The meteoric rise of GLP-1 receptor agonists (GLP-1 RAs) may seem like a bolt from the blue. In reality, the household-name success of Ozempic – one of several semaglutide brands – reflects a broader wave of innovation in peptide therapeutics.
Peptides occupy a fascinating middle ground between small molecules and biologics – they can be engineered with highly tunable pharmacokinetic and pharmacodynamic profiles, while offering advantages like lower immunogenicity, better tissue penetration, and reduced manufacturing costs compared to antibodies or proteins. The excitement around GLP-1 RAs is especially driven by their proven clinical success in type 2 diabetes and obesity, and the fact that they’re now moving far beyond glucose control. Their ability to promote weight loss and reduce cardiovascular risk has expanded their appeal and created a very dynamic development pipeline. We’re also seeing rapid growth in multi-agonist therapies that expand beyond traditional single-target approaches. Tirzepatide, the first dual GIP/GLP-1 RA, is a good example; and newer candidates like retatrutide (phase III trials), which targets three receptors, are pushing the boundaries even further by demonstrating significant weight loss, improved glycemic control, and enhanced fat metabolism. Many other GLP-1 RAs are currently under clinical investigation, exploring additional therapeutic areas such as liver disease, kidney disease, and even neurodegenerative disorders.
Another strong trend is innovation in drug delivery. Traditionally, peptides have faced challenges with oral bioavailability and stability, but with the development of technologies like permeation enhancers and structural modifications, we’re now seeing oral formulations become a reality –, as well as extended lifetimes that enable the development of long-acting injectable formulations. This is a real game-changer for the development of GLP-1 RAs and peptide therapeutics more broadly. Oral semaglutide (Rybelsus) marks an important milestone, and we’re seeing intense activity around alternative delivery platforms, such as implantables, depots, and even transdermal systems, that could simplify treatment and improve adherence. On the molecular design side, modifications such as lipidation, PEGylation, cyclization, and fusion (e.g., with albumin, IgG Fc fragment) are being used to extend half-life, improve stability, and enhance receptor selectivity. These chemical strategies are what make once-weekly dosing possible for drugs like semaglutide, and they’ll continue to drive next-generation development. It’s important to highlight that peptides like GLP-1 RAs are now being integrated into broader therapeutic strategies, such as fixed-dose combinations with insulin, dual or triple agonists. Patient-centric considerations are equally important, with a growing demand for treatments that are easier to use –, whether that means less frequent dosing, oral options, or alternatives to injections.
Nevertheless, innovation in GLP-1 RAs isn’t driven by science alone. The remarkable commercial success of drugs like semaglutide and tirzepatide has created strong market momentum, driving significant investment in next-generation molecules and accelerating development cycles. Companies are racing to differentiate their candidates not only by efficacy, but also by safety, convenience, and route of administration.
In short, the trends that stand out are clear: multi-target designs, smarter delivery systems, structural innovations that expand the therapeutic scope and stability of peptides.
The complex reality of peptide analysis
What initially drew us to work with GLP-1 RAs was the combination of their exciting therapeutic potential and the complex analytical challenges they present, demanding fit-for-purpose analytical procedures. This applies not only to GLP-1 RAs but broadly across peptide therapeutics.
Peptides are inherently more complex than small molecules, which leads to a broader spectrum of product-related impurities arising during synthesis, purification, formulation, or storage. Many of these impurities (e.g., truncated sequences, modified residues, amino acid insertions, isomerization, aggregation products, etc.) may have physicochemical properties very similar to the therapeutic peptide. This makes detection and quantification particularly challenging: sensitive and highly selective methods are often required to distinguish these impurities from the main peptide. Complex formulations and multi-agonist designs further complicate characterization and quantification. When dealing with dual or triple agonists, or fixed-dose combinations with insulin, it is critical to develop assays that can accurately measure each component without interference.
Techniques like ion-pairing reversed-phase liquid chromatography (IP-RPLC) are widely used to monitor peptide critical quality attributes (CQAs), with hyphenated LC-MS approaches playing a key role in enhancing sensitivity and selectivity, enabling quantification of low-level impurities. SEC with UV or MALS detection is widely used to assess oligomers, monomeric purity, and aggregation.
