Scholarly article on topic 'On-bead antibody-small molecule conjugation using high-capacity magnetic beads'

On-bead antibody-small molecule conjugation using high-capacity magnetic beads Academic research paper on "Chemical sciences"

CC BY
0
0
Share paper
Academic journal
Journal of Immunological Methods
OECD Field of science
Keywords
{"Antibody drug conjugates" / "Antibody labeling" / "Antibody purification" / "Antibody internalization" / "Cell based assay"}

Abstract of research paper on Chemical sciences, author of scientific article — Nidhi Nath, Becky Godat, Hélène Benink, Marjeta Urh

Abstract Antibodies labeled with small molecules such as fluorophore, biotin or drugs play an important role in various areas of biological research, drug discovery and diagnostics. However, the majority of current methods for labeling antibodies is solution-based and has several limitations including the need for purified antibodies at high concentrations and multiple buffer exchange steps. In this study, a method (on-bead conjugation) is described that addresses these limitations by combining antibody purification and conjugation in a single workflow. This method uses high capacity-magnetic Protein A or Protein G beads to capture antibodies directly from cell media followed by conjugation with small molecules and elution of conjugated antibodies from the beads. High-capacity magnetic antibody capture beads are key to this method and were developed by combining porous and hydrophilic cellulose beads with oriented immobilization of Protein A and Protein G using HaloTag technology. With a variety of fluorophores it is shown that the on-bead conjugation method is compatible with both thiol- and amine-based chemistry. This method enables simple and rapid processing of multiple samples in parallel with high-efficiency antibody recovery. It is further shown that recovered antibodies are functional and compatible with downstream applications.

Academic research paper on topic "On-bead antibody-small molecule conjugation using high-capacity magnetic beads"

ELSEVIER

Contents lists available at ScienceDirect

Journal of Immunological Methods

journal homepage: www.elsevier.com/locate/jim

On-bead antibody-small molecule conjugation using high-capacity magnetic beads

Nidhi Nath *, Becky Godat, Hélène Benink, Marjeta Urh

Promega Corporation, 2800 Woods Hollow Rd, Madison, W! 53711, United States

ARTICLE INFO ABSTRACT

Antibodies labeled with small molecules such as fluorophore, biotin or drugs play an important role in various areas of biological research, drug discovery and diagnostics. However, the majority of current methods for labeling antibodies is solution-based and has several limitations including the need for purified antibodies at high concentrations and multiple buffer exchange steps. In this study, a method (on-bead conjugation) is described that addresses these limitations by combining antibody purification and conjugation in a single workflow. This method uses high capacity-magnetic Protein A or Protein G beads to capture antibodies directly from cell media followed by conjugation with small molecules and elution of conjugated antibodies from the beads. High-capacity magnetic antibody capture beads are key to this method and were developed by combining porous and hydrophilic cellulose beads with oriented immobilization of Protein A and Protein G using HaloTag technology. With a variety of fluorophores it is shown that the on-bead conjugation method is compatible with both thiol- and amine-based chemistry. This method enables simple and rapid processing of multiple samples in parallel with high-efficiency antibody recovery. It is further shown that recovered antibodies are functional and compatible with downstream applications.

© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license

(http://creativecommons.org/licenses/by/4.0/).

CrossMark

Article history:

Received 14 May 2015

Received in revised form 19 August 2015

Accepted 20 August 2015

Available online 24 August 2015

Keywords:

Antibody drug conjugates Antibody labeling Antibody purification Antibody internalization Cell based assay

1. Introduction

Labeled antibodies are cornerstone in biological research and diagnostic testing with applications in immunodetection, immunoassays, immunohistochemistry, cell imaging and many others. An indication of the usefulness of the labeled antibodies is the variety of small molecules that have been conjugated to the antibodies including biotin, fluorescent dyes, proteins, nanoparticles and radioactive molecules for a variety of applications (Silverstein, 2004). More recently, labeled antibodies are also finding therapeutic applications, where drug molecules are attached to antibodies, resulting in improved disease treatments (Drachman and Senter, 2013; Panowksi et al., 2014). The terms labeling and conjugation are used interchangeably throughout the text.

Along with the increasing use of labeled antibodies as reagents there is a growing need for proper validation of the antibodies and related reagents to reduce the chances of false results (Bradbury and Pluckthun, 2015; Colwill, 2011). Although simple in concept, making validated labeled antibodies requires a lengthy process of generating antibodies, optimizing labeling chemistry and screening labeled antibodies for intended applications. Thus there is a desire for easy and robust methods that will allow multiple antibodies to be labeled and screened in relevant assays, preferably during the early stages of the antibody development process.

* Corresponding author. E-mail address: nidhi.nath@promega.com (N. Nath).

Currently, the two most common antibody labeling methods use amine and thiol groups present at lysine and cysteine amino acids on the antibody (Flygare et al., 2013). These methods are solution-based multistep reactions involving several incubations and buffer exchange steps hence limiting the throughput of the method. In addition, solution-based conjugation requires purified antibody at relatively high concentrations (> 1.0 mg/ml) thus limiting its utility during the early monoclonal screening stage where sample volume is limited and antibody titers are low (50 |ag/ml). To alleviate some of these limitations, on-bead labeling methods were developed (Lyon et al., 2012; Strachan et al., 2004), where antibody is captured onto the nonmagnetic bead surface followed by labeling and elution to recover labeled antibody. These methods have the advantage of combining purification and labeling in a single workflow but are done in a batch or column format, which limits the number of samples. More recently a magnetic bead-based antibody labeling methodology was developed to label up to twelve polyclonal antibodies in parallel (Dezfouli et al., 2014). The method is an improvement over non-magnetic bead-based labeling methods and shows nanogram-scale labeling of polyclonal rabbit antibodies.

To further broaden the utility of magnetic bead-based antibody labeling methods to include sample types such as hybridoma cell media containing various concentrations and isotypes of antibodies, high-capacity antibody-binding magnetic beads were developed, and a simple workflow was optimized for on-bead antibody purification and

http: //dx.doi.org/10.1016/j.jim.2015.08.008

0022-1759/© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/40/).

