A simple yet effective method that employs magnetic nanoparticles to detect and enrich antigen-reactive B cells for functional and phenotypic analysis is described.
B cells reactive with a specific antigen usually occur at a frequency of <0.05% of lymphocytes. For decades researchers have sought methods to isolate and enrich these rare cells for studies of their phenotype and biology. Approaches are inevitably based on the principle that B cells recognize native antigen by virtue of cell surface receptors that are representative in specificity of antibodies that will eventually be secreted by their differentiated daughters. Perhaps the most obvious approach to the problem involves use of fluorochrome-conjugated antigens in conjunction with fluorescence-activated cell sorting (FACS). However, the utility of these methods is limited by cell frequency and the achievable rate of analysis and isolation by electronic sorting. A novel method to enrich rare antigen-specific B cells using magnetic nanoparticles that results in high yield enrichment of antigen-reactive B cells from large starting cell populations is described. This method enables improved monitoring of the phenotype and biology of antigen reactive cells before and following in vivo antigen encounter, such as after immunization or during development of autoimmunity.
Limiting dilution analyses of antibody-secreting cell precursor frequency have suggested that B cells reactive to a particular antigen typically occur at a frequency of 0.05 to 0.005% in the normal repertoire, depending on vaccination status and size/number of epitopes present on the antigen. The low frequency of these cells has made it difficult to study changes in their status during development of immune responses, such as following vaccination or exposure to a foreign antigen, or during development of autoimmunity. Previously, researchers have undertaken isolation of antigen-reactive B cells using techniques ranging from antigen coated plates or column adsorbents, to antigen-coated red blood cell resetting, to fluorescence-activated cell sorting1,2,3,4,5,6. Though these techniques have been successful in identifying and isolating antigen-reactive B cells, the results have varied in terms of yield, purity and scalability. Recently we developed a novel method to both detect and enrich rare B lymphocyte subpopulations using magnetic nanoparticles. The method enables enrichment with relatively high yield and purity from large starting populations, and is compatible with analysis of responses to antigen. By enriching from populations of cells in suspension, the method eliminates constraints that are associated with the geometry of antigen-coated plates or columns, and limit throughput. Finally, since enriched cells remain associated with antigen and a fluorescent reporter, they can be further purified by FACS sorting. As described herein we have used this approach for study of peripheral blood tetanus toxoid-reactive B cells before and following immunization of human subjects, as well as autoantigen-reactive B cells from subjects with various autoimmune disorders, including type 1 diabetes, Graves' disease, and Hashimoto's disease7. The method works equally well in mouse and human, and is compatible with analysis of antigen-reactive B cells from a variety of tissues (manuscript in preparation).
In its basic format, peripheral blood mononuclear cells are first incubated with biotinylated antigen along with antibodies to cell surface antigens required for phenotypic analysis. This labeling step is followed by washing and fixation, and addition of streptavidin coupled to far-red-fluorescent dye for detection of the biotinylated-antigen binding cells (Figure 1). Previous studies have identified antigen-specific B cells in a similar manner but using antigens directly conjugated to a fluorochrome8,9,10,11. Although this is a worthy approach, use of biotinylated antigens in conjunction with streptavidin enables greater signal amplification (hence better differentiation of binding and non-binding cells), particularly when antigens are small12,13,14. An additional consideration is the use of streptavidin instead of avidin because streptavidin is deglycosylated, decreasing non-specific binding. Further, we use far-red-fluorescent dye as the fluorochrome due to its photostability, quantum yield (brightness), and its small size (~1.3 kD). Protein fluorochromes such as phycoerythrin (~250 kD) and allophycocyanin (~105 kD)15 are not optimal because they potentially contain many antigenic epitopes. Use of a small organic fluorescent dye composed of a single epitope, such as far-red-fluorescent dye, reduces complexity of the isolated cell population.
Once cells are biotinylated-antigen and far-red-fluorescent dye-streptavidin adsorbed, they are enriched using anti-far-red-fluorescent dye-conjugated magnetic nanoparticles. Single nanoparticles are not detected by most flow cytometers and therefore need not be removed prior to purification by FACS sorting and downstream assays16. Magnetic selection for antigen-specific B cells enriches the population of interest, eliminating the time and cost of sorting rare events using a flow cytometer.