Accurate sequence confirmation, including differentiation of isomers or isobaric variants, and structural elucidation of low-level impurities often require more advanced MS/MS techniques, such as electron-capture dissociation (ECD) or electron-transfer dissociation (ETD). High-resolution MS systems may also be necessary to achieve precise identification and characterization of impurities. It is worth noting that access to high-end MS technologies can be challenging in QC settings, due to the high cost, operational complexity, and need for specialized expertise, which can limit their routine use for CQA monitoring and low-level impurity characterization. Ion mobility-mass spectrometry (IM-MS) has also been applied to distinguish closely related peptides, such as those containing isoAsp residues. Recovery can also be particularly challenging in complex formulations, such as solid dosage forms compared to solutions, or multi-agonist combinations versus single therapies, requiring optimized sample preparation to ensure efficient extraction, minimize losses, and preserve peptide integrity. Closely linked to this, MS techniques, particularly for complex formulations, are highly susceptible to matrix effects that can affect recovery and ionization efficiency of both the peptide and its impurities. Internal standards (e.g., isotopically labeled peptides) may be needed to compensate for these matrix effects, ensuring precise and accurate quantification in LC-MS analyses. In addition, the availability of reference standards is essential to support reliable and accurate quantification in general.
Since we’ve touched on chromatography for peptide analysis, it’s worth noting that peptides are prone to adsorption onto surfaces during analysis (often due to ionic or dipole–dipole interactions with metals in the LC flow path), reducing sensitivity and compromising the accurate quantification of low-level impurities. To mitigate this, low-adsorption flow paths, such as low adsorption columns and system components, are increasingly used to minimize peptide-surface interactions, improve recovery, and enhance overall method sensitivity and accuracy.
Another important point to highlight is the impact of delivery-related modifications, such as lipidation, PEGylation, or Fc/albumin fusion. While these modifications are essential for half-life extension, they significantly increase the complexity of analytical workflows. They introduce additional CQAs that must be carefully monitored, including the proportion of conjugated versus unconjugated peptide, the presence of free form of the conjugate moiety (e.g., lipids, PEG chains), and the occurrence of di- or poly-PEGylation. Conjugation is typically achieved via chemical linkers, adding another layer of complexity to characterization and control, as well as necessitating highly selective and sensitive analytical methods to accurately assess the modified species and ensure consistent product quality. Furthermore, these modifications can affect solubility, chromatographic behavior, and mass spectrometry response, making method development and validation even more challenging.
The use of orthogonal analytical approaches, such as HILIC and 2D-LC, can significantly enhance analytical workflows by providing complementary selectivity. These strategies help confirm findings, resolve co-eluting species, and enable a more comprehensive characterization of CQAs.
Finally, there’s a strong push toward high-throughput, life cycle-oriented analytics. Companies want methods that not only meet regulatory requirements, but also support process development, in-process control, and long-term stability monitoring. Balancing sensitivity, selectivity, and robustness across the full peptide lifecycle remains a key challenge.
The importance of Impurity profiling
Impurity profiling is especially critical for GLP-1 RAs because of the vast diversity of impurities – both structurally related and unrelated. These drugs can range from active pharmaceutical ingredients (APIs) composed purely of natural amino acids, as in the case of exenatide, to more complex molecules that incorporate non‑canonical amino acids and are conjugated to small molecules (such as fatty acids), as in liraglutide, semaglutide, and tirzepatide – resulting in a unique set of properties not observed with earlier peptide therapeutics. This creates a distinct and more challenging impurity profile that requires rigorous characterization.
These impurity profiles are particular to a manufacturing route and may require control at both intermediate stages and in the final product. GLP‑1 RAs can be manufactured entirely via chemical synthesis or via a hybrid approach – combining the latter with biosynthetic steps, as in the case of semaglutide. Within each production route, alternative approaches are possible, further expanding the range of possible impurities for a given API. Beyond their direct impact on safety, efficacy and quality, impurity profiles can serve as fingerprints, reflecting variations in the (bio)synthetic process and the quality of employed starting materials.
It is worth noting that, as a consequence of the exceptional clinical success of GLP-1 RA drugs, and the emergence of more convenient administration routes (e.g., oral), the demand for these drugs has risen sharply, leading to compounding and counterfeiting of marketed drug products using APIs supplied by different manufacturers. This has resulted in drug substances or products that presenting different impurity profiles compared to the original drug. Another important aspect is that, because compounding oversight is typically less stringent than for commercial drug manufacturers, the risk of substandard products is significantly higher. Beyond this, cases of fraudulent compounded GLP-1 RAs have been reported by the FDA in the US, including products with false or misleading information on their labels. In some instances, the compounding pharmacies listed on the labels didn't exist at all! In other instances, the labels falsely named licensed pharmacies that, according to FDA findings, had not compounded the products in question. This situation – which has been exacerbated by the ease with which these medicines can be accessed through (sometimes unregulated) online suppliers – is extremely important for the scientific, healthcare, and patient communities to be aware of.
