, we and others have demonstrated that co‐delivery of gene‐encoded molecular adjuvants such as enzymes, cytokines, and chemokines can augment immune responses [31, 32, 33, 34, 35].
Here, using DNA‐launched nanoparticle vaccines as a model, we demonstrate that linkage of a cytokine to a multivalent antigen termed nanoparticle‐immunoadjuvant complexes (NICs) drives more potent antigen‐specific GC responses. As proof‐of‐principle, we designed a chimeric nanoparticle using the eOD‐GT8‐60mer (herein abbreviated GT8‐60mer) bearing scaffolded IL‐21, termed GT8‐IL‐21‐NIC. We sought to understand whether IL‐21 scaffolded on an immunogen could enable improved GC responses in terms of magnitude, expansion, diversity, and durability. Immunization with the GT8‐IL‐21‐NIC led to improved long‐lived serum antibody responses as well as augmented germinal center responses superior to monomeric IL‐21 co‐delivery. In the T cell compartment, GT8‐IL‐21‐NIC improved functional Tfh cell responses as determined through activation‐induced markers and suppressed T follicular regulatory (Tfr) cells. Examining GC kinetics, the GT8‐IL‐21‐NIC enabled earlier responsiveness as evidenced by an increased frequency of GCs at day 7 post immunization relative to later timepoints. Single cell RNA sequencing (scRNAseq) showed that the GT8‐IL‐21‐NIC enhanced genes and pathways associated with GC B cell formation and selection. In transgenic mice harboring VRC01‐class bnAb precursor heavy (V_H_) and light chain (V_K_) variable genes, immunization with GT8‐IL‐21‐NIC led to improved priming of CD4 binding site (CD4bs)‐specific responses compared to the GT8‐60mer. BCR sequencing revealed that the GT8‐IL‐21‐NIC drove increased diversity, clonal expansion, and somatic hypermutation in the GC. Crucially, among cells with transgenic human BCRs, we observed a robust increase in the number of key VRC01‐class bnAb mutations with GT8‐IL‐21‐NIC immunization. Thus, utilizing IL‐21 as a NIC enables the generation of robust antigen‐specific GC responses to promote adaptive immunity.
Results
2
Multimeric GT8‐IL‐21‐NICs Assemble and Functionally Induce B Cell Differentiation In Vitro
2.1
We hypothesized that direct linkage of antigen to a cytokine as a molecular adjuvant would significantly augment GC interactions in an antigen‐specific manner compared to antigen and cytokine adjuvant delivered separately. To do so, we leveraged plasmid DNA‐delivered nanoparticle immunogens as a proof‐of‐concept to scaffold both antigen and cytokine as a NIC. We utilized the GT8‐60mer as a model antigen that we have characterized previously using a DNA vaccine approach [26]. For a molecular adjuvant, we focused on IL‐21 as a key GC cytokine implicated in both Tfh and GC B cell proliferation and differentiation [8, 36]. We designed DNA plasmid cassettes to express a hybrid nanoparticle immunogen displaying native mouse IL‐21 on the GT8‐60mer as a NIC via genetic fusion using glycine‐serine linkers (Figure 1A). We first demonstrated that the antigen‐cytokine fused nanoparticle, GT8‐IL‐21‐NIC, assembled in vitro using size exclusion chromatography (Figure 1B). Negative stain electron microscopy revealed comparable nanoparticle size and assembly between the GT8‐60mer and GT8‐IL‐21‐NIC (Figure 1C; Figure S1A). Next, we determined that IL‐21 and GT8 were indeed co‐displayed in a conformationally correct manner on the GT8‐IL‐21‐NIC using an ELISA‐based method to interrogate binding to both mouse IL‐21 receptor (IL‐21R) and VRC01, a broadly neutralizing antibody recognizing GT8 (Figure 1D). Additionally, using VRC01 enabled us to determine that the key germline‐targeting epitope on GT8 was not occluded on the surface of the GT8‐IL‐21‐NIC. To determine whether IL‐21 is functional when co‐displayed with antigen on the NIC, we purified follicular (FO) B cells from B6.Blimp1^+/GFP^ reporter mice and stimulated them with the GT8‐60mer or GT8‐IL‐21‐NIC alone or in combination with recombinant mouse IL‐4 and IL‐5. We observed robust induction of Blimp‐1+ plasma cells with the GT8‐IL‐21‐NIC relative to the GT8‐60mer alone, indicating that IL‐21 is presented in its native capacity and can sufficiently drive signaling and subsequent proliferation and differentiation (Figure 1E,F). Additionally, we observed robust activation using the markers CD86 and CD80 with the GT8‐IL‐21‐NIC relative to the GT8‐60mer (Figure 1G,H). The GT8‐IL‐21‐NIC was also able to induce STAT3 phosphorylation in purified FO B cells, supporting its capacity to signal through IL‐21R (Figure 1I,J). Taken together, these data demonstrate that multimeric NICs can be engineered to display antigen and cytokine in their native conformation and drive robust signaling in B cells capable of plasma cell differentiation and STAT3 phosphorylation.
Development and characterization of GT8‐IL‐21‐NIC. (A) Model depicting GT8‐IL‐21‐NIC generated using Alphafold and Pymol; GT8 (light blue), IL‐21 (red), lumazine synthase scaffold (gray). IgE denotes signal sequence; GS denotes glycine‐serine linker. (B) Size exclusion chromatography (SEC) trace of GT8‐IL‐21‐NIC run at 215 nm. (C) Negative stain electron microscopy (nsEM) of the GT8‐60mer and GT8‐IL‐21‐NIC. (D) Binding ELISA of recombinant GT8‐60mer or GT8‐IL‐21‐NIC to IL‐21R probed with VRC01. (E) Representative contour plot of follicular B (FO B) cells labelled with Cell Trace Yellow from a Blimp‐1 reporter mouse cultured in vitro with recombinant GT8‐60mer, GT8‐IL‐21‐NIC, or co‐cultured with IL‐4/5. Plots show Blimp‐1+ plasma cells at 72 h post co‐culture. (F) Frequency of Blimp‐1+ plasma cells. (G,H) CD86 and CD80 geometric MFI (gMFI) of cultured follicular B cells with indicated stimuli. (I) Histogram showing pSTAT3 expression in purified FO B cells after culture with 10 µg/mL of the GT8‐60mer or GT8‐IL‐21‐NIC. (J) pSTAT3 gMFI in purified FO B cells with varying concentrations of the GT8‐60mer or GT8‐IL‐21‐NIC. Bars show mean with SD. Each dot represents one technical replicate (F–H); n = 3 technical replicates representative from one or two independent experiments. One‐way ANOVA (F‐H) or two‐way ANOVA (J) adjusted for multiple comparisons with Bonferroni correction used to compare groups; *** p < 0.001; **** p < 0.0001.
GT8‐IL‐21‐NIC Generates Robust Humoral and GC Responses
2.2
Eliciting durable antibody titers through the generation of long‐lived plasma cells (LLPCs) is a crucial for long‐term protective humoral immunity. To understand the impact of GT8‐IL‐21‐NICs on serum antibody titers, we immunized mice once with 0.1 or 0.5 µg of plasmid DNA encoding either the GT8‐60mer or GT8‐IL‐21‐NIC (Figure 2A; Figure S2A). Mice immunized with GT8‐IL‐21‐NIC generated more robust GT8‐specific IgG antibody titers relative to the GT8‐60mer alone (Figure 2B). The GT8‐IL‐21‐NIC elicited titers approximately ten‐fold higher than GT8‐60mer beginning two weeks post immunization that persisted through 4 months post immunization. Antibody titers after immunization with 0.5 µg were more comparable, suggesting saturation of the humoral immune response at this dose (Figure S2B). We next asked whether GT8‐IL‐21‐NIC generated improved antibody‐secreting cell (ASC) responses in the bone marrow (BM). We observed two‐fold higher GT8‐specific IgG+ ASC responses in the BM of immunized mice 16 weeks post vaccination with the GT8‐IL‐21‐NIC relative to the GT8‐60mer (Figure 2C).