Below we show representative results from enrichment of tetanus-toxoid-specific B cells from a subject before and seven days after tetanus toxoid booster immunization. We chose this particular application as an example in order to demonstrate the ability of this method to enrich antigen-specific B cells following acute in vivo stimulation. When coupled with flow cytometry, this method is capable of enriching and differentiating antigen-specific naïve, memory, and plasmablast B cells and allows the researcher to follow changes in their frequency over time. In addition, we include another possible downstream assay, e.g. an ELISPOT assay, which demonstrates that cells retain the ability to secrete antibody following enrichment. Another application of this method could involve adoptive transfer of enriched cells into a host. We have previously shown cells maintain the ability to act as antigen presenting cells to antigen-specific T cells following isolation and transfer (data not shown). Hence, there are a number of possible downstream assays that could be coupled to the method, which together informs the understanding of the antigen-specific immune response. We have described the method below, including controls to determine overall yield, purity, cell specificity, and fold-enrichment.
1. Isolation of Human PBMCs
2. Staining of Cells
3. Magnetic Nanoparticle-based Enrichment
4. Magnetic Enrichment Using LS or LD Columns
Analysis of purity, yield, and fold-enrichment using flow cytometry
Populations enriched as described above inevitably contain contaminating cells that have not bound streptavidin- far-red-fluorescent dye but are trapped in the matrix. These impurities can be removed from enriched populations by FACS sorting. To estimate purity of enriched populations, gate on live cells based on forward and side scatter and/or live/dead stain and analyze far-red-fluorescent dye + cell frequency. The percentage within this latter gate is indicative of purity. Further analysis of the specificity and antigen affinity of these cells requires FACS sorting of single cells, followed by cloning and sequencing of expressed immunoglobulin genes and recombinant antibody production and analysis.
Yield refers to the number of isolated antigen-binding B cells recovered in Fraction A relative to the number in the starting population (Fraction D). Since the same number of cells were in Fraction D as Fraction A, one can directly determine the yield (# of antigen-binding B cells in Fraction A ÷ # of antigen-binding B cells in Fraction D).
Fold-enrichment (E) is reflected by the percentage of far-red-fluorescent dye + B cells in the "positively" enriched Fraction A relative to that in Fraction D. This value indicates the proportional increase in antigen-specific B cells achieved by nanoparticle enrichment. The value can be found using the following equation:
To demonstrate this technique, we show representative results obtained using tetanus-toxoid (Tet-Tox) as a model antigen to detect and enrich for Tet-Tox-reactive B cells in the peripheral blood of subjects who have been vaccinated against tetanus at varying time points. Figure 1 is a schematic of the method and the antigen adsorbent. Figure 2a shows the frequency of Tet-Tox-reactive B cells from an individual who was last vaccinated against tetanus more than a decade ago. Shown are the enriched ("positive") and depleted ("negative") populations from Fraction A and the baseline frequency of Tet-Tox-specific B cells from Fraction D. Hence, in this example we are able to enrich binding cells by around 7-fold and achieved a purity of 4% in the enriched fraction.
Figure 2b shows the specificity of the reaction using blood from a subject who had been vaccinated ~3 years earlier. When we added a 50-fold excess (50 µM) of unlabeled tetanus-toxoid to the cells before addition of the biotinylated antigen (Fraction B) we are able to block binding by 83%, demonstrating the specificity of antigen binding. When we omitted biotinylated antigen from the procedure (Fraction C), a small number (~ 0.2%) of B cells still bound; this small fraction of B cells presumably reflects binding to some non-antigen in the adsorbant, such as the streptavidin–far-red-fluorescent dye, the anti–far-red-fluorescent dye antibody or the magnetic beads.
To demonstrate use of the method to analyze changes in antigen-binding cell phenotype following immunization we show in Figure 3 the frequency of Tet-Tox-specific B cells in the naïve, memory, and plasmablast populations in blood drawn before and 7 days after booster vaccination of a healthy subject. While the percentage of naïve B cells among total Tet-Tox-binding B cells decreased from 84% to 64% after boost, the percentage of memory and plasmablast populations among total Tet Tox-reactive cells increased from 16% to 36% and 0% to 40%, respectively. Figure 3b shows representative ELISPOT data from peripheral blood from the 7 days post-boost healthy subject. Isolation of the Tet-Tox-binding B cells led to 4-fold enrichment in tetanus toxoid antibody secreting-cell frequency.
Figure 1: Method for detection and enrichment of tetanus toxoid (Tet Tox)-binding B cells. (A) Protocol for enrichment and isolation of Tet-Tox-binding B cells. (B) Diagram of adsorbent employed to identify and enrich Tet-Tox-binding B cells. Please click here to view a larger version of this figure.
