Abstract
Antigen-specific B cells bifurcate into antibody secreting cells (ASC) and memory B cells after infection or vaccination. ASCs or plasmablasts have been extensively studied in humans but less is known about B cells that get activated but do not differentiate into early plasmablasts. Here, we define the phenotype and transcriptional program of an antigen-specific B cell subset, referred to as activated B cells (ABC), that is distinct from ASCs and is committed to the memory B cell lineage. ABCs were detected in humans after infection with Ebola virus or influenza virus and also after vaccination. By simultaneously analyzing antigen-specific ASCs and ABCs in human blood after influenza vaccination we interrogated the clonal overlap and extent of somatic hypermutation (SHM) in the ASC (effector) and ABC (memory) lineages. Longitudinal tracking of vaccination-induced HA-specific clones revealed minimal increase in SHM over time suggesting that repeated annual immunization may have limitations in enhancing the quality of influenza-specific antibody.
INTRODUCTION
Antigen-specific B cells proliferate and differentiate into antibody secreting cells (ASCs) or maintain the B cell program and become memory cells after infection or vaccination1,2. The fate of responding B cells is tightly controlled by the expression levels of a network of transcription factors; Blimp-1, IRF4 and XBP-1 drive the differentiation of B cells into ASCs and induce the expression of the molecular machinery required for antibody secretion, whereas Pax5, IRF8 and Bcl-6 are among a set of transcription factors that help B cells maintain their identity3.
We previously demonstrated that a transient burst of antigen-specific ASCs could be detected in peripheral blood shortly after influenza vaccination and after influenza and dengue viral infections in humans4–6. Defining the phenotype of the ASC subset paved the way for our group and others to interrogate the repertoire of B cell responses in humans. ASCs have also been used to rapidly generate human monoclonal antibodies (mAbs) that can be utilized as therapeutics and for diagnostics6–8. In the past few years, the dynamics, magnitude, and specificity of the ASC response in peripheral blood after a variety of vaccinations and infections in humans have been extensively studied9.
Apart from ASCs, antigen-specific human memory B cells (MBCs) can also be detected in peripheral blood6,10,11. These memory cells are usually identified either indirectly by culturing peripheral blood mononuclear cells (PBMCs) under various conditions to drive their differentiation into ASCs, or by using fluorescently labeled antigens. In this study, we define a subset of antigen-specific B cells that is phenotypically, transcriptionally and functionally distinct from ASCs, that we refer to as activated B cells (ABCs). We demonstrate that ABCs could be detected in peripheral blood after influenza virus and Ebola virus infections and after influenza immunization in humans. Sequencing of antibody heavy chain (IGH) rearrangements showed that clones present among hemagglutinin (HA)-specific ABCs isolated at days 7 or 14 post-influenza immunization were detected in the MBC compartment three months after vaccination, at higher frequency than in pre-vaccination samples. The magnitude and duration of the ABC response compared to the ASC response was substantially larger when considered over the course of the infection or vaccination, indicating that ABCs may be a particularly useful population for capturing the comprehensive repertoire of vaccine-induced antigen-specific B cell responses.
RESULTS
Defining B cell subsets after influenza immunization
We previously demonstrated that antigen-specific ASCs peak in peripheral blood around day seven post-influenza immunization in humans6. We wanted to examine whether any other subset of antigen-specific B cells distinct from ASCs could be detected in the blood after vaccination. We hypothesized that, similar to ASCs, all recently antigen-engaged B cells would be actively proliferating. We used CD71 as a surface marker to label proliferating peripheral blood B cells after immunization with the inactivated seasonal trivalent influenza vaccine (TIV). From the isotype-switched, proliferating B cells, defined as CD3−CD14−CD16−CD19+IgD−CD71hi cells isolated 7 days after vaccination, we identified two subsets: CD38hiCD20− and CD38int-loCD20hi cells (Fig. 1a). Markers of the former population are typical of those of ASCs, while we referred to the other population as activated B cells or ABCs. Both populations were positive for Ki67, confirming that cells were dividing (Fig. 1b).
We confirmed the identity and specificity of the ASC subset by direct ex vivo ELISPOT (Fig. 1c and Supplementary Fig. 1a). The majority (~82%) of IgG+ ASCs were antigen-specific. ABCs, however, did not spontaneously secrete antibodies under the same conditions (Fig. 1c). ABCs differentiated into IgG-secreting cells when cultured in the presence of autologous CD19− cells as feeders (Fig. 1d and Supplementary Fig. 1b). Approximately 45% of the IgG+ ABCs were specific to influenza vaccine (Fig. 1d).
Similar to ASCs, ABCs were larger in size compared to naïve B cells (Fig. 1e). In contrast to ASCs, ABCs did not downregulate CD19 relative to naïve B cells (Fig. 1e). ABCs showed higher expression of CD27 and lower expression of CD21 relative to naïve B cells (IgD+CD27−) (Fig. 1e).
Down-regulation of the B cell transcription factor Pax5 is one of the earliest events of ASC differentiation12–14. Pax5 expression measured by intracellular staining and flow cytometry was significantly lower in ASCs in comparison to naïve B cells and ABCs (Fig. 1f). In contrast, IRF4 expression was highest in ASCs compared to naïve B cells and ABCs (Fig. 1g). IRF8 is a transcription factor that antagonizes plasma cell differentiation15. Similar to Pax5, IRF8 expression was significantly lower in ASCs in comparison to naïve B cells and ABCs (Fig. 1h). These data demonstrate that following influenza immunization in humans, two functionally and phenotypically distinct antigen-specific B cell subsets can be detected in peripheral blood.
