From Complexity to Clarity: Expanding Metabolome Coverage With Innovative Analytical Strategies

ABSTRACT

Metabolomics, a powerful discipline within systems biology, aims at comprehensive profiling of small molecules in biological samples. The challenges of biological sample complexity are addressed through innovative sample preparation methods, including solid‐phase extraction and microextraction techniques, enhancing the detection and quantification of low‐abundance metabolites. Advances in chromatographic separation, particularly liquid chromatography (LC) and gas chromatography (GC), coupled with high‐resolution (HR) mass spectrometry (MS), have significantly improved the sensitivity, selectivity, and throughput of metabolomic studies. Cutting‐edge techniques, such as ion‐mobility mass spectrometry (IM‐MS) and tandem MS (MS/MS), further expand the capacity for comprehensive metabolite profiling. These advanced analytical platforms each offer unique advantages for metabolomics, with continued technological improvements driving deeper insights into metabolic pathways and biomarker discovery. By providing a detailed overview of current trends and techniques, this review aims to offer valuable insights into the future of metabolomics in human health research and its translational potential in clinical settings. Toward the end, this review also highlights the biomedical applications of metabolomics, emphasizing its role in biomarker discovery, disease diagnostics, personalized medicine, and drug development.

1. Introduction

Metabolomics, a recent addition to the “omics” field, focuses on the quantitative and qualitative analysis of metabolites central to biological processes [1]. It targets compounds with molecular weights ranging from 50 to 2000 Da [2], aiming to comprehensively measure and characterize each molecule in complex biological matrices like serum, plasma, and urine [3]. Metabolomics enhances our understanding of both normal physiological functions and pathological conditions. Its global application facilitates rapid metabolite identification across diverse biological systems, owing to advancements in biomedical instrumentation. Various mass spectrometry (MS) and chromatography techniques, such as capillary electrophoresis (CE), gas chromatography (GC), liquid chromatography (LC), time‐of‐flight (TOF) MS, Orbitrap, and Fourier transform ion cyclotron resonance (FTICR) MS, have significantly bolstered metabolomics capabilities [4, 5]. In addition, nuclear magnetic resonance (NMR) spectroscopy [6, 7] and chromatography coupled with high‐resolution (HR) MS [8, 9, 10] are pivotal methods enabling HR identification of molecular species [11]. Untargeted metabolomics using LC–MS proves valuable in uncovering disease‐related metabolic changes and identifying potential biomarkers for early disease detection and therapeutic monitoring [12]. In metabolomics research, although several analytical techniques are available, LC–MS has emerged as the most widely used for several reasons. The predominance of LC–MS can be attributed to its superior sensitivity, specificity, and versatility in analyzing a wide range of metabolites [13]. LC–HR–MS with newest generation of instruments offers exceptional mass accuracy and the ability to handle complex biological samples, providing high‐throughput capabilities crucial for large‐scale metabolomic studies. However, other techniques, like NMR, GC–MS, and CE–MS, are often perceived as less widespread due to certain limitations when compared to LC–MS. Ultimately, the widespread adoption of LC–MS can also be attributed to the extensive commercial availability of LC–MS platforms, standardized protocols, and broad applicability across various types of metabolomics research [14]. Thus, although other techniques have their strengths, LC–MS has become the method of first choice due to its ability to provide a comprehensive metabolomic profile in a wide range of biological samples [15].

Catabolism and anabolism are metabolic processes that generate small molecule metabolites, which encompass a variety of compounds, including sugars, nucleic acids, organic acids, amino acids, lipids, fatty acids, and peptides. These metabolites serve as critical links between an individual's genotype and phenotype, reflecting interactions with their environment and their transcriptomic, proteomic, and genomic profiles [16]. They act as significant bioindicators throughout cellular processes and are essential to biological pathways [17, 18, 19]. Identifying biomarkers is pivotal for disease diagnosis, prognosis, and personalized treatment strategies. Additionally, it aids in exploring and elucidating the molecular pathways and mechanisms implicated in disease progression or influenced by therapeutic interventions. A crucial step in the biomarker discovery process is the identification of active metabolites [20, 21, 22, 23, 24].

Figure 1 illustrates the critical steps involved in a metabolomics study. The Metabolomics Standards Initiative (MSI), established in 2007, has set stringent reporting criteria across the metabolomics workflow to ensure data integrity and comparability in the field [25, 26]. The overall metabolite profile, which consists of amino acids, lipids, carbohydrates, and organic acids, is influenced by various factors, including environmental conditions, nutrition, toxicity, infections, inflammation, genetics, microbiota composition, and disease states [1]. The analytical workflow in metabolomics typically encompasses several vital steps: sample collection, sample preparation, instrumental analysis, metabolite annotation, and the analysis of metabolic pathways. These processes collectively contribute to a comprehensive understanding of the metabolic landscape in biological systems [27].

FIGURE 1.

Workflow for metabolic profiling using a metabolomics approach. The figure illustrates the critical steps involved in a metabolomics study. First, biological samples, such as blood, plasma, or cells/tissues, are collected under controlled conditions to preserve their metabolic integrity. Next, the samples undergo preparation and extraction using techniques like protein precipitation, liquid–liquid extraction, or solid‐phase extraction to isolate metabolites. The processed samples are analyzed using advanced analytical platforms, including liquid chromatography–mass spectrometry (LC–MS), gas chromatography–mass spectrometry (GC–MS), and capillary electrophoresis–mass spectrometry (CE–MS), which provide comprehensive metabolite profiling. Finally, data analysis and processing are performed using specialized software to identify, quantify, and interpret metabolite patterns, offering insights into biochemical pathways and biological processes.

The thorough examination of the metabolome presents numerous analytical challenges, primarily due to significant variations in physicochemical properties, including polarity, solubility, pKa values, and molecular mass; the extensive dynamic range required to analyze both trace metabolites and those present in high abundance, spanning up to many orders of magnitude; and the existence of various isomers that share structural similarities but exhibit considerable differences in biological activity, particularly among lipid‐based signaling molecules [28]. This complexity underscores the necessity for advanced analytical methodologies to effectively address these challenges, enabling both qualitative and quantitative evaluations of the metabolome while maximizing metabolic coverage through enhanced resolution and selectivity [29].

In this context, metabolomics has dramatically benefited from recent chromatography and MS advancements over the past two decades. The adoption of reversed‐phase LC (RP‐LC) columns featuring sub‐2‐µm fully porous particles (ultra‐high performance LC, UHPLC) or sub‐3‐µm superficially porous particles (core–shell technology) has emerged as a well‐established technique within metabolomics, delivering significant improvements in resolution and throughput compared to traditional high‐performance LC (HPLC) [29, 30, 31]. On the other hand, recent innovations in LC and MS, including hydrophilic interaction liquid chromatography (HILIC), supercritical fluid chromatography (SFC), multidimensional LC, ion‐mobility MS (IM–MS), and data‐independent acquisition (DIA) approaches, remain underutilized in metabolomics, despite their potential to enhance metabolite coverage. This review article aims to explore the latest advancements in chromatography and MS, emphasizing their capability to improve metabolome coverage.

Polar and non‐polar metabolites exhibit significant differences in their chemical properties, leading to the development of specialized analytical techniques to study them. Polar metabolites, such as amino acids, sugars, and nucleotides, are hydrophilic and readily dissolve in water [32]. Their separation and analysis often require techniques like MS hyphenated with HILIC [33] or RP‐LC [34] that provide excellent resolution and sensitivity for these compounds. On the other hand, non‐polar metabolites, such as lipids, are hydrophobic and exhibit poor solubility in aqueous environments. Their structural diversity, including various glycerophospholipids, glycerolipids, sphingolipids, and sterols, poses challenges for traditional metabolomic techniques. These unique characteristics led to the emergence of lipidomics, which is a specialized branch of metabolomics focusing on the comprehensive analysis of lipids. Lipidomics employs advanced methodologies like SFC [35] and [36], specifically designed to handle the complex nature of lipid molecules. The differentiation between polar and non‐polar metabolites is crucial because these two groups require distinct sample preparation, chromatographic, and detection methods. Lipidomics arose to address the limitations of conventional metabolomics techniques, which were not optimized for the comprehensive profiling of non‐polar metabolites, especially lipids.

2. Fast Sampling and Metabolism Quenching

In metabolomics, achieving sample representativeness requires that the sampling process effectively halts metabolism. Quenching is used to rapidly suppress endogenous enzymatic activity, thereby stopping metabolism and preventing alterations in the metabolic profile during sampling. This is especially critical when analyzing cells and tissues to capture intracellular (endometabolome) and extracellular (exometabolome) metabolic states. Effective quenching strategies must meet some criteria [37], including that metabolic deactivation must occur faster than the sample's metabolic alterations. The quenching process must be efficient because many primary metabolites have turnover rates around 1 mM/s. Metabolite concentrations vary widely, from just a few molecules per cell for signaling molecules to millimolar concentrations for primary metabolites like glucose [38, 39]. In addition, the sample's integrity, especially in cellular studies, should be preserved to prevent intracellular metabolite leakage. Further, the quenching process must not significantly alter the metabolites’ concentration or chemical and physical properties, and the quenched sample should allow for seamless continuation of subsequent analytical procedures.

