Peroxide-cleavable linkers for antibody–drug conjugates

Antibody–drug conjugates containing peroxide-cleavable arylboronic acid linkers are described, which target the high levels of reactive oxygen species (ROS) in cancer. The arylboronic acid linkers rapidly release a payload in the presence of hydrogen peroxide, but remain stable in plasma. Anti-HER2 and PD-L1 peroxide-cleavable ADCs exhibited potent cytotoxicity in vitro.


Bioconjugation
To a solution of trastuzumab or durvalumab (40 μL, 16.9 μM, 2.5 mg/mL) in tris-buffered saline (TBS) (25 mM Tris HCl pH 8, 25 mM NaCl, 0.5 mM EDTA) was added TCEP (10 eq.). The mixture was vortexed and incubated at 37 °C for 1 h with shaking at 400 rpm. A solution of DVP-linker-payload (20 mM in DMSO) was added with additional DMSO (final concentration of 0.61 mM, for 40 eq of linker or 1.22 mM for 80 eq of linker, 10% DMSO (v/v)) and the reaction mixture incubated at 37 °C for 4 h with shaking at 400 rpm. The excess reagents were removed by size-exclusion chromatography with a Zeba Spin desalting column (40K MWCO, 0.5 mL) and exchanged into PBS with an Amicon-Ultra centrifugal filter (10K MWCO, Merck Millipore).

SDS-PAGE
Non-reducing Tris-Glycine SDS-PAGE with12% acrylamide with 4% stacking gel was performed as standard. Broad range molecular weight marker (10-200 kDa, New England BioLabs) was run in all gels. Samples (5 µL, 2.5 µg unless stated otherwise) were prepared with reducing loading dye (5 μL, containing β-mercaptoethanol) and heated to 90 °C for 5 min before loading. Gels were run at constant voltage (200 V) for 45-60 min in 1 x Laemmli running buffer (LRB). All gels were stained with Coomassie blue dye and imaged on a Syngene gel imaging system. Figure S1: Reducing SDS-page gel for the bioconjugation reactions a) to form ADC 1 (from 1f), ADC 2 (from 2f) and b) to form ADCs 4-6. The molecular weight ladder is shown to the left of the lanes ran on the same gel and images have been cropped to remove irrelevant lanes. c) representations of the different antibody species present on the gel and structure of each linker. SDS-PAGE analysis demonstrates that the major product of bioconjugation is the half-antibody species, which has a molecular weight between that of the full trastuzumab/durvalumab antibody and the heavy chain. Figure S1-a: Raw, uncropped, un-processed SDS-Page Gel bioconjugation reactions a) to form ADC 1 (from 1f), ADC 2 (from 2f). Lanes not labelled in this raw image were not of relevance to this publication.

Size-Exclusion Chromatography
Analytical size-exclusion chromatography (SEC) was carried out on an AKTA pure chromatography system using a Superdex™ 200 Increase 10/300 GL column. Samples were injected at a concentration of 1 mg/mL and eluted with TBS pH 8 (25 mM Tris HCl, 200 mM NaCl, 0.5 mM EDTA) at a flow rate of 0.5 mL/min. Figure S2: Size-exclusion chromatography of ADCs 1, 2 & 3, highlighting the low levels of ADC aggregation (< 1%). No aggregation was observed for ADCs 4-6.

Hydrophobic Interaction Chromatography
Analytical hydrophobic interaction chromatography (HIC) was carried out on a Tosoh Bioscience TSKgel Butyl-NPR column (3.5 cm  4.6 mm, 2.5 m). Samples were injected at a concentration of 1 mg/mL and eluted with a linear gradient of solvent A in solvent B (solvent A: 1.5 M ammonium sulfate, 25 mM NaPi, pH 7 and solvent B: 25% isopropyl alcohol in 25 mM NaPi, pH 7 at a flow rate of 0.6 mL/min. The drug-to-antibody ratio was calculated by the integration of the signals at 280 nm: = ( 1 + 2 × 2 + 3 × 3 + 4 × 4 + 5 × 5 ) ( 0 + 1 + 2 + 3 + 4 + 5 ) Figure S3: Hydrophobic interaction chromatography of the product of bioconjugation reactions of trastuzumab to form ADCs 1-3. Graphs are annotated with the elution volume (mL) and the grey lines represent integrations of the curve used to calculate average drug-to-antibody ratio. Figure S4: Hydrophobic interaction chromatography of the product of bioconjugation reactions of durvalumab to form ADCs 4-6. Graphs are annotated with the elution volume (mL) and the grey lines represent integrations of the curve used to calculate average drug-to-antibody ratio.