Fig. 1. Schematic of magnetic on-bead antibody conjugation using (A) thiols on cysteine residues and (B) amines on lysine residues.

labeling (Fig. 1). The method is compatible with both thiol- and amine-based chemistry and offers the flexibility of handling multiple samples in a manual or automated fashion. A key advantage of using high-capacity magnetic beads in this workflow is that a very small amount of beads is sufficient for efficient capture of antibody from dilute cell media. This results in concentrated antibody on the bead and leads to efficient conjugation reactions. In addition, the low amount of beads means that the conjugated antibodies can be eluted in a small volume resulting in a fairly high concentration amenable for direct use in downstream applications. Lastly, in this method, removal of unincorporated labels is accomplished by simple washing steps while antibody remains bound to the beads, eliminating the need for buffer exchange.

High-capacity antibody-binding magnetic beads were key to the method and developed by combining two novel approaches. First was the selection of porous magnetic cellulose beads (30 ^m-80 |am) because they offer high surface area and the hydrophilic cellulose shows low non-specific protein binding. Second was the choice of HaloTag technology for the covalent and oriented immobilization of Protein A or Protein G (Fig. 2), the two commonly used affinity ligands for capture and purification of antibodies. HaloTag is a 34 kDa protein fusion tag that forms a specific and covalent bond with synthetic HaloTag ligands and has been used previously to orient the protein of interest on a variety of surfaces (Urh and Rosenberg, 2012; Los et al., 2008). The choice of oriented immobilization of Protein A and Protein G instead of random attachment chemistry using lysines was driven by literature reporting better protein functionality using oriented immobilization (Weinrich

et al., 2010; Colombo et al., 2012; Ha et al., 2007; Jung et al., 2007). By combining porous cellulose beads and oriented protein attachment chemistry, magnetic Protein A and magnetic Protein G beads were developed. These beads have a high antibody-binding capacity similar to that of many commercial non-magnetic beads (~25 mg Human IgG/ ml of beads).

High-capacity antibody-binding magnetic beads were subsequently used to optimize methods for on-bead labeling of antibodies. Using several fluorescent dyes it was demonstrated that on-bead labeling of antibodies is compatible with amine- and thiol-based chemistries as well as with a variety of mouse and human antibody isotypes either purified or present in cell media. Finally, data is presented to show that antibodies with various dye-to-antibody ratios (DAR) can be made using this method, and these conjugated antibodies maintain their functionality as demonstrated by several downstream applications.

2. Materials and methods

2.1. Development of high-capacity magnetic Protein G and magnetic Protein A beads

2.1.1. Protein expression and purification

To generate Protein G-HaloTag fusion, the coding sequence for three domains of Protein G that bind the IgG Fc domain (from CAA27638 Streptococcus (Group G Strain G148)) were synthesized with an N-terminal (HQ)5 tag and a C-terminal HaloTag. For making Protein A-

Fig. 2. Schematic showing oriented and covalent immobilization of HaloTag fusion of Protein G or Protein A on magnetic cellulose beads.

HaloTag protein, coding sequence from five domains of Protein A that bind the Fc section of IgG (GenBank: EFB98015.1) were synthesized with an N-terminal (HQ)5 tag and a C-terminal HaloTag. Sequences were flanked by Sgf I and Pme I sites and transferred to the pF1K T7 Flexi Vector (Accession Number AY753577). The resulting vector was transformed into Single Step (KRX) Competent Cells (Promega). Proteins were expressed using the late auto-induction conditions (0.15% glucose/0.2% rhamose) in LB media at 25 °C for 18-24 h. Cells were harvested, lysed and proteins purified through an HQ tag using HisLink Protein Purification Resin (Promega) following the manufacturer's protocol followed by a Q-sepharose column. Protein was quantitated using the Bradford assay, and purity was checked using SDS-PAGE gel analysis.

2.1.2. Development of magnetic Protein G and Protein A beads

Magnetic Protein G and Protein A beads were made by incubating

purified Protein G-HaloTag and Protein A-HaloTag, respectively, with Magne HaloTag Beads (Promega). Magne HaloTag Beads are magnetic cellulose beads activated with HaloTag Ligands and can be used for various applications including HaloTag protein purification and protein:protein interaction studies (Banks et al., 2014; Saul et al., 2014). To optimize conditions for Protein A-HaloTag and Protein G-HaloTag conjugation, Magnetic HaloTag Beads were pipetted into 1.5 ml microcentrifuge tubes, washed with HEPES buffer (pH 7.2) and incubated with different amounts of Protein A-HaloTag or Protein G-HaloTag in HEPES buffer. The ratio of sample volume to settled bead volume was 4:1, and binding was done at room temperature for 2 h with constant end-over-end rotation. Following incubation, the amount of protein in the flow-through was determined using absorbance at 280 nm and used to calculate the amount of protein captured on the Magnetic HaloTag Beads. Beads were washed three times with HEPES buffer and stored in 20% ethanol as 20% slurry at 4 °C. For various washing steps described in this and the following sections, samples with magnetic beads were placed in the magnetic stand (MagneSphere® Technology Magnetic Separation Stand (twelve-position; Promega)), which pulls the beads to one side and allows solution to be easily removed using a pipette without any loss of magnetic beads.

2.1.3. Antibody binding capacity of magnetic Protein G and Protein A beads

Maximum binding capacities of magnetic Protein A and Protein G

beads were determined by incubating 10 |al (bead volumes mentioned in this paper refer to the settled volume) of each bead type with 400 |ag of purified polyclonal Human IgG (Sigma or Rockland Immunochemicals) in 200 of PBS for 30-60 min. After washing (three times with 200 of PBS), bead bound-antibody was eluted twice with 50 |al of elution buffer (10 mM glycine-HCl buffer, pH 2.2-2.7) and immediately neutralized with 10 of neutralization buffer (2 M Tris buffer, pH 7.5). To test the binding capacity of various isotypes of antibodies under conditions similar to that in the early hybridoma screening stage, purified antibodies from different species (Sigma or Rockland Immunochemicals) were spiked in PBS at 50 |ag/ml, and 1.0 ml samples were incubated with 10 ^l of either magnetic Protein A or Protein G beads for 30-60 min. After washing, antibodies were eluted twice with 50 of elution buffer and immediately neutralized with 10 ^l of neutralization buffer. The amount of antibody recovered was calculated using absorbance at 280 nm.