Detecting ABCs after influenza and Ebola virus infections
After characterizing ABCs following influenza vaccination, we sought to define this population after influenza infection. We examined cryopreserved PBMCs from two patients with confirmed 2009 pandemic H1N1 influenza virus infection16. Blood was first collected at the day of enrollment (Day 0) and at several time points thereafter (on or about days 4, 7, 11, and 27 after enrollment). At days 3 and 11 after enrollment, ASCs (CD20−CD71+) and ABCs (CD20hiCD71+) could be distinguished among isotype-switched B cells (Fig. 2a).
B cell responses to influenza in adults are dominated by recall responses6,17. We wanted to test whether we could detect ABCs and ASCs in blood in the course of a primary immune response in humans. We had the unique opportunity to examine the immune responses in four patients (EVD02, EVD05, EVD09 and EVD15) infected with the Ebola virus18,19. The four patients were cared for in the Serious Communicable Diseases Unit at Emory University Hospital18,19. Among CD19+ cells, we were able to clearly distinguish both ASCs (CD20−CD71hi) and ABCs (CD20hiCD71+) during the early phase of the infection (Fig. 2b). A massive B cell response was observed in all patients approximately two to three weeks after onset of symptoms, with ABCs and ASCs comprising almost half of peripheral blood B cells in three out of the four patients (Fig. 2b). The percentages of ABCs and ASCs (of CD19+ cells) were 14.8 and 29.3 in EVD02, 31.4 and 33.3 in EVD05, 14.4 and 36.8 in EVD09, and 16.2 and 4.14 in EVD15, respectively (Fig. 2b). As the infection progressed, the percentage of B cells with the ABC phenotype increased while those with the ASC phenotype decreased (Fig. 2c). These data clearly demonstrate that ABCs could be detected in peripheral blood shortly after viral infection in humans, including after primary infections. It also shows that ABCs can originate from memory B cells (influenza) or naïve B cells (Ebola).
ABCs are transcriptionally distinct from ASCs and MBCs
To define the transcriptional program of ABCs, we analyzed the gene expression profile of ABCs and compared it with those of ASCs, naïve B cells and resting MBCs. We isolated these B cell populations by flow cytometry from four individuals seven days after immunization with the 2013/14 TIV (Supplementary Fig. 2). Unsupervised hierarchical clustering showed that both proliferating B cell subsets (ABCs and ASCs) grouped differently from the resting ones (Naïve and MBCs) (Fig. 3a). Importantly, the four ABC samples formed a distinct cluster, indicating that the differences in gene expression between ABCs and all other B cell subsets were sufficiently robust to be reproducible in the four individuals (Fig. 3a).
Expression analysis revealed 192 genes that were differentially upregulated in the ABC subset (Fig. 3b, Supplementary Table 1). A substantial number of these genes encode proteins that are involved in antigen presentation (for example, HLA-DRA, HLA-DMA, HLA-DPA1, NCF2, CD1C and CTSH), and in B cell antigen receptor (BCR) signaling (CD19, RFTN1, CSK and PRKCB). To validate the array results we examined the expression of some of these differentially expressed genes at the protein level (Fig. 3c). Consistent with the array results, surface protein expression of CD20, TLR10 and CD52 was highest on ABCs (Fig. 3c). Conversely, there were 760 genes differentially upregulated in ASCs (Fig. 3b). These genes included PRDM1, IRF4 and XBP1, which encode the key transcription factors that control ASC generation, maintenance, and antibody secretion. Genes encoding proteins involved in extracellular protein synthesis and secretion (TRAM1, SEC61B and KDELR1) were also substantially upregulated in ASCs (Fig. 3b).
To better understand the interactions between these differentially expressed genes and to put them into biological context, we performed gene set enrichment analysis (GSEA) on pre-ranked gene lists using previously described Blood Transcription Modules (BTMs) as gene sets20,21. This analysis confirmed that the differentially expressed genes in ABCs were mostly associated with antigen processing and presentation and BCR signaling (Fig. 3d,e). Differentially expressed genes in ASCs were associated with a plasma cell signature and the molecular machinery essential for antibody secretion (Fig. 3d). We next examined the genes that were differentially expressed between ABC and resting MBC subsets. We found that those genes were mainly associated with cell cycle regulation, antigen presentation, BCR signaling and MAP kinase signaling (Fig. 3f). These data clearly demonstrate that ABCs represent a transcriptionally distinct subset from ASCs and resting MBCs.
Dynamics of influenza HA-specific B cell responses
To further examine the specificity and dynamics of ABCs after influenza vaccination, we used the HA protein of influenza virus to label antigen-specific cells after seasonal influenza vaccination. The HA used was derived from the 2009 pandemic H1N1 virus (A/California/7/2009), which has been included as a component in all human seasonal influenza vaccines formulations since the 2009/10 season. Two populations of HA-binding cells were clearly identified among isotype-switched B cells seven days after vaccination: HA-positive CD20− cells (or HA-positive ASCs) and HA-positive CD20hi (or HA-positive ABCs) (Fig. 4a–e). Lower HA binding intensity observed on HA-positive ASCs in comparison to HA-positive ABCs is probably a result of the lower density of surface BCR expression on ASCs compared to ABCs (Fig. 4a). HA-positive ABCs were Pax5hi and IRF4lo while HA-positive ASCs were Pax5lo and IRF4hi (Fig. 4b).
