Automated workflow for intact antibody drug conjugates (ADCs) and DAR analysis in Chromeleon CDS

Significance of the topic

Antibody–drug conjugates (ADCs) combine monoclonal antibody specificity with cytotoxic payloads and are a rapidly expanding class of targeted therapeutics. Accurate, high-throughput characterization of critical quality attributes such as drug load distribution (DLD) and average drug-to-antibody ratio (DAR) is essential across development and manufacturing to ensure efficacy, stability, and safety. Native intact mass analysis provides a fast, direct measure of ADC heterogeneity and drug occupancy without peptide-level digestion, improving confidence in attribute assignment and supporting process development and quality control needs.

Goals and study overview

This application note describes the development and demonstration of a fully integrated, automated workflow in Thermo Scientific Chromeleon Chromatography Data System (CDS) for intact ADC analysis and automated DAR calculation. The objectives were to combine instrument control, high-resolution native LC‑MS acquisition, automated deconvolution/identification, and a dedicated Chromeleon report template to enable high-throughput sample-to-report processing with minimal manual intervention.

Methodology

Commercially available ADCs (Enhertu, Polivy, Adcetris, Aidixi) representing lysine- and cysteine-linked conjugates were prepared at 1.0 mg/mL in ultrapure water and analyzed under native SEC-LC conditions. Key operational features:

• Short native SEC method: NativePac OBE-1 column, isocratic 100 mM ammonium acetate, 100 µL/min, 3 min LC run (4.5 min injection-to-injection).
• Injection: 2.0–3.0 µL (2–3 µg protein).
• Orbitrap Exploris 240 mass spectrometer with BioPharma Option operating in native intact protein mode (m/z 2,000–8,000, Orbitrap resolution 60,000, output mass range ~145–160 kDa).
• Data acquisition controlled by Chromeleon CDS 7.3.2 MUe (workable in MUb and later).
• Automated intact deconvolution using the ReSpect algorithm (Sliding Windows approach) implemented as a Chromeleon processing step immediately after acquisition.

The automated workflow uses a Chromeleon sequence table augmented with five custom columns that define AntibodyMass, DrugName, DrugMass, MaxNumberDrugs, and TargetNumberDrugs so expected component masses (antibody + glycoforms + discrete drug additions) are generated and matched against deconvoluted spectra. Identifications are made by matching observed masses to the predefined modification mass ranges within a mass error tolerance; an INDEX/MATCH spreadsheet-style logic underpins assignment. DAR is computed as an intensity-weighted average: sum(drugs × intensity)/sum(intensity) across identified components. A fractional abundance filter (example: ≥0.5%) can be used to exclude minor components from the DAR calculation.

Used instrumentation

Instrumentation reported in the workflow:

• Thermo Scientific Vanquish Flex Binary UHPLC System (with System Base, Binary Pump F, Split Sampler FT, Column Compartment C, VWD-F).
• Thermo Scientific NativePac OBE-1 SEC column (online buffer exchange column for native analysis).
• Thermo Scientific Orbitrap Exploris 240 mass spectrometer with BioPharma Option.
• Chromeleon CDS 7.3.2 with intact protein deconvolution plugin and ReSpect algorithm.
• Consumables: ultrapure water (18.2 MΩ·cm), ammonium acetate (5 M stock), Vanquish Viper MS connection kit.

Data processing details

Key processing elements implemented in Chromeleon CDS:

• Automatic post-acquisition intact deconvolution (Sliding Windows ReSpect template with minor parameter adjustments).
• Custom sequence columns to generate expected masses for antibody + glycan + payload permutations.
• Component identification via mass matching within specified tolerances; identified components are annotated with glycoform and drug load (e.g., A2G0F/A2G1F 4×MMAE).
• Report template creates a per-injection page (UV, m/z spectrum, deconvoluted mass spectrum, drug load distribution plot, component table with measured mass, ID, mass accuracy, abundance, deconvolution score, and Drugs×intensity used for DAR computation).
• Sequence overview page summarizes the most intense component per injection across the batch.

Main results and discussion

The workflow was validated with four ADCs and four blanks across 50 injections (10–15 replicates per ADC). Highlights:

• Throughput: 50 injections—including acquisition, automated deconvolution, identification, and reporting—processed in under four hours using 3-minute LC-MS runs (4.5 min cycle).
• Resolution and specificity: Native-MS conditions improved spectral separation of glycoforms and drug-load variants; overlapping charge states and closely spaced species (example m/z 5,520–5,560) were resolvable to permit confident assignment.
• Mass accuracy: For the most abundant components mass errors were typically <10 ppm after deconvolution.
• Comprehensive DAR readout: Deconvoluted spectra routinely identified species from D0 to D8 (depending on the ADC) and allowed calculation of an intensity-weighted average DAR per injection. Aidixi was shown as an example where D4 species (with specified glycoforms) were among the most abundant.
• Automated filtering and scoring enabled bespoke reporting (e.g., fractional abundance threshold, deconvolution confidence) so only reliable components contribute to final DAR values.

Benefits and practical applications

The integrated Chromeleon CDS solution offers multiple practical advantages for biopharma laboratories:

• End-to-end automation: single environment for instrument control, data processing (deconvolution), identification, and reporting reduces manual steps and potential for error.
• High throughput: short native SEC-LC runs coupled with automated processing supports rapid batch analysis in development and process screening contexts.
• Improved data confidence: native intact MS enhances separation of glyco- and drug-related variants, supporting robust DAR determination without peptide mapping for routine monitoring.
• Customizable reporting: per-injection and sequence summary pages simplify review and archiving; user-adjustable filters and expected-modification lists adapt the workflow to specific ADC constructs.

Future trends and potential applications

Possible directions to extend and apply this workflow include:

• Regulatory QC adoption: further validation and robustness testing to support release assays or lot-to-lot comparability studies in GMP environments.
• Broader payload/linker chemistries: expanding expected-modification libraries to cover novel linkers, site-specific conjugations, and alternative conjugation chemistries.
• Integration with LIMS and automated sample handling to enable fully hands-off, high-throughput screening pipelines.
• Hybrid strategies: combining intact native MS DAR readouts with orthogonal peptide-level or middle-up analyses for complete structural confirmation.
• Data analytics: applying machine learning to improve component annotation, deconvolution confidence scoring, and anomaly detection across large development datasets.

Conclusion

The presented Chromeleon CDS–centered workflow demonstrates a practical, high-confidence approach for automated intact ADC characterization and DAR calculation using native LC‑MS on an Orbitrap Exploris 240. The solution consolidates acquisition, deconvolution, identification, and reporting into a single, customizable environment, enabling high-throughput processing with robust mass accuracy and clear representation of glycoform and drug-load heterogeneity. This approach supports accelerated decision-making in ADC development and can be adapted for broader biopharmaceutical analytics and QC implementation.

Reference

1. Zhou M., Ma Z., Chen J., et al. The next frontier in antibody-drug conjugates: challenges and opportunities in cancer and autoimmune therapy. Cancer Drug Resistance. 2025;8:34.
2. Drugs@FDA: FDA-Approved Drugs database. (Accessed as cited in original application note.)
3. Liu W., Memmi S., Zhang T., et al. Thermo Fisher Scientific Technical Note 001259: Rapid online buffer exchange. (Technical note cited in original application.)