GT8‐IL‐21‐NIC augments humoral immunity and improves GC priming. (A) Wild‐type mice were immunized with GT8‐60mer and GT8‐IL‐21‐NIC for serology (0.1 µg each) or for GC responses (0.5 µg each). (B) GT8‐specific serum IgG titers by ELISA. (C) GT8‐specific bone marrow antibody‐secreting cells at 16 weeks post immunization by ELISpot. (D–F) Frequency of GC B and T cell populations were assessed at Day 7 post immunization in the DLNs using flow cytometry. Representative contour plots and frequency shown. (D) Total GC B cells (E) GT8‐specific GC B cells. (F) Activated Tfh cells. (G) Frequency of CD40L+ Tfh cells at Day 7 after peptide stimulation. (H) Representative contour plot of Foxp3 expression in PD‐1hi CXCR5+ cells. (I) Ratio of Foxp3‐ Tfh to Foxp3+ Tfr. (J–L) Frequency of GC B and T cell populations were assessed at Day 7, 14 and 28 post immunization in the DLNs. DPI = days post immunization. (J) Total GC B cells. (K) GT8‐specific GC B cells. (L) Activated Tfh cells. (M) Wild‐type mice were immunized with the GT8‐60mer (0.5 µg), the GT8‐60mer (0.5 µg) with plasmid‐encoded monomeric IL‐21 (0.5 or 5 µg), or the GT8‐IL‐21‐NIC (0.5 µg). GC B cell populations were assessed using flow cytometry. (N) Representative contour plot and frequency of total GC B cells. (O) Frequency of GT8‐specific GC B cells. Bars show mean with SD, serum titers in (B) shown as geometric mean with geometric SD. Each dot in bar plots represents one mouse; n = 4–14 mice per group. Data combined from 2–3 independent experiments. Day 7 animals in (J–L) recapitulated from Figure 2 and Figure S3. Non‐parametric Mann Whitney U Test or one‐way ANOVA adjusted for multiple comparisons with Bonferroni correction used to compare groups; * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.
To investigate the impact of GT8‐IL‐21‐NIC immunization on resulting GCs, mice were immunized with 0.5 µg of plasmid encoding the GT8‐60mer or GT8‐IL‐21‐NIC. Immunization with the GT8‐IL‐21‐NIC resulted in a three‐fold increase in the frequency of total GC B cell responses (Figure 2D) and a six‐fold increase in the frequency of GT8‐specific GC B cell responses (Figure 2E) in the draining lymph nodes (DLNs) 7 days post immunization. Similarly, immunization with a low dose of 0.1 µg of the GT8‐IL‐21‐NIC led to an increased frequency of total and GT8‐specific GCs compared to the GT8‐60mer alone, albeit to a lesser magnitude (Figure S2C–F).
Along with its role in GC B cell dynamics, IL‐21 is paramount to Tfh proliferation and differentiation [8]. We characterized the Tfh response after GT8‐IL‐21‐NIC immunization relative to the GT8‐60mer. GT8‐IL‐21‐NIC immunized mice had a two‐fold increase in the frequency of activated (CD44+) Tfh cells in the DLNs (Figure 2F). Cognate help signals include the presence of costimulatory molecules on the Tfh cell surface as well as secreted cytokines. To determine whether GT8‐IL‐21‐NICs modulate the functional capacity of induced Tfh, we stimulated lymph node suspensions with peptides spanning the GT8 and lumazine synthase domains and assessed various activation‐induced markers. GT8‐IL‐21‐NIC immunization drove a 2‐fold increase in the frequency of CD40L+ Tfh cells relative to GT8‐60mer alone (Figure 2G). CD40L was distinct among activation‐induced markers, as we did not observe differences in CD25, OX‐40, or ICOS expression between GT8‐60mer and GT8‐IL‐21‐NIC (Figure S3A–C).
T follicular regulatory (Tfr) cells suppress Tfh in the germinal center, and co‐express Foxp3 alongside surface expression of PD‐1 and CXCR5. IL‐21 has been demonstrated to restrict Tfr proliferation [37]. To determine whether the GT8‐IL‐21‐NIC negatively regulates Tfr cells in the GC, we used intracellular staining to discriminate Foxp3+ Tfr cells from Foxp3‐ Tfh cells within the heterogeneous population of PD‐1 and CXCR5 co‐expressing cells 7 days post immunization. There was a significant decrease in the frequency of Tfr with GT8‐IL‐21‐NIC immunization relative to the GT8‐60mer (Figure 2H; Figure S4A). GT8‐IL‐21‐NICs led to a two‐fold increase in the ratio of Tfh (Foxp3‐) to Tfr (Foxp3+) (Figure 2I). However, we did not observe a significant change in the frequency of total non‐GC regulatory T cells (Tregs) between the two treatment groups (Figure S4B). Together, these data indicate that GT8‐IL‐21‐NICs specifically regulate the Tfh‐Tfr axis in a pro‐Tfh manner.
We next interrogated the architecture of induced GCs in vivo using immunofluorescent microscopy to confirm bona fide GC formation. Anatomically, the GC exists within a B cell follicle of IgD+ naïve B cells, outside of the T cell zone; zonal distribution can be inferred through CR2/CR1+ FDCs localized in the light zone [38]. Mice were immunized with either the GT8‐60mer or GT8‐IL‐21‐NIC and iliac lymph nodes were harvested 7 days post immunization. We observed GC formation after immunization with both the GT8‐60mer and GT8‐IL‐21‐NIC by microscopy, with strong co‐localization of CR2/CR1 and GL7. These data extend our flow cytometry data that bona fide GCs are formed at this timepoint (Figure S5A).
GT8‐IL‐21‐NIC Induces an Earlier GC Response Peak at Day 7
2.3
Natural acute infection or vaccinations that generate a T‐dependent GC response in secondary lymphoid organs peak approximately 14‐ to 21‐days post immunization and subsequently wane [39, 40, 41]. To understand the kinetics of the GC response after GT8‐IL‐21‐NIC immunization, we immunized cohorts of mice and examined the DLNs 7‐, 14‐, and 28‐days post immunization. Timepoints beyond 7‐days post immunization did not exhibit a dramatic difference between the GT8‐60mer and GT8‐IL‐21‐NIC groups in the frequency of total (Figure 2J) or GT8‐specific GC B cell responses (Figure 2K). This was also evident among activated Tfh cells in the GC, where an early increase Day 7 contracted to a similar frequency between the antigen‐only GT8‐60mer and GT8‐IL‐21‐NIC at later timepoints (Figure 2L). These data demonstrate that immunization with the GT8‐IL‐21‐NIC augments the magnitude of early antigen‐specific GC responses.
GT8‐IL‐21‐NIC Induces Superior GC Responses Relative to Co‐Delivery of IL‐21 Monomer
2.4
We determined the capacity of GT8‐IL‐21‐NIC to augment germinal center responses relative to monomeric plasmid‐encoded IL‐21 as a molecular adjuvant. We immunized mice with 0.5 µg of plasmid‐encoded GT8‐60mer, co‐immunized with a low (0.5 µg) or high (5 µg) dose of plasmid‐encoded monomeric IL‐21, or with 0.5 µg of plasmid‐encoded GT8‐IL‐21‐NIC (Figure 2M). We observed gradual increases in both total (Figure 2N) and GT8‐specific (Figure 2O) GC B cells with increasing amounts of plasmid‐encoded monomeric IL‐21. However, we observed significant increases in the frequency of total and GT8‐specific GC B cells with the GT8‐IL‐21‐NIC, demonstrating that scaffolding of IL‐21 on the NIC leads to improved potency despite 10‐fold higher delivery of plasmid encoding IL‐21 monomer. Extending these analyses to Day 14 post immunization, we observed a normalization in the GC response among IL‐21 monomer co‐immunized mice to levels similar to the GT8‐60mer among total GC B, GT8‐specific GC B, Tfh, and Tfr populations (Figure S6A–E). This follows the pattern observed with the GT8‐IL‐21‐NIC at later GC timepoints.
