Chemoenzymatic glycan labelling as a platform for site-specific IgM-antibody drug conjugates

Analytical Biochemistry 584 (2019) 113385

Contents lists available at ScienceDirect

Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio

Chemoenzymatic glycan labelling as a platform for site-specific IgMantibody drug conjugates

T

Edward S.X. Moh, Nima Sayyadi, Nicolle H. Packer* Department of Molecular Sciences and ARC Centre of Excellence for Nanoscale BioPhotonics, Macquarie University, Sydney, Australia

A R T I C LE I N FO

A B S T R A C T

Keywords: IgM Antibody-drug conjugate Click chemistry Drug-antibody ratio Glycans Glycopeptides

Immunoglobulin M (IgM) type antibodies play a significant role in complement activation, cellular debris clearance and cell quality control, and have the potential to be used as a therapeutic or targeting/delivery antibody. However, this potential has not been explored thoroughly due to its high molecular weight, polymeric structure and large number of glycosylation sites. Site-specific antibody-drug-conjugates (ADC) are considered the next generation protein biotherapeutic drugs and currently all, in clinical trials and approved, are of the IgG isotype. As existing methods for the development and characterization of IgG-ADCs are not compatible with IgMADC, we describe a platform methodology suitable for site specific IgM-ADC using a chemoenzymatic method targeting the glycans on the IgM. Azide functionalized sialic acids were incorporated onto IgM glycans using sialyltransferase for biocompatible conjugation using “click” chemistry. The number of azide groups incorporated onto the IgM glycans were characterized by mass spectrometry of the enzymatically released glycans and glycopeptides. Quantitation of the azide incorporation showed an azide antibody ratio of 8 (glycan data) and 6–10 (glycopeptide data) which translates to a high drug antibody ratio based on IgG-ADC standards. This platform methodology can be readily adapted for any human IgM produced in a mammalian cell expression system.

1. Introduction Immunoglobulin M (IgM) type antibodies have untapped potential, both as a biotherapeutic and targeting/delivery molecule. IgMs are secreted by immune cells as the first antibody response against an antigen, and serve as a cell quality control molecule, killing “rogue” cells in the body and clearing cellular debris [1]. Germline encoded natural IgM antibodies have also been reported to offer protection against cancer, inflammation and auto-immunity [2]. These natural IgMs target host epitopes that are not displayed on the cell surface of a normal cell (either intracellular or inaccessible in conformation), but get overexpressed or exposed in a diseased/apoptotic state [2]. Several IgM type antibodies, including natural IgM [3,4], immune IgM [5] and class switched IgM [6,7], demonstrate the potential of IgM type antibodies as a therapeutic antibody. Monoclonal antibody products make up the majority of the biopharmaceutical market, predominantly IgG type antibodies, partly due to the extensive research and development into IgG production, purification and characterization methods. IgG exists as a monomer consisting of 2 heavy and light chains, with each heavy chain containing a single glycosylation site. The total molecular weight is only

*

approximately 150 kDa, enabling quick analytical characterization with minimal sample preparation mostly using mass spectrometry techniques [8,9]. IgGs have provided the platform for development of IgGantibody drug conjugates (ADC) with defined drug-antibody ratio (DAR) for advanced therapeutic treatments [10–12]. Site-specific conjugation of drugs for IgG-ADC is the currently preferred method for creating IgG-ADC, as the theoretical maximum of drug loading can be defined by the conjugation site. Cysteines, incorporated non-natural amino acids and the single Fc glycosylation site are often the target sites as they offer unique functional groups in defined locations of the antibody [10,11]. For example, Adcetris®, an FDA-approved ADC for the treatment of classical Hodgkin lymphoma, is brentuximab antibody conjugated with the drug vedotin targeted to the cysteine side chains on the antibody [12]. As the biotherapeutic potential of IgM gains attention [7,13,14] and large scale methods for IgM expression and purification develop [15–17], we have investigated the potential of creating site-specific glycan conjugated IgM-ADCs. IgM is the largest antibody in humans, existing predominantly as pentameric antibodies (5 IgM monomer of 2 heavy and light chains + J-chain), reaching approximately 950 kDa in molecular weight. Hexameric IgM can also be formed without the J-

Corresponding author. Department of Molecular Sciences Macquarie University, North Ryde NSW, 2109, Australia. E-mail address: [email protected] (N.H. Packer).

https://doi.org/10.1016/j.ab.2019.113385 Received 14 June 2019; Received in revised form 30 July 2019; Accepted 1 August 2019 Available online 02 August 2019 0003-2697/ © 2019 Published by Elsevier Inc.

Analytical Biochemistry 584 (2019) 113385

E.S.X. Moh, et al.

chain, reaching approximately 1.1 MDa in molecular weight. It is also heavily glycosylated (51 N-glycans per pentameric IgM = 5 sites per 10 IgM heavy chains + 1 on J-chain; 60 N-glycans per hexameric IgM = 5 sites per 12 IgM heavy chain) and is heavily dependent on cysteine disulphide bonds to maintain its polymeric state. Currently, only one recent study has explored the synthesis of IgM-ADC, by conjugating site-specifically the engineered selenocysteines at the C-terminus of each heavy chain, for potential treatment of chronic lymphocytic leukemia [18]. In this study, we explored the site-specific conjugation of IgM-ADC by targeting the multiple glycosylation sites. Glycosylation site targeting for ADC can be performed by modifying the glycans to contain unique functional groups not present in the protein backbone. For IgG antibodies, this is generally done by two means, oxidation of the vicinal diols in the glycan chain to create aldehyde groups for aldehyde-amine or aldehyde-hydrazine conjugation [19,20], or by transferring a modified monosaccharide containing an azide group using a specific glycosyltransferase, followed by azide-alkyne or azidephosphine ligation [21–23]. These unique functional groups can then be used for specific chemical ligation reactions with the drug or small molecule target at known and quantifiable sites. No glyco-targeted IgG-ADC are currently in trials, with only 1 in pre-clinical stage, and 1 still in development [10]. Both products use methods that require multiple enzymatic treatments of the single glycan on the heavy chain to increase the conjugation yield; galactosidase treatment to generate numerous acceptor terminal glycans, followed by either galactosyltransferase (GalT-Y289L) that has been mutated to add azide-UDP-GalNAc, or by galactosyltransferase and sialyltransferase followed by the periodate oxidation of sialic acids [10]. This results in a maximum DAR of 4 as each IgG heavy chain only contains 1 glycosylation site with a bi-antennary glycan (2 conjugates per glycan). For IgM, this problem is circumvented as it contains many more glycans per antibody (51 per pentameric IgM compared to 2 per IgG). Based on our recently published glycopeptide characterization of monoclonal IgM [24], 31 of the 51 N-glycosylation sites (10x Asn171, 10x Asn332, 10x Asn395 and J-chain Asn71) contain predominantly galactosylated and sialylated complex N-glycans, suggesting that employing glycosyltransferases that can directly modify these complex glycans could result in a high DAR. In our previous work, we characterized the site-specific glycosylation of the monoclonal IgM PAT-SM6 [24] and found that approximately 40–45% of PAT-SM6 glycans contained terminal galactose residues, that could serve as substrates for the sialyltransferase, ST6GAL1 (Table 1). Hence, in this work, we quantitatively and specifically functionalized the monoclonal IgM with a one-step reaction using ST6GAL1 and azide-functionalized CMP-sialic acid, that was further modified using copper-free click chemistry with a strain-alkyne functionalized small fluorescent molecule as a proof of concept (Fig. 1).

Table 1 Summary of glycan compositions and their relative abundance on the respective glycopeptides that are potential substrates for ST6GAL1 based on glycopeptide quantitation data from Ref. [24], and previously unpublished data on J-chain glycopeptide quantitation. Glycan compositions are listed in linear nomenclature; prefix F→core fucose, man→no. of mannose, A→no of antenna GlcNac, G→no. of galactose, F→no. of outer arm fucose, S→no. of sialic acid. % contribution in pentameric IgM was calculated by scaling the summed abundance against the total number of glycosylations sites in pentameric IgM. Identified glycan compositions that are potential substrates for ST6GAL1

No. of free galactose for ST6GAL1

man5 A1G1 1 F man5 A1G1 1 A1G1 1 FA1G1F1 1 A2G2S1 1 FA2G1 1 FA2G1F1 1 FA2G2F1S1 1 FA2G2S1 1 A3G3S2 1 FA3G2 1 FA3G2F1 1 FA3G2F1S1 1 FA3G2S1 1 FA3G3S2 1 A2G2 2 FA2G2 2 FA2G2F1 2 A3G3S1 2 A3G3S1F1 2 FA3G1 2 FA3G3S1 2 A3G3 3 FA3G3 3 FA3G3F1 3 A4G4 4 Relative % at each glycosylation site % contribution to intact Pentameric IgM

Glycopeptide site IgM heavy chain Asn171

Asn332

J-chain Asn395

2.71 6.13

Asn563

Asn71

0.83 0.72

1.08

16.08 19.15 0.90

4.34 2.25 15.62 34.81

7.87 2.47 10.86 16.84

2.92 1.05

6.84 1.50

1.60 2.69

1.74 2.06

5.04 5.39

13.25 9.99

2.16

24.18

1.25 0.51

9.97

0.89

20.02 20.02

0.44

4.74 5.35

4.77 0.10 3.21 7.75

5.04 2.46

1.28 2.52

1.12

56.58

90.48

82.04

6.45

1.95 95.21

11.1

17.7

16.1

1.3

1.9

13.08 1.22

2.2. Click reaction of azide-sialic labelled glycans with Cy5-DBCO Azide sialic acid incorporated IgM was purified from the reaction mixture using a 10 mm × 300 mm Superdex200 Increase size exclusion column (GE Healthcare, 28990944), with a mobile phase of PBS containing 50 mM L-arginine and 5% (v/v) glycerol, at a flow rate of 0.4 ml/min. The void volume fractions containing the IgM were combined and concentrated using an Amicon 10 kDa MWCO centrifugal filter (Merck Millipore, UFC501024) to a volume of approximately 80 μl. 20 μg equivalent of the concentrated IgM was removed for quantitation of azide incorporation by released glycans (10 μg) and by intact glycopeptides (10 μg) analysis. The removed volume was replaced with an equal volume of Cy5-DBCO (Sigma Aldrich, 777374) to a final concentration of 100 μM, and incubated at room temperature overnight, in the dark. It is important to note that while Sigma catalog no. 777374 is sold as DBCO-Cy5, it is in fact the water soluble, sulphated DBCO-Cy5. The excess dye was removed by 4x 400 μl wash of PBS containing 50 mM L-arginine and 5% (v/v) glycerol in an Amicon 10 kDa MWCO centrifugal filter.

2. Materials and methods 2.1. Azide incorporation into IgM glycans using azido-CMP-sialic acid and ST6GAL1 enzyme 100 μg of PAT-SM6 IgM antibody with anti-cancer targeting capabilities [25] (Patrys Ltd, recombinantly produced by Patheon), was mixed with 2.5 μg of recombinant human ST6GAL1 (R&D systems, 7620-GT-010) in a 100 μl reaction mixture containing 10 mM HEPES pH 7.5, 10 mM MnCl2, 5 mM CaCl2, 5 mM MgSO4 and 100 μM of azidoCMP-sialic acid (R&D systems, ES102-100), and incubated overnight at 37 °C. The reaction was terminated by adding an equal volume of phosphate buffered saline (PBS, 10 mM phosphate buffer pH 7.4, 137 mM NaCl, 2.7 mM KCl); the phosphate ions react with the manganese required for enzyme activity to form insoluble manganese phosphate that was removed after the reaction by centrifugation at 14000 rpm for 1 min.

2.3. Confirmation of azide-sialic acid incorporation onto IgM glycans Glycan analysis was carried out as described previously [26]. In brief, the 10% aliquot of azide incorporated IgM was dot-blotted onto a PVDF membrane and glycans released with Rapid™ PNGaseF (NEB, 2

Analytical Biochemistry 584 (2019) 113385

E.S.X. Moh, et al.

Fig. 1. Schematic representation of the IgM functionalisation procedure. The azide functional group is first transferred onto the IgM glycans by the enzyme ST6GAL1 and can be specifically ligated with a cyclic-octyne functionalised probe using copper-free click chemistry.

P0711S) according to manufacturer formulation overnight at 37 °C. The glycans were then collected, reduced by 1 M NaBH4 in 50 mM KOH and desalted for analysis by PGC-LC-MS, using a 5 μm particle size, 2.1 mm × 100 mm Hypercarb (ThermoFIsher Scientific, 35005–102146) column with 10 mM ammonium bicarbonate aqueous solution (solvent A) and 10 mM ammonium bicarbonate aqueous solution in 70% acetonitrile (v/v) (solvent B) as mobile phases at a flow rate of 4 μL/min. The gradient program: 0 min, 2.6% B for 8 min; linear increase up to 13.5% B for 2 min; linear increase up to 33% B for 45 min; linear increase up to 64% B for 15 min; linear increase up to 98% B for 1 min; held at 98% B for 5min; and then equilibrated at 2.6% B for 7 min, before the next injection—giving a total LC run time of 83 min. The mass spectrometric data was acquired using a Thermo LTQ Velos, with parameters as previously described [27]. Glycan mass and fragmentation were analyzed manually and identified structures were quantified using Skyline [27].

azidoacetylneuraminic acid [28]. When using this methodology, the resulting mass difference between N-azidoacetylneuraminic acid and Nacetylneuraminic acid is 41 Da, from the addition of the azide group onto the acetyl group on the 5th carbon (R–COCH3 → R–COCH2N3). The resulting mass difference of 41 Da (accurate mass of 41.0014 Da) can cause misidentification of the glycan composition as the mass difference between naturally occurring N-acetylhexosamine and hexose is also 41 Da (accurate mass of 41.0266 Da). A recent publication developed a 9-azido sialic acid [29], replacing the –OH group on the 9th carbon with the –N3 group (Fig. 2, in-set), resulting in a mass difference of +25 Da (accurate mass of 25.0065 Da) between 9-azido-N-acetylneuraminic acid and native N-acetylneuraminic acid that is not a mass difference between any human glycan monosaccharide masses. The nucleotide sugar group (CMP) was then added by enzymatic activity to form azido-CMP-sialic acid and is currently commercially available (R& D systems, ES102-100) and used in this work.

2.4. Confirmation of azide-sialic acid incorporation onto IgM glycopeptides

3.2. Identification of azide incorporated glycans on PNGaseF released IgM glycans by MS

Glycopeptide analysis was carried out as described previously [24]. In brief, the 10 μg equivalent of azide incorporated IgM was reduced in 10 mM DTT at 50 °C for 1 h and alkylated with 25 mM IAA in the dark for 1 h. The mixture was split equally for protein digestion using 0.1 μg of trypsin (Promega, V5111) or 0.1 μg of GluC (Promega, V1651) in a 20 μl reaction volume containing 50 mM ammonium bicarbonate, pH 8, incubated overnight at 37 °C. Digested glycopeptides were added with 80 μl of 1% (v/v) trifluoroacetic acid in acetonitrile for glycopeptide enrichment using ZIC-HILIC columns as described previously [24]. The enriched glycopeptides were then analyzed by RP-LC-MS with CID fragmentation as described previously [24].

To verify that the azido-CMP-sialic acid could be transferred onto IgM glycans, the IgM was incubated with ST6GAL1 and azido-CMPsialic acid and purified by size exclusion chromatography. The N-glycans were subsequently released by PNGaseF, reduced by sodium borohydride and desalted for analysis by PGC-LC-MS/MS [26]. Azide-incorporated glycans released from the PAT-SM6 IgM antibody were readily identified by mass spectrometry, carrying an additional 25Da (12.5Da on a doubly charged ion) compared to the native sialylated glycans of identical monosaccharide composition (Fig. 2). It is important to note that as part of the glycan sample preparation in order to avoid separation of the α and β reducing end anomers, sodium borohydride was used as a reducing agent to reduce the alcohol to an alditol at the reducing end of the glycan. In aqueous conditions and in absence of a catalyst such as palladium on carbon [30], reduction of the alkyl-azide group to alkyl-amine by sodium borohydride is a low yielding process [31] and has not been observed to complicate the MS data interpretation. Fragmentation of the glycan composition precursors ions containing the +25Da azide showed that the +25Da added mass corresponded to sialic acid containing fragment ions (Fig. 3), validating that the transfer of the azido-sialic acid onto galactose terminating glycans was successful.

2.5. Visualization of Cy5-labelled IgM by SDS-PAGE An aliquot corresponding to approximately 5 μg of concentrated Cy5-labelled IgM was ran on an SDS-PAGE gel for fluorescent visualization of the conjugation. The SDS-PAGE gel was first fixed in 50% (v/ v) ethanol, 10% (v/v) acetic acid for 15 min, washed twice with distilled water and visualized using the Typhoon Trio imager (GE Healthcare) at 633 nm with the photomultiplier tube set at 250 V for imaging of Cy5 fluorescence. Subsequently, the same gel was stained with Coomassie Blue for detection of protein bands for comparison to the Cy5 fluorescence.

3.3. Site-identification of azide incorporated into IgM glycopeptides 3. Results Based on the site-specific glycosylation analysis of IgM PAT-SM6 described in our previous work [24], the incorporation of the azidoCMP-sialic acid could only occur at three out of the five N-glycosylation sites on the IgM heavy chain (Asn171, Asn332, Asn395) and the single glycosylation site on the J-chain (Asn71), as these sites carry > 95% of the precursor galactose terminating glycan substrates of ST6GAL1

3.1. Incorporation of azide into glycans for validation by MS In vivo metabolic incorporation of azides onto sialic acids has been previously reported by feeding per-acetylated N-azidoacetylmannosamine to cells, where cellular activity converts it into N3

Analytical Biochemistry 584 (2019) 113385

E.S.X. Moh, et al.

Fig. 2. Averaged mass spectrum of complex-type N-glycans over retention time on PGC-liquid chromatography. The unique mass shift of 25 Da was readily observed for glycans containing the azide-functionalized sialic acid transferred by ST6GAL1.

and were reported to be a rapid and high yielding reaction [30,32]. However, we did not find spectral data to suggest that this azide reduction occurred; low amounts, approximately 5% relative compared to the azido-sialic acid form, of a precursor mass corresponding to a potentially aminated-sialic acid containing glycopeptide was observed, but there were no tandem MS fragmentation masses indicating the presence of an aminated-sialic acid oxonium ion. Compared with the previous reports describing the azide to amine reduction [30,32], the experimental conditions used in this work and the azido-moiety were

(Table 1). Using CID fragmentation, which predominantly fragments glycopeptides at the glycosidic bonds, the azido-sialic acid containing oxonium ions (m/z 299 = azido-sialic acid-18, m/z 317 = azido-sialic acid, m/z 682 = azido-sialyl-lacNac) were strong diagnostic ions for the identification of azido-glycopeptides at all four above mentioned sites (Fig. 4). As part of the glycopeptide sample preparation, the IgM was reduced by dithiothreitol prior to proteolytic digestion and mechanisms of di-thiol reduction of azides to amines have been described in the 80's

Fig. 3. Tandem MS fragmentation of mass of m/z 1038.92− corresponding to the bi-antennary monosialylated glycan with a core fucose, and the +12.5 Da counterpart mass of m/z 1051.42− of the same composition with an incorporated azide. The +25 Da mass increase in only the sialic acid containing fragment ions validates that the azide-incorporation via ST6GAL1 activity was successful. 4

Analytical Biochemistry 584 (2019) 113385

E.S.X. Moh, et al.

Fig. 4. Tandem MS CID fragmentation of the most abundant azide incorporated glycopeptide identified at each of the complex glycosylation site. The azido-sialic acid containing oxonium fragment ions were diagnostic and were identified in each of the tandem MS fragments of azide incorporated glycopeptides. The most abundant glycoforms identified correspond to the most abundant azide incorporated glycans (as shown in Fig. 2).

calculated for each glycosylation site (area under curve of all azide incorporated glycopeptide/area under curve of total glycopeptide), along with the distribution of these sites on the pentameric IgM. From this approach, the AAR was found to be in the range of 6–10 (1–2 for Asn171, 3–4 for Asn332, 2–3 for Asn395 and 0–1 for J-chain), which agreed with the estimate based on the released glycan analysis. A potential DAR of 6–10 is higher compared to the IgG-ADC, which usually has a DAR of 2–4 [11,33,34]. Although the incorporation of the azidesialic acid was incomplete in this example (based on the terminal galactose substrate availability of 31) optimization of AAR will need to be determined for the addition of a conjugated azide-monosaccharide depending on the efficiency and accessibility of the specific glycosyltransferase and IgM used for ADC applications.

different and we believe there is no evidence to suggest that the azide that was incorporated into the glycopeptides was reduced.

3.4. Azide-antibody ratio (AAR) quantitation Estimation of the Drug-Antibody Ratio (DAR) for antibody-drug conjugates is important, as it relates to drug loading and dosage [11]. In this work, we quantified the number of azide groups incorporated onto the glycans and glycopeptides to calculate the azide-antibody ratio (AAR). Click chemistry reaction between azides and strain-promoted cyclic alkynes has been well established and is known to be fast, high yielding and can be driven to completion. Hence, under the assumption that AAR will be equal to a DAR, we performed relative quantitation of the azide incorporation onto the glycans and glycopeptides (Fig. 5) from the MS data. By calculating the relative abundance of the azide-sialic acid incorporated glycans against the total abundance of complex glycans using area under curve quantitation, we calculate that approximately 26% of the complex glycans incorporated the azido-sialic acid. As described in our previous site-specific IgM glycopeptide quantitation data [24], a total of 31 glycosylation sites contributed to carrying > 95% of the complex glycan species, and when distributed across the 31 glycosylation sites, this glycan abundance approximates to an AAR of 8. From the glycopeptide quantitation after the azide incorporation, the relative abundance of azide-sialic acid incorporated glycopeptides was

3.5. Fluorescence visualization of IgM-Cy5 conjugate by SDS-PAGE To verify that the azide-incorporated IgM could be labelled by the “click” chemistry addition of an alkyne to the azido–sialic acid, a fluorescent small molecule Cy5-Dibenzocyclooctyne (DBCO) was ligated to the azide by copper-free click chemistry, followed by visualization by SDS-PAGE. Cy5 label is commonly used in two-dimensional fluorescence difference gel electrophoresis (2D-DIGE) and is not sensitive to conditions used in SDS-PAGE (boiling temperatures and reducing agents) [35]. Only the IgM heavy chain, where the majority of the complex glycosylation sites are located was highly fluorescent by 5

Analytical Biochemistry 584 (2019) 113385

E.S.X. Moh, et al.

Fig. 5. Relative quantitation of azide incorporation into complex glycans (pie chart) and complex glycopeptides (bar graph). For the complex glycans, the relative abundances of the azide incorporated complex glycans are shown. The result was directly correlated with the analysis of the glycopeptides which carry these complex glycans on the IgM. From the glycopeptide data, the relative abundance of the azide containing glycopeptides was charted against the number of glycosylation sites on the pentameric IgM for estimation of the AAR.

approaches [9,11,36,37], with either intact native IgG, deglycosylated IgG or disulphide-reduced IgG material. While these methodologies are desirable as they involve minimal sample preparation, they are not suitable for characterization of IgM-ADC, due to the higher level of complexity and size of the IgM molecule. Glycosylation of proteins increases the heterogeneity of the protein sample, resulting in one protein having more than one mass, due to different numbers of glycans and/or glycan compositions. With IgG, there had been much success in topdown mass spectrometry [9,36,37], as IgG is only approximately 150 kDa in size and there is only 2 glycans per IgG molecule, with predominantly complex bi-antennary glycans. Pentameric IgM, however, is close to 1 MDa in size, with 51 glycosylation sites, and a wide spread of glycan compositions ranging from high mannose, hybrid to biand tri-antennary glycans [24]. The biggest challenge in top-down mass spectrometry is the deconvolution of the protein charge envelope, that converts the observed m/z signal of the intact protein to its molecular weight. Under disulphide-reduced conditions, while a deglycosylated single IgM heavy chain is approximately 75 kDa but with the presence of 5 sites of glycosylation with numerous compositions on each site, resolving mass differences between heavy chain glycoforms is not meaningful. Using the bottom-up approach for quantitation of azide incorporation, we showed that the AAR could be quantified by either analysis of PNGaseF released glycans, or by enriched, intact glycopeptide analysis. Under existing sample preparation protocols used in this work, the azide groups were unreduced, charge neutral and could be detected by mass spectrometry. Their accessibility to conjugation was validated by fluorescent alkyne (DBCO-Cy5) reactivity by click chemistry. In the case of the attachment of a small molecule drug or probe that is more hydrophobic or contains charge carrying groups, the ionization efficiency of the glycan or glycopeptides may change and render area under curve quantitation less accurate after conjugation. Hence, we suggest that bottom up quantitation based of AAR may provide a more accurate quantitation if required compared to bottom up quantitation of DAR. Based on the glycan and glycopeptide analysis, we found that the AAR was 8 (glycan analysis) or within the range of 6–10 (glycopeptide analysis). Correlating fluorescence intensities of a Cy5-DBCO standard curve with UV absorption of known IgM concentrations found that the Cy5:IgM ratio (equivalent to DAR in this case) was also within the range of the estimated AAR (data not shown). Having synthesized ADC, their use for targeted drug treatment requires three major steps, the binding of the ADC to the target cell, internalization of the ADC and the release of the drug within the target cell by the same mechanisms, such as cleavable linkers, that are currently used with IgG mediated ADCs [38]. Importantly, ADC created by antibody glycan functionalization does not affect antibody binding as all the IgM glycosylation sites are away from the Fab region. Internalization of an antigen-bound IgM into the target cell has been previously reported with the IgM used in this work, PAT-SM6 [4], confirming the potential of developing an active site-specific ADC. In the previous reported example of IgM Fcμ-ADC [18], in which only one

Fig. 6. SDS-PAGE of azide labelled PAT-SM6, visualized by Coomassie blue or Cy5 fluorescence (633 nm) of the “Click” reaction between DBCO-Cy5 and azide-sialic acid labelled and unlabeled IgM. Cy5 fluorescence was only observed for the azide labelled IgM on both the heavy chain and the J-chain.

detection at 633 nm (Fig. 6), as compared to the non-enzyme treated IgM. The J-chain, which was not visible under Coomassie stain due to the much lower abundance (1 J-chain for every 10 IgM heavy/light chain), was also observed as being fluorescently labelled, suggesting that despite the low abundance of azide-incorporation onto the J-chain, the click reaction was still successful. 4. Discussion For all existing ADC, in trials or in development, characterization of the DAR of the ADC is crucial. With IgG-ADC, DAR estimation is mostly performed by top-down or middle-down mass spectrometric 6

Analytical Biochemistry 584 (2019) 113385

E.S.X. Moh, et al.

selenocysteine was incorporated into each pentameric Fcμ (i.e DAR 1), the Fcμ-ADC increased the potency of killing chronic lymphocytic leukemia cells by 1000-fold in vitro. Applying our method of conjugating IgM could increase the DAR, and the efficacy even further. In conclusion, we describe site specific IgM-ADC using a single step enzymatic reaction, followed by the highly efficient click chemistry quantitative addition of a drug or of an imaging molecule for microscopy applications. We also identified the location and amount of azide incorporated onto the IgM to quantify the theoretical DAR.

[14]

[15]

[16] [17]

Declaration of interest [18]

None.

[19]

Acknowledgements

[20]

PAT-SM6 antibody is proprietary to Patrys Limited (www.patrys. com), and we would like to thank Ms Valentina Dubljevic her provision of the antibody material. This work was supported by the Australian Research Council (CE140100003).

[21]

[22]

Appendix A. Supplementary data

[23]

Supplementary data to this article can be found online at https:// doi.org/10.1016/j.ab.2019.113385.

[24]

References [25]

[1] S. Panda, J.L. Ding, Natural antibodies bridge innate and adaptive immunity, J. Immunol. 194 (2015) 13–20. [2] C. Gronwall, J. Vas, G.J. Silverman, Protective roles of natural IgM antibodies, Front. Immunol. 3 (2012) 66. [3] H. Peter Vollmers, R. O'Connor, J. Müller, T. Kirchner, H.K. Müller-Hermelink, SC1, a functional human monoclonal antibody against autologous stomach carcinoma cells, Cancer Res. 49 (1989) 2471–2476. [4] S. Brandlein, N. Rauschert, L. Rasche, A. Dreykluft, F. Hensel, E. Conzelmann, H.K. Muller-Hermelink, H.P. Vollmers, The human IgM antibody SAM-6 induces tumor-specific apoptosis with oxidized low-density lipoprotein, Mol. Cancer Ther. 6 (2007) 326–333. [5] R.F. Irie, D.W. Ollila, S. O'Day, D.L. Morton, Phase I pilot clinical trial of human IgM monoclonal antibody to ganglioside GM3 in patients with metastatic melanoma, Cancer Immunology, Immunotherapy 53 (2004) 110–117. [6] S. Wolbank, R. Kunert, G. Stiegler, H. Katinger, Characterization of human classswitched polymeric (immunoglobulin M [IgM] and IgA) anti-human immunodeficiency virus type 1 antibodies 2F5 and 2G12, J. Virol. 77 (2003) 4095–4103. [7] X. Piao, T. Ozawa, H. Hamana, K. Shitaoka, A. Jin, H. Kishi, A. Muraguchi, TRAILreceptor 1 IgM antibodies strongly induce apoptosis in human cancer cells in vitro and in vivo, OncoImmunology 5 (2016) e1131380. [8] Y. Yang, G. Wang, T. Song, C.B. Lebrilla, A.J.R. Heck, Resolving the micro-heterogeneity and structural integrity of monoclonal antibodies by hybrid mass spectrometric approaches, mAbs 9 (2017) 638–645. [9] R.Y.C. Huang, G. Chen, Characterization of antibody–drug conjugates by mass spectrometry: advances and future trends, Drug Discov. Today 21 (2016) 850–855. [10] F. Tian, D. Jackson, Y. Bai, Site-specific antibody-drug conjugates, in: M. Damelin (Ed.), Innovations for Next-Generation Antibody-Drug Conjugates, Springer International Publishing, Cham, 2018, pp. 241–265. [11] P. Agarwal, C.R. Bertozzi, Site-specific antibody–drug conjugates: the nexus of bioorthogonal chemistry, protein engineering, and drug development, Bioconjug. Chem. 26 (2015) 176–192. [12] N. Dan, S. Setua, V. Kashyap, S. Khan, M. Jaggi, M. Yallapu, S. Chauhan, Antibodydrug conjugates for cancer therapy: chemistry to clinical implications, Pharmaceuticals 11 (2018) 32. [13] O. Duramad, B. Wang, F. Zheng, L. Keyt, C. Repellin, L. Beviglia, N. Bhat, M. Bieber, N. Teng, B. Keyt, Abstract 645: IGM-55.5, a novel monoclonal human recombinant

[26] [27]

[28]

[29] [30] [31] [32] [33] [34] [35] [36]

[37]

[38]

7

IgM antibody with potent activity against B cell leukemia and lymphoma, Cancer Res. 74 (2014) 645. B. Wang, T. Kothambawala, P. Hinton, D. Ng, A. Saini, R. Baliga, B. Keyt, Multimeric anti-DR5 IgM antibody displays potent cytotoxicity in vitro and promotes tumor regression in vivo, American Association for Cancer Research Annual Meeting 2017 (2017) Abstract 1702. A. Tchoudakova, F. Hensel, A. Murillo, B. Eng, M. Foley, L. Smith, F. Schoenen, A. Hildebrand, A.R. Kelter, L.L. Ilag, H.P. Vollmers, S. Brandlein, J. McIninch, J. Chon, G. Lee, M. Cacciuttolo, High level expression of functional human IgMs in human PER.C6 cells, mAbs 1 (2009) 163–171. C. Valasek, J. Cole, F. Hensel, P. Ye, M. Conner, Production and purification of a PER. C6-expressed IgM antibody therapeutic, BioProcess Int 9 (2011) 28–37. J. Lee, H.T. Gan, S.M. Latiff, C. Chuah, W.Y. Lee, Y.S. Yang, B. Loo, S.K. Ng, P. Gagnon, Principles and applications of steric exclusion chromatography, J. Chromatogr. A 1270 (2012) 162–170. B. Vire, M. Skarzynski, J.D. Thomas, C.G. Nelson, A. David, G. Aue, T.R. Burke, C. Rader, A. Wiestner, Harnessing the Fcμ Receptor for Potent and Selective Cytotoxic Therapy of Chronic Lymphocytic Leukemia, Cancer Res, 2014. G.T. Hermanson, Chapter 2 - Functional Targets for Bioconjugation, Bioconjugate Techniques, third ed., Academic Press, Boston, 2013, pp. 127–228. D.J. O'Shannessy, M.J. Dobersen, R.H. Quarles, A novel procedure for labeling immunoglobulins by conjugation to oligosaccharide moieties, Immunol. Lett. 8 (1984) 273–277. Q. Zhou, J.E. Stefano, C. Manning, J. Kyazike, B. Chen, D.A. Gianolio, A. Park, M. Busch, J. Bird, X. Zheng, H. Simonds-Mannes, J. Kim, R.C. Gregory, R.J. Miller, W.H. Brondyk, P.K. Dhal, C.Q. Pan, Site-specific antibody-drug conjugation through glycoengineering, Bioconjug. Chem. 25 (2014) 510–520. B.M. Zeglis, C.B. Davis, R. Aggeler, H.C. Kang, A. Chen, B.J. Agnew, J.S. Lewis, Enzyme-mediated methodology for the site-specific radiolabeling of antibodies based on catalyst-free click chemistry, Bioconjug. Chem. 24 (2013) 1057–1067. P.K. Qasba, E. Boeggeman, B. Ramakrishnan, Site-specific linking of biomolecules via glycan residues using glycosyltransferases, Biotechnol. Prog. 24 (2008) 520–526. E.S. Moh, C.H. Lin, M. Thaysen-Andersen, N.H. Packer, Site-specific N-glycosylation of recombinant pentameric and hexameric human IgM, J. Am. Soc. Mass Spectrom. 27 (2016) 1143–1155. T. Pohle, S. Brandlein, N. Ruoff, H.K. Muller-Hermelink, H.P. Vollmers, Lipoptosis: tumor-specific cell death by antibody-induced intracellular lipid accumulation, Cancer Res. 64 (2004) 3900–3906. P.H. Jensen, N.G. Karlsson, D. Kolarich, N.H. Packer, Structural analysis of N- and O-glycans released from glycoproteins, Nat. Protoc. 7 (2012) 1299–1310. C. Ashwood, C.-H. Lin, M. Thaysen-Andersen, N.H. Packer, Discrimination of isomers of released N- and O-glycans using diagnostic product ions in negative ion PGC-LC-ESI-MS/MS, J. Am. Soc. Mass Spectrom. 29 (6) (2018) 1194–1209. J.M. Baskin, J.A. Prescher, S.T. Laughlin, N.J. Agard, P.V. Chang, I.A. Miller, A. Lo, J.A. Codelli, C.R. Bertozzi, Copper-free click chemistry for dynamic in vivo imaging, Proc. Natl. Acad. Sci. 104 (2007) 16793–16797. Z.L. Wu, X. Huang, A.J. Burton, K.A.D. Swift, Glycoprotein labeling with click chemistry (GLCC) and carbohydrate detection, Carbohydr. Res. 412 (2015) 1–6. I.L. Cartwright, D.W. Hutchinson, V.W. Armstrong, The reaction between thiols and 8-azidoadenosine derivatives, Nucleic Acids Res. 3 (1976) 2331–2340. E.F.V. Scriven, K. Turnbull, Azides: their preparation and synthetic uses, Chem. Rev. 88 (1988) 297–368. A.L. Handlon, N.J. Oppenheimer, Thiol reduction of 3′-azidothymidine to 3′-aminothymidine: kinetics and biomedical implications, Pharm. Res. 5 (1988) 297–299. S. Panowski, S. Bhakta, H. Raab, P. Polakis, J.R. Junutula, Site-specific antibody drug conjugates for cancer therapy, mAbs 6 (2014) 34–45. A. Beck, L. Goetsch, C. Dumontet, N. Corvaïa, Strategies and challenges for the next generation of antibody–drug conjugates, Nat. Rev. Drug Discov. 16 (2017) 315. N.S. Tannu, S.E. Hemby, Two-dimensional fluorescence difference gel electrophoresis for comparative proteomics profiling, Nat. Protoc. 1 (2006) 1732–1742. R.Y.-C. Huang, E.G. Deyanova, D. Passmore, V. Rangan, S. Deshpande, A.A. Tymiak, G. Chen, Utility of ion mobility mass spectrometry for drug-to-antibody ratio measurements in antibody-drug conjugates, J. Am. Soc. Mass Spectrom. 26 (2015) 1791–1794. H.J. Wang, T. Zhang, B.J. Agnew, R. Viner, J. Josephs, Comprehensive Characterization of Site-specific Engineered Antibody Drug Conjugate by Orbitrap Mass Spectrometer Poster Note 64423 at 63rd ASMS Conference on Mass Spectrometry and Allied Topics, (2015). J.D. Bargh, A. Isidro-Llobet, J.S. Parker, D. R Spring, Cleavable linkers in antibodydrug conjugates, Chem. Soc. Rev. (2019), https://doi.org/10.1039/C8CS00676H (In press).