Standard quenching methods involve rapid changes in the sample's pH or temperature. pH quenching involves moving to extreme pH levels, either with high alkali (e.g., KOH, NaOH) or strong acids (e.g., perchloric, hydrochloric, or trichloroacetic acid) [37]. Temperature quenching typically involves cooling the sample to temperatures below −20°C, provided the cold shock does not compromise sample integrity [40]. Cold methanol quenching is a widely used approach that rapidly halts metabolic activity within seconds and can distinguish between intracellular and extracellular metabolites if appropriately designed. Methanol is the preferred organic solvent for quenching due to its miscibility with water, low freezing point (lower than glycerine and ethanol), and low viscosity in methanol–water mixtures [41]. These properties make methanol superior to other organic solvents used in quenching procedures [42]. It is important to note that some cell types, such as bacteria, are susceptible to osmotic changes in their surrounding environment, leading to membrane rupture and altered concentrations of internal metabolites [42].

3. Sample Preparation and Extraction Methods

In metabolomics research, a diverse array of biological samples has been employed, including plasma [43], serum [44], urine [45], milk [46], cells [47], tissues [45], saliva [48], and cerebrospinal fluid [49]. Among these, serum and plasma are the most commonly utilized biofluids. It is important to note, however, that metabolite concentrations can vary between these two sample types [50]. Factors such as the use of anticoagulants, differences in centrifugation processes, and variations in sampling tubes contribute to these differences. Ultra‐high performance LC–MS (UHPLC–MS) studies have identified specific metabolite discrepancies between plasma and serum samples [51]. For instance, serum samples exhibit higher 1‐lysophosphatidylethanolamine and 1‐acyl‐sn‐glycero‐3‐phosphate (lysophosphatidic acid) concentrations than plasma. In addition, slight differences in amino acid profiles have been observed between the two. The choice of anticoagulant is critical in metabolomic studies as it can significantly impact the quality and accuracy of the data obtained. By carefully selecting and managing the anticoagulant, the precision and reliability of metabolomic analyses can be enhanced.

Several factors, such as temperature, pH, and exposure to light and air, can significantly influence the stability of metabolites in biological samples [52]. The type of storage container is also a critical consideration [53, 54, 55, 56, 57, 58]. Samples should be stored in sealed, chemically inert containers resistant to contamination and leaching. Although glass or plastic vials are commonly used, it is essential to ensure the material is compatible with the specific sample being stored [59, 60]. Guidelines from organizations like the MSI provide valuable recommendations for proper sample preparation in metabolomics studies [61]. These standards encompass protocols for sample collection, processing, and storage, along with quality assurance and control measures to ensure the accuracy and reproducibility of metabolomics data. Selecting an appropriate sample size is a critical aspect of the extraction process in metabolomics. The sample volume collected after extraction should exceed the minimum required for analysis. Generally, 1–100 mg of tissue and 106–250 µL of biofluids are considered adequate for extraction; however, these values may vary depending on the analytical instruments used. Therefore, determining the minimum sample volume before extraction is essential. The sample preparation process should be efficient and involve as few steps as possible, and in untargeted metabolomics, it should remain non‐selective. Because different sample types and analytical platforms are employed in metabolomics research, each requires specific sample preparation protocols tailored to its unique characteristics.

Direct sample introduction is not feasible in MS due to the sensitivity of the column and detector systems [62]. Proper sample preparation is crucial for ensuring the accurate detection, quantification, and characterization of analytes in MS [63]. Direct injection of complex biological samples, such as blood or tissue extracts, is typically impractical because these matrices contain a wide array of interfering substances, including salts, proteins, and lipids. These matrix components can cause issues such as ion suppression, where the signal of the target analyte is diminished due to competition with coeluting compounds, or ion enhancement, which artificially inflates the analyte signal, both of which compromise quantification accuracy [64]. In addition, injecting raw samples can lead to contamination of the chromatographic column or mass spectrometer ion source, reducing performance and requiring more frequent maintenance. The sensitivity of MS detectors also means that unprocessed samples can overload the system with unwanted compounds, masking the analyte of interest and introducing excessive noise [65]. Moreover, such interference affects reproducibility, introducing inconsistencies between sample runs. To address these challenges, sample preparation techniques are employed to remove matrix components, enrich the target analyte, and reduce matrix effects. These techniques help improve the signal‐to‐noise ratio, protect the instrument, and ensure reliable, reproducible data. Therefore, although MS is a highly sensitive technique, appropriate sample preparation is essential for ensuring its full potential and generating high‐quality, accurate results.

Samples must be processed to remove proteins, salts, and specific lipids [66, 67]. A common technique for sample preparation in untargeted metabolomics, particularly with LC–MS, is solvent‐based protein precipitation that is typically used to isolate proteins from biological matrices by using one or more solvents in a single‐step extraction process. However, the choice of extraction technique must be tailored to the specific analytes of interest, as not all methods are universally applicable. The solvent selection is critical, as it influences the range of metabolites extracted based on the solvent's polarity index. Solvents, like hexane, dichloromethane, acetone, methanol, acetonitrile, and water, are ranked by increasing polarity. Highly polar solvents can hinder the effective extraction of lipid‐derived metabolites, so the choice of solvent must align with the targeted metabolite profile [68]. Simple and rapid “dilute and shoot” approaches are frequently used in metabolomics, especially for urine and targeted metabolomics, as they require minimal sample preparation. This method involves diluting the urine with a solvent and directly injecting it into the analytical system to reduce matrix effects [63, 69]. It is most effective for metabolites in high concentrations with low variability, as dilution has minimal impact on the analytical signal. It is also beneficial when sample volumes are limited, or the sorbent used during sample preparation fails to retain the metabolites of interest [70]. It is also beneficial when sample volumes are limited or the sorbent used during sample preparation fails to retain the metabolites of interest [70]. However, caution is needed with this approach, as the complexity of the urine matrix and the broad range of metabolite concentrations can hinder the detection of low‐abundance compounds. In the context of syringe filters, the extraction process's success depends on the filter type used. In untargeted metabolomics, syringe filters (typically 0.22 µm nylon filters) are widely used to remove particulates and contaminants from biological samples such as blood, plasma, and urine, as these contaminants can interfere with chromatographic separation in LC–MS, lowering sensitivity and accuracy [27, 71, 72, 73].

Organic solvents can be used under different conditions to optimize the extraction of specific metabolites. These conditions can include variations in pH (e.g., primary or acidic environments) and temperature (high or low) to enhance extraction efficiency. Solvents may be used in their pure forms, such as methanol or ethanol, or as mixtures (e.g., methanol–chloroform) to improve the solubility and recovery of diverse metabolites based on their polarity and chemical characteristics [38, 74]. Although these methods can effectively remove specific compounds, alcohols, particularly methanol, are considered the most efficient extractants for polar and intermediate polar metabolites like amino acids, organic acids, and nucleotides [75]. Methanol is advantageous because it avoids introducing unwanted salts into MS analysis, precipitates proteins efficiently, penetrates cells effectively, evaporates quickly, and allows for rapid sample concentration.

Ideally, the extraction process should achieve the following objectives: (1) efficiently release metabolites from the sample; (2) remove interfering substances (e.g., salts and proteins) that can complicate the analysis; (3) ensure the extract is compatible with the analytical method; and, if necessary, (4) concentrate trace metabolites before analysis. Because metabolite profiling extraction is largely non‐selective, this stage focuses on eliminating macromolecules (such as proteins) and salts as potential interferents. The extraction process yields clean, concentrated extracts. Solid–liquid extraction techniques used with solid samples include methods such as microwave‐assisted extraction, ultrasound‐assisted extraction, supercritical fluid extraction, and Soxhlet extraction. The specific biological sample being analyzed primarily dictates the extraction technique. However, for liquid samples, the suitable extraction techniques are solid‐phase extraction (SPE), liquid–liquid extraction (LLE), and solid‐phase microextraction (SPME) to obtain clean samples [76]. Table 1 provides the extraction techniques for biological sample preparation and identification of the metabolite biomarkers using analytical techniques.

TABLE 1.

Extraction techniques for biological sample preparation and identification of the metabolite biomarkers using analytical techniques.

