Precursor Frequency and Affinity Determine B Cell Competitive Fitness in Germinal Centers, Tested with Germline-Targeting HIV Vaccine Immunogens

Abstract

How precursor frequencies and antigen affinities impact interclonal B cell competition is a particularly relevant issue for candidate germline-targeting HIV vaccine designs because of the in vivo rarity of naive B cells that recognize broadly neutralizing epitopes. Knowing the frequencies and affinities of HIV-specific VRC01-class naive human B cells, we transferred B cells with germline VRC01 B cell receptors into congenic recipients to elucidate the roles of precursor frequency, antigen affinity, and avidity on B cell responses following immunization. All three factors were interdependently limiting for competitive success of VRC01-class B cells. In physiological high-affinity conditions using a multivalent immunogen, rare VRC01-class B cells successfully competed in germinal centers (GC), underwent extensive somatic hypermutation, and differentiated into memory B cells. The data reveal dominant influences of precursor frequency, affinity, and avidity for interclonal GC competition and indicate that germline-targeting immunogens can overcome these challenges with high-affinity multimeric designs.

Keywords: HIV; T follicular helper (Tfh); VRC01; affinity maturation; antibody; avidity; humoral immunity; immunoglobulins; interclonal competition; vaccine.

Figure 1. Development of a VRC01^gHL B Cell Transfer Model to Assess Precursor Frequency Effects on Immune Responses

(A) pERK induction in VRC01^gHL B cells upon stimulation with eOD-GT8 60-mer. n = 3 mice. (B) 10⁶ CFSE-labeled heterozygous VRC01^gHL B cells were transferred into CD45.1⁺ hosts. Host mice were immunized with eOD-GT8 60-mer. Splenocytes were harvested 3 days later and VRC01^gHL B cell division was analyzed by flow cytometry. Gated on scatter/singlet/live (SSL), B220⁺, CD4⁻, CD45.2⁺, CD45.1⁻. n = 3 mice. (C) Schematic of the VRC01^gHL B cell transfer system used for (D)–(J). (D) Frequency of total VRC01^gHL B cells (SSL, CD4⁻, B220⁺, CD45.2⁺, CD45.1⁻) on day 8 in spleens of mice immunized as in (B), seeded with different VRC01^gHL cell precursor frequencies. n = 4 mice/group. (E) Representative plot of (red) eOD-GT8 60-mer binding to VRC01^gHL GC B cells (SSL, CD4⁻, B220⁺, CD45.2⁺, CD45.1⁻, GL7⁺, BCL6⁺) or (blue) naive host B cells (SSL, CD4⁻, B220⁺, CD45.2⁻, CD45.1⁺, GL7⁻, BCL6⁻) on day 8 post-immunization. Host mice were seeded with a 1:10⁶ VRC01^gHL B cell precursor frequency prior to immunization. n = 4 mice/group. (F) Representative day 8 serum IgG ELISA data. Blue = WT eOD-GT8 60-mer; Red = CD4bs-KO eOD-GT8 60-mer. n = 4 mice/group. (G–I) Day 8 GC B cells in mice immunized as in (D). n = 4 mice/group. Representative flow cytometry (G) and quantification (H) of GC B cells. VRC01^gHL B cells (CD45.2⁺) and endogenous B cells (CD45.1⁺) are indicated within the GC compartment. Pre-gated on SSL, B220⁺, CD4⁻. (I) Quantitation of VRC01^gHL B cells among GC B cells from G. * = p < 0.05. Full gating shown in Figure S1P. (J) Day 8 spleen GC histology from a host mouse seeded with a 1:10⁶ VRC01^gHL B cell precursor frequency. Green, B220; red, CD45.2; white, GL7. Independent Experiments = E. (A) E = 3; (C) E = 2; (D) E = 2; (E) E = 2; (F) E = 2; (G) E = 2; (H) E = 2; (I) E = 2; (J) E = 2. Data from one experiment are shown in each panel. See also Figures S1 and S2. Error bars are SEM.

Figure 2. Antigen Affinity and Precursor Frequency Are Limiting for Early GC Occupancy by VRC01^gHL B Cells

(A) Relative monovalent affinities of gVRC01 for eOD-GT1 (GT1), eOD-GT2 (GT2), and eOD-GT5 (GT5) monomers (left axis, “eOD-GT series”), in comparison to affinities of human naive B cell VRC01-class Abs for eOD-GT8 (right axis and data points, “Human naive VRC01-class B cell affinities”). Human B cell data points from (Jardine et al., 2016a). (B) Frequency of VRC01^gHL B cells among GC B cells on day 8 following immunization of host mice with eOD-GT1, -GT2, or –GT5 60-mer nanoparticles. Host mice started with the different VRC01^gHL precursor frequencies (PF) indicated. GC B cells gated as SSL, B220⁺, CD4⁻, BCL6⁺, GL7⁺. LOD = limit of detection (see STAR Methods for calculations). Each immunogen shown was conducted as an independent experiment. n = 3–4 mice/group. (C) Quantification of VRC01^gHL B cells among GC B, as gated in B. See methods for additional details. n = 3–4 mice/group. (B) E = 2; (C) E = 2. Data from one experiment are shown in each panel. See also Figure S3. Error bars are SEM.

