|Home | About | Journals | Submit | Contact Us | Français|
We describe herein a protocol for the production of antigen-specific human monoclonal antibodies (hmAbs). Antibody-secreting cells (ASCs) are isolated from whole blood collected 7 d after vaccination and sorted by flow cytometry into single cell plates. The antibody genes of the ASCs are then amplified by RT-PCR and nested PCR, cloned into expression vectors and transfected into a human cell line. The expressed antibodies can then be purified and assayed for binding and neutralization. This method uses established techniques but is novel in their combination and application. This protocol can be completed with as little as 20 ml of human blood and in as little as 28 d when optimal. Although previous methodologies to produce hmAbs, including B-cell immortalization or phage display, can be used to isolate the rare specific antibody even years after immunization, in comparison, these approaches are inefficient, resulting in few relevant antibodies. Although dependent on having an ongoing immune response, the approach described herein can be used to rapidly generate numerous antigen-specific hmAbs in a short time.
This protocol is derived from strategies developed in our recent study characterizing the human B-cell response to influenza1. By this technique, it is possible for a lab experienced with the process to produce milligrams of human monoclonal antibodies (hmAbs) in as little as 28 d. This ability to express and characterize antigen-specific hmAbs is extremely useful for a variety of applications. These range from elucidating the interactions of particular antibodies and antigens to exploring basic B-cell immunology or to producing valuable therapeutics. Because of the wide epitope specificity of the antibodies produced by this method, large numbers of high-affinity antibodies can be produced quickly, yielding panels of diagnostics for rapid antigen screens.
HmAbs can be produced by several methods, including immortalization of B cells with Epstein–Barr virus2,3, and the production of B-cell hybridomas4, humanization of antibodies from other species5, using phage display libraries6 or generating antibodies recombinantly from isolated single B cells7,8. However, the technique described herein is more suited for the rapid development of a large library of antibodies with a range of specificities against a particular immunogen. In methods requiring immortalized B-cell lines, the extensive subcloning and overall shotgun approach limit the number of useful antibodies that can be produced even over extensive periods of time9. Current phage display and related platforms spend extensive amounts of time identifying the few candidate antibodies present and a significant portion of these turn out to be of low affinity9. Although phage display technology uses fully human heavy and light chain variable genes, the heavy and light chains are randomly paired in vitro, and so are more likely to induce anaphylactic responses as foreign proteins or to be auto-reactive if therapeutic uses are the goal. The mAbs generated by in vitro methods or in other species do not provide a true evaluation of the epitope specificities that humans generate in vivo, limiting the use of these techniques for applications such as epitope discovery and vaccine development or evaluation. These same applications have been hindered by technologies using immortalized B-cell lines because of the relatively few specific antibodies isolated that can be generated. Finally, for potential therapeutic applications, the Fab that is produced by phage display libraries or in other species (mice) must be cloned and fused to a human Fc backbone and expressed in a human cell line. These humanizing techniques represent a significant outlay of time and resources9.
There are limitations to this method that are balanced by the advantages. The other approaches described above (Epstein–Barr virus transformation, phage display, etc.) are memory B-cell–based hmAb technologies that allow a retrospective evaluation of the entire history of previous antigen exposures. This allows mAbs to be isolated even 80 years after exposure to the pathogen as recently illustrated by the cloning of antibodies against the 1918 influenza pandemic strain from a 95-year-old donor10. These methods negate the need for obtaining fresh samples (frozen peripheral blood mononuclear cells (PBMCs) are suitable for use) and avoid the logistical difficulties of obtaining B cells from people with active immune responses. Conversely, the power of the ASC-based hmAb approach derives from that very limitation: the approach relies on isolating activated plasmablasts at the peak of the immune response such that the majority of the hmAbs isolated are antigen specific. Thus, although a human vaccine must be available, as well as a donor to receive the vaccination and donate blood, this allows an unprecedented efficiency to generate many specific mAbs. In addition, the process provides a window directly into ongoing immune responses. For example, we have observed expansions of the ASC population during natural infections (unpublished observations). Therefore, it is likely that the procedure can be used to make antibodies from ASCs induced during or soon after natural infections, or to make anti-self antibodies from patients with certain autoimmune disorders. Finally, our method as described herein relies on transient transfection for production of the antibodies, allowing rapid screening of many antibodies but making large-scale antibody preparations difficult. However, should production need to be scaled up, methods of producing stable transfectants and manipulation of the production cell lines could easily be adapted.
