(M1330-09-56) Lipid Nanoparticle Impurity Monitoring Using Single Quadrupole Mass Detection for Regulated Environments

Purpose: Lipid nanoparticles (LNPs) represent a novel solution to deliver gene-based therapeutics that is challenging to execute. LNPs are a multi-component biologic comprised of lipid species with varying properties impacting chromatography and detector response. New approaches are needed for characterization and monitoring to reduce the associated risks and ensure active pharmaceutical ingredients are delivered in a safe and efficacious manner. We present an LC-based optical/MS dual-detector platform with improved sensitivity/diagnostic power for LNP workflows. A single quadrupole mass detector provides complementary mass data for raw material impurity screening, MS spectral library-based lipid confirmation, and stability monitoring. Supported by compliant-ready informatics, this workflow is easy to deploy in both non-regulated and regulated environments alike for efficient method development and migration.

Methods: Five lipid nanoparticle components were analyzed, including cholesterol (CHO), distearoylphosphatidylcholine (DSPC), PEGylated lipid DMG-PEG 2000, and ionizable lipids SM-102 and Dlin-MC3-DMA. All lipids were used exclusively for research and demonstration purposes. Lipid stocks were prepared at 1 mg/mL in methanol and diluted to the appropriate concentration using 90/10 water/methanol (v/v) diluent. Lipid samples were separated with ultra high performance liquid chromatography using a phenyl-hexyl column (1.7 µm, 2.1 mm × 50 mm) at 50°C with a 12-min gradient. MS acquisition was performed on a single quadrupole mass detector. Optical data were acquired with an evaporative light scattering detector (ELSD) and a UV detector. Data acquisition and analysis were performed in an integrated compliant-ready software platform.

Results: Using a split-flow configuration, optical (ELSD) and MS data were acquired for individual LNP components and a mixture containing a spike-in to evaluate dynamic range and sensitivity of the MS-enabled workflow. MS results indicated a four order of magnitude-increase in overall detector sensitivity (LOD) could be obtained for the majority of LNP components, enabling detection of low abundance species that were otherwise undetected in optical workflows. This dynamic range and sensitivity of the MS-enabled workflow were evaluated in a practical setting for raw material screening of the formulated drug substance. Raw material screening studies indicated the MS-enabled workflow could detect co-eluting impurities that accounted for up to a total 5% of impurities (peak area %) using extracted ion chromatograms. Furthermore, MS-enabled workflows for LNP mixtures were effective in tentatively assigning stability-indicating degradants and enabling detection of spike-in impurities at levels of 0.1%, which were 10-fold lower than optical-only workflows. The benefit of MS-enabled workflows was realized through integrated spectral library functionality within compliant-ready informatics. In this dual detector workflow, spectral information is automatically linked to optical data allowing users access to spectral data for orthogonal confirmation of lipids and identification of impurities. For example, a spectral library is created for a lipid formula. The spectral matching is performed in the processing method by cross-referencing spectral information associated with the optical results against the spectral library. The best “spectral match” for each peak is tabulated. Out of specification results that exceed defined impurity thresholds (e.g. NMT 0.1 %) or detected mass range (e.g. ±1 m/z) are automatically flagged. This workflow allows users to quickly assess impurities from LNP samples using integrated impurity processing, confirm peak identities, and identify new impurities based on spectra stored in spectral libraries, all via a compliant-ready software to ease the method transfer to regulated environments.

Conclusion: MS increases limit of detection and offers orthogonal structure confirmation of lipid and impurities for LNP process development and manufacturing.

Acknowledgements:Ann Gray is an employee of Waters Corporation and and is presenting the poster on behalf of the company. Duanduan Han, Kellen DeLaney, Robert E. Birdsall, and Ying Qing Yu are employees of Waters Corporation.

Figure 1. Four lipid components of LNP. Cholesterol, DSPC and DMG-PEG2000 provide structural stability and are typically outsourced. The ionizable lipids are often proprietary and produced in-house to protect intellectual property.

Figure 2. Detector type and response. Detector response is shown across 3 detector types for a panel of lipids representative of components used in the production of lipid nanoparticles. Peaks 1-5 are SM102, DOTMA, cholesterol, DMG-PEG2000, and DSPC. SM102, cholesterol, DMG-PEG200, and DSPC were prepared in a molar ratio of 50:38.5:1.5:10 representative of a formulated LNP sample. DOTMA was spiked-in at half the concentration relative to SM102.