key: cord-0283448-jiy7xljp
authors: Lee, Jeong Hyun; Sutton, Henry; Cottrell, Christopher A.; Phung, Ivy; Ozorowski, Gabriel; Sewall, Leigh M.; Nedellec, Rebecca; Nakao, Catherine; Silva, Murillo; Richey, Sara T.; Torres, Jonathan L.; Lee, Wen-Hsin; Georgeson, Erik; Kubitz, Michael; Hodges, Sam; Mullen, Tina-Marie; Adachi, Yumiko; Cirelli, Kimberly M.; Kaur, Amitinder; Allers-Hernandez, Carolina; Fahlberg, Marissa; Grasperge, Brooke F.; Dufour, Jason P.; Schiro, Faith; Aye, Pyone P.; Carnathan, Diane G.; Silvestri, Guido; Shen, Xiaoying; Montefiori, David C.; Veazey, Ronald S.; Ward, Andrew B.; Hangartner, Lars; Burton, Dennis R.; Irvine, Darrell J.; Schief, William R.; Crotty, Shane
title: Long-lasting germinal center responses to a priming immunization with continuous proliferation and somatic mutation
date: 2021-12-21
journal: bioRxiv
DOI: 10.1101/2021.12.20.473537
sha: 0d1e47b09f09d98f4c34a49901ac505335ff0351
doc_id: 283448
cord_uid: jiy7xljp

Germinal centers (GCs) are the engines of antibody evolution. Using HIV Env protein immunogen priming in rhesus monkeys (RM) followed by a long period without further immunization, we demonstrate GC B cells (BGC) lasted at least 6 months (29 weeks), all the while maintaining rapid proliferation. A 186-fold BGC cell increase was present by week 10 compared to a conventional immunization. Single cell transcriptional profiling revealed that both light zone and dark zone GC states were sustained throughout the 6 months. Antibody somatic hypermutation (SHM) of BGC cells continued to accumulate throughout the 29 week priming period, with evidence of selective pressure. Additionally, Env-binding BGC cells were still 49-fold above baseline 29 weeks after immunization, suggesting that they could be active for significantly longer periods of time. High titers of HIV neutralizing antibodies were generated after a single booster immunization. Fully glycosylated HIV trimer protein is a complex antigen, posing significant immunodominance challenges for B cells, among other difficulties. Memory B cells (BMem) generated under these long priming conditions had higher levels of SHM, and both BMem cells and antibodies were more likely to recognize non-immunodominant epitopes. Numerous BGC cell lineage phylogenies spanning the >6-month GC period were identified, demonstrating continuous GC activity and selection for at least 191 days, with no additional antigen exposure. A long prime, adjuvanted, slow delivery (12-day) immunization approach holds promise for difficult vaccine targets, and suggests that patience can have great value for tuning GCs to maximize antibody responses.