Biomarker	Biological matrix	Extraction technique	Sorbent material/solvents	Recovery (%)	Analytical technique	Sensitivity	References
Neuron‐specific enolase (NSE) and progastrin‐releasing peptide (ProGRP)	Serum	MISPE	NLLGLIEAK and ELPLYR	ProGRP—72 and NSE—93	LC–MS	0.1–0.3 nM	[77]
8‐OHdG and 8‐oxoG	polystyrene/polypyrrole (PS/PPY) electronspun nanofibers	PFSPE	polystyrene/polypyrrole (PS/PPY) electronspun nanofibers	88.8–104.9	HPLC	0.195 ng/mL	[78]
Volatile organic metabolites (VOMs)	URINE	Solid‐phase microextraction (SPME)	CAR/PDMS	NA	GC–MS	NA	[79]
3‐Nitrotyrosine (3‐NT)	Plasma	MCX 96‐well SPE plate	Mixed mode cationic exchange polymer	97	LC–MS/MS	5 pg/mL	[80]
Huntington biomarkers	Plasma	SPE–CE–MS	C18 bonded phase	>70%	SPE–CE–MS	<10 ng/mL	[81]
VOC markers	Breath sample	Solid‐phase extraction	DNPH‐coated silica gel	NA	LC–MS/MS	0.25–50 µg/L	[82]
Sarcosine, alanine, leucine, and proline	Urine	DDLLME	Acetonitrile, pyridine, and carbon tetrachloride	93.8%–106%	GC–MS and LC–MS	0.05–0.1 ng/mL	[83]
Human plasma metabolites	Plasma	LLE	Methanol, chloroform, chloroform	NA	LC/MS	NA	[84]
Phosphatidylethanol (PEth)	Blood	Ultrasound‐assisted dispersive liquid–liquid microextraction	Acetonitrile‐dichloromethane mixture (2.5:1, v/v)	99.87%	LC–MS/MS	1.04–1.87	[85]
Multi‐mycotoxin	Urine	SALLE	Ethyl acetate/FA (99/1, v/v)	70–108	LC–MS/MS	0.07–3.3 ng/mL	[86]
Glyoxal (GO) and methylglyoxal (MGO) are oxoaldehydes	Urine	SALLE and DLLME	Acetonitrile and carbon tetrachloride	GO—95.5–98.7 MGO—101–107	GC–MS	0.12 and 0.06 ng/mL	[87]

Abbreviations: CAR/PDMS, carboxen/polydimethylsiloxane; DDLLME, dispersive derivatization liquid–liquid micrextraction; DLLME, dispersive liquid–liquid microextraction; DNPH, 2,4‐Dinitrophenylhydrazine; DNPH, 2,4‐dinitrophenylhydrazine; FA, formic acid; GC–MS, gas chromatography–mass spectrometry; HPLC, high performance liquid chromatography; LC–MS/MS, liquid chromatography–tandem mass spectrometry; LC–MS, liquid chromatography–mass spectrometry; LLE, liquid–liquid extraction; MCX, mixed‐mode cation exchange; MISPE, molecularly imprinted solid‐phase extraction; NA, not available; PFSPE, packed‐fiber solid‐phase extraction; SALLE, salting‐out assisted liquid–liquid extraction; SPE, Solid phase extraction; SPE–CE–MS, online solid‐phase extraction coupled to capillary electrophoresis–mass spectrometry; VOC, volatile organic compounds.

Because most biological fluids are complex, suitable sample preparation is typically required for analysis via MS, LC–MS, or CE–MS [88]. Biological fluids such as urine, microdialysates, and digestive fluids often require basic processing steps like dilution, buffering, centrifugation, and evaporation [89, 90]. Although direct analysis can minimize metabolite loss, nonvolatile residues, such as salts, can negatively affect instrument performance, leading to adduct formation and ion suppression. To address these challenges, a practical extraction step, such as LLE [91] or SPE [92], can clean and desalt the sample, improving analysis quality. Direct sample analysis is often preferred for NMR analysis, as it requires minimal preparation, such as buffering with deuterated solutions. However, when analyzing complex fluids like serum or plasma, proteins can interfere with mass spectra, making detecting low molecular weight metabolites difficult, even though many of these metabolites exist only in trace amounts.

Sample preparation is a critical step in metabolomics workflows, significantly influencing the accuracy and reproducibility of analytical outcomes. SPE is a widely used technique for targeted metabolomics due to its effectiveness in sample clean‐up and preconcentration of specific metabolite classes. However, its application in untargeted metabolomics remains less common, primarily because of the potential biases introduced by selective enrichment. Despite this limitation, recent advancements have demonstrated that SPE can also be effectively adapted for untargeted metabolomics, particularly for the enrichment of polar, neutral, and ionic compounds. Cerrato et al. employed a novel SPE method to analyze zwitterionic and positively charged compounds in urine for untargeted metabolomics studies, specifically in prostate cancer research. This approach enhanced metabolite recovery while maintaining analytical reproducibility, thereby showcasing the feasibility of SPE in untargeted workflows [93]. Similarly, another study demonstrated that SPE combined with nano‐LC and nanoelectrospray ionization MS (nano‐LC–nano‐ESI–MS) can significantly improve the global analysis of urine metabolites. This integrated approach provides better sensitivity and resolution, facilitating the comprehensive analysis of metabolites in complex biological matrices [94]. Another recent innovation in sample preparation is the use of chemoselective enrichment tags. The authors developed a tagging strategy that enhances the resolution of highly polar metabolites, which are notoriously difficult to retain using conventional RP chromatography. The metabolites were tagged with a hydrophobic p‐Cl‐phenylalanine residue, improving their retention and resolution in chromatographic separations. This strategy allowed for the minimization of overlap in tandem MS (MS/MS) profiles, facilitating the identification and structure determination of polar metabolites. Moreover, the chlorine atom in the enrichment tag enabled the differentiation of tagged metabolites from background noise, which is particularly useful for profiling complex biological samples. This method has shown great potential in enriching and profiling challenging fractions of the metabolome, especially in untargeted studies [95]. These studies highlight the growing applicability of enrichment techniques, including SPE, in expanding the scope of untargeted metabolomics, allowing researchers to achieve improved coverage of complex metabolomes while retaining a broad analytical perspective. Although SPE remains more common in targeted metabolomics, these recent innovations demonstrate its utility in untargeted approaches when carefully optimized. Such advancements emphasize the importance of tailoring sample preparation methods to suit the goals of untargeted metabolomics while minimizing potential biases.

Advanced supercritical fluids, microwaves, or ultrasound methods can significantly speed up the extraction process, which may take minutes to hours, depending on the sample's characteristics. Focused microwave‐assisted Soxhlet extraction (FMASE) works with the same principle that involved in the traditional Soxhlet extraction technique, but additionally it involves auxiliary energy in the form of microwaves and has proven particularly effective for accelerated extraction of non‐polar and weakly polar metabolites from biological solids [96]. This method enables quantitative extraction of target metabolites within minutes, compared to the hours required by conventional techniques. Given the complexity of solid sample preparation, processes must be carefully optimized to avoid altering or degrading metabolites due to enzymatic activity or harsh extraction conditions. The optimal extraction conditions depend on the specific goals of the analysis.

4. Advancements in the Chromatographic Separation of Metabolites

4.1. Hydrophilic Interaction Chromatography

Although RP‐LC has been widely used in metabolomics due to the diverse range of available column chemistries, ease of operation, and consistent retention times, it is often ineffective for polar or ionizable metabolites. These compounds, including amino acids, small organic acids, nucleosides, phosphate derivatives, and saccharides, exhibit poor retention with RP‐LC. These polar metabolites are critical in various physiological and pathological processes, necessitating alternative analytical methods. HILIC, introduced by Alpert in 1990, is particularly effective for separating polar compounds [97]. In this technique, retention occurs through a multimodal separation process involving a polar stationary phase and a hydrophobic mobile phase, typically consisting of an aqueous‐organic mixture with a high organic content, typically acetonitrile. When the mobile phase contains 5%–40% water in acetonitrile, a water‐rich layer forms on the stationary phase's surface, allowing for analyte partitioning between this layer and the bulk mobile phase. The precise mechanisms of retention and separation are not entirely understood, but they are primarily attributed to hydrophilic partitioning, dipole–dipole interactions, hydrogen bonding, and electrostatic interactions, depending on the specific chemistry of the stationary phase [30].

Normal phase LC, although effective in separating highly polar analytes, often requires non‐aqueous mobile phases, making it less versatile and less compatible with biological sample matrices. In contrast, HILIC operates efficiently with aqueous‐organic mobile phases, enhancing its applicability for biological studies. Another emerging alternative is ion chromatography (IC), which offers exceptional separation for ionic metabolites but requires specialized instrumentation and suppressor systems, potentially limiting its widespread adoption in metabolomics workflows [98]. Ongoing innovations in HILIC technology have further strengthened its position as a pivotal technique in polar metabolomics. The development of hybrid stationary phases, which combine silica‐based substrates with polymeric or zwitterionic functionalities, addresses several traditional limitations of HILIC, including poor retention of weakly polar compounds and peak tailing. For instance, zwitterionic phases have demonstrated improved selectivity and reduced matrix effects when analyzing complex biological samples [99]. In addition, the emergence of sub‐2 µm and core–shell particles in stationary phases has significantly enhanced chromatographic resolution and sensitivity, making HILIC highly suitable for untargeted metabolomics [100].

Prinsen et al. developed an HILIC–MS method for the rapid quantification of underivatized amino acids in plasma. This method achieved excellent separation of 36 amino acids within an 18‐min run time, providing baseline separation for isomeric amino acids such as leucine and isoleucine. The study underscores HILIC's capability in efficiently resolving polar metabolites in complex biological matrices [101]. Another comparative study assessed various chromatographic techniques for carbohydrate analysis and found that although RP‐LC–MS with derivatization provided the highest sensitivity and repeatability, HILIC–MS enabled the analysis of underivatized carbohydrates, simplifying sample preparation. This underscores HILIC's capability in efficiently resolving polar metabolites in complex biological matrices [102]. These examples substantiate the claims that ongoing advancements in HILIC are not only improving analytical performance but are also broadening its utility in metabolomics. By integrating such innovations and critical comparisons, HILIC continues to evolve as a robust technique, addressing the challenges posed by alternative high‐polarity separation methods and expanding its applicability in metabolomics.