22. Antibody-small molecule conjugation on magnetic Protein G and Protein A beads (on-bead conjugation)

Uses of optimized high-capacity magnetic Protein A (Magne Protein A Beads from Promega) and Protein G beads (Magne Protein G Beads from Promega) for on-bead purification and conjugation of antibodies were shown using several fluorescent dyes. Both the amine and thiol reactions were tested, and several cell media samples expressing various isotypes of mouse antibodies were used to demonstrate the utility of the workflow.

22.1. Purification and conjugation of mouse antibodies with AlexaFluor 532 using amine reaction

Combined on-bead purification and conjugation of mouse antibodies using amine chemistry was performed with three different cell media samples containing mIgG1, mIgG2a and mIgG2b (Rockland Immunochemicals) at 50-100 |ag/ml. One milliliter samples were incubated for 60 min with 10 ^l of magnetic Protein G beads with constant mixing. After washing twice with 200 |al of PBS to remove nonspecifical-ly bound proteins, beads were incubated for 1 h with 25 |ag (~35 ^mol) of amine-reactive AlexaFluor 532 Dye (Life Technologies) in 50 |al of 10 mM phosphate buffer (PB) (pH 8.0). Beads were again washed three times with 10 mM PB (pH 7.0), and labeled antibodies were eluted twice using 50 of elution buffer and immediately neutralized by 10 of neutralization buffer. Two aliquots were combined before quantitat-ing antibody concentration and DAR as suggested by the dye provider. To determine if various chemical conjugation steps led to any loss of antibody, a simple purification of antibody from three cell media samples was also performed, and the antibody recovered was compared to the antibody recovered after conjugation. The amount of purified antibody was calculated using absorbance at 280 nm, and purity was checked using SDS-PAGE gel analysis.

222. Purification and conjugation of mouse antibodies with AlexaFluor 532 using thiol reaction

Combined on-bead purification and conjugation using thiol chemistry was performed using the same three samples used in the previous section. One milliliter samples were incubated for 60 min with 10 ^l of magnetic Protein G beads with constant mixing. After washing three times with 10 mM PB (pH 7.0) to remove nonspecifically bound proteins, bead-bound antibodies were incubated for 60 min with 2.5 mM DTT in PBE (10 mM PB with 1 mM EDTA (pH 7.0)) to selectively reduce inter-chain disulfide bonds. Following three washings with 200 |al of PBE to remove DTT, reduced antibodies on the beads were reacted for 1 h with ~35 |amol of maleimide-activated AlexaFluor 532 (Life Technologies) in 50 of PBE. Beads were washed three times with PBE, and labeled antibodies were eluted and neutralized as before. A separate purification step of the three antibodies was also performed to compare the recoveries with and without the conjugation step. Antibody recovery, purity and DAR were calculated as described in the previous section.

2.2.3. Purification and conjugation of mlgG2a with AlexaFluor 532, AlexaFluor 647 and Fluorescein using thiol reaction

Three different maleimide dyes, AlexaFluor 532, AlexaFluor 647 and Fluorescein, were used to label mIgG2a. To simulate cell media samples with high antibody concentration, cell media containing mIgG2a was concentrated about two- to threefold (Rockland Immunochemicals). The conjugation reaction was performed as described in the previous section using thiol chemistry with the following changes; to accommodate the higher amount of antibody, conjugation was performed using 20 |al of both magnetic Protein A and Protein G beads, and the amount of dye used was about 70 ^mol in 100 of buffer. Elution and neutralization buffer amounts were doubled.

2.3. Antibody functional assays

To demonstrate that antibody conjugated using the on-bead conjugation method retains its functional activity we labeled an anti-HER2 antibody (Trastuzumab) with AlexaFluor 647 and tested it in cell-based assays.

2.3.1. Conjugation of Trastuzumab with AlexaFluor 647 using amine reaction

Four hundred micrograms of purified Trastuzumab was captured on 30 |al of magnetic Protein G beads and reacted with amine-reactive AlexaFluor 647 dye as described before. Conjugated Trastuzumab was

eluted twice using 150 of elution buffer and neutralized with 30 of neutralization buffer.

23.2. Cell-based ELISA ofTrastuzumab

SKBR3 cells overexpressing HER2 and MDA-MB-231 cells with very low expression of HER2 were plated at 15,000 cells per well in a 96-well polystyrene plate and grown to confluence by overnight incubation at 37 °C. Cells were fixed using 4% paraformaldehyde and incubated with various dilutions of Trastuzumab or Trastuzumab labeled with AlexaFluor 647 for 1 h. Dilutions were made in PBS containing 10 mg/ml BSA (PBSB). After washing with PBS containing 0.05%Tween 20 (PBST), plates were incubated for 1 h with Anti-Human-IgG (H + L)-HRP (horseradish peroxidase) Conjugate (Promega) diluted 1:5000-fold in PBSB. After washing with PBST, TMB (3,3', 5,5'-tetramethylbenzidine) (Promega) was used as the HRP substrate. The colorimetric reaction was stopped by adding 1 N HCl and plates were read at 595 nm.

2.3.3. Cell-based internalization studies ofTrastuzumab

SKBR3 cells were plated at 30,000 cells per well in 8-well chambered cover glass and incubated for 48 h. Trastuzumab or Human IgG labeled with fluorescent dyes was added at 30 nM in complete media and incubated for 24 h. Cells were washed with complete media, and plates were imaged on a Nikon Confocal Microscope using the appropriate filter sets for each dye. MDA-MB-231 cells with very low expression of HER2 were used as negative-control cells.