IL‐21 signaling in GC B cells regulates specific zonal localization of GC B cells to the DZ, as IL‐21R deficient GC B cells demonstrated skewed localization to the LZ of the GC [8]. To further dissect the GC B cell response, we examined the zonal localization of GT8‐specific GC B cells. In contrast with prior findings, GT8‐IL‐21‐NIC immunization resulted in a significant increase in GT8‐specific GC B cells in the LZ (and a decrease in DZ) (Figure S6F–H). We found a dose‐dependent increase in the frequency of GT8‐specific GC B cells in the LZ (and a decrease in the DZ), with effect size magnified using the NIC platform. With monomeric IL‐21 delivery, this largely normalized by Day 14 post immunization, in contrast with the GT8‐IL‐21‐NIC (Figure S6I,J). Thus, increasing exogenous IL‐21 through nanoparticulate display increases antigen‐specific GC B cell responses and zonal reorganization in a dose‐dependent manner.
scRNAseq Demonstrates that GT8‐IL‐21‐NIC Drives a Pro‐Selection Signature in GC B Cells
2.5
Considering the early GC skew after immunization with NICs, we sought to gain additional mechanistic insight using scRNAseq on the transcriptional profile of GT8‐specific GC B cells 7 days post immunization (Figure 3A). Seurat clustering revealed 9 distinct clusters containing heterogeneous subpopulations of GC B cells (Figure 3B). We identified clusters using cell cycle state, pseudotime analysis, and top upregulated genes in each cluster (Figure S7A–C) [42]. Based on gene expression profiles, we identified top upregulated genes such as Il4i1 and Cd83 belonging to a LZ signature in cluster 1, and a DZ signature in cluster 3 with upregulated genes such as Mki67 (Figure S7A) [43, 44]. Additionally, we observed numerous clusters (cluster 2, 4, 5) with a heterogeneous gene expression profile in different phases of the cell cycle. Among DZ‐like clusters, we observed one primarily in G2/M (cluster 3) and one primarily in S phase (cluster 2) with active DNA replication genes. Cluster 4 had a DZ signature with cells primarily in G1, indicative of exit from the DZ and active cell division. In contrast, Cluster 5 had a gene signature/cell cycle profile of cells entering active cell division and thus into the DZ. Clusters 6–9 were primarily in G1 and represented either pre‐ or post‐GC B cell fates. Cluster 6 was defined as pre‐memory due to robust expression of Ccr6, Sell, and Gpr183 [45]. Cluster 7 had high expression of Ighd and thus was defined as naïve B cells. Cluster 8 had upregulation of genes associated with an activated B cell/pre‐GC B program (Bach2, Pax5, and Zbtb20). Cluster 9 had high expression of specific immunoglobulin variable genes, Irf4, and Jchain, indicative of a pro‐plasma cell differentiative program [46, 47].
GT8‐IL‐21‐NIC engineers GC kinetics and promotes a pro‐selection phenotype. (A) Mice were immunized with 0.5 µg of GT8‐60mer or GT8‐IL‐21‐NIC and pooled antigen‐specific GC B cells were sorted and subjected to single‐cell RNA sequencing via the 10x platform. (B) UMAP separated by sample colored by cluster. (C) Frequency of clusters 1–5 in each sample. (D) UMAP density plot of indicated genes of interest. (E–G) Upregulated IPA regulators in cluster 1 (LZ) (E), cluster 3 (DZ) (F), and cluster 5 (activated/selected LZ) (G) in GT8‐IL‐21‐NIC relative to the GT8‐60mer. (H) Upregulated genes of interest in cluster 1 (LZ). n = 10 mice pooled. Wilcoxon rank sum used to compare groups (H). p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.*
In line with our flow cytometry data, we observed an increased frequency of cells in the LZ (cluster 1) and decreased frequency of cells in the DZ (clusters 2 and 3) with GT8‐IL‐21‐NIC immunization relative to the GT8‐60mer (Figure 3C). Interestingly, we observed an increase in cells in an activated/selected LZ B cell cluster (cluster 5) with GT8‐IL‐21‐NIC immunization; this cluster co‐localizes with Myc expression, a transcription factor transiently induced after GC B cell selection canonically from Tfh [5] (Figure 3C,D). Additional density plots of gene expression further confirmed GC vs. pre‐ or post‐GC states. We also observed decreases in the frequency of cells in differentiated states such as pre‐memory (cluster 6) and plasma cell (cluster 9) but an increase in pre‐GC like cells (cluster 8), indicating a pro‐GC transcriptional landscape within the GC. (Figure S7D).
To further understand pathways of interest within select GC B clusters, we examined differentially expressed genes (DEGs) and pathways upregulated after immunization with GT8‐IL‐21‐NIC relative to the GT8‐60mer alone. In the LZ (cluster 1), Ingenuity Pathway Analysis (IPA) of regulators demonstrated upregulated expression of genes/signaling pathways implicated in GC B cell commitment and maintenance such as SMARCA4, IRF8, and POU2F2 [48, 49, 50], an activation marker TNFRSF8 (CD30) [51], and pathways associated with T cell help such as IL‐4, and CD40 [52, 53] (Figure 3E). The LZ also had two Myc‐associated regulators upregulated compared with the control GT8‐60mer, indicative of a pro‐selection/activation phenotype. Notably, cells in the DZ (cluster 3) after GT8‐IL‐21‐NIC immunization had IPA regulators activated reminiscent of a LZ signature; strong Myc (and related Myc effector) activation and STAT3 induction [54] (Figure 3F). Additionally, PI3K related regulators and downstream AKT1 were both upregulated, all signaling cascades known to negatively regulate Foxo1, a transcription factor crucial for DZ identity [55, 56, 57]. In line with the increase in activated/selected B cells (cluster 5), we observed upregulation in Myc or Myc‐related regulators using IPA, in addition to other regulators associated with transcriptional activation and BCR signaling (Figure 3G). Further examination of individual DEGs in the LZ (cluster 1) demonstrated upregulated expression of genes associated with GC formation (Plcg2, Id3, Irf8, Pou2f2), maintenance (Bach2, Pax5), and T cell help responsiveness (Cd40) after GT8‐IL‐21‐NIC immunization in line with our IPA analysis (Figure 3H). We also observed increased Ccnd3 upregulation, which is important for GC B cell cycling between light and dark zones [58]. Together, these data indicate that immunization with GT8‐IL‐21‐NIC generates a pro‐selection and pro‐GC maintenance transcriptional phenotype, supporting zonal remodeling toward the Tfh‐centric light zone.
GT8‐IL‐21‐NIC Primes CD4bs‐Specific Responses in Human VH1‐2/VK1‐33 Transgenic Mice
2.6
eOD‐GT8 is a germline‐targeting priming immunogen designed to elicit VRC01‐class HIV CD4bs bnAb precursors bearing V_H_1‐2 heavy chains (HCs) and 5 amino acid kappa light chain (LC) CDRL3s [59, 60, 61, 62]. To understand whether the GT8‐IL‐21‐NIC could improve CD4bs‐specific priming responses, we examined the ability of GT8‐IL‐21‐NICs to prime bnAb precursors in a relevant mouse model such as V_H_1‐2R^JH2^/V_K_1‐33R^CSΔ/hTdT^ transgenic mice (henceforth abbreviated V_H_1‐2/V_K_1‐33 Tg). These mice harbor human V_H_1‐2, J_H_2, and V_K_1‐33 gene segments with the addition of a human terminal deoxynucleotidyl transferase (hTdT), allowing for the generation of VRC01‐class rearranging HC and LC variable regions with improved repertoire diversity [63]. This rearrangement enables a more physiologically relevant model for studying responses to the eOD‐GT8‐60mer. We primed V_H_1‐2/V_K_1‐33 Tg mice with 0.5 µg of either the GT8‐60mer or GT8‐IL‐21‐NIC and assessed GC responses in DLNs 7 days post immunization (Figure 4A). We observed a robust increase in total GC B cell responses with the GT8‐IL‐21‐NIC relative to the GT8‐60mer (Figure 4B,C). Importantly, we observed an approximately two‐fold improvement in CD4bs‐specific GC B cells using epitope knock‐out bait GT8‐KO11 to interrogate specificity (Figure 4B,D). Together, these data demonstrate that the GT8‐IL‐21‐NIC augments GC responses to the GT8‐60mer in a relevant mouse model to elicit CD4bs‐specific responses.
GT8‐IL‐21‐NIC generates GC responses with increased diversity, clonal expansion, and somatic hypermutation in VH1‐2/VK1‐33 transgenic mice. (A) VH1‐2/VK1‐33 mice were immunized with GT8‐60mer and GT8‐IL‐21‐NIC (0.5 µg each) to assess GC priming. (B) Representative flow cytometry gating strategy to identify CD4bs‐specific GC B cells. (C) Frequency of total GC B cells. (D) Frequency of GT8‐specific GC B cells. (E) Average number of sorted CD4bs‐specific GC B cells in each treatment group. (F) Pie charts showing proportion of clones vs. singlets among sorted cells and absolute number of sorted cells (center). (G) Frequency of top expanded VH/VL pairs (count >100) from the GT8‐60mer and GT8‐IL‐21‐NIC. Number of pairs indicated in the center. (H) Counts of shared top VH/VL pairs from (G). (I,J) Counts of shared (I) and unique (J) IGKVs between GT8‐60mer and GT8‐IL‐21‐NIC that pair with human VH1‐2, excluding singlets. Number on top of bar in (J) indicates the number of IGKVs. (K,L) Counts of shared (K) and unique (L) IGHVs between GT8‐60mer and GT8‐IL‐21‐NIC that pair with human VK1‐33, excluding singlets. Number on top of bar in (L) indicates the number of IGHVs. Bars show mean with SD or absolute counts. Each dot in bar plots represents one mouse; n = 5‐6 mice from two independent experiments (A–D) or n = 4–5 mice pooled (E–L). Non‐parametric Mann Whitney U Test correction used to compare groups; * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.