The identity and specificity of HA-positive ASCs was confirmed by ex vivo ELISPOT where more than 90% of the IgG-secreting cells were specific to the H1 HA (Supplementary Fig. 3a). In contrast, no IgG-secreting cells were detected with sorted HA-positive ABCs confirming that they are not differentiated into ASCs (Supplementary Fig. 3a). However, after culturing HA-positive ABCs, we found that 62% of these cells secreted HA-binding antibodies (Supplementary Fig. 3b).
At day 14, HA-positive ASCs were no longer detectable in blood. In contrast, the frequency of HA-positive CD20hi isotype-switched B cells was highest at this time point (Fig. 4c). Both HA-positive ASCs and HA-positive ABCs detected at day 7 showed high expression of the proliferation marker CD71 (Fig. 4d). CD71 and CD20 expression remained relatively high on HA-positive ABCs at days 14 & 30 in comparison to naïve B cells (Fig. 4d). Three months after vaccination, CD71 expression on HA-positive memory B cells was indistinguishable from that on naïve B cells (Fig. 4d). We next examined the dynamics of HA-positive ABCs and HA-positive ASCs following influenza infection. Four days after enrollment, HA-positive ASCs (CD20−) and HA-positive ABCs (CD20hi) could be identified among isotype-switched B cells (Fig. 4e). These data demonstrate that HA-positive ABCs and ASCs could be clearly distinguished in peripheral blood shortly after influenza vaccination and infection in humans.
HA-specific clones span ABCs and ASCs
We next wanted to determine whether similar or distinct B cell clones constitute the ABC and ASC compartments isolated after influenza vaccination. To address this question, we sequenced the immunoglobulin heavy chain (IGH) genes from HA-positive and HA-negative ABCs at days 7 and 14, and from ASCs isolated at day 7 (peak of ASC response) from three donors following immunization with the 2014/15 seasonal TIV (Supplementary Fig. 4). Down-regulation of surface Ig expression in differentiating ASCs likely results in underestimation of the number of antigen-specific cells. Approximately 90% of IgG-secreting ASCs from day 7 are specific to TIV by ELISPOT (data not shown), therefore we considered IgG lineages from the ASC pool to be representative of the vaccine-induced clonal lineages in subsequent analyses. B cell clonal lineages were defined by IGH rearrangements that shared variable and joining gene segment usage (referred to hereafter as IGHV and IGHJ), and encoding complementarity-determining region 3 (CDR3) of equal length and similar CDR3 nucleotide sequence, clustering at a 90% identity threshold using single linkage clustering.
An average of 33% (range: 20.1% – 41%) of the day 7 HA-positive ABC clonal lineages were also present among day 7 IgG+ ASCs, while 20% (range: 10.4% – 26.7%) of the day 14 HA-positive ABC clonal lineages were detected in day 7 IgG+ ASC clonal lineages (Fig. 5a, top). H1 HA-positive ABCs represent only a fraction of the total response to all the different antigens derived from the three influenza strains (H1N1, H3N2 and influenza B) included in seasonal TIV preparations. Therefore, we also examined the clonal overlap between total (HA-positive and HA-negative) ABCs and IgG+ ASCs. Among total ABC clonal lineages, we found that 34.4% (range: 21.8% – 52.6%) of day 7 ABCs, and 24% (range: 16% – 29.1%) of day 14 ABCs were present in the day 7 IgG+ ASCs (Fig. 5a, bottom). These data show that many B cell clones have members in both of these two distinct phenotypic subsets, but the majority of clones are found in one subset only.
We next determined the frequency of each of the clonal lineages within the ABC and ASC pools. The clone frequencies within each pool were positively correlated, suggesting that the extent of clonal expansion in each compartment was similar (Fig. 5b). The accumulation of somatic hypermutations (SHMs) in the antibody IGH rearrangement provides information about the cell division history within a clonal lineage. The average IGHV SHM frequency within responding ASCs and ABCs was 7% and 6.5%, respectively (Fig. 5c). These ABC and ASC SHM frequencies were similar to those observed among pre-existing memory B cells, suggesting that MBCs are the likely precursors for both cell subsets. We then assessed the SHM frequencies within the clonal lineages shared between the ABC and ASC pools, and as shown in Fig. 5c the SHM frequencies were strikingly similar between the two compartments. This match in SHM frequencies was observed regardless of the extent of SHM frequency (Fig. 5c). Collectively, these data suggest that part of the vaccine-induced ABC and ASC clonal lineages have originated from the same memory B cells and later diverged into distinct fates.
