key: cord-0915351-8m18e0t6
authors: Yang, Fan; Nielsen, Sandra C. A.; Hoh, Ramona A.; Lee, Ji-Yeun; Pham, Tho D.; Jackson, Katherine J. L.; Roskin, Krishna M.; Liu, Yi; Ohgami, Robert S.; Osborne, Eleanor M.; Niemann, Claus U.; Parsonnet, Julie; Boyd, Scott D.
title: Shared B cell memory to coronaviruses and other pathogens varies in human age groups and tissues
date: 2020-12-02
journal: bioRxiv
DOI: 10.1101/2020.12.01.407015
sha: f759eb2548af9a8cf99d67c18ff4f7c3f3710743
doc_id: 915351
cord_uid: 8m18e0t6

Vaccination and infection promote the formation, tissue distribution, and clonal evolution of B cells encoding humoral immune memory. We evaluated convergent antigen-specific antibody genes of similar sequences shared between individuals in pediatric and adult blood, and deceased organ donor tissues. B cell memory varied for different pathogens. Polysaccharide antigen-specific clones were not exclusive to the spleen. Adults’ convergent clones often express mutated IgM or IgD in blood and are class-switched in lymphoid tissues; in contrast, children have abundant class-switched convergent clones in blood. Consistent with serological reports, pre-pandemic children had class-switched convergent clones to SARS-CoV-2, enriched in cross-reactive clones for seasonal coronaviruses, while adults showed few such clones in blood or lymphoid tissues. These results extend age-related and anatomical mapping of human humoral pathogen-specific immunity. One Sentence Summary Children have elevated frequencies of pathogen-specific class-switched memory B cells, including SARS-CoV-2-binding clones.

Clonal proliferation of lymphocytes is critical for the adaptive immune system to respond to pathogens. The clonal identity of a B cell can be traced by the sequence of its B cell receptor (BCR), which determines its antigen specificity (1-3). Immunoglobulin (Ig) sequences are 5 formed via irreversible V(D)J gene segment rearrangement and can be diversified through somatic hypermutation (SHM) (4) and class-switch recombination (CSR) (5) . Convergent or "public" Ig sequences showing similar antibody gene rearrangements can be found in the primary BCR repertoires of naive B cells, owing to biases in V(D)J gene recombination mechanisms, but are enriched in individuals exposed to the same antigen, reflecting an antigen- 10 driven selection for particular sequence features in the Ig heavy (IGH) or Ig light chain complementarity-determining region (CDR) loops of the BCR (6) (7) (8) (9) (10) (11) (12) . These Ig sequences, when detected in antigen-experienced B cells, document previous infections and allow for the reconstruction of the immunological memory of an individual. Recently, high-throughput sequencing (HTS) has been applied to study convergent IGH repertoires across individuals in 15 response to vaccination and infection (11) (12) (13) (14) . It is still unclear, however, how our immune memories to different antigens distribute in immunological tissues and change during an individual's lifespan.

The immune system gradually matures and accumulates immunological memory from infancy 20 onward (15) . The immune responses to some pathogens appear to differ between young children and adults. For example, young children infected with human immunodeficiency virus (HIV) can achieve neutralizing antibody breadth relatively early after infection, via different mechanisms than adults (16). Data emerging from the current coronavirus disease 2019 pandemic show that children usually have milder symptoms than adults (17) (18) (19) (20) (21) (22) , potentially due to differences of viral receptor expression and other host factors, but also, as a recent serological study suggests, elevated pre-pandemic frequencies of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) binding and neutralizing antibodies in children, stimulated by other coronavirus infections (23, 24) . During the COVID-19 pandemic, infected children 5 compared to adults have lower SARS-CoV-2 antibody titers and a predominance of IgG antibodies specific for the S protein but not the N protein, with these differences attributed to faster clearance of virus limiting the amount of viral antigens produced in the body (25) . The ways in which B cell clonal populations specific for coronaviruses and other pathogens may differ between children and adults is unknown. Most studies of human B cells rely on peripheral 10 blood mononuclear cells (PBMCs), which represent a small part of an individual's full IGH repertoire, with secondary lymphoid tissues in lymph nodes, spleen, and the gastrointestinal tract harboring larger numbers of B cells and being major sites for SHM and CSR (26, 27) .

Specialized immune responses in some lymphoid tissues have been reported; for example, antibody responses to polysaccharide antigens are decreased in the absence of functional splenic 15 tissue (28) (29) (30) (31) (32) (33) . It is still unclear if B cell clones are widely dispersed in different secondary lymphoid tissues or enriched in particular tissues depending on their antigen or pathogen specificity.