GT8‐IL‐21‐NIC Drives Increased Clonal Expansion and Diversity in the GC
2.7
Considering the robust GC responses primed in V_H_1‐2/V_K_1‐33 Tg mice, we sought to understand whether the NIC influenced diversity of recruited B cells and/or expanded similar clonotypes as the GT8‐60mer. To do this, we performed V(D)J sequencing to examine the repertoire of induced GT8‐specific cells (Figure 4A). We sorted CD4bs‐specific germinal center B cells from V_H_1‐2/V_K_1‐33 Tg mice immunized with either the GT8‐60mer (n = 5) or GT8‐IL‐21‐NIC (n = 4). We saw a substantial (greater than ten‐fold) increase in the average number of sorted cells (Figure 4E), consistent with our flow cytometry data. We recovered 2085 and 9774 sequences from GT8‐60mer and GT8‐IL‐21‐NIC respectively (Figure 4F). Among sequenced cells, the majority (95% and 98% respectively for the GT8‐60mer and GT8‐IL‐21‐NIC) were clones, with the remaining being singlets. Immunization with the GT8‐60mer and GT8‐IL‐21‐NIC induced 5 and 11 V_H_/V_L_ pairs that were expanded (>100 cells), respectively (Figure 4G). The most expanded V_H_/V_L_ pair in both treatment groups was human V_H_1‐2/V_K_1‐33. Among the other expanded V_H_/V_L_ pairs seen with GT8‐60mer immunization, all were seen after GT8‐IL‐21‐NIC immunization, although the NIC induced at least a two‐fold expansion relative to the GT8‐60mer (Figure 4H). We next sought to examine how many murine LC genes (IGKV) paired with human V_H_1‐2. A total of 16 IGKVs that paired with human V_H_1‐2 were shared between the GT8‐60mer and GT8‐IL‐21‐NIC, although the majority of these were more expanded after GT8‐IL‐21‐NIC immunization (Figure 4I; and Table S1). Among IGKVs unique to each treatment group, there were 3 and 11 IGKVs unique to the GT8‐60mer and GT8‐IL‐21‐NIC, respectively (Figure 4J; and Table S2). A similar pattern was observed when examining murine HC genes (IGHV) that paired with human V_K_1‐33. We observed greater expansion of shared IGHVs as well as an increase in the number of unique IGHVs that paired with human V_K_1‐33 after immunization with the GT8‐IL‐21‐NIC (Figure 4K,L; and Tables S3 and S4). These data show that the GT8‐IL‐21‐NIC is generating both increased expansion of recruited B cells in the GC as well as increased diversity of V_H_/V_L_ pairs.
Acquisition of Key VRC01‐Class bnAb Mutations with GT8‐IL‐21‐NIC Immunization
2.8
To understand whether GT8‐IL‐21‐NIC immunization led to the acquisition of somatic mutations associated with VRC01‐class bnAb development (VRC01‐class mutations) in V_H_1‐2 and V_K_1‐33, we analyzed V_H_1‐2/V_K_1‐33 clonotypes and their mutational profile. We observed an increased number of cells belonging to each clonotype after immunization with the GT8‐IL‐21‐NIC relative to the antigen‐only GT8‐60mer (Figure 5A). Among shared V_H_1‐2/V_K_1‐33 clonotypes, there was greater clonal expansion with the GT8‐IL‐21‐NIC relative to the GT8‐60mer (Figure 5B). Additionally, among unique V_H_1‐2 and V_K_1‐33 clonotypes, there was an increased number unique to GT8‐IL‐21‐NIC immunization (Figure 5C). To understand the mutational profile of recruited cells, we next examined somatic hypermutation (SHM) among V_H_1‐2/V_K_1‐33 clonotypes. We observed a significant increase in the number of nucleotide and amino acid mutations across HC and LC sequences after immunization with the GT8‐IL‐21‐NIC (Figure 5D,E). To understand the maturation of B cell lineages, we generated a phylogenetic lineage tree for cells from each treatment group of the largest phylogeny containing V_H_1‐201 and J_H_201 among V_H_1‐2/V_K_1‐33 clonotypes. We observed increased expansion of branches with a greater number of mutations with the GT8‐IL‐21‐NIC (Figure 5F), supporting the robust clonal expansion and somatic hypermutation profile seen. To assess whether immunization with the GT8‐IL‐21‐NIC could induce mutations associated with bnAb development, we examined a number of key paratope and non‐paratope VRC01‐class bnAb mutations (Figure 5G, [64]). Critically, we observed a greater number of cells with 1 or 2 key mutations with the GT8‐IL‐21‐NIC relative to the antigen‐only GT8‐60mer (Figure 5H). We also observed cells with more than 3 key mutations with the GT8‐IL‐21‐NIC which was not observed with the GT8‐60mer. To visualize each of the paratope and non‐paratope VRC01‐class mutations, we next calculated the 90^th^ percentile number of key mutations in each immunization group. For the GT8‐IL‐21‐NIC, the 90^th^ percentile represents approximately the top 15% most mutated sequences. As a more representative comparison, we utilized the top 15% of sequences from the 90^th^ percentile in the GT8‐60mer immunization group (as the 90^th^ percentile included cells with only 1 mutation). Strikingly, we observed a dramatic increase in the number of amino acid positions with key VRC01‐class mutations in the GT8‐IL‐21‐NIC (Figure 5I). Mutations were spread across a number of key residues suggesting that the GT8‐IL‐21‐NIC can elicit VRC01‐class bnAb mutations after a single immunization.
Increased SHM and acquisition of key VRC01‐class bnAb mutations with GT8‐IL‐21‐NIC. (A) Counts of each clonotype harboring human VH1‐2/VK1‐33. (B,C) Counts of shared (B) and unique (C) VH1‐2/VK1‐33 clonotypes between the GT8‐60mer and GT8‐IL‐21‐NIC. Number on top of bar in (C) indicates the number of clonotypes. (D,E) Average number of nucleotide (left) and amino acid (AA, right) mutations in the heavy chain (D) or light chain (E) V regions across all human VH1‐2/VK1‐33 clonotypes. SHM = somatic hypermutation. (F) Lineage tree of the largest phylogeny containing VH1‐201 and JH201 in the GT8‐60mer (top) or GT8‐IL‐21‐NIC (bottom) group. Length of each branch represents the expected number of nucleotide substitutions per codon. Tree scale = 0.1 nucleotide substitutions per codon. (G) Key non‐paratope and paratope VRC01‐class bnAb mutations, defined by Cottrell et al. [64]. (H) Counts of key VRC01‐class bnAb mutations after treatment with the GT8‐60mer (top) or GT8‐IL‐21‐NIC (bottom). (I) Key VRC01‐class heavy chain residues in the GT8‐60mer (top panel, top 15% of the 90th percentile values) and the GT8‐IL‐21‐NIC (bottom panel, 90th percentile values). Each row represents a single sequence, and each colored line indicates the presence of a mutation (specific residues as columns).
Discussion
3
GC reactions are crucial to humoral immunity against T‐dependent antigens, and control over GCs via vaccination remains a key focus for generating responses against complex pathogens. Strategies employing germline targeting immunogens have tailored GC responses to elicit broadly neutralizing antibody (bnAbs) precursors to the CD4 binding site and V3 glycan patch through priming rare lineages [11, 12, 65, 66]. Importantly, what underlies these approaches are the formation of robust GC responses. As such, engineering vaccination modalities to drive GC responses upon priming that result in improved clonal expansion, diversity, and durability of humoral immune responses would be of interest. Here, we studied how linking a canonical GC cytokine to a multimeric immunogen would influence priming responses. We demonstrate that scaffolding cytokine IL‐21 to an immunogen as a NIC can robustly increase humoral response magnitude and GC responses. A prior study has examined ligands implicated in GC responses (CD40L, ICOS, and IL‐21) on a nanoparticle immunogen, however this was in the context of Tfh‐deficient mice [67]. Additionally, a recent study investigated Notch ligands as a nanoparticle adjuvant which improved T cell and IgG2c responses in mice, but required multiple, high (20–100 µg) intravenous doses accompanied by separate administration of antigen in Complete and Incomplete Freund's adjuvant, complicating dosing [68]. Our study demonstrated that GT8‐IL‐21‐NICs delivered using low doses (0.1–0.5 µg) of plasmid DNA were able to elicit robust enhancements in both GC and humoral immune responses. Using a nucleic acid vaccine modality enables rapid replacement of specific domains through mutagenesis or subcloning, reducing the purification constraints of recombinant nanoparticle platforms. Critically, immunization with NICs improved immune responses relative to both antigen only nanoparticles and co‐delivery of cytokine monomer.
