Continuous germinal center invasion contributes to the diversity of the immune response

Abstract

Antibody responses are characterized by increasing affinity and diversity over time. Affinity maturation occurs in germinal centers by a mechanism that involves repeated cycles of somatic mutation and selection. How antibody responses diversify while also undergoing affinity maturation is not as well understood. Here, we examined germinal center (GC) dynamics by tracking B cell entry, division, somatic mutation, and specificity. Our experiments show that naive B cells continuously enter GCs where they compete for T cell help and undergo clonal expansion. Consistent with late entry, invaders carry fewer mutations but can contribute up to 30% or more of the cells in late-stage germinal centers. Notably, cells entering the germinal center at later stages of the reaction diversify the immune response by expressing receptors that show low affinity to the immunogen. Paradoxically, the affinity threshold for late GC entry is lowered in the presence of high-affinity antibodies.

Keywords: antibody affinity; antibody diversity; germinal center.

Figure 1.. Acquisition of activated and GC markers

(A) Schematic of the experiment: 2 × 10⁵ CTV-labeled B1-8^hi AID-GFP CD45.1 congenic B cells were transferred on day −1 to C57BL/6J mice intravenously. Recipient mice were immunized with NP-OVA (footpad injection) on day 0 and B cells from draining lymph nodes were analyzed according to the number of divisions on day 7 (left). Representative flow cytometry profiles and gating strategy for tracking of transferred cells (right). (B) Representative flow cytometry plots of transferred (CD45.1⁺ Igλ⁺, blue) and host (CD45.2⁺, gray) B cells after the number of divisions indicated above the plots. CD95 and GL7 (top), CD38 and IgD (middle), and AID-GFP expression (bottom). (C) Graphs summarizing the geometric mean fluorescence intensities (gMFI) of naïve (CD38 and IgD) and GC B cells (GL7, CD95 and AID) markers and the frequency of AID-expressing cells with cell division. Representative of 2 independent experiments, each circle represents one mouse, n=5-6, horizontal lines represent means.

Figure 2.. Activated B and GC cell differentiation over time after immunization.

(A) Schematic representation of the experiment: Aicda^GEP/+ Vav^Tg Col1a^mCherry/+ mice were immunized with TM4-Core (footpad injection) on day 0. Mice were treated with doxycycline (Dox) by intraperitoneal injection (2 mg) and by supplementing Dox in drinking water (1 g/l) for one or two weeks at the times depicted in (A). Analysis was performed 1, 2, 3, 4, and 6 weeks after immunization. (B) Graph showing the absolute numbers of GC and activated B cells at indicated timepoints. 2 independent experiments, each circle represents one mouse, n= 7, one-way ANOVA, *p ≤ 0.05, ***p ≤ 0.001, ****p ≤ 0.0001. (C) Schematic representation of the experiment: Aicda^CreERT2/+ R26^eYFP Vav^Tg Col1a^mCherry/+ mice were immunized with TM4-Core (footpad injection) on day 0. Mice were treated with Dox as above, for time indicated in diagram. Tamoxifen (TAM) was administered by oral gavage on day 0, 7, 12, or 19, as depicted. Analysis was performed 9, 14, or 21 days after immunization. (D) Graph showing the percentage of mCherry^hi activated B cells on day 9, 14 and 21 after TM4-Core immunization following a long (0→9, 7→14, and 12→21) or a short (7→9, 12→14, and 19→21) interval after tamoxifen administration. 4 independent experiments, each circle represents one mouse, n=10-17, one-way ANOVA, **p ≤ 0.01, ****p ≤ 0.0001. (E-G) C57BL/6J mice were immunized with TM4-Core (footpad injection) on day 0. (E) Representative flow cytometry profiles and graph showing the percentage of antigen binding TM4-Core⁺ GC cells on day 12, 27, and 42 after immunization. 6 independent experiments, each circle represents one mouse, n = 18, one-way ANOVA, ***p ≤ 0.001. (G) Graph showing the number of somatic mutations (nucleotides, VH + VL chains) in the antibodies obtained from TM4-Core⁺ GC cells (black) and TM4-Core⁻ GC cells (white) purified 12, 27, and 42 after TM4-Core immunization. 3 independent experiments, each circle represents one antibody gene sequenced, n=85-232, two-way ANOVA, **p ≤ 0.01, ****p ≤ 0.0001. Horizontal lines and height of bars indicate means.