3. Results

3.1. Development of high-capacity magnetic Protein G and Protein A beads

Oriented and covalent attachment of Protein A and Protein G on a high-surface area porous magnetic bead was critical for development of high-capacity antibody-binding beads. The approach was to use HaloTag fusions of Protein A and Protein G for covalent and oriented attachment to a porous magnetic cellulose bead activated with the HaloTag Ligand. HaloTag fusions of Protein A and Protein G expressed very well in Escherichia coli and were purified using His-Tag beads followed by a Q column to obtain proteins at >95% purity. Incubation of purified HaloTag fusion proteins with Magne HaloTag Beads allowed specific, covalent and oriented capture of Protein A and Protein G. Unlike other affinity tags where binding is reversible and equilibrium-based, the covalent binding between HaloTag and its ligand allows efficient and quantitative capture on the magnetic beads. This advantage is important as it allowed us to exactly tune the amount of protein at the surface simply by incubating the beads with increasing amounts of the HaloTag fusion as shown for Protein G-HaloTag in Fig. 3A. Subsequently, these magnetic beads charged with varying amounts of Protein G-HaloTag were used for human IgG purification (Fig. 3B). Maximum

human IgG capture of about 30 mg/ml of bead (200 nmol assuming m.wt =150 kDa) was reached at around 10 mg of Protein G-HaloTag/ml of bead (128 nmol assuming m.wt = 58 kDa) giving a ratio of about 1.2 antibody molecules per molecule of Protein G. This ratio is similar to that reported in the literature for solution-phase reactions (Lund et al., 2011) between Protein G and antibody, and it indicates that oriented attachment of proteins maintains the functionality of the protein. Almost quantitative binding to magnetic HaloTag beads was again seen with the HaloTag fusion of Protein A. The peak antibody recovery of about 21 mg of Human IgG (140 nmol) was achieved at around 5.5 mg of Protein A-HaloTag on the bead (77 nmol assuming m.wt = 70 kDa), and antibody recovery actually decreased at higher loading of Protein A probably due to ste-ric inhibition (data not shown).

Optimized magnetic Protein G beads and Protein A beads were subsequently tested for capture and recovery of various isotypes of mouse and human IgG from a 1.0 ml sample spiked with 50 ^g/ml of antibody. This concentration is typical of cell media samples during the early antibody discovery phase. Recoveries of various antibodies ranged from 60 to 80% of the input amount (Table 1).

3.2. Purification and on-bead conjugation of mouse antibodies using amine reaction

To test the ability of high-capacity magnetic Protein G and Protein A beads to capture antibodies from cell media followed by on-bead chemical conjugation, antibody conjugation was tested first through amines on the lysine residues of the antibodies. These sites are the most common reactive groups on antibodies and are frequently used for conjugation with dyes, drug molecules and peptides. AlexaFluor 532 was used with an amine-reactive succinimidyl ester (AlexaFluor 532-SE) to label various mouse antibody isotypes (mIgG1, mIgG2a and mIgG2b) present in the cell media at a typical expression level of ~50-100 ^g/ml (0.33-0.66 nmol/ml).

Ten microliters of magnetic Protein G beads was used to capture antibodies from the samples, washed to remove any non-specifically bound proteins and then incubated with 25 |ag (35 nmol) of AlexaFluor 532-SE dye in 50 of 10 mM PB (pH 8.0). Following another washing step to remove excess free dye, conjugated antibodies were eluted from the beads using a quick low-pH (pH 2.2) wash, and eluted antibodies were neutralized immediately. Efficiency of labeling was calculated using DAR as suggested by the manufacturer. Results showed a DAR of 4.0-5.0 (Table 2), which is in the range of 2-4 labels per antibody and has been shown to be optimal for maintaining antibody activity (Vira et al., 2010). In fact, two antibody drug conjugates (ADCs) approved by the FDA contain an average of 3.5 drugs per antibody (Drachman and Senter, 2013; Panowksi et al., 2014). It is worth mentioning that AlexaFluor 532-SE also reacts with lysines on the Protein G-HaloTag, but in the presence of a large excess of reactive dye, it was possible to consistently label antibodies with high DARs.

Fig. 3. (A) Capture efficiency of Protein G-HaloTag on Magne HaloTag Beads. (B) Antibody recovery as a function of Protein G-HaloTag on the bead. Each point is the average of two

Table 1

Antibody recovery using magnetic Protein G and Protein A beads.

Amount of IgG purified from 50 |og input sample (|g)

Species Isotype Magnetic Protein G beads Magnetic Protein A beads

Human IgA 0.4 ± 0.3 4.8 ± 2.1

IgG1 39.4 ±1.8 35.2 ± 1.9

IgG2 32.6 ± 0.7 34.4 ± 0.7

IgG3 37.6 ± 2.4 0

IgG4 38.0 ± 8.6 28.8 ± 2.75

IgM 0 8.2 ± 0.7

Mouse IgG1 30.4 ± 4.85 18.6 ± 7.8

IgG2a 31.6 ± 1.2 33.2 ± 3.0

IgG2b 31.4 ± 0.3 29.6 ± 1.4

IgG3 9.2 ± 1.2 16.0 ± 2.1

Rat IgG1 34.2 ± 1.6 29.4 ± 1.6

IgG2a 33.0 ± 0.6 0

IgG2b 31.2 ± 1.6 0

Sample: 1.0 ml sample spiked with 50 |og of antibody. Beads: 10 |lbead.

Average and standard deviation were calculated from triplicate samples.

In addition to the DAR, the efficiency of recovery for labeled antibodies was compared. For this comparison, antibodies from duplicate samples were captured using 10 |l of Protein G beads, and after washing, antibodies were eluted from one set of beads to obtain purified antibody. A second set of beads with captured antibodies was used to label and recover antibodies as described above. The efficiency of labeled antibody recovery was calculated by comparing the amount of antibody recovered from two different sets and was 70% for mIg2a and mIgG2b and 50% for mIgG1 (Table 2). This recovery using on-bead conjugation is much higher compared to about 30% reported for the solution-based conjugation reaction (Acchione et al., 2012). High antibody recovery can be attributed to the high affinity between antibody and Protein G, which is strong enough to withstand multiple reaction and washing steps without antibody loss. In addition, excellent magnetic response of the magnetic Protein G beads minimizes any loss of beads during purification. Furthermore, the conjugated antibody was eluted in 120 |l, which results in antibody concentration of 200-500 |g/ml compared to 50-100 |g/ml present in the initial sample. It is worth mentioning that all of the purification and conjugation steps were easily performed in triplicate and in parallel (n = 18) because of the ease of handling of magnetic beads.

Since the capture and conjugation of antibody was performed directly from cell media, it is important to check the purity of the conjugated antibody, which was done by resolving the conjugated antibody on a SDS-PAGE gel. The gel was fluorescently scanned to detect AlexaFluor 532 conjugated to the antibody heavy and light chains and then Coomassie stained to test for purity (Fig. 4A). The results indicate that both the heavy (55 kDa) and the light chains (25 kDa) are fluores-cently labeled and the conjugated antibodies are pure. Also, no Protein G leaching from the beads was observed under the conditions used for conjugation.

