Advances in Analysis of Therapeutic Oligonucleotides with Chromatography Coupled to Mass Spectrometry

Abstract

Introduction

Therapeutic oligonucleotides (ONs) have emerged as a promising drug modality for a broad spectrum of diseases, particularly those previously deemed undruggable by small molecules or monoclonal antibodies, such as spinal muscular atrophy, polyneuropathy of hereditary transthyretin-mediated amyloidosis, and acute hepatic porphyria. , Since the first FDA approval in 1998, a total of 21 therapeutic oligonucleotide products have received FDA approval. , Additionally, two products, alicaforsen − and olezarsen, are currently in the preregistration stage, and one product, volanesorsen, , has been approved by the EMA. These include 15 antisense oligonucleotides (ASOs), 6 small interfering RNAs (siRNAs), 2 aptamers, and one formulation comprising a mixture of single- and double-stranded polydeoxyribonucleotides (Figure ). While early approvals were largely restricted to rare diseases, the approval of inclisiran in 2020 for the treatment of atherosclerotic cardiovascular disease marked a significant milestone, highlighting the broader therapeutic potential of this modality.

1.

Approved therapeutic oligonucleotides (vaccines excluded). All approvals are by FDA, except ^aVolanesorsen, which is approved by the European Medicines Agency (EMA) only. For ^bAlicaforsen and ^cOlezarsen, the dates refer to the preregistration phase by FDA. AMD: Age-related macular degeneration. FH: Familial hypercholesterolemia. hATTR: Hereditary transthyretin-mediated amyloidosis.

Therapeutic oligonucleotides exert their effects by targeting mRNA (mRNA) directly or indirectly, thereby modulating protein expression at the transcriptional or post-transcriptional level. Their sequence-based design enables high target specificity, while chemical modifications and advanced delivery strategies are employed to improve stability, cellular uptake, and pharmacokinetic properties. Structurally, ONs typically consist of approximately 20 nucleotides per strand and can be single- or double-stranded, depending on their mechanism of action. This diverse class includes antisense single-stranded oligonucleotides (ASOs), small interfering RNAs (siRNAs), microRNAs (miRNAs), and aptamers.

To enhance their drug-like properties, therapeutic ONs are extensively modified at the nucleobase, sugar, and backbone levels. Common backbone modifications include phosphorothioate (PS) linkages − and phosphorodiamidate morpholino (PMO) motifs, while frequently used sugar modifications involve 2′-fluoro, 2′-O-methyl, and 2′-O-methoxyethyl (MOE) substitutions (Figure ). In addition, bioconjugation strategies, such as attachment of N-acetylgalactosamine (GalNAc) or lipids are widely applied to enhance tissue targeting (e.g., hepatocytes) and cellular uptake. While these modifications substantially enhance efficacy and stability, they also introduce significant structural complexity. In particular, phosphorothioate linkages generate chiral centers at each modified phosphorus atom, producing complex mixtures of diastereomers that may influence biological activity and toxicity in vivo. ,

2.

Chemical modifications encountered in antisense and siRNA approved therapeutic oligonucleotides and their impact on pharmacokinetic properties and cellular delivery. Reproduced from ref . Copyright 2024 American Chemical Society.

Within this landscape, siRNAs represent a distinct and analytically challenging subclass of therapeutic ONs. siRNAs are double-stranded RNA molecules, typically ∼ 21 base pairs in length, consisting of a pharmacologically active antisense strand and a complementary sense (passenger) strand. − The hydrophilic nature and relatively large size (∼13 kDa) of siRNAs make it difficult for them to enter cells without the help of specialized carriers. This need has driven the development of delivery systems, most notably lipid nanoparticles (LNPs). − The clinical success of LNP-formulated siRNAs, exemplified by the FDA approval of Onpattro, the first LNP-encapsulated siRNA therapeutic, and the widespread use of LNPs in mRNA COVID-19 vaccines, underscores the critical role of formulation in siRNA therapeutics.

The rapid progress of therapeutic oligonucleotides through clinical pipelines, has intensified the demand for robust, sensitive, and orthogonal analytical methods to support quality control. These methods are essential for identity confirmation, assay determination, analysis of PS diastereomers, impurity profiling, and stability assessment. ,,− Analytical characterization is particularly challenging due to the structural diversity of ONs and the wide range of potential process- and degradation-related impurities. These impurities include shortmers (n-1, n-2, etc.), longmers (n+1, n+2, etc.), phosphodiester (PO) linkages in PS ONs, depurination and deamination products, and residual protecting group adducts, such as N³-(2-cyanoethyl)­thymine (CNET). −

Comprehensive identification and quantification of these impurities are critical to ensuring the quality, safety, and efficacy of ON-based therapeutics. A detailed overview of oligonucleotide impurity formation and mass spectrometric characterization has been provided by Pourshahian (2019), including characteristic mass shifts associated with common impurity classes. For conjugated ONs, such as GalNAc- or lipid-modified species, it is also essential to monitor the distribution of unconjugated material, fully conjugated target product, and free conjugating agents. In the case of double-stranded siRNAs, analytical strategies must address both single-strand purity and duplex integrity, further increasing analytical complexity. ,,

In addition, regulatory expectations (FDA/EMA) place strong emphasis on predefined quality attributes, including identity, purity and impurity profiles, assay, and stability, which require highly selective and orthogonal analytical workflows. − The recently published EMA guideline on oligonucleotide development and manufacture indicates that identification and qualification thresholds of 1.0% and 1.5%, respectively, are generally acceptable. When defining a reporting threshold, the primary consideration is the lower limit of quantification (LOQ) of the analytical method used to measure impurities. Capaldi et al. noted that although the specific value depends on the size and complexity of the oligonucleotide, a reporting threshold of approximately 0.2% should be achievable in most cases. Meeting these requirements is further complicated by the limited availability of qualified reference standards, especially for low-abundance or structurally related impurities, making accurate quantification difficult. These challenges underscore the need for reliable reference materials and for validated methods capable of addressing the specific chemical diversity and degradation pathways of therapeutic ONs.

Mass spectrometry (MS) has become an indispensable tool for the identification and structural characterization of ONs and their impurities. However, efficient MS-based analysis relies heavily on the availability of compatible and selective chromatographic separations.

Ion-pair reversed-phase liquid chromatography (IP-RPLC) has been extensively studied and remains the gold standard for ON analysis, combining high selectivity and separation efficiency with MS compatibility, although mobile phase additives may introduce some constraints. , Anion-exchange chromatography (AEX) is another well-established technique that provides high resolution, although historically its use has been limited by poor MS compatibility due to nonvolatile salts.

The structural complexity of therapeutic oligonucleotides, arising from chemical modifications, conjugation moieties, and diastereomeric mixtures, together with the challenge of resolving closely related impurities, clearly demonstrates that no single chromatographic technique can address all analytical requirements. Comprehensive characterization therefore necessitates orthogonal analytical approaches capable of probing distinct physicochemical properties and resolving species that may coelute under a single separation mode.