Figure 3.. Antigen-binding and non-binding GC B cells

(A) Schematic representation of the experiment: S1pr2^CreERT2/+ R26^tdT/+ mice were immunized with TM4-Core (footpad injection) on day 0, treated with tamoxifen on day 5, and popliteal lymph node GC B cells were analyzed on days 12, 27, and 42. (B-C) Representative flow cytometry profiles and graph summarizing the percentage of tdT⁺ GC B cells on days 12, 27, and 42 after TM4-Core immunization. 6 independent experiments, each circle represents one mouse, n = 17- 20, one-way ANOVA, **p ≤ 0.01, ****p ≤ 0.0001. (D) Graph showing the number of somatic mutations (nucleotides, VH + VL chains) in the antibodies obtained from tdT⁺ GC cells (red) and tdT⁻ GC cells (black) purified on day 12, 27, and 42 after immunization. 3 independent experiments, each circle represents one antibody gene, n = 49-215, two-way ANOVA, ****p ≤ 0.0001. (E) Representative flow cytometry profiles and graph summarizing the percentage of TM4-Core⁺ tdT⁺ (red) or TM4-Core⁺ tdT⁻ (black) GC cells on days 12, 27 and 42 after immunization. 6 independent experiments, each circle represents one mouse, n = 17-22, two-way ANOVA, **p ≤ 0.01, ****p ≤ 0.0001. (G) Graph depicting the percentage of TM4-Core⁺ tdT⁺ (red), tdT⁺ TM4-Core⁻ (pink), tdT⁻ TM4-Core⁺ (black) and tdT⁻ TM4-Core⁻ (gray) GC cells on days 12, 27 and 42 after immunization. 6 independent experiments, each bar represents the mean, n = 17-22. (H) Pie charts depicting the distribution of paired antibody sequences (IgH+IgK) for GC B cells sorted on day 12, 27, and 42 after TM4-Core immunization from 3 individual lymph nodes at each time point. The number in the inner circle indicates the number of sequences analyzed. colored slices indicate expanded clones and white slices indicate proportion of singlets. Pie slices are proportional to the size of individual clones. Horizontal bar graph slices depict the distribution of TM4-Core⁺ tdT⁺ (red), TM4-Core⁺ tdT⁻ (black), TM4-Core⁻ tdT⁺ (pink) and TM4-Core⁻ tdT⁻ (gray) clones for each lymph node. (I) Graph showing the number of somatic mutations (nucleotides, VH + VL chains) in the antibodies obtained from TM4-Core⁺ tdT⁺ (red), TM4-Core⁻ tdT⁺ (pink) or TM4-Core⁻ tdT⁻ (gray) clones on days 12, 27, and 42 after immunization. 3 independent experiments, each circle represents one antibody gene, n = 15-228, two-way ANOVA, **p ≤ 0.01, ****p ≤ 0.0001. Horizontal lines indicate means.

Figure 4.. Naïve B cells produce new GC B cells throughout the immune response.

