Open Access

Activation-Induced Cytidine Deaminase Impacts the Primary Antibody Repertoire in Naive Mice

Katherine Bao, Juan Zhang, Alexis Scherl, James Ziai, Azi Hadadianpour, Daqi Xu, Christopher Dela Cruz, John Liu, Yuxin Liang, Lucinda Tam, Cesar A. Corzo, Merone Roose-Girma, Soren Warming, Zora Modrusan, Wyne P. Lee, Kam Hon Hoi and Ali A. Zarrin
J Immunol June 15, 2022, 208 (12) 2632-2642; DOI: https://doi.org/10.4049/jimmunol.2101193

Key Points

  • Inducible AID-knockout model shows that AID impacts the primary B cell repertoire in naive mice.

  • Natural Ags and microbiome may potentiate AID footprints.

Abstract

Genetic and environmental cues shape the evolution of the B cell Ig repertoire. Activation-induced cytidine deaminase (AID) is essential to generating Ig diversity through isotype class switching and somatic mutations, which then directly influence clonal selection. Impaired B cell development in AID-knockout mice has made it difficult to study Ig diversification in an aging repertoire. Therefore, in this report, we used a novel inducible AID-knockout mouse model and discovered that deleting AID in adult mice caused spontaneous germinal center formation. Deep sequencing of the IgH repertoire revealed that Ab diversification begins early in life and evolves over time. Our data suggest that activated B cells form germinal centers at steady state and facilitate continuous diversification of the B cell repertoire. In support, we identified shared B cell lineages that were class switched and showed age-dependent rates of mutation. Our data provide novel context to the genesis of the B cell repertoire that may benefit the understanding of autoimmunity and the strength of an immune response to infection.

This article is featured in Top Reads, p.

Introduction

The rearrangement of the V, D, and J region gene segments at the Ig locus is a process known as V(D)J recombination, which generates a repertoire of B cell receptors that can recognize an immeasurable number of Ags, including self-antigens (1). Nearly 90% of this nascent B cell repertoire is deleted or anergized to remove autoreactive B cells during central and peripheral tolerance (2, 3). Those that persist mature into naive IgM+ B cells [estimated at 1011 cells in human (4) and 108 in mouse (5)] and continue to diversify in response to Ag. As expected, each individual has a B cell repertoire indicative of their unique historical Ag exposure (infections, immunizations, or diseases) (6). However, considerable similarities exist across donor repertoires as well, which may be explained by exposures to common Ags (7, 8). For instance, colonization of mucosal membranes by microbes shortly after birth immediately shapes the host immune response (9). Mice housed in a germ-free environment are more susceptible to pathogenic infections than mice colonized with specific pathogen-free microbes, in part due to the lack of homeostatic antimicrobial Igs that buffer the host and microbiome interface (10). Dietary and environmental Ags similarly induce immune priming (11). Such innocuous factors or “natural Ags” collectively shape the “preimmune” B cell repertoire and can prime the Ab response to pathogenic diseases.

B cells activated by Ag undergo isotype class switching (to IgG, IgA, or IgE) and affinity maturation to gain specialized effector functions (12). Through this process, B cells produce neutralizing Abs to provide long-term humoral immunity to foreign pathogens (13). Ags introduced by infection or immunization are trafficked to the draining lymphoid organs to induce Ag-specific germinal center (GC) reactions. These GCs are composed of B cells and T follicular helper (Tfh) cells that drive B cell diversification and selection (14). B cell diversity is regulated by activation-induced cytidine deaminase (AID) encoded by the gene Aicda, which induces both class-switch recombination (CSR) and somatic hypermutation (SHM) (15, 16). Meanwhile, selection is mediated by competition for Tfh cells that provide “help” in the form of costimulation and survival factors to B cells with higher affinity, and, consequently, B cells with lower affinity are deleted by neglect (17). The surviving B cells enter iterative rounds of selection and diversification, leading to their differentiation into memory B cells and plasma cells that provide potent isotype-switched Ab-mediated recall immunity (17).

Although less is known about their origin, isotype-switched B cells are also present in the absence of pathogenic infection and immunization (11, 18). Abs against commensal Ags are produced in the mesenteric lymph nodes (MLNs) and Peyer’s patches (PPs) of healthy animals due to persistent GC activation by gut microbes (17, 19). Such Abs are often somatically mutated IgA and provide protection at mucosal barriers (19). Natural Abs found in naive animals providing frontline immunity against infections mainly originate from B1 cells (20).

Molecular mimicry between dysbiotic microbial communities and self-antigens as well as epitope spreading contribute to autoimmunity (6, 21). Accordingly, antibiotic treatments that reduce pathobionts have been shown to lower autoantibody production (21) and repertoire analyses suggest that microbes may activate B cells derived from common ancestors predisposed for autoreactivity, leading to autoreactive B cell lineages in the matured repertoire (6). Additional links to autoreactivity also exist among dietary Ags, such as egg, peanut, and shellfish, as well as environmental Ags, such as seasonal pollen, which can trigger Ab-mediated allergic responses (22). Capturing the initiating events that lead to such pathologies has been difficult due to the ongoing nature of natural Ag exposure. Nevertheless, understanding the coevolution of the immune repertoire and its environment can offer meaningful insight into the role of B cells in homeostasis and immunogenicity (7, 18).

In this report, we show that deletion of AID in mature B cells caused spontaneous GC activity and halted repertoire diversification for activated lineages in the B cell repertoire. Our study has elucidated a novel mechanism of GC-mediated B cell diversification at steady state that emphasizes the critical role of AID activity in the development of the B cell repertoire over time.

Materials and Methods

Mice

All mice used are of the C57BL/6N background. Aicdako/ko mice were previously generated (23). Aicdafl/fl mice were generated using the same genetic site-directed strategies as those used to generate Aicdako/ko mice (detailed below). Aicdawt/wt and Aicdafl/fl mice were crossed onto a Rosa26.CreERT2 background (24). Mice were maintained under specific pathogen-free conditions at Genentech. Experimental protocols were approved by Institutional Animal Care and Use Committee of Genentech Lab Animal Research and performed in an American Association for Accreditation of Laboratory Animal Care international–accredited facility in accordance with the Guide for the Care and Use of Laboratory Animals.

Generation of Aicdafl/fl;Rosa26CreERT2 mice

Homologous recombination and mouse embryonic stem (ES) cell technology was used to generate the Aicdafl/fl mouse strain. A gene-targeting vector was constructed with a 2,386-bp arm of 5' homology corresponding to GRCm38/mm10 chr6: 122,558,398–122,560,783 and a 5,496-bp arm of 3' homology corresponding to chr6:122,561,657–122,567,152. The 873-bp region flanked by loxP sites (exon 3) corresponds to chr6: 122,560,784–122,561,656. The final vector (confirmed by DNA sequencing) was linearized to target C2 (C57BL/6N) ES cells using standard methods (G418-positive and ganciclovir-negative selection). C57BL/6N C2 ES cells were electroporated with 20 µg of linearized targeting vector DNA and cultured under drug selection. Positive clones were identified using long-range PCR followed by sequence confirmation. Correctly targeted ES cells were subjected to karyotyping. Euploid gene-targeted ES cell clones were treated with Adeno-FLP to remove Pgk1-neomycin, and ES cell clones were tested to identify clones that lacked the neomycin cassette and the sequence of the targeted allele was verified. Presence of the Y chromosome was confirmed before microinjection into albino C57BL/6N embryos. Germline transmission was obtained after crossing resulting chimeras with C57BL/6N females. Genomic DNA obtained from pups born was screened by long-range PCR to verify the desired gene targeted structure before the mouse colony was expanded. Heterozygous progeny carrying the final flox allele was then crossed onto the Rosa26.CreERT2 background. Genotyping was carried out using 0.3 µM of each primer, Aicda1, Aicda2, and Aicda3 (Supplemental Table I), in a 20-µl PCR reaction as follows: 1 cycle of 94°C (4 min), 30 cycles of 94°C (1 min), 60°C (30 s), and 72°C (1 min), and 1 cycle of 72°C (10 min). Alleles were identified by amplicon size: wild-type, 372 bp; knockout, 509 bp; and loxP, 406 bp.

Anti-CD40L treatment

On days 1, 3, and 5, mice were treated i.p. with 25 µg of Armenian hamster IgG isotype control or anti-CD40L (MR-1; BioXCell). Mice were sacrificed on day 6, and spleens (SP) were harvested.