Looking ahead, the introduction of generic drugs may present specific challenges, making the accurate identification of impurities and assurance that they remain below safe levels critically important.
HILIC: a tool for next-generation workflows?
At RIC group, we have been searching for orthogonal separation mechanisms and streamlined workflows for GLP-1 RAs. This led us to HILIC, which we have explored for impurity profiling and beyond.
We have used a HILIC column to successfully separate polar and ionic/ionizable analytes – which include various excipients, the API, and related impurities – before detection using an evaporative light scattering detector (ELSD) and MS.
HILIC provides complementary information to RPLC: usually offering higher retention for polar analytes that typically have poor retention and often elute near the void volume in RPLC mode. A clear example of non-retained compounds in RPLC includes the buffering and tonicity agents commonly found in liquid formulations of GLP-1 RAs, such as sodium chloride, sodium phosphate, both present in top -selling drugs Mounjaro (tirzepatide) and Ozempic. Separation of these ionic salts and other polar excipients (e.g., glycerol, mannitol, sucrose, etc.) is feasible and straightforward using HILIC.
Furthermore, we evaluated and compared HILIC columns with conventional stainless‑steel (SST) hardware and low adsorption, corrosion-resistant hardware. The use of low adsorption LC column hardware has shown to be crucial for achieving good peak shape and high sensitivity for several metal-sensitive GLP-1 RA APIs, related impurities and excipients – enabling the development of sensitive methods for impurity profiling with good recovery. It’s worth noting that the use of low adsorption LC systems may further help improve sensitivity in challenging situations.
We found, to our surprise, excessive tailing of GLP-1 RAs. Many GLP-1 peptides and their impurities are moderately to strongly metal-sensitive, resulting in non-specific adsorption (NSA) to surfaces containing metals, such as stainless-steel (SST) surfaces. Comparison of low adsorption and traditional LC columns hardware provided clear evidence of the impact of this NSA on chromatographic efficiency, leading to poor peak shapes and significantly reduced sensitivity – an issue that analytical scientists strive to minimize when monitoring impurities. The improvement in peak shape and increased sensitivity with the use of the low adsorption LC column hardware was striking!
Further inspection of the impurity profiles made evident that conventional SST hardware may struggle to detect several low-level impurities. Particularly, peaks eluting near the API peak tail were hardly detectable with conventional SST hardware. The extent of this effect varies among different peptides and their related impurities. For example, peptides containing more acidic amino acid residues (which may be negatively charged in solution) tend to be more sensitive and exhibit lower efficiency due to stronger NSA. Interestingly, lipidation can shield potential interaction sites on the molecule, reducing the impact of NSA. This was evident when comparing exenatide and semaglutide chromatographic profiles: exenatide, a non-lipidated peptide with several acidic residues, was strongly affected by NSA (likely due to ionic and dipole-ionic interactions with metal surfaces), leading to decreased efficiency. In contrast, semaglutide, a lipidated peptide with fewer acidic residues, showed similar chromatographic behavior on both conventional and low adsorption LC hardware. Similarly, the efficiency of analyzing excipients was also influenced by the type of column hardware. For instance, low sensitivity and poor recovery of phosphate on conventional SST hardware were observed. Although this phenomenon has been well described in the literature, we expected it to be less pronounced at the concentration levels used as an excipient, which was not the case.
Beyond NSA-involving metals in the LC flow path, we also believe that ionic interactions between positively charged sites on peptides or related impurities and residual siloxide groups on commonly used silica particles can exacerbate NSA, leading to sensitivity issues and poor peak shapes. In this context, peptides or impurities containing more basic amino acids are particularly affected. The development of technologies to mitigate NSA in HILIC mode is therefore critically important for GLP-1 RA analysis and, more broadly, for peptide analysis in general.