(A) Schematic representation of the experiment: Cd55^CreERT2/+ R26^tdT/+ mice were immunized with TM4-Core (footpad injection) on day 0 and treated with tamoxifen (TAM) on day 11. Draining lymph nodes were analyzed on days 18, 25, and 32 after immunization. (B) Representative flow cytometry profiles showing the percentage of tdT⁺ naïve B cells (left) and tdT⁺ GC B cells (right). (C) Graph showing the percentage of tdT⁺ naïve (white) and GC (black) B cells over time. Green circles show frequency of tdT⁺ GC B cells normalized for the labeling efficiency of the naïve B cell compartment. 3 independent experiments, circle indicates mean, error bars indicate S.D., n = 7-11. (D) Graph showing the proportion of TM4-Core binding cells among tdT⁻ and tdT⁺ GC populations over time as measured by flow cytometry. 3 independent experiments, circle indicates mean, error bars indicate S.D., n = 7-11. (E) Graph showing the number of somatic mutations (nucleotides, VH + VL chains) in the antibodies obtained from tdT⁺ and tdT⁻ GC B cells. Lymph nodes from 6 mice were pooled for sequencing on day 32 after immunization. Each circle represents one paired antibody gene, n = 75-99, Mann-Whitney, ****p ≤ 0.0001. (F) Pie charts depicting clonal diversity in tdT⁺ and tdT⁻ GC B cell populations. Colored slices indicate clones identified by combined IgH and IgL identity. White slices represent singlets. Number in the middle of pie chart represents number of sequences analyzed per indicated population. Lymph nodes from 6 mice were pooled for sequencing on day 32 after immunization.

Figure 5.. Naïve B cells invade established GCs

(A) Schematic representation of the experiment: Aicda^GFP/+ Cd55^CreERT2/+ R26^tdT/+ mice were immunized with HIV-1 TM4-Core (footpad injection) on day 0 and gavaged with tamoxifen on day 11. Draining lymph nodes were analyzed on day 32 after immunization. Draining lymph nodes from Aicda^GFP/+ Cd55^CreERT2/+ R26^tdT/+ mice were dissected into fragments containing single GCs using vibratome sectioning combined with multiphoton microscopy, followed by single-cell sorting and Ig sequencing. (B-C) Representative multiphoton images of draining lymph nodes with magnified insets (C) shown on the right. Red is tdT; green is GFP. Scale bar represents 200 μm (B) and 100 μm (C), respectively. Representative of 2 independent experiments. (D) Flow cytometry profiles of 6 isolated GCs as in (A), showing proportions of tdT⁺ cells among GFP⁺ GC B cells. (E) Graph showing the number of somatic mutations (nucleotides, VH + VL chains) in the antibodies obtained from tdT⁻ (resident) and tdT⁺ (invader) GC B cells. Pool of 6 individual GCs, each circle represents one antibody gene, n = 54-184, Mann-Whitney, ****p ≤ 0.0001, horizontal lines indicate means. (F) Pie charts depicting clonal diversity of the GCs analyzed in (D). Number in center of chart indicates total number of sequences obtained per GC. Green and yellow slices indicate expanded resident and invader clones, respectively. White and gray slices indicate proportion of singlet resident and invader sequences, respectively.

Figure 6.. Affinity of resident and invader GC B Cells

(A-B) Biolayer interferometry (BLI) traces assessing binding of Fabs isolated from resident GC B cells (tdT⁺ GC cells isolated from S1pr2^CreERT2/+ R26^tdT/+ mice, A) and invader GC B cells (tdT⁺ GC cells isolated from CD55^CreERT2/+ R26^tdT/+ mice, B) on day 27 to 32 after immunization. Colored curves show Fabs with binding profile higher than ED38 control Fab and gray curves represent Fabs with affinities similar or lower than ED38 Fab. Left graph: biosensor chips were loaded with TM4-Core immersed in solutions containing resident GC Fabs. Curves in red represent Fabs with biding higher than controls. Numbers show the number of red and gray curves. Middle graphs: resident (A) and invader (B) GC Fabs were immobilized to biosensor chips and immersed in solutions containing TM4-Core trimer. Curves in green represent Fabs with biding higher than controls. Numbers show the number of green and gray curves. Right graphs: resident (A) and invader (B) GC Fabs were immobilized to biosensor chips and immersed in solutions containing TM4-Core tetramerized trimers. Curves in blue represent Fabs with biding higher than controls. Numbers show the number of blue and gray curves. (C) Graphs showing the proportions of resident and invader GC Fabs binding in (A) or (B). (D-E) Biolayer interferometry (BLI) traces as in (A-B) assessing binding of Unmutated Common Ancestor (UCA) Fabs corresponding to resident GC B cells (tdT⁺ GC cells isolated from S1pr2^CreERT2/+ R26^tdT/+ mice, D) and invader GC B cells (tdT⁺ GC cells isolated from CD55^CreERT2/+ R26^tdT/+ mice, E) isolated on day 27 to 32 after immunization. (F) Graphs showing the proportions of UCA resident and UCA invader GC Fabs binding in (D) or (E).