Within this framework, the need for MS-compatible, orthogonal, and more sustainable separation strategies has stimulated growing interest in hydrophilic interaction liquid chromatography (HILIC). HILIC represents an attractive alternative particularly for highly polar oligonucleotides and offers complementary selectivity relative to IP-RPLC. , However, its broader implementation is still limited by complex retention mechanisms and, at the current stage of development, by challenges related to robustness and method selectivity. Nevertheless, HILIC can provide complementary selectivity and enable the separation of impurity classes that cannot be simultaneously resolved by IP-RPLC, while employing more environmentally friendly and MS-compatible mobile phases. Finally, multidimensional LC strategies, combining complementary modes such as IP-RPLC or AEX in the first dimension and HILIC in the second, may offer substantial potential to increase peak capacity and improve impurity profiling. However, their broader adoption is currently limited by the technical complexity of implementation and the demanding instrumental setup required.

In this context, the present review summarizes recent advances in chromatographic methods coupled with MS for the characterization of therapeutic oligonucleotides. It focuses on IP-RPLC, HILIC, and multidimensional LC, discussing their strengths, limitations, complementarity, and the analytical challenges posed by chemically modified, conjugated, and structurally complex ONs. A dedicated section addresses MS detection for structural characterization, including charge-state distributions, adduct formation, in-source fragmentation, and the need for optimized acquisition strategies and tailored data-processing workflows. Beyond drug substance characterization, bioanalytical applications are reviewed, highlighting the role of LC–MS in pharmacokinetic and biodistribution studies and in monitoring metabolites and degradation products. Finally, future directions toward more robust, sustainable, MS-compatible, and bioanalysis-ready analytical strategies are discussed to support continued development, regulatory compliance, and industrial implementation.

Ion-Pairing Reversed-Phase Liquid Chromatography (IP-RPLC)

Effective analysis of oligonucleotide therapeutics relies on the development of robust, sensitive, and selective methods that can accurately identify, separate, and quantify impurities. This task is particularly challenging because many oligonucleotide-related impurities are structurally similar to the parent sequence in both size and physicochemical properties, making chromatographic separation difficult not only from the main component but also among individual impurity species. These challenges are further amplified for PS oligonucleotides, as the presence of multiple diastereomers results in peak broadening and reduced chromatographic resolution. As a result, high selectivity is crucial during method development to reliably distinguish the parent oligonucleotide from closely related impurities.

Historically, AEX ,− and capillary gel electrophoresis (CGE) ,− have been widely used for oligonucleotide analysis due to their high resolving power. However, their lack of compatibility with mass spectrometry (MS) limits their applicability for comprehensive structural characterization. To overcome this limitation, IP-RPLC has emerged as the predominant technique, combining high chromatographic resolution with MS compatibility. ,

IP-RPLC exploits the negatively charged oligonucleotide backbone, which forms ion pairs with positively charged alkylamine reagents added to the mobile phase. These ion-paired complexes exhibit increased hydrophobicity and enhanced affinity for the reversed-phase stationary phase, leading to improved retention and separation. Although ion-pair formation can reduce the direct influence of sequence composition, chemical modifications, or conjugated groups on chromatographic behavior, ion-pairing agents remain essential to enhance retention and achieve adequate resolution. Under optimized conditions, IP-RPLC enables effective separation of truncated species, sequence variants, and conjugated oligonucleotides, although certain degradation products, such as depurinated species, may remain difficult to resolve.

Alkylamines used at concentrations of approximately 5–20 mM as ion-pairing (IP) reagents enable compatibility with electrospray ionization mass spectrometry (ESI-MS). In IP-RPLC, the choice of alkylamine primarily governs oligonucleotide retention and chromatographic resolution, whereas the buffering acid largely determines MS sensitivity. The ion-pairing strength of alkylamines is mainly dictated by their hydrophobicity and volatility, which together define overall chromatographic performance. , Among physicochemical descriptors, boiling point has proven to be a reliable descriptor of alkylamine hydrophobicity, allowing classification into weak (e.g., triethylamine, propylamine), moderate (e.g., dibutylamine, dimethylcyclohexylamine), and strong (e.g., tributylamine) ion-pairing systems.

Weak IP reagents typically provide limited retention and selectivity, whereas more hydrophobic amines enhance retention and resolution but may compromise MS sensitivity due to increased ion suppression. , Even alkylamines with comparable ion-pairing strength can differ substantially in performance, as variations in chain length, branching, or cyclicity influence interactions with the stationary phase and the oligonucleotide, as well as reagent volatility and ionization efficiency. , Consequently, careful selection of the IP reagent is essential to balance chromatographic resolution and MS compatibility.

To enhance selectivity, IP-RPLC methods are adopting dual ion-pairing strategies, moving beyond the limitations of single-reagent systems. This concept was introduced early by Levin et al., who demonstrated that combining two or more IP reagents in the mobile phase can significantly improve peak shape and resolution compared to the use of a single IP reagent. Building on this principle, subsequent developments have focused on dynamically modulating ion-pairing strength during the separation.

One effective implementation of dual IP concepts involves ion-pairing gradients, in which a weak IP reagent is progressively replaced by a stronger one as the organic solvent strength increases. This simultaneous modulation of solvent composition and ion-pairing strength broadens the elution window and enhances resolution for complex oligonucleotide mixtures relative to conventional single-IP systems. Systematic studies by Fekete et al. demonstrated that neither weak nor strong IP systems alone are universally optimal: weak systems frequently suffer from coelution, whereas strong systems may widen the elution window without fully resolving closely related variants. In contrast, controlled weak-to-strong IP gradients substantially reduced coelution and enabled baseline separation across diverse oligonucleotide mixtures. However, it should be noted that this approach has not yet been applied to the real impurity profiling of therapeutic oligonucleotides.

A very recent study by the same authors introduced a highly flexible IP-RPLC strategy for oligonucleotide separations, decoupling the cosolvent and IP reagent gradients using a quaternary pump. This approach allows independent optimization of solvent strength and ion-pairing transitions, enabling segmented gradients and flexible weak-to-strong IP modulation. As a result, both selectivity (Figure ) and retention predictability were markedly improved for complex oligonucleotide separations.

3.

Combined segmented and multilinear gradients for both the cosolvent and weak-to-strong IP reagent transitions. (A1, A2, A3) Traditional linear gradient program (20–80% MeOH, 9–0 mM TEA, 1–15 mM DAA). (B1, B2, B3) Two-segmented gradient. (C1, C2, C3) Linear cosolvent gradient combined with concave/convex transitions in the weak-to-strong IP reagent gradient. Adapted with permission from ref . Copyright 2025 Elsevier.

Regarding counterions, although acetate was initially used in IP-RPLC, 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) has now become the standard. HFIP offers several advantages in ESI-MS applications: its high volatility promotes efficient droplet desolvation, reducing ion suppression and increasing sensitivity, often by 2–3 orders of magnitude. It also sharpens chromatographic peaks, improving resolution in IP-RPLC, in part by suppressing separation of PS diastereomers. Additionally, HFIP contributes buffering capacity, helping to stabilize mobile phase pH.