To better understand how the antigen-specific B cell memory compartment changes through the 20 human lifespan, and distributes itself across immunological tissues, we systematically characterized the convergent IGH repertoires specific to six common pathogens or vaccine types, and two viruses not encountered by the participants (Ebola virus and SARS-CoV-2) in pre-COVID-19 pandemic individuals from newborns to 87 years of age, and in tissues including cord blood, peripheral blood, spleen, and lymph nodes. For blood B cell analysis, we analyzed 12 cord blood samples, 93 peripheral blood samples from 51 children (34) , 122 healthy human adult peripheral blood samples (34) , and eight blood samples from deceased organ donors (table S1)).

All children were known to be exposed to Haemophilus influenzae type b (Hib), Pneumococcus 5 pneumoniae (PP), tetanus toxoid (TT), and influenza virus (Flu) antigens via vaccination, and it is highly likely that all were exposed to respiratory syncytial virus (RSV) in their first three years of life (35) . The children did not receive Neisseria meningitidis (NM) vaccination. The detailed vaccination histories of the adults are unknown. To identify convergent IGH sequences for the six pathogens in blood samples, we clustered IGH reads with known pathogen-specific reference 10 IGH sequences (table S2) based on their IGHV and IGHJ gene segment usage, CDR-H3 length, and at least 85% CDR-H3 amino acid sequence similarity (see Methods).

We hypothesized that cord blood samples would show the least evidence of convergent pathogen-specific IGH gene sequences, if these IGH were in fact truly stimulated by pathogen or 15 vaccine exposures. Indeed, we found only low frequencies of convergent B cell clones, and only expressed as unmutated IgM or IgD, in the cord blood repertoires (Fig. 1A , age zero years old).

In contrast, convergent clones for Hib, NM, PP, TT, RSV, and Flu in children (one to three years of age) and adults (17-87 years of age) usually expressed mutated IgM or IgD or class-switched isotypes ( Fig. 1A and fig. S1 ). There was a higher frequency of Hib, PP, TT and RSV convergent 20 clones in the blood from children compared to adult blood (Fig. 1B) . The childhood vaccination schedule (table S3) ensured immunization of the children to Hib, PP, and TT with immunizations at approximately 2, 4, 6 months, and again between 12-15 months of life; NM is not given as an early childhood vaccine, and showed low convergent clone frequencies (Fig. 1A) . The timing of vaccination events compared to the detection of convergent IGH did not show significant correlation ( fig. S2-S4) , suggesting that children have persistently elevated frequencies of peripheral blood B cells expressing convergent clones for these pathogens. SHM frequency of each convergent cluster expressing each isotype. Clones with median SHM frequency less than or equal to 1% are colored in pink, and those with median SHM frequency more than 1% are in blue. The SHM frequencies of convergent clones expressing IgG or IgA were lower in children than in adults, p-value = 6.50e-13 and 1.96e-8, respectively, by

Wilcoxon-Mann-Whitney test.