HA-specific ABCs are clonally related to MBCs
To determine the fate of ABCs, we tracked ABC clonal lineages for three months following TIV vaccination in two separate cohorts (Supplementary Fig. 5). In the first cohort we sorted total ABCs and ASCs from three donors seven days after immunization with the 2013/14 TIV. We also sorted HA-positive ABCs from day 14 (peak of ABC expansion). To gather mutational data from additional B cell clone members, we analyzed the IGH repertoire within total PBMCs from day 0 (before vaccination) and days 28 and 90 post-vaccination. We tracked HA-positive clonal lineages (sorted at day 14) at various time points (Supplementary Fig. 6). A relatively small percentage (average 1%, range: 0.34% – 2%) of the HA-positive day 14 ABC lineages were detectable prior to vaccination. This percentage increased at one-week post-vaccination; an average of 11.6% (range: 4.4% – 23.3%) of the HA-positive day 14 ABC lineages were detectable among D7 ABC lineages (Supplementary Fig. 6). 10% and 8.8% of day 14 HA-positive ABC clonal lineages persisted in peripheral blood and were detectable at days 28 and 90 post-vaccination, respectively, suggesting that these clones had now become part of the resting MBC population (Supplementary Fig. 6).
To further explore the kinetics and clonal relationship between ABC and resting MBCs in humans after influenza vaccination, we analyzed the corresponding repertoires in three additional donors during the 2014/15 influenza season. HA-positive ABCs were sorted from days 7 and 14, as well as MBCs from day 0 and days 14, 28 and 90 after vaccination (Fig. 6a and Supplementary Fig. 5b). HA-positive clonal lineages present in the ABC populations were found to persist as resting memory B cells, with an average of 36.25% (range: 12 – 55.6%) of day 7 HA-positive ABC lineages and an average of 26.7% (range: 21.5 – 36.0%) of day 14 HA-positive ABC lineages detected at day 90 after vaccination (Fig. 6a,b). We examined the persistence of total ABC-derived lineages from day 7 and day 14, and found that the percentages of total ABC lineages still detectable at day 90 post-vaccination were similar to the percentages of HA-positive ABC lineages (Fig. 6a,b). Both ASC and ABC lineages were detectable at a low percentage among pre-vaccination MBCs. An average of 11.7%, (range 4.6% – 19.2%) of IgG+ ASC lineages, and 8%, (range 2.1% – 12.1%) of day 14 ABC lineages were detected in the pre-vaccination memory pool (Fig. 6a, b).
Importantly, the ABC and ASC clonal lineages were largely distinct from resting MBCs sorted at day 14 (Fig. 6b). The percentage of total ABC lineages that included day 14 resting MBCs ranged from 0.3% (donors 4 and 8) – 3.2% (donor 6). The frequency of HA-positive lineages in the day 14 resting MBC pool was very low, and ranged from 0.4% (donor 4) – 4.7% (donor 6) (Fig. 6b). The data, therefore, indicate that the ABC clones are enriched for vaccine-induced lineages, and that few members of these clonal lineages remain in a non-activated state in the blood at day 14 after vaccination. Taken together, our data clearly show that ABC clonal lineages contribute to the resting MBC population.
Minimal impact on SHM frequency by seasonal TIV vaccination
We next wanted to investigate how the extent of SHM of HA-positive ABCs changes over time after seasonal TIV vaccination. To address this question we longitudinally tracked HA-positive ABC clonal lineages and then compared the SHM frequencies within the IGHV segments measured at various time points before and after vaccination. HA-positive ABCs substantially expanded from day 0 to day 7, and reached their peak at day 14 after vaccination (Fig. 7a). From the six donors we analyzed, we did not detect any substantial differences in IGHV SHM frequencies measured prior to vaccination, early after vaccination (days 7 and 14) and those measured at day 90 (Fig. 7b and Supplementary Fig. 7a). This result was consistent with the SHM frequencies of individual HA-positive ABC clonal lineages from donors 4 and 6 (Fig. 7c, d), and from donors 157, 162, 163 and 8 (Supplementary Fig. 7b) measured at various time points before and after vaccination. The initial SHM frequency within individual HA-positive ABC clonal lineages did not seem to have an impact on whether the descendants of that particular lineage will gain more mutation at later time points after vaccination. For example, all five HA-positive ABC clonal lineages tracked from baseline to day 90 after vaccination in donor 4 did not substantially gain any increase in SHM frequency despite the differences in their starting frequencies (Fig. 7c). Longitudinal tracking of HA-specific ABC clones revealed minimal gain in SHM following seasonal influenza immunization.
DISCUSSION
Memory B cells and long-lived plasma cells provide a remarkably stable humoral immune memory that can last for a lifetime22,23. Appropriate induction of both cell populations is therefore essential for optimal efficacy of prophylactic vaccines. Plasmablasts or ASCs in peripheral blood following infection and vaccination of humans have recently been extensively characterized9. In the current study we define an antigen-specific human activated B cell subset (referred to as ABCs) that is phenotypically, transcriptionally and functionally distinct from ASCs. ABCs could be detected within seven days of influenza vaccination in adults, and shortly following influenza and Ebola viral infections. High-throughput sequencing of the ABC immunoglobulin gene repertoire allowed us to track ABC clonal lineages, revealing that these lineages persisted in peripheral blood up to 90 days after vaccination (the latest time point measured in these experiments), suggesting that ABCs contribute to the memory B cell pool. Similar to ASCs, ABCs could be used for unbiased interrogation of the antigen-specific B cell repertoire following infection and vaccination of humans, and also as a source for procuring therapeutic mAbs. These ABC-derived mAbs may differ from those derived from ASCs in terms of quality and specificity24.