However, HFIP has notable limitations. It promotes higher and multiple charge states in MS, complicating spectra and potentially reducing detection sensitivity. As a perfluoroalkyl substance (PFAS), HFIP raises toxicological and environment concerns due to persistence, bioaccumulation, and toxicity. Additionally, it is relatively expensive, which may limit routine use in high-throughput settings, and can cause persistent contamination in LC-MS systems, requiring extensive cleaning or dedicated instrumentation. Finally, HFIP also exhibits limited solubility in acetonitrile at higher concentrations, sometimes requiring methanol as the organic modifier. One effective strategy to mitigate these drawbacks is to reduce HFIP concentration by operating IP-RPLC at higher mobile-phase pH, enabled by hybrid silica-based C18 phases. Elevated pH lowers dependence on HFIP, reducing PFAS burden, system contamination, and MS spectral complexity while maintaining chromatographic selectivity. Alternatively, a very recent study demonstrated that HFIP can be replaced by an acetate/triethylamine/glycine mobile phase. Glycine, a hydrophilic amino acid, significantly enhances electrospray ionization, providing MS sensitivity comparable to HFIP/TEA. These findings highlight hydrophilic additives as a promising, environmentally safer, and cost-effective approach for HFIP-free IP-RPLC-MS of oligonucleotides. Nevertheless, the use of nonvolatile additives may pose long-term risks for instrument compatibility and limit broader applicability, indicating that further optimization may be needed.

There is no universal combination of alkylamine and buffering acid that guarantees optimal separation for all oligonucleotides. Ion-pairing efficiency strongly depends on oligonucleotide length and composition. Strong, hydrophobic alkylamines often provide better separation for longer oligonucleotides at low concentrations (typically 5–10 mM), but their solubility in water is restricted. In such cases, acetic acid is often used instead of HFIP, which is commonly paired with weaker IP reagents.

A recent study by Obata et al. demonstrated that a tributylamine/heptanoic acid system achieves optimal selectivity for separating PS oligonucleotides from common impurities, including phosphodiester (PO) byproducts and n–1 truncated sequences. Using tertiary or quaternary alkylamines minimized diastereomeric resolution, reducing PS peak broadening. Additionally, the use of hydrophobic alkyl carboxylic acids further enhanced selectivity between PS oligonucleotides and impurities. This combination enabled clear separation of PS oligonucleotides from both PO and n–1 variants, even in the presence of positional isomers (Figure ), with reliable quantitative performance across 0.2–5% impurity levels. However, the use of heptanoic acid can pose challenges for MS detection, as its relatively low volatility compared to shorter-chain acids may increase source contamination and suppress ionization, potentially reducing sensitivity.

4.

Analysis of spiked impurities in a NUS sample using the IP system tributylamine (TBA)/heptanoic acid. NUS was a fully modified 18-mer PS oligonucleotide with 2′-MOE modifications, including 5-methylcytosine and 5-methyluracil. (a) Effect of PO and n-1 position on the separation of 1% spiked impurity. (b) Quantification of spiked impurities (0.2%–5%). L-column3 C18 column (150 × 2.0 mm, 2 μm, stainless steel hardware). Mobile phase A: 10 mM TBA, pH 7.0 (adjusted with heptanoic acid); mobile phase B: 50% ACN in mobile phase A (1:1, v/v); initial ACN content: 88%; gradient slope: 0.2%ACN/min. Adapted with permission from ref . Copyright 2025 Elsevier.

Mobile phase pH is another critical parameter that influences retention, selectivity, and MS sensitivity. It can be adjusted by varying the concentrations of triethylamine (TEA), which increases pH, and HFIP, which lowers it. A pH range of 8–9 often represents an optimal compromise, offering enhanced oligonucleotide ionization, which in turn improves retention, resolution, and selectivity. At lower pH values, nonspecific adsorption and ion suppression are more pronounced due to higher additive concentrations. Conversely, at high pH values, alkylamine ionization decreases, aggregates may form, and mobile phase lifetime is reduced.

C18 columns are generally the most widely used in reversed-phase chromatography and remain the preferred choice for oligonucleotide separations. While porous silica-based particles are standard, alternative materials, such as nonsilica polystyrene-divinylbenzene resins , and superficially porous particles, have also shown effective separation performance. , Various alkyl- and aryl-based stationary phases, including C4, C8, C18, and phenyl ligands immobilized on silica, have been evaluated in IP-RPLC for oligonucleotide analysis. , Overall, as noted by experts, the type of reversed-phase sorbent has only a minor impact on selectivity for routine analytical purposes, since oligonucleotide retention is predominantly governed by ion-pairing interactions, particularly for more hydrophobic amines. , Minor differences between stationary phases may arise under specific conditions, such as the partial, often undesired, resolution of phosphorothioate diastereomers or during preparative-scale separations. Overall, C18 remains dominant over phenyl or other phases due to widespread adoption and robustness, but C4, C8, and phenyl phases generally provide similar selectivity when the type and concentration of ion-pairing agent, gradient, and mobile phase conditions are properly optimized.

Nonspecific adsorption (NSA) onto various components of the LC system, especially the column hardware, is a major limitation in oligonucleotide liquid chromatography. It is especially critical for trace-level LC–MS analyses (<pmol per injection), where adsorption to metal surfaces can cause analyte loss, peak distortion, and poor reproducibility. , NSA is generally less critical in routine LC–UV applications at higher loadings (>100 pmol per injection). To address these issues, conventional stainless-steel columns have been replaced by a range of bioinert materials, such as titanium-lined, PEEK-lined, and hybrid organic/inorganic surfaces. Comparative studies with stainless-steel columns demonstrated that adsorption of oligonucleotides occurs predominantly within the column hardware, particularly on internal surfaces such as column walls and frits, which account for most of the sample flow path and present a substantial metal oxide surface area. Bioinert materials markedly reduce NSA from the very first injection, whereas stainless-steel columns often require multiple injections to saturate active adsorption sites. Furthermore, although stainless-steel column conditioning can temporarily mitigate NSA, this effect is transient and can be easily reversed during routine use (e.g., high-pH flushing), limiting both long-term applicability and method transferability.

Another important parameter in ON separation is the column temperature. Ideally, separations should be performed at elevated temperatures, typically around 60 °C. In some cases, even higher temperatures (80–90 °C) are necessary to disrupt stable secondary structures, particularly in G-rich sequences that can form G-quadruplexes, which otherwise complicate chromatographic analysis. Operating at high temperatures offers several advantages: it linearizes ONs, thereby improving separation efficiency; in the case of PS ONs, it also suppresses diastereomeric separation, resulting in sharper peaks. Additionally, higher temperatures lead to lower backpressure and faster analyses. Column temperature is especially a critical factor in the analysis of siRNA duplexes, as it directly influences the melting behavior of the duplexes within the column. The effect of temperature depends on the specific melting temperature (T _m) of each duplex. , Nondenaturing conditions, achieved at temperatures below T _m, preserve duplex integrity, allowing resolution from any excess single-stranded oligonucleotides in the sample. In contrast, denaturing conditions, applied at temperatures well above T _m, result in fully single-stranded oligonucleotides, enabling assessment of strand purity. LC separations performed at temperatures close to T _m should be avoided, as partial on-column duplex melting may lead to peak broadening, distortion, and reduced resolution.

