LC-MS Analysis of Therapeutic Oligonucleotide and Related Products - A comparison of TQ and Q-TOF Systems

Background

The purpose of this study is the qualitative analysis and quantitation of a phosphorothioate oligonucleotide. Therapeutic oligonucleotides are synthetic oligonucleotides composed of 10-30 nucleic acid bases. They are aimed directly at targets such as mRNAs, miRNAs or target proteins that are involved in disease pathogenesis. Both sample preparation and method development are simplified using MS detection to analyse oligonucleotides when compared to immune-based detection methodologies such as ELISA. MS detection also provides the added benefit of being the only method capable of detecting slight differences in modifications and resembling impurities based on accurate molecular weight information.

In this study, a molecular weight determination and quantitation of a therapeutic oligonucleotide and related products was performed utilising liquid chromatography-mass spectrometry. The therapeutic oligonucleotides used in the analysis were Mipomersen (Kynamro) and its analogues. From the results of the charge state deconvolution, the molecular weight of the oligonucleotides was confirmed with a mass error of less than 1 ppm.
In the development of a quantitative method using Q-TOF LC-MS or Triple Quadrupole LC-MS, a product ion derived from phosphorothioate was selected as a multi-reaction monitoring (MRM) transition, and a highly-sensitive MRM-quantification was successfully achieved. Since the phosphorothioate modification is one of the most common methods used to increase the stability of oligonucleotides in vivo, it is expected that quantitative methods constructed in the same way as this can be applied to a variety of therapeutic oligonucletotides.

Introduction

Therapeutic oligonucleotides are synthetic oligonucleotides with a chemical backbone structure composed of deoxyribonucleic acid (DNA) or ribonucleic acid (RNA) [1]. They are ‘molecular-targeted drugs’ aiming specific mRNAs, miRNAs and proteins related to the onset of diseases and regulating their expression and function. Although they work as a molecular target drug, these therapeutic oligonucleotides share the favourable attribute of antibody therapeutics and low-molecular drug molecules in that they can be chemically synthesised in the laboratory, making it possible to reduce manufacturing costs compared to an antibody. They are expected to be the next-generation medicines combining the properties of both low-molecular drugs and biopharmaceutics [2].
Therapeutic oligonucleotides are classified into antisense, siRNA, miRNA, aptamer, decoy and CpG oligo, depending on their structure and mechanism of action. Currently available therapeutic oligonucleotides include antisense and siRNAs targeting mRNAs and miRNAs in cells, aptamers that interact directly with proteins outside cells, and CpG oligos that activate innate immunity.
Natural oligonucleotides are unstable to enzymes in vivo such as nucleases, and it is known that their membrane permeability is low due to their molecular weight and negative charge [3]. These oligonucleotides are chemically modified, and phosphorothioated (Sizing) in which an oxygen atom of a phosphate moiety is replaced with a sulphur atom, modification of a sugar moiety to the 2’position such as 2’-O-methoxyethyl (2’ -MOE) and 2’-O-methyl (2’-OMe), and 2’,4’-BNA/LNA in which crosslinking is performed to the 2’-4 ‘position are known.
The molecular weight of these oligonucleotides’ ranges from approximately 1000 to several tens of thousands, larger than that of small molecule drugs and smaller than that of antibody drugs. The molecular weight of antisense drugs is approximately 1000. Ligand binding assays (LBA) have been used to quantify therapeutic oligonucleotides with chromatography and mass spectrometry providing other modes of analysis. In general, LBA has the advantage of high sensitivity and high throughput yet takes time to establish these methods and they experience difficulty in identifying metabolites and impurities [4].
Liquid chromatography-mass spectrometry based methods on the other hand can be developed in a short period of time and can identify molecules with different molecular weights or partial structures. These favourable traits support the use of LC-MS in both the screening environment in the drug discovery stage and metabolite analysis in the development stage.
To establish an LC-MS based analytical method for quality characterisation and kinetic evaluation of therapeutic oligonucleotides using a liquid chromatograph mass spectrometer, molecular weight confirmation is reported by multivalent ion mass spectrometry deconvolution of Mipomersen and its related compounds and quantitative analysis by MRM.

Method

The oligonucleotide to be analysed was Mipomersen, a drug for treatment of familial hypercholesterolemia. Mipomersen is a 20-base antisense oligonucleotide targeted to apolipoprotein B-100 mRNA that inhibits transcription and has a molecular weight of approximately 7000. Three types of analogues of Mipomersen in total were used: an oligonucleotide obtained by substituting all the 2’-O-(2-Methoxyethyl)-modified bases contained in Mipomersen with DNA base; an oligonucleotide obtained by substituting it with a 2’-O-methyl nucleic acid base; and an oligonucleotide obtained by substituting it with a locked nucleic acid (LNA) base. The respective oligonucleotide sequences and modifications are shown in Table 1.