Table 3

Dye-to-antibody ratios of mouse IgG2a using thiol chemistry.

Magnetic Protein G beads Magnetic Protein A beads

Antibody Dye to Antibody Dye to

recovery antibody recovery antibody

(| g) ratio ( | g) ratio

(DAR) (DAR)

1 Purification 263.7 ±10.4 0 269.4 ± 5.1 0

2 AlexaFluor532 182.9 ± 15.3 5.3 ± 0.04 189.1 ± 6.9 6.9 ± 0.2

3 AlexaFluor647 192.5 ±2.9 3.3 ± 0.1 112.3 ± 11.4 3.6 ± 0.2

4 Fluorescein 179.5 ±5.3 6.8 ± 0.1 201.5 ±6.5 6.8 ± 0.1

20 | l of bead slurry was used with a 1.0 ml sample and a dye concentration of 67 | mol.

3.3. Purification and on-bead conjugation ofmouse antibodies using thiol reaction

Thiol chemistry is another approach used for attaching small molecules to antibodies (Doronina et al., 2003). Antibodies have intra-chain disulfide bonds as well as inter-chain disulfide bonds; however, only the inter-chain disulfide bonds in the antibody hinge region are solvent-accessible and can be reduced to free thiol groups using reducing agents like DTT (Dithiothreitol) or TCEP (Tris(2-Carboxyethyl) phosphine hydrochloride). For on-bead conjugation using thiol chemistry, antibodies were captured from the cell culture media as described previously using 10 |l of beads. Antibodies were reduced using 2.5 mM DTT, and after washing to remove DTT, reacted with AlexaFluor 532 dye containing a thiol-reactive maleimide group (AlexaFluor 532-ME). Beads were washed to remove excess free ligand and labeled antibody eluted and neutralized. DAR of 5.5 was obtained for mIgG2a and mIgG2b, while mIgG1 had a lower DAR of 1.6. A range of DAR is expected depending on antibody isotype, conjugation chemistry and the dye. Further optimization could improve the DAR of specific antibodies if needed. Recovery of conjugated antibody was > 70% as seen with amine chemistry, and the eluted antibody was concentrated. Purity of the antibody and fluorescent labeling of the antibody was confirmed using SDS-PAGE gel (Fig. 4B) and is similar to that obtained with amine chemistry.

3.4. On-bead conjugation of mouse IgG2a with multiple fluorescent dyes

The study was further expanded to test the robustness and flexibility of the on-bead conjugation method using (a) two additional thiol reactive dyes, AlexaFluor 647 and Fluorescein, and (b) cell media samples containing mIgG2a at three- to fourfold higher antibody titer than previously used. On-bead conjugation was performed on both the magnetic Protein G and Protein A beads because, even though Protein G has wider antibody binding specificity, Protein A is widely used in academic research and the biologics industry. In addition, although both Protein A and Protein G bind to the Fc region of the antibody, their mechanisms of binding are different (Sauer-Eriksson et al., 1995). Antibody Fc

Table 2

On-bead antibody conjugation using amine and thiol chemistry.

Well Isotype Amine Reaction Thiol Reaction

Antibody Recovery ( | g) Dye to Antibody Ratio (DAR) Antibody Recovery (|g) Dye to Antibody Ratio (DAR)

1 Mouse IgG1 Purification 54.8 ± 1.8 0 49.2 ± 2.0 0

2 Conjugation 27.4 ± 3.2 4.2 ± 0.2 46.2 ± 3.5 1.6 ± 0.2

3 Mouse IgG2A Purification 85.9 ± 6.4 0 75.3 ± 8.0 0

4 Conjugation 60.7 ± 3.2 4.4 ± 0.1 61.5 ± 7.4 5.5 ± 0.4

5 Mouse IgG2B Purification 79.9 ± 6.9 0 73.3 ± 1.4 0

6 Conjugation 66.7 ± 0.5 5.6 ± 0.1 55.7 ± 0.9 5.2 ± 0.5

Fig. 4. Gel images of antibody (mIgG1, mIgG2a and mIgG2b) coupled using amine (A) or thiol (B) chemistry. Gels were first scanned for fluorescence to detect AlexaFluor 532 labeled heavy and light chains and subsequently stained with Coomassie and scanned again. Lanes 1-6 correspond to samples in Table 2.

interaction with Protein A is more hydrophobic, whereas the interaction with Protein G is ionic and may influence on-bead conjugation.

Results indicate that on-bead conjugation worked very well for all three fluorescent dyes with DAR in the range of 3-7 dyes per antibody (Table 3). Lower DAR for AlexaFluor 647 dye may possibly be due to its structure, which is based on carbocyanine, whereas AlexaFluor 532 and Fluorescein are rhodamine-based dyes. Overall recoveries of conjugated antibodies were around ~ 70%, except for AlexaFluor 647 conjugation on magnetic Protein A beads, for which recovery was about 42%. In addition, no significant difference is seen in conjugation performed on magnetic Protein G and Protein A beads for mIgG2a that binds equally well to both Protein A and Protein G. However, in cases where antibodies, for example mIgG1, Rat Ig2a and Rat IgG2b, don't bind efficiently to Protein A, use of magnetic Protein G may result in better recoveries.

This study clearly demonstrated that the on-bead conjugation method is suitable for combining antibody labeling and purification in a single step. It was also shown that this method is compatible with labeling several antibody isotypes present over a range of antibody titers with a variety of fluorescent dyes.

3.5. On-bead conjugation and effect on antibody functionality

After showing the feasibility of on-bead conjugation, it was further investigated whether on-bead conjugation affected the functionality of the antibodies. An anti-HER2 antibody (Trastuzumab) was selected as a proof-of-concept because both the unlabeled antibody (Herceptin) and antibody labeled with the drug Maytansinoids (Kadcyla) have been approved by the FDA for cancer treatment (T-DM1) (Burris et al., 2011) and are known internalizing antibodies. Human IgG was used as a negative control, and both antibodies were labeled with AlexaFluor 647 using amine chemistry. A DAR of ~3.5 was obtained for each, and the conjugation of dye to heavy and light chain was confirmed by using SDS-PAGE gel analysis (Fig. 5A). Two different assays were used to test the functionality of Trastuzumab.