The IP-RPLC-UV-MS platform plays a central role in the identification, assay, and impurity profiling of both active substances and finished oligonucleotide products, particularly antisense oligonucleotides (ASOs) and siRNAs. Public assessment reports from the European Medicines Agency (EMA) Committee for Medicinal Products for Human Use (CHMP) highlight the analytical procedures employed by manufacturers to ensure product quality prior to regulatory approval. , While UV detection remains the standard for quantitation, the compatibility of IP-RPLCwith mass spectrometry enables accurate mass measurement and detailed MS/MS characterization. This is particularly useful for identifying and quantifying impurities that are not well separated from the main oligonucleotide peak by chromatography alone. As a result, IP-RPLC/ESI-MS has become the standard method for routine analysis of ONs. ,,

The growing clinical importance of PS oligonucleotides has intensified interest in diastereomer separation. PS modifications increase stereochemical complexity, generating thousands of possible isomers in antisense oligonucleotides. Regulatory agencies now emphasize advanced analytical tools to assess diastereomer distribution and batch-to-batch reproducibility, recognizing their potential impact on drug efficacy and safety. ,

Historically, method development focused on minimizing diastereomer influence during impurity profiling. More recent approaches aim for direct separation and characterization of PS diastereomers, with Sp and Rp configurations increasingly recognized as key determinants of oligonucleotide conformation and chromatographic behavior. , A 2023 review summarizes advances in separating PS diastereomers, highlighting how oligonucleotide structure, sugar modifications, and nucleotide sequence can either enhance or hinder resolution. Using a 2′-OMe-modified CUUUU marker, SpRp (or RpSp) diastereomers eluted after SpSp forms, emphasizing the impact of sugar modifications on conformation and retention. Chiral assignments were based on the predominance of Sp configurations generated during solid-phase synthesis with tetrazole activators such as BTT.

Complementary work by Gilar et al. demonstrated that diastereomer separation in IP-RPLC depends on nucleobase composition, with guanosine providing the greatest resolution and uridine the least. Studies have further shown that hydrophilic ion-pairing systems enhance diastereomer separation, whereas stronger hydrophobic ion-pairing agents tend to suppress it. Recent work by Vosáhlová et al. investigated the influence of counterion selection (acetate vs HFIP) and alkylamine properties on the separation of PS diastereoisomers of varying lengths (2–6 mers and 21 mers) with different chemical modifications. The study confirmed that both alkylamine and counterion choice critically affect diastereomeric resolution. Hydrophobic alkylamines reduced diastereomer separation, producing narrower peaks and improved resolution of n/n-x oligonucleotide mixtures (Figure ). Substituting acetate with HFIP further reduced PS diastereomer separation. Column temperature was also shown to play a key role: elevated temperatures suppressed diastereomeric resolution while simultaneously improving the separation of n/n-x species.

5.

Effect of alkylamine type (triethylammonium acetate (TEAA), dipropylammonium acetate (DPAA), and hexylammonium acetate (HAA)) as ion-pairing reagents on the diastereomeric separation of fully phosphorothioated (all-PS) oligonucleotides (2–6 mers). Gradient conditions: TEAA: 0–20 min = >4–29% MP A; DPAA: 0–20 min = >8–33% MP A; HAA: 0–20 min = >25–50% MP A. MP A contains 10 mM alkylammonium acetate in ACN/water, 80/20 (v/v). MP B is fully aqueous with the same IP reagent concentration. Column: Acquity UPLC BEH Phenyl. Temperature: 60 °C. Adapted with permission from ref . Copyright 2024 Elsevier.

Focusing on siRNAs, Enmark et al. reported that increasing the ionic strength of the IP-RPLC mobile phase with phosphate-buffered saline (PBS) enhanced duplex stability and improved diastereomer separation. They also emphasized that oligonucleotide structure strongly influences diastereomer resolution, consistent with earlier observations.

Hydrophilic Interaction Chromatography (HILIC)

Hydrophilic Interaction Liquid Chromatography (HILIC) has emerged as a powerful method for separating polar and charged compounds, including ONs. Since Alpert’s pioneering 1990 study on oligothymidylic acids (12–30 bases), HILIC has become a widely adopted alternative to IP-RPLC. Unlike IP-RPLC, HILIC does not require ion-pairing agents or HFIP additives, simplifying mobile phase composition, reducing cost and toxicity, and enhancing environmental sustainability. This simplicity also makes HILIC well-suited for scale-up to preparative HPLC applications.

Retention in HILIC primarily arises from partitioning between a water-rich layer immobilized on the polar stationary phase and the predominantly organic mobile phase (typically acetonitrile). Water acts as the strongest eluent, so retention is maximized under weakly aqueous conditions, where hydrophilic partitioning dominates. Additional electrostatic and hydrogen-bonding interactions further contribute to retention, particularly for highly charged ONs. −

For oligonucleotides, the mobile phase often contains 30–70% water, which can disrupt the water layer, diminishing hydrophilic partitioning and making electrostatic interactions (often repulsive) and hydrogen bonding the main retention drivers. , Retention is influenced by nucleotide composition, backbone chemistry (phosphorothioate vs phosphodiester), sugar modifications (e.g., 2′-OMe, 2′-F), and mobile phase parameters. In general, ONs elute in order of increasing chain length, with retention following base hydrophilicity (A < T < G < C < U), roughly the reverse trend of IP-RPLC. ,, Sugar- and PS-modified oligonucleotides generally exhibit lower retention, and PS species often produce broader peaks than phosphodiester ONs, likely due to diastereomeric heterogeneity.

HILIC separations of ONs commonly use neutral stationary phases, such as amide ,,,,− and diol groups. ,, Zwitterionic phases have also shown promise for ON characterization. The BEH amide column is widely used, and its bioinert variant, the PREMIER BEH amide column, enhances analyses by minimizing nonspecific adsorption (NSA). Lardeux et al. systematically identified column hardware, specifically walls and frits, as the primary sources of NSA, with adsorption more pronounced under HILIC than IP-RPLC, particularly for long ONs (40–100 bases).

Unlike IP-RPLC, column temperature has minimal impact on HILIC retention. Typical operating ranges are 40–60 °C, ,,, with only slight increases in retention observed at higher temperatures. , For siRNAs, temperature plays a critical role in duplex stability and separation selectivity, as previously highlighted in a recent application note. Using a siRNA standard, the study showed that the duplex remained stable below its T _m, yielding a single sharp peak. As the temperature approached T_m, partial dissociation occurred, producing a broad, early eluting zone, while temperatures above T _m led to complete strand separation and the appearance of two distinct peaks. These results demonstrate that temperature can be strategically used to switch between nondenaturing and denaturing HILIC conditions, directly modulating retention behavior and separation outcomes.

Buffered, ion-pairing-free mobile phases, such as ammonium acetate or formate, are widely used in HILIC for MS compatibility. Increasing salt concentration enhances retention by stabilizing the water layer, which strengthens hydrophilic partitioning, and by reducing electrostatic repulsion between ONs and silanol groups. However, to maintain MS sensitivity, moderate concentrations (5–25 mM) are preferred. Mobile phase pH also influences retention by altering the ionization states of both residual silanols and ON phosphate groups: at low pH (∼3.5) adsorption is favored, while higher pH (5.5–7.5) increases negative charge, enhancing electrostatic repulsion and reducing retention. CAN is the preferred organic modifier, as protic solvents like methanol may compete for hydrogen-bonding sites and disrupt the water layer, if present, though MeOH can provide distinct selectivity and remains underexplored.