20–30% of the ABC clonal lineages are shared with the ASC compartment, indicating that while some clones adopt both phenotypes, it may be more common for individual clones to be polarized to an ASC or ABC phenotype in a vaccine response. This percentage could be an underestimate due to the limited number of individual B cell members of antigen-specific clones that are captured in the volume of blood sampled for research studies. In our previous influenza vaccination and infection studies4,6,17 we reported that antigen-specific IgG-secreting cells that possessed high IGH SHM frequencies dominated the ASC response, and attributed this result to the likely MBC origin of these ASCs. Here, we show that HA-positive ABC lineages possessed SHM loads similar to both ASCs and to the pre-vaccination MBC pool, suggesting that ABCs are also derived from MBCs. This notion was confirmed by tracking of HA-positive lineage members from the pre-vaccination MBC compartment into the post-vaccination ABC, ASC and MBC populations.
Memory B cells, by definition, are quiescent25. This is why we make the distinction between the Ki67+ and CD71hi ABC subset and resting MBCs. This is consistent with our observation that very few ABC clonal lineages were detectable in the resting MBC compartment isolated at day 14 after vaccination. Conversely, starting from days 28 and 90 after vaccination (after the immune reaction had subsided) we detected up to 60% of the ABC clonal lineages in the resting MBC compartment. The delayed kinetics of ABCs are consistent with previous reports demonstrating delayed peaking of antigen-specific MBCs in comparison to ASCs following tetanus and influenza immunization6,26. This difference in timing following infection or vaccination may reflect a preference for the immune system to first rapidly generate ASCs whose antibodies directly engage the foreign antigens. ABCs could have greater capacity to mutate their BCRs as indicated by the higher expression of AID relative to ASCs (data not shown), however, we did not detect substantial mutational load differences between these two subsets. Overall, our data supports a model where antigenic stimulation drives naïve or resting memory B cells to undergo a rapid clonal expansion. Some of of the expanded cells terminally differentiate into ASCs. The rest of the progeny, ABCs, are released into the circulation more slowly. Over time ABCs down-regulate the activation markers, and join the resting MBC pool.
Perhaps the most intriguing finding in the current study is that the majority of responding B cell clones did not experience a substantial increase in SHM frequency over time following TIV immunization. These data suggest that while the tens of millions of seasonal influenza vaccine doses administered every year have the capability of boosting B cell responses quantitatively, the impact on the quality of these responses is minimal. One possible explanation is that the TIV-induced B cells could have originated from pre-existing MBCs with already high affinity receptors. Our cohorts were all adults with probable existing MBCs elicited by prior exposures to influenza antigens, such as the 2009 pandemic H1 HA, included in the seasonal influenza vaccine used in the current study. So perhaps the lack of novel epitopes being engaged by responding B cells has contributed to the minimal changes observed in SHM frequencies. Alternatively, the vaccine antigen amount may have been too limited to support a persistent germinal center reaction from which more highly mutated clones could have emerged. Finally, it is important to note that the TIV vaccine used was non-adjuvanted. It will be important to examine whether adjuvants could change the dynamics of affinity maturation associated with seasonal influenza vaccination in humans. Defining and studying ABCs along with ASCs in humans will reveal the intrinsic differences between the cells constituting these two compartments, which in turn would enhance our understanding of the longevity of vaccine-induced B cell responses.
Methods
Patients
All studies were approved by the Institutional Review Boards of Emory University, St. Jude Children’s Research Hospital, the University of Tennessee Health Science Center/Le Bonheur Children’s Hospital, and by the US Centers for Disease Control (CDC). Written consent was obtained from all patients. All work with Ebola specimens was performed in the CDC BSL-4 laboratory space. For the influenza vaccination studies, all subjects were healthy adults 18–49 years of age. Peripheral blood mononuclear cells (PBMCs) were isolated using Vacutainer tubes (BD) for immediate use or cryopreserved.
Vaccines
Trivalent, inactivated seasonal influenza vaccines or TIV (2012/13, 2013/14 and 2014/15 seasons) were obtained from Sanofi Pasteur Inc.
Antigens
For ELISPOT, we either used the TIV vaccine (diluted 1:20) or recombinant H1 HA protein derived from the pH1N1 (A/California/06/2009) influenza virus. HA was obtained from the Biodefense and Emerging Infections (BEI) research repository. For HA staining, we used the HA1 subunit from H1 HA of A/California/06/2009 (H1N1) that was expressed in 293 T cells. HA molecules were biotinylated using the EZ-Link Micro NHS-PEG4-Biotinylation kit (Thermo-Fisher). Excess biotin was removed using the Zeba Spin Desalting Columns (Pierce).
Flow cytometry and cell sorting
Flow cytometry analysis was performed using either fresh or cryo-preserved PBMCs. For the Ebola samples, whole blood was incubated with appropriate surface antibody mix for 30 min then 1x FACS/Lyse (BD) was added to each sample. After a 5-min incubation, the sample was centrifuged to pellet cells and pellets were washed twice in PBS prior to data collection. For influenza vaccination and infection samples, PBMCs were resuspended in PBS supplemented with 2% FBS and 1mM EDTA. Cells were stained for 30 min at 4 °C for surface markers. For analyzing the intracellular IRF4, IRF8 and Pax5 molecules, cells were further fixed and permeabilized (eBioscience), and then incubated with antibodies for 1 h at 25°C. All reagents for flow cytometry staining are listed in Supplementary Table 2. Flow cytometry data were analyzed using FlowJo software.
