Analytical Quality by Design Method Development and Optimization of an Impurity Profiling Method for GLP-1 Receptor Agonist Exenatide

Significance of the topic

Control and characterization of process- and product-related impurities are essential for the safety, efficacy and regulatory approval of peptide therapeutics. Developing robust, high-resolution impurity-profiling methods for peptides such as exenatide is challenging due to closely related degradants and variable charge states. Applying an Analytical Quality by Design (AQbD) approach aligned with ICH Q14 provides a structured, risk-based route to define method performance requirements, understand critical method parameters, and deliver a method operable design region (MODR) suitable for routine quality control.

Objectives and study overview

The study aimed to develop and optimize a reversed-phase liquid chromatography (RPLC) impurity-profiling method for exenatide generated by thermal stress (50 °C, 7 days) using an AQbD workflow. Goals included systematic identification of critical method parameters (CMPs), screening and optimization via design of experiments (DoE), establishment of a MODR that meets predefined system suitability criteria, and transfer/scaling from UPLC screening conditions to an HPLC configuration appropriate for routine use.

Methodology and experimental design

Key elements of the workflow:

• Four-phase AQbD process: (1) define analytical target profile (ATP) and acceptance criteria, (2) risk assessment to identify CMPs, (3) DoE-driven screening and optimization, (4) establish and verify MODR.
• Screening DoE: evaluated column chemistry (CSH vs BEH), organic modifier (acetonitrile vs methanol), acidic modifier (0.1% TFA vs 0.1% formic acid), and gradient time (25–50 min) to maximize detectable impurity peaks.
• Optimization DoE: focused multivariate study of flow rate (0.650–0.900 mL/min), gradient time (50–100 min), final % organic (40–55%), and column temperature (50–80 °C) to maximize baseline resolution while minimizing run time; followed by a final DoE exploring temperature (65 °C vs 80 °C), gradient time (20–60 min), and initial % organic (20–28%).
• System suitability metrics: tangent-based USP resolution and peak-to-valley (peak height / valley height) ratio were used in combination to better discriminate partially coeluting peptide impurities.
• Data management and modeling: Fusion QbD Software used for DoE design, Pareto analysis, resolution maps, and MODR visualization; Empower CDS used for chromatographic data processing.

Used instrumentation

The study employed complementary UPLC and HPLC platforms and mass detection to accelerate and validate method development:

• Screening platform: ACQUITY Premier UPLC System with TUV (220 nm) and ACQUITY QDa II Mass Detector (400–1500 m/z) for impurity tracking and charge-state information.
• Routine platform (scaled method): Alliance iS Bio HPLC System with TUV detection (220 nm).
• Columns used: ACQUITY Premier Peptide CSH C18 and BEH C18 (1.7 µm, 2.1 × 100 mm) for UPLC screening; scaled to XSelect Premier Peptide CSH C18 (2.5 µm, 4.6 × 150 mm) for HPLC.
• Software: Empower 3.8.1 CDS and Fusion QbD (S‑Matrix) for experimental design and modeling.

Main results and discussion

Key findings from the AQbD-guided development:

• Screening: Acetonitrile-based mobile phases with the ACQUITY Premier Peptide CSH C18 column produced more detectable impurity peaks than methanol-based alternatives; TFA improved the number of detected peaks in many cases.
• Mass detection advantage: The ACQUITY QDa II detector enabled consistent peak tracking across DoE runs by monitoring m/z and charge-state distributions, which varied with acidic modifiers (e.g., different +3/+4 charge state intensities for certain degradants when using FA vs TFA).
• Optimization: Combined use of USP tangent resolution and peak-to-valley metrics improved discrimination between partially resolved peaks and guided selection of operating regions balancing resolution and run time.
• MODR selection and verification: A MODR centered at 65 °C (with specific gradient and % organic ranges) provided consistent separation for the majority of impurities. Experimental verification at five MODR points showed comparable performance: 20 impurity peaks with ≥0.05% area were observed, with only four peaks showing USP resolution <1.5 (two process-related and two early-eluting oxidized species that remained unresolved under evaluated conditions).
• Scaling: UPLC screening conditions were successfully scaled to an HPLC configuration following USP <621> guidance by preserving stationary phase chemistry and column L/particle-size ratio; resulting HPLC conditions delivered similar selectivity suitable for routine QC.

Benefits and practical applications

Practical advantages demonstrated by the study:

• AQbD framework aligned with ICH Q14 provided a science-based, regulatory-friendly approach to method development with quantified design space and verified MODR.
• Complementary mass detection simplified peak identification and tracking across many DoE runs, reducing risk of misassignment during method optimization.
• Efficient workflow: fast UPLC screening reduced development time and solvent use, while scaling to HPLC enabled long-term implementation in regulated QC labs with appropriate error-reduction features.
• Combination of resolution metrics (USP and peak-to-valley) improved method understanding for partially coeluting peptides and guided robust method selection.

Future trends and applications

Potential developments and broader uses informed by this work:

• Wider adoption of AQbD and integrated DoE/software workflows across peptide and biologics method development to accelerate and de-risk impurity profiling.
• Increased use of compact, easy-to-use mass detectors during development to support peak tracking without requiring full high-resolution MS systems.
• Further refinement of performance metrics and automated analytics (e.g., machine-learning assisted peak deconvolution) to handle complex coelutions typical of peptide therapeutics.
• Application of the demonstrated scaling approach to transfer methods between modern UPLC screening and legacy HPLC systems in regulated environments.

Conclusion

The study shows that an AQbD-driven, DoE-supported method development strategy—using complementary UV and mass detection and multivariate modeling—can efficiently deliver a robust impurity-profiling RPLC method for exenatide. The approach enabled identification and control of CMPs, detection of twenty impurity peaks, establishment of a verified MODR (preferentially at 65 °C), and successful scaling from UPLC screening to an HPLC method suitable for routine QC. Combining orthogonal system suitability metrics improved discrimination for partially resolved peaks and strengthened confidence in the final method.

Reference

1. International Council for Harmonisation. ICH Q14 Guideline on Analytical Procedure Development. November 2023.
2. Han D, Ippoliti S, Birdsall RE, Nyholm K. Application of LC-UV/MS Workflows to Increase Efficiency in Impurity Profiling of GLP-1 Analogs. Waters Application Note. May 2025.
3. Han D, Ippoliti S, Birdsall RE, Nyholm K. Accelerating Method Development and Manufacturing of GLP-1 Analogs with LC-UV/MS. Waters Application Note. May 2025.
4. United States Pharmacopeia. Chromatography. USP General Chapter <621>. 2022.