ESI-MS is the benchmark for ON identification and impurity detection, producing multiply charged negative ions due to phosphate deprotonation. HILIC’s ion-pairing-free, high-organic mobile phases reduce LC and source contamination and can enhance negative-ion ESI. However, despite early expectations, HILIC generally provides lower sensitivity than IP-RPLC with alkylamine/HFIP systems (typically 5–50 mM), , although it better preserves ON structural integrity, resulting in narrower charge-state distributions compared to IP-RPLC (Figure ). Kilanowska et al. systematically evaluated HILIC stationary phases and mobile phase conditions for various DNA-based antisense oligonucleotides and observed that IP-RPLC consistently produced higher MS signal intensity. The lower sensitivity of HILIC-MS was attributed to increased alkali metal adduction, ion suppression from acetonitrile, and suboptimal desorption during electrospray ionization at lower pH. Similarly, Huang et al. demonstrated that using moderate concentrations of ammonium acetate at higher pH improved HILIC-MS/MS signal for most ONs, with accurate mass determination achievable despite low-abundance adduct peaks. Although promising, comparisons between IP-RPLC and HILIC-MS must be interpreted with care, as they often reveal differing charge state distributions that can affect both sensitivity and data interpretation. Comparative studies by Lobue et al. further indicate that HILIC-MS can, in somes cases, yield steeper response curves and improved limits of quantification compared to IP-RPLC-MS, highlighting its potential as a complementary MS-compatible separation approach.

6.

Representative mass spectra of oligonucleotides acquired under HILIC and IP-RP conditions. (A) GCAAG­CUGAC­CCUGAA­GUUCAU and (B) T*C*G*T*G*­C*T*T*T*T*­G*T*T*G*­T*T*T*T*­C*G*C*G*T*T. * indicates a phosphorothioate linkage. For each pair, spectra are displayed using the same absolute y-axis scale. Adapted with permission from ref . Copyright 2019 Elsevier.

HILIC has also been shown to provide orthogonal selectivity relative to AEX and IP-RPLC, making it a valuable complementary technique for comprehensive oligonucleotide analysis. This complementarity was further demonstrated by Gilar et al., who studied a fully phosphorothioated 25-mer oligonucleotide and its truncated (n–1) and PO impurities. HILIC improved separation of G- and C-containing truncated products and reversed the elution order of PO impurities, which appeared after the full-length product. This behavior contrasts with IP-RPLC, where the hydrophobicity of alkylamine ion-pairing agents often causes PO impurities to coelute with truncated species, complicating analysis. In this study, the 24-mer n–1 impurities eluted before the full-length 25-mer, while PO impurities appeared afterward, illustrating how HILIC enables oligonucleotide resolution based on apparent hydrophilicity and improves chromatographic separation of both truncated and PO impurities (Figure ).

7.

HILIC separation of oligonucleotide impurities. (A) Separation of a 25-mer PS oligonucleotide from n-1 impurities with a loss of each common nucleotide (green); impurities containing 1 PO linkage (red) at different locations within the sequence (1–4)); impurities containing 2 PO linkages (blue) at different locations within the sequence (5–7). (B) Separation of mixtures of standards (blue line: FLP with -G, 4, and 7; black line: FLP with -C, 3, and 6; green line: FLP with -A and 2; orange line: FLP with -T and 1; pink line: FLP with 5). HILIC conditions: ACQUITY Glycan BEH Amide Column (50 × 2.1 mm, 1.7 μm); mobile phase: 10 mM ammonium acetate buffer, pH 6.9, with ACN as organic solvent; column temperature: 60 °C. Adapted with permission from ref . Copyright 2023 Elsevier.

To exploit the orthogonality of chromatographic modes, Goyon and Zhang recently developed a multidimensional LC–HRMS approach for impurity profiling of fully PS oligonucleotides. This 2D-LC method couples either AEX or IP-RP LC in the first dimension with HILIC in the second, which simultaneously serves as an effective desalting step. By optimizing sample loop volumes and column dimensions, the authors achieved high-resolution separation and detected impurities down to 0.3%. IP-RPLC provided superior resolution in the first dimension compared to AEX, while HILIC in the second dimension enabled further separation of additional impurities, including PO and dithioate species. Notably, this 2D-LC setup allowed direct coupling of AEX or IP-RP to HILIC without requiring solvent modulation or specialized instrumentation.

Similarly to IP-RP HPLC, HILIC suffers from peak broadening due to partial separation of PS diastereomers, which can be detrimental to the analysis of phosphorothioate oligonucleotides. Strategies in HILIC should therefore be considered to more effectively suppress diastereomer separation. Vosáhlová et al. studied nusinersen, an 18-mer antisense oligonucleotide with sugar and phosphate backbone modifications, using a bioinert amide column to minimize nonspecific adsorption. To reduce diastereomer separation, they optimized mobile phase pH, salt concentration, and column temperature, identifying a neutral to slightly basic pH and 40 °C as optimal. Despite these adjustments, PS diastereomers still exhibited ∼ 25% broader peaks than phosphodiester analogs, indicating incomplete suppression of diastereomer effects.

Conversely, HILIC can enhance diastereomer resolution under certain conditions. Goyon et al. demonstrated improved separation for oligonucleotides with rigid ribose modifications, including 2′-methoxy, 2′-fluoro, or locked nucleic acids, although separation efficiency depends on oligonucleotide length, the type and position of chemical modifications, and PS linkage location. Lardeux et al. further advanced PS siRNA analysis, identifying higher-order RNA structures as the key determinant of diastereomer separation in HILIC. By combining folding predictions, melting profiles, and mass spectrometry, they established a direct link between siRNA conformation and diastereomer separation. This work not only provides deeper insight into the underlying mechanisms but also presents a robust strategy to control and fine-tune diastereomer separation, enabling reliable batch-to-batch monitoring of full-length PS siRNAs and providing a robust strategy to control diastereomer resolution.

HILIC has also proven useful for bioconjugated oligonucleotides. For example, a C16 lipid-conjugated 16-mer MOE gapmer ASO with a fully PS backbone was analyzed at 75 °C using 25 mM ammonium acetate to minimize diastereomeric peak splitting. Under these conditions, two isobaric species and an additional unidentified impurity were successfully resolved, illustrating HILIC’s potential for distinguishing hydrophilic-related impurities in lipid-conjugated oligonucleotides.

Finally, in HILIC analysis of oligonucleotides, sample diluent selection presents a major challenge. While water is commonly used to dilute samples in IP-RPLC ensuring oligonucleotide solubility, in HILIC it acts as a strong solvent, compromising retention and causing peak distortion and loss of efficiency. − Dissolving samples in 70–80% acetonitrile improves chromatographic performance but may be limited by oligonucleotide solubility. Reducing injection volume allows aqueous samples to be injected but at the cost of sensitivity. Bracketing injection strategies, such as the Performance Optimizing Injection Sequence (POISe), offer a promising by coinjecting a defined volume of weak solvent with the sample to enhance retention during loading, thereby reducing or eliminating the impact of a strong sample diluent. However, these approaches remain underutilized for oligonucleotides. Notably, despite its critical impact on chromatographic performance, the influence of sample diluent is rarely investigated in HILIC studies of oligonucleotides. ,,

Despite its promise and demonstrated complementarity to IP-RPLC, broader adoption of HILIC remains limited. Its selectivity is generally lower than IP-RPLC, particularly for oligonucleotides longer than ∼ 35 nucleotides. Moreover, HILIC-MS sensitivity typically falls 1–2 orders of magnitude below optimized IP-RPLC methods using alkylamine/HFIP mobile phases, and sodium/potassium adduct formation is more pronounced at low analyte concentrations, complicating data interpretation. Additionally, HILIC retention mechanisms are complex, involving a combination of adsorption, partitioning, and ionic interactions, which are not yet fully understood. These factors make method development less predictable and more time-consuming. Addressing these challenges will be crucial to fully exploiting the potential of HILIC for routine analysis of therapeutic oligonucleotides.