Antibiotics treatment

Mice were treated by mouth with 200 µl drinking water or antibiotic mixture daily for 14 consecutive days, adapted from a previous publication (25). The antibiotic mixture was made fresh every 2 d and consisted of 6 mg/ml ampicillin, 3 mg/ml vancomycin, 6 mg/ml metronidazole, and 6 mg/ml neomycin in drinking water provided by Lab Animal Resources at Genentech.

Tamoxifen treatment

Mice were weighed prior to tamoxifen treatments and treated by mouth with 100 µl 200 mg/kg tamoxifen (Sigma-Aldrich) resuspended in 5% EtOH and 95% sunflower oil three times per week (Monday, Wednesday, and Friday) for 2 wk. Diet was supplemented with DietGel and HydroGel (ClearH2O) during and up to 1 wk after treatment.

Histopathology

SP were fixed in 10% neutral buffered formalin for 24 h, then bisected longitudinally, and paraffin-embedded. The 4-μm sections were cut onto SuperFrost Plus glass slides. Sections were deparaffinized and rehydrated, and H&E staining was performed according to a routine standard operating procedure using a Leica Autostainer XL (Leica Microsystems, Buffalo Grove, IL). Hematoxylin (American Mastertech, Lodi, CA) was applied for 8 min followed by a 30-s differentiation in 0.5% acid alcohol and bluing agent (Richard-Allan Scientific, Kalamazoo, MI) for 1 min, after which eosin (Eosin Y; American Mastertech) was applied for 30 s. The number and size of GCs from H&E-stained sections per mouse were evaluated by two pathologists independently, blinded to treatment group.

Flow cytometry

Single-cell suspensions were prepared from the MLNs, SP, and PPs. MLNs and SP were directly homogenized; PPs were digested with 2 mg/ml Collagenase D (Roche) for 30 min with rotation. SP was treated with ACK lysis buffer to lyse RBCs. Surface staining was performed with the following: Abs against B220 (RA3-6B2), CD4 (GK1.5), CD45 (30-F11), and GL7 (GL7) were from BioLegend; IgG (Fab2 MinX) from Bethyl Laboratories; CD95 (Jo2) from BD Biosciences; IgA (polyclonal) from SouthernBiotech; and GL7, IgM (Il/41), and LIVE/DEAD Viability marker from Thermo Fisher Scientific. Live lymphocytes were gated by viability dye exclusion, size and granularity based on forward and side scatter, and singlet gates and B cells were further gated on CD45+CD4−B220+ cells. Cell numbers were determined with CountBright beads (Invitrogen), according to the manufacturer’s guidelines. Data were collected on an FACSymphony (BD Biosciences) and analyzed with FlowJo software (Tree Star). For sorting: B cells were enriched using the MACS B cell Isolation Kit (Miltenyi Biotec) according to the manufacturer’s guidelines. Up to 750,000 cells were sorted of either B220+CD95+GL7+ GC B cells or B220+CD95−GL7− non-GC B cells. All other experiments used up to ∼106 cells in bulk from each indicated organ.

Preparation and sequencing of VH gene amplicons

Cells were collected as indicated, pelleted, and resuspended in lysis buffer to extract RNA according to the manufacturer’s protocol in the RNeasy Mini Kit (Qiagen). mRNA was converted to cDNA using unique molecular identifier (UMI)–tagged, degenerate IGHV- (that can bind to either the Cγ or Cα region of the H-chain at the Ig locus) or IgM-specific primers according to the manufacturer’s protocol in the Superscript IV kit (Life Technologies). Briefly, a 20-µl reverse-transcriptase reaction was performed using 2 µM primer, 0.5 mM 2'-deoxynucleoside 5'-triphosphate, up to 200 ng of RNA, and the rest of the recommended buffer components. The primer, 2'-deoxynucleoside 5'-triphosphate, and RNA were first incubated at 65°C for 5 min and then on ice for 1 min. The remaining reaction ingredients were added, and the following reverse-transcription reaction was performed: 50°C (60 min), 95°C (10 min), and 4°C (∞). IgH genes were next amplified in a 50-µl reaction: 2 µl cDNA from the previous reaction, 0.3 µM IGHV_F primers (pooled), 0.3 µM Universal_R primer, and 25 µl PrimeSTAR Master Mix (Takara Bio) in a touchdown PCR as follows: 1 cycle of 98°C (5 min), 4 cycles of 98°C (10 s); 50°C (10 s); 72°C (10 s), 4 cycles of 98°C (10 s); 55°C (10 s); 72°C (10 s), 25 cycles of 98°C (10 s); 58°C (10 s); 72°C (10 s), and 1 cycle of 72°C (5 min). Primer sequences are provided in Supplemental Table I. Amplicons were run on a 1.5% agarose gel, then extracted, and libraries were sequenced on the Illumina MiSeq platform.

Sequence analysis

Paired-end sequencing reads (2 × 300 bp) were first merged using FLASH to reconstruct the full-length variable domain (26). UMIs were extracted from merged sequences to tag and group those derived from the same cDNA template. Only groups with three or more sequences were used for consensus error-correction to recover original cDNA sequences. Only productive full-length sequences were retained and subjected to framework and CDR determination using position-weighted motif and gene-segment germline assignment using IgBlast (27) and IMGT (28). Germline sequences, region annotations, and read counts were tabulated and analyzed using pandas in Python (29), where each row in the data table represents a unique full-length amino acid VH sequence from a given sample within a given campaign. Isotype was determined using the C-region sequence captured on each cDNA. Mutation count was based on AID hotspot (AIDHS; WRC or WGCW motifs) mutation, and only the unique amino acid sequences were considered instead of multiplying the mutation by read counts. Lineages were defined as sequences using the same VJ recombination with ≤80% CDR3 amino acid identity, accounting for somatically mutated progeny (or clones) from common ancestors. Clones were defined as sequences using the same VJ recombination with 100% CDR3 amino acid identity. Unique amino acid sequences were used to distinguish sequence variants. Raw sequencing counts were determined by the number of UMIs associated with each unique amino acid sequence. When comparing between different cohorts where appropriate—for example, the VJ recombination—the count numbers were normalized on a per-animal basis to avoid biases by the high sequence read yielding animal. The pyCircos package (https://github.com/ponnhide/pyCircos) was used to visualize VJ recombination in Circos plots.

Data availability

The data sets generated for the current study are available in the GitHub repository (https://github.com/kamhonhoi-gne/AIDKO-MS.git). Raw sequencing data of the VH gene amplicons are deposited at National Institutes of Health’s Sequence Read Archive under the accession number PRJNA812943 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA812943). Mice generated by Genentech are available through a material transfer agreement.

Results

Temporal deletion of AID in adult mice promotes spontaneous GC activity and interrupts the evolution of B cell diversity

Spontaneous GCs develop when routine GC function is disrupted and the GC undergoes expansion without resolution. Even in the absence of overt infection or immunization, spontaneous GCs have been observed as a clinical autoimmune pathology and in preclinical animal models of disease (30), suggesting that such GCs may be initiated by commonly innocuous Ags. Spontaneous GCs have been identified in AID-deficient mice, representing an opportune model to study B cell activation at steady state (15, 31, 32). However, because low levels of AID expression have been detected in nascent B cells, it was suggested that AID deficiency may impair central tolerance mechanisms and allow self-reactive B cells to escape to the periphery (33, 34). Although this escape may address the rise of clones relevant to autoimmunity, in this report, we wanted to elucidate only GC activity stimulated by natural Ags in healthy animals. To do so, we created a novel Aicdafl/fl mouse line bred onto the Rosa26.CreERT2 background. Tamoxifen treatment of Aicdafl/fl;Rosa26.CreERT2 mice causes the systemic deletion of AID; because AID expression is restricted to B cells, deletion is intrinsically consequential only to B cells. Therefore, the Aicdafl/fl;Rosa26.CreERT2 mice offer a novel advantage over its germline AID-deficient counterpart by enabling the temporal deletion of AID in mature B cells in adult mice.