The first assay was a cell-based sandwich ELISA assay to compare the binding affinity of unconjugated and AlexaFluor 647-conjugated Trastuzumab. Results show a similar limit of detection (1.0 ng/ml) for unconjugated and conjugated Trastuzumab, but a slight decrease in the affinity ofTrastuzumab was observed after dye coupling (Fig. 5B).

Antibody Concentration (|ig/ml)

Fig. 5. (A) Gel images ofTrastuzumab and Human IgG conjugated to AlexaFluor 647 fluorescent dye. The gel was first scanned for fluorescence to detect fluorescently labeled heavy and light chain and subsequently stained with Coomassie and scanned again. (B) ELISA ofTrastuzumab and Trastuzumab conjugated to AlexaFluor 647. HER2 overexpressing SKBR3 cells and MDA-MB-231 cells with low HER2 proteins fixed to the bottom of a 96-well plate was used for binding.

The small loss in the binding activity upon conjugation with the fluorescent molecule is not uncommon especially with lysine chemistry and has been reported previously (Vira et al., 2010). This loss can be minimized by changing the conjugation chemistry and reaction conditions. Binding of Trastuzumab to HER2-expressing SKBR3 cells was specific as determined by low binding of both conjugated and unconjugated Trastuzumab to MDA-MB231 cells with very low expression of HER2.

The second assay explored the receptor-mediated internalization of the Trastuzumab-AlexaFluor 647 conjugate. In this assay, fluorescent dye-labeled antibodies bind to their specific receptor expressed on the cells and upon internalization traffic into endocytic vesicles, which appear as small fluorescent dots or punctate structures within the cytoplasm when seen using microscopy. Trastuzumab-AlexaFluor 647 incubated with HER2-positive SKBR3 cells results in antibody binding to HER2, which is then slowly internalized over 24 h (Fig. 6A-C) and can be seen as fluorescent punctate structures within the cells. In-ternalization is specific because MDA-MB-231 cells treated with Trastuzumab-AlexaFluor 647 (Fig. 6E) and SKBR3 cells treated with human IgG-AlexaFluor 647 do not show any internalization (Fig. 6D).

It is known that conjugating small molecules to the antibody may impact the antigen-binding affinity of the antibody (Vira et al., 2010; Lundberg et al., 2007; Shrestha et al., 2012), but the goal is to minimize the deleterious effect. These results clearly demonstrate that the on-bead conjugation method has the advantage of being a simple and robust method for labeling antibodies with a variety of small molecules directly from cell media with no significant impact on the antibody functionality.

4. Discussion

Developing validated labeled antibody requires the ability to prepare and screen a large library of labeled antibodies for desired

downstream biological application. Here we present an on-bead conjugation method using high-capacity magnetic Protein A and Protein G beads that enables single-step combined purification and conjugation of antibodies directly from cell media and allows processing of several samples in parallel. Furthermore, the method is compatible with two commonly used labeling chemistries and can be used to label various isotypes of mouse and human antibodies. The labeled antibodies are functional and can be used directly in the downstream applications required to characterize antibodies.

For successful implementation of on-bead antibody conjugation high-capacity antibody-binding Protein G and Protein A magnetic beads first were developed by combining hydrophilic and porous magnetic cellulose beads with covalent and oriented attachment using HaloTag technology. The choice of HaloTag technology was based on their previous use for protein immobilization (Hoppe et al., 2012; Wang et al., 2013) and the advantages it offers, including: (1) it allows the same HaloTag ligand-activated magnetic bead to be used for making both Protein A and Protein G beads; (2) the capture of the HaloTag fusion protein is quantitative, which fine tunes the amount of Protein A and Protein G needed for maximum antibody binding; (3) it allows robust covalent attachment of Protein A and Protein G, which prevents leaching as demonstrated by the absence of Protein G in antibody samples analyzed using SDS-PAGE gel. The observation is similar to another report, where HaloTag was used to prevent protein losses from surface under harsh denaturating conditions (Wang et al., 2013); and (4) the oriented attachment should overcome the limitation of the traditional random protein immobilization method, which often results in loss of protein activity and requires time-consuming protein-specific optimization (Rusmini et al., 2007; Wilchek and Miron, 2003).

Efficient recovery of antibodies is a key aspect of the antibody purification and labeling method and is especially vital when samples contain low amounts of antibody. Thus the performance of on-bead

Trastuzumab-AlexaFluor 647 SKBR-3

Trastuzumab-AlexaFluor 647 SKBR-3

Trastuzumab-AlexaFluor 647 SKBR-3

Fig. 6. Cell internalization of AlexaFluor 647 conjugate of Trastuzumab made using amine chemistry. SKBR3 cells overexpressing HER2 were incubated with 30nM of Trastuzumab-AlexaFuor 647 and imaged at (A) 0 h, (B) 3 h and (C) 24 h to monitor internalization. (D) Human IgG-AlexaFluor 647 (negative control) was incubated with SKBR3 cells and imaged at 24 h. (E) The MDA-MB-231 cell line expressing very low level of HER2 was incubated with Trastuzumab-AlexaFuor 647 and used as an additional negative control.

conjugation was carefully compared to the conventional solution-based conjugation methods. One of the major concerns was that on-bead conjugation involves multiple washing and incubation steps and may result in a significant loss of antibody. However, recovery of 50-70% using on-bead conjugation compare very favorably with the reported solution-based method of 30-55% (Acchione et al., 2012) and is probably due to the robust binding between antibody and Protein A and Protein G. High antibody recovery is especially attractive because it is achieved with small amounts (50-100 |g) of antibody. In addition to efficient recovery, the DARs (1.6-6.9) obtained using on-bead conjugation correlate well with the range of 2-4 that has been determined to be optimum (Hamblett et al., 2004; McDonagh et al., 2006) for therapeutic ADCs and for maintaining antibody activity (Vira et al., 2010). Further, the ability to label low-concentration antibody with a therapeutically relevant DAR of 2-4 using on-bead conjugation is especially useful for amine-based chemistries because the amine-reactive succinimidyl ester group is susceptible to hydrolysis in aqueous conditions. To minimize the hydrolysis side reaction, traditional solution-based methods require a high concentration of purified antibody (1-10 mg/ml) in order to drive the reaction towards antibody conjugation (Strachan et al., 2004).