Multidimensional LC

The complementarity of different HPLC modes in oligonucleotide analysis has been extensively investigated. For example, a 23-mer synthetic oligonucleotide bearing phosphorothioate and ribose modifications (2′F, 2′OMe) was used to evaluate chromatographic orthogonality across IP-RP, HILIC, and AEX modes for nine common impurities. Retention time comparisons revealed that coupling HILIC with either AEX or IP-RP provided the highest orthogonality, reflecting the distinct retention behavior of hydrophilic nucleobases under HILIC conditions. While IP-RP achieved the best overall resolution, HILIC and AEX exhibited more coelution. Coefficients of determination (R²) between modes ranged from 0.54 to 0.84 (IP-RP vs HILIC: 0.64; IP-RP vs AEX: 0.84; AEX vs HILIC: 0.54), indicating varying levels of orthogonality.

The complementarity between HILIC and IP-RPLC is further illustrated by Gilar et al., who showed that n/n–1 separations of modified nucleotides are strongly influenced by the hydrophobicity of adjacent bases. Specifically, the loss of a hydrophilic nucleotide at the n position, when adjacent to a hydrophobic nucleotide at n–1, reduces resolution in IP-RPLC but improved separation in HILIC. Conversely, truncation of a hydrophobic nucleotide next to a hydrophilic one produces the opposite effect. The extent of this effect depends on the relative hydrophobicity difference between the involved nucleotides. Additionally, PO impurities eluted later than the PS oligonucleotide in HILIC, whereas the reverse order was observed in IP-RPLC.

To exploit this orthogonality for increased peak capacity, several studies have implemented two-dimensional LC (2D-LC) for oligonucleotide analysis. Li et al. developed a comprehensive 2D method for di- to deca-oligonucleotides, using HILIC in the first dimension and IP-RPLC in the second, with a C18 trap column and aqueous makeup flow for effective ON immobilization between dimensions. Using 27 oligonucleotide standards, the method achieved a high peak capacity of 500 and enabled ESI-MS detection in negative ion mode for clear ON identification.

Goyon and Zhang extended this approach to ASO impurity profiling, developing a versatile multiple heart-cutting 2D-LC method combining AEX or IP-RPLC in the first dimension with MS-compatible HILIC (25 mM ammonium acetate) in the second, coupled to HRMS. This setup allowed detection of impurities down to 0.3%, with particularly high-resolution using IP-RPLC in the first dimension. Optimization of sample loop volumes and column dimensions enabled seamless coupling without solvent modulation. Moreover, the HILIC second dimension provided online desalting of AEX and IP-RPLC eluates, preventing TEA/HFIP contamination in the MS source.

Despite these successful demonstrations of two-dimensional LC for oligonucleotide analysis, several inherent limitations merit consideration. Orthogonality between IP-RPLC, HILIC, and AEX modes is moderate, as retention in all three techniques is primarily governed by oligonucleotide length rather than other physicochemical properties. While even limited orthogonality can increase peak capacity, this underscores that oligonucleotides remain particularly challenging analytes for 2D LC approaches. Moreover, online 2D LC setups are technically complex, and most methods are still performed offline, limiting routine applicability.

Mass Spectrometry

Mass spectrometry is an essential tool for the structural characterization of therapeutic oligonucleotides and their impurities. Nevertheless, MS detection of ONs presents several analytical challenges that can compromise sensitivity, spectral clarity, and data interpretation. ONs typically produce complex MS spectra characterized by broad charge-state distributions and extensive adduct formation, often accompanied by low signal-to-noise ratios. These issues are particularly critical when analyzing low-abundance impurities on single-quadrupole instruments, which lack the sensitivity and resolving power of high-resolution Orbitrap-based systems. Therefore, sufficient sample loading and careful optimization of acquisition parameters are essential for reliable analysis.

Broad m/z acquisition windows are generally required for intact mass confirmation but inherently dilute signal intensity across multiple charge states, reducing sensitivity. As a result, quantitative impurity analysis is commonly performed using narrow m/z windows targeting a single, abundant charge state for extracted-ion chromatograms (EICs), typically combined with higher sample loads (≈20–50 pmol). In practice, intact mass deconvolution over wide m/z ranges is used for component identification, whereas targeted EIC-based approaches are preferred for sensitive impurity detection and quantification.

Adduct formation further complicates spectral interpretation and deconvolution. Alkali metal adducts (Na⁺, K⁺, NH₄ ⁺) are particularly prevalent in HILIC-MS, while IP-RPLC-MS spectra often show adducts involving triethylamine (TEA⁺) and HFIP-related species. The extent of adduction is strongly influenced by solvent purity, glassware contamination, LC system memory effects, and ion-source conditions. Mitigation strategies include rigorous system cleaning (e.g., methane-sulfonic acid flushing), the use of high-purity solvents and glassware, and careful control of source parameters. Partial removal of TEA/HFIP adducts can also be achieved through increased in-source activation by raising gas flows or temperatures.

However, overly harsh source conditions may induce in-source fragmentation (e.g., base loss) or oxidation, generating artifacts that are difficult to distinguish from process- or degradation-related impurities, thereby complicating impurity profiling. Such species typically coelute with the full-length product (FLP), complicating impurity profiling. A careful balance must therefore be struck between adduct suppression and analyte integrity. Typical starting conditions include ion transfer tube temperatures around 320 °C, vaporizer gas temperatures near 350 °C, and source CID energies of approximately 10 eV, which can then be refined as needed.

Validated workflows often combine intact mass deconvolution for identification with EIC-based quantification of impurities. To streamline this process, several HPLC–MS suppliers have developed automated data-processing solutions that perform intact mass deconvolution of the FLP and its impurities and generate EIC-based quantitative reports, providing relative abundance information for all detected species in a standardized format.

Tandem mass spectrometry (MS/MS) plays a central role in the structural elucidation of therapeutic ONs, enabling sequence confirmation, impurity identification, and investigation of degradation pathways. ,, Fragmentation is most commonly induced by collision-induced dissociation (CID) or higher-energy collisional dissociation (HCD), yielding informative backbone fragments. However, fragmentation behavior is highly dependent on precursor charge state, sequence length, chemical modifications, and applied collision energy. As a consequence, MS/MS spectra often exhibit incomplete sequence coverage, dominant neutral losses, or preferential fragmentation pathways. Targeted precursor selection, charge-state control, and careful optimization of collision energies are therefore essential to obtain reproducible and structurally meaningful spectra.

Despite its power, MS/MS analysis of oligonucleotides remains challenging. Spectra are frequently congested with multiple charge states, metal adducts, and modification-dependent fragmentation patterns, hindering automated data interpretation. Moreover, most commercially available software tools are not fully adapted to the chemical diversity of therapeutic ONs, often necessitating extensive manual validation of fragment assignments, particularly for phosphorothioate/phosphodiester chimeras, LNA- and PMO-based chemistries, or conjugated constructs. These limitations highlight the need for optimized MS/MS acquisition strategies and continued development of dedicated bioinformatics tools.