Vaccine platforms that generate robust GC responses, such as lipid nanoparticle‐formulated mRNA, display altered GC kinetics with B cell responses peaking 7‐days post immunization [69, 70]. Immunization with GT8‐IL‐21‐NICs led to dramatically enhanced antigen‐specific GC B cell responses 7‐days post immunization, that contract to levels comparable with antigen alone from 14 days post immunization onward. Responses were skewed toward the LZ, where Tfh‐driven selection occurs. This is in contrast to prior studies suggesting that IL‐21 is crucial for DZ localization using IL‐21R KO mice [7, 8, 36, 71]. Our data suggests that additional exogenous IL‐21 drives a unique LZ skewed program in GC B cells, indicating the value of exploring the mechanism of cytokine signaling through this platform. However, DNA delivery has unique kinetics of protein production compared to a recombinant protein bolus [72]; which may play a role in the phenotype observed.
Immunization with the GT8‐IL‐21‐NIC also had an impact on the T cell compartment, through expanding the frequency of total and effector Tfh cells. In its native state, IL‐21 has an independent role on Tfh cells through paracrine signaling mechanisms to promote proliferation and differentiation [8, 73, 74]. Additionally, there is a role for IL‐21 in suppressing Tfr [37, 75]. We demonstrate that the GT8‐IL‐21‐NIC selectively enhances Tfh and negatively regulates Tfr. This, paired with dramatic increases in GT8‐specific GC B cell and antibody responses, shows local regulation of the GC response by the NICs to promote antigen‐specific immunity.
Immunization with GT8‐IL‐21‐NICs induced a pro‐selection and pro‐GC fate in GC B cells. The transcription factor Myc is a key mediator of the LZ to DZ transition; increased Myc expression in LZ B cells leads to increased cell divisions in the DZ and is proportional to T cell help [5, 76, 77]. LZ skewing was recapitulated through transcriptomics, with pathway analysis of Myc being upregulated with GT8‐IL‐21‐NIC immunization relative to GT8‐60mer alone in numerous clusters. These data suggest the GT8‐IL‐21‐NIC functions similarly to a synthetic Tfh cell, inducing a Myc signature in GC B cells. Additionally, genes associated with the cell cycle, such as Ccnd3, were significantly upregulated in numerous clusters with GT8‐IL‐21‐NIC immunization. Ccnd3 has been shown to be crucial for inertial cell cycling of GC B cells in response to potent Tfh help [58]. Additionally, we observed a gene program in DZ GC B cells reminiscent of the LZ in mice immunized with GT8‐IL‐21‐NIC; this could be due to direct interaction with the immunogen and B cells. Dendritic cells are capable of presenting unprocessed antigen to B cells within the germinal center, with the potential for direct DC‐B cell interaction at the follicle border [78, 79]. Further studies to visualize GT8‐IL‐21‐NIC distribution within the B cell follicle are warranted.
The V_H_1‐2/V_K_1‐33 transgenic mouse model is ideal for our study as it allows for physiologically‐relevant V(D)J rearrangement and enables us to examine the maturation of primed B cells when using eOD‐GT8‐based immunogens. In this model, the GT8‐IL‐21‐NIC drove improved GC responses that had more diversity, clonal expansion, and somatic hypermutation relative to the GT8‐60mer alone. This finding is crucial for vaccine design, as bnAb clones are often at low frequency in the naïve repertoire and require extensive SHM to become mature bnAbs. The NIC platform is potentially beneficial in this context as it provided further support of recruited B cells to participate in the GC reaction, leading to increased SHM. For rapidly mutating and/or structurally complex pathogens such as HIV, a competitive GC environment is recruited to get the necessary number of mutations for affinity maturation and clonal expansion. An approach using the NIC platform to improve upon these aspects could impact numerous vaccine modalities, especially germline targeting. Critically, we observed the acquisition of multiple key VRC01‐class mutations in recruited V_H_1‐2/V_K_1‐33 B cells with the GT8‐IL‐21‐NIC at only 7 days post immunization. Previous studies have demonstrated the acquisition of similar mutations in the memory B cell compartment more than 1 month post prime‐boost immunization [64]. This finding has important implications for sequential immunization schemes; the ability to prime B cells harboring these key mutations can potentially shorten the duration or even number of immunizations for a germline targeting approach.
There are some limitations in our study. We utilized electroporated plasmid DNA as our vaccine modality; additional studies to understand the immunogenicity of NICs using other delivery mechanisms such as recombinant protein in varying adjuvants, mRNA‐LNPs, or DNA‐LNPs [80] are warranted. Experiments were performed in numerous mouse models, but larger animal models and the inclusion of species‐specific IL‐21 may impact immunogenicity. The NICs generated in this work scaffolded IL‐21 to understand its impact on promoting GC responses. Further characterization of additional GC cytokines, chemokines, or costimulatory molecules as NICs may have varied potency due to optimal levels required in the GC reaction to potentiate proliferation and differentiation. IL‐21 non‐binding mutants can also be developed to further confirm the mechanism of action. Additionally, we characterized responses after a single immunization; understanding whether GT8‐IL‐21‐NIC priming or utilizing a sequential immunization strategy with a GT8‐IL‐21‐NIC prime could elicit neutralizing antibody responses is of importance, considering this is a critical goal of any prophylactic HIV vaccination effort. It will also be important in future work to understand the affinity of these recruited cells, and what their post‐GC fate is (such as differentiation into MBCs or LLPCs). Taken together, our findings show that the inclusion of GC cytokine IL‐21 enables robust serum antibody and germinal center responses with enhanced affinity maturation, which may provide value in the context of sequential immunization, germline targeting, or other vaccination approaches.
Experimental Section
4
Design and Modeling of Antigen‐Cytokine NIC Immunogens
4.1
Plasmid DNA‐encoded GT8‐60mer was developed in a previous work [26]. GT8 NICs were designed by genetic fusion of mouse IL‐21 (Uniprot E9PX58; GT8‐IL‐21‐NIC) to the C terminus of GT8. Models of the NIC nanoparticle were created using LS (PDB ID: 1HQK/7X7M) as the seeding base. The relevant sequence of GT8 and IL‐21 were predicted using ColabFold, an online accessible version of Alphafold as well as AlphaFold 2 [81] and AlphaFold 3 that are locally installed on the Wistar Institute High‐Performance Cluster. The Wistar Institute has an AlphaFold 3 [82] non‐commercial ShareAlike license and has signed the Alphafold3 Model Parameters Access Terms of Use policy. Predictions were verified using the closest crystal structure, if available [81, 83]. A single subunit of the relevant combinations of antigen and cytokine was created with GS linkers and then copied to the biological assembly of LS to create 60‐mers. Multiple geometries were tested, and each final nanoparticle version consisted of subunits sampled from two different geometries. The nanoparticle core, antigens, and cytokines are all shown as surface‐filled models, and GS linkers are shown in cartoon representation.
In Vitro Production and Size Exclusion Chromatography
4.2
All nanoparticles were expressed in Expi293F cells maintained in Expi293 expression medium (ThermoFisher) via Expifectamine transfection (Gibco, A14524) of plasmid following the manufacturer's protocol. Enhancers were added 18–22 h after transfection. The Supernatant was harvested after 6 days by centrifuging (4000 × g, 15 mins) and filtering (0.2 um Nalgene Rapid‐Flow Filter). Affinity purification was performed by lectin purification using lectin beads (Vector Laboratories) and lectin elution buffer (1 m Methyl alpha‐Dmannopyranoside). The nanoparticles were then purified over a size‐exclusion chromatography (SEC) Superdex S6 column (Cytiva 28990944). The Nanoparticles were aliquoted and flash‐frozen in thin‐walled PCR tubes prior to use.