The steps intrinsic to our process of producing recombinant hmAbs are a modification of a system that has been used to elucidate basic mechanisms of B-cell immunology and autoimmunity by both us11,12 and others7,8. However, the protocol described here is the first integration of this process that yields high-affinity hmAbs specific to an antigen of interest relatively quickly and critically, independent of any specific antigen staining. Thus far, this technique has produced anti-influenza, anti-anthrax toxin antibodies and anti-pneumoccocal hmAbs and can be adapted easily to any immunogen with which humans are vaccinated. For example, another group has generated anti-tetanus antibodies using variable genes from ASCs that were expressed in Escherichia coli13. This versatility can provide a library of human antibodies against targets that require a rapid response, such as bioterrorism agents or viral epidemics, or large libraries of hmAbs such as in vaccine evaluation.
There is a common misconception that active immunization precludes the need for passive immunization. Even though a vaccine is necessary to produce antibodies by the method we describe, the monoclonal antibodies produced by our system could be used as passive immunotherapy treatments in a large variety of cases. Pooled human immunoglobulin is currently used as a treatment for several agents, including hepatitis B, tetanus and rabies14. Pooled sera carry risks, including potential anaphylactic responses or autoimmune reactions, that could be avoided by a single effective neutralizing monoclonal antibody. Furthermore, in cases of a bioterrorist attack with a pathogen such as anthrax that the general public is not vaccinated against, hmAbs could provide rapid protection against the pathogen while the antibiotics begin to decrease the bacterial load. Similarly, monoclonals against toxins such as botulinum neurotoxin could aid in treating those exposed, as active immunization with vaccine will take at least 2 weeks to confer protection. Finally, there are many immunocompromised populations in which vaccines are ineffective15. In these cases, including the elderly and the very young, monoclonal passive immunotherapies could be crucial in treating infectious diseases. A final potential application is for the development of therapeutic antibodies to treat chronic or antibiotic-resistant infectious diseases. Some key examples include the substantial effort now being invested to isolate the rare broadly neutralizing antibodies that can control various strains of HIV16. Although these reagents could be used directly as an adjunct to antiviral drugs in controlling viremia, the more important application may be the ability to evaluate many of these antibodies to understand how a vaccine can elicit them. A second example is the potential to generate therapies against antibiotic-resistant bacteria directly from patients that are clearing the infection: new antibiotics are rare, but antibodies can clear these infections despite drug resistance17. Neutralizing antibodies from these patients could be used directly or pharmaceutical targeting of the neutralizing epitopes discovered could substantially increase our treatment options. In all of these cases, primarily by the ability to isolate many specific hmAbs rapidly, our technique greatly increases the potential for using monoclonal antibody therapeutics for a wide variety of infectious diseases and bioterrorist agents.
A flow chart briefly describing all stages in this protocol is shown in Figure 1.
In this protocol, antibody-secreting cells (ASCs) are first isolated from whole blood collected 7 d after vaccination with an immunogen. We have successfully made antibodies following vaccination with Fluvirin (2005–2006, 2006–2007 and 2007–2008), Pneumo-vax23 and Biothrax. PBMCs are isolated using a standard lymphocyte separation protocol. The frequency of antigen-specific ASCs is analyzed using a standard ELISpot protocol18 (see Box 1). This assay enumerates the number of IgG-producing ASCs, as well as antigen-specific ASCs. The percentage of antigen-specific, IgG-producing ASCs is a useful measure of the donor’s response to the vaccine and therefore the approximate quantity of high-affinity antibodies produced.
The cells are then sorted by flow cytometry. First, the live cell gate, including larger blasting cells, is set using forward versus side scatter. The ASCs are bulk sorted by first gating on CD19high/CD20low to neg/CD3neg and then on CD27high/CD38high cells as shown in Figure 2. The appropriate IgG, IgM and IgD gates are set to obtain IgG-producing ASCs, although it is also possible to use this method to isolate IgM-producing ASCs as well. Finally, the purified ASCs are single cell sorted into single cell PCR plates loaded with catch buffer containing RNase inhibitor.
Using both RT-PCR and nested PCR, the antibody genes in each cell are amplified on a per cell basis. The RT-PCR is accomplished using a cocktail of nine primers, designed to cover all of the families of variable (V) genes possible (Table 1). The nested PCR is performed to amplify the DNA enough to obtain sequences of the heavy and light chain V genes. This is necessary for the cloning PCR. In this step, highly specific primers for each V gene family are used to amplify the DNA for cloning. The ‘cloning PCR’ primers are designed both to incorporate the cloning restriction sites and to place the VDJ heavy or VJ light chain genes in frame with the signal peptide sequences and constant region genes within the respective cloning vectors. Cloning sites were incorporated into the vectors that are specific for the particular heavy or light chain vectors to allow proper, in-frame incorporation of the variable gene inserts. The inserts and vectors are then digested and purified for cloning. The heavy and light chain DNA from each single cell is then cloned into separate vectors and transformed. At least four colonies from the transformation are grown, mini-prepped and sequenced. The sequences from each colony are compared and the colony most closely matching the consensus is then chosen for further amplification to maxi scale.