To understand how natural Ags influence B cell activation, we began by characterizing GCs in naive mice. Mice were housed in a conventional specific pathogen-free facility with controlled exposure to environmental and dietary Ags. Eight-week-old Aicdawt/wt;Rosa26.CreERT2 and Aicdafl/fl;Rosa26.CreERT2 mice were optimized and treated with tamoxifen and aged to elucidate the long-term effects of AID deletion, avoiding undesired effects of tamoxifen (Fig. 1A). In vivo tamoxifen-induced excision was validated by PCR analysis from genomic DNA derived from the SP, MLNs, and PPs (Supplemental Fig. 1A, 1B). Because we used common primers that span the wild-type and deleted Aicda allele, we could quantify the proportion of deletion. Our analyses showed up to 89% deletion in lymphoid tissues. Correspondingly, deletion was also validated functionally by ex vivo CSR assays using splenocytes that confirmed up to a 90% defect in class switching to IgG3 (Supplemental Fig. 1C). Although it was possible that residual AID expression and activity persisted post–tamoxifen treatment, the profound effect on AID deletion was evident. On day 0 (D0), prior to tamoxifen treatment, we found similar frequencies and numbers of B220+ B cells across Aicdawt/wt;Rosa26.CreERT2 and Aicdafl/fl;Rosa26.CreERT2 mice in the MLNs, SP, and PPs (Supplemental Fig. 1D). At this time point, both groups also displayed comparable levels of GC B cells (Fig. 1B, 1C). On D120 after tamoxifen treatment, tissue-specific numbers of B220+ B cells remained similar between Aicdawt/wt;Rosa26.CreERT2 and Aicdafl/fl;Rosa26.CreERT2 mice (Supplemental Fig. 1D). We observed an age-related upward trend of GC B cells in Aicdawt/wt;Rosa26.CreERT2 mice at D120 (Fig. 1B, 1C), collectively hereafter referred to as “6mo-Aicda+” mice (Fig. 1A). Importantly, the increase in GC B cells at D120 in Aicdafl/fl;Rosa26.CreERT2 mice (Fig. 1B, 1C), hereafter referred to as “6mo-AicdaiKO mice” (Fig. 1A), further exceeded the δ of that observed in their wild-type counterparts (6mo-Aicda+ mice). Histologic analyses of the gut lymphoid tissues showed an increase in the size of lymphoid follicles accompanied by more prominent marginal zones in MLNs, PPs, and submucosa colonic lymphoid tissues following temporal AID deletion (Supplemental Fig. 2A, 2B). Necropsy was performed to assess additional immune pathologies induced by temporal AID deletion and revealed no other noticeable abnormalities in the Aicdafl/fl;Rosa26.CreERT2 mice (Supplemental Fig. 2A, 2B). Therefore, our data demonstrate that AID deletion in adult mice primarily caused increased GC formation in secondary lymphoid organs.

FIGURE 1.
FIGURE 1.

Temporal AID deletion causes increased GC activity in naive mice. Aicdawt/wt;Rosa26.CreERT2 and Aicdafl/fl;Rosa26.CreERT2 mice were treated with tamoxifen and aged for 120 days. (A) Experimental design and status of AID deletion. (B) Flow plots were gated through live, single cells that were CD45+CD4−B220+ B cells and show proportions of CD95+GL7+ GC B cells in the MLNs, SP, and PPs on D0 or D120 before or after tamoxifen treatment, respectively. (C) Quantification of GC B cell proportions (top) or cell numbers (bottom) on D0 or D120 after tamoxifen treatment in the MLNs, SP, and PPs. (D) Flow plots were gated on live, single cells that were CD45+CD4−B220+ GCs B cells [as in (B)] found in Aicdafl/fl;Rosa26.CreERT2 mice on D120 after tamoxifen treatment to show proportions of IgG+ GC B cells in the MLNs, SP, and PPs. (E) Quantification of IgG+ GC B cells found in Aicdawt/wt;Rosa26.CreERT2 and Aicdafl/fl;Rosa26.CreERT2 in (D). (F) Flow plots were gated on live, single cells that were CD45+CD4−B220+ GCs B cells [as in (B)] found in Aicdafl/fl;Rosa26.CreERT2 mice on D120 after tamoxifen treatment to show proportions of IgA+ GC B cells in the MLNs, SP, and PPs. (G) Quantification of IgA+ GC B cells found in Aicdawt/wt;Rosa26.CreERT2 and Aicdafl/fl;Rosa26.CreERT2 in (F). Aicdawt/wt;Rosa26.CreERT2 (filled circles, pink) and Aicdafl/fl;Rosa26.CreERT2 (open circles, orange). Data represent 1 experiment with n = 5–10 mice/group (outliers excluded from graphs); error bars represent SEM. *p ≤ 0.05, two-tailed t test.

After tamoxifen treatment, both IgM+ (Supplemental Fig. 2C, 2D) and class-switched GC B cells were present in the MLNs, SP, and PPs in 6mo-AicdaiKO mice (Fig. 1D–G). Notably, there were substantially greater proportions and numbers of IgM+ GC B cells in 6mo-AicdaiKO mice compared with 6mo-Aicda+ mice (Supplemental Fig. 2C, 2D); therefore, temporal AID deletion caused a clear GC B cell effect. Interestingly, we found class-switched GC B cells in both 6mo-Aicda+ and 6mo-AicdaiKO mice. Although we expected that AID deletion would inhibit class switching, we were intrigued to find IgG+ GC B cells were increased in the MLNs and PPs and decreased in the SP (Fig. 1E). Similarly, IgA+ GC B cells were also found in both groups (Fig. 1F); the number of IgA+ GC B cells trended higher in the MLNs, SP, and PPs of 6mo-AicdaiKO mice compared with those of 6mo-Aicda+ mice (Fig. 1G). Together, our data show that pre– and post–class-switched GC B cells were found at 120 d after tamoxifen treatment in both 6mo-Aicda+ and 6mo-AicdaiKO mice. More intriguing to our study, however, was that IgG+ and IgA+ GC B cells were present 120 d after AID deletion, often at elevated levels. Because AID deletion renders B cells incapable of isotype class switching, we hypothesized that such class-switched GC B cells existed at the time of tamoxifen treatment (D0) and were maintained until D120. Taking these findings into consideration, we consolidated our subsequent analyses to class-switched B cells to study only B cells in which AID deletion occurred after maturity.

AID deletion in mature B cells demonstrates that the B cell repertoire is continuously diversified over time

We next wanted to understand how inhibiting diversification by deleting AID in mice at 8 wk of age would impact the aging B cell repertoire in naive animals. We compared three B cell repertoires from the study in the previous section (Fig. 1A). Specifically, we obtained repertoires from: 1) Aicdafl/fl;Rosa26.CreERT2 mice (five mice, prior to tamoxifen, D0); 2) Aicdawt/wt;Rosa26.CreERT2 mice (six mice, treated with tamoxifen, D120); and 3) Aicdafl/fl;Rosa26.CreERT2 mice (nine mice, treated with tamoxifen, D120) and will be, respectively, referred to as: 1) “2mo-Aicda+”; 2) “6mo-Aicda+”; and 3) “6mo-AicdaiKO” mice (Fig. 1A). Individual next-generation sequencing (NGS) libraries were prepared from ∼106 cells isolated in bulk from the MLNs, SP, and PPs from each mouse. Libraries were prepared from bulk tissue homogenate with the intention of reducing bias for or against GC B cells because this population was limited (and sometimes nonexistent, depending on the tissue of interest).

The consolidated B cell repertoire from all tissues isolated from 2mo-Aicda+ mice was comprised of ∼87% IgM+ and 13% class-switched Ig (swIg) (Fig. 2A, top). In 6mo-Aicda+ mice, the distribution of swIg shifted dramatically in 6mo-Aicda+ mice to ∼88% showing evidence of B cell activation (Fig. 2A, middle). Interestingly, in 6mo-AicdaiKO mice, the distribution reverted back to be predominantly IgM with ∼76% IgM and 14% swIg (Fig. 2a, bottom). Together, our data demonstrate that tamoxifen-mediated AID deletion significantly impacted the proportionality of IgM to swIg when compared with age-matched tamoxifen-treated Aicda+ mice. We next examined the diversity of unique VJ recombinations found within the swIg repertoires of 2mo-Aicda+ or 6mo-Aicda+ as a measure of selection and expansion (1, 3). Across 2mo-Aicda+ mice we found 10 VJ recombinations that each represented >2.5% of the entire swIg repertoire, whereas the vast majority of the remaining recombinations was each represented at <1% (Fig. 2B). In contrast, we found greater diversity in the swIg of 6mo-Aicda+ mice demonstrated by a wider distribution of VJ recombinations, whereas only one represented >2.5% of the swIg repertoire (Fig. 2B). Together, our data revealed that the swIg repertoire in 2mo-Aicda+ mice, though limited, showed enrichment for a subset of VJ recombinations (>2.5%), whereas increased VJ diversity in 6mo-Aicda+ mice suggested that older mice experienced more complex Ag exposure over time.