Figure 2: Representative cytograms demonstrating enrichment of Tet Tox-binding B cells from human peripheral blood. (A) Cytograms of enriched and depleted Tet Tox-reactive B cells from Fraction A and total Tet-Tox-reactive B cells from Fraction D from a normal subject last vaccinated >10 years previously. (B) Cytograms of PBMCs from the "positive" populations from a subject vaccinated ~3 years previously and enriched as in A (Fraction A), after pre-incubation with excess of 50 uM unlabeled Tet-tox (Fraction B), or enriched with Tet-Tox-biotin omitted during the procedure (Fraction C). All analyses were gated on CD19+ lymphocytes. Percentages represent the percent of Tet-Tox-binding cells of all B cells (CD19+) in the final fraction. Please click here to view a larger version of this figure.
Figure 3: Examples of post-enrichment analysis of antigen-binding cells. (A) Representative flow cytometric analysis of cell phenotype following enrichment of Tet-Tox-binding B cells. Cytograms depict identification and analysis of Tet-Tox-reactive naïve (CD27–), memory (CD27+), and plasmablasts (CD27++ CD38++) in B cell subpopulations of a subject before and 7 days after booster vaccination against tetanus. All cells were gated on CD19+ lymphocytes from the enriched "positive" populations. (B) Representative ELISPOT analysis of antibody secreting cells in enriched, depleted, and total unenriched B cell fractions from peripheral blood of a healthy subject 7 days after boost. For the ELISPOT assay, wells of a 96 well plate were first coated with a 10 µg/ml solution of tetanus-toxoid. Two-fold serial dilutions of cell suspensions were made in wells (across the plate) starting with cells at equal concentration from the enriched, depleted, and total unenriched fractions. Anti-tetanus antibody secreting cells were detected with biotinylated anti-human IgG (H&L), followed by Streptavidin-AP. The plate was developed with ELISPOT development buffer and the reaction was stopped by washing the plate three times with double-distilled H2O. The number of spots at a cell dilution in the linear range was determined, and the number of ASCs was calculated. Please click here to view a larger version of this figure.
Here we describe a novel method to accomplish isolation and enrichment of antigen-binding B cells from human peripheral blood. The method is readily applicable to mice and to other tissues, such as the spleen and lymph nodes, and is compatible with post-enrichment analysis of cell phenotype and function (manuscript in preparation).
The user should be cognizant of a number of variables that can affect success of this procedure. From experience dead cells tend to stick to the magnetic beads and "take up" fluorescent dyes, such as the Streptavidin-far-red-fluorescent dye, leading to non-specific increased background. It is critical to remove as many dead cells from samples as possible before beginning this procedure. Use of stains that enable live/dead discrimination to gate out dead cells from the analysis may be helpful, but presence of dead cells may compromise interpretation of assays of antibody secreting cell frequency and function following adoptive transfer.
In addition, we have found that careful titration of the biotinylated antigen added is essential to distinguishing antigen-specific and non-specific binding. Use of too little antigen will lead to underestimation of the frequency of antigen-binding cells, while too much antigen can cause false positive detection/enrichment. Use of antigen concentrations that are too high may can be indicated by apparent binding to non-B cells in the samples. Further, apparent detection of an excessively high frequency (>1%) of antigen reactive cells may be an indication of binding specificity issue. The ultimate goal when determining the optimal amount of antigen to use is to add a little bit more antigen than is present at physiological concentrations, such that the majority of the B cell receptors with an affinity of <10-6 M, for example, for the antigen are saturated. This ensures that antigen-specific B cells are being detected with an affinity that has pathogenic potential. Addition of too little antigen, i.e. at or below physiological concentrations, may lead to detection of very few antigen-specific B cells due to decreased antigen occupation of the BCRs. On the other hand, addition of too much antigen, far exceeding physiological concentrations, can lead to increased non-specific binding which will include B cells with a low affinity for the antigen (>10-6 M) and are actually ignorant of the antigen in vivo. Hence, it is vital to first titrate the biotinylated antigen across a few logs in order to be confident that only true antigen-reactive B cells are being detected. Similarly, we have found that use of too much streptavidin-far-red-fluorescent dye can increase background.
It is also important to note that biotinylation of antigen can affect availability of antigenic epitopes. Hence, optimization using different biotinylation methods may be important. Alternatively biotin can be coupled to primary amines, tertiary amines, sulfhydryls, or carboxyl groups. Use of extended spacer arm lengths can improve epitope availability by reducing steric hindrance.
