MABS 2016, VOL. 8, NO. 5, 916–927 http://dx.doi.org/10.1080/19420862.2016.1170263

REPORT

A novel antibody discovery platform identifies anti-influenza A broadly neutralizing antibodies from human memory B cells Xiaodong Xiaoa, Yan Chena, Reena Varkeya, Nicole Kallewaardb, Adem C. Koksala, Qing Zhub, Herren Wua, Partha S. Chowdhurya, and William F. Dall’Acquaa a Department of Antibody Discovery and Protein Engineering, MedImmune, Gaithersburg, MD, USA; bDepartment of Infectious Diseases and Vaccines, MedImmune, Gaithersburg, MD, USA

ABSTRACT

ARTICLE HISTORY

Monoclonal antibody isolation directly from circulating human B cells is a powerful tool to delineate humoral responses to pathological conditions and discover antibody therapeutics. We have developed a platform aimed at improving the efficiencies of B cell selection and V gene recovery. Here, memory B cells are activated and amplified using Epstein-Barr virus infection, co-cultured with CHO-muCD40L cells, and then assessed by functional screenings. An in vitro transcription and translation (IVTT) approach was used to analyze variable (V) genes recovered from each B cell sample and identify the relevant heavy/light chain pair(s). We achieved efficient amplification and activation of memory B cells, and eliminated the need to: 1) seed B cells at clonal level (1 cell/well) or perform limited dilution cloning; 2) immortalize B cells; or 3) assemble V genes into an IgG expression vector to confirm the relevant heavy/light chain pairing. Cross-reactive antibodies targeting a conserved epitope on influenza A hemagglutinin were successfully isolated from a healthy donor. In-depth analysis of the isolated antibodies suggested their potential uses as anti-influenza A antibody therapeutics and uncovered a distinct affinity maturation pathway. Importantly, our results showed that cognate heavy/light chain pairings contributed to both the expression level and binding abilities of our newly isolated VH1-69 family, influenza A neutralizing antibodies, contrasting with previous observations that light chains do not significantly contribute to the function of this group of antibodies. Our results further suggest the potential use of the IVTT as a powerful antibody developability assessment tool.

Received 17 December 2015 Revised 17 March 2016 Accepted 19 March 2016

Introduction Antibody isolation directly from human B cells has distinct advantages in harnessing rare antibodies with desirable functions. Following the early successes with the isolation of monoclonal antibodies (mAbs) 4E10 and 2F5,1,2 there has been a rapid growth in the number of potent HIV-neutralizing antibodies as a result of the application of B cell-based platforms.3 Potent antibodies targeting severe acute respiratory syndrome coronavirus,4 influenza virus,5-7 respiratory syncytial virus8 as well as other viruses9 have also been isolated. In addition, B cell-derived antibodies against human self-antigens have helped in the understanding of autoimmune diseases.10 These advances highlight the potential of B cell-based antibody discovery platforms, while underscoring the technical challenges in using such a strategy.11 Unlike from rodent B cells, hybridoma generation from human B cells has faced various difficulties, and alternative approaches have been sought after.12 Human antibody discovery using primary B cells faces 2 main obstacles. The first is the ability to maintain and screen the antibody-producing B cells. The primary approaches to overcome this challenge are B-cell immortalization and transient B-cell activation (reviewed in ref. 12). These strategies remain topics of active research because successful studies adopted methods that were often proprietary.13 The second

Founder mutation; germline; HA (hemagglutinin); IVTT (in vitro transcription and translation); influenza broadly neutralizing antibody; P52aG; VH1-69

obstacle is the capacity to recover antibody genes from as few as one cell. Technologies have advanced sufficiently to allow such a practice, but recombinant IgG cloning and recombinant expression procedures are labor intensive and time consuming, especially when the number of samples needing V gene rescue is large. This necessitates clonal B cell culture or single B cell sorting prior to V gene recoveries by all current protocols.13 Memory B cell immortalization by Epstein-Barr virus (EBV) infection is a low efficiency process that involves a balancing act among many regulatory elements.14 Clone losses are common prior to achieving the true “immortalization.” However, the initial outgrowth stage after infection is robust and should be long enough for screening B cells of interest. We have found that the number of candidate B cell samples needed for gene recovery could be reduced through a carefully designed screening strategy, and that it is not always necessary to achieve immortalization when EBV is used. Following this strategy, we developed a new platform that allows functional screenings of EBV-activated B cells seeded in non-clonal format, which can then be followed by in vitro transcription and translation (IVTT) Fab expression to quickly identify the functional heavy/light chain pairs. The IVTT Fab expression procedure does not require the assembly of a single operon containing both heavy chain

CONTACT William F. Dall’Acqua [email protected] Partha S. Chowdhury and Protein Engineering, MedImmune, One MedImmune Way, Gaithersburg, USA © 2016 Taylor & Francis Group, LLC

KEYWORDS

[email protected]

Dept. of Antibody Discovery

MABS

917

(HC) and light chain (LC), an attrition-causing step, or the construction of expression vectors. These attributes should lead to significantly increased efficiencies. Influenza virus infections continue to be a health threat and economic burden despite decades of vaccine and therapeutics development.15,16 B cell-based platforms and phage panning both have contributed to the identification of broadly protective antibodies.5,17-20 We attempted to validate our platform by recovering anti-hemagglutinin (HA) neutralizing antibodies from a healthy donor. This effort led to the isolation of several broadly neutralizing antibodies, one of which utilizes a distinct affinity maturation pathway that has not been reported previously. Furthermore, this platform revealed that accurate heavy and light chain pairings may be an important step in the assemblies of selected antibodies. This platform can therefore be used as a unique tool to assess the developabilities for some antibody therapeutics.

Results Design of the platform Clonal B cell culture achieved through immortalization and limited dilution cloning (LDC) preceding V gene recovery poses significant technical challenges due to the high attrition rate, and has been performed routinely by only a limited number of groups.4,8 Transient activation of B cells followed by functional screening and V gene rescue protocol yielded the first 2 of a group of broadly neutralizing anti-HIV antibodies, but it demanded “near clonal density” (1.3 cells/well) of B cells at seeding to assist V gene rescue.21 To circumvent those restrictions, we adopted 2 critical improvements to develop our platform. The first is the initiation of gene recovery without achieving B cell immortalization or LDC to reduce the cell loss and save time, and the second is the V gene matrix pairings through a novel IVTT platform to allow non-clonal B cell seeding. The experimental procedure is summarized in Fig. 1a. In a typical campaign, a 7-step protocol will yield a sufficient amount of antigen-specific, recombinant IgG for in-depth functional analysis in an estimated 1–1.5 month time frame. The IVTT step (step 6), which is unique to this platform, is illustrated in detail in the subsequent V gene recoveries from sample 1N23. During IVTT template assembly, a 2-step nested PCR ensures maximum V gene recovery efficiency and the addition of T7 promoter and ribosomal binding site (RBS). The third PCR step allows the attachments of: 1) CH1 to the VHs and CK and Cλ to the VLs, and 2) translational terminator to the 30 of the templates (Fig. 1b). This step also adds c-Myc and Flag affinity tags to the LC and HC, respectively allowing Western blot or ELISA detections. The primers used for PCR amplification of V genes were designed based on the IMGT® database (www.imgt.org/) with the inclusion of appropriate sequence degeneracies to cover all major V genes and are listed in Table 1. V gene rescue efficiency of 80% or better was achieved consistently on sorted single memory B cells from various healthy donors, and sequencing of randomly selected, rescued VH, VK, and Vλ confirmed their uniqueness and diversities (data not shown).

