Ex vivo characterization and isolation of rare memory B cells with antigen tetramers

Abstract

Studying human antigen-specific memory B cells has been challenging because of low frequencies in peripheral blood, slow proliferation, and lack of antibody secretion. Therefore, most studies have relied on conversion of memory B cells into antibody-secreting cells by in vitro culture. To facilitate direct ex vivo isolation, we generated fluorescent antigen tetramers for characterization of memory B cells by using tetanus toxoid as a model antigen. Brightly labeled memory B cells were identified even 4 years after last immunization, despite low frequencies ranging from 0.01% to 0.11% of class-switched memory B cells. A direct comparison of monomeric to tetrameric antigen labeling demonstrated that a substantial fraction of the B-cell repertoire can be missed when monomeric antigens are used. The specificity of the method was confirmed by antibody reconstruction from single-cell sorted tetramer(+) B cells with single-cell RT-PCR of the B-cell receptor. All antibodies bound to tetanus antigen with high affinity, ranging from 0.23 to 2.2 nM. Furthermore, sequence analysis identified related memory B cell and plasmablast clones isolated more than a year apart. Therefore, antigen tetramers enable specific and sensitive ex vivo characterization of rare memory B cells as well as the production of fully human antibodies.

Figure 1

Overview of the technique. (A) Soluble antigen or antigen domain is expressed with a BirA tag for site-specific biotinylatation and tetramerization with fluorescently labeled streptavidin. (B) B cells are stained with tetramer and a panel of monoclonal antibodies. Tetramer-positive class-switched memory B cells (CD19⁺ CD27⁺ IgM⁻) are single-cell sorted into PCR strips. (C) mRNA preamplification is performed with T7 RNA polymerase. Single-stranded cDNA is synthesized by the use of a primer with a single-stranded T7 RNA polymerase site. Conversion to double-stranded cDNA enables an in vitro transcription reaction with T7 RNA polymerase, which provides sufficient amounts of RNA for RT-PCR from resting, recirculating memory B cells. (D) Sequencing of PCR products is carried out directly from 300- to 400-bp PCR products by the use of second-round forward and reverse primers. (E) Overlap PCR is used for construction of full-length IgG1 heavy chain and κ light sequences, which are cloned into separate vectors. These vectors are transiently transfected into CHO-S cells for expression of fully human recombinant antibodies.

Figure 2

Detection of tetanus toxoid specific plasmablasts after vaccination. (A) Plasmablasts were identified on the basis of intermediate CD19 and high CD27 expression. Fluorescence-activated cell sorter plots were gated on total CD19⁺ B cells, which were negative for a panel of exclusion markers (CD3, CD14, CD16, and 7AAD). Numbers adjacent to the gate represent the percentage of plasmablasts within the CD19⁺ B-cell population. (B) Identification of TTCF tetramer-positive cells within the plasmablast population (gated in panel A). Numbers adjacent to gate represent the percentage of TTCF tetramer-positive within the plasmablast population (identified in panel A). (C) Frequency of tetramer-positive plasmablasts within the total CD19⁺ B-cell population during the first 2 weeks after vaccination. (D) Variable gene segment usage by unique B-cell clones for heavy (n = 26) and light (n = 25) chains from single cell sorted tetramer-positive plasmablasts isolated from days 6 and 7. (E) Scatter plot of somatic mutations detected in variable gene segments of heavy and light chains with mean values (V_H = 25.3 and V_K = 16.2) indicated by vertical solid lines.

Figure 3

Detection of tetanus toxoid specific memory B cells. (A) Gating scheme for detection of TTCF tetramer-positive B cells among class-switched memory B cells. (B) Identification of TTCF tetramer-positive class-switched memory B cells from 4 donors. The elapsed time since the last tetanus boost is indicated (top). Labeling with a control tetramer (membrane-proximal domain of CD80) is shown for each donor (bottom). Frequencies represent tetramer-positive cells among CD19⁺ CD27⁺ IgM⁻ B cells. (C) Scatter plot of tetramer⁺ cells detected per 1 × 10⁶ B cells, with the mean indicated by a solid vertical line. The mean frequency was 52.0 cells per 1 × 10⁶ B cells with the TTCF tetramer, and 4.3 cells per 1 × 10⁶ B cells with the control tetramer. (D) Scatter plot showing percentage of tetramer-positive cells within the memory B-cell population. The mean frequency of TTCF tetramer-positive was 0.0447% (0.0032% for the control tetramer).

Figure 4

Comparison of monomeric and tetrameric antigen for identification of memory B cells. (A) Mono-biotinylated TTCF or CD80 antigens were directly labeled with the Alexa-488 fluorophore; tetramers were generated with unlabeled streptavidin. Enriched B cells from each donor were split into 3 fractions and stained with control CD80 tetramer, TTCF monomer, or TTCF tetramer at the same total antigen concentration of 0.125 μg/mL. Fluorescence-activated cell sorter plots depict CD19⁺ CD27⁺ IgM⁻ class-switched memory B cells; numbers adjacent to the gate represent the percentage of the parental gate. (B) Frequencies of tetramer⁺ memory B cells detected in 3 different donors. Numbers are calculated as tetramer-positive cells per 1 × 10⁶ CD19⁺ memory B cells.

Figure 5

Clonally related B cells detected in both memory and plasmablast populations. (A) Alignment of CDR3 protein sequences as well as V_H-D-J_H and V_κ-J_κ gene segment use of clonally related memory B cells and plasmablasts. Amino acid differences are shaded in gray. (B) Variable gene segments were aligned at the nucleotide level for clonally related memory B cells and plasmablasts. Solid vertical lines represent coding/replacement mutations, and dashed lines silent mutations per codon, compared with the most homologous germline segment. Asterisks denote coding mutations that occurred at the same codon positions but resulted in different amino acids between the aligned sequences.

Figure 6

High-affinity binding of TTCF by antibodies generated from plasmablasts and memory B cells. Saturation binding experiments were performed to determine the affinities of recombinant antibodies. TTCF antigen was labeled with europium, which emits a strong fluorescent signal at 615 nm on incubation with a chelating reagent. Antibodies were immobilized in a 96-well plate and incubated with TTCF-europium (100nM to 4pM) for 2 hours at 37°C. Fluorescence counts at 615 nm were recorded and K_D calculated by the use of nonlinear regression analysis. A control antibody (clone 8.18.C5) that was also produced in CHO-S cells was included in all experiments. (A) Recombinant TTCF Abs 1 and 2 were generated from TTCF tetramer-positive plasmablasts (donor 1). (B) TTCF Abs 3, 4, and 5 originated from TTCF tetramer-positive memory B cells of 3 different donors.