Negative Stain Electron Microscopy Sample Preparation and Data Collection
4.3
Nanoparticles were diluted to 0.025 mg/mL in TBS and adsorbed onto carbon‐coated Cu400 EM grids that were glow‐discharged with an EasiGlow Glow Discharging System (Ted Pella) instrument for 60 s. Grids were blotted with Whatman filter paper and rinsed and blotted 3 times with 6 µL TBS. Grids were then stained with 3 µL 2% uranyl formate and immediately blotted and stained again for 90 s, and then blotted again. Grids were imaged using a Talos 120 kV microscope and Cata 16 M camera at 120 kV, 57,000× magnification, and around −1.6 µm defocus.
In Vitro Stimulation of Isolated Follicular B Cells
4.4
Spleens were harvested from B6.Blimp1^+/GFP^ (Prdm1^+/GFP^) mice by originally provided by Stephen Nutt (Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia), minced between frosted glass slides, and lysed with ACK lysis buffer to remove red blood cells. Single cell suspensions were labeled with Cell Trace Yellow according to manufacturer's protocol. Follicular B cells were purified by CD23 positive selection using MACS columns (Miltenyi). Purified cells were plated in 96‐well plates at 2 × 10^5^ cells per well with culture media containing RPMI 1640 supplemented with 10% Fetal Bovine Serum, 1% penicillin streptomycin, L‐glutamine, HEPES, non‐essential amino acids, sodium pyruvate, and 0.1% 2‐mercaptoethanol and gentamicin. Cells were stimulated with indicated concentrations of GT8‐60mer or GT8‐IL‐21‐NIC protein alone or in combination with IL‐4 (100 ng/mL) and IL‐5 (100 ng/mL) for 72 h incubated at 5% CO^2^ and 37°C. Stimulation conditions were plated at culture initiation. Downstream assessment was performed by harvesting cells from 96‐well plate using PBS + 0.1% BSA + 1 mm EDTA and staining with BV421 anti‐mouse CD86 (Clone GL1, Biolegend) and BUV395 anti‐mouse CD80 (Clone 16‐10A1, BD) for 30 min at 4°C before washing and resuspending in 1X ToPro‐3 solution. Flow cytometry was performed on a BD Symphony A3 Lite and analyzed using FlowJo analysis software. To assess STAT3 phosphorylation, cells were purified as above from wild‐type C57BL/6 mice before culture in RPMI 1640 with 10% FBS and 1% penicillin streptomycin containing the GT8‐60mer or GT8‐IL‐21‐NIC. Cells were cultured for 30 min before washing in PBS and incubation with Zombie Aqua Viability dye (BioLegend) and BV650 anti‐mouse CD19 (Clone 1D3, Biolegend) for 10 min at 37°C. Cells were subsequently washed and incubated with BD Cytofix buffer (12 min 37°C) and subsequently Biolegend Perm Buffer III (25 min on ice) to fix and permeabilize the cells respectively. Cells were then incubated with anti‐mouse pSTAT3 (Tyr705) (Clone 13A3‐1, Biolegend) for 30 min at RT. Samples were subsequently acquired on a BD Symphony A5.
Animal Studies
4.5
All animal studies were conducted under protocols 201399 and 201214 approved by the Wistar Institute Institutional Animal Care and Use Committee (IACUC). Six‐to‐twelve‐week‐old female BALB/cJ (Jackson Laboratory) and six‐to‐eight‐week‐old male and female V_H_1‐2R^JH2^/V_K_1‐33R^CSΔ/hTdT^ mice (from Frederick Alt) were housed in the Wistar Institute Animal Facility. Mice were immunized with indicated plasmid DNA constructs at 0.1 µg or 0.5 µg in the tibialis anterior muscle formulated in water followed by in vivo adaptive electroporation using the CELLECTRA 3P device (Inovio Pharmaceuticals). For serological experiments, mice were bled weekly or biweekly by submandibular bleed. For germinal center experiments, lymph nodes were harvested at indicated timepoints post vaccination.
Tissue Processing
4.6
Mice were euthanized using CO_2_ at timepoints where indicated and popliteal and iliac draining lymph nodes were harvested.
Microscopy
4.6.1
Iliac lymph nodes (LNs) were placed into 10% formalin for 1 h at RT. Iliac LNs were subsequently placed in 15% sucrose 0.1% sodium azide (NaN_3_) in PBS for 6 h at 4C before final incubation in 30% sucrose 0.1% NaN_3_ in PBS overnight (ON) at 4C. The following day, lymph nodes were washed once in PBS before freezing in Optimal Cutting Temperature (OCT) compound on dry ice. OCT blocks were subsequently sectioned by microtome at 10 µM thickness onto slides before subsequent staining.
Flow cytometry
4.6.2
Popliteal and iliac LNs were pooled into RPMI media containing 10% FBS and 1% Penicillin/Streptomycin (R10). LNs were then mechanically dissociated through a 40 µm strainer before washing with R10. Samples were plated for subsequent flow cytometry staining or downstream assays.
Immunofluorescent Microscopy
4.7
Iliac lymph nodes sections were blocked in 5% donkey serum in PBS‐Tween‐20 for 30 min at RT with FcR block, followed by staining with anti‐mouse IgD (Clone 11–26c), anti‐mouse CD3 (Clone 17A2), anti‐mouse CD21/35 (Clone 7E9), and anti‐mouse GL7 (Clone GL7) for 1 h at RT. Slides were washed and mounted using ProLongTMDiamond antifade (Invitrogen) and imaged using a Zeiss LSM 980 Confocal microscope. Images were analyzed using Fiji.
Flow Cytometry
4.8
Germinal Centers
4.8.1
LNs were resuspended in Fixable Viability Dye eF780 (eBioscience) in 0.2% BSA in PBS (FACS) for 10 min at RT. Cells were next washed before resuspension with Biotin anti‐mouse CXCR5 (Clone SPRCL5, eBioscience) or Biotin anti‐mouse CXCR4 (Clone 2B11, eBioscience) for 30 min at RT in FACS. Cells were subsequently washed and resuspended with a surface stain cocktail for 30 min at RT containing a subset of the antibodies or proteins listed in Table S7. Probes were made via direct conjugation of fluorochromes to recombinant ferritin‐scaffolded GT8 (GT8‐24mer), biotinylated Avi‐tagged monomeric eOD‐GT8, or biotinylated Avi‐tagged monomeric GT8‐KO11 made in house as in [26] or using streptavidin. Lightning Link Kits for FITC (ab102884) or PE (ab102918) were used according to manufacturer's protocol. Cells were then washed and resuspended in FACS buffer before acquisition. If staining for transcription factors, cells were fixed and permeabilized using the eBioscience Foxp3/Transcription Factor Staining set (Thermo Fisher) for 1 h at room temperature, followed by washing with 1x Permeabilization Buffer before incubation with PE‐eF660 anti‐mouse Foxp3 for 30 min at room temperature. Cells were then washed and resuspended in FACS buffer before acquisition. All samples were acquired on a FACSymphony A3 or A5 SE analyzer (BD). FCS files were exported and analyzed using FlowJo (Treestar).
Activation Induced Marker Stimulation
4.8.2
LNs were first incubated with a pool of overlapping peptides spanning the GT8 and lumazine synthase domains for 18 h at 37°C with APC anti‐mouse CD40L (Clone MR1, Biolegend). DMSO and Cell Stimulation Cocktail (eBioscience) were used as positive and negative controls, respectively. After stimulation, cells were first washed in FACS before incubation with Fixable Viability Dye eF780 (eBioscience) for 10 min at RT. Cells were then washed in FACS before incubation with a surface stain cocktail for 30 min at RT containing a subset of antibodies or proteins listed in Table S7. After incubation, cells were then washed in FACS before acquisition on a FACSymphony A3 or A5 SE analyzer (BD). FCS files were exported and analyzed using FlowJo (Treestar).
ELISA
4.9
IL‐21R‐VRC01 Binding ELISA
4.9.1
96‐well half area plates (Cat# 3690, Corning) were coated with 1 µg/mL mouse IL‐21R (Cat# 51184‐M08H, Sino Biologicals) overnight at 4°C. The following day, plates were blocked with 5% non‐fat dry milk in PBS for 1 h at RT. Plates were then incubated with serially‐diluted recombinant GT8‐60mer or GT8‐IL‐21‐NIC for 1 h at RT. Plates were washed and incubated with VRC01 (obtained through BEI resources, NIAID, NIH, ARP‐12033) at 1 µg/mL for 1 h at RT. Plates were subsequently washed and incubated with goat anti‐human IgG‐Fc (Cat# A80‐304P, Bethyl) for 1 h at RT. Plates were washed and developed with TMB Ultra Substrate (Thermo) for 5 min at RT before quenching with 1N H_2_SO_4_. Washing occurred between each step using 0.05% Tween‐20 in PBS. Plates were read on a Biotek Synergy plate reader at 450 and 570 nm, and background optical density at 570 nm was subtracted from 450 nm.