Transiently transfected human kidney epithelial cells (the HEK293 cell line19) are used to produce the antibody. Polyethyleneimine-based transfection is used with equimolar amounts of heavy and light chain vector according to standard protocols20. The cells are allowed to produce antibody for 5 d. The transfection media containing the hmAbs are then purified using protein A agarose beads and concentrated using commercial protein concentrators. During the final stage, the hmAbs are analyzed for concentration, purity and reactivity.
!CAUTION Toxic; carcinogen.
! CAUTION Highly toxic.
As indicated in Supplementary Figure 1 online, the expression vectors contain a murine immunoglobulin signal peptide sequence and variable-gene cloning sites upstream of the appropriate human immunoglobulin constant regions followed by an SV40 polyadenylation sequence. Transcription is under the HCMV (human cytomegalovirus immediate-early) promoter and clones are selected based on ampicillin resistance. The antibody variable-heavy and variable-light rearranged genes from each single cell are cloned into the respective vectors in frame with the signal peptide and constant region genes. These vectors are then co-transfected into the 293A cell line for expression. The resultant antibodies are properly trafficked and secreted after cleavage of the signal peptide, resulting in fully human IgG/κ or IgG/λ amino-acid sequences. The vector sequences are available through the NCBI GenBank (accession numbers: FJ475055, FJ475056 and FJ517647), and the vectors themselves are available upon request.
An aliquot of 250 ml each of sterile RPMI and DMEM; 3.75 ml of antibiotic/antimycotic and 5 ml each of L-glutamine (200 mM), 100× Nutridoma and sodium pyruvate (100 mM) was used. Basal media must be made fresh every 7 d. L-Glutamine can be stored at −20 °C for up to 1 year, Nutridoma can be stored at room temperature (20–25 °C) for up to 1 year and sodium pyruvate can be stored for up to 6 months at 4 °C.
Here 0.1 M glycine solution equilibrated to pH 2.7 with 12 M HCl and filter sterilized. Solution can be stored up to 60 d at room temperature.
Here 1 M Tris solution equilibrated to pH 9.0 with HCl and filter sterilized. Solution can be stored up to 60 d at 4°C.
Here 0.15 M NH4Cl, 10 mM KHCO3 and 0.1 mM Na2EDTA. Adjust pH to 7.2–7.4 with 1 M HCl and filter sterilized. Solution can be stored up to 1 year at room temperature (20–25 °C).
Dissolve LB agar in dH2O according to package directions and autoclaved. When cooled to 45 °C, add 50 μg ml−1 ampicillin. Dispense 20–25 ml agar solution into 100 mm × 15 mm petri dishes. Cool and store at 4 °C for up to 6 months.
Prepare AEC stock (20 mg ml−1 AEC in dimethylformamide). Dilute AEC from stock to 0.3 mg ml−1 in 0.1 M sodium acetate buffer (pH 5.0) just prior to use. Filter sterilized with a 0.45-μm syringe filter. The stock solution may be made and stored for up to 2 months. The diluted solution must be made fresh each time used.
To 5 ml of RNAse-free water, add 50 μl of 1 M Tris pH 8.0 and 125 μl of Rnasin. Keep on ice. This makes enough for 10 half plates. Catch buffer must be made fresh each time used.
It was prepared by 1 mg ml−1 PEI in 80 °C dH2O. Adjust pH to 7.2 with HCl. Filter sterilize with a 0.45-μm syringe filter. Store at −20 °C for up to 1 year.
For 10 ml of a 10% SDS-PAGE resolving gel (vol/vol), combine 3.3 ml 30% acrylamide mix (wt/vol), 2.5 ml 1.5 M Tris (pH 8.8), 100 μl 10% ammonium persulfate (wt/vol) and 4 ml water. Mix well, add 4 μl TEMED and mix again. Pour into gel casting apparatus. When 10% gel is set (approximately 30 min), make the 5% stacking gel. For 5 ml, combine 830 μl 30% acrylamide mix (wt/vol), 630 μl 1.0 M Tris (pH 6.8), 50 μl 10% SDS (wt/vol), 50 μl 10% ammonium persulfate (wt/vol) and 3.4 ml water. Mix well, add 5 μl TEMED and mix again. Pour on top of resolving gel. 1% (wt/vol) agarose gel Dissolve 0.3 g of agarose in 30 ml of boiling Tris-acetate EDTA solution. Cool slightly, add 2 μl of ethidium bromide solution. Pour into gel caster.