Figure 1. (a) Flow chart of a typical antibody discovery campaign using human B cell sources and the IVTT platform. The time needed for each step is estimated based on current protocol. (b) Schematic representation of the assembly of the IVTT template via a 3 step PCR process for Fab expression. In the first 2 steps the VH, VK, and Vλ fragments are amplified by a nested PCR with the T7 promoter and ribosome binding site attached to the 50 end. An additional overlapping PCR step adds the constant region (CH1, CK or Cλ), an affinity tag and the T7 terminator at the 30 end to generate a complete template for IVTT Fab expression.

Establishment of in vitro transcription and translation of Fab IVTT was used previously for a-glycosylated Fab and IgG productions.22 We optimized the critical parameters for the IVTT platform by using known antibodies so this platform can be suitable for rapid screening of cognate HC and LC pairs (Fig. 2a, b). The amount of template, inclusion of disulfide bond enhancer, and reaction temperature were determined for maximum Fab yield. However, for screening purposes »125 ng of each of the HC and LC templates was used routinely because this amount was easily achievable through the RT-PCR process and the Fab yield was not significantly improved by using more templates (Fig. 2a). The yield of the Fab was typically 1–10 mg/ml in a 25 ml

918

X. XIAO ET AL.

Table 1. Primers for IVTT template assembly. VH-PCR1 50 VH-1 50 VH-2 50 VH-3 50 VH-4 50 VH-5 50 VH-6 50 VH-7 50 VH-8 50 VH-9 VK-PCR1 50 VK-1 50 VK-2 50 VK-3 50 VK-4 50 VK-5 50 VK-6 50 VK-7 VL-PCR1 50 VL-1 50 VL-2 50 VL-3 50 VL-4 50 VL-5 50 VL-6 50 VL-7 50 VL-8 50 VL-9 50 VL-10 PCR2 50 HL2

PCR3

50 HL3

GGAGGAAAAAATATGGCC CAG RTG CAG CTG GTG CAR T GGAGGAAAAAATATGGCC SAG GTC CAG CTG GTR CAG T GGAGGAAAAAATATGGCC CAG RTC ACC TTG AAG GAG T GGAGGAAAAAATATGGCC SAG GTG CAG CTG KTG GAG GGAGGAAAAAATATGGCC CAG GTG CAG CTA CAG CAG T GGAGGAAAAAATATGGCC CAG GTA CAG CTG CAG CAG TC GGAGGAAAAAATATGGCC GAR GTG CAG CTG GTG CAG GGAGGAAAAAATATGGCC CAG STG CAG CTG CAG GAG TC GGAGGAAAAAATATGGCC GAG GTG CAG CTG TTG GAG TCT GGAGGAAAAAATATGGCC GAC ATC CAG WTG ACC CAG TCT GGAGGAAAAAATATGGCC GAT ATT GTG ATG ACC CAS ACT GGAGGAAAAAATATGGCC GAA ATT GTG WTG ACR CAG TCT GGAGGAAAAAATATGGCC GAT RTT GTG ATG ACW CAG TCT GGAGGAAAAAATATGGCC GAA ACG ACA CTC ACG CAG TCT GGAGGAAAAAATATGGCC GAA ATT GTG CTG ACT CAG TCT GGAGGAAAAAATATGGCC GMC ATC CRG WTG ACC CAG TCT CC GGAGGAAAAAATATGGCC CAG TCT GTG YTG ACK CAG CCR GGAGGAAAAAATATGGCC CAG TCT GCC CTG ACT CAG CCT GGAGGAAAAAATATGGCC TCC TAT GWG CTG ACT CAG CYA GGAGGAAAAAATATGGCC CWG CCT GTG CTG ACT CAG CCM GGAGGAAAAAATATGGCC TCT TCT GAG CTG ACT CAG GAC GGAGGAAAAAATATGGCC AAT TTT ATG CTG ACT CAG CCC GGAGGAAAAAATATGGCC CAG YCT GTA CTG ACT CAA CCG GGAGGAAAAAATATGGCC CAG RCT GTG GTG ACY CAG GAG GGAGGAAAAAATATGGCC CAG CYT GTG CTG ACT CAA TC GGAGGAAAAAATATGGCC CAG GCA GGG CTG ACT CAG CCA CTAGGCGAATTAATACGACTCACTATAGGGCTTAAGTATAAGGAGG AAAAAATATGGCC

GCGAATTAATACGACTCACTATAG

reaction volume, which is sufficient for several rounds of ELISA-based preliminary screening. Mimicking a higher throughput process, we used the platform first to recover V genes from sorted single B cells in a plate format (Fig. 2c). We then assembled the IVTT in a pairwise fashion between the VH and VL from each B cell directly from the RT-PCR products. 80% of the samples displayed robust Fab expression falling into the range observed with known antibodies (Fig. 2d). Memory B cell activation and maintenance for functional screenings EBV infection was proposed as a useful tool for human B cell immortalization.23 Although it has been attempted by many groups, the outcomes have been inconsistent and mostly poor. Modifications are constantly being made to increase the “immortalization” rate.4 We used irradiated CHO-muCD40L instead of allogeneic peripheral blood mononuclear cell (PBMC) as the feeder layer to: 1) provide consistent CD40L stimulation,24 2) prevent a potential T cell response,25 and 3) eliminate false positive signal from the PBMC feeder layer during ELISA and molecular cloning. At four memory B cells C 10,000 CHO-muCD40L cells/well seeding density in a 384-well plate, we observed robust outgrowth and B cell cluster formation in 85% of the wells 10 d post seeding, and detected human IgG production in 100% of the wells, including those without cluster formation but live B cells. The supernatant from each well was used in ELISA for HA binding.