Figure 7.. Persistent Selection for Affinity

(A) Schematic representation of the experiment: S1pr2^CreERT2/+ R26^tdT/+ mice were immunized with TM4-Core (footpad injection) on day 0 and treated with tamoxifen 5 days later. 5 × 10⁶ B1-8^hi CD45.1⁺ congenic B cells were transferred on day 12 intravenously and popliteal lymph node GC B cells were analyzed on day 27. (B-C) Representative flow cytometry profile and graph showing the percentage of Igλ⁺ cells among host CD45.2⁺ Follicular B cells, CD45.2 tdT⁺, and CD45.2 tdT⁻ GC B cells. 3 experiments, each circle represents one mouse, n = 12, one-way ANOVA, **p ≤ 0.01. (D-E) Representative flow cytometry profile and graph showing the percentage of Igλ⁺ cells among transferred B1-8^hi CD45.1⁺ B cells at day 12 and CD45.1⁺ invader GC cells at day 27. 3 experiments, each circle represents one mouse, n = 12, t test, ****p ≤ 0.0001. (F) Graph showing the percentage of invading CD45.1⁺ GC cells at day 27. 3 experiments, each circle represents one mouse, n = 12. (G) Schematic representation of the experiment: C57BL/6J mice received 100 μg each of 3BNC117 and 10-1074 (HIV specific antibodies) or 100 μg each of C135 and C144 (SARS-CoV-2 specific antibodies) and were immunized one day later with TM4-Core (footpad injection). GC formation and affinity were analyzed 13 days after antibody injection. (H-J) Representative flow cytometry profile and graph showing the percentage of GC B cells among total B cells and TM4-Core⁺ cells among GC B cells. Green circles represent samples sequenced in (K and L). 3 experiments, each circle represents one mouse, n = 12, T test, **p ≤ 0.01. (K) Graph showing the number of somatic mutations (nucleotides, VH + VL chains) in the antibodies obtained from TM4-Core⁺ and TM4-Core⁻ GC B cells after injection of high affinity SARS-CoV-2 or HIV-1 antibodies. 3 experiments, each circle represents one paired antibody gene, n = 45-179, two-way ANOVA. (L) Pie charts depicting clonal diversity in TM4-Core⁺ and TM4-Core⁻ GC B cell populations isolated from 10 lymph nodes 13 days after injection of high affinity SARS-CoV-2 or HIV antibodies in 3 experiments. White (TM4-Core⁻) and red (TM4-Core⁺) slices indicate clones identified by shared IgH and IgL identity. Hatched white (TM4-Core⁻) and red (TM4-Core⁺) slices represent singlets. Number in the middle of pie chart represents number of sequences analyzed for each sample. Percentages indicate the proportion of TM4-Core⁺ GC cells. (M) Graph summarizes the relative proportions of clonal (TM4-Core⁻ white and TM4-Core⁺ red) or singles (TM4-Core⁻ hatched white and TM4-Core⁺ hatched red) for all samples pooled from (L). (N) Biolayer interferometry (BLI) traces assessing binding of Fabs isolated from TM4-Core⁻ GC cells isolated from C56BL/6J mice on day 13 after HIV antibody injection. Colored curves show Fabs with binding profiles higher than polyreactive ED38 control Fab and gray curves represent Fabs with affinities similar or lower than ED38. Left graph: biosensor chips were loaded with TM4-Core immersed in solutions containing Fabs. Curves in red represent Fabs binding better than controls. Numbers show the number of red and gray curves. Right graph: GC Fabs were immobilized on biosensor chips and immersed in solutions containing TM4-Core trimer. Curves in green represent Fabs binding better than controls. Numbers of green and gray curves for each respective experimental condition are shown.