A possible concern with on-bead conjugation is that binding to Protein A or Protein G in close proximity of the bead surface may sterically hinder reactive sites on the antibody. This concern is partially addressed by the observation that by using thiol chemistry an average DAR of 6.9 for mouse IgG2A out of a maximum of 10 thiols that are solvent-accessible could be achieved (Lyon et al., 2012). A related concern is the possibility that distribution of dye on the antibody, especially using lysine chemistry, may be different than that which results from solution-based chemistry. However, even with solution-based chemistry, distribution of antibodies with different numbers of dyes per antibody is a major challenge, and there are concerted efforts to develop site-specific conjugation methods (Panowksi et al., 2014; Junutula et al., 2008; Rabuka et al., 2012; Wu et al., 2009).

Labeled antibodies often need to be dialyzed into buffer that is appropriate for a downstream application, for example ELISA, internaliza-tion studies, cell toxicity assays and others. However, due to the high capacity of magnetic beads, small volumes of beads can capture and concentrate antibody from diluted (0.05-0.1 |g/ml) cell media and allow elution of conjugated antibody in a small volume at a relatively high concentration (0.2-0.5 mg/ml). Availability of high-concentration fluorescent antibodies allowed functional assays such as ELISA and in-ternalization assays to be run without any additional buffer exchange.

A key parameter for any antibody-labeling method is to maintain the antigen binding affinity of the conjugated antibody. Using Trastuzumab as one example it was shown that the on-bead conjugation method maintained the functionality of the antibody for both binding to its antigen (HER2) in vitro as well as in a cell-based internalization assay. The small shift in IC50 of labeled Trastuzumab reported for the ELISA assay is not unexpected especially with amine chemistry since dyes can bind to lysines present in proximity to antigen binding and thus modify the affinity for antigen binding. Similar observations have been made with other antibodies used in this manner (Vira et al., 2010).

Several advantages of on-bead antibody purification and conjugation directly from cell media have been highlighted; however, the use of cell media poses few challenges. One limitation is that the antibody titer and antibody isotypes in early hybridoma samples may not be known. As a result, using a single concentration of reactive small molecule may over- or under-label the antibodies. In addition, DARs and antibody recoveries will also vary depending on the antibody titer and antibody isotypes. Only two different fluorescent dyes were tested; however, other cytotoxic drugs used in ADCs are hydrophobic and at high labeling ratios may aggregate or stick to the beads resulting in lower recoveries. However the solubility problem is not unique to on-bead conjugation and has been reported for the traditional solution-based labeling method (Wakankar et al., 2010). This problem can be

mitigated by optimizing the elution buffer. Finally, scalability of the conjugation method is important as larger scale production of labeled antibodies is required. The on-bead conjugation was scaled to 50 ml of cell media with similar DAR and antibody recoveries (data not shown), but at larger scales, the conjugation process will have to be transitioned to solution-based chemistry and may need re-optimization at various steps. Notwithstanding these limitations, on-bead conjugation is an enabling approach for primary screening of antibody and label together in the context of the desired biological application and should allow selection of well characterized antibody reagents.

5. Conclusions

In conclusion we have developed an on-bead antibody-small molecule conjugation process using high-capacity magnetic Protein G and Protein A beads that may simplify the process of selecting high-quality validated labeled antibody reagents. Key highlights of this method are (a) antibody present in the cell media can be conjugated without prior purification, (b) small volumes of cell media samples (1 ml) containing low amount of antibodies (50 |g/ml) can be labeled, (c) multiple antibody samples can be labeled in parallel, (d) conjugated antibodies are functional and (e) labeled antibodies are eluted at high concentration suitable for a variety of downstream applications. Finally, given the advantages of the on-bead conjugation method, it is easy to imagine that this approach can be extended to other areas of biological research that use antibodies labeled with small molecules, proteins, enzymes, polymer and nanoparticles among others.

References

Acchione, M., Kwon, H.,Jochheim, C.M., Atkins, W.M., 2012. Impact of linker and conjugation chemistry on antigen binding, Fc receptor binding and thermal stability of model antibody-drug conjugates. mAbs 4 (3), 362. Banks, C.A., Lee, Z.T., Boanca, G., Lakshminarasimhan, M., Groppe, B.D., Wen, Z., Hattem, G.L., Seidel, C.W., Florens, L., Washburn, M.P., 2014. Controlling for gene expression changes in transcription factor protein networks. Mol. Cell. Proteomics 13 (6), 1510. Bradbury, A., Pluckthun, A., 2015. Reproducibility: standardize antibodies used in research. Nature 518 (7537), 27. Burris III, H.A., Tibbitts, J., Holden, S.N., Sliwkowski, M.X., Lewis Phillips, G.D., 2011. Trastuzumab emtansine (T-DM1): a novel agent for targeting HER2 + breast cancer. Clin. Breast Cancer 11 (5), 275. Colombo, M., Mazzucchelli, S., Montenegro, J.M., Galbiati, E., Corsi, F., Parak W.J., Prosperi, D., 2012. Protein oriented ligation on nanoparticles exploiting O6-alkylguanine-DNA transferase (SNAP) genetically encoded fusion. Small 8 (10), 1492. Colwill, K., 2011. Renewable protein binder working G, Graslund S: a roadmap to generate

renewable protein binders to the human proteome. Nat. Methods 8 (7), 551. Dezfouli, M., Vickovic, S., Iglesias, M.J., Nilsson, P., Schwenk, J.M., Ahmadian, A., 2014. Magnetic bead assisted labeling of antibodies at nanogram scale. Proteomics 14 (1), 14. Doronina, S.O., Toki, B.E., Torgov, M.Y., Mendelsohn, B.A., Cerveny, C.G., Chace, D.F., DeBlanc, R.L., Gearing, R.P., Bovee, T.D., Siegall, C.B., et al., 2003. Development of potent monoclonal antibody auristatin conjugates for cancer therapy. Nat. Biotechnol. 21 (7), 778.