One of the longstanding challenges in metabolomics is the dynamic range of detection, where highly abundant metabolites can obscure the detection of low‐abundance metabolites. This issue can lead to incomplete profiling of the metabolome, especially for metabolites involved in disease processes or those present in trace amounts. Recent advancements, such as the use of HILIC coupled with HRMS, have significantly improved the dynamic range for lipidomic and polar metabolite analysis. Zhang et al. developed an HILIC–MS/MS‐based lipidomics platform that quantified 608 lipid species across 19 subclasses in human plasma, improving the dynamic range of lipidomic profiling. The integration of multi‐internal standards for each lipid class and post hoc correction techniques allowed for accurate quantification, making it possible to detect low‐abundance lipids even in complex plasma samples. This approach was applied to study the lipid profile of patients with COVID‐19, revealing differential lipid features associated with disease severity, demonstrating the power of this method in clinical research [103]. Similarly, another study highlights the utility of HILIC for improving lipidomics analysis. HILIC's ability to separate lipids based on their polarity allows for better retention and resolution of metabolites, thereby reducing the impact of dynamic range limitations [104]. This advancement ensures more consistent detection of both low‐ and high‐abundance metabolites, leading to more comprehensive lipidomic analyses in biological systems.

A wide range of phase chemistries for HILIC is now commercially available, utilizing silica or polymer‐based materials functionalized with polar groups such as aminopropyl, amine, amide, diol, triazole, sulfobetaine, phosphorylcholine, hydroxyethyl, and sulfoethyl [29]. Unlike RP‐LC, where analyte retention can be predicted to aid method development, retention in HILIC is more complex and less predictable. The stationary phase chemistry and the composition of the mobile phase highly influence chromatographic selectivity. Therefore, it is recommended that extensive screening of various conditions using a broad set of representative metabolites be conducted during method development to ensure comprehensive metabolite coverage. Modern computer‐assisted method development tools, such as the predictive elution window shifting and stretching (PEWS²) approach, can help accelerate this process [105]. This PEWS² strategy tolerates retention errors ranging from 5% to 10%. Even with moderately accurate predictions for the initial and final peaks, this method can generate a variety of distinct gradient conditions, enabling the first and last peaks to be shifted and stretched differently along the time axis. Comparative studies on different stationary phases for metabolomics suggest that diol, amide, and zwitterionic phases generally offer the best metabolite coverage, making them suitable starting points for method development [106, 107]. Compounds like small organic acids, sugar phosphates, and nucleosides, which are challenging to analyze using RP‐LC, show better retention in HILIC mode, particularly with polymeric zwitterionic phases that can operate at higher pH (9–10) due to the polymeric nature of the stationary phase [108]. In addition, incorporating phosphate at micromolar concentrations into the mobile phase has enhanced peak shape and sensitivity for these metabolites when analyzed with a zwitterionic phase [109].

In addition to the stationary phase chemistry, the mobile‐phase composition plays a critical role in determining the selectivity and quality of separation in HILIC. Acetonitrile is the preferred organic solvent due to its water solubility and aprotic nature, which makes it ideal for HILIC. On the other hand, protic solvents, like methanol, isopropanol, and ethanol, are not recommended as they compete with water for solvation of the stationary phase, potentially reducing analyte retention. The high proportion of acetonitrile in the mobile phase lowers its viscosity compared to the mixtures used in RP‐LC, allowing for longer columns (which enhance separation efficiency), increased electrospray ionization (ESI) sensitivity, and improved volatility. Buffer composition, specifically salt concentration and pH, significantly impacts both selectivity and retention time repeatability. The concentration of buffer (typically ≤50 mM to prevent salt precipitation in acetonitrile) affects the thickness of the water layer on the stationary phase, thus influencing hydrophilic and electrostatic interactions. Ammonium formate and ammonium acetate buffers are commonly used because they are compatible with MS and provide better peak shapes than their corresponding acid forms. However, due to its strong ion suppression effects, trifluoroacetic acid is not recommended for HILIC–MS. Lastly, maintaining a consistent and appropriate buffer pH is essential for reproducibility in HILIC. Variations in buffer pH can lead to increased retention variability, highlighting the importance of precise and repeatable buffer preparation.

HILIC chromatography, on the other hand, avoids isobaric overlaps among different classes as it enables lipid class separation by hydrophilic interaction. Lipid classes share a common hydrophilic headgroup so that hydrophobic differences of lipids within one class can be neglected in HILIC chromatography. Hence, class‐specific coelution of all lipids occurs in HILIC and was successfully applied for polar lipid analysis, for example, glycerophospholipids, glycosphingolipids, and phosphosphingolipids [110, 111, 112, 113]. Polar metabolites are separated using RP‐LC or HILIC. However, HILIC is more efficient for very polar metabolites, whereas RP‐LC is for less polar metabolites. For instance, carnitine (XlogP = −4.9) can be detected using both RP‐LC and HILIC. However, in RP‐LC, the elution is close to a void volume where ion suppression can be expected (elution of other very polar metabolites, salts), whereas HILIC provides better retention and separation selectivity [114]. Mobile phases containing water, acetonitrile, or methanol are used to analyze polar metabolites. In contrast, for RP‐LC‐based lipidomics, stronger mobile phases with a high percentage of isopropanol are needed. The separation time usually ranges from 10 to 30 min.

4.2. Monolithic Columns

In metabolomics research, distinguishing and identifying compounds with similar physical properties can be challenging. Unger et al. explored advancements in packed and monolithic columns for HPLC [115, 116]. Significant efforts have been directed at enhancing column efficiency by extending column lengths, reducing particle sizes, and replacing packed beds with monolithic structures. For instance, reducing particle size to less than 2 µm increases column back pressure, shortens analysis time, and improves separation efficiency. This development led to the creation of specialized equipment capable of handling higher pressures than those used in conventional HPLC, culminating in the development of UHPLC. However, higher pressure drops in UHPLC can result in viscous heating effects, negatively impacting distribution equilibria, mass transfer kinetics, and solvent solubility or miscibility [116]. To provide fast and efficient separations without these pressure limitations, monolithic columns have been developed, allowing the use of standard HPLC equipment.

Monolithic columns have been designed to enhance mass transfer kinetics, reduce column back pressure, and achieve faster separation speeds, which eliminate the need for column frits. The primary advantage of silica monolithic columns is their low column pressure drop. Monolithic silica columns feature continuous porous beds with bimodal structures composed of large through pores and small mesopores, especially characteristic of silica‐based monoliths [117, 118, 119]. These have been widely used for separating small molecules. However, the performance of polymeric monoliths in small molecule separations has been limited due to their low surface area and swelling in LC mobile phases. Recent advances, however, have demonstrated the improved effectiveness of polymer‐based monoliths in separating small molecules, as observed in metabolomics studies [120, 121]. Svec [122] and Nischang et al. [123] examined methods to enhance the performance of polymeric monoliths, focusing on factors such as porogen selection, polymerization time, crosslinking, hyper crosslinking, and incorporating carbon nanostructures to increase surface area. Urban et al. studied the retention behaviors of large and small molecules in organic polymeric monoliths using gradient and isocratic elution [124]. Additional strategies based on a systematic investigation of factors such as porogen selection, polymerization time, degree of crosslinking, and hyper crosslinking [125], as well as the incorporation of carbon nanostructures [126, 127], have been employed to enhance the surface area of polymeric monoliths. Besides, capillary columns are also shown to provide the best separation for small molecules, as radial diffusion contributes less to peak broadening than standard 4.6 mm internal diameter column. Radial diffusion, unlike other forms of diffusion, positively impacts chromatographic separation by aiding equilibration between phases, thereby enhancing peak sharpness and resolution.

4.3. Two‐Dimensional (2D) Chromatographic Separations

An effective strategy to enhance the metabolome coverage of highly complex samples or closely related metabolites is introducing an additional separation dimension, improving selectivity. Similar to SFC advancements, online 2D‐LC has experienced substantial progress in recent years, thanks to theoretical developments and improved instrumentation [128]. In online 2D‐LC, 2 distinct LC separations are integrated, typically using a 4‐port duo valve, a 10‐port valve with 2 sampling loops, or specialized valves equipped with multiple sample parking loops [129]. Comprehensive 2D‐LC (LC × LC) involves transferring all fractions from the first dimension into the second dimension for further analysis [128, 130]. In contrast, multiple heart‐cutting 2D‐LC targets specific fractions from the first dimension, which are then subjected to HR analysis in the second dimension. Selective comprehensive 2D‐LC (sLC × LC) represents an intermediate approach, where specific regions of the first‐dimension chromatogram are comprehensively transferred as consecutive fractions and further analyzed in the second dimension, providing enhanced selectivity and resolution for complex samples. This approach has proven particularly useful in metabolomics, as it allows targeted analysis of specific metabolite classes while maintaining high coverage of the metabolome [128, 130, 131].

The comprehensive LC × LC method is particularly advantageous in untargeted metabolomics, which enables profiling hundreds of features within a single analysis. A wide range of chromatographic modes can be combined in 2D‐LC, including RP‐LC, HILIC, normal‐phase LC, ion‐exchange chromatography (IEX), ion‐pairing chromatography, and porous graphitized carbon (PGC) columns. The goal is to achieve the highest orthogonality in separating the 2D by utilizing different stationary phases and mobile‐phase compositions. The choice of separation dimensions depends on several factors: the analytes of interest, the miscibility and compatibility of mobile‐phase solvents, compatibility with the detection system, and the use of a faster technique (such as UHPLC) for the second dimension [29, 130].