GT8 Binding ELISA
4.9.2
96‐well half area plates were coated at 1 µg/mL with anti‐His capture antibody (Cat#A01857, Genscript) overnight at 4°C. The following day, plates were blocked with 5% non‐fat dry milk in PBS for 1 h at RT. Plates were incubated with 1 µg/mL His‐tagged GT8 monomer (in house, see [26]) for 2 h at RT. Mouse samples were then added in serial dilutions for 2 h at RT. Plates were then incubated with HRP‐conjugated goat anti‐mouse IgG heavy+light chain (Cat# A90‐516A, Bethyl) for 1 h at RT. Plates were then developed and read as above. Endpoint titers were calculated against naïve mouse serum. The endpoint titer was defined as the highest dilution where the OD value was greater than cutoff determined using the following formula: Average(NaïveMice)+(4×SD(NaïveMice)).
ELISpot
4.10
Bone marrow (BM) was flushed from femurs and tibia from each mouse using a 23G X 3 /4″ needle and syringe into FACS buffer and filtered through 63‐micron Nitex mesh. Red blood cells were lysed in ACK lysis buffer for 5 min on ice. The resulting cells were counted using a Nexcelom Cellometer Auto 2000 Cell Viability Counter Profiler (Nexcelom Bioscience LLC). MultiScreenHTS IP Filter Plate, 0.45 µm (Millipore Sigma), was coated with 1 µg/mL anti‐His (Cat#A01857, Genscript) followed by His‐tagged GT8 monomer at 1 µg/mL (GT8 immunogens) in sodium carbonate/sodium bicarbonate buffer (pH 9.6) (35 mm NaHCO_3_ and 15 mm Na_2_CO_3_) for 1 h at 37°C. Plates were then washed with 200 µL PBS/well three times and blocked at 37°C in complete RPMI + 10% FBS for 30 min. BM cells were plated in eight halving dilutions beginning at 2.5 million cells per well and incubated overnight in complete RPMI + 10% FBS. Plates were then washed with wash buffer (1x PBS + 0.1% Tween 20) five times and incubated with biotinylated polyclonal goat anti‐mouse anti‐IgG detection antibody (SouthernBiotech) in PBS + 2% BSA at RT for 1 h. Plates were once again washed five times, and streptavidin‐alkaline phosphatase (1:20,000 dilution in PBS + 2% BSA) was added prior to incubation at RT for 30 min. Plates were then washed five times with wash buffer, and 50 µL/well BCIP/NBT single solution (SigmaAldrich) was added for 5 min or until spots developed, at which time the reaction was quenched with 100 mL 1 m sodium phosphate monobasic solution. After plates were rinsed with dH2O and dried overnight, they were scanned and counted using Mabtech IRIS.
Cell Sorting
4.11
Mice were immunized and GC B cells were sorted from DLNs 7 days post immunization. Briefly, DLNs were mechanically dissociated (as above) and stained with Fixable Viability Dye eF780 (eBioscience) in 0.2% BSA in PBS (FACS) for 10 min at RT. Cells were washed and subsequently stained with a cocktail of the following antibodies listed in Table S8 (for transcriptomics experiment in BALB/c mice in Figure 3) or Table S9 (for V(D)J sequencing in V_H_1‐2R^JH2^/V_K_1‐33R^CSΔ/hTdT^ mice in Figure 4) for 30 min at RT. Probes made using Avi‐tagged Biotinylated eOD‐GT8 monomer and GT8‐KO11 monomer were done on protein produced in house. Cells were then sorted using a MoFlo Astrios or Symphony S6 sorter for downstream processing.
Single Cell RNA Sequencing Sample Preparation, Transcriptomics
4.12
Antigen‐specific germinal center B cells were sorted and subsequently uniquely barcoded using the 10× Genomics Chromium single‐cell platform, and complementary DNA (cDNA) libraries were prepared for Next Generation Sequencing according to the manufacturer's protocol (Chromium NextGem Single Cell 3' V2 Library and Gel Bead Kit, 10× Genomics, USA). Cell suspensions of each sample, reverse transcription master mix, and partitioning oil were loaded on a single‐cell “G” chip with a targeted cell output of 10,000 cells per library and then run on the Chromium Controller. A total of 4 lanes were used on the G chip, 1 lane/sample. Reverse transcription was performed within the droplets at 53°C for 45 min and newly synthesized cDNA was amplified for 11 cycles on a Veriti Thermal Cycler (Thermofisher, USA). cDNA size selection was performed using SPRIselect beads (Beckman Coulter, USA) following manufacturer's protocol. cDNA was analyzed on an Agilent Tapestation High Sensitivity D5000 (Agilent, USA) for qualitative and quantitative control purposes. cDNA was then used to make libraries according to manufacturer's protocol (Chromium Next GEM Single Cell 3’ Reagent Kit v3.1). cDNA was fragmented using the proprietary fragmentation enzyme blend for 5 min at 32°C, followed by end‐repair and A‐tailing at 65°C for 30 min. cDNA was double‐sided size selected using SPRIselect beads. Sequencing adaptors were ligated to the cDNA at 20°C for 15 min. and after a round of post‐ligation SPRIselect bead clean‐up, cDNA was amplified for 14 cycles using a sample‐specific index oligo as a primer. A final round of double‐sided size selection using SPRIselect beads followed. Final library size and quantity was determined using an Agilent Tapestation High Sensitivity D5000 (Agilent, USA) and a Qubit dsDNA High Sensitivity Assay kit (Thermofisher, USA), respectively. Additional library quantification was done using the Kapa Library Quantification kit for Illumina Libraries (Roche, USA).
Single Cell RNA Sequencing Sample Preparation, V(D)J Sequencing in VH1‐2/VK1‐33 Tg Mice
4.13
Antigen‐specific germinal center B cells (live CD4‐ CD19+ CD38‐ Fas+ eOD‐GT8++ KO11‐) were sorted and subsequently uniquely barcoded using the 10x Genomics Chromium single‐cell platform. Complementary DNA (cDNA) libraries were prepared for Next Generation Sequencing according to the manufacturer's protocol (Chromium GEM‐X Single Cell 5’ V3 Library and Gel Bead Kit, 10× Genomics, USA). Cell suspensions of sample ID: GT8‐60mer and sample ID: GT8‐IL21‐NIC, reverse transcription master mix, and partitioning oil were loaded on a GEM‐X 5’ chip with a targeted cell output of 2,000 (GT8‐60mer) and 20,000 (GT8‐IL21‐NIC) cells per library. The chip was then run on the Chromium X instrument in which a total of 2 lanes were used on the GEM‐X 5’ chip, 1 lane/sample. Reverse transcription was performed within the droplets at 48°C for 45 min and newly synthesized cDNA was amplified for 16 (GT8‐60mer) and 12 (GT8‐IL21‐NIC) cycles on a Veriti Thermal Cycler (Thermofisher, USA). cDNA size selection was performed using SPRIselect beads (Beckman Coulter, USA) following manufacturer's protocol. cDNA was analyzed on an Agilent Tapestation High Sensitivity D5000 (Agilent, USA) and the Qubit dsDNA High Sensitivity Assay kit (Thermofisher, USA), for qualitative and quantitative control purposes, respectively. Mouse V(D)J region PCR amplification was performed using BCR amplification primers (Single cell mouse BCR amplification kit, 10× Genomics, USA) according to manufactures instructions. Specifically, two rounds of PCR amplification were done, each with a total of 8 cycles and a double‐sided size selection using SPRIselect beads, post‐amplification. Amplified mouse VDJ cDNA was analyzed on an Agilent Tapestation High Sensitivity D5000 and the Qubit dsDNA High Sensitivity Assay kit for qualitative and quantitative control purposes. Libraries for sequencing were generated according to manufacturer's protocol (Chromium GEM‐X Single Cell 5’ V3 Library and Gel Bead Kit, 10× Genomics, USA). VDJ cDNA was fragmented using the proprietary fragmentation enzyme blend for 2 min at 32°C, followed by end‐repair and A‐tailing at 65°C for 30 min. cDNA was double‐sided size selected using SPRIselect beads. Sequencing adaptors were ligated to the cDNA at 20°C for 15 min. and after a round of post‐ligation SPRIselect bead clean‐up, cDNA was amplified for 7 cycles using a sample‐specific index oligo as a primer. A final round of double‐sided size selection using SPRIselect beads followed. Final library size and quantity was determined using an Agilent Tapestation High Sensitivity D5000 and a Qubit dsDNA High Sensitivity Assay kit, respectively. Additional library quantification was done using the Kapa Library Quantification kit for Illumina Libraries (Roche, USA). Two libraries were pooled and sequenced on the Nextseq 2000 (Illumina, San Diego, CA) using a P1 100 cycle kit (100 M reads total, Illumina), paired end run with the following run parameters: 26 base pair × 8 base pair (dual index) × 90 base pair.