Drachman, J.G., Senter, P.D., 2013. Antibody-drug conjugates: the chemistry behind

empowering antibodies to fight cancer. Hematology 2013, 306. Flygare, J.A., Pillow, T.H., Aristoff, P., 2013. Antibody-drug conjugates for the treatment of

cancer. Chem. Biol. Drug Des. 81 (1), 113. Ha, T.H., Jung, S.O., Lee, J.M., Lee, K.Y., Lee, Y., Park, J.S., Chung, B.H., 2007. Oriented immobilization of antibodies with GST-fused multiple Fc-specific B-domains on a gold surface. Anal. Chem. 79 (2), 546. Hamblett, K.J., Senter, P.D., Chace, D.F., Sun, M.M., Lenox, J., Cerveny, C.G., Kissler, K.M., Bernhardt, S.X., Kopcha, A.K., Zabinski, R.F., et al., 2004. Effects of drug loading on the antitumor activity of a monoclonal antibody drug conjugate. Clin. Cancer Res. 10 (20), 7063.

Hoppe, S., Bier, F.F., von Nickisch-Rosenegk M., 2012. Microarray-based method for

screening of immunogenic proteins from bacteria. J. Nanobiotechnol. 10,12. Jung, Y., Lee, J.M., Jung, H., Chung, B.H., 2007. Self-directed and self-oriented immobilization of antibody by protein G-DNA conjugate. Anal. Chem. 79 (17), 6534. Junutula, J.R., Raab, H., Clark S., Bhakta, S., Leipold, D.D., Weir, S., Chen, Y., Simpson, M., Tsai, S.P., Dennis, M.S., et al., 2008. Site-specific conjugation of a cytotoxic drug to an antibody improves the therapeutic index. Nat. Biotechnol. 26 (8), 925. Los, G.V., Encell, L.P., McDougall, M.G., Hartzell, D.D., Karassina, N., Zimprich, C., Wood, M.G., Learish, R., Ohana, R.F., Urh, M., et al., 2008. HaloTag: a novel protein labeling technology for cell imaging and protein analysis. ACS Chem. Biol. 3 (6), 373. Lund, L.N., Christensen, T., Toone, E., Houen, G., Staby, A., St Hilaire, P.M., 2011. Exploring variation in binding of Protein A and Protein G to immunoglobulin type G by isothermal titration calorimetry. J. Mol. Recognit. 24 (6), 945.

Lundberg, E., Sundberg, M., Graslund, T., Uhlen, M., Svahn, H.A., 2007. A novel method for reproducible fluorescent labeling of small amounts of antibodies on solid phase. J. Immunol. Methods 322 (1-2), 40.

Lyon, R.P., Meyer, D.L., Setter, J.R., Senter, P.D., 2012. Conjugation of anticancer drugs through endogenous monoclonal antibody cysteine residues. Methods Enzymol. 502,123.

McDonagh, C.F., Turcott, E., Westendorf, L., Webster, J.B., Alley, S.C., Kim, K., Andreyka, J., Stone, I., Hamblett, K.J., Francisco, JA, et al., 2006. Engineered antibody-drug conjugates with defined sites and stoichiometries of drug attachment. Protein Eng. Des. Sel. 19 (7), 299.

Panowksi, S., Bhakta, S., Raab, H., Polakis, P., Junutula, J.R., 2014. Site-specific antibody drug conjugates for cancer therapy. mAbs 6(1), 34.

Rabuka, D., Rush, J.S., deHart, G.W., Wu, P., Bertozzi, C.R., 2012. Site-specific chemical protein conjugation using genetically encoded aldehyde tags. Nat. Protoc. 7 (6), 1052.

Rusmini, F., Zhong, Z., Feijen, J., 2007. Protein immobilization strategies for protein biochips. Biomacromolecules 8 (6), 1775.

Sauer-Eriksson, A.E., Kleywegt, G.J., Uhlen, M., Jones, T.A., 1995. Crystal structure of the C2 fragment of streptococcal protein G in complex with the Fc domain of human IgG. Structure 3 (3), 265.

Saul, J., Petritis, B., Sau, S., Rauf, F., Gaskin, M., Ober-Reynolds, B., Mineyev, I., Magee, M., Chaput, J., Qiu, J., et al., 2014. Development of a full-length human protein production pipeline. Protein Sci. 23 (8), 1123.

Shrestha, D., Bagosi, A., Szollosi, J., Jenei, A., 2012. Comparative study of the three different fluorophore antibody conjugation strategies. Anal. Bioanal. Chem. 404 (5), 1449.

Silverstein, A.M., 2004. Labeled antigens and antibodies: the evolution of magic markers and magic bullets. Nat. Immunol. 5 (12), 1211.

Strachan, E., Mallia, A.K., Cox, J.M., Antharavally, B., Desai, S., Sykaluk, L., O'Sullivan, V., Bell, P.A., 2004. Solid-phase biotinylation of antibodies. J. Mol. Recognit 17 (3), 268.

Urh, M., Rosenberg, M., 2012. HaloTag, a platform technology for protein analysis. Curr. Chem. Genom. 6, 72.

Vira, S., Mekhedov, E., Humphrey, G., Blande, P.S., 2010. Fluorescent-labeled antibodies: balancing functionality and degree of labeling. Anal. Biochem. 402 (2), 146.

Wakankar, A.A., Feeney, M.B., Rivera, J., Chen, Y., Kim, M., Sharma, V.K., Wang, Y.J., 2010. Physicochemical stability of the antibody-drug conjugate Trastuzumab-DM1: changes due to modification and conjugation processes. Bioconjug. Chem. 21 (9), 1588.

Wang, J., Barker, K., Steel, J., Park, J., Saul, J., Festa, F., Wallstrom, G., Yu, X., Bian, X., Anderson, K.S., et al., 2013. A versatile protein microarray platform enabling antibody profiling against denatured proteins. Proteomics Clin. Appl. 7 (5-6), 378.

Weinrich, D., Lin, P.C., Jonkheijm, P., Nguyen, U.T., Schroder, H., Niemeyer, C.M., Alexandrov, K., Goody, R., Waldmann, H., 2010. Oriented immobilization of farnesylated proteins by the thiol-ene reaction. Angew. Chem. 49 (7), 1252.

Wilchek, M., Miron, T., 2003. Oriented versus random protein immobilization. J. Biochem. Biophys. Methods 55 (1), 67.

Wu, P., Shui, W., Carlson, B.L., Hu, N., Rabuka, D., Lee, J., Bertozzi, C.R., 2009. Site-specific chemical modification of recombinant proteins produced in mammalian cells by using the genetically encoded aldehyde tag. Proc. Natl. Acad. Sci. U. S. A. 106 (9), 3000.