Unlocking the Complexity of Antibody-Drug Conjugates: A Cutting-Edge LC-HRMS Approach to Refine Drug-to-Antibody Ratio Measurements with Highly Reactive Payloads

Abstract

The complexity of therapeutic proteins like antibody-drug conjugates (ADCs) holds a tremendous analytical challenge. Complementary mass spectrometry approaches such as peptide mapping and intact mass analysis are required for the in-depth characterization of these bioconjugates. Cysteine-linked ADCs have shown a unique challenge for characterization, mainly when the conjugation is carried out on interchain cysteines, because their intact analysis requires native mass spectrometry conditions to preserve non-covalent binding between antibody chains. In this work, two different approaches were proposed. Specifically, a full scan data-independent all ion fragmentation (FS-AIF) and a full scan data-dependent targeted MS2 (FS-ddtMS2) were applied to generate complementary datasets for a cysteine-linked ADC characterization with a highly reactive payload. These two methods were applied to in vitro plasma stability and in vivo PK samples to calculate and refine mean drug-to-antibody ratio over time. Using this approach, we successfully characterized an ADC containing a hydrolysis-sensitive payload and refined the "active" drug-to-antibody ratio on in vitro stability and in vivo samples. These two methods allowed the confirmation of the different ADC species and potential metabolites of conjugated payload attached to the antibody backbone in a single analysis without needing a dedicated method for the conjugated payload metabolite identification.

Keywords: antibody–drug conjugates; drug discovery; drug-to-antibody ratio; liquid-chromatography mass spectrometry.

Figure 1

Schematic structure of an antibody–drug conjugate (ADC) and potential strategies for mass spectrometry characterization. The most common strategy is to characterize ADCs via intact protein or bottom-up LC-MS analyses, making it harder to identify antibody-conjugated payload metabolites (MS1-only strategy). An alternative strategy can be applied by taking advantage of the cleavability feature of the linker. This approach promotes the formation of payload fragments in the gas phase during mass spectrometry analysis. By triggering MS2 fragmentation, this method enables the potential identification of small molecule metabolites derived from the ADC payload.

Figure 2

(A) HCD/CID-based linker fragmentation pattern in mass spectrometer Ion Routing Multiple for ADC linker-payload system. (B) Potential spots on linker-payload that are sensitive to hydrolysis in plasma. The succinimide ring is highlighted in blue. Linker payload is conjugated on interchain cysteines (one linker payload is shown for simplification).

Figure 3

Total Ion Chromatogram for conjugated LC and HC. (A) ADC-A glycosylated Heavy chain MS spectrum at 0 h in mouse plasma. (B) ADC-A light chain MS spectrum at 0 h in mouse plasma. (C) Total Ion chromatogram of separated conjugated HC (R_T = 21.8 min), LC (R_T = 23.40 min) and half-antibody (R_T = 25.84 min) of ADC-A at 0 h in mouse plasma in vitro stability. (D) ADC-A glycosylated half-antibody MS spectrum at 0 h in mouse plasma. (E) Deconvoluted spectrum of ADC-A light chain MS spectrum at 0 h in mouse plasma. (F) Deconvoluted spectrum of ADC-A heavy chain and half-antibody MS spectrum at 0 h in mouse plasma.

Figure 4

Hydrolysis rate evaluation for ADC-A LC-conjugated chain. (A) MS1 spectra of different in vitro mouse plasma stability samples, zooming around m/z 1364.25, corresponding to the unmodified LC DAR 1 ADC species. (B) Targeted-MS2 fragmentation ion mass spectra for the three identified LC-conjugated payload species in the survey MS1 scan used to calculate the payload hydrolysis rate. From 24 h on, the hydrolyzed antibody-conjugated payload species started to appear (mainly in bi-hydrolyzed species).

Figure 4

Hydrolysis rate evaluation for ADC-A LC-conjugated chain. (A) MS1 spectra of different in vitro mouse plasma stability samples, zooming around m/z 1364.25, corresponding to the unmodified LC DAR 1 ADC species. (B) Targeted-MS2 fragmentation ion mass spectra for the three identified LC-conjugated payload species in the survey MS1 scan used to calculate the payload hydrolysis rate. From 24 h on, the hydrolyzed antibody-conjugated payload species started to appear (mainly in bi-hydrolyzed species).

Figure 5

Different DAR calculation hypotheses impact and comparison to actual MS2-refined DAR mean value for ADC-A in vitro mouse plasma stability samples. (A) ADC-A MS2-refined LC DAR analysis. LC can be conjugated up to 1 linker payload (LC DAR 1). LC DAR 0 is the unmodified light chain or LC with inactive hydrolyzed payload. Different hypotheses for DAR calculation can be made in advance. Red trace considered mono-hydrolyzed LC species as fully “active” species (e.g., DAR1 + 1 hydrolysis, namely DAR1_1hydr) and bi-hydrolyzed species as fully “inactive” species (e.g., DAR1 + 2 hydrolysis, DAR1_2hydr, assigned as DAR0). Green trace considered mono and bi-hydrolyzed species as inactive species (e.g., both DAR1_1hydr and DAR1_2hydr species as DAR0, then considering the payload more hydrolysis-sensitive than succinimide ring linker). Purple trace considered mono and bi-hydrolyzed species as active species (e.g., both DAR1_1hydr and DAR1_2hydr species as DAR1). Cyan trace is an unrealistic scenario of having mono-hydrolyzed species as fully inactive (e.g., DAR0) and bi-hydrolyzed species as fully active (e.g., DAR1). The blue trace is the experimentally MS2-refined DAR value for ADC-A conjugated LC using targeted DDA MS/MS experiments. (B) Hydrolyzed over total payload ratio for the mono and bi-hydrolyzed LC DAR1 species. The red line represents the average total hydrolyzed payload (mono + bi-hydrolyzed LC DAR1 species) over the total payload ratio for ADC-A LC. (C) ADC-A MS2-refined HC and half-antibody DAR analysis. (D) Hydrolyzed over total payload ratio for the identified mono-hydrolyzed HC DAR3 G0 and bi-hydrolyzed glycosylated half-antibody (LC-HC DAR2 G0 2hyd). Black and blue lines represent the total hydrolyzed payload over the total payload ratio for ADC-A half-antibody and HC, respectively. (E) Total ADC DAR and refined mean DAR of individual ADC chains.

Figure 6

DAR decay regression analysis of ADC-A in in vitro mouse plasma stability. DAR loss over time regression analysis for LC (left) and HC and half-antibody (right) ADC species, slope represents the DAR loss unit/h in the time frame of in vitro mouse plasma stability at 37 °C and 5% CO₂.

Figure 7

In vivo PK MS2-refined DAR analysis of ADC-A from FS-AIF experiments. (Left) Mean DAR value for the three main ADC conjugated species, LC, HC and half-antibody (n = 3) from three C57BL/6N treated mice. Nominal DAR of the chains for this ADC is shown and compared to the ADC standard in buffer (STD in the plot). Bars show 95% confidence interval (2 standard deviations). The asterisk indicates the statistical significance of p-value (* corresponds to 95% significance, α = 0.05). (Right) Percentage of hydrolyzed conjugated payload for the three main ADC conjugated species, LC, HC and half-antibody (n = 3) from three C57BL/6N treated mice. Bars show 95% confidence interval (2 standard deviations).