In analyzing antibody isotype expression by convergent clones, we classified clonal lineages into three groups: 1) those containing only unmutated IgM or IgD (unmutM/D), most likely derived and IgD expressing (IgD+) clone members (Fig. 1C) . Both adults and children had a high fraction of CS convergent clones for influenza (Fig. 1A) . 20 Median SHM frequencies for IgM+ or IgD+ convergent clones for each pathogen were similar between adults and children (from 2% to 5%) whereas SHM frequencies of IgG+ or IgA+ convergent clones were significantly lower in children (Fig. 1D ), in agreement with our prior observations of total IGH repertoires for these isotypes (34) . We observed an increase of SHM up to age 87 years in IgG+ convergent clones for Flu, likely due to the recurrent exposures to influenza antigens by infection, or annual influenza vaccination, in contrast to the other pathogens. 5 Potential explanations for the lower frequencies of CS B cell convergent clones in adult compared to pediatric blood could be that the CS clone members gradually die off after childhood exposures, or that they are localized elsewhere in the body in adults, or that the clones specific for individual vaccines or pathogens are diluted by the larger overall pools of memory 10 clones in adults and therefore at too low a frequency to be detected. To test the hypothesis that these cells were in other tissues in adults, we analyzed IGH repertoires in the spleen, mediastinal lymph node (MDLN), and mesenteric lymph node (MSLN) in eight adult deceased organ donors (table S1) . 15 In each deceased organ donor, we compared overall clone distributions within and between secondary lymphoid tissues and to those found in blood. As expected, blood B cells had higher frequencies (mean proportion: 43.8%) of naïve-like clones expressing unmutated IgM or IgD, suggesting that clones that have been stimulated the most by prior antigen exposures become the most widely distributed in the body. 5 The frequency of B cell convergent clones for Hib, NM, PP, TT, RSV, and Flu in lymph nodes and spleen of deceased organ donors was significantly higher than in blood ( Fig. 2A) . Prior clinical reports also suggest that different lymphoid tissues may have specialized functions in protective immunity, such as enrichment of B cells specific for bacterial capsular 10 polysaccharides in splenic tissue, leading to vulnerability to these pathogens in splenectomized patients (28) (29) (30) (31) (32) (33) . We tested whether each tissue site showed tissue-specific convergent IGH to particular pathogens, for example, whether spleen had more clonal lineages specific for polysaccharide antigens. Surprisingly, the frequencies of convergent clones for Hib, NM and PP in the two lymph node sites were similar or even higher than in spleen, indicating that the spleen 15 may be a major reservoir of such clones, but is not the only source of them. Convergent IGH for polysaccharide antigens were usually expressed as IgM or IgD, and less often IgG (Fig. 2B) . S8 ). Although a single example does not permit generalization, we had anticipated that the infection would have stimulated expansion of convergent clones for influenza; it seems possible that the lack of such clones was associated with an ineffective immune response to the infection. We hypothesized that some differences in humoral immune responses in children compared to 20 adults could be related to changes in antigen-specific B cell repertoires with development and aging. To test this, we evaluated whether distinct subsets of the reference IGH sequences specific for each pathogen or vaccine were enriched in particular age groups. In fact, IGH sequences from the two lymph nodes of the deceased organ donors, and the blood from the children were more likely to match the same known pathogen-specific IGH sequences, compared to IGH sequences from adult blood (including both the eight donor blood and the other 114 adult blood samples), indicating that localization of these clones differs between children and adults ( Fig. 2C   and fig. S9 ). Mean CDR-H3 lengths ( fig. S10A ) and IGHV gene usage ( fig. S10 , B and C) of 5 convergent clones were not significantly different between cord blood, pediatric or adult blood, or adult lymphoid tissues.

Prompted by a recent serological report of cross-reactive SARS-CoV-2 binding antibodies in the blood of children prior to the COVID-19 pandemic (24), we next tested if we could detect infants, children and adults in this study had not been exposed to SARS-CoV-2 or Ebola virus (EBOV). We detected rare examples of convergent clones for EBOV in the unmutM/D compartment in both adult and pediatric blood, and in adult lymphoid tissues (Fig. 3, A and B) .

In stark contrast, convergent clones with IGH similar to SARS-CoV-2 binding and neutralizing 20 antibodies (Table 1, table S4 ) were common in the blood of children one to three years old, and in 37 of 51 children showed SHM with or without class-switching, providing evidence of prior antigen exposure (Fig. 3A, C) . Cord blood lacked these mutated and class-switched sequences, further supporting a role for antigen stimulation in their development (Fig. 3A) . Adults had lower frequencies of SARS-CoV-2 convergent clones in blood compared to the children, with some IgM or IgD clones having SHM, but almost no class-switched examples (Fig. 3A) . We tested whether convergent SARS-CoV-2 B cell clones in adults might be localized to lymphoid tissues rather than circulating in the blood, but found similarly low clone frequencies, with comparable 5 SHM frequencies to those in the blood, without class switching, in lymph nodes and spleen (Fig.   3B, C, D) . Notably, about 7% of the SARS-CoV-2 convergent IGHs in children were similar to known antibodies that cross-react with other human coronaviruses (HCoVs), such as HKU1, NL63, and 229E, which are continuously circulating in the population (37) (Fig. 3E) . None of the convergent clones for SARS-CoV-2 in adults were similar to IGH previously reported to bind 10 other HCoVs, with the exception of SARS-CoV. These results suggest that children have higher frequencies than adults of cross-reactive spike-binding memory B cells stimulated by other HCoV exposures, leading to the pre-pandemic cross-reactive SARS-CoV-2 binding and neutralizing antibody titers recently described in children(24). We note that both adults and children had relatively high frequencies of convergent SARS-CoV-2 antibody clones detected in 15 naive B cell-derived unmutated IgM and IgD, indicating that these clonotypes have high generation frequencies in the primary B cell repertoires (Fig. 3A) . The cross-reactive B cell memory and serum antibodies could contribute to the milder illnesses caused by SARS-CoV-2 in children, in combination with other factors such as decreased expression of the angiotensinconverting enzyme 2 (ACE2) viral receptor protein in the airway (39) . Relatively rapid decreases 20 in antibody titers specific for various coronaviruses in adults have been previously reported (40, 41) , suggesting a lack of long-lived plasma cells producing secreted antibodies. Given that the adults in our study were once children, and are likely to have been exposed periodically to community coronaviruses through their lifetimes, another implication of our data is that coronavirus infections may not stimulate long-lasting memory B cell responses, particularly of class-switched B cells. IGHM  IGHD  IGHG  IGHA  IGHM  IGHD  IGHG  IGHA  IGHM  IGHD  IGHG  IGHA  IGHM  IGHD  IGHG responses (42). Our data show that individuals with diverse underlying genotypes can respond to common pathogens and vaccine antigens in a conserved manner (producing some highly similar convergent clones in response to the same antigen), and those convergent clones are found at higher frequency and more often as class-switched isotypes in the blood of children compared to adults. Among adults, these convergent clones are instead enriched in lymph nodes and spleen. B 10 cell memory to antigens that are frequently encountered by infection or vaccination, such as those of influenza viruses, show similar clone frequencies and isotype switching between children and adults, but demonstrate progressive increases in SHM frequencies over the human lifespan. In contrast to ideas of tissue-enriched B cell populations for particular antigens, such as bacterial capsular polysaccharide-specific B cells in the spleen, the B cell clones likely to be specific for these antigens are found at similar frequencies in the spleen and lymph nodes. 5 Our data also highlight important differences in potential cross-reactive B cell responses to novel pathogens in children and adults. Young children and adults both showed very little evidence for IGH specific for EBOV, consistent with their lack of exposure to Ebola or related viruses.