FIGURE 2.
FIGURE 2.

AID-mediated CSR and SHM shape the repertoire of naive mice. NGS was performed on the IGHV repertoire of Aicdafl/fl;Rosa26.CreERT2 (2mo-Aicda+) mice prior to tamoxifen treatment and Aicdawt/wt;Rosa26.CreERT2 (6mo-Aicda+) and Aicdafl/fl;Rosa26.CreERT2 (6mo-AicdaiKO) mice 120 days after tamoxifen treatment. (A) Pie charts depict percentages of each isotype (IgM, IgG, and IgA) across the composite B cell repertoire obtained from the MLNs, SP, and PPs. Each chart was respectively comprised of 19,756 sequence reads in the 2mo-Aicda+ cohort (top), 305,573 sequence reads in the 6mo-Aicda+ cohort (middle), and 597,237 sequence reads in the 6mo-AicdaiKO cohort (bottom). Analyses were confined to swIg (B–F). (B) Manhattan plot shows the representation of unique VJ recombinations (dots) in either the 2mo-Aicda+ (left) or 6mo-Aicda+ (right) data sets. Treadline is set at 2.5% (dotted line). (C) Quantification of AIDHS mutations found across the swIg repertoire in 2mo-Aicda+ (left) and 6mo-Aicda+ (right) mice. Mann–Whitney U test, *p ≤ 0.05. (D) Manhattan plot shows the representation of unique VJ recombinations (dots) from 6mo-AicdaiKO (left) or 6mo-Aicda+ (right) mice. Treadline is set at 2.5% (dotted line). (E) Quantification of AIDHS mutations found across the swIg repertoire in 6mo-AicdaiKO (left) and 6mo-Aicda+ (right) mice. Mann–Whitney U test, *p ≤ 0.05. (F) A list of the top 20 VJ recombinations within the swIg repertoires per data set. Duplicate VJ recombinations across data sets are highlighted in green. NGS data represent 1 experiment with n = 5–10 mice/group.

Concurrent with B cell activation, Igs undergo Ag-specific affinity optimization in the GC through AID-dependent SHM. Specifically, V-genes each contain three CDRs (CDRs), CDR1, CDR2, and CDR3, that are subject to affinity maturation. CDR3 contributes most to binding specificity and is highly enriched for AID binding motifs or AIDHS (35). AID targets AIDHS: WRC(Y)/(R)GYW (W = A/T; R = A/G;Y = C/T) (underlined letters specify deaminated nucleotides) for deamination. Deaminated nucleotides are then repaired by the DNA damage response to generate a point mutation (35). We used AIDHS to quantify the rate of AID-generated mutations and, on average, observed approximately two mutations per H-chain in the swIg repertoire of 2mo-Aicda+ mice (Fig. 2C). The average increased to approximately four mutations in 6mo-Aicda+ mice, although the range was broader (Fig. 2C). Together, we observed higher rates of SHMs among isotype-switched B cells in older mice.

We next wanted to understand how the deletion of AID changed the B cell repertoire in 6-mo-old mice. We showed earlier that spontaneous GCs containing IgG+ and IgA+ B cells emerged in 6mo-AicdaiKO mice (Fig. 1D–G); this finding signified Ag-specific B cell activation and diversification in healthy animals. Therefore, to find patterns in activation-driven expansion, we investigated whether AID deletion caused a differential usage of VJ recombinations. Indeed, the VJ profile looked different from 6mo-Aicda+ mice, as we found seven recombinations in 6mo-AicdaiKO mice that each represented ≥2.5% of the repertoire (Fig. 2D). Furthermore, the average mutation rate in 6mo-AicdaiKO mice was reduced to approximately two to three SHMs (Fig. 2E) comparable to that found in 2mo-Aicda+ mice. Together, AID deletion in 6mo-AicdaiKO mice arrested the diversification of the B cell repertoire and resulted in patterns of clonal enrichment and SHM coinciding with that of 2mo-Aicda+ mice over age-matched and tamoxifen-treated 6mo-Aicda+ mice (Fig. 2B–E). Additionally, we found 5 overlapping swIg VJ recombinations across the top 20 in each data set (Fig. 2F, highlighted in green). These findings support that commonalities exist and drive B cell repertoire development in the absence of exogenous immunological challenge.

Germline AID deficiency reveals a unique repertoire of spontaneous GC B cells that persist after antibiotic treatment

After showing that spontaneous GCs formed following tamoxifen-mediated AID deletion (Fig. 2B), we were curious about their origin. Spontaneous GCs are often correlated with autoimmunity, but the causal mechanisms that lead to GC activity remain incompletely defined (30). Canonically, GCs require Ag to persist (17). For example, microbes in the gut provide constitutive GC stimulation to the MLNs and PPs (19, 36). Therefore, we hypothesized that natural Ags in AID-sufficient animals may have generated low levels of GC activity intensified by AID deficiency to cause spontaneous GC formation (Fig. 1B). Thus, due to the overlap in genetics and biology, we characterized the expansive spontaneous GCs found in germline AID-deficient (AicdaKO) mice. B220+ B cells of AicdaKO mice trended upwards in both the MLNs and PPs but not in the SP when compared with wild-type mice (Supplemental Fig. 3A). Of the B220+ B cell compartment, we found spontaneous GCs in the MLNs and SPs of AicdaKO mice and an upward trend in total GC B cells in the PP (Supplemental Fig. 3B). Together, our findings are supported in the literature showing that AID-knockout mice have spontaneous GC expansion in the absence of exogenous immunological challenge (15, 37).

Next, we treated AicdaKO mice with antibiotics to help us discriminate between Ags that cause spontaneous GC activity (Fig. 3A). Antibiotic treatment targeting both Gram-positive and Gram-negative bacteria had no effect on B220+ B cells in the MLNs or SP. However, it caused reduced numbers found in the PPs of AicdaKO mice, which may reflect the disproportionate bacterial stimuli engaged in the PPs compared with other lymphoid tissues (Fig. 3B). We found reductions in GC B cells across all tissues with antibiotic treatment, although not all trends were statistically significant (Fig. 3C, 3D) (37). To confirm that the spontaneous GC activity in AicdaKO mice was T cell–dependent, AicdaKO mice were treated with anti-CD40L Ab to disrupt T cell–B cell interactions (38), and this treatment effectively decreased both GC B cell (Supplemental Fig. 3C) and Tfh (Supplemental Fig. 3D) compartments compared with its isotype control. Moreover, histopathologic review of H&E-stained slides confirmed a reduction in both GC frequency and size in the SP after treatment with anti-CD40L Ab compared with its isotype control (Supplemental Fig. 3E). Together, these data show that antibiotic treatment of AicdaKO mice caused a global reduction in T-dependent GC activity in the MLNs, SP, and PPs (37). Notably, persistent GC activity after bacterial depletion implicates additional sources of Ag stimulation at steady state.

FIGURE 3.
FIGURE 3.

Spontaneous GCs found in AicdaKO mice are enriched for B cells that respond to natural Ags. Twenty-week-old AicdaKO mice were treated daily with either drinking water (open circles, blue) or an antibiotic cocktail (ampicillin, vancomycin, neomycin, and metronidazole dissolved in drinking water) (open circles, orange) for 14 days. (A) Experimental design. (B) Quantification of total B220+ B cells from the MLNs, SP, and PPs. (C) Flow plots were gated on B220+ to show CD95+GL7+ GCs B cells in the indicated tissues. (D) Quantification of proportion and cell numbers found in flow plots in (C). Flow data are representative of 3 experiments with n = 4–6 mice/group; error bars represent SEM. *p ≤ 0.05, two-tailed t test (B–D). GC B cells were isolated, and IGHV NGS libraries were prepared (E–H). (E) Sequences were pooled across tissues, and the top five shared VJ recombinations were identified per treatment group (water or antibiotics) to compare trends in expansion or contraction as a result of bacterial depletion. Top five shared VJ recombinations analyzed are summarized in legend (top right, those that appeared in both treatment groups are underlined; those that trended up or down from water to antibiotics are highlighted in blue or red, respectively). (F) Analyses performed in (E) separated into subset by tissue. Top five shared VJ recombinations that were found across two or more tissues are boxed. (G and H) Top five shared VJ recombinations were also represented by Circos plots in which each labeled arc segment at the perimeter represents a unique V- or J-gene proportionally scaled to its prevalence in the respective repertoire. The ribbon connecting a V-gene segment with a J-gene segment represents its corresponding VJ recombination. Repertoires of mice treated with water (G) or antibiotics (H) found in the MLNs (left), SP (center), and PPs (right). NGS data represent 1 experiment with n = 4 to 5 mice/group.