|Reagent for PCR||Volume (μl) for each (25 μl sample)||Final concentration (with template)|
|Taq DNA Polymerase (added last)||0.25||50 U ml−1|
|dNTPs (10 mM each, combined)||0.5||200 μM|
|Forward primer: VH3a and VH3b or PanVκ||0.5||1.2 μM|
|Reverse primer: PW-Cgamma or CK494-516||0.5||1.2 μM|
|dH2O (nuclease free)||17.25–19.25 (to 24 μl total volume)||—|
|Reagent for PCR||Volume (μl) for each (25 μl total)||Final concentration (with template)|
|Taq DNA polymerase (added last)||0.25||50 U ml−1|
|dNTPs (10 mM each, combined)||0.5||200 μM|
|5′ AgeI—VH or VK primer||0.5||1.2 μM|
|3′ SalI-JH or 3′ BsiWI-JK primer||0.5||1.2 μM|
|dH2O (nuclease free)||19.75||—|
Steps 1–9, lymphoprep and B-cell enrichment: 2 h
Box 1, ELISpot: ≥30 h
Steps 10–20, staining and flow cytometry: ~5 h
Steps 21–28*, RT, nested and cloning PCRs: ~3 d
Step 29, PCR purification: 10 min
Steps 30–33, first digestion of gamma, kappa, or lambda chains: 5–20 h
Step 34, digestion purification: 10 m
Steps 35–37, second digestion: 5–20 h
Steps 38 and 39, gel purification: 1 h
Steps 40–42, ligation: 2.5–18 h
Steps 43–46, transformation: 2 d
Steps 47–49*, miniprep: 2 d
Steps 50–54, maxiprep: ≥34 h
Steps 55–63, transfection: 5 d
Steps 64–78, protein purification: 7–24 h
Step 79, protein quantification: 2–3 h
Step 80, protein qualification: 5 h
*Assumes overnight turnaround on DNA sequencing or synchronization of clones processed so that sequencing delays are avoided.
In the examples provided herein, two quite different vaccine formulations (Fluvirin or Pneumovax23) were used to generate mAbs illustrating the similar utility of this procedure. Fluvirin is primarily influenza HA and NA proteins, whereas the Pneumovax23 is produced from highly purified capsular polysaccharides from the 23 most prevalent or invasive pneumococcal types of Streptococcus pneumoniae21. When purifying PBMCs from 30 ml of blood 7 d after vaccination, it should be possible to isolate several thousand IgG-positive ASCs. Because the single cell sorting process is highly efficient, typically 6–10 half plates can easily be sorted in this manner per donor. A half plate of cells (42 wells because Row H of the plate is left open for controls) yields about 20 antibodies. When the light chain of interest is kappa, typically, 70% of the antibodies will be kappa positive, the remaining being lambda and thus unamplified (29 antibodies). Of these, the heavy chain PCR efficiency is also close to 70% (20 antibodies). These RT-PCR efficiencies likely arise from a variety of factors, including stability of the RNA template from only a single cell, calibration of the flow cytometer to err on the side of having no cell rather than two cells within a well and limitations of the PCR that we have never overcome (such as occasional V genes that are not primed by the set of primers). Generally, several antibodies will not PCR correctly from the cloning PCR and a few others will be lost through the cloning process. We have found that for anti-influenza antibodies, approximately 40% of the ASCs were clonally related (from the same progenitor B cell) but with their antibody sequences differing by accumulated somatic hypermutations.1 Other vaccines or acute immunizations may have more or less clones. In our hands, even with variations in the frequency of mutations, two antibodies from the same clonal expansion are quite similar in binding characteristics. Thus, unless relevant to the experiment only one of a clone need be expressed. The end yield of transfectable antibodies containing a valid heavy and light chain will be 10–16 per half plate of cells sorted. Almost all will yield enough antibody (>50 μg ml−1) upon transfection for use in further assays. As an example, when using either Fluvirin or Pneumovax23, an average of 70% of the antibodies bound to the immunizing antigen(s) as measured by ELISA assays (see Fig. 3 and ref. 1). For typical antibodies, transfecting four plates of 293A cells will yield a final concentration of 100–500 μg ml−1 of purified antibody.
Variations in the immune systems of the donors utilized can cause variations in the yield of antibodies produced. Because of these variations, the ELISpot procedure is a valuable adjunct to the antibody production procedure. Certain donors will respond poorly to the vaccination, perhaps having only 10–30% antigen-specific ASCs, thus the yield of antigen-specific antibodies will also be low. When attempting to make antibodies to a new vaccine, the ELISpot results will accurately predict the final yield of antibodies.
We thank Ken Wilson, Matt Jared and Christina Helms for their technical efforts. We also thank Jennifer Morris and Christina Helms for their help with editing and formatting. This study was funded, in part, by NIH grant numbers HHSN266200500026C (P.C.W.) and P20 RR018758 (P.C.W.), NIH/National Institute of Allergy and Infection Diseases (NIAID) U19-AI057266-04 (R.A.), NIH/NIAID HHSN266200700006C Center of Excellence for Influenza Research and Surveillance (R.A.) and NIH/NIAID N01-AI-50025-02 (R.A.). J.W. was supported by a postdoctoral fellowship from The Swedish Research Council.