(M1030-01-02) Identification and Characterization of Impurities in Lipid Nanoparticle Components Using TOF-MS with Predictive In Silico Fragmentation and Automated Data Processing

Purpose: The development of lipid nanoparticles (LNPs) as delivery vehicles has enabled the success of mRNA-based vaccines and other nucleic-acid based therapeutics. These lipid shells are comprised of four components that spontaneously self-assemble around genetic material when mixed at the correct ratios. Before an LNP formulation can be released into the market, impurities must be screened for and appropriately characterized to ensure both safety and efficacy of the released product. Regardless of whether these LNP components are synthesized in-house or procured from third-party vendors, rigorous analysis and screening of impurities is critical to ensuring products are of an acceptable quality and purity. Here, we show how TOF-MS can be used with integrated software to facilitate the identification of impurities in LNPs.

Methods: All lipids in this study were used exclusively for research and demonstration purposes. Five lipid nanoparticle components were analyzed in this study, including cholesterol (CHO), distearoylphosphatidylcholine (DSPC), PEGylated lipid DMG-PEG 2000, and ionizable lipids SM-102 and Dlin-MC3-DMA. Stocks of each lipid were prepared in methanol at 1 mg/mL and diluted to the appropriate concentration at 90/10 water/methanol (v/v). Samples were separated with UPLC using a Phenyl-Hexyl column (1.7 µm, 2.1 mm × 50 mm) at 50°C over a 12-min gradient. MS acquisition was performed on a benchtop TOF mass spectrometer using Data Independent Acquisition (MS^E), using alternate high low collision energy switching. Data acquisition and analysis were performed in an integrated software platform.

Results: In LNP workflows, the addition of MS provides valuable information to facilitate identification of impurities. However, analysis of MS data can be burdensome due to the volume of information generated. The purpose of this study was to use MS informatics tools to streamline analysis of impurities in LNP components. MS^E was used to acquire low and high energy spectra across the chromatogram, enabling acquisition of precursor and fragment ion spectra of all impurity peaks present in the chromatogram. To facilitate the interpretation of fragmentation spectra and aid in the identification of analyte molecules, in-silico fragmentation of known structures was performed within the informatics and the resulting predicted fragment ions were matched to experimentally-obtained fragment ions. With this software feature enabled, detected analytes were rapidly screened at both the precursor and fragment ion levels to provide more confident identification of LNP components and impurities in a mixture. In addition to improving identification confidence, the tool was used to unambiguously localize modifications on the impurity’s molecular structure. As an example, Dlin-MC3-DMA was investigated, as there are two plausible types of oxidation that can be observed for the lipid, including amine oxidation on the polar headgroup and epoxidation of one of the double bonds. Both possible structures were added to the library of structures and used to process data from an oxidized sample of Dlin-MC3-DMA. Reviewing the fragmentation spectra with the in-silico fragmentation tool revealed that the actual location of oxidation is the amine headgroup. The spectra showed three key fragment ions that confirmed the location of the oxygen, which were matched to theoretical fragments of the amine-oxidized structure. These results show how MS^E spectra analyzed with in-silico fragmentation tools can be used for structural elucidation.

Conclusion: Rigorous analysis of impurities is necessary throughout the development and manufacturing processes to ensure the safety and efficacy of released LNP products. Accurate mass data of both precursor ions and fragment ions offers both sensitive detection and confident identification of unknown peaks in an LNP profile. This analysis workflow facilitates both accurate mass screening and interpretation of fragmentation data by taking advantage of instrument and informatics tools.