To understand how the activated B cell repertoire responded to bacterial depletion, we prepared H-chain sequencing libraries using GC B cells isolated from AicdaKO mice treated with either water or antibiotics (Fig. 3A). B cell diversity was evaluated by the frequency of the “top five” most used shared VJ recombinations (VJ recombinations found in three or fewer mice) from each treatment group assessed for their expansion or contraction as a function of bacterial Ag. Across tissues, eight shared VJ recombinations were identified as the top five across both treatment groups (Fig. 3E). In proportion to the repertoire of total GC B cells, we observed that IGHV1-72:J2 decreased dramatically followed by IGHV1-81:J2, IGHV1-80:J1, and IGHV1-53:J4 in response to antibiotics. Conversely, the proportions of IGHV1-53:J2, IGHV5-6:J2, and IGHV1-4:J3 increased with antibiotic treatment. Overall, we found striking directional trends of VJ usage in the GC B cells of AicdaKO mice treated with antibiotics.

Next, within tissue-specific GC B cell repertoires, we found antibiotic-dependent trends among the respective top five VJ recombinations from each treatment group (Fig. 3F). Alternatively, to give an expanded yet legible view of the VJ-recombination usage, Circos plots were made based on the respective top 10 VJ-recombinations from each treatment group (Fig. 3G, 3H). In the MLNs, nine unique VJ recombinations were identified, and antibiotic treatment resulted in the most dramatic decrease in IGHV1-53:J4 and IGHV1-80:J1 followed by IGHV1-81:J2, IGHV1-55:J4, and IGHV5-17:J4 (Fig. 3F–H, left panel in each); conversely, IGHV1-53:J2 and IGHV1-72:J2 were enriched. In the SP, eight unique VJ-recombinations were identified, and we observed a substantial decrease in IGHV1-81:J2, followed by IGHV-1-80:J1, IGHV1-55:J2, and IGHV5-17:J4 (Fig. 3F–H, center panel in each); in the opposite direction, antibiotic treatment resulted in the greatest increase in IGHV5-17:J3 followed by IGHV5-6:J2. In the PPs, we found six unique VJ recombinations (only three shared VJ recombinations were found in the data set for water treatment) (Fig. 3F–H, right panel in each). As in the MLNs and SP, IGHV1-81:J2 decreased after antibiotic treatment; conversely, treatment with antibiotics caused the greatest increase in IGHV1-53:J2 followed by IGHV1-15:J3 and IGHV1-55:J2. Together, our data revealed tissue-specific patterns in GC B cell responses to antibiotic treatment (Fig. 3D). In particular, IGHV1-81:J2 decreased substantially with antibiotic treatment across all tissues. Also interesting was antibiotic-induced enrichment of IGHV1-53:J2 in the MLNs and PPs. Additionally, IGHV1-55:J2 was highly represented in the both the SP and PPs but differentially decreased in the SP and increased in the PPs after antibiotic treatment. IGHV1-80:J1 also decreased with antibiotics in the MLNs and SP (Fig. 3F), an apparent trend when sequences were pooled across tissues, suggesting that the magnitude of its representation was especially high (Fig. 3E). Overlapping VJ recombinations between treatment groups (underlined in (Fig. 3E, 3F) suggested continuous usage and enrichment of such VJ recombinations irrespective of bacterial stimulation. Together, these data reveal particular patterns in VJ usage at steady state and that the depletion of bacterial Ags in AicdaKO mice led to notable tissue-specific trends in the repertoire of GC B cells.

Common lineages may represent Ag-specific B cell selection, subject to diversity with age

After identifying several VJ recombinations in AicdaKO mice that were responsive to bacterial depletion (IGHV1-53:J2, IGHV1-53:J3, IGHV1-55:J3, IGHV1-72:J2, IGHV1-81:J2, and IGHV5-17:J4) (Fig. 3F), we next characterized their associated B cell lineages. To enrich for those most relevant (and avoid caveats introduced by germline AID deficiency), we selected VJ recombinations that were also found in the swIg repertoires of 2mo-Aicda+, 6mo-Aicda+, and 6mo-AicdaiKO mice (Fig. 2F, green). 2mo-Aicda+ mice were tamoxifen-naive, whereas 6mo-Aicda+ and 6mo-AicdaiKO mice had been exposed to tamoxifen. Therefore, to account for the potential for tamoxifen-induced bias in our analyses, we specifically focused on IGHV1-53:J2, IGHV1-55:J3, IGHV1-72:J2, and IGHV1-81:J2, concentrating on the swIg, which were represented both before and after tamoxifen treatment to study the preactivated repertoire. Combined, each VJ recombination was associated with between 49 and 114 lineages (defined as same VJ recombination and ≤80% CDR3 amino acid identity) and 88 and 169 clones (defined as same VJ recombination and 100% CDR3 amino acid identity). In total, we retrieved 32–123 unique sequences from the 2mo-Aicda+ data set, 2287–9813 from the 6mo-Aicda+ data set, and 936–9581 from the 6mo-AicdaiKO data set (Fig. 4A).

FIGURE 4.
FIGURE 4.

Select lineages activated in naive mice suggest common epitopes elicit B cell responses in the B cell repertoire. The respective swIg repertoires of four highly represented VJ recombinations found across the 2mo-Aicda+ (blue), 6mo-Aicda+ (orange), and 6mo-AicdaiKO (green) data sets were characterized. (A) Table illustrates the numbers of unique lineages, clones, and sequences found within the swIg repertoire for each VJ recombination. (B) CDR3 amino acid (AA) lengths were calculated per unique IGHV sequence and separated into subsets by data set in each violin plot. Median value is shown by a white dot. (C) Manhattan plots depict the representation of unique clones (dots) associated with each VJ recombination across data sets. (D) Quantification of the rate of AIDHS mutations found among the clones associated with each VJ recombination. (E) Same as in (D) separated by isotype: IgM (light blue), IgG (dark blue), and IgA (light green). NGS data represent 1 experiment with n = 5–10 mice/group.

Because increased CDR3 length has been associated with autoimmune pathology (39), we measured that of each respective VJ recombination to ask whether AID affected a trend in selection against this characteristic. We found a constant median CDR3 length of 9–13 aa across 2mo-Aicda+, 6mo-Aicda+, and 6mo-AicdaiKO mice. However, the ranges displayed an upward trend with age, correlated with AID activity in lineages using IGHV1-53:J2, IGHV1-72:J2, and IGHV1-81:J2, although this was not true for IGHV1-55:J3. Notably, we also observed bimodal or differential distributions of CDR3 lengths across repertoires suggestive of age-dependent, selection-based expansion.

We were therefore interested in finding trends of clonal expansion in the swIg repertoires using these VJ recombinations. Similar to the greater data sets (Fig. 2B, 2D), we found enriched clonal expansion per VJ recombination in 2mo-Aicda+ mice; clonal diversity was more distributed in 6mo-Aicda+ mice, but seemed to be reverted by AID deletion in 6mo-AicdaiKO mice (Fig. 4C). This was likely due to impaired mechanisms of selection normally instructed by mutation-driven diversification. Therefore, we next quantified the SHMs and found an average rate of approximately three mutations per IgH in 2mo-Aicda+ mice, which increased to approximately six in 6mo-Aicda+ mice and seemed dependent on AID activity (Fig. 4D). These results support that AID deletion would inhibit further mutational B cell diversification, which would likely have profound effects on B cell selection. As expected, IgM was least mutated when compared with either IgG or IgA, regardless of VJ recombination or age (with the exception of IGHV-1-55-IGHJ3) (Fig. 4E). IgA trended with the greatest number of average mutations among each group of mice: 2mo-Aicda+, 6mo-Aicda+, and 6mo-AicdaiKO mice (Fig. 4E). Overall, we detected AID-dependent diversification among several highly represented VJ recombinations throughout this study, which provide evidence for its importance in shaping the B cell repertoire.