Protein LC-MS
Protein LC-MS was performed on a Xevo G2-S TOF mass spectrometer coupled to an Acquity UPLC system using an Acquity UPLC BEH300 C4 column (1.7 μm, 2.1 × 50 mm). H 2 O with 0.1% formic acid (solvent A) and 95% MeCN and 5% H 2 O with 0.1% formic acid (solvent B) were used as the mobile phase at a flow rate of 0.2 mL/min. The gradient was programmed as follows: 95% A for 0.93 min, then a gradient to 100% B over 4.28 min, then 100% B for 1.04 minutes, then a gradient to 95% A over 1.04 min. The electrospray source was operated with a capillary voltage of 2.0 kV and a cone voltage of 190 V, 150V or 40 V. Nitrogen was used as the desolvation gas at a total flow rate of 850 L/h. Total mass spectra were reconstructed from the ion series using the MaxEnt 1 algorithm preinstalled on MassLynx 4.2 software according to the manufacturer's instructions. Trastuzumab samples were deglycosylated with PNGase F (New England Biolabs) prior to LC-MS analysis. Only the region of the total ion chromatogram (TIC) between 3.25-3.75 min was analysed. Peaks outside of this range did not contain proteinogenic signals and were excluded. Analysis was conducted in the same way for all protein LC-MS traces.

Peroxide-Cleavage Studies
Data from all cleavage studies were processed using GraphPad Prism Version 9.3.0. Human and mouse plasma were obtained from Sigma Aldrich.
Release from model linkers 1c, 2e and 3b with 0, 1, 5, 10 equivalents of hydrogen peroxide Model linkers 1c, 2e, 3b (7.2 μL, 2.5 mM DMSO) were vortexed with PBS (280.8 μL) before adding 72.0 μL of hydrogen peroxide solution (1, 5, 10 equivalents -final concentration of 50 μM, 250 μM and 500 μM respectively). 340 μL of the resulting solution was added to a 96-well plate (Greiner, black, clear flat bottomed). Fluorescence intensity was monitored at λ abs = 441 nm over 20 h at 37 °C. An adhesive film (BioRad) was used to prevent solvent evaporation. Readings were taken at 30-35 second intervals. The reactions were performed in triplicate, with three technical replicates for each set of conditions. Control wells contained no hydrogen peroxide but 72.0 μL of water instead.
Release in the presence of hydrogen peroxide scavenger CAT Model linkers 1c (7.2 μL, 2.5 mM DMSO) was vortexed with either PBS (280.8 μL) containing ~2000 U/mL catalase (CAT) from bovine liver. 72.0 μL hydrogen peroxide solution (10 equivalents) (or 72.0 μL water for control wells) was then added to the solutions and 340 μL of the resulting solution was added to a 96-well plate (Greiner, black, clear flat bottomed). Fluorescence intensity was monitored at λ abs = 441 nm over 25 h at 37 °C. An adhesive film (BioRad) was used to prevent solvent evaporation.
Readings were taken at 1.5-minute intervals. The reactions were performed in triplicate with three technical replicates for each set of conditions. Catalase was obtained from MERCK LIFE SCIENCE UK LTD.
Long term release in the presence of CAT scavenger Once complete, the plate containing scavenger CAT was covered in foil and placed in an incubator at 37 °C for 7 days. Additional fluorimetry measurements were taken at t = 24, 48, 72, 120 and 240 h. This was performed in duplicate, with three technical replicates of each condition. Figure S13: The analysis of fluorescence release from 1c in the presence of 10 equivalents hydrogen peroxide and ROS scavenger Catalase (~2000 U/mL) over 10 days.

Stability Studies
Stability of 1c over 20 hours in human and mouse plasma by fluorimetry Model linker 1c (36.0 μL, 2.5 mM DMSO) were vortexed with PBS (774 μL) and 153 μL of the resulting solution was added to a 96-well plate (Greiner, black, clear flat bottomed). Then, 187 μL human plasma was added to each well. Fluorescence intensity was monitored at λ abs = 441 nm over 20 h at 37 °C. An adhesive film (BioRad) was used to prevent solvent evaporation. Readings were taken at 30-35 second intervals. The reactions were performed in triplicate, with three technical replicates for each set of conditions.  Figure S14: The fluorescence intensity of 1c when incubated with human or mouse plasma. Also shown is 1c with 10 equivalents of hydrogen peroxide for comparison.
Stability of 1c vs Val-Cit 2e in human and mouse plasma over 10 days (HPLC) 1c and 4e (10 mM) were incubated with PBS (149 μL), human/mouse plasma (265 μL), DMSO (86 μL) and caffeine (10 μL, 15 mg/mL, internal standard) at 37 °C for 10 days. 50 μL aliquots of the mixture were taken at t = 1, 2, 3, 4, 5, 10 days and flash frozen in liquid nitrogen and stored at -20 °C before HPLC analysis of all samples on the same day. The linker peak area was calibrated to the caffeine internal standard for quantification of degradation over time. One representative dataset is shown of two independent replicates. The decrease in fluorescence of 2e over time is thought to be due to photo-bleaching of the released fluorophore. Model linker 1c (7.2 μL, 2.5 mM DMSO) was vortexed with unconditioned cell media (280.8 μL) and then 272 μL was added to a 96-well plate (Greiner, black, clear flat bottomed). Water or hydrogen peroxide solution (10 equiv.) (68 μL) was then added to the appropriate wells and absorbance was monitored at λ abs = 441 nm over 24 h at 37 °C. An adhesive film (BioRad) was used to prevent solvent evaporation. The reactions were performed in triplicate, with three technical replicates for each set of conditions.
Long term release in the presence of unconditioned cell media Once complete, the plate containing 1c in unconditioned cell media was covered in foil and placed in an incubator at 37 °C for the remaining 3 days. Additional fluorimetry measurements were taken at t = 48, 72 and 96 h. This was performed in triplicate, with three technical replicates of each condition.