Recent technological advancements in 2D‐LC instruments have greatly expanded the number of experimental parameters that can be optimized during method development. This includes column dimensions, stationary phase chemistry, particle sizes, mobile‐phase composition, gradient settings, sample loop volumes, injection volumes, flow rates, and modulation times. Although optimizing these parameters can be complex and time‐consuming, modern 2D‐LC systems provide greater flexibility and efficiency. Setting up a complete 2D‐LC method may require significant effort, especially with the need for dedicated instruments (although existing one‐dimensional LC systems can often be upgraded to 2D‐LC with minimal investment). Furthermore, coupling 2D‐LC with MS presents additional challenges, as introducing a second separation dimension can lead to sample dilution, potentially impacting the sensitivity of MS detection [130].

The indecision to adopt 2D‐LC in metabolomics, despite its potential to enhance metabolome coverage significantly, can be attributed to challenges similar to those encountered with techniques like HILIC and SFC. Inexperienced users often face difficulties obtaining reproducible results due to insufficient theoretical and practical knowledge to exploit these techniques fully. Recent advancements in instrumentation have addressed some significant concerns, such as issues with MS detector sensitivity and the incompatibility of certain mobile phases, through active solvent modulation techniques. In addition, the availability of software tools to aid in method development, along with continuous improvements in algorithms for processing 2D chromatograms, is expected to promote wider use of 2D‐LC in clinical metabolomics. Most applications have focused on heart‐cutting approaches and proof‐of‐concept studies rather than being applied in clinical settings. However, the results from these studies underscore the potential of 2D‐LC in metabolomics [132, 133, 134, 135, 136]. For instance, a twofold increase in the detection of intracellular energy metabolites was demonstrated by combining RP‐LC with PGC [137], and simultaneous metabolomic and lipidomic data acquisition was achieved in a single analysis using 2D‐LC [134]. These findings highlight the significant promise of 2D‐LC in expanding analytical coverage in metabolomic research.

5. Advanced MS‐Based Analytical Techniques in Metabolomics

Metabolomics primarily focuses on the detection, identification, and (relative or absolute) quantitative analysis of metabolites. MS is one of the leading analytical techniques used in this field [138]. Meanwhile, MS has received immense attraction in metabolomics studies due to its ability to detect a wide range of metabolites and analyze complex biofluid samples from biomarker identification to structural elucidation [139]. MS is highly effective in accurate mass determination and sensitive quantification and provides valuable structural insights through fragmentation patterns, particularly for smaller molecules. Although NMR remains the preferred method for comprehensive structural elucidation due to its detailed connectivity information, MS offers complementary capabilities that enhance metabolite profiling, especially in complex samples where high sensitivity is essential [140]. In addition, MS instruments are cost‐efficient while providing high levels of sensitivity and resolution, making MS the most widely adopted technique in metabolomics. MS offers distinct advantages in sensitivity, selectivity, and multiplexing capabilities, which can lead to long‐term cost efficiencies, especially in high‐throughput settings or when analyzing complex matrices that require extensive separation and quantitation [141]. In addition, advancements in MS technology, such as the development of HR mass spectrometers (e.g., Orbitrap and QTOF), improvements in ionization techniques (e.g., nanoelectrospray and desorption ESI [DESI]), and innovations in automated sample preparation systems have contributed to reducing operational costs through enhanced automation, reduced reagent consumption, and lower labor requirements [142]. Direct injection MS (DIMS technique) is a technique that directly introduces the sample into the MS without any chromatographic separation. This technique helps metabolomic fingerprinting, whereas factors such as low ionization efficiency, complexity, and ion suppression limit this technique for biomarker discovery [143]. Integrating chromatographic separation with mass spectrometric detection lowers detection limits compared to other methods, such as UV detection, and simplifies the interpretation of analytical results. This combination has become a key technology for multi‐parallel analysis of low molecular weight compounds in biological fluids and tissues. When coupled with LC, MS enables sensitive, reproducible, and accurate analysis of numerous metabolites in metabolite profiling studies [144, 145, 146, 147, 148]. MS detection provides molecular weight and structural data and offers an additional dimension of separation selectivity, particularly for coeluting molecules with different nominal masses [149]. Typically, MS is coupled with separation technologies, such as CE, GC, and LC, which facilitate enhanced metabolite identification and quantification by simplifying complex biological samples. Recent advances have significantly improved chromatographic separation efficiency, mass spectral resolution, detector sensitivity, data acquisition capabilities, and the integration interfaces between separation systems and MS platforms [150, 151]. Recent advancements have led to the prominence of atmospheric‐pressure ionization (API) and matrix‐assisted laser desorption ionization (MALDI) in MS‐based metabolite profiling. Among the API techniques, ESI is the most commonly employed in metabolite profiling, generating ions by transitioning them from the liquid to the gas phase. Besides, MALDI is advantageous in imaging MS [152]. Table 2 provides the MS‐based metabolite profiling studies in biomarker identification.

TABLE 2.

Biomarker identification by mass spectrometry‐based analytical techniques.