Ion mobility mass spectrometry (IMS) is increasingly recognized as a powerful yet still emerging analytical tool for oligonucleotide characterization. ,, By introducing a gas-phase separation based on ion size, shape, and charge, IMS adds a structural dimension that complements conventional LC-MS strategies. Specifically, IMS enables discrimination of conformers, positional isomers, or sequence variants that are otherwise indistinguishable by m/z alone. As demonstrated in a recent study, the integration of IMS within an online LC–IMS–CID–MS workflow enables the simultaneous separation of multiple charge states in a single injection, facilitating charge-state-specific fragmentation and significantly enhancing sequencing information content and analytical throughput. Nevertheless, limitations remain, including incomplete sequence coverage for longer or chemically modified oligonucleotides, as well as the need for analyte-specific optimization of collision energies and advanced data-processing workflows.

Earlier work by Demelenne et al. demonstrated that IMS-derived collision cross-section (CCS) measurements provide valuable, sequence-independent structural information, enabling discrimination between phosphodiester and phosphorothioate backbones while supporting the assessment of oligonucleotide length and conformational heterogeneity. More recently, Bill et al. showed that although intact siRNA strands exhibit limited ion mobility differences, a fragment-based IMS–MS/MS approach enables gas-phase resolution and quantitative analysis of PTO diastereomers. These studies, together with other recent reports, highlight that IMS, particularly when combined with gas-phase fragmentation, offers a unique and powerful complement to LC-MS methods for the characterization of PTO diastereomers that would otherwise be challenging to resolve. ,,,

Despite its clear advantages, IMS remains costly, technically demanding, and not yet widely accessible. Fully integrated LC–IMS–MS platforms require advanced instrumentation, specialized expertise, and complex data analysis, limiting routine adoption, particularly in regulated environments. IMS separations are highly sensitive to ion charge state and experimental conditions and often rely on indirect strategies, such as gas-phase fragmentation, rather than direct separation of intact species. Furthermore, drift-time reproducibility, calibration strategies, and standardized data-processing workflows are still under development for pharmaceutical QC applications. Nevertheless, these studies demonstrate that IMS provides a substantial analytical advantage and strong future potential for oligonucleotide analysis, especially for complex, modified, and stereochemically heterogeneous systems where conventional LC–MS approaches reach their limits.

Bioanalysis for PK/PD Study

The bioanalysis of ONs has progressed rapidly to support the growing needs of preclinical and clinical pharmacokinetic/pharmacodynamic (PK/PD) studies. LC–MS has become the dominant analytical platform due to its ability to provide both quantitative sensitivity and structural selectivity, particularly for differentiating full-length ONs from structurally similar nuclease-generated metabolites.

IP-RPLC–MS remains the most established approach, delivering high sensitivity at subng/mL levels and improved, though still challenging, selectivity between parent oligonucleotides and truncated metabolites. − Indeed, chromatographic resolution remains a major bottleneck, especially for closely related or isobaric metabolites, and the selection of optimal ion-pairing reagents continues to rely largely on empirical, trial-and-error approaches.

As a result, ion-pair-free alternatives such as HILIC–MS have gained increasing attention. ,, HILIC methods employing amide stationary phases and ammonium-based mobile phases have demonstrated promising performance for ON bioanalysis. However, this approach is currently limited by relatively low chromatographic resolution, highlighting the need for further development of HILIC sorbents specifically tailored for oligonucleotide separation.

Sample preparation continues to be a critical determinant of assay performance in complex biological matrices. Ion-pair-free extraction strategies, including solid-phase extraction and hybridization-based methods, are increasingly explored to reduce ion suppression and improve MS compatibility. However, these approaches require further optimization of sorbent chemistry and extraction mechanisms, as insufficient recovery and poor reproducibility can still limit their broader application. ,

Overall, current limitations in therapeutic ON bioanalysis stem primarily from insufficient chromatographic resolution and suboptimal extraction performance, particularly for low-abundance metabolites. In this context, high-resolution mass spectrometry offers a clear advantage over unit-resolution instruments by improving selectivity, mitigating crossed-ion interferences, and enabling retrospective metabolite identification. Continued advances in chromatographic materials and high-resolution quantitative MS are expected to further strengthen PK/PD and biodistribution studies of therapeutic oligonucleotides.

Perspectives

The analysis of therapeutic oligonucleotides faces both growing challenges and exciting opportunities, driven by the rapid expansion of this field. As oligonucleotide-based therapies continue to evolve in terms of chemical modifications, formulation complexity, and regulatory expectations, there is an increasing need for analytical methods that are robust, high-performance, and adaptable. At the same time, there is a strong push toward greener and more sustainable practices, particularly regarding the use of ion-pairing reagents and perfluoroalkyl substances in IP-RPLC, which raise environmental and operational concerns. To address these issues while maintaining MS compatibility, alternative strategies such as ion-pair-free reverse-phase and HILIC approaches are being explored. Notably, a recently developed ammonium bicarbonate–based non-IP-RP-LC-MS method achieves sensitivity comparable to conventional IP-RPLC-MS while avoiding environmentally problematic ion-pairing reagents.

Alternative reversed-phase strategies are being developed to reduce reliance on toxic reagents, with phenyl-based stationary phases emerging as promising MS-compatible options that maintain high selectivity and separation performance. , Retention is influenced by oligonucleotide secondary structure and π–π interactions, highlighting the need to consider molecular properties in method design. In parallel, greener chromatographic approaches, such as micro-LC systems operating at low flow rates, offer the potential to substantially reduce solvent consumption, lower costs, and minimize the environmental footprint of oligonucleotide analysis.

Therapeutic oligonucleotides commonly incorporate chemical modifications or bioconjugations, including lipid ,, and sugar conjugates, ,, that substantially alter their chromatographic behavior. In IP-RPLC, more lipophilic sugar modifications, such as locked nucleic acids LNAs, 2′-O-methyl, 2′-MOE, and 2′-fluoro substitutions, typically increase retention relative to unmodified DNA or RNA. , This effect scales with modification density and is particularly pronounced for LNAs, owing to their rigid C3′-endo sugar conformation. In contrast, in HILIC, increasing modification lipophilicity leads to decreased retention and reshapes selectivity, revealing an inverse trend relative to IP-RPLC. Moreover, HILIC is highly sensitive to the interplay between nucleotide hydrophobicity and hydration, making highly modified oligonucleotides especially difficult to resolve in this mode. These analytical difficulties are further amplified for charge-neutral backbones such as PMOs, where separation of full-length species from n–1 deletion impurities is particularly challenging and depends largely on nucleobase ionization. , Taken together, despite increasing interest and recent progress, there remains a clear need for more systematic studies explicitly linking chemical modifications and bioconjugations with chromatographic behavior, analytical performance, and selectivity. Likewise, systematic impurity profiling of these conjugates is still limited, representing a significant analytical gap as such chemistries gain prominence in next-generation oligonucleotide therapeutics. ,,

Multidimensional LC strategies and high-resolution MS offer enhanced separation, selectivity, and structural characterization, yet their routine adoption in industrial and QC settings remains limited due to complexity, instrumentation requirements, and cost.