Discussion

The B cell Ab repertoire can potentiate protection during an immune response; however, it can also influence the development of autoimmunity. Unlike T cells, the generation and selection of Ab repertoire are more promiscuous (3). Although the nascent B cell repertoire is determined by genetics and V(D)J recombination, its interaction with the environment can lead to extensive diversification. These changes may begin in the absence of exogenous immunological challenge and evolve over time; the precise mechanisms that underly this development, however, are poorly understood. Specifically, the role of SHM through AID in shaping the baseline Ab repertoire is poorly understood. AID is essential for SHM, making it challenging to examine its role during adulthood. In this study, repertoire analyses using a new in vivo model of inducible AID deletion reveal novel insight into the role of AID in the evolution of the B cell repertoire.

We found that tamoxifen-induced AID deletion in adult mice recapitulates several phenotypes observed in germline AID-deficient animals, including increased IgM-producing cells, increased GCs, and reduced CSR and SHM (15, 16). Others have shown that multiple mechanisms may contribute to increased GCs, including reduced apoptosis (31, 32), dysfunctional Ig production in GC B cells (40), and reduced elimination of B cells via AID-induced locus suicide recombination (41). We observed large numbers of IgM+ GC B cells in 6mo-AicdaiKO mice, which we hypothesize were newly AID-deficient B cell bone marrow emigrants mirroring the tolerance-challenged B cells found in germline AicdaKO mice (33). Although we did not further characterize the mechanisms that led to these phenotypes, our studies revealed a new paradigm for the role of AID in shaping the B cell repertoire. SHM analyses of AID footprints suggested that impaired diversification may directly interfere with GC resolution. The presence of class-switched GC B cells validated that B cell maturation and class switching occurred prior to tamoxifen treatment. Class-switched GC B cells 120 d post–tamoxifen treatment also suggested chronic antigenic stimulation. Our functional ex vivo CSR and genomic PCR analyses revealed significant (up to 95%) but not complete AID deletion in our studies; therefore, escape of rare AID-positive clones should be considered in this model system. We believe, however, that because spontaneous GC activity is not observed in wild-type, AID-sufficient animals, we expect that the spontaneous GCs observed in our models did not result from AID-sufficient B cells. However, we cannot rule out escaping AID-positive clones may also contribute to observed phenotype. Together, our analyses of challenge-naive mice, aged up to 6 mo, suggests that the baseline B cell repertoire evolves to incorporate chronic environmental Ags. This concept was supported by antibiotic treatment in AicdaKO mice, which led to reduced spontaneous GC activity. We hypothesize that low levels of transient GC activity may be overlooked in healthy animals but were captured by impaired GC resolution caused by AID deletion.

Antibiotic treatment of AicdaKO mice significantly reduced the GC B cell compartment, demonstrating the contraction of GC activity after Ag depletion. Furthermore, we found patterns of preferential usage of VJ recombinations within the H-chain repertoire of GC B cells from AicdaKO mice in response to antibiotic treatment. In particular, antibiotics caused a dramatic reduction in IGHV1-81:J2 usage within the MLNs, SP, and PPs, suggesting that a common bacterial Ag was prevalent among these animals. Correspondingly, VJ recombinations that increased in representation following antibiotic treatment, such as IGHV1-53:J2, may be activated by Ags of nonbacterial origin. These trends identified in response to alterations in the microbiome provided rationale to characterize the associated lineages in greater depth. We explored 4 VJ recombinations that both responded to antibiotic treatment in AicdaKO mice and were identified in the top 20 found across the 2mo-Aicda+, 6mo-Aicda+, and 6mo-AicdaiKO data sets. Although CDR3 length did not change significantly as a function of age, we observed a slight trend of longer CDR3 length in older mice that correlated with AID activity. Furthermore, we found that attenuating AID activity caused a broad defect in clonal expansion compared with that observed of aged, AID-sufficient animals. The resulting expansion profile of 6mo-AicdaiKO mice thus resembled that of 2mo-Aicda+ mice. Lastly, we observed that clones associated with these VJ recombinations accrued greater rates of AID-dependent SHM in 6mo-Aicda+ mice. Due to the nature of tamoxifen treatments, some B cell activity found in 6-mo-old mice may have been directed toward tamoxifen in addition to natural environmental Ags; therefore, we focused on lineages activated both before (2 mo) and after (6 mo) tamoxifen, in an effort to exclude tamoxifen-specific clones. Together, we hypothesize that these activated and highly represented VJ recombinations in the repertoire are recurrently engaged by natural Ags causing both AID-mediated CSR and SHM in their evolution. Chronic stimulation via gut Ags such as in the microbiome is only one such source of recurrent B cell engagement that seed chronic PP GC responses (19). Analogous to infection-induced GC reactions, we propose that these activities occur within “spontaneous GCs” of naive animals, which present a novel extension on the origins of the Ab repertoire.

Defining the interaction between the B cell repertoire and its environment can be used to predict the magnitude and quality of an immune response as well as guide therapeutic design. As an example, developing broadly neutralizing Abs against harmful viruses, such as HIV (42) and severe acute respiratory syndrome coronavirus 2 (43), may be heavily influenced by the preactivated status of the B cell repertoire. Our findings provide context and rationale for the presence of mutated and swIg found in healthy animals. A recent study using CD20-cre and AIDCAflox showed that the conditional deletion of AID between challenges of West Nile, Japanese encephalitis, Zika, and dengue viruses did not affect recall responses, suggesting cross-neutralizing Abs are rare and may already exist prior to heterologous infections (44). Intriguingly, the analysis of Ag-specific memory B cells suggests that recall responses are limited by preexisting clonal diversity. Ab repertoire deep sequencing and population genetic signatures before and after influenza vaccine suggests that a broad spectrum of evolutionary processes acts on preexisting clonal lineages (45). In contrast to these Ag-challenged Ab repertoire analyses, our studies elucidate the role of AID in generating B cell diversity in the absence of exogenous immune challenge. Although our data represent an undersampling of its true complexity, improved technologies such as single-cell RNA-sequencing and lineage tracing will reduce such limitations in Ig repertoire analyses in the future. From our data, however, we have extracted several shared lineages that were highly represented across mice; thus, the continued characterization to elucidate their respective epitopes may reveal important factors that drive autoimmunity and offer guidance for vaccine design. Finally, AID inhibition is actively being investigated as potential therapeutic intervention in autoimmune diseases (23, 46), and our model can facilitate preclinical studies to assess the systemic effects of short- and long-term AID deletion in mice.

Disclosures

Authors are or were employees of Genentech. A.A.Z. is currently an employee of TRex Bio.

Acknowledgments

We thank G. Lazar for suggestions and M. Reich and Laboratory Animal Resources, flow, pathology, and sequencing cores for technical assistance.

Footnotes

  • This research was funded by Genentech.

  • K.B., K.H.H., and A.A.Z. designed the study; K.B. performed the experiments with assistance from J. Zhang, A.H., D.X., J.L., C.A.C., W.P.L., and K.H.H.; K.B. and K.H.H. performed bioinformatic analyses; Y.L. and Z.M. performed next-generation sequencing; A.S. and J. Ziai analyzed histological samples; L.T., M.R.-G., and S.W. designed and generated the Aicdafl/fl mouse line; C.D.C. managed the mouse colonies; K.B., K.H.H., and A.A.Z. interpreted the data and wrote the manuscript with input from all authors; all authors read and approved the manuscript; A.A.Z. conceived the research.

  • The online version of this article contains supplemental material.

  • Abbreviations used in this article:

    AID
    activation-induced cytidine deaminase
    AIDHS
    AID hotspot
    CSR
    class-switch recombination
    D
    day
    ES
    embryonic stem
    GC
    germinal center
    MLN
    mesenteric lymph node
    NGS
    next-generation sequencing
    PP
    Peyer’s patch
    SHM
    somatic hypermutation
    SP
    spleen
    swIg
    class-switched Ig
    Tfh
    T follicular helper
    UMI
    unique molecular identifier

  • Received January 4, 2022.
  • Accepted April 7, 2022.

This article is distributed under The American Association of Immunologists, Inc., Reuse Terms and Conditions for Author Choice articles.