30 CH1 GGAAGTAGTCCTTGACCAGGCAG

30 CK1

CAGCAGGCACACAACAGAGG

30 CL1

GTGTGGCCTTGTTGGCTTGAAG

30 JH-1 30 JH-2 30 JH-3 30 CK2 30 CL2 30 HL3

GAAGACGGATGGGCCCTTGGTCGACGC TGA GGA GAC RGT GAC CAG GGT GAAGACGGATGGGCCCTTGGTCGACGC TGA AGA GAC GGT GAC CAT TGT GAAGACGGATGGGCCCTTGGTCGACGC TGA GGA GAC GGT GAC CGT GGT TTCCAGATTTCAACTGCTCATCAG GTTGGCTTGAAGCTCCTCAGAGGA AAACCCCTCCGTTTAGAGAGG

From 9 (plates) £ 384 (wells/plate) £ 4 (cells/well) samples screened (»14,000 memory B cells), 34 HA (H1, CA09) positive wells were identified in the first screening. This represents a »10¡3 HA-positive rate, which is in line with similar surveys using healthy donors prior to receiving vaccination.26,27 Half of the cells were collected for V gene recovery, and the other half was transferred to a 96-well plate for continued culturing and monitoring. We observed a steady decrease in antigen-specific IgG production over time, with 50% of the original wells losing the HA binding activity after 3 months of culture. Well 1E18 displayed robust outgrowth and maintained HA reactivity throughout the process. However, attempts to generate clonal culture through LDC were not successful. These observations underscored the importance of performing gene recovery at early time points and the challenges to generate truly “immortalized” B cell clones by EBV infection. 28,29 V gene recovery from wells of interest Influenza A viruses are serologically divided into different HA subtypes, which are subdivided into 2 distinct phylogenetic groups, group 1 and group 2. Currently, H1 and H3 HA subtypes are associated with seasonal human disease, although other more diverse HA subtypes are thought to be of concern for future pandemics.5 Single-point ELISA screening of the 34 HA (H1, CA09) positive wells against an expanded panel of recombinant HAs indicated that a few of them were broadly reactive against influenza A HA proteins. We failed to observe any sample well that was

MABS

919

Figure 2. Optimization and evaluation of the IVTT platform for Fab production. For IVTT optimization, a known antibody was used and the Fab production was evaluated in a direct antigen binding ELISA. (a) Correlation between the amount of IVTT template input and the Fab yield. (b) Temperature and disulfide bond enhancer are critical factors for higher Fab yield. In both a and b, antigen was coated on an ELISA plate and the bound Fab was detected using anti-Fab-HRP. (c) Single B cells were sorted based on IgGCKappaC or IgGClambdaC into 96-well PCR plates containing the lysis buffer from the RT-PCR kit. Each well was subjected to VH and VL recovery as described in Methods. Randomly picked RT-PCR products were sampled on agarose gels to assess the gene recovery efficiency. All samples (VH, VK, Vλ) were picked randomly from different 96-well plates and no sample-sample correspondence (HC/LC pairing, etc.) was intended by way of corresponding locations on the 3 agarose gels. (d) For IVTT evaluation, 60 pairs of HC and LC IVTT templates assembled from 60 sorted single B cells were used directly in Fab production in a plate format. The single data points represent the sandwich ELISA signal from each IVTT sample after 30 fold dilution. For the sandwich ELISA to detect appropriately assembled Fab level, heavy chains were captured by an anti-Fd antibody, and light chains were detected by HRP-conjugated anti-kappa or lambda antibodies. Over 80% of the samples showed robust Fab expression.

cross reactive to both influenza A and B isolates, confirming the rarity of such B cells. Eight of the 34 were reactive against H5, with 5 of those also recognizing H2 and H6 proteins. Among all samples tested, wells 1E18 and 1N23 displayed the most robust pan group 1 HA activities and well 2H4 was the only one that recognized all group 1 and 2 HAs tested. Based on these observations, wells 1E18, 1N23, 2H4, and a mono-reactive well, 7G16 were chosen as test cases for antibody recovery using the new platform. Cell pellets collected immediately after the HA binding screenings as described in the “activation and screening” section were applied to the IVTT platform. After confirming the successful HC and LC template assembly (Fig. 3a) and HA binding activities by IVTT products from each well against HA (H1, CA09) (Fig. 3b), we proceeded to identify the original HC/LC pair by performing matrix pairings. Although we set out to seed memory B cells at 4 cells/ well, TA cloning and sequencing indicated that between 3– 13 VHs had been recovered from the wells of interest due to variations in manual operation. There existed a reverse correlation between the number of V genes recovered and the ELISA signal in the first round of IVTT from each sample well. Wells 7G16, 1E18, and 1N23 each contained 3, 4, and 11 unique VHs, respectively. This might have contributed to

the weaker IVTT signal from well 1N23 compared to those of 7G16 and 1E18 (Fig. 3b). V gene recovery from well 1N23 is presented as an example, illustrating in detail how matrix pairings are used in step 6 of the platform (Fig. 1a). There were 11 VH and 7 VKs in well 1N23 and they were grouped according to their subfamilies (Fig. 3c left panels) and paired in the first step of IVTT (IVTT1) (Fig. 3c right panel). We recovered more VHs (11) than VLs (7) from sample well 1N23; this could be attributed to the fact that we did not attempt to rescue lambda LCs due to the lack of contribution to the binding by lambda LCs (Fig. 3b). After identifying the positive groups from reaction # 2 pairing groups B and d (Fig. 3c, right upper panel), we proceeded to pair the individual heavy and light chains from the positive B and d groups in a second step (IVTT2). The exact pairing between HC #5 and LC #3 was resolved in 2 steps involving fewer than 20 IVTT reactions. The increase in specific ELISA signal was significant from step 1 to step 2 as the specific HC/ LC pair was identified (Fig. 3c, lower panel). Wells 7G16 and 1E18 were subjected to the same IVTT process, and the HC/ LC pairs recapitulating the parental well activities were identified in one step (data not shown). Well 2H4 recovery was atypical as it had a very poor KD and 13 unique VHs leading to weak IVTT signals. In order

920

X. XIAO ET AL.

Figure 3. Identification of the original HC/LC pairs through the IVTT. (a) Agarose gels showing the assembled HC and LC templates from the wells of interest (#1–8). Note that from each well both kappa and lambda chains were recovered. (b) IVTT products from 8 HA positive wells were tested in an antigen binding ELISA. Both kappa (red bar) and lambda (blue bar) Fabs were assembled in the IVTT and tested. Wells #1–7 contain kappa functional clones and well #8 has a lambda functional clone. Wells corresponding to 1E18 (well #2), 1N23 (well #3), and 7G16(well #7) were indicated. 2H4 well was recovered from a separate experiment and not shown here. (c) HC and LC were grouped according to their germline subfamilies and used in the first IVTT experiment (IVTT1) to narrow down the search of the original HC/LC pair to groups B and d. This was followed by a final step of matrix pairing using HCs and LCs from groups B and d (IVTT2) to identify the original HC/LC pair as VHseq5/VLseq3.

to recover this rare cross-reactive clone, we applied less stringent screening criteria (e.g., the ELISA signal was counted as positive if it was 50% above negative control instead of the typical 3 folds (300%) in the first step), and

succeeded in identifying the functional VH and VL pair in 2 steps involving 14 IVTT reactions. Its recovery process suggested that the B cell number/well should not exceed a limit if high recovery rate is preferred over throughput.