the total dose of the MD39 plus SMNP formulation (50 µg protein, 375 µg adjuvant per side) was split 66 between 7 gradually increasing doses, delivered every other day for a total of 12 days (Extended Data 67 Fig. 1a) . Group 3 was designed with an unusual "long prime" period to assess the durability of GCs after 68 a primary immunization. Each animal in the study was immunized bilaterally, thereby doubling the 69 number of lymph nodes (LNs) and GCs that could be tracked over time. GCs were sampled every 2-3 70 weeks by LN fine needle aspiration (FNA) of inguinal LNs (ILNs) (Fig. 1a) . 71 Following the priming immunization, conventional MD39 plus alum-bolus immunized animals 72 (Group 1) exhibited an increase in total BGC cell (CD71 + CD38 -) % at week 3 post-immunization (Fig. 1b) ; 73 the frequency of Env-binding BGC cells (CD71 + CD38 -/Env +/+ ) peaked at week 3 and declined thereafter 74 (Env-binding BGC cells as % of total B cells. Fig. 1c ). Both total and Env-binding BGC cells were 75 substantially increased in MD39 plus SMNP ED-immunized RMs compared to RMs that received 76 conventional protein plus alum bolus immunization (Groups 2 & 3 combined vs. Group 1, Fig. 1b -c, 77 Extended Data Fig. 1b, 2a) . Median peak BGC cell frequencies observed were 24-33% compared to 3.5% 78 (Mann-Whitney P < 0.0001, week 3 Groups 2 & 3 combined vs. Group 1. Fig. 1b) . Median Env-binding 79 BGC cell frequencies were approximately 7.8-times greater at week 3 (Mann-Whitney, P < 0.0025, Groups 80 2 & 3 combined vs Group 1. Fig. 1c) . Strikingly, in contrast to the conventionally primed Group 1, 81 frequencies of Env-binding BGC cells in Groups 2 and 3 continued to increase, resulting in a 186-fold GC 82 difference by week 10 (Mann-Whitney, P < 0.0001, Group 2 & 3 combined vs Group 1. Fig. 1c ) 83 Tracking of the priming immune response continued for Group 3 animals beyond week 10 84 (Group 1 & 2 animals were boosted at week 10, Fig. 1a) . Surprisingly, GC responses were still active at 85 weeks 13, 16, 21, 25, and 29 ( Fig. 1b -d, Extended Data Fig. 2a-b) . The median magnitude of these Env- 86 binding BGC cells at week 29 was still 27-fold higher than the peak Env-binding BGC cells observed after 87 conventional alum immunization, and it was also greater than the post-boost Env-binding BGC cell 88 response to conventional alum immunization (Fig. 1c) . 191 days (27 weeks) after the end of the priming 89 dose (29 weeks from day 0), median Env-binding BGC cell frequencies in ILNs were still ~49-fold higher 90 than baseline (Fig. 1b-d, Extended Data Fig. 2a-b) . Thus, GCs were capable of continuous activity for > 91 191 days with no additional antigen introduced. Total GC-TFH cell frequencies in ILNs changed over the course of the priming period (Extended Data fig.   94 2c), however longitudinal quantitation of Env-specific GC-TFH cells was not possible due to limiting FNA 95 samples. At 6 weeks post-boost, increased Env-specific GC-TFH cell frequencies trended higher in the 96 long prime Group 3 (Extended Data Fig. 2d-e) 16 . Long-lasting prime GCs may contribute to an improved 97 antigen-specific GC-TFH response after the booster immunization.

Enhanced antibody response quality 100 Group 3 RMs Env-binding serum IgG titers remained stable from week 3 to 29 of the priming 101 phase in the absence of a booster immunization (Fig. 2a) . After boosting, Group 2 and 3 animals 102 generated similar peak binding antibody titers (2 to 3 weeks post-boost), but Group 3 animals 103 maintained significantly higher Env-binding IgG titers at week 6 post-boost (Fig. 2b ). The quality of the 104 antibody responses was next evaluated in terms of ability to neutralize the tier-2 autologous BG505 105 pseudovirus. Notably, autologous tier-2 neutralizing antibodies were detectable in all long prime Group head-to-head neutralization assays, limited tier-2 neutralization breadth was observed with serum from 120 an earlier RM study given ED immunization and two booster immunizations with a similar Env trimer 121 (Olio6) and an earlier ISCOMs adjuvant (SMNP without MPLA) 12 (Extended Data Fig. 3d- Fig. 4) . Group 2 and 3 animals generated antibody responses to V5/C3 and V1/V3 129 epitopes associated with autologous BG505 SHIV protection 21 . Antibody responses in conventional 130 bolus plus alum-immunized animals were largely restricted to the Env trimer base (Fig. 2g) . In sum, 131 employing a 12-day ED immunization strategy and vaccine formulation with SMNP was associated with 132 substantially improved epitope breadth and quality of neutralizing antibodies. 