Biomarker	Disease	Biological matrix	Analytical technique	Chromatographic separation	Ionization source	Analyzer	Observation	References
N‐acetyl‐d‐tryptophan, 2‐arachidonoylglycerol, pipecolic acid and oxo‐glutaric acid, N‐acetyl‐d‐tryptophan and 2‐arachidonoylglycerol	Breast cancer	Serum	LC–MS/MS	HSS T3 column (2.1 mm × 150 mm × 1.8 µm) 0.1% formic acid in water (v/v) and (B2) 0.1% formic acid in acetonitrile (v/v) as mobile phase	ESI	Quadrupole‐orbitrap	4‐Metabolite panel and the 2‐integrated biomarker panel both provided better diagnostic value than either one alone	[153]
Palmitic acid, oleic acid, cis‐8,11,14‐eicosatrienoic acid, docosanoic acid, and the ratio of oleic acid to stearic acid	Breast cancer	Serum	GC–MS	HP‐5MS (5% phenyl methyl silox) capillary column (30 m × 0.25 mm i.d., film thickness 0.25 µm) Helium carrier gas	EI	Quadrupole	FFAs (C16:0, C22:0, C24:0) in serum were significantly higher in breast cancer patients	[154]
Cytidine triphosphate, 11‐ketoetiocholanolone, saccharopine, nervonic acid, and erucic acid	Diabetic cardiomyopathy	Plasma	LC–MS	ACQUITY UPLC BEH C18 column 0.1% formic acid in water (phase A) and ACN with 0.1% formic acid (phase B)	ESI	Q–TOF	Higher levels of N8‐acetylspermidine, 3‐hydroxytetradecanedioic acid, 17α‐ethynyl‐estradiol, phytanic acid, methyl‐guanosine, 11‐β‐hydroxyandrosterone‐3‐glucuronide, 19,20‐DiHDPA, thromboxane B3, and d‐CTP in both the DCM and T2DM groups	[155]
Isopropanol and 2,3,4‐trimethylhexane, 2,6,8‐trimethyldecane, tridecane, and undecane	Type 2 diabetes mellitus	Exhaled breath	GC–MS	60 m (0.32 mm ID, 1.8 µm phase thickness) DB‐624 polysiloxane capillary column Helium carrier gas	EI	Ion trap	The experimental data indicated that T2DM patients had higher levels of 2,3,4‐trimethylhexane, 2,6,8‐trimethyldecane, tridecane, and undecane than healthy controls	[156]
Phenylalanine, vanillactic acid, 3b–hydroxy‐5‐cholenoic acid, glycoursodeoxycholic acid, lysopc(18:2), PA(18:2/15:0), valeric acid, 2‐octenoic acid, docosene, carnitine, 2‐methylbutyroylcarnitine, 4‐hydroxybenzaldehyde, adrenochrome, leukotriene B3, cytidine 2′,3′‐cyclic phosphate, 3‐methylene–indolenine, heptanoylcholine	Parkinson's disease	Plasma	LC–MS	C18 column (3.0, i.d. × 150 mm, 1.8 µm) Water (solvent A, modified by the addition of 0.5% acetic acid) and ACN (solvent B)	ESI	Q–TOF–MS analyzer	First, bile acid dysregulation may directly cause lipid metabolism dysfunction. Second, perturbations of carnitine metabolism may directly lead to lower LCFA levels. Third, the abnormal metabolite levels found in this study may collectively disrupt energy production	[157]
Desmosterol	Alzheimer's disease	Plasma	LC–MS	XR‐ODS column Solvent A (water‐methanol, 50:50) and solvent B (methanol)	APCI Atmospheric‐pressure chemical ionisation	LTQ Orbitrap mass spectrometer	Analyzed the concentration of desmosterol in human AD and elderly controls and have shown that desmosterol plasma level and the desmosterol/cholesterol ratio in the same patients were significantly decreased in AD	[158]
Tetra hydrocortisone, cortolone, urothion, and 20‐oxo‐leukotriene E4	Medulloblastoma	Urine	LC–MS	HSS C18 column (3.0 mm × 100 mm, 1.7 µm) Mobile phase was 0.1%formic acidinH2O and acetonitrile. At a flow rate of 0.3 mL/min. Flow rate of 0.3 mL/min	ESI	LTQ Orbitrap Velos Pro mass spectrometer	Tetrahydrocortisone and cortolone were upregulated in MB patients. Metabolites involved in leukotriene B4 metabolism showed higher levels in MB	[159]
Perillic aldehyde and octa‐decanal	Parkinson's	Sebum	GC–MS	HP‐5MS Ultra Inert 30 m × 0.25 mm × 0.25 µm column. The column flow was kept at 1 mL/min. The oven ramp was programmed as follows: 40°C held for 5 min, 10°C/min–170°C, 8°C/min–250°C, 10°C/min–260°C held for 2 min for a total run time of 31 min. The transfer line to the MS was kept at 300°C. Dry nitrogen was used as the purge gas	EI	Quadrupole	Identified a distinct volatiles‐associated signature of PD, including altered levels of perillic aldehyde and eicosane, the smell of which was then described as being highly similar to the scent of PD by our “Super Smeller”	[160]
Methionine, lysine, glycine, phenylalanine, citrulline, 3‐methyladenine, histidine, isoleucine, serine, proline, tryptophan, alanine, arginine, valine, and asparagine	Epileptic seizures	Plasma	CE–MS	Fused‐silica capillary	ESI	TOF–MS	Study emphasized the suitability of CE–MS to analyze volume‐restricted plasma samples with an acceptable variation. Additionally, the method is sensitive enough to measure a suitable profile of polar components, differentiating the samples before and after an evoked seizure, and defining some amino acids involved in evoked seizures	[161]
Salivary and urinary polyamines	Colorectal cancer	Saliva	CE–MS/MS	Fused‐silica capillaries 1 M formic acid solution	ESI	Triple‐quadrupole MS/MS system, TOFMS	This method is sensitive, selective, and quantitative, and its utility was demonstrated by screening polyamines in 359 salivary samples within 360 min, resulting in discrimination of colorectal cancer patients from noncancer controls	[162]
Malonate, formate, N‐methylnicotinamide, m‐hydrox‐yphenylacetate, and alanine	Depression	Plasma	LC–MS/MS	C 8 BEH column (1.7 µm, 2.1 mm × 100 mm) and a T3 HSS column (1.8 µm, 2.1 mm × 100 mm) 0.1% Formic acid in water, 0.1%formic acid in acetonitrile were used in positive ionization mode, while (C) 6.5 mM NH4 HCO3 in water and (D) 6.5 mM NH4 HCO3 in 95% methanol/water were used in negative ionization mode	ESI	Triple TOF	A combinational biomarker—carnitine C10:1, PE‐O 36:5, LPE 18:1 sn‐2, and tryptophan—was discovered and validated by means of binary logistic regression model with high sensitivity and specificity for both moderate and severe MDD diagnosis	[163]
Dihydroceramide, ceramide, PC(16:1(9Z)/14:1(9Z)), lysopc (22:5), thromboxane A2, nicotinamide riboside, 5‐HTP	Hypertension	Serum	HPLC–MS	Halo‐C18 column (2.1 mm × 100 mm, 2.7 µm, America Advanced Material Technology Cor.) with a binary solvent system (solvent A: water with 0.1% formic acid; solvent B: acetonitrile with 0.1% formic acid)	NA	HPLC–TOF	7 potential biomarkers were identified. By comparison with the normal group, 5 metabolites (dihydroceramide, ceramide, LysoPC (22:5), thromboxane A2, 5‐HTP) were increased in the model group, and the 2 metabolites (PC(16:1(9Z)/14:1(9Z)), nicotinamide riboside) were decreased	[164]
Ethanolamine, azelaic acid, histidine, threitol, 2,4‐dihydroxypyrimidine, levulinic acid, glyceric acid, methylmalonic acid, hippuric acid, pyruvic acid, acetic acid, sucrose, threitol, aminomalonic acid	Depression	Urine	GC–MS	Used the HP‐5 MS fused silica capillary column (30 m × 0.25 mm × 0.25 µm, Agilent, USA) to do separation. The helium carrier gas was set a flow rate of 1 mL/min	EI	Quadrupole	Significant increase in levels of azelaic acid, histidine, threitol, and levulinic acid in dHB	[165]
Arachidic acid, glycerol, urea	Asthma	Serum	GC–MS	Rtx‐5MS capillary column (0.25 mm × 30 mm × 0.25 µm) in a split mode (1:30). Helium was used as the carrier gas, and flow rate was kept constant at 1.0 mL/min	EI	Quadrupole	Saturated fatty acids were identifiedas main metabolites associated with asthma	[166]
Hypoxanthine, p‐chlorophenyl‐alanine, l‐glutamine, glycerophosphocholine, inosine, hypoxanthine, succinate, xanthine, arachidonic acid (peroxide free), l‐pyroglutamic acid, indoxyl sulfate, theophylline, l‐valine, l‐norleucine, bilirubin, l‐leucine, inosine, palmitic acid, l‐phenylalanine	Asthma	Serum	LC–MS	ACQUITY UPLC BEH amide (2.1 mm × 100 mm, 1.7 µm, Waters, USA) as the chromatographic column. Binary mobile‐phase system (phase A: water containing 25 mmol/L ammonium acetate and 25 mmol/L ammonium hydroxide; phase B: acetonitrile)	ESI	Triple quadrupole time‐of‐flight	In the ESI+ mode, the levels of hypoxanthine, l‐pipecolic acid, p‐chlorophenylalanine, and acetyl‐carnitine were significantly higher in the asthmatic patients, whereas the levels of alpha‐N‐phenylacetyl‐l‐glutamine, 1‐methyladenosine, glycochenodeoxycholate, l‐citrulline, and l‐glutamine were significantly decreased in asthmatic patients compared with those of the COPD patients. In the ESI mode, the levels of linoleic acid and hypoxanthine were significantly higher in asthmatic patients, whereas the levels of pseudo‐uridine, alpha N‐phenylacetyl‐l‐glutamine, succinate, l‐citrulline, and glycochenodeoxycholate were significantly decreased in asthmatic patients compared with those in the COPD patients	[167]

Abbreviations: ACN, acetonitrile; APCI, atmospheric‐pressure chemical ionisation; CE–MS/MS, capillary electrophoresis–mass spectrometry/mass spectrometry; CE–MS, capillary electrophoresis–mass spectrometry; COPD, chronic obstructive pulmonary disease; DCM, dilated cardiomyopathy; dCTP, deoxycytidine triphosphate; dHB, HBV‐infected patients with depression; EI, electron ionization; ESI, electron spray ionization; FFA, free fatty acids; GC–MS, gas chromatography–Mass spectrometry; HDPA, high‐density polyethylene; HPLC–MS, high performance liquid chromatography–mass spectrometry; HPLC–TOF, high performance liquid chromatography–time‐of‐flight; HSS, high strength silica; HTP, hydroxytryptophan; LCFA, long‐chain fatty acids; LC–MS/MS, liquid chromatography–mass spectrometry/mass spectrometry; LC–MS, liquid chromatography–Mass spectrometry; LPE, lyso‐phosphatidylethanolamines; LTQ, linear ion trap quadrupole; MB, medulloblastoma; NH₄ HCO₃, ammonium bicarbonate; ODS, octadecyl—silica; PD, Parkinson's disease; PE‐O, phosphatidyl‐ethanol‐amines oxidative metabolite; QToF MS, quadrupole time‐of‐flight mass spectrometry; QToF, quadrupole time‐of‐flight; T2DM, type 2 diabetes mellitus; TOF–MS, time of flight–mass spectrometry; UPLC BEH C18 column, ultra performance liquid chromatography ethylene bridged hybrid c18 column.

5.1. GC–MS

The aim of metabolomics analysis is to gain a comprehensive understanding of all metabolites in biological samples [168]. Numerous effective platforms have been established to support this objective. GC–MS‐based metabolomics typically demands high‐throughput capabilities to process substantial sample volumes and precise peak identification through reference retention times and mass spectra [169, 170]. GC–MS‐based metabolite profiling of biological fluids and tissues is rapidly becoming a cornerstone of functional genomics and systems biology due to its high chromatographic resolution, sensitivity, specificity, and reproducibility. GC–MS is a hyphenated technique that supports detecting low molecular weight volatile compounds and polar molecules, rendering them suitable for metabolomic analysis [171]. It involves separating volatile and thermostable compounds by GC, where eluted compounds’ ions are generated via electron impact (EI), where sample solvents are independently removed before introducing the vaporized analyte into the ionization source. At optimal ionization energy, the vaporized metabolites collide with a stable electron flow, resulting in ion formation [172]. Various mass analyzers, including single quadrupole, TOF, and orbitrap, are used for mass identification in GC–MS [138]. As one of the earliest and most popular techniques in metabolomics, GC–MS offers excellent separation performance, ease of use, and cost‐efficiency. It benefits from a comprehensive metabolite spectrum library, enabling rapid and accurate qualitative analysis of metabolites, with the ability to simultaneously analyze 100 of compounds [66]. However, to study polar and nonvolatile metabolites, such as organic acids, sugars, and amino acids, chemical derivatization is often required to enhance volatility and thermal stability before analysis [17, 173]. Even though derivatization shows high specificity sensitivity, providing efficient and reproducible analysis [174], it is time‐consuming, resulting in less stable products [29]. Sometimes, these derivatization steps can lead to changes in the studied metabolites.