LC/MS conditions

Analysis using a high-performance liquid chromatography-quadrupole time-of-flight mass spectrometer (Q-TOF-LC-MS)
QTOF-LC-MS is a high-performance liquid chromatography (LC) tandem mass spectrometer (LC-MS/MS) combined with a quadrupole time-of-flight mass spectrometer (Q-TOF). Utilising high mass resolution, it is used mainly for quality characterisation and quality control in the analysis of nucleic acid drugs. The body of this work utilised a high-performance liquid chromatograph (Nexera) coupled with a Q-TOF system (LCMS-9030) both by Shimadzu. In a typical reversed-phase separation of oligonucleotides, a mobile phase containing an ion-pair reagent is generally applied. Here, the ion pair reagents hexafluoro-2-propanol (HFIP) and N,N-diisopropylethylamine (DIPEA) were selected, which can measure phosphorothioate oligonucleotides with higher sensitivity, was used. The analytical conditions for both the HPLC and MS components are shown in Table 2.
 
Analysis using a high-performance liquid chromatography-triple quadrupole mass spectrometer (TQ-LC-MS)
TQ-LC-MS is an LC-MS/MS combining LC and a triple quadrupole mass spectrometer (TQ). It is possible to carry out selective and supersensitive analysis by multi-reaction monitoring (MRM), and it can be utilised for analysis of pharmacokinetics etc. in the analysis of therapeutic oligonucleotides. In this research, a high-performance liquid chromatograph (Nexera) and a TQ system (LCMS-8060) by Shimadzu were used as the TQ-LC-MS. As with Q-TOF-LC-MS, the mobile phase with ion pair reagents HFIP and DIPEA was utilised in the LC separation (Table 3).

Data analysis

The mass spectra of Mipomersen and its analogues obtained by Q-TOF-LC-MS were confirmed for their molecular weights using analysis software (Insight Explore CSD, Shimadzu) for mass spectrum deconvolution of multivalent ions. The analysis algorithm applied for the deconvolution was ReSpect (Positive Provability Ltd). Quantitative analysis by TOF-LC-MS and TQ-LC-MS was performed by measuring MRM using a common fragment ion (PSO2-, m/z 94.9358) derived from the phosphorothioate skeleton of Mipomersen as a monitored ion.

Results and discussion

Molecular weight confirmation of Mipomersen and its analogues
Using Q-TOF-LC-MS, MS1 spectra of Mipomersen and the analogues were obtained, and TIC chromatograms were compared. Although each analogue was separated from Mipomersen, peak separation of the analogues was not possible (Figure 1). In ESI-MS analysis, oligonucleotides are generally detected as multivalent ions with a charge state distribution.
Since the spectrum tends to become much more complicated than that of a low-molecular compound, charge-state deconvolution is generally performed to simplify the raw spectrum for analysis, and the molecular weight of a therapeutic oligonucleotide is confirmed using the uncharged spectrum. The mass spectrum of each oligonucleotide was obtained near the peak apex in the TIC chromatogram corresponding to Mipomersen and its analogues, and deconvolution was then executed using the spectrum.
In QTOF-LC-MS, the charge state distributions of Mipomersen and its analogues were from 4 to 8 (Figure 2 - left), and the isotope distribution of Mipomersen analogs were also observed with high mass accuracy of around 1 ppm (Figure 2 - right).
The result of charge state deconvolution using the ReSpect algorithm is shown in Figure 3.
The mass error of the theoretical molecular weight calculated from the composition formula was less than 1 ppm for all oligo nucleic acids (Table 4).
From the above results, if the difference between the target oligonucleotide and the impurities in the therapeutic agent is the structural difference only due to a slight modification, it is assumed that they cannot be distinguished by LC-UV detection because the chromatographic separation is difficult. QTOF-LC-MS is considered to be capable of highly selective detection due to its mass resolution and mass accuracy, even when chromatographic separation is not possible.