References

  1. 1.↵
    2. Jung D.
    3. C.Giallourakis
    4. R.Mostoslavsky
    5. F. W.Alt
    . 2006. Mechanism and control of V(D)J recombination at the immunoglobulin heavy chain locus. Annu. Rev. Immunol. 24: 541–570.
    OpenUrlCrossRefPubMed
  2. 2.↵
    2. Wardemann H.
    3. S.Yurasov
    4. A.Schaefer
    5. J. W.Young
    6. E.Meffre
    7. M. C.Nussenzweig
    . 2003. Predominant autoantibody production by early human B cell precursors. Science 301: 1374–1377.
    OpenUrlAbstract/FREE Full Text
  3. 3.↵
    2. Nemazee D.
    2017. Mechanisms of central tolerance for B cells. Nat. Rev. Immunol. 17: 281–294.
    OpenUrlCrossRefPubMed
  4. 4.↵
    2. Georgiou G.
    3. G. C.Ippolito
    4. J.Beausang
    5. C. E.Busse
    6. H.Wardemann
    7. S. R.Quake
    . 2014. The promise and challenge of high-throughput sequencing of the antibody repertoire. Nat. Biotechnol. 32: 158–168.
    OpenUrlCrossRefPubMed
  5. 5.↵
    2. Collins A. M.
    3. K. J. L.Jackson
    . 2018. On being the right size: antibody repertoire formation in the mouse and human. Immunogenetics 70: 143–158.
    OpenUrlCrossRefPubMed
  6. 6.↵
    2. Bashford-Rogers R. J. M.
    3. L.Bergamaschi
    4. E. F.McKinney
    5. D. C.Pombal
    6. F.Mescia
    7. J. C.Lee
    8. D. C.Thomas
    9. S. M.Flint
    10. P.Kellam
    11. D. R. W.Jayne, et al.
    2019. Analysis of the B cell receptor repertoire in six immune-mediated diseases. Nature 574: 122–126.
    OpenUrlCrossRefPubMed
  7. 7.↵
    2. Briney B.
    3. A.Inderbitzin
    4. C.Joyce
    5. D. R.Burton
    . 2019. Commonality despite exceptional diversity in the baseline human antibody repertoire. Nature 566: 393–397.
    OpenUrlPubMed
  8. 8.↵
    2. Soto C.
    3. R. G.Bombardi
    4. A.Branchizio
    5. N.Kose
    6. P.Matta
    7. A. M.Sevy
    8. R. S.Sinkovits
    9. P.Gilchuk
    10. J. A.Finn
    11. J. E.Crowe Jr.
    . 2019. High frequency of shared clonotypes in human B cell receptor repertoires. Nature 566: 398–402.
    OpenUrlCrossRefPubMed
  9. 9.↵
    2. Gensollen T.
    3. S. S.Iyer
    4. D. L.Kasper
    5. R. S.Blumberg
    . 2016. How colonization by microbiota in early life shapes the immune system. Science 352: 539–544.
    OpenUrlAbstract/FREE Full Text
  10. 10.↵
    2. Macpherson A. J.
    3. B.Yilmaz
    4. J. P.Limenitakis
    5. S. C.Ganal-Vonarburg
    . 2018. IgA function in relation to the intestinal microbiota. Annu. Rev. Immunol. 36: 359–381.
    OpenUrlCrossRefPubMed
  11. 11.↵
    2. Hooijkaas H.
    3. R.Benner
    4. J. R.Pleasants
    5. B. S.Wostmann
    . 1984. Isotypes and specificities of immunoglobulins produced by germ-free mice fed chemically defined ultrafiltered “antigen-free” diet. Eur. J. Immunol. 14: 1127–1130.
    OpenUrlCrossRefPubMed
  12. 12.↵
    2. Hackney J. A.
    3. S.Misaghi
    4. K.Senger
    5. C.Garris
    6. Y.Sun
    7. M. N.Lorenzo
    8. A. A.Zarrin
    . 2009. DNA targets of AID evolutionary link between antibody somatic hypermutation and class switch recombination. Adv. Immunol. 101: 163–189.
    OpenUrlCrossRefPubMed
  13. 13.↵
    2. Bernasconi N. L.
    3. E.Traggiai
    4. A.Lanzavecchia
    . 2002. Maintenance of serological memory by polyclonal activation of human memory B cells. Science 298: 2199–2202.
    OpenUrlAbstract/FREE Full Text
  14. 14.↵
    2. Shlomchik M. J.
    3. W.Luo
    4. F.Weisel
    . 2019. Linking signaling and selection in the germinal center. Immunol. Rev. 288: 49–63.
    OpenUrl
  15. 15.↵
    2. Muramatsu M.
    3. K.Kinoshita
    4. S.Fagarasan
    5. S.Yamada
    6. Y.Shinkai
    7. T.Honjo
    . 2000. Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell 102: 553–563.
    OpenUrlCrossRefPubMed
  16. 16.↵
    2. Revy P.
    3. T.Muto
    4. Y.Levy
    5. F.Geissmann
    6. A.Plebani
    7. O.Sanal
    8. N.Catalan
    9. M.Forveille
    10. R.Dufourcq-Labelouse
    11. A.Gennery, et al.
    2000. Activation-induced cytidine deaminase (AID) deficiency causes the autosomal recessive form of the Hyper-IgM syndrome (HIGM2). Cell 102: 565–575.
    OpenUrlCrossRefPubMed
  17. 17.↵
    2. Mesin L.
    3. J.Ersching
    4. G. D.Victora
    . 2016. Germinal center B cell dynamics. Immunity 45: 471–482.
    OpenUrlPubMed
  18. 18.↵
    2. Tiller T.
    3. M.Tsuiji
    4. S.Yurasov
    5. K.Velinzon
    6. M. C.Nussenzweig
    7. H.Wardemann
    . 2007. Autoreactivity in human IgG+ memory B cells. Immunity 26: 205–213.
    OpenUrlCrossRefPubMed
  19. 19.↵
    2. Chen H.
    3. Y.Zhang
    4. A. Y.Ye
    5. Z.Du
    6. M.Xu
    7. C. S.Lee
    8. J. K.Hwang
    9. N.Kyritsis
    10. Z.Ba
    11. D.Neuberg, et al.
    2020. BCR selection and affinity maturation in Peyer’s patch germinal centres. Nature 582: 421–425.
    OpenUrlCrossRef
  20. 20.↵
    2. Ochsenbein A. F.
    3. R. M.Zinkernagel
    . 2000. Natural antibodies and complement link innate and acquired immunity. Immunol. Today 21: 624–630.
    OpenUrlCrossRefPubMed
  21. 21.↵
    2. Manfredo Vieira S.
    3. M.Hiltensperger
    4. V.Kumar
    5. D.Zegarra-Ruiz
    6. C.Dehner
    7. N.Khan
    8. F. R. C.Costa
    9. E.Tiniakou
    10. T.Greiling
    11. W.Ruff, et al.
    2018. Translocation of a gut pathobiont drives autoimmunity in mice and humans. Science 359: 1156–1161.
    OpenUrlAbstract/FREE Full Text
  22. 22.↵
    2. Sampson H. A.
    2001. Utility of food-specific IgE concentrations in predicting symptomatic food allergy. J. Allergy Clin. Immunol. 107: 891–896.
    OpenUrlCrossRefPubMed
  23. 23.↵
    2. Sun Y.
    3. I.Peng
    4. K.Senger
    5. K.Hamidzadeh
    6. M.Reichelt
    7. M.Baca
    8. R.Yeh
    9. M. N.Lorenzo
    10. A.Sebrell
    11. C.Dela Cruz, et al.
    2013. Critical role of activation induced cytidine deaminase in experimental autoimmune encephalomyelitis. Autoimmunity 46: 157–167.
    OpenUrlCrossRefPubMed
  24. 24.↵
    2. Seibler J.
    3. B.Zevnik
    4. B.Küter-Luks
    5. S.Andreas
    6. H.Kern
    7. T.Hennek
    8. A.Rode
    9. C.Heimann
    10. N.Faust
    11. G.Kauselmann, et al.
    2003. Rapid generation of inducible mouse mutants. Nucleic Acids Res. 31: e12.
    OpenUrlCrossRefPubMed
  25. 25.↵
    2. Hagan T.
    3. M.Cortese
    4. N.Rouphael
    5. C.Boudreau
    6. C.Linde
    7. M. S.Maddur