Various derivatization agents have been explored for metabolite profiling, including diazomethane, N,O‐bis(trimethylsilyl)trifluoroacetamide (BSTFA), and N‐methyl‐N‐(trimethylsilyl)trifluoroacetamide (MSTFA) [175]. Among these, BSTFA and MSTFA are the most commonly employed agents, primarily for the silylation (trimethylsilylation) of organic acids [176, 177, 178, 179]. However, silylation can induce conversion reactions, such as the transformation of arginine into ornithine, when reacting with trimethylsilylating agents like MSTFA [180]. Other derivatizing agents, such as ethyl chloroformate, methyl chloroformate, and isobutyl chloroformate, have also been utilized and reported in metabolite profiling studies [139, 143, 181]. The use of ethyl chloroformate, remarkably, is increasing in urine sample metabolomics due to its reactivity in aqueous media, unlike trimethylsilylating agents [182]. Additionally, tert‐butyldimethylsilyl (TBDMS) derivatives are employed in metabolite profiling. Although TBDMS derivatives are less susceptible to hydrolytic degradation by moisture than tri‐methyl silylation derivatives [123], their use significantly increases the molecular mass when multiple derivative groups are involved [173].

GC–MS has been used for non‐targeted metabolomic analysis, particularly hydrophilic metabolite detection [169]. Employing GC–MS, Nam et al. identified nine altered metabolomic pathways from BC patients [174]. Among all altered pathways, four metabolite biomarkers (homo‐vanillate, 5‐hydroxyphenylacetate, 5‐hydroxyindoleacetate, and urea) differed in BC compared to healthy controls. In another study, GC–MS with quadrupole as a mass analyzer (GC–MS) was used to screen salivary volatiles for putative discrimination between healthy and breast cancer [183]. Volatile metabolomic signature of human BC cultured cell lines (T‐47D, MDA‐MB‐231, and MCF‐7) showed that 2‐pentanone, 2‐heptanone, 3‐methyl‐3‐buten‐1‐ol, ethyl acetate, 2‐methyl butanoate, and ethyl propanoate have variable concentrations compared to control cell lines, suggesting their utilization as biomarkers [184]. HRGC is the preferred method for highly volatile compounds, especially when standard GC–MS cannot improve conventional library matching [185]. HRGC excels in sensitivity and metabolic coverage, making it a superior choice for investigating volatile metabolites. 2D GC (GC × GC) coupled with MS is an emerging technique to increase the analytical performance of conventional GC as it offers superior sensitivity and high resolving power [186]. As a modern technique, GC × GC–TOF–MS detected organic aciduria biomarkers for six inherited metabolic disorders, making it superior to a conventional GC–Q–MS method [187].

5.2. LC–MS

Although both GC–MS and LC–MS are utilized in metabolomics, LC–MS overcomes some limitations inherent to GC–MS by enabling similar analysis and detection of a broader range of metabolites. Among all hyphenated techniques, LC–MS offers high versatility in separation procedures and a wide range of mass analyzers for analysis [188]. In contrast to GC–MS, it is particularly well suited for metabolites with moderate volatility and low thermal stability, making it an essential platform for metabolomics research [189, 190, 191]. RP separation techniques commonly used in LC–MS further enhance its compatibility with metabolomics samples. In addition, LC–MS allows for partial sample injection directly onto the chromatographic column without extensive pre‐treatment [192]. Highly polar compounds, such as sugars and amino acids, can be effectively analyzed using HILIC [93, 193, 194]. In contrast, mid‐ and low‐polarity compounds are separated using RP gradient elution [195].

This technique allows analysis in both positive and negative ion modes [196]. In ESI, ions are transferred from the liquid to the gas phase, making it particularly suitable for MS analysis in LC and automated sample handling. Although rapid polarity switching between positive and negative modes is a feature available in triple quadrupole (QqQ) mass spectrometers (used primarily for targeted analysis), recent advancements in MS technology have improved this functionality by enabling faster switching and reducing analysis time. These improvements, which have become more widely available in modern instruments, contribute to enhanced efficiency, minimizing the need for multiple analyses and ultimately reducing operational costs. Although atmospheric‐pressure chemical ionization (APCI) is less frequently used in metabolite profiling, it remains valuable for analyzing neutral molecules such as lipids.

In targeted metabolite profiling, the choice of mass analyzers can be limited by their mass accuracy and resolution. However, analyzers like QqQ and Q‐Trap mass spectrometers, operating in selective ion scanning modes (such as precursor ion scanning, neutral loss, and multiple reaction monitoring [MRM]), enable highly sensitive targeted metabolite profiling [197, 198, 199]. Given the low abundance of biomarkers in biological fluids, highly sensitive techniques are necessary for their detection. With its high throughput, sensitivity, and selectivity, MS/MS is particularly suited for detecting biomarkers in biological samples. HRMS, including FT‐ICR, Q–TOF, and orbitrap‐based instruments, offer high mass resolution [200] and exhibit a significant role in the qualitative analysis of biomarkers and distinguish isobaric molecular ions [192, 193, 194, 195, 196, 197, 198, 199, 200, 201]. For instance, the amino acids glutamine (MW 146.0689 Da) and lysine (MW 146.1052 Da) can be differentiated with a TOF detector [17]. HRMS overcomes the main problems faced with the quadrupole instrument by providing high spectrum potential and a large platform for quantification in a single MS experiment [201, 202, 203]. Willmann et al. analyzed and identified 58 endo‐metabolites and 92 annotated exo‐metabolites of BC cell lines and breast epithelial cell line MCF‐10A using UHPLC–ESI–QTOF and HPLC–ESI–QqQ [204]. LC–ESI–QTOF MS/MS was utilized to identify 1269 metabolites with different concentrations present in plasma from healthy control and breast cancer patients. Of 1269 metabolites, 35 supported cell growth by providing energy for synthesizing a biomolecule associated with breast cancer [205].

A lipidomic study to distinguish between human breast cancer and surrounding tissue was performed by HPLC–electrospray mass ionization (HPLC–ESI–MS). A significant increase in lipid concentration was observed for phosphatidylcholine, phosphatidylinositol, phosphatidylethanolamine, and lysophosphatidylcholine [206]. Continuous advancement in chromatography enables UHPLC combined with MS for high resolution and more metabolite identification. Using UHPLC MS/MS, nine novel blood plasma biomarkers, including adenine, arginine–aspartate, folic acid, guanosine, methyladenine, methylguanine, phenylacetylglycine, thiamine pyrophosphate, and uridine, were identified to discriminate between BRCA1 and non‐BRCA1‐like breast cancer profiles.

UHPLC and nano‐flow LC represent advanced generations of LC systems. Commercially available UHPLC systems include ultra‐performance LC, high‐speed LC, ultrafast LC, rapid resolution LC, fast LC, and rapid separation LC. These systems achieve improved chromatographic separation and sensitivity due to reduced particle size while offering faster run times and robust performance compared to traditional HPLC. These advancements have proven highly beneficial in metabolite profiling studies [202, 203, 204]. Nano‐flow LC systems, which operate at flow rates between 200 and 1000 nL/min, are primarily used in proteomics but have recently been applied to metabolomics when coupled with MS [205]. Nano‐flow LC–ESI–MS systems provide enhanced sensitivity, a broader dynamic range, reduced carryover, lower ion suppression, and shorter analysis times. These systems utilize miniaturized emitters (10–100 µm, i.d.) to produce smaller eluent droplets, which improve ionization efficiency and minimize analyte degradation in the MS source [206, 207, 208]. For highly polar metabolites that exhibit minimal retention or elute early in RP‐LC, HILIC is advantageous. Following initial RP‐LC profiling, re‐profiling with HILIC allows for comprehensive coverage of highly polar metabolites. In contrast to RP‐LC, HILIC employs a gradient elution with a high proportion of organic solvents in the initial chromatography conditions, making it suitable for profiling polar compounds from aqueous extracts [209].

5.3. CE–MS

CE–MS is a highly efficient and emerging technique for separating charged metabolites, delivering exceptional resolution of analytes and primarily targeting polar or ionic compounds found in biological fluids [210, 211]. CE–MS is an established alternative to LC–MS for targeted metabolite profiling, particularly for compounds like amino acids. It has been widely applied in biomarker discovery. CE–MS offers several advantages over GC–MS, including direct analysis without the need for sample derivatization and minimal sample preparation. It also provides a complementary method to LC–MS by enabling the simultaneous analysis of both charged and neutral molecules in a single run. CE–MS facilitates the comprehensive quantification of cationic and anionic metabolites and low molecular weight energy metabolites in cells, tissues, and biological fluids. However, CE–MS is less effective for separating neutral molecules and large macromolecules, such as sugars, long‐chain peptides, and steroidal hormones. Its strength lies in analyzing samples with limited volume, making it a preferred technique. Wild et al. performed an untargeted cationic and anionic metabolite profiling using CE–MS in human plasma and urine [212].