Data Processing, Transcriptomics
4.14
CellRanger suite (pipeline v7.0.0 https://support.10xgenomics.com) was used with refdata‐gex‐mm10‐2020‐A transcriptome as a reference to map reads on the mouse genome. The barcodes, genes, and counts were imported into R package Seurat [84] where quality control was performed by discarding cells with over 5% mitochondrial content and those with fewer than 250 genes with reads. R package SingleR [85] was used to identify individual cell types with the ImmGen mouse dataset as a reference and around 110 cells that weren't predicted as B cells were further discarded. After these steps, there were 7822 cells in the GT8_60mer sample and 8831 cells in the GT8_IL21_NIC sample. Seurat's shared nearest neighbor was then used to unbiasedly cluster the remaining cells and UMAPs were used for visualization. Sample integration was not necessary as a batch effect was not observed. Germinal center B cells were subset, and reclustered to further identify transitionary states [42]. Slingshot was used for pseudotime analysis with the LZ cluster as the initial node [86]. Differential Expression analysis was performed using Wilcoxon Rank Sum Test and FDR < 5% was used to determine statistical significance. Qiagen IPA (QIAGEN Inc., https://digitalinsights.qiagen.com/IPA) was used to perform enrichment analysis to identify over‐represented pathways and regulators. Visualization was done on RStudio with ggplot2 and scCustomize.
Data Processing, V(D)J Sequencing in VH1‐2RJH2/VK1‐33RCSΔ/hTdT Mice
4.15
Illumina next‐generation sequencing (NGS) data were processed using the Cell Ranger V(D)J pipeline (v9.0.0, 10x Genomics), which performs read alignment, contig assembly, gene annotation, and heavy/light chain pairing. A custom V(D)J reference was used to accommodate the engineered IGHV1‐2/IGKV1‐33 mouse model. This reference was generated by modifying the mouse (GRCm39) V(D)J reference, replacing the IGHV5‐2, IGHJ, and IGKV3‐2 gene segments with the human IGHV1‐2, IGHJ2, and IGKV1‐33 gene segments from the GRCh38 human reference dataset, respectively. Assembled contigs generated by Cell Ranger were subsequently re‐annotated using IgBLAST (v1.22.0, NCBI), for which an IMGT germline gene database was generated for the engineered IGHV1‐2/IGKV1‐33 mouse as previously described. Downstream processing of annotated contigs was performed using the Immcantation framework (Change‐O v1.3.1‐2025.03.27) with the customized IMGT germline gene sequences and clonal assignment was performed with a clonal grouping threshold of 0.15. Additional processing was carried out using a custom adaptation of the IgPipeline v3 workflow with the aforementioned IMGT germline gene database. Phylogenetic analysis of B cell clonal lineages was performed using the Immcantation framework, building upon the previously described clonal assignments. A single germline sequence per clone was reconstructed using the heavy chain IMGT germline gene database customized for the engineered IGHV1‐2/IGKV1‐33 mouse. To reduce ambiguity from junctional diversity, the D gene segment was masked during reconstruction. Reconstructed germline and clonal sequences were formatted for phylogenetic analysis and used as input to IgPhyML (v2.1.0). Initial tree inference was performed using the GY94 codon substitution model, followed by refinement with the HLP19 model, which incorporates features of somatic hypermutation. Trees were visualized using iTOL (Interactive Tree Of Life).
Statistical Analysis
4.16
All statistical tests were performed using GraphPad Prism 10. Graphs and error bars represent means ± SD or geometric mean ± geometric SD where indicated in the figure legends. Non‐parametric Mann‐Whitney U tests, one‐way ANOVA adjusted for multiple comparisons, or two‐way ANOVA adjusted for multiple comparisons were used to compare groups where indicated. The number of samples in each graph was notated in the figure legend. In all datasets, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.
Author Contributions
N. J. Tursi, D. W. Kulp, and D. B. Weiner conceptualized the project and designed experiments. N. J. Tursi, C. J. Agostino, J. Huang, T. Kannan, N. Laenger, J. Londregan, K. Lederer, M. Helble, N. Bedanova, C. Livingston, E. N. Gary, M. McCanna, M. T. Marqis, R. Habib, and I. R. Relano performed experiments and analyzed data. I. Maillard, A. Patel, D. Allman, A. Kossenkov, and A. Escolano provided crucial reagents or overisght. N. J. Tursi, C. J. Agostino, J. Huang, T. Kannan, and A. Kossenkov analyzed the sequencing datasets. N. J. Tursi, C. J. Agostino, J. Huang, T. Kannan, D. W. Kulp, and D. B. Weiner wrote the manuscript. All authors contributed to editing the manuscript. D. W. Kulp and D. B. Weiner jointly supervised this work.
Conflicts of Interest
D.B.W. has received grant funding, participates in industry collaborations, has received speaking honoraria, and has received fees for consulting, including serving on scientific review committees. Remunerations received by D.B.W. include direct payments and equity/options. D.B.W. also discloses the following associations with commercial partners: Geneos (consultant/advisory board), AstraZeneca (advisory board, speaker), Inovio (board of directors, consultant), Sanofi (advisory board), BBI (advisory board), Pfizer (advisory Board), and Advaccine (consultant). I.M. has received research funding from Genentech and Regeneron, and has been a member of Garuda Therapeutics's scientific advisory board. All other authors declare no competing interests.
Supporting information
Supporting File: advs73554‐sup‐0001‐SuppMat.docx.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1G. D. Victora and M. C. Nussenzweig , “Germinal Centers,” Annual Review of Immunology 40 (2022): 413–442.10.1146/annurev-immunol-120419-02240835113731 · doi ↗ · pubmed ↗
- 2R. Nakagawa and D. P. Calado , “Positive Selection in the Light Zone of Germinal Centers,” Frontiers in Immunology 12 (2021): 661678.33868314 10.3389/fimmu.2021.661678 PMC 8044421 · doi ↗ · pubmed ↗
- 3S. Crotty , “T Follicular Helper Cell Differentiation, Function, and Roles in Disease,” Immunity 41 (2014): 529–542.25367570 10.1016/j.immuni.2014.10.004PMC 4223692 · doi ↗ · pubmed ↗
- 4S. Crotty , “T Follicular Helper Cell Biology: A Decade of Discovery and Diseases,” Immunity 50 (2019): 1132–1148.31117010 10.1016/j.immuni.2019.04.011PMC 6532429 · doi ↗ · pubmed ↗
- 5D. Dominguez‐Sola , G. D. Victora , C. Y. Ying , R. T. Phan , M. Saito , and M. C. Nussenzweig , “The Proto‐Oncogene MYC is Required for Selection in the Germinal Center and Cyclic Reentry,” Nature Immunology 13 (2012): 1083–1091.23001145 10.1038/ni.2428 PMC 3711534 · doi ↗ · pubmed ↗
- 6C. Jandl and C. King , “Cytokines in the Germinal Center Niche,” Antibodies (Basel) 5 (2016): 5.31557986 10.3390/antib 5010005 PMC 6698856 · doi ↗ · pubmed ↗
- 7D. Zotos , J. M. Coquet , Y. Zhang , et al., “IL‐21 Regulates Germinal Center B Cell Differentiation and Proliferation through a B Cell–Intrinsic Mechanism,” Journal of Experimental Medicine 207 (2010): 365–378.20142430 10.1084/jem.20091777 PMC 2822601 · doi ↗ · pubmed ↗
- 8I. Quast , A. R. Dvorscek , C. Pattaroni , et al., “Interleukin‐21, Acting Beyond the Immunological Synapse, Independently Controls T Follicular Helper and Germinal Center B Cells,” Immunity 55 (2022): 1414–1430.35896116 10.1016/j.immuni.2022.06.020 · doi ↗ · pubmed ↗