Figure 3. Rare VRC01^gHL B Cells Compete Poorly after Monomeric Protein Immunization

(A–D) VRC01^gHL B cells were transferred as in Figure 1B to generate CD45.1/VRC01^gHL mice with either high (1:10³ B cells) or low (1:10⁶ B cells) precursor frequencies. Mice were immunized with the monomeric eOD-GT proteins indicated. (A and B) Flow cytometric plots (A) and quantitation (B) of total GC B cells in CD45.1/VRC01^gHL mice on day 8 following immunization with the monomeric immunogens indicated. Pre-gated on SSL, B220⁺, CD4⁻. n = 3 mice/group. (C and D) (C) Flow cytometric plots and (D) quantitation of VRC01^gHL B cells among GC B cells in CD45.1/VRC01^gHL mice on day 8 following immunization with the monomeric immunogens indicated. Pre-gated on SSL, B220⁺, CD4⁻, Bcl6⁺, GL7⁺. n = 3 mice/group. (A–D), E = 2. Data from one experiment shown in each panel. See also Figure S3. Error bars are SEM.

Figure 4. Antigen Affinity Is Limiting for GC Fitness of VRC01^gHL B Cells

(A) Frequency of GC B cells over time in eOD-GT5 60-mer (top), or eOD-GT2 60-mer (bottom) immunized CD45.1/VRC01^gHL mice, starting with a host mouse VRC01^gHL precursor frequency of 1:10⁶ B cells prior to immunization. Dotted line represents the average splenic GC B cell frequency in unimmunized controls. n = 4 mice/group. (B) Flow cytometric plots showing VRC01^gHL (CD45.2⁺) GC B cells and endogenous (CD45.1⁺, “host”) GC B cells among total GC B cells of mice in (A). Cells pregated as in Figure 2B. n = 4 mice/group. (C and D) Quantitation of VRC01^gHL GC B cells in eOD-GT5 60-mer-immunized mice (C, purple) or eOD-GT2 60-mer-immunized mice (D, blue) as a percentage of total GC B cells. n = 4 mice/group. (E) Quantitation of VRC01^gHL B cells among total B cells in the mice shown in (A)–(D). n = 4 mice/group. (F) eOD-GT2 or eOD-GT5 and eOD-GT8-CD4bs-KO probe binding by endogenous GC B cells in mice immunized as in (A). Gated as SSL, B220⁺, CD4⁻, BCL6⁺, GL7⁺, CD45.2⁻, CD45.1⁺. n = 4 mice/group. (G) Quantitation of non-CD4bs-specific (left) and CD4bs-specific (right) endogenous (“host”) GC B cell responses in mice immunized with eOD-GT5 60-mer as in (A). Gated as in (F). n = 4 mice/group. (H) Quantitation of non-CD4bs-specific (left) and CD4bs-specific (right) endogenous (“host”) GC B cell responses in mice immunized with eOD-GT2 60-mer as in (A). Gated as in (F). n = 4 mice/group. (I and J) Distribution of eOD-GT CD4bs-specific HC-LC pairs from endogenous GC B cells on day 8 (I) and day 36 (J). Each circle represents the fraction of BCRs that utilize the same HC V-D-J and LC V-J genes. Circles are colored by VK gene, and are text labeled with the corresponding VH gene. Smallest circle = 1 sequence, largest circle = 7 sequences. V gene color code shown in Figure S4. Day 8 data from GT5 and GT2 were pooled (eOD-GT5 and –GT2 n = 4 mice each). Day 36 data were eOD-GT5 (n = 4 mice). See Figure S4 for additional analysis. (K) The distribution of unpaired VK gene usage in eOD-GT5 60-mer-immunized mice on day 8 (four mice, VK n = 141) and day 36 (four mice, VK n = 130). The most commonly represented VK genes are shown. VK gene usage in naive C57BL/6J mice is shown for reference. (L) Change in frequently represented IGHV gene usage in eOD-GT5 60-mer-immunized mice between days 8 (four mice, n = 169 sequences) and 36 (four mice, n = 171). Baseline usage frequency of the listed VH genes in naive C57BL/6J mice (Collins et al., 2015) is shown by the color matched dotted lines. (A and B) E = 2; (C) E = 2; (D) E = 2; (E) E = 2; (F) E = 2; (G) E = 2; (H) E = 2; (I) E = 2; (J) E = 1. Data from one representative experiment shown in (A)–(H). Aggregated data are shown in (I)–(L). See also Figure S4. Error bars are SEM.