In Vitro Cytotoxicity
General Cell Viability Protocol Cells were seeded in 96-well plates for 24 h at 37 °C with 5% CO 2 . SKBR3 cells were seeded at 15,000 cells/well, BT474 cells were seeded at 20,000 cells/well, MCF7 cells were seeded at 7,500 cells/well and MDA-MB-231 cells were seeded at 5000 cells/well. Serial dilutions of ADCs or trastuzumab/durvalumab were added to the cells in complete growth medium and incubated at 37 °C with 5% CO 2 for 96 h. Cell viability was determined using a CellTiter-Glo viability assay (Promega) according to the manufacturer's instructions. Cell viability was plotted as a percentage of that of untreated cells. Each measurement was taken in triplicate. Three independent replicates were performed. Data was processed using GraphPad Prism Version 9.3.0 and best-fit IC 50 values of each compound were calculated using the log (inhibitor) vs response (variable slope) function. IC 50 s are labelled "N/A" when it was higher than the range of concentrations examined. Cells were seeded in 96-well plates for 24 h at 37 °C with 5% CO 2 . MCF7 cells were seeded at 7,500 cells/well and MDA-MB-231 cells were seeded at 5000 cells/well. Cell media was then removed and replaced with 100 µL complete growth media containing catalase (400-1000 U/well) and incubated at 37 °C for 2 h. Then serial dilutions of ADCs in 100 µL complete growth media were added and cell growth was monitored on an IncuCyte® S3 Live-Cell Analysis System for 96 h with incubation at 37 °C and 5% CO 2 . Cell viability was determined using a CellTiter-Glo viability assay (Promega) according to the manufacturer's instructions. Cell viability was plotted as a percentage of that of untreated cells. Each measurement was taken in triplicate. Two independent replicates were performed. Data was processed using GraphPad Prism Version 9.3.0 and best-fit IC 50 values of each compound were calculated using the log (inhibitor) vs response (variable slope) function. Confluency increase was calculated using the equation: whereby confluency initial is [ -] * 100 the % cell confluency at t = 3h after ADC treatment, and confluency final is the % cell confluency at approx. t = 96 h. Amplex™ Red Hydrogen Peroxide/Peroxidase Assay Kit (Invitrogen) was used following the manufacturers instructions. Briefly, 20,000 cells/well were seeded in 100 µL F-12 Ham's nutrient mixture (for the no cell control, wells contained only F-12 Ham's nutrient mixture). After seeding for 24 h at 37 °C with 5% CO 2 , 100 µL of the Amplex™ Red reaction mixture was added to each well and the fluorescence intensity monitored on a CLARIOstar plate reader over 30 minutes. Data shown is the reading at 30 minutes.
For catalase scavenging, after seeding the media was removed and replaced with Ham's nutrient mixture containing catalase (approx. 400-1000 U/well) (100 µL) and incubated at 37 °C, 5% CO 2 for 2h. Then, 100 µL of the Amplex™ Red reaction mixture was added to each well and the fluorescence intensity monitored on a CLARIOstar plate reader over 30 minutes. Data shown is the reading at 30 minutes.  Figure S17: The fluorescence intensity from Amplex™ red reagent reaction with extracellular hydrogen peroxide of antigen-negative cell lines without scavenging of hydrogen peroxide (left) and with catalase scavenging (right). The fluorescence of Amplex red™ was compared to the fluorescence in the absence of cancer cells. Data was analysed by an ordinary one-way ANOVA whereby * = p < 0.05, ** = p < 0.01, *** = p < 0.001, **** = p < 0.001. ns = non-significant, p >0.05.