One of the limitations to this technique is its reduced effectiveness for enrichment of antigen-binding B cells from cryopreserved compared to fresh cells. This could relate to greater cell death and associated "stickiness" in frozen samples. Nonetheless, the method can be applied to frozen samples but it is probably important not to directly compare results from enrichment and analysis of fresh and frozen cells. Another limitation is the purity of antigen-specific B cells in the enriched sample. From experience, there are always some cells that "tag along", which includes both non-specific B cells and T cells. This is to be expected given the rarity of the cell population on is trying to enrich. However, the purity of the cells can be increase by ensuring that a single cell suspension is obtained before adding the cells to the column, enriching a second time on a new column, or by FACS sorting the antigen-specific B cells. Depending on the type of study, the researcher can decide which option is best and whether it is necessary.
As we optimized this technique we envisioned that it would be amenable to discrimination of high affinity versus low affinity antigen-reactive cells based on antigen titration or staining intensity. However, in a previous study, we showed antigen-reactive B cells with up to a 1,000-fold difference in affinity to their antigen were equally enriched using this method7. Hence, formal measurement of the affinity of enriched cells using surface plasmon resonance analysis of expressed recombinant antigen receptors is required if knowledge of the affinity is pertinent to the study.
Inherent in this technique is the limitation that addition of the biotinylated antigen and streptavidin can activate the cells by cross-linking the BCR. This could pose a problem for particular downstream functional assays, such as those that seek to compare stimulated cells to a resting control. This problem may be mitigated by keeping control samples cold to prevent transduction of signals. It is also possible that BCR signals could affect an ongoing biological response. From experience antigen-enriched cells maintain their ability to secrete antibody, as indicated by ELISPOT analysis (Figure 3b) and function as antigen presenting cells to T cells in vitro and after adoptive transfer into mice (data not shown). A final limitation to this technique is the use of anti-fluorochome antibodies to accomplish nanoparticle binding to antigen adsorbed cells. Anti- far-red-fluorescent dye cross-reacts with certain structurally related fluorochromes, including Cy7 and Cy5.5. This can limit the options of antibody conjugates available for measurement of phenotypic markers. However, with increased availability of new fluorochromes, this is becoming less of an issue.
Using this detection and enrichment method, it is possible to easily identify antigen-binding B cells in ex vivo peripheral blood, spleen, or other tissues. This method is particularly useful in studying the immune response of antigen reactive cells following immunization or exposure to foreign antigens and in the setting of autoimmune disease. Although no method to detect antigen-reactive B cells has proven to be without its faults, this method is fast, simple, and produces a relatively high purity, yield, and enrichment of antigen-reactive cells that can be coupled to a variety of downstream assays. In addition, due to the basic principles of the method, this procedure could easily be tailored to a variety of experimental needs.
The authors have nothing to disclose.
This work was supported by grants from the JDRF (1-2008-994, 27-2012-450) and the National Institutes of Health (R01DK096492-05, R21AI124488-01, T32OD012201, and F30OD021477).
Antigen of interest | variable | variable | At least 100 μg to biotinylate easily; if using protein try to use protein that has been validated by ELISA. It must be carrier protein free. |
Biotin for labeling; e.g. EZ link Sulfo-NHS-LC-Biotin | e.g. Thermo Scientific | e.g. 21335 | Biotin is available in different formulations, such as those containing various length spacers, so the type used should be determined by the researcher |
Streptavidin-Alexa Fluor 647 | Invitrogen | S21374 | Can obtain from other suppliers. |
Anti-Cy5/Anti-Alexa Fluor 647 Microbeads | Miltenyi Biotech | 130-091-395 | |
LS Columns | Miltenyi Biotech | 130-042-401 | |
MACS manual separators | Miltenyi Biotech | variable | |
Formaldehyde | Dilute to 2% with PBS; optional if downstream assay requires live cells | ||
PBS without calcium and magnesium | |||
Ficoll-Paque PLUS | GE Healthcare | 17-1440-02 | |
Whole blood in heparinized collection tubes | |||
FACS buffer (PBS + 1% BSA + 0.01% sodium azide) | |||
Separation buffer (PBS + 0.5% BSA + 2mM EDTA) | |||
50 mL conical tubes | |||
15 mL conical tubes | |||
1.5 mL Eppendorf tubes | |||
Surface marker reactive antibodies, Fc Block, live/dead discriminating stain, if needed | |||
ELISPOT supplies, if needed |