HA staining
293 cells-expressed HA1 subunit of the pH1N1 H1 HA was biotinylated (EZ-Link™ Sulfo-NHS-Biotin, Life Technologies). PBMCs were incubated the biotinylated HA for 30 min at 4 °C prior to staining with APC-conjugated streptavidin.
ELISPOT and memory B cell assay
Direct ELISPOT to enumerate the number of either total IgG-secreting, vaccine-specific and recombinant HA-specific plasmablasts present in the PBMC samples. Basically, dilutions of washed PBMCs incubated in supplemented RPMI were incubated in ELISPOT plates for 18 h. After washing the plates, secreted antibodies were detected with anti-human IgG-biotin (Jackson ImmunoResearch) and avidin-D-HRP (Vector Laboratories) and developed with AEC substrate (Sigma) before analysis on an ELISPOT counter (Cellular Technologies Ltd.). Differentiation of ABCs to ASCs was done by culturing cells with autologous 105 CD19− PBMCs R-10 supplemented with phosphothiolated CpG ODN-2006, CD40L (R&D) and IL-21 (Peprotech). After culture for 6 d the cells were washed and quantified by ELISPOT assay.
Gene expression array
From four subjects, 104 cells of either naïve B cells, memory B cells (MBCs), ABCs, ASCs were sorted seven days after influenza vaccination into RLT-Plus buffer (Qiagen). Total RNA was isolated using the Rneasy micro kit from Qiagen. After RNA amplification (Nugen), the genome-wide transcriptional profiles of both populations were assessed using the HumanHT-12 v4 Expression Bead Chip (Illumina). Data were quantile normalized before analysis.
Gene expression analysis
Hierarchical cluster analysis was performed using the top 15,000 probes with highest variance across all samples (Euclidean distance and Ward’s criterion). To assess the uncertainty in cluster analysis, we run Pvclust R package with 5,000 bootstrap resampling27. To identify genes whose expression was specifically high in one B cell subset compared to others, we performed a student’s t-test (P < 0.05 and mean fold-change > 1.5) between any given two cell subsets28. Genes were classified as highly expressed in one subset if they were up-regulated in this subset compared to at least 2 other subsets (out of 3 possible comparisons). We performed gene set enrichment analysis (GSEA) on pre-ranked gene lists using Blood Transcription Modules (BTMs) as gene sets. Expression of each gene across all samples was normalized by z score transformation and the mean z score of each subset was used to generate the pre-ranked gene lists. Probes representing the same gene symbol were collapsed by taking the z score with highest absolute value. Nominal P-values and normalized enrichment scores for the BTMs were generated running GSEA with 1,000 permutations. Cytoscape was used to display the gene network29.
Deep sequencing of IGH repertoires from influenza vaccine response
For the first cohort of vaccinees (2012–13 and 2013–14 seasons), RNA was purified with a Qiagen Rneasy micro kit from: 1) PMBCs from pre-vaccination, days 28 and 90 post-vaccination, 2) Flow sorted ASCs and ABCs isolated at day 7 post-vaccination, and 3) Flow sorted H1-specific ABCs isolated at day 14 post-vaccination. For the second cohort of vaccinees (2014–15 season) RNA was purified from: 1) PBMCs from pre-vaccination day 0, and post-vaccination days 7, 14, 28 and 90, 2) Flow sorted H1-positive and H1-negative ASCs and ABCs from day 7 post-vaccination, 3) Flow sorted HA-positive and HA-negative ABCs from day 14 post-vaccination, 4) Flow sorted memory B cells isolated prior to vaccination and at days 14, 28 (sorted for HA-positive and HA-negative) and 90 after vaccination, and 5) Naïve B cells flow sorted prior to vaccination. cDNA was synthesized with Superscript II enzyme (Life Technologies) and priming by random hexamers. Templates were amplified by PCR using Biomed IGHV primers in the framework 1 (FR1) region and isotype-specific primers for IgG, IgM and IgA located in the first exon of the constant regions. These primers also encoded approximately half of the Illumina linker sequences needed for cluster generation and sequencing on the Illumina MiSeq instrument. Sample identity was encoded by 8-nucleotide multiplex identifier barcodes in each primer. For Illumina cluster recognition, four randomized nucleotides were encoded in the primers immediately after the Illumina linker sequence in the constant region primers. Each antibody isotype for each sample was amplified in a separate PCR reaction, to prevent formation of cross-isotype chimeric PCR products. PCR was carried out with AmpliTaq Gold (Roche) following the manufacturer’s instructions, and used a program of: 94°C 7 min, 35 cycles of (94°C 30 sec, 58°C 45 sec, 72°C 120 sec), and final extension at 72°C for 10 min. A second PCR step was used to add the remaining portion of the Illumina linkers to the amplicons, and was carried out with the Qiagen Multiplex PCR kit (Qiagen) according to the manufacturer’s instructions, using 0.4 microliter of the first PCR product as template in a 30-microliter reaction. The PCR program for the second PCR step was 94°C 15 min, 12 cycles of (94°C 30 sec, 60°C 45 sec, 72°C 90 sec), and final extension at 72°C for 10 min. The products of each PCR reaction were quantitated, pooled in equimolar amounts, electrophoresed on agarose gels, and gel extracted with QIAquick kits (Qiagen). High-throughput sequencing was performed on an Illumina MiSeq instrument using 600-cycle sequencing kits.