To directly assess functionality of the GCs over these extended time periods, multiple 158 experimental approaches were employed comparing Groups 2 and 3 RMs. We performed BCR 159 sequencing of Env-binding LN FNA-derived BGC cells from nine different timepoints to assess SHM over 160 time, as well as clonal diversity and mutational patterns in clonal lineages ( Fig. 4 and Extended Data Fig.   161 7-10). Env-binding BGC cell heavy chain (HC) nucleotide (NT) mutations increased significantly between 162 week 3 and 10 (Mann-Whitney, P < 0.0001 for both Group 2 & 3, Fig. 4a-b) . Notably, BGC cells continued 163 to accumulate mutations in the absence of another immunization through week 29 in Group 3 RMs, at 164 which point the median number of HC mutations was 17, with the top 25% of BGC cells containing 22 to 165 45 HC mutations ( Fig. 4a-b) . The difference in SHM in the long prime (week 29) versus 10-week prime 166 was highly significant (Mann-Whitney, P < 0.0001, Fig. 4b ; P < 0.0009, Fig. 4c ), and the difference in 167 median mutations between weeks 10 and 29 was nearly as great as the difference between weeks 3 and Here we demonstrate clearly that GCs can last for at least 191 days in the absence of new antigen, using an experimental system taking advantage of protein 220 immunization, a 12-day ED immunization strategy, a robust adjuvant, and the use of Env probes to 221 identify antigen-specific BGC cells. Furthermore, we demonstrate the GCs are remarkably robust and 222 functional for six months. The BGC cells maintain proliferation, SHM, and affinity maturation, and long-223 lasting GCs can produce high autologous tier-2 neutralizing antibody titers, heterologous neutralizing 224 antibody titers, and highly somatically mutated circulating Env-specific BMem cells to non-Env base 225 epitopes.

A 12-to 14-day slow delivery (ED or osmotic pump) immunization regimen can result in BMem cells during the long prime (Fig 4i) , as opposed to a normal 10-week boost that appears to recall 239 more base-specific B cells. Thus, the improved autologous and heterologous neutralization by Group 3 240 animals is likely partly owing to the diversity of B cells recruited and partly due to increased affinity 241 maturation from extensive GC responses. 242 We have shown that GCs can persist for greater than six months in response to a priming 243 immunization, with a number of notable outcomes. These findings indicate that patience can have great 

All statistics were calculated in Prism 9 or R. The statistical tests used are indicated in the 460 respective figure legends and utilize a two-tailed test. All graphs were generated in Prism 9 or R.

Geometric mean and geometric standard deviation (SD) are shown for data plotted on a Log10 axis.

Mean and SD are plotted for data graphed on a linear axis. 4 : Clonal competition and affinity maturation occurs in antigen-specific BGC cells identified in long lasting GCs. a, Number of nucleotide (NT) mutations in the HC (VH + JH) of Env-binding BGC cells after priming, spaghetti plots track mutations per animal. b, Number of pre-and postboost NT mutations in Env-binding BGC cells. c, Comparison of BGC mutations at the last pre-boost time point (pre-boost; week 10 and 29 for Group 2 & 3 respectively) and 3 weeks post-boost (week 13 and 33 for Group 2 & 3 respectively). 2-way ANOVA multiple comparisons test. d, BGC population diversity at post-prime time points (Chao1). e, BGC cell pre-boost population diversity. pre-boost; week 10 and 29 for Group 2 & 3 respectively. f, Frequency of Env-binding BMem cells in blood. 2-Way ANOVA multiple comparisons test. g, Clonal diversity of BMem cells after boosting. h, The number of mutations in week 6 post-boost BMem cells. i, Quantitation of Env-binding BMem cells that do not bind the trimer-base epitope. j, Serum titers of non-base directed Env-binding antibodies detected by ELISA. k, Clonal trees of 2 unique B cell lineages from two different long prime Group 3 animals. The tree on the left is color coded by time points while the tree on the right is color coded by Env binding. Each ring indicates 5 HC mutations from the predicted germline. For all graphs, Mann-Whitney test was used unless otherwise indicated. ns > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Broadly Neutralizing Antibodies to HIV and Their Role in Vaccine 580

Multiple roles for HIV broadly neutralizing 582 antibodies

Helper Cell Biology: A Decade of Discovery and Diseases

Cell Responses: Cell Interaction Dynamics and Decisions

T cell help controls the speed of the cell cycle in germinal center B cells

Clonal selection in the germinal centre by 596 regulated proliferation and hypermutation