MABS

Anti-HA antibodies display cross binding and broadly neutralizing activity by targeting a conserved epitope The identified V genes were used to construct IgG1 expression vectors for all 4 new mAbs for further functional analysis. Recombinant IgG1s 1E18 and 1N23 displayed broad binding activity against group 1 HAs, whereas IgG1 2H4 showed activities against all group 1 and 2 HAs tested, albeit with poor affinity (Fig. 4a). Their KDs against the HA from H1 (CA 09) isolate were assessed using a bio-layer interferometry (BLI) assay. Both mAbs 1E18 and 1N23 displayed rapid on-rate and slow off-rate resulting in excellent KDs (0.09 and 0.1 nM, respectively). mAb 2H4 had a slow onrate and a fast off-rate, leading to a very poor KD (349 nM, Fig. 4b and Table 2). These KDs correlated directly with their neutralizing ability in a micro-neutralization assay. Recombinant IgGs 1E18 and 1N23 neutralized all tested group 1 isolates effectively, with activities comparable to that of FI6v3, a broadly neutralizing antibody isolated previously from a plasma cell after vaccination (Fig. 4c).5 In contrast, IgG1 2H4 did not show significant neutralization against any of the viruses tested. However, at the highest antibody concentration used (100 mg/ml), we did see modest reductions (< 50%) in the viral growth of some of the isolates (data not shown). In an epitope binning experiment, mAbs 1E18, 1N23, and 2H4 appeared to share the similar epitope as FI6v3,5 which recognizes a highly conserved epitope in the HA2 stem region (Fig. 4d). mAb 7G16, an antibody that was isolated through the same platform but displayed binding to only one HA (H1 CA 09) recognized a different epitope.

Sequence alignment revealed distinct antibody features Sequence analysis revealed that mAbs 1E18, 1N23, and 2H4 each belonged to a unique group of influenza A neutralizing antibodies. mAbs 1E18 and 1N23 both use a Vh1–69 heavy chain, a VH family preferred by a large number of influenza A, group 1 broadly neutralizing antibodies, including F10 and CR6261 isolated from phage panning, and a series of others directly from B cells17,18,30,31 (Fig. 5). Their light chains belong to the family most preferred by Vh1–69, Vk3.32 Unlike mAb 1E18, however, mAb 1N23 does not have 98Y in HCDR3, which is seen in the majority of Vh1–69 stem binding antibodies.30,31 mAb 2H4 uses a VH subfamily Vh3–483 as compared to FI6, which is cross reactive to both group 1 and 2 isolates and uses Vh3–30.5 mAb 2H4 has significantly fewer somatic hypermutation (SHM) compared to FI6 (Table 3). Recent studies have highlighted the importance of a tripartite of residues, 98Y (HCDR3), P52aA (HCDR2), and 54F (HCDR2), which represent the respective roles of VDJ recombination, founder mutations, and polymorphism in the development of VH1–69 using HA stem binding broadly neutralizing antibodies.30,31 Notably, mAb 1E18 has a 98Y, but a P52aG mutation instead of a P52aA, which was found to be predominant in all 98Y using influenza A group 1 neutralizing antibodies.31 Strikingly, mAb 1N23 does not possess 98Y and has a P52aG mutation, a combination that has not previously been described in the maturation process.

921

mAb 1N23 represents a new affinity maturation branch To understand the significance of 52a mutations in the maturation process of the newly identified antibodies, we created the germline predecessors of both mAbs 1E18 and 1N23 and their variants carrying the mutations (Fig. 5). Both P52aA and P52aG were capable of restoring the germline antibody binding to several HAs tested, confirming their roles as the “founder mutation” (Fig. 6, upper 2 panels). For mAb 1E18, “G” was more effective than “A,” supporting a modeling-based prediction that a smaller residue at position 52a would be beneficial.30 We then investigated if this finding is true for 98Y anti-HA antibodies identified from another study. In the absence of UCA9 light chain information, we constructed UCA9H/1E18L hybrid germline predecessor, its mutant variants and tested them in the same ELISA assay.31 We found that “P52aG” was also equally or more efficient than “P52aA” in restoring the germline antibody bindings by UCA9H/1E18L (Fig. 6, lower panel). Significantly, P52aA and P52aG both restored germline 1N23 binding effectively. This is an observation that has not been made with any non-98Y antibody. This result necessitates the addition of a new affinity maturation sub-branch typified by the non-98Y usage and a P52aG mutation to those of VH1–69, HA stem binding antibodies proposed in a previous publication.31

The HC/LC pairing is regulated at the Fab assembly level It has been reported that broadly influenza neutralizing antibodies arising from the VH1–69 family displayed no LC contributions to their binding activities.18,31 However, when identifying the unique LCs for mAbs 1E18 and 1N23 during the matrix pairings, we saw preferential LC pairings. In the first round of IVTT before conducting matrix pairings, LC usages for sample wells 1E18, 1N23, and a few others were either kappa or lambda, but not both (Fig. 3b). While pairing the mAb 1N23 HC with all kappa LCs recovered from the same sample well, we found that the Fab expression level varied significantly, but the functional pair was the highest (Fig. 7a). The dramatic differences in the fully assembled Fab level were not due to the expression levels of individual heavy and light chains, as all the chains were expressed at comparable levels when measured by Western blot (Fig. 7b). Furthermore, all the LCs also paired equally well with another irrelevant HC #6 (VH1–2 subfamily) in assembling into Fab (Fig. 7c). We then tested several pairs in the IgG format. IgG expression retained the same trend even though the differences were more moderate (Fig. 7d). Importantly, the “mismatched” LC also led to a reduction in binding ability, in contrast to previous reports 18,31 (Fig. 7e). It is worth noting that LCs #3 and #4 are from the same Vk3 germline family (Vk3–20) and only differ in the SHM, whereas LCs #1 and 2 are from the subfamilies Vk1–9 and Vk1–5, respectively. The differences in LCs #3 and 4 did not affect IgG binding, yet contributed to different Fab and IgG productions. The biased Fab expressions for the functional pairs were also observed during mAb 1E18 and 2H4 V gene recoveries (data not shown). To investigate the contribution of the LC to similar antibodies from other donors, we generated UCA5H/1N23L hybrid

922

X. XIAO ET AL.

Figure 4. Newly identified mAbs recognize a highly conserved, neutralizing epitope. (a) Recombinant IgGs 1E18, 1N23, and 2H4 were tested in ELISAs against indicated HA antigens. One HA from a B isolate Brisbane was included as a control. (b) The binding kinetics of recombinant IgGs were estimated using a Bio-Layer Interferometry (BLI) assay as described in the methods. (c) mAbs 1E18 and 1N23 displayed potent neutralization against various influenza A, group 1 isolates indicated. (d) They all recognized a similar epitope as mAb FI6v35 as revealed by a BLI analysis described in the methods.

germline predecessor and found that it lost completely the reported binding ability of UCA5 to the HA from H1 (CA) isolate (data not shown),31 supporting our finding that the LC is critical for maintaining the binding activity of this family of antibodies. Therefore, IVTT reduced HC/LC pairing ambiguity during gene recovery through at least 2 mechanisms, i.e., assembly efficiency and binding affinity.

Discussion We demonstrated in this study the development of a novel IVTT-based antibody discovery platform. The unique characteristics of the isolated antibodies validated the quality and efficiency of the new platform and added new information to the current understandings of anti-influenza humoral responses.

