A presentation by Jingshu Xu, PhD, Senior Scientist in Mass Spectrometry, AstraZeneca
This presentation was given by Jingshu Xu, PhD, Senior Scientist in Mass Spectrometry, AstraZeneca at the virtual TIDES Europe 2020 and summarized by David Orchard-Webb.
Dr. Jingshu Xu talked through AstraZeneca’s approach to MS data processing for oligos.
To support the new modalities portfolio and handle more complex molecules, AstraZeneca has a range of MS capabilities. For oligo analysis, a number of mass spectrometers offer various resolutions.
They offer different modes of ionization, dissociation, and different types of mass analyzers. The Waters Q-TOF comes with a mobility, for example, giving 3D resolution. The Thermo Fusion Orbitrap is a trap read that comes with a choice of three dissociation methods; CID/HCD/ETD.
The new modalities are also supported by the Bruker UltrafleXtreme, a MALDI-TOF/TOF instrument, and ScimaX which is an ESI MRMS instrument that comes with extremely high resolution. Both ESI and MALDI source for versatility offering further flexibility.
Most modern pharmaceutical oligonucleotides are resistant to enzymatic digestion precluding the typical DNA sequencing methods. The principle of sequence confirmation by MALDI-TOF or LC-HRMS/MS is quite straightforward.
The intact molecule is ionized and fragmented using available dissociation methods and the fragments ions are measured by the mass analyzer. The sequence can be deduced by working out the differences between the fragment ions.
For an efficient sequence confirmation workflow, AstraZeneca wanted a simplified MS method that provides good sequence coverage through comprehensive fragmentation but also a more streamlined way to deal with the complex output data.
As opposed to the LC-HRMS method, fragmentation in source was coupled with single stage MS to take advantage of the ultra-high resolution of the Orbitrap. To deal with the complex data analysis, an in-house automated data process and interpretation workflow was developed.
Automated steps include the calculation of expected fragment ions based on the McLuckey fragmentation scheme. It also models the isotopic patterns to allow the choice of either monoisotopic or the most abundant mass.
The hit list from the MS is searched against the calculated theoretical fragment ions in an exhaustive manner for matches within a given parameter.
The specialist is required to focus on the methods with no match in the automated search. A sequence coverage plot has been created to visualize the output in a simple yet informative manner.
Automated data processing has allowed fast data interpretation through exhaustive search and systematically applied filtering, which in turn gives consistent spectrum annotation and also consistently informative reports for a large amount of data in a very short time.
Furthermore, the visualization reveals sequence coverage, and fragment ion distribution from each experiment at a glance.
To demonstrate sensitivity, switch sequences of Danvartirsen were made by swapping two bases around in position. All switched sequences were analyzed using the simplified LC-MS method.
From the MS, differences were identified, especially in the center region of the spectrum.
The hit list from each mass spectrum was searched against the calculated theoretical fragment ions of the correct Danvartirsen sequence.
Essentially, this is an alignment of how much the spectrum matches that of the correct sequence. In this way, it becomes obvious which fragment ions do not match those that are expected of the correct sequence, which gives a good indication of the position of the base mismatch.
This approach was used to quantify the level of deamination impurity in Danvartirsen. In Danvartirsen, there are three sites with a high probability of deamination. Deamination is effectively a loss of ammonia in exchange for water resulting in a mass shift of one dalton.
By fragmenting the molecule, small fragment ions (around 700-1500 dalton) that contain those deamination sites individually can be monitored and mass-charge shifts detected.
Xu would always start with more than one fragment ion to monitor for each deamination site and also use site specific deamination impurity standards to aid the quantification itself.
In a recent study evaluating the risk associated with terminal sterilization, the described method was used to quantify lower levels of deamination.
From the result, an intriguing observation was made: the terminal methylC residues showed lower levels of deamination than the one in the middle of the sequence.
Increasing the throughput of data analysis is a critical part of achieving an efficient analytical platform for oligonucleotides and AstraZeneca is currently in the process of further developing automated data processing workflows for oligos. Quantification of degradation impurity is another area of focus.
EMA guidelines require a risk assessment for terminal sterilization, even though deamination is a low risk for aseptic processes. AstraZeneca was able to quantify low levels of deamination impurities using the newly developed method.
Having a sensitive and site-specific quantification method also provides an exciting opportunity to investigate the site-specific reaction kinetics and formulation parameters in the future. Finally, in-source fragmentation-based methods are also attractive for chemical ID tests.
Read our full overview of TIDES Europe 2020 here.