Analysis of vaccine response lGH clonal lineages
Paired-end reads from Illumina sequencing were merged using FLASH30 and samples were de-multiplexed according to the primer-encoded barcodes. Only exact, full-length matches to barcodes were permitted. Reads were aligned using a local installation of IgBLAST31 to determine features such as germline IGHV, IGHD and IGHJ gene usage, to define framework and complementarity determining regions and to identify mutation positions. Non-productive sequences that encoded either stop codons or had out-of-frame IGHJ were discarded from analysis, along with those that encoded CDR3 segments shorter than five residues (as these could not be reliably assigned to lineages). Antibody isotype subclasses were identified by exact matching to the CH1 sequence upstream of the primer used for amplification. For each donor, the clonal lineages were inferred from IGH rearrangements across the available samples by single linkage clustering of the CDR3 nucleotide sequences with a 90% identity threshold for IGH sequences that shared IGHV and IGHJ gene assignment (ignoring IGHV or IGHJ allele) and CDR3 length. HA-binding lineages were defined as clonal lineages for which a member in any phenotypic B cell subset was shown to bind HA in flow cytometry experiments; for lineages that had members detected in both HA-positive and HA-negative populations from flow sorting, we considered the lineage to be HA-binding if at least 10 reads were detected in the HA-positive samples, and at least 10% of total reads from the lineage in sorted cells were found in the HA-positive populations. IgG-expressing ASC lineages were defined as those containing members expressing IgG subtypes IgG1, IgG2 or IgG3. IGHV mutation percentages were calculated for the region spanning FR1 through the end of FR3. In summarizing IGHV mutation frequencies in B cell subsets, the mean mutation frequency was calculated for each B cell clone, then the median of these means over all clones was determined. The distribution of clonal lineage members across the different sample points was visualized using Circos32.
Statistics
Statistical analyses (described in context) were performed using GraphPad Prism and R.
Supplementary Material
Acknowledgments
We thank Robert Karaffa and Sommer Durham for excellent technical assistance. This work was funded in parts by National Institutes of Health (NIH)/National Institute of Allergy and Infectious Diseases (NIAID) contract number HHSN266200700006C, grants 1P01AI097092, U19AI117891 and Advanced Immunization Technologies (ADITEC) #280873, European Union (to R.A.), and by grants U19AI09525801, UM1AI100663, and U01AI104342 (to S.D.B.). A.E. is supported by the training grant (T32AI074492) from the National Institute of Allergy and Infectious Diseases. H.I.N. receives investigator fellowship from CNPq, Brazil. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Allergy and Infectious Diseases or the National Institutes of Health or that of the Centers for Disease Control and Prevention.
Footnotes
Accession codes. GEO: Microarray data, GSE68245. SRA: BCR deep sequencing data, SRP075992.
AUTHOR CONTRIBUTIONS
A.E., K.J.L.J., S.D.B. and R.A. designed the study, interpreted data and wrote the paper. A.E. performed most of the experiments. K.J.L.J. performed NGS sequencing of IGH repertoire and devised, undertook and interpreted repertoire data analysis. K. M. R. processed repertoire sequence data. H.K. and H.N. analyzed the microarray data. C.W.D. and A.K.M. helped in data analysis and interpretation. C.O., R.E., S.T., G.L., C.S., A.M., and P.T. helped in collecting and processing the clinical samples.
COMPETING FINANCIAL INTERESTS
The authors declare no competing financial interests.
References
- 1.Tangye SG, Tarlinton DM. Memory B cells: effectors of long-lived immune responses. Eur J Immunol. 2009;39:2065–2075. doi: 10.1002/eji.200939531. [DOI] [PubMed] [Google Scholar]
- 2.McHeyzer-Williams M, Okitsu S, Wang N, McHeyzer-Williams L. Molecular programming of B cell memory. Nat Rev Immunol. 2012;12:24–34. doi: 10.1038/nri3128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Singh H, Khan AA, Dinner AR. Gene regulatory networks in the immune system. Trends Immunol. 2014;35:211–218. doi: 10.1016/j.it.2014.03.006. [DOI] [PubMed] [Google Scholar]
- 4.Wrammert J, et al. Broadly cross-reactive antibodies dominate the human B cell response against 2009 pandemic H1N1 influenza virus infection. J Exp Med. 2011;208:181–193. doi: 10.1084/jem.20101352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Wrammert J, et al. Rapid and massive virus-specific plasmablast responses during acute dengue virus infection in humans. J Virol. 2012;86:2911–2918. doi: 10.1128/JVI.06075-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Wrammert J, et al. Rapid cloning of high-affinity human monoclonal antibodies against influenza virus. Nature. 2008;453:667–671. doi: 10.1038/nature06890. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Corti D, et al. A neutralizing antibody selected from plasma cells that binds to group 1 and group 2 influenza A hemagglutinins. Science. 2011;333:850–856. doi: 10.1126/science.1205669. [DOI] [PubMed] [Google Scholar]