MABS

Table 2. KD measurement of the newly identified mAbs by Octet. mAb

Kon(£ 103 1/Ms)

Koff(£ 10¡6 1/s)

KD (£ 10¡9 M)

R2

X2

1E18 1N23 2H4

360.0 560.0 0.5

34.8 58.5 172.7

0.097 0.105 349

0.99 0.99 0.98

0.87 0.47 1.55

By eliminating the expression plasmid construction step, this platform transforms the process through which V genes are recovered from B cells. Almost all current protocols require near clonal culture to minimize the number of expression plasmids to be constructed for antibody expression and confirmation, a critical attrition-inducing and rate-limiting step. We demonstrated that the matrix pairings in the IVTT platform is sensitive and accurate in identifying the functional pairings against a complex background with more than 10 B cells. This strategy theoretically allows the seeding of B cells in a nonclonal format, ideally with a density of 4–10 cells/well, thus reducing the burden of screening by corresponding folds and eliminating the needs for the stringent and time-consuming single B cell culture. The exact seeding densities can be determined according to the estimated hit rate and the sample size to strike a balance between the B-cell screening and V gene recovering efforts. Furthermore, the experience can be significantly improved if FACS sorting is used to dispense B cells into cell culture plates instead of manual operations, which were used in this study and introduced variations and unnecessary complexities. Deep sequencing offers the opportunity to gain insight into the genetic repertoire of an individual’s humoral response with unprecedented scale.11 However, to take full advantage of this information, knowledge must be obtained of the functions of the encoded antibodies. Attempts have been made to construct scFv operons maintaining the original HC/LC pairs before library construction, but the efficiency and quality are relatively

923

low and the attrition rate is high.33 Our platform provides the foundation to develop a fully automated, high throughput process to functionally probe a large number of B cells. Compatibilities with a plate format and Fabs production in a cell-free environment, the prerequisites for an automation process, have been demonstrated in this study. The IVTT platform might be a useful tool to study the mechanistic basis for antibody maturation. Of interest are the contributions of light chains to the VH1–69 using influenza A neutralizing antibodies. This is to our knowledge the first attempt to use primary B cells to demonstrate that efficient HC/LC pairing may play a role in B cell maturation. Our findings support a hypothesis that antibody expression level is correlated with thermostability, and may be one critical factor in determining the B cell fate.34 IVTT represents a sensitive tool to assess the stability of therapeutic antibodies under development. However, it is also apparent from our data that the impact from light chains is antibody dependent (Fig. 7C), and caution should be taken in extrapolating to other antibodies. High efficiency is critical to a B cell platform because the number of desirable B cells is small. Tremendous efforts were required to generate a single broadly neutralizing antibody against both influenza A group 1 and 2 isolates.5,19,17 In comparison, we were able to isolate the influenza A cross-reactive antibody 2H4 from only 10^4 memory B cells from a healthy donor. mAb 2H4 recognizes the same highly conserved epitope in the HA2 stem region as antibodies from the above described successful campaigns, but displays only modest neutralizing activity at high concentrations. This is expected as mAb 2H4 has a very poor KD profile, which may reflect its nature as a group 1 and 2 cross-reactive antibody in the early stage of maturation. The repeated immunizations preceding B cell isolation by all the above described campaigns seem to support this hypothesis. An anti-influenza therapeutic may result from mAb 2H4 if affinity is improved through antibody engineering.

Figure 5. Sequence comparison of parental mAbs 1E18, 1N23 (shaded sequences) with other broadly neutralizing antibodies including mAbs F10,18 CR6261,17 UCA1, 5, and 9.31 mAb 1E18 and 1N23 germlines and germlines with single mutations P52aG and P52aA are also included. SHMs from germlines are color highlighted and P52aG is colored in red.

924

X. XIAO ET AL.

Table 3. Comparison between mAbs 2H4 and FI6.5 Abbreviations: SHM, somatic hypermutation; nt, nucleotide; aa, amino acid. Heavy chain V gene 2H4 FI6



3–48 03 3–3018

CDR3 length (aa) 18 20

N1 nt 2 22

D gene 

3–22 01 3–901

Light chain N2 nt 31 11

J gene 

4 02 402

In addition to mAb 2H4, 2 group 1 neutralizing antibodies with unique features were isolated in the same campaign. Given the moderate starting memory B cell number and the even more restricted HA cross reactive B cell number, this platform displayed high efficiency. We expect this platform to perform superbly if immunized/patient sources are readily accessible. Previous work has established that the majority of group 1 broadly neutralizing antibodies use VH1–69 heavy chain for binding HA and are further divided into 2 groups distinguished by the 98Y and non-98Y usage.30,31 P52aA was found to be a major “founder mutation” for 98Y group, accounting for more than 80% of all members.31 Surprisingly, mAbs 1E18 and 1N23 both possess a P52aG instead of a P52aA. P52aG was observed with higher frequencies in phage-derived 98Y antibodies, but a P52aG in a non-98Y background has never been reported, which makes mAb 1N23 a very unique finding.18,30 This dramatic shift away from a recently published model analyzing nearly 200 mAbs31 mostly from a single donor raises several important questions. Are these differences attributed to donor difference, infection and vaccination-induced bias, or stochastic choosing of pathways? Given the limits of the small sample size

SHM nt (aa) 23 (13) 26 (17)

V gene

CDR3 length (aa)



K1–39 01 K4–101

9 9

N nt 0 0

J gene 

2 01 101

SHM nt (aa) 15 (7) 14C6 (9C2)

of this study and a general lack of anti-influenza mAbs derived directly from B cells from different donors, we believe that both vertical (following an individual over time in response to different flu strain challenges) and lateral (different donors) studies are needed to answer these questions.

Materials and methods Donor and HA specific B cell selection by ELISA, recombinant HAs Sera from healthy donors were collected with informed consent and used in ELISA-based screening of anti-Influenza HA activities. The blood collection procedure has been approved by the Chesapeake Institutional Review Board (protocol number 2010–001). All ELISAs for donor screenings and subsequent analysis were performed as described,35 and all secondary antibodies were prepared according to manufacturers’ suggested dilutions. Briefly, Nunc MaxiSorp 96-well plates (Corning, PA) were coated with 2 mg/ml recombinant HA (MedImmune, Gaithersburg, MD) isolated from influenza A/California/09 pandemic strain in phosphate-buffered

Figure 6. Contributions to binding by mutations at P52a position. Recombinant mAbs in their parental form (WT), their germlines (GL) and germline mutational variants as indicated were tested in ELISAs against a panel of HAs indicated on top of each of the graphs. The mAb variants are in different colors. In the lower panel, mAb 1E18 parental (WT) was used as the parental control (WT) for hybrid mAb UCA9H/1E18L due to the lack of UCA9 light chain information, and the mutations at P52a were generated against the UCA9H/1E18L background.