Sustained antigen availability during germinal center initiation enhances 598 antibody responses to vaccination

Elicitation of Robust Tier 2 Neutralizing Antibody Responses in Nonhuman 600 Primates by HIV Envelope Trimer Immunization Using Optimized Approaches

Slow Delivery Immunization Enhances HIV Neutralizing Antibody and 603

Germinal Center Responses via Modulation of Immunodominance

Correlates of adjuvanticity: A review on 606 adjuvants in licensed vaccines

HIV Vaccine Design to Target Germline Precursors of Glycan-Dependent 609

A particulate saponin/TLR agonist vaccine adjuvant alters lymph flow and 611 modulates adaptive immunity

Helper Cells in Immunized Nonhuman Primates Using a Live Cell Activation-614 Induced Marker Technique

Murine Antibody Responses to Cleaved Soluble HIV-1 Envelope Trimers Are

Highly Restricted in Specificity

Mapping the immunogenic landscape of near-native HIV-1 envelope trimers 618 in non-human primates

Electron-Microscopy-Based Epitope Mapping Defines Specificities of 620

Polyclonal Antibodies Elicited during HIV-1 BG505 Envelope Trimer Immunization

Tfh cells and HIV bnAbs, an immunodominance 623 model of the HIV neutralizing antibody generation problem

Mapping Neutralizing Antibody Epitope Specificities to an HIV Env Trimer in 625 Immunized and in Infected Rhesus Macaques

Novel specialized cell state and spatial compartments within the germinal 627 center

Single-cell analysis of germinal-center B cells informs on lymphoma cell of 629 origin and outcome

Tunable dynamics of B cell selection in gut germinal centres

CD4 T follicular helper cell dynamics during SIV infection

Temporal dynamics of persistent germinal centers and memory B cell 635 differentiation following respiratory virus infection

SARS-CoV-2 infection generates tissue-localized immunological memory in 637 humans

SARS-CoV-2 mRNA vaccines induce persistent human germinal centre 639 responses

Germinal center responses to SARS-CoV-2 mRNA vaccines in healthy and 641 immunocompromised individuals

Germinal centre-driven maturation of B cell response to SARS-CoV-2 vaccination

Endocytosis and Recycling of Immune Complexes by Follicular Dendritic 645 Cells Enhances B Cell Antigen Binding and Activation

Vaccine genetics of IGHV1-2 VRC01-class broadly neutralizing antibody 648 precursor naïve human B cells. npj Vaccines

Mapping Polyclonal Antibody Responses in Non-human Primates Vaccinated 652 with HIV Env Trimer Subunit Vaccines

Appion: An integrated, database-driven pipeline to facilitate EM image 654 processing

New tools for automated high-resolution cryo-EM structure determination in 656 RELION-3

Rhesus and cynomolgus macaque immunoglobulin heavy-chain 658 genotyping yields comprehensive databases of germline VDJ alleles

Integrated analysis of multimodal single-cell data

PRESTO: A toolkit for processing high-throughput sequencing raw 663 reads of lymphocyte receptor repertoires

Human B cell lineages associated with germinal centers following influenza 665 vaccination are measurably evolving

Sergushichev, A. A. An algorithm for fast preranked gene set enrichment analysis using 668 cumulative statistic calculation

Fig 2: Long priming enhanced antibody quality. a, Env-binding serum IgG titers determined by ELISA. ET: Endpoint titer. b, Env-binding IgG titers following boost. Triangles indicate the boost time point. c, BG505 pseudovirus neutralization titers at 50 % inhibition (ID50)

Heterologous tier-2 virus neutralization titers. ID50 ≤ 30 was considered non-neutralizing (NN). f, Number of tier-2 heterologous viruses neutralized (ID50 > 50) in the 12-virus panel by week 3 post-boost serum. g, EMPEM of polyclonal plasma Fabs post-boost. Group 1 at week 2 and Groups 2 & 3 at week 3 post-boost. The Env trimer is shown in gray. Graphs quantify number of animals recognizing each indicated epitope. The color of the Fabs in the EM map match that of the epitope colors in the bar graph. Fisher's exact test comparing number of epitopes recognized vs