- 8.Thomson CA, et al. Pandemic H1N1 Influenza Infection and Vaccination in Humans Induces Cross-Protective Antibodies that Target the Hemagglutinin Stem. Front Immunol. 2012;3:87. doi: 10.3389/fimmu.2012.00087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wilson PC, Andrews SF. Tools to therapeutically harness the human antibody response. Nat Rev Immunol. 2012;12:709–719. doi: 10.1038/nri3285. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Pinna D, Corti D, Jarrossay D, Sallusto F, Lanzavecchia A. Clonal dissection of the human memory B-cell repertoire following infection and vaccination. Eur J Immunol. 2009;39:1260–1270. doi: 10.1002/eji.200839129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Frölich D, et al. Secondary immunization generates clonally related antigen-specific plasma cells and memory B cells. J Immunol Baltim Md 1950. 2010;185:3103–3110. doi: 10.4049/jimmunol.1000911. [DOI] [PubMed] [Google Scholar]
- 12.Medina KL, Singh H. Genetic networks that regulate B lymphopoiesis. Curr Opin Hematol. 2005;12:203–209. doi: 10.1097/01.moh.0000160735.67596.a0. [DOI] [PubMed] [Google Scholar]
- 13.Nutt SL, Heavey B, Rolink AG, Busslinger M. Commitment to the B-lymphoid lineage depends on the transcription factor Pax5. Nature. 1999;401:556–562. doi: 10.1038/44076. [DOI] [PubMed] [Google Scholar]
- 14.Nutt SL, Kee BL. The transcriptional regulation of B cell lineage commitment. Immunity. 2007;26:715–725. doi: 10.1016/j.immuni.2007.05.010. [DOI] [PubMed] [Google Scholar]
- 15.Xu H, et al. Regulation of bifurcating B cell trajectories by mutual antagonism between transcription factors IRF4 and IRF8. Nat Immunol. 2015;16:1274–1281. doi: 10.1038/ni.3287. [DOI] [PubMed] [Google Scholar]
- 16.Oshansky CM, et al. Mucosal immune responses predict clinical outcomes during influenza infection independently of age and viral load. Am J Respir Crit Care Med. 2014;189:449–462. doi: 10.1164/rccm.201309-1616OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Li GM, et al. Pandemic H1N1 influenza vaccine induces a recall response in humans that favors broadly cross-reactive memory B cells. Proc Natl Acad Sci U S A. 2012;109:9047–9052. doi: 10.1073/pnas.1118979109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lyon GM, et al. Clinical care of two patients with Ebola virus disease in the United States. N Engl J Med. 2014;371:2402–2409. doi: 10.1056/NEJMoa1409838. [DOI] [PubMed] [Google Scholar]
- 19.McElroy AK, et al. Human Ebola virus infection results in substantial immune activation. Proc Natl Acad Sci U S A. 2015;112:4719–4724. doi: 10.1073/pnas.1502619112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Li S, et al. Molecular signatures of antibody responses derived from a systems biology study of five human vaccines. Nat Immunol. 2014;15:195–204. doi: 10.1038/ni.2789. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Subramanian A, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102:15545–15550. doi: 10.1073/pnas.0506580102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Amanna IJ, Carlson NE, Slifka MK. Duration of humoral immunity to common viral and vaccine antigens. N Engl J Med. 2007;357:1903–1915. doi: 10.1056/NEJMoa066092. [DOI] [PubMed] [Google Scholar]
- 23.Crotty S, et al. Cutting edge: long-term B cell memory in humans after smallpox vaccination. J Immunol Baltim Md 1950. 2003;171:4969–4973. doi: 10.4049/jimmunol.171.10.4969. [DOI] [PubMed] [Google Scholar]
- 24.Purtha WE, Tedder TF, Johnson S, Bhattacharya D, Diamond MS. Memory B cells, but not long-lived plasma cells, possess antigen specificities for viral escape mutants. J Exp Med. 2011;208:2599–2606. doi: 10.1084/jem.20110740. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Takemori T, Kaji T, Takahashi Y, Shimoda M, Rajewsky K. Generation of memory B cells inside and outside germinal centers. Eur J Immunol. 2014;44:1258–1264. doi: 10.1002/eji.201343716. [DOI] [PubMed] [Google Scholar]
- 26.Odendahl M, et al. Generation of migratory antigen-specific plasma blasts and mobilization of resident plasma cells in a secondary immune response. Blood. 2005;105:1614–1621. doi: 10.1182/blood-2004-07-2507. [DOI] [PubMed] [Google Scholar]
- 27.Suzuki R, Shimodaira H. Pvclust: an R package for assessing the uncertainty in hierarchical clustering. Bioinforma Oxf Engl. 2006;22:1540–1542. doi: 10.1093/bioinformatics/btl117. [DOI] [PubMed] [Google Scholar]
- 28.Nakaya HI, et al. Systems biology of vaccination for seasonal influenza in humans. Nat Immunol. 2011;12:786–795. doi: 10.1038/ni.2067. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Saito R, et al. A travel guide to Cytoscape plugins. Nat Methods. 2012;9:1069–1076. doi: 10.1038/nmeth.2212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Magoč T, Salzberg SL. FLASH: fast length adjustment of short reads to improve genome assemblies. Bioinforma Oxf Engl. 2011;27:2957–2963. doi: 10.1093/bioinformatics/btr507. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Ye J, Ma N, Madden TL, Ostell JM. IgBLAST: an immunoglobulin variable domain sequence analysis tool. Nucleic Acids Res. 2013;41:W34–40. doi: 10.1093/nar/gkt382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Krzywinski M, et al. Circos: an information aesthetic for comparative genomics. Genome Res. 2009;19:1639–1645. doi: 10.1101/gr.092759.109. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.