MABS

925

Figure 7. The impact of light chain on antibody expression and binding ability. (a) mAb IN23 heavy chain VHseq5 was paired with all 7 kappa light chains from well 1N23 and the Fab assembly was measured with a sandwich ELISA as described in method. (b) No significant variations were observed with heavy and light chain expressions. One heavy (#5) and 7 light (#1–7) chains (Fig. 3c) were expressed by IVTT in pairwise fashion and the heavy and light chain expressions from each pair were evaluated by Western blot using anti-Flag and anti-c-Myc antibodies, respectively. (c) Heavy chains showed different light chain biases in an IVTT. Heavy chain #5 preferred light chain #3 over light chains #1, 2, and 4, whereas heavy chain #6 showed no bias as examined with a sandwich ELISA showing Fab assembly. (d) The light chain preference by heavy chain #5 was confirmed in the recombinant IgG format. The recombinant IgGs using heavy chain #5 and indicated light chains were expressed transiently in 293F cells and their titers were estimated in an Octet instrument using protein A probes on days 3 and 6. (e) Light chain #1 usage led to reduced antigen binding by recombinant IgG using heavy chain #5. H1 (CA) and H5 were the ELISA antigens.

saline (PBS). Serially diluted sera were added to each well and horseradish peroxidase (HRP)-conjugated goat anti-human IgG Fc secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) was used for detection. The blocking and dilution buffer for all steps is PBST (PBS C 0.1% Tween20) C 3% dry milk. TMB (KPL, Gaithersburg, MD) substrate was added and color development was stopped by the addition of 0.1 N HCl. The HA-specific B cell screening procedure is essentially the same, except that Nunc MaxiSorp 384-well plates (Corning, PA) were used for HA coating. Recombinant HAs from Influenza A isolates including H1 (A/South Dakota/06/2007 H1N1, A/California/7/2009 H1N1), H2 (A/swine/Missouri/42964224/2006 H2N3), H5 (A/Vietnam/1203/2005 H5N1), H6 (A/mallard/Alberta/89/65 H6N2), H3 (A/Perth/2009 H3N2) and H7 (A/British Columbia/CN-6/ 2004 H7N3), and one HA from a B isolate (B/Brisbane/60/ 2008) were prepared at MedImmune as described previously.36 Memory B cell activation and amplification through EBV infection and CD40L stimulation B cells were isolated from whole blood of donors with good HA titer using a combination of EasySepTM Direct Human B Cell Isolation Kit (Stem Cell technologies, Canada) and Ficoll density gradient centrifugation (Histopaque 1077–1, Sigma, St. Louis, MO). IgGC and IgAC memory B cells were isolated using a

Switched memory B cell isolation kit (Milteyni Biotech Inc., California) following the manufacturer’s suggested protocol. Memory B cells were then infected with EBV (ATCC VR-1492) as described.37 A total of 1.2 £ 106 memory B cells were infected in RPMI medium supplemented with 10% FBS, 1% PenStrep (Invitrogen, CA) and stimulated using TLR agonist CpG ODN2006 (Invivogen, San Diego, CA). The cells were incubated at 5% CO2 at 37 C for 24 hours. Cells were then seeded at 4 cells per well in 384-well plates with 10,000 irradiated CHO-muCD40L cells per well. CHO-muCD40L stable cell line was generated at MedImmune. CHO-muCD40Lcells were irradiated at 8000 rads in a Model 68A irradiator using Caesium-137 as the radiation source before seeding. Outgrowth of B cells was estimated by microscopic observation and human IgG expression detection. V gene rescue and Fab template assembly for in vitro transcription and translation For validation experiment using single B cells, memory B cells were sorted based on the IgGCkappaC or IgGClambdaC phenotypes into 96-well PCR plate containing 5 ml/well lysis buffer from the SuperScript® III One-Step RT-PCR system (Invitrogen, Carlsbad, CA). The plates were used either directly for RT-PCR or stored at ¡80 C until use. For VH and VL recovery from EBV stimulated B cell samples, cells were collected from the selected wells into Eppendorf tubes, washed once with PBS, and pelleted. The cell

926

X. XIAO ET AL.

pellets were either lysed directly for RT-PCR or stored at ¡80 C until use. The SuperScript® III One-Step RT-PCR system (Invitrogen, Carlsbad, CA) was used for first strand cDNA synthesis. A two-step nested PCR procedure was used to amplify the VH, VK, and Vλ fragments with T7 promoter and RBS attached to the 50 end. This was followed by an overlapping PCR to add the CH1, CK, Cλ, affinity tags and terminator at the 30 end. The primers used for each step is summarized in Table 1. IVTT was performed using the PURExpress kit from NEB (New England Biolabs, Ipswich, MA). Each reaction (25 ml) contained »250 ng of total template DNA (HCCLC), 1 ml of Disulfide Bond Enhancer (New England BioLabs), and 1 ml of RNaseOut (40U/ ml, Invitrogen). The reactions were incubated at 37 C for 3 hours. The IVTT products were evaluated by both Fab expression and antigen binding ELISAs. In the capture ELISA to measure the Fab expression, the NUNC Maxisorp plates (Corning, PA) were coated with a Goat anti-Fd at 2 mg/ml in PBS (SouthernBiotech, Birmingham, AL) to capture the HC, the fully assembled Fabs were detected by goat anti-LC antibodies (horseradish peroxidase (HPR)-conjugated anti-human kappa or antihuman lambda, SouthernBiotech). Purified Kappa and lambda Fab proteins with known concentrations were used as the standards in the ELISA to determine the Fab expression level. In the antigen binding ELISA, the plates were coated with antigens. Fabs that bound to antigens were detected by HPR-conjugated goat anti-human Fab antibody (Jackson ImmunoResearch Laboratories, West Grove, PA). Bio-layer interferometry assays A ForteBio Octet QK384 instrument was used to study the kinetics of recombinant IgGs 1E18, 1N23, and 2H4 binding to HA (H1, CA09). All the assays were done at 200 ml/well in ForteBio 10x kinetic buffer (part no. 18–1092) at 25 C. 0.3 mg/ml of biotinylated-HA was loaded on the surface of streptavidin biosensors (SA) for 400 s, reaching capture levels between 1.0 and 1.5 nM, followed by a 300 s washing step. Association of HA on the biosensor to the individual mAbs in solution (0.274 –200 nM) was analyzed for 600 s. Dissociation of the interaction was probed for 600 s. Any systematic baseline drift was corrected by subtracting the shift recorded for a sensor loaded with ligand, but incubated with no analyte. Octet Data Analysis software version 8.0 was used for curve fitting with 1:1 and 2:1 interaction models, resulting in similar binding kinetics. Global analyses were done using nonlinear least squares fitting. Goodness of fit for the data was assessed by the generated residual plots, R2 and x2 values. Epitope binning was done on a ForteBio Octet QK384. Biotinylated-HA (H1, CA09) was captured onto streptavidin biosensors and coated with FI6v3 at a saturating concentration of 200 nM for 600 seconds. The epitopes of 1E18, 1N23, and 2H4 were probed in relation to FI6v3 by assaying the biotinylated-HAFI6v3 coated biosensors in 100 nM each of 1E18, 1N23, 2H4, FI6v3 (control mAb), and 7G16 (control mAb) separately for 1200 seconds. All graphs were overlaid and aligned at the baseline. Influenza micro-neutralization assay The microneutralization assay was performed as described previously.31 Briefly, 60 TCID50 of virus was added to 3-fold serial

dilutions of recombinant mAb IgGs in a 384-well plate in complete MEM medium containing 0.75 mg/ml L-(tosylamido-2phenyl) ethyl chloromethyl ketone (TPCK) treated trypsin (Worthington) in duplicate wells. After 1 hour incubation at 33 C, 5% CO2, 2 £ 104 Madin-Darby Canine Kidney (MDCK) cells/well were added to the plate. Plates were incubated at 33 C, 5% CO2 incubator for approximately 40 hr, and the NA activity was measured by adding a fluorescently-labeled substrate, methylumbelliferyl-N-acetyl neuraminic acid (MUNANA) (Sigma) to each well and incubated at 37 C for 1 hr. Virus replication represented by NA activity was quantified by reading fluorescence in Fluorometer Envison (PerkinElmer) using the following settings: excitation 355 nm, emission 460 nm; 10 flashes per well. The concentrations of mAbs required for a 50% reduction in viral replication (IC50) was calculated using a non-linear fit algorithm curve fit in Graph Pad Prism. If neutralization curve was not complete, then value was assigned the highest concentration of inhibitor tested. Reported IC50 values are an average of 4 independent experiments.