CE–MS combines electrophoresis with advanced micro‐column separation techniques, offering efficient and rapid compound separation [138]. Using tiny bore capillaries (20–200 µm), CE enables highly effective separation of both large and small molecules [213, 214]. High voltage is applied during separation to enhance the electroosmotic and electrophoretic flow of ionic species and buffered solutions through the capillary [215, 216]. The high resolution, versatility, and sensitivity of CE–MS make it particularly suitable for metabolite analysis. CE offers sharp peaks and high mass sensitivity, primarily due to the absence of eddy diffusion and mass transfer resistance. However, it still experiences longitudinal molecular diffusion and electromigration dispersion, which contribute to band broadening. Despite these factors, the lack of significant mass transfer resistance allows for highly efficient separation and improved sensitivity. In CE–MS, metabolites are first separated based on charge and size by CE. Subsequently, MS is used for detection across a wide mass‐to‐charge ratio (m/z) range (70–1027 Da). CE can be paired with various mass analyzers, with the QqQ mass analyzer being the most common. ESI is the preferred technique in CE–MS for metabolite profiling, especially for more polar compounds. This method is characterized by low reagent consumption, cost‐effectiveness, high sensitivity, and high throughput, with minimal sample handling required [217]. The typical CE–MS setup involves a sheath‐flow interface, which is both simple to construct and highly reproducible. However, the development of sheathless interfaces offers promising potential for improving metabolite coverage in biological samples. Boizard et al. utilized a beveled tip sheath‐liquid interface, initially developed by Tseng et al., as an alternative to the conventional sheath‐liquid interface [218]. This CE–MS technique, when applied to pooled human urine samples, revealed a slightly more significant number of metabolite features compared to the standard CE–MS method. Furthermore, by calculating the mean intensity for each metabolite observed across multiple runs with both methods, they determined that the beveled tip CE–MS approach yielded an average sensitivity enhancement of threefold over the classical CE–MS method. CE–MS has notable limitations in repeatability and sensitivity. The small volume of injected sample restricts sensitivity, and temporal drift can affect the reproducibility of metabolic phenotyping [138].

5.4. SFC–MS

Milton et al.’ development of open‐tubular capillary column SFC in the 1980s represented a significant milestone in separation science [219]. SFC provides enhanced separation efficiency, particularly for non‐polar and moderately polar compounds, due to its supercritical CO₂‐based mobile phase, which offers lower viscosity and higher diffusivity than conventional LC solvents that are beneficial for high‐throughput analyses requiring both speed and resolution [220]. In addition, SFC is more environmentally friendly, as it reduces reliance on organic solvents, aligning with green chemistry principles, especially valuable in large‐scale metabolomic analyses [29]. For specific analytes, SFC shows improved selectivity and sensitivity, making it preferable for lipidomics and hydrophobic metabolites such as lipids, fatty acids, and fat‐soluble vitamins, which often exhibit low retention in conventional RP‐LC [221]. By offering orthogonal selectivity, SFC complements LC–MS in metabolomic studies, enhancing coverage for a more comprehensive analysis [222].

SFC SFC–MS is emerging as an alternative approach in this field, offering unique advantages. When integrated with multivariate statistical analysis, this method allows for identifying spectral features that highlight biochemical variations, facilitating the discovery of potential biomarkers [223, 224, 225]. SFC–MS has been successfully applied in diverse areas of metabolite profiling, including the examination of lipids in biological fluids [226], metabolic phenotyping [225], and estrogen metabolite analysis [227]. SFC is a hybrid technique combining features of GC and HPLC, offering a unique range of separation modes not achievable with either GC or HPLC alone. By precisely controlling temperature and back pressure, SFC enables high‐throughput and HR analysis, characterized by low theoretical plate heights even at high flow rates. This makes it particularly suitable for metabolomics, where both hydrophilic metabolites and lipids, which function as signaling molecules in biological processes, are commonly targeted [228].

One of the primary strengths of SFC–MS lies in its ability to separate polar and non‐polar metabolites more effectively than traditional LC, which is critical for lipidomics and glycolipid profiling. In particular, SFC–MS allows for better resolution of lipid isomers, which can be difficult to distinguish using LC–MS due to their structural similarities and low ionization efficiency. Recent advancements demonstrate its potential as a powerful platform for lipidomic and metabolomic analysis when coupled with ESI–MS. In a research study, an SFC–MS method that efficiently analyzes metabolites of varying polarities, employing two consecutive injections from biphasic methyl tert‐butyl ether (MTBE) extractions [229]. The method utilizes a diol column and gradient programs involving methanol, water, and ammonium formate modifiers. The primary focus of this method development was on optimizing system pressure limits, ensuring long‐term repeatability, and improving chromatographic performance. We optimized various parameters, such as flow rate programs, modifier composition, gradient conditions, and solvent selection for the injections, to achieve high performance. The resulting SFC–MS approach enabled the rapid and comprehensive analysis of both lipids and polar metabolites from plasma samples, completing a full cycle in just 24 min using a single platform, column, and mobile phase. However, a comparison of SFC–MS with traditional LC–MS revealed differences in the separation selectivity for metabolites. Although SFC–MS provided a fast alternative, it showed lower sensitivity compared to LC–MS, attributed to the higher flow rates and less efficient chromatographic resolution associated with SFC. Nonetheless, this method showcases the ability of SFC–MS to provide a comprehensive, time‐efficient solution for lipidomic and metabolomic analyses, especially for profiling metabolites across a broad polarity range [229].

The polarity of the SFC mobile phase can be adjusted by adding polar organic solvents, such as methanol, which allows for the elution of more polar analytes, thus broadening the range of compounds that can be analyzed [230, 231]. The supercritical state, achieved when temperature and pressure surpass their critical points, endows fluids with unique properties, combining the density of liquids with the viscosity and diffusivity of gases, making supercritical fluids highly desirable for separation processes. These characteristics facilitate efficient, high‐throughput separation, positioning SFC–MS as a valuable tool in metabolomics research [232]. In addition, using a 2 µm particle size in the stationary phase enhances the resolution for more polar compounds. Two types of SFC systems are available: open‐tubular column SFC (OT‐SFC), which resembles GC, and packed‐column SFC (PC‐SFC), similar to LC. SFC offers several critical advantages over RP‐LC, including faster mass transfer, shorter analysis times, and improved separation capacity and resolution due to the higher diffusion coefficient and lower viscosity of sCO₂ [233]. These features make SFC an increasingly attractive option in metabolite profiling, and its combination with MS enhances sensitivity and selectivity. SFC is particularly effective for separating isomers and enantiomers, with chiral SFC–MS emerging as a powerful technique for characterizing chiral mixtures [234]. It allows for simultaneous chiral and achiral separations, comprehensively analyzing complex mixtures [235]. In addition, SFC–ESI–MS offers a promising alternative to RP‐LC–MS for metabolite profiling, with the higher CO₂ content in the mobile phase improving the evaporation process during ionization. However, the decompression of CO₂ between the SFC and MS interface can lead to analyte precipitation and reduced retention time reproducibility. To mitigate this, pre‐back pressure regulator (BPR) splitting interfaces are now commonly used in SFC–ESI–MS coupling [236]. The significant benefits of SFC–MS include enhanced kinetic performance, reduced consumption of organic solvents, more efficient chiral separations, and increased productivity.

5.5. Ion‐Mobility MS

Despite being a relatively novel technique, IM–MS has gained significant acceptance within the metabolomics community. IM–MS introduces an additional orthogonal separation dimension between chromatographic separation and MS detection without extending analysis time. This separation occurs on the millisecond timescale, making IM–MS highly compatible with both rapid LC and high‐throughput MS approaches, particularly TOF mass analyzers, which facilitate quick cycles [237]. Figure 2 illustrates the operational stages of an IM–MS system, highlighting the process from sample ionization to mass detection. As a gas‐phase technique, IM–MS separates ions as they traverse an ion‐mobility cell under the influence of an electric field and in the presence of an inert buffer gas. Ions are distinguished based on their mobility or drift time, which is inherently related to their size, shape, and charge. Assuming constant experimental conditions such as drift‐tube length, gas pressure, temperature, and electric field, ion drift time is proportional to the rotationally averaged collision cross‐section (CCS). The CCS measures the effective area involved in the interaction between an ion and the buffer gas, providing a unique and reproducible value for each analyte. This characteristic reflects the chemical structure and 3D conformation of the molecules, underscoring the utility of IM–MS in metabolomics. In particular, in untargeted metabolomics, CCS values can complement traditional parameters such as retention time, mass‐to‐charge ratio, and fragmentation patterns, thereby enhancing the reliability of metabolite identification [238].

FIGURE 2.

Schematic representation of the working principle of ion‐mobility mass spectrometry (IM–MS). The sample, introduced with a carrier gas, is ionized in the ionization source, generating charged analyte particles. These ions pass through the desolvation region, where solvent molecules are removed. The ion gate modulates the entry of ions into the drift region, ensuring discrete ion packets for analysis. Within the drift region, ions travel under the influence of an electric field through a gas introduced via the drift gas inlet. The separation of ions is based on their collisional cross‐section and mobility in the gas phase. Smaller, more mobile ions move faster, whereas larger ions lag behind. Separated ions are guided through the pressure interface into the time‐of‐flight mass spectrometer (TOF–MS), where they are further analyzed for their mass‐to‐charge ratio (m/z). A reflector enhances resolution by correcting kinetic energy discrepancies. Finally, ions are detected by the microchannel plate (MCP) detector, generating a signal for data interpretation.

Several IM–MS technologies are currently available commercially, including (1) drift‐tube ion‐mobility spectrometry (DTIMS), (2) traveling‐wave ion‐mobility spectrometry (TWIMS), (3) field‐asymmetric ion‐mobility spectrometry (FAIMS), also known as differential mobility spectrometry (DMS), (4) differential mobility analyