Methods

Sample preparation: The lysine-linked ADC sample (adotrastuzumab emtansine, T-DM1) was denatured with 7.2M guanidine hydrochloride, 100mM Tris buffer pH 7.2, followed by reduction with 10mM DL-dithiothreitol and alkylation with 30 mM iodoacetamide. Digestion was performed with trypsin/Lys-C enzyme at 37°C for 16 h.

Chromatography: 4 µL (4 µg) of the trypsin/Lys-C digest were separated with a CSH C18 column (2.1×100 mm, 1.7 μm, 130 Å, Waters) using an ExionLC AD system. The mobile phase A consisted of water with 0.1% formic acid, while the organic phase B was acetonitrile 0.1% formic acid. A gradient profile was used at a flow rate of 300 μL/min (Table 1). The column temperature was maintained at 50°C.here

Mass spectrometry: Data were acquired with an informationdependent acquisition (IDA) method using the SCIEX ZenoTOF 7600 system. The electron energy for EAD cell was set to a value of 7 eV. Detailed method parameters are summarized in Table 2.

Data processing: Data were processed using Byos software (Protein Metrics Inc.) using customized modifications. 

The what, why and how

Ado-trastuzumab emtansine (T-DM1) was amongst the first ADC therapeutics, receiving approval by the FDA in 2013 for the treatment of human epidermal growth factor receptor 2 (HER2)- positive, metastatic breast cancer. T-DM1 consists of a monoclonal antibody, trastuzumab, which is covalently conjugated to the cytotoxic agent emtansine (DM1) via a noncleavable linker (Figure 2). Combining target-specificity of a mAb with the high potency of a cytotoxic drug takes advantage of the best of two worlds, minimizing side effects.³ T-DM1 is linked to amines, such as in the side chain of lysine residues of trastuzumab. Previous intact mass studies show that the average DAR for T-DM1 is around 3.5.^5,6 However there are 88 lysine residues and 4 N-terminal groups in trastuzumab, which could result in more than 4.5 million unique molecules.⁵ The site and structure of payload will directly affect drug efficacy and safety, which are therefore classified as critical quality attributes (CQA) and require comprehensive characterization and tight monitoring during development. Currently, bottom-up approaches are the method of choice for the characterization of product quality attributes, enabling the simultaneous identification and localization of modifications. LC-MS/MS utilizing CID is commonly used to verify conjugation sites, as each DM1 will lead to a ~957 amu mass shift. However, besides the dissociation of the peptide backbone, CID results in a series of small fragments from the payload drug, such as m/z of 547.221, 485.224, and 453.199, which increase the spectra complexity.⁵ Although alternative fragmentation technology such as electron capture dissociation (ECD) is expected to provide orthogonal information on drug conjugated peptides, the application has not been explored extensively. 

With the SCIEX ZenoTOF 7600 system, a robust alternative fragmentation technique is introduced, enabling scientists to get an in-depth picture of their samples by using analytical-flow LC separation in combination with a fast scanning DDA method and processing using Protein Metrics Inc. software. This breakthrough technology realizes the dream of answering complex questions in a routine and reproducible manner.

Analysis of conjugated peptides

The study focused on the characterization of a commercialized ADC: T-DM1. The conjugation reaction between DM1 and trastuzumab is of stochastic nature, targeting amines. Out of the 88 lysine residues on trastuzumab, 40 are solvent exposed6 and therefore susceptible to conjugation. Multiple approaches on intact and subunit mass have been explored to study the DAR. These approaches, however, cannot reveal the exact sites of the conjugation. 