Figure 5. VRC01^gHL B Cell Rapidly Dominate Individual GCs under Higher Affinity Conditions

(A) Representative histological images of splenic cryosections from mice immunized with eOD-GT5 60-mer. The day post-immunization is indicated. Host mice VRC01^gHL precursor frequency was 1:10⁶ B cells prior to immunization. Blue = B220, Green = TCR-β, Red = GL7, White = CD45.2. High magnification images of representative GCs indicated by white boxes are shown below. (B) The fraction of GCs containing any VRC01^gHL B cells (CD45.2⁺) on the indicated day. (C) Quantitation of percentage GC occupancy by VRC01^gHL B cells (CD45.2⁺) for each VRC01^gHL+ GC. Percentage occupancy was calculated based on area. Each dot represents an individual GC. Each shade of color represents an individual mouse. Each color (blue, green, purple) represents an individual experiment. Red bar indicates median percentage VRC01^gHL GC occupancy, with number shown above each violin plot. Violin plots show distribution frequencies. Results are pooled from two or three independent experiments (n = 8 total per time point, except d36 where n = 12). Total numbers of GCs analyzed were: d8 = 453, d12 = 598, d16 = 302, d20 = 404, d36 = 254. Numbers of VRC01^gHL+ GCs assessed for percentage area occupancy by VRC01^gHL B cells were: d8 = 153, d12 = 156, d16 = 66, d20 = 105, d36 = 45. ** = p < 0.01, and **** = p < 0.0001. Additional histology examples shown in Figure S5 for reference.

Figure 6. VRC01^gHL B Cells Undergo Extensive SHM

(A–G) VRC01^gHL BCR sequence data from mice immunized with eOD-GT5 60-mer. Host mice VRC01^gHL precursor frequency was 1:10⁶ B cells prior to immunization. (A) Circle charts represent the fraction of HC and LC sequences that acquired the indicated number of aa mutations at day 8, 16, and 36 post-immunization. The total number of individual B cell sequences is shown at the center of the circle. (B) Frequency of observed HC and LC aa mutations per residue position at day 8, 16, and 36 post-immunization from sequences in (A). Residue positions are listed sequentially. CDRs are highlighted in gray. Asterisks indicate residues analyzed in (F) and (G). (C–E) bnAb-type aa HC mutations in VRC01^gHL B cells from (A and B), shown for d8 (C), d16 (D), and d36 (E). The red diagonal line indicates a 100% efficiency of VRC01-class bnAb-type HC mutations. The black stair step indicates a calculated VH1-2 antigen-agnostic mutation distribution, which might include Ab structure stabilizing mutations. HCDR3 mutations were not included in this analysis, since HCDR sequences vary among different VRC01-class bnAbs. (F) Circle charts of commonly observed day 36 HC and LC mutations. The aa position (Kabat numbering) and residue in gVRC01 is shown at the center of the circle. VRC01-class bnAb mutations are shown in shades of blue, with the light blue segment representing an exact mutation that occurs in the mature VRC01 bnAb. Non-VRC01-class mutations are colored in shades of yellow and red. (G) Distribution of select VRC01^gHL B cell HC aa mutations over time. Composite data from all mice are shown in each panel. (A)–(G), day 8, 16, 36 E = 2. Day 8, n = 5 mice (HC and LC); day 16, n = 6 mice (HC), n = 5 mice (LC); day 36, n = 4 mice (HC and LC). See also Figure S6 and Tables S1, S2, and S3.

Figure 7. VRC01^gHL B Cells Form Memory

(A and B) VRC01^gHL memory B cell formation (B220⁺IgD^lo/−CD38⁺GL7⁻) in CD45.1/VRC01^gHL (1:10⁶ precursor frequency) host mice on day 36 following immunization with the immunogens indicated. (A) Representative flow cytometric plots and gating strategy. Prior gating on SSL, B220⁺, CD4⁻, CD8⁻. (B) Quantitation of VRC01^gHL memory B cells. LOD = limit of detection (see STAR Methods for calculations). n = 4 mice/group. Data were pooled from 4 (GT5) and 2 (GT2) experiments. (C) Expression of surface markers on VRC01^gHL memory B cells (blue) in eOD-GT5 60-mer immunized mice, gated as in A. Orange, total non-GC B cells (SSL, CD4⁻, CD8⁻, B220⁺, CD38⁺, GL7⁻). n = 4 mice/group. (D) Quantitation of B cells gated as in (C). n = 4 mice/group. E = 2. (E and F) CSR to IgG1 for VRC01^gHL memory B cells, in eOD-GT5 60-mer immunized mice, gated as in A. (E) Representative flow cytometry plot, with total endogenous B cells as a black contour plot and VRC01^gHL memory B cells overlaid as a blue dot plot. (F) Quantitation of IgG1⁺ memory B cells. n = 4 mice/group. E = 2. See also Figure S6. Error bars are SEM.