Disclosure of potential conflicts of interest No potential conflicts of interest were disclosed

References 1. Purtscher M, Trkola A, Gruber G, Buchacher A, Predl R, Steindl F, Tauer C, Berger R, Barrett N, Jungbauer A, et al. A broadly neutralizing human monoclonal antibody against gp41 of human immunodeficiency virus type 1. AIDS Res Hum Retroviruses 1994; 10:1651-8; PMID:7888224; http://dx.doi.org/10.1089/aid.1994.10.1651 2. Stiegler G, Kunert R, Purtscher M, Wolbank S, Voglauer R, Steindl F, Katinger H. A potent cross-clade neutralizing human monoclonal antibody against a novel epitope on gp41 of human immunodeficiency virus type 1. AIDS Res Hum Retroviruses 2001; 17:1757-65; PMID:11788027; http://dx.doi.org/10.1089/08892220152741450 3. Burton DR, Ahmed R, Barouch DH, Butera ST, Crotty S, Godzik A, Kaufmann DE, McElrath MJ, Nussenzweig MC, Pulendran B, et al. A Blueprint for HIV Vaccine Discovery. Cell Host Microbe 2012; 12:396-407; PMID:23084910; http://dx.doi.org/10.1016/j.chom.2012.09.008 4. Traggiai E, Becker S, Subbarao K, Kolesnikova L, Uematsu Y, Gismondo MR, Murphy BR, Rappuoli R, Lanzavecchia A. An efficient method to make human monoclonal antibodies from memory B cells: potent neutralization of SARS coronavirus. Nat Med 2004; 10:871-5; PMID:15247913; http://dx.doi.org/10.1038/nm1080 5. Corti D, Voss J, Gamblin SJ, Codoni G, Macagno A, Jarrossay D, Vachieri SG, Pinna D, Minola A, Vanzetta F, et al. A neutralizing antibody selected from plasma cells that binds to group 1 and group 2 influenza A hemagglutinins. Science 2011; 333:850-6; PMID:21798894; http://dx.doi.org/10.1126/science.1205669 6. Wrammert J, Smith K, Miller J, Langley WA, Kokko K, Larsen C, Zheng NY, Mays I, Garman L, Helms C, et al. Rapid cloning of highaffinity human monoclonal antibodies against influenza virus. Nature 2008; 453:667-71; PMID:18449194; http://dx.doi.org/10.1038/ nature06890 7. Yu X, Tsibane T, McGraw PA, House FS, Keefer CJ, Hicar MD, Tumpey TM, Pappas C, Perrone LA, Martinez O, et al. Neutralizing antibodies derived from the B cells of 1918 influenza pandemic survivors. Nature 2008; 455:532-6; PMID:18716625; http://dx.doi.org/10.1038/ nature07231 8. Kwakkenbos MJ, Diehl SA, Yasuda E, Bakker AQ, van Geelen CM, Lukens MV, van Bleek GM, Widjojoatmodjo MN, Bogers WM, Mei H, et al. Generation of stable monoclonal antibody-producing B cell receptor-positive human memory B cells by genetic programming.

MABS

9.

10.

11.

12. 13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

Nat Med 2010; 16:123-8; PMID:20023635; http://dx.doi.org/10.1038/ nm.2071 Corti D, Bianchi S, Vanzetta F, Minola A, Perez L, Agatic G, Guarino B, Silacci C, Marcandalli J, Marsland BJ, et al. Cross-neutralization of four paramyxoviruses by a human monoclonal antibody. Nature 2013; 501:439-43; PMID:23955151; http://dx.doi.org/10.1038/nature12442 Piccoli L, Campo I, Fregni CS, Rodriguez BM, Minola A, Sallusto F, Luisetti M, Corti D, Lanzavecchia A. Neutralization and clearance of GM-CSF by autoantibodies in pulmonary alveolar proteinosis. Nat Commun 2015; 6:7375; PMID:26077231; http://dx.doi.org/10.1038/ ncomms8375 Wine Y, Horton AP, Ippolito GC, Georgiou G. Serology in the 21st century: the molecular-level analysis of the serum antibody repertoire. Curr Opin Immunol 2015; 35:89-97; PMID:26172290; http://dx.doi. org/10.1016/j.coi.2015.06.009 Beerli RR, Rader C. Mining human antibody repertoires. mAbs 2010; 2:365-78; PMID:20505349; http://dx.doi.org/10.4161/mabs.12187 Huang J, Doria-Rose NA, Longo NS, Laub L, Lin CL, Turk E, Kang BH, Migueles SA, Bailer RT, Mascola JR, et al. Isolation of human monoclonal antibodies from peripheral blood B cells. Nat Protoc 2013; 8:1907-15; PMID:24030440; http://dx.doi.org/10.1038/nprot.2013.117 Szekely L, Pokrovskaja K, Jiang WQ, Selivanova G, Lowbeer M, Ringertz N, Wiman KG, Klein G. Resting B-cells, EBV-infected B-blasts and established lymphoblastoid cell lines differ in their Rb, p53 and EBNA-5 expression patterns. Oncogene 1995; 10:1869-74; PMID:7753563 Thompson WW, Shay DK, Weintraub E, Brammer L, Bridges CB, Cox NJ, Fukuda K. Influenza-associated hospitalizations in the United States. Jama 2004; 292:1333-40; PMID:15367555; http://dx.doi.org/ 10.1001/jama.292.11.1333 Thompson WW, Shay DK, Weintraub E, Brammer L, Cox N, Anderson LJ, Fukuda K. Mortality associated with influenza and respiratory syncytial virus in the United States. Jama 2003; 289:179-86; PMID:12517228; http://dx.doi.org/10.1001/jama.289.2.179 Dreyfus C, Laursen NS, Kwaks T, Zuijdgeest D, Khayat R, Ekiert DC, Lee JH, Metlagel Z, Bujny MV, Jongeneelen M, et al. Highly conserved protective epitopes on influenza B viruses. Science 2012; 337:1343-8; PMID:22878502; http://dx.doi.org/10.1126/science.1222908 Sui J, Hwang WC, Perez S, Wei G, Aird D, Chen LM, Santelli E, Stec B, Cadwell G, Ali M, et al. Structural and functional bases for broadspectrum neutralization of avian and human influenza A viruses. Nat Struct Mol Biol 2009; 16:265-73; PMID:19234466; http://dx.doi.org/ 10.1038/nsmb.1566 Nakamura G, Chai N, Park S, Chiang N, Lin Z, Chiu H, Fong R, Yan D, Kim J, Zhang J, et al. An in vivo human-plasmablast enrichment technique allows rapid identification of therapeutic influenza A antibodies. Cell Host Microbe 2013; 14:93-103; PMID:23870317; http:// dx.doi.org/10.1016/j.chom.2013.06.004 Wrammert J, Koutsonanos D, Li GM, Edupuganti S, Sui J, Morrissey M, McCausland M, Skountzou I, Hornig M, Lipkin WI, et al. Broadly crossreactive antibodies dominate the human B cell response against 2009 pandemic H1N1 influenza virus infection. J Expl Med 2011; 208:181-93; PMID:21220454; http://dx.doi.org/10.1084/jem.20101352 Walker LM, Phogat SK, Chan-Hui PY, Wagner D, Phung P, Goss JL, Wrin T, Simek MD, Fling S, Mitcham JL, et al. Broad and potent neutralizing antibodies from an African donor reveal a new HIV-1 vaccine target. Science 2009; 326:285-9; PMID:19729618; http://dx.doi. org/10.1126/science.1178746 Yin G, Garces ED, Yang J, Zhang J, Tran C, Steiner AR, Roos C, Bajad S, Hudak S, Penta K, et al. Aglycosylated antibodies and antibody fragments produced in a scalable in vitro transcription-translation system. mAbs 2012; 4:217-25; PMID:22377750; http://dx.doi.org/10.4161/ mabs.4.2.19202 Pope JH, Horne MK, Scott W. Transformation of foetal human keukocytes in vitro by filtrates of a human leukaemic cell line containing