Notably, young children have class-switched and mutated clones with similarity to known SARS-CoV-2 specific IGH, particularly for clones known to be cross-reactive for binding the 

Tracing cells for tracking cell lineage and clonal behavior

B lymphocytes: how they develop and function

Immune Cell Identity: Perspective from a Palimpsest

Mechanism and control of V(D)J recombination at the immunoglobulin heavy chain locus

Mechanism and control of V(D)J recombination versus class switch recombination: similarities and differences

Longitudinal Analysis of the Human B Cell Response to Ebola Virus Infection

Affinity maturation in a human humoral response to influenza hemagglutinin

Multi-Donor Longitudinal Antibody Repertoire Sequencing Reveals the Existence of Public Antibody Clonotypes in HIV-1 Infection

Antibodies to a Conserved Influenza Head Interface Epitope Protect 15 by an IgG Subtype-Dependent Mechanism

Preferential induction of cross-group influenza A hemagglutinin stem-specific memory B cells after H7N9 immunization in humans

Human responses to influenza vaccination show seroconversion 20 signatures and convergent antibody rearrangements

Convergent antibody signatures in human dengue

Restricted, canonical, stereotyped and convergent 25 immunoglobulin responses

Identification of antigen-specific B cell receptor sequences using public repertoire analysis

Evolution of the immune system in humans from infancy to old age

Early development of broadly neutralizing antibodies in HIV-1-infected infants

COVID-19 in children and adolescents in Europe: a multinational, multicentre cohort study

SARS-CoV-2 Infection in Children

Clinical Characteristics of Children with Coronavirus Disease

A Case Series of children with 2019 novel coronavirus infection: clinical and epidemiological features

Children with Covid-19 in Pediatric Emergency Departments in Italy

SARS-COV-2 infection in children and newborns: a systematic review

The immune system of children: the key to understanding SARS-CoV

Preexisting and de novo humoral immunity to SARS-CoV-2 in humans

Distinct antibody responses to SARS-CoV-2 in children and adults across the COVID-19 clinical spectrum

Mutation drift and repertoire shift in the maturation of the immune 5 response

Intraclonal generation of antibody mutants in germinal centres

Clinical implications of the specialised B cell response to polysaccharide encapsulated pathogens

Immaturity of the human splenic marginal zone in infancy. Possible contribution to the deficient infant immune response

Marginal zone B cells: virtues of innate-like antibodyproducing lymphocytes

Marginal zone B cells respond both to polysaccharide antigens and protein antigens

Complement-mediated follicular localization of T-independent type-2 antigens: the role of marginal zone macrophages revisited

Bacterial Infections Following Splenectomy for Malignant and Nonmalignant Hematologic Diseases

Shaping of infant B cell receptor repertoires by environmental factors and infectious disease

Impact of respiratory syncytial virus: the 25 nurse's perspective

Targets of T Cell Responses to SARS-CoV-2 Coronavirus in Humans with COVID-19 Disease and Unexposed Individuals

Broad neutralization of SARS-related viruses by human monoclonal antibodies

Stereotypic Neutralizing V<sub>H</sub> Clonotypes Against SARS-CoV-2 RBD in COVID-19 Patients and the Healthy Population. bioRxiv

Nasal Gene Expression of Angiotensin-Converting Enzyme 2 in Children and Adults

The time course of the immune response to experimental coronavirus infection of man

Seasonal coronavirus protective immunity is short-lasting

Stereotypic Immune System Development in Newborn Children