In addition, higher-order oligonucleotide structuring, particularly in lipid-conjugated systems, remains an important yet challenging aspect to characterize. Techniques such as circular dichroism, spectroscopic methods, melting temperature analysis, and phosphorus NMR provide valuable insights into oligonucleotide secondary structure and stability. −

Stereochemistry adds another layer of analytical complexity, as phosphorothioate diastereomers can significantly affect efficacy, pharmacokinetics, and safety. ,,− While recent advances have improved chromatographic separation, confident stereochemical assignment remains challenging. Retention trends in IP-RPLC often correlate with Sp content, but exceptions arising from certain sugar modifications underscore the need for confirmatory approaches, such as stereoselective nuclease digestion or stereopure reference standards. However, nuclease activity can be affected by sugar chemistry, and access to stereopure markers remains limited by cost, yield, and achievable oligonucleotide length. Looking ahead, the separation of diastereoisomers of bioconjugated oligonucleotides is expected to present an additional analytical challenge. On the production side, high-resolution preparative HPLC combined with advances in stereocontrolled synthesis is expected to enable the generation of stereochemically defined oligonucleotides. Together, these analytical and synthetic developments are essential for the advancement of next-generation stereochemically pure oligonucleotide therapeutics, a field in which alicaforsen remains the only compound to have reached preregistration. ,

MS detection is undeniably a key tool for oligonucleotide analysis to detect and characterize the structure of low-abundance impurities and metabolites. HRMS provides a decisive advantage for the analysis of therapeutic oligonucleotides by enabling accurate mass measurements and confident discrimination of closely related impurities, metabolites, and full-length sequences. However, HRMS alone cannot fully overcome limitations arising from chromatographic coelution, ion suppression, or in-source artifacts, particularly for quantitative applications, underscoring the need for highly selective and orthogonal analytical methods to ensure robust identification and quantification. In addition, the high cost and limited availability of HRMS in quality control laboratories currently restrict routine use. Nevertheless, as affordability and automation improve, HRMS is expected to become more widely implemented, shaping future regulatory and quality control strategies for oligonucleotide therapeutics.

In bioanalysis, accurate quantification of therapeutic oligonucleotides and their metabolites in complex biological matrices remains a major challenge, particularly to meet the growing demands of both preclinical and clinical development. ,, Chromatographic resolution is often limiting when distinguishing full-length sequences from closely related nuclease-derived metabolites, while matrix effects, ion suppression, and low-abundance metabolites further compromise sensitivity and reproducibility. Sample extraction is essential to reduce ion suppression, but suboptimal extraction strategies can adversely affect recovery and reproducibility, representing a critical bottleneck in bioanalytical workflows. Robust, highly sensitive methods are therefore essential for pharmacokinetic and biodistribution analyses. In this context, high-resolution MS, combined with optimized chromatographic and extraction strategies, offers a clear advantage by improving selectivity and enabling confident metabolite identification.

Looking ahead, the pharmaceutical industry requires analytical methods that combine robustness, speed, and automation to meet current regulatory standards while anticipating future requirements. To enable efficient technology transfer and routine implementation, methods must be not only scientifically rigorous but also compatible with high-throughput and industrial workflows. A major challenge remains the lack of standardized and harmonized analytical protocols for oligonucleotide analysis, which hampers interlaboratory comparability and limits widespread adoption in quality control environments.

As oligonucleotide manufacturing scales up, analytical strategies must also be adapted to preparative purification workflows that employ cleaner, more industrially compatible conditions, while ensuring compliance with Good Manufacturing Practices (GMP). In parallel, the emergence of new oligonucleotide modalities, including mRNA, CRISPR guide RNAs, and aptamers, introduces additional complexity due to their distinct physicochemical properties and impurity profiles, underscoring the need for novel and adaptable analytical solutions. Overall, the future of oligonucleotide analysis will rely on integrated, high-resolution, and flexible strategies that balance analytical performance, regulatory compliance, scalability, and environmental sustainability.

Conclusion

The field of therapeutic oligonucleotides has witnessed unprecedented growth in recent years, driven by advances in chemical modifications, delivery strategies, and clinical success stories. This expansion has created parallel challenges in the analytical domain, where precise characterization of increasingly complex structures and impurities remains critical. Traditional approaches such as IP-RPLC continue to provide a robust foundation, but their reliance on ion-pairing reagents and HFIP additives raises sustainability concerns and can compromise MS compatibility. Complementary MS-compatible techniques such as HILIC, IP-free RPLC, along with multidimensional LC–MS workflows, are emerging as powerful tools to address selectivity, sensitivity, and orthogonality needs. Importantly, the rise of bioconjugated, sugar- or lipid-modified oligonucleotides highlights the necessity of tailored analytical solutions rather than universal methods. At the same time, the drive toward greener and more cost-effective practices is redefining oligonucleotide analysis in laboratories. Moving forward, progress will depend on integrating complementary chromatographic strategies with advanced mass spectrometry, while continuously adapting to evolving regulatory expectations. Ultimately, developing robust, sensitive, and eco-friendly methods will be key to ensuring the safety, effectiveness, and accessibility of next-generation oligonucleotide therapeutics.

Biographies

Sandy Al Bardawil is currently a PhD student in Analytical Chemistry within the DNAlytics group at the ARNA Research Unit (INSERM U1212, CNRS 5320, University of Bordeaux, France), under the supervision of Dr. Ludivine Ferey and Prof. Philippe Barthélémy, and supported by a SiTH doctoral fellowship. She obtained her Master’s degree in Analytical Chemistry from the University of Bordeaux in 2024. Her research focuses on developing liquid chromatography–mass spectrometry methods for the characterization of therapeutic oligonucleotides, with particular emphasis on structural identification, impurity profiling, and degradation product analysis.

Philippe Barthélémy is a French chemist and professor whose research focuses on nucleic acid–based nanomaterials and amphiphilic oligonucleotides for biomedical applications. He is recognized for developing lipid–oligonucleotide conjugates that self-assemble into functional nanostructures for enhanced gene regulation and therapeutic delivery, particularly in cancer or neurological diseases.

Ludivine Ferey is currently an Assistant Professor in the Department of Analytical Chemistry at the UFR of Pharmaceutical Sciences, University of Bordeaux, France, where she has held a faculty position for 12 years, and a researcher at the ARNA Research Unit (INSERM U1212, CNRS 5320, University of Bordeaux, France). She is part of the direction of the DNAlytics group within ARNA. She obtained her PhD in Analytical Chemistry from AgroParisTech in 2013. Her research focuses on the development of innovative and environmentally friendly separation methods for pharmaceutical compounds, with particular emphasis on LC-MS and CE for the analysis of therapeutic oligonucleotides.

Sandy Al Bardawil: Writing – review and editing. Philippe Barthélémy: Writing – review and editing, Supervision, Project administration, Funding acquisition. Ludivine Ferey: Writing – original, Writing – review and editing, Supervision, Funding acquisition

The authors declare the following competing financial interest(s): P.B. holds equity in NAXIA Discovery, a start-up company. All other authors have no competing interests.

Published as part of Analytical Chemistry special issue “Fundamental and Applied Reviews in Analytical Chemistry 2026”.