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

927

herpes-like virus. Int J Cancer Journal international du cancer 1968; 3:857-66; PMID:4894385; http://dx.doi.org/10.1002/ijc.2910030619 Wiesner M, Zentz C, Mayr C, Wimmer R, Hammerschmidt W, Zeidler R, Moosmann A. Conditional immortalization of human B cells by CD40 ligation. PloS One 2008; 3:e1464; PMID:18213373; http://dx. doi.org/10.1371/journal.pone.0001464 Gudgeon NH, Taylor GS, Long HM, Haigh TA, Rickinson AB. Regression of Epstein-Barr virus-induced B-cell transformation in vitro involves virus-specific CD8C T cells as the principal effectors and a novel CD4C T-cell reactivity. J Virol 2005; 79:5477-88; PMID:15827162; http://dx.doi.org/10.1128/JVI.79.9.5477-5488.2005 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; 39:1260-70; PMID:19404981; http://dx.doi.org/10.1002/eji.200839129 Corti D, Suguitan AL, Jr., Pinna D, Silacci C, Fernandez-Rodriguez BM, Vanzetta F, Santos C, Luke CJ, Torres-Velez FJ, Temperton NJ, et al. Heterosubtypic neutralizing antibodies are produced by individuals immunized with a seasonal influenza vaccine. J Clin Invest 2010; 120:1663-73; PMID:20389023; http://dx.doi.org/10.1172/JCI41902 Sugimoto M, Furuichi Y, Ide T, Goto M. Incorrect us of “immortalization” for B-lymphoblastoid cell lines transformed by Epstein-Barr virus. J Virol 1999; 73:9690-1; PMID:10577062 Sugimoto M, Tahara H, Ide T, Furuichi Y. Steps involved in immortalization and tumorigenesis in human B-lymphoblastoid cell lines transformed by Epstein-Barr virus. Cancer Res 2004; 64:3361-4; PMID:15150084; http://dx.doi.org/10.1158/0008-5472.CAN-04-0079 Avnir Y, Tallarico AS, Zhu Q, Bennett AS, Connelly G, Sheehan J, Sui J, Fahmy A, Huang CY, Cadwell G, et al. Molecular signatures of hemagglutinin stem-directed heterosubtypic human neutralizing antibodies against influenza A viruses. PLoS Pathog 2014; 10:e1004103; PMID:24788925; http://dx.doi.org/10.1371/journal.ppat.1004103 Pappas L, Foglierini M, Piccoli L, Kallewaard NL, Turrini F, Silacci C, Fernandez-Rodriguez B, Agatic G, Giacchetto-Sasselli I, Pellicciotta G, et al. Rapid development of broadly influenza neutralizing antibodies through redundant mutations. Nature 2014; 516:418-22; PMID:25296253; http://dx.doi.org/10.1038/nature13764 Jayaram N, Bhowmick P, Martin AC. Germline VH/VL pairing in antibodies. Protein Eng Des Sel 2012; 25:523-9; PMID: NOT_FOUND; http://dx.doi.org/10.1093/protein/gzs043 Meijer PJ, Andersen PS, Haahr Hansen M, Steinaa L, Jensen A, Lantto J, Oleksiewicz MB, Tengbjerg K, Poulsen TR, Coljee VW, et al. Isolation of human antibody repertoires with preservation of the natural heavy and light chain pairing. J Mol Biol 2006; 358:764-72; PMID:16563430; http://dx.doi.org/10.1016/j.jmb.2006.02.040 Wang F, Sen S, Zhang Y, Ahmad I, Zhu X, Wilson IA, Smider VV, Magliery TJ, Schultz PG. Somatic hypermutation maintains antibody thermodynamic stability during affinity maturation. Proc Natl Acad Sci U S A 2013; 110:4261-6; PMID:23440204; http://dx.doi.org/ 10.1073/pnas.1301810110 Xiao X, Chen Y, Mugabe S, Gao C, Tkaczyk C, Mazor Y, Pavlik P, Wu H, Dall’Acqua W, Chowdhury PS. A Novel Dual Expression Platform for High Throughput Functional Screening of Phage Libraries in Product like Format. PloS One 2015; 10:e0140691; PMID:26468955; http://dx.doi.org/10.1371/journal.pone.0140691 Benjamin E, Wang W, McAuliffe JM, Palmer-Hill FJ, Kallewaard NL, Chen Z, Suzich JA, Blair WS, Jin H, Zhu Q. A broadly neutralizing human monoclonal antibody directed against a novel conserved epitope on the influenza virus H3 hemagglutinin globular head. J Virol 2014; 88:6743-50; PMID:24696468; http://dx.doi.org/10.1128/ JVI.03562-13 Traggiai E. Immortalization of human B cells: analysis of B cell repertoire and production of human monoclonal antibodies. Methods Mol Biol 2012; 901:161-70; PMID:22723100; http://dx.doi.org/10.1007/ 978-1-61779-931-0_10

A novel antibody discovery platform identifies anti-influenza A broadly neutralizing antibodies from human memory B cells.

Monoclonal antibody isolation directly from circulating human B cells is a powerful tool to delineate humoral responses to pathological conditions and...
1MB Sizes 0 Downloads 8 Views