Here, a DDA approach in combination with Zeno EAD was chosen. With this approach, routine peptide mapping analyses can be performed, while EAD enables advanced characterization in the same, single analysis. Furthermore, the detection of fragment ions, and thus the correct identification of low abundant species, is enhanced by Zeno EAD. This approach allowed for the straight-forward data interpretation using Protein Metrics Inc. software. It is the first exploration of SCIEX EAD technology for an ADC sample. Figure 3 demonstrates an example of the fragmentation pattern observed on a conjugated peptide, SCDK[DM1]THTCPPCPAPELLGGPSVFLFPPKPK. The peptide without linker and drug or part of it was not observed in the analysis indicating a full conjugation. High quality MS/MS spectra was achieved for the peptide, allowing for 96.6% MS/MS sequence coverage for this particular peptide. One of the more dominant fragments was derived from the payload with an m/z greater than 500 (see labeling in Figure 3). The major cleavage site observed for the payload structure was the COO-C bond of the DM1. This fragmentation pattern is different from previous published data leveraging CID, which produced a series of small fragments.⁵ Larger fragments of the drug can be used as signature fragments to confirm the existence of the payload more specifically, and can be leveraged to confirm the payload structure. Furthermore, fragments from the peptide backbone were also well detected by applying Zeno EAD for enhanced fragment ion detection, providing information on the molecular integrity of the peptide. The existence of conjugated drugs on the protein can lead to more missed cleavages during enzymatic sample preparation, due to steric hinderance of the enzyme. In addition, the conjugation process between lysine residues and the payload is a stochastic reaction. The occupancy of the conjugation is not always 100%, which results in diverse and low abundant forms. When multiple potential positions exist in one peptide, it can be a challenge to identify the correct site of linkage. The peptide ASQDVNTAVAWYQQKPGKAPK is another example of this type of challenging peptide (Figure 4). It contains a miscleavage site and a lysine N-terminal to a proline, leading to multiple options for the site of conjugation. However, with the rich, high-quality MS/MS spectra derived from EAD, an automatic assignment of the drug localization was achieved (Figure 4A). Because the payload is close to the C-terminus of the peptide, more abundant c ions are detected than z ions (Figure 4A), while the unconjugated peptide shows abundant fragments from both the C-terminus and N-terminus (Figure 4B). A full series of c fragments from c3 to c17 was detected, except for c15, since it is well known that electron capture dissociation techniques do not dissociate the N-terminal side of prolines.⁷ This provides solid evidence that K15 is not linked to the drug. In addition, z4, z5 and z7 show that K18, not K21, is the correct site of drug conjugation.

This is further supported by the fact that conjugated lysines are not cleavable with trypsin⁵, excluding K21 as the site of attachment. The most dominant species in the spectrum in Figure 4A (m/z 547.221) can be linked to the dissociation of the payload. The isotope pattern supports this finding since it fits a compound containing a halogen element (Cl). The associated counterpart linked to the peptide was also observed (Figure 4A). Unambiguous characterization, including the identification and localization of conjugated peptides in a single DDA run, was achieved with the SCIEX ZenoTOF 7600 system and Zeno EAD. This is an example of how ADC analysis, which was previously thought a challenge by LC-MS/MS, can be simplified by Zeno EAD.

Conclusions
 

• The exact site of drug conjugation in peptides with multiple potential locations was achieved with the novel fragmentation technique of EAD 
  
  
• MS/MS fragment detection was significantly enhanced compared to traditional MS/MS analysis. Utilizing Zeno EAD resulted in remarkable data quality for confident fragment assignment, even for precursors with medium or very low intensities, such as low abundant conjugated peptides 
  
  
• The robust, reproducible and easy-to-use alternative fragmentation of the SCIEX ZenoTOF 7600 system enables users to answer challenging analytical questions in a streamlined manner 
  
  
• Automatic data processing using Protein Metrics Inc. software enables the reproducible, routine and advanced characterization of complex biotherapeutics 

References
 

1. Baba T et al. (2015) Electron capture dissociation in a branched radio-frequency ion trap, Anal Chem, 87, 785−792. 
   
   
2. Baba T et al. (2021) Dissociation of biomolecules by an intense low-energy electron beam in a high sensitivity time-of-flight mass spectrometer. Accepted by JASMS. 
   
   
3. Beck, Alain et al. (2010) Strategies and challenges for the next generation of therapeutic antibodies. Nat Rev Immunol 10(5):345-52. 
   
   
4. Botzanowski, Thomas et al. (2017) Insights from native mass spectrometry approaches for top- and middle- level characterization of site-specific antibody-drug conjugates. Mabs, 9(5):801-811. 
   
   
5. Chen, Liuxi et al. (2016) In-depth structural characterization of Kadcyla (ado-trastuzumab emtansine) and its biosimilar candidate. Mabs, 8(7): 1210-1223. 
   
   
6. Harris Reed et al. (2015) Native mass spectrometry and ion mobility characterization of trastuzumab emtansine, a lysinelinked antibody drug conjugate Protein Sci 24(8):1210-1223. 
   
   
7. Fort, Kyle et al (2018) Exploring ECD on a benchtop Q Exactive orbitrap mass spectrometer. J Proteome Res 17(2) :926-933.