High-resolution MRM for quantitative determination of Mipomersen and related compounds
In Mipomersen, phosphorothioate is introduced into all phosphate-ester linkages between nucleotide bases, and a signal amplification effect in MS detection can be expected by multiple generations of a PSO2-fragment through a single MS/MS fragmentation. Using fragment ions derived from the phosphorothioate oligonucleotide backbone of Mipomersen (PSO2 -: m/z 94.9362) as a monitored ion, a high-resolution MRM using Q-TOF was developed for the determination of Mipomersen and the analogues. As the precursor ion, an octavalent ion with the highest valence was selected (Table 5), which was expected to generate PSO2-fragment more frequently.
Table 5: MRM conditions of Mipomersen analogues (QTOF-LC-MS).
Mipomersen -2’-deoxy was detected with the highest sensitivity. A linear calibration curve was obtained in the concentration range of 2 - 1000 ng/mL with a limit of quantification of 2 ng/mL (Figure 4).
Figure 4: Quantitative analysis of Mipomersen-2’-deoxy using QTOF-LC-MS: MRM chromatogram (left) and calibration curve (right).

The calibration curve range and sensitivity of the analogues are shown in Table 6.

Table 6: MRM conditions of Mipomersen analogues (TQ-LC-MS).

Quantitative analysis of Mipomersen and related compounds by TQ-LC-MS
MRM measurement of TQ-LC-MS was performed using fragment ions (PSO2 -: m/z 94.9362) derived from the phosphorothioate oligonucleotide backbone of Mipomersen in the same manner as high-resolution MRM. Detailed optimisation of the MRM conditions was executed using the parameter optimisation software LabSolutions. The optimised parameters are shown in Table 6.
 
Table 7: Results of molecular weight confirmation for Mipomersen analogues.

The limit of quantification for Mipomersen -2’-deoxy was 1 ng/mL, and good linearity was obtained in the calibration curve concentration range of 1 - 1000 ng/mL (Figure 5).

The lower limit of quantifications of all analogues were 2 to 10 times higher than that of QTOF-LC-MS (Table 7). The sample volume used for injection in this measurement was 2 μL, smaller than that of QTOF-LC-MS. These results suggest that MRM measurement by TQ-LC-MS can achieve more sensitive determination.

Summary and conclusion

The results of charge state deconvolution using QTOF-LC-MS suggested that the molecular weight of the impurities generated by elimination or modification of therapeutic oligonucleotide could be confirmed with high mass accuracy of 1 ppm or less. Deconvolution with the ReSpect algorithm could achieve the multiple detection of a target oligonucleotide and the impurities even when they could not be separated by LC and were present in the same spectrum (data not shown). Therefore, more accurate purity determination of therapeutic oligonucleotides is expected even in samples that cannot be detected by conventional chromatographic methods such as LC-UV.
Phosphorothioate modifications have been used in many kinds of therapeutic oligonucleotides to confer resistance to nuclease activity. MRM measurement utilising a PSO2-fragment ion is therefore applicable to a wide range of oligonucleotide therapeutics other than Mipomersen. In this study, the Q-TOF system was shown to have higher resolution and mass accuracy, while the TQ system is more suitable for more sensitive analysis. In addition, it was shown that the optimised parameters could be diverted to each other since a common platform was used in ion introduction, ion optics etc.
Figure 6 shows a comparison between the precursor scan and the product ion scan of the LCMS-9030 and the LCMS-8060. Although there is a slight difference in intensity between the valence ions or between the product ions, the main ion species detected were almost the same for both. This makes it possible to use common analytical conditions in all stages of drug discovery, including quality characterisation of therapeutic oligonucleotide, quality control in drug substance manufacturing, metabolite identification and its concentration measurement, and pharmacokinetic testing. This is expected to contribute to cost reduction in establishing analytical methods.

References

1.Stein C.A., Krieg A.M.
Applied Antisense Oligonucleotide Technology
Wiley, ISBN: 978-0-471-17279-6 (1998)
 
2.  Tozaki T., Karasawa K., Ishii H., Kikuchi M., Kakoi H., Hirota K., Kusano K., Nagata S. Detection of phosphorothioated (PS) oligonucleotides in horse plasma using a product ion (m/z 94.9362) derived from moiety for doping control.
BMC Res Notes, 11, 770 (2018)
 
3.  Hughes J.A., Bennett C.F., Cook P.D., Guinosso C.J.,
Mirabelli C.K., Juliano R.L.
Lipid membrane permeability of 2′ modified derivatives of phosphorothioate oligonucleotides
J Pharm Sci, 84(4), 597-600 (1993)
 
4.  Norris D.A., Post N., Rosie Z Yu R.Z., Greenlee S., Wang Y.
Bioanalysis considerations on the pharmacokinetic evaluation of antisense therapeutics, Bioanalysis
11(21), 1909–1912 (2019)