    8. J.Das
    9. H.Wang
    10. J.Guthmiller
    11. N. Y.Zheng, et al.
    2019. Antibiotics-driven gut microbiome perturbation alters immunity to vaccines in humans. Cell 178: 1313–1328.e13.
    OpenUrlCrossRefPubMed
  26. 26.↵
    2. Magoč T.
    3. S. L.Salzberg
    . 2011. FLASH: fast length adjustment of short reads to improve genome assemblies. Bioinformatics 27: 2957–2963.
    OpenUrlCrossRefPubMed
  27. 27.↵
    2. Reddy S. T.
    3. X.Ge
    4. A. E.Miklos
    5. R. A.Hughes
    6. S. H.Kang
    7. K. H.Hoi
    8. C.Chrysostomou
    9. S. P.Hunicke-Smith
    10. B. L.Iverson
    11. P. W.Tucker, et al.
    2010. Monoclonal antibodies isolated without screening by analyzing the variable-gene repertoire of plasma cells. Nat. Biotechnol. 28: 965–969.
    OpenUrlCrossRefPubMed
  28. 28.↵
    2. Giudicelli V.
    3. P.Duroux
    4. C.Ginestoux
    5. G.Folch
    6. J.Jabado-Michaloud
    7. D.Chaume
    8. M. P.Lefranc
    . 2006. IMGT/LIGM-DB, the IMGT comprehensive database of immunoglobulin and T cell receptor nucleotide sequences. Nucleic Acids Res. 34: D781–D784.
    OpenUrlCrossRefPubMed
  29. 29.↵
    2. McKinney W.
    2010. Data Structures for Statistical Computing in Python. In Proceedings of the 9th Python in Science Conference (SCIPY 2010), June 28–July 3. p. 56–61.
  30. 30.↵
    2. Domeier P. P.
    3. S. L.Schell
    4. Z. S.Rahman
    . 2017. Spontaneous germinal centers and autoimmunity. Autoimmunity 50: 4–18.
    OpenUrlCrossRef
  31. 31.↵
    2. Boulianne B.
    3. O. L.Rojas
    4. D.Haddad
    5. A.Zaheen
    6. A.Kapelnikov
    7. T.Nguyen
    8. C.Li
    9. R.Hakem
    10. J. L.Gommerman
    11. A.Martin
    . 2013. AID and caspase 8 shape the germinal center response through apoptosis. J. Immunol. 191: 5840–5847.
    OpenUrlAbstract/FREE Full Text
  32. 32.↵
    2. Zaheen A.
    3. B.Boulianne
    4. J. Y.Parsa
    5. S.Ramachandran
    6. J. L.Gommerman
    7. A.Martin
    . 2009. AID constrains germinal center size by rendering B cells susceptible to apoptosis. Blood 114: 547–554.
    OpenUrlAbstract/FREE Full Text
  33. 33.↵
    2. Kuraoka M.
    3. T. M.Holl
    4. D.Liao
    5. M.Womble
    6. D. W.Cain
    7. A. E.Reynolds
    8. G.Kelsoe
    . 2011. Activation-induced cytidine deaminase mediates central tolerance in B cells. Proc. Natl. Acad. Sci. USA 108: 11560–11565.
    OpenUrlAbstract/FREE Full Text
  34. 34.↵
    2. Meyers G.
    3. Y. S.Ng
    4. J. M.Bannock
    5. A.Lavoie
    6. J. E.Walter
    7. L. D.Notarangelo
    8. S. S.Kilic
    9. G.Aksu
    10. M.Debré
    11. F.Rieux-Laucat, et al.
    2011. Activation-induced cytidine deaminase (AID) is required for B-cell tolerance in humans. Proc. Natl. Acad. Sci. USA 108: 11554–11559.
    OpenUrlAbstract/FREE Full Text
  35. 35.↵
    2. Rogozin I. B.
    3. N. A.Kolchanov
    . 1992. Somatic hypermutagenesis in immunoglobulin genes. II. Influence of neighbouring base sequences on mutagenesis. Biochim. Biophys. Acta Gene Struct. Expr. 1171: 11–18.
    OpenUrlCrossRefPubMed
  36. 36.↵
    2. Victora G. D.
    3. M. C.Nussenzweig
    . 2012. Germinal centers. Annu. Rev. Immunol. 30: 429–457.
    OpenUrlCrossRefPubMed
  37. 37.↵
    2. Fagarasan S.
    3. M.Muramatsu
    4. K.Suzuki
    5. H.Nagaoka
    6. H.Hiai
    7. T.Honjo
    . 2002. Critical roles of activation-induced cytidine deaminase in the homeostasis of gut flora. Science 298: 1424–1427.
    OpenUrlAbstract/FREE Full Text
  38. 38.↵
    2. Han S.
    3. K.Hathcock
    4. B.Zheng
    5. T. B.Kepler
    6. R.Hodes
    7. G.Kelsoe
    . 1995. Cellular interaction in germinal centers. Roles of CD40 ligand and B7-2 in established germinal centers. J. Immunol. 155: 556–567.
    OpenUrlAbstract/FREE Full Text
  39. 39.↵
    2. Meffre E.
    3. M.Milili
    4. C.Blanco-Betancourt
    5. H.Antunes
    6. M. C.Nussenzweig
    7. C.Schiff
    . 2001. Immunoglobulin heavy chain expression shapes the B cell receptor repertoire in human B cell development. J. Clin. Invest. 108: 879–886.
    OpenUrlCrossRefPubMed
  40. 40.↵
    2. Mayer C. T.
    3. A.Gazumyan
    4. E. E.Kara
    5. A. D.Gitlin
    6. J.Golijanin
    7. C.Viant
    8. J.Pai
    9. T. Y.Oliveira
    10. Q.Wang
    11. A.Escolano, et al.
    2017. The microanatomic segregation of selection by apoptosis in the germinal center. Science 358: eaao2602.
    OpenUrlAbstract/FREE Full Text
  41. 41.↵
    2. Péron S.
    3. B.Laffleur
    4. N.Denis-Lagache
    5. J.Cook-Moreau
    6. A.Tinguely
    7. L.Delpy
    8. Y.Denizot
    9. E.Pinaud
    10. M.Cogné
    . 2012. AID-driven deletion causes immunoglobulin heavy chain locus suicide recombination in B cells. Science 336: 931–934.
    OpenUrlAbstract/FREE Full Text
  42. 42.↵
    2. Escolano A.
    3. J. M.Steichen
    4. P.Dosenovic
    5. D. W.Kulp
    6. J.Golijanin
    7. D.Sok
    8. N. T.Freund
    9. A. D.Gitlin
    10. T.Oliveira
    11. T.Araki, et al.
    2016. Sequential immunization elicits broadly neutralizing anti-HIV-1 antibodies in Ig knockin mice. Cell 166: 1445–1458.e12.
    OpenUrlCrossRefPubMed
  43. 43.↵
    2. Shivarov V.
    3. P. K.Petrov
    4. A. D.Pashov
    . 2020. Potential SARS-CoV-2 preimmune IgM epitopes. Front. Immunol. 11: 932.
    OpenUrl
  44. 44.↵
    2. Wong R.
    3. J. A.Belk
    4. J.Govero
    5. J. L.Uhrlaub
    6. D.Reinartz
    7. H.Zhao
    8. J. M.Errico
    9. L.D’Souza
    10. T. J.Ripperger
    11. J.Nikolich-Zugich, et al.
    2020. Affinity-restricted memory B cells dominate recall responses to heterologous flaviviruses. Immunity 53: 1078–1094.e7.
    OpenUrlCrossRef
  45. 45.↵
    2. Horns F.
    3. C.Vollmers
    4. C. L.Dekker
    5. S. R.Quake
    . 2019. Signatures of selection in the human antibody repertoire: Selective sweeps, competing subclones, and neutral drift. Proc. Natl. Acad. Sci. USA 116: 1261–1266.
    OpenUrlAbstract/FREE Full Text
  46. 46.↵
    2. Galicia G.
    3. B.Boulianne
    4. N.Pikor
    5. A.Martin
    6. J. L.Gommerman
    . 2013. Secondary B cell receptor diversification is necessary for T cell mediated neuro-inflammation during experimental autoimmune encephalomyelitis. PLoS One 8: e61478.
    OpenUrlCrossRefPubMed
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools

Jump to section