Looking ahead, we see advanced HILIC-based methods playing an increasingly crucial role in analyzing highly polar peptides and their impurities, which can be difficult to separate using conventional reversed-phase LC. HILIC may provide complementary selectivity, enabling the resolution of low-level impurities, including isomers that may co-elute in RPLC. These methods are also highly useful for monitoring polar/ionic excipients, such as peptide counter-ions, salts, and buffers, often with minimal sample preparation. To address detectability challenges with UV detection, HILIC can be hyphenated with ELSD, allowing effective monitoring of excipients like sodium, phosphate, and other components. In addition, HILIC is usually run with MS-compatible solvents and buffers, which can provide a solution when IP-RPLC conditions cause ionic suppression –, further improving sensitivity during LC-MS analysis of low-level impurities. This approach strengthens the analytical toolkit to going beyond gold-standard techniques like IP-RPLC, enabling efficient tracking of both excipient- and peptide-related CQAs, which help to ensure overall drug product quality. In general, available pharmacopeial methods commonly used in GMP settings for peptides primarily rely on IP-RPLC, often using non–MS-compatible ion-pairing reagents, with only a few methods employing more MS-compatible alternatives.
It is worth noting that robustness and repeatability can be more challenging in HILIC analyses due to its complex retention mechanism. In RPLC, retention is primarily driven by hydrophobic interactions (with partitioning and adsorption due to other intermolecular interactions also playing a role) which are relatively stable and less influenced by minor external factors. (By the way, this is precisely what makes HILIC an orthogonal technique to RPLC!) By contrast, retention in many HILIC columns depends on the semi-immobilized water layer formed at the interface of the polar stationary phase and the organic-rich mobile phase. Small disturbances to this water layer can significantly affect retention, making HILIC more susceptible to variability. Consequently, HILIC method development and validation require thorough risk assessment during method development to identify and minimize sources of variability and ensure robust methods suitable for GMP settings (which is entirely possible!). NSA, which can compromise efficiency, is also a concern in HILIC analysis. Mitigation strategies, such as using LC columns with low adsorption technologies, may be needed to address this issue (or at least it will make the analytical scientist’s life much easier!).
For GLP-1 RAs and other peptides, we see the potential of developing HILIC methods to become part of the analytical workflows for peptides CQA monitoring in GMP and non GMP settings, complementing the information provided by other more common techniques.
Toward integrated and robust workflows
Thinking more broadly, as GLP-1 drugs continue to expand in both therapeutic scope and commercial scale, analytical strategies will need to evolve to keep up with demand and complexity. Today,current instrumental capabilities and methods can address many analytical challenges for GLP-1 RAs. However, the parallel execution of different method platforms, along with consistent data handling and interpretation, remains fragmented and not implemented effectively. The future of analytical strategies lies in the integration and development of methodologies that streamline workflows and can measure multiple attributes simultaneously, expediting both drug development and quality control.
For example, LC-based approaches can leverage a battery of orthogonal separation mechanisms (e.g., RPLC, HILIC, SEC, IEX), hyphenated to different detection systems (e.g., UV, ELSD, RI, MALS, MS) to expand the coverage of measurable properties.
Rather than increasing workflow complexity, simplicity should be prioritized by developing robust, and ideally multi-attribute methods that can be implemented in R&D and quality control laboratories, and extended to in-line continuous monitoring platforms during production. It is worth mentioning that sample preparation remains a challenging bottleneck in various workflows. This crucial aspect of the analytical method is still predominantly performed manually or with limited automation; as a result, smooth progress and transition to fully automated workflows remain in need of further development.
Finally, ongoing and future implementation of modeling, risk assessment, and other AI tools to assist method development, data processing, and technological development, workflows will strengthen decision-making capabilities. This enables the design of more robust methods and appropriate analytical control strategies to ensure procedures remain fit-for-purpose along the analytical procedure lifecycle. The good news is that several guidelines have been published recently which are paving the way for the implementation of risk-based method development approaches and the management of procedure lifecycle closely to the product lifecycle. Think of ICH Q14 on analytical procedure development and USP <1220> on the analytical procedure lifecycle. But that’s a whole other story for another article!
To conclude, it is an exciting time to be in the GLP-1 RA space – and more broadly across peptide therapeutics, which hold great promise for treating numerous diseases while offering several advantages over antibodies and proteins. Equally exciting are innovative peptide delivery systems capable of targeting specific cells or tissues, transporting nucleic acids, proteins, radionuclides, or drugs, and enhancing cellular uptake through approaches such as peptide–drug conjugates, cell-penetrating peptides, and functionalized nanoparticles. Tackling the challenges of increasingly complex GLP-1 RAs – developing robust analytical procedures to characterize low-level impurities and ensure consistent product quality – demands creative thinking and innovative solutions, and ultimately supports the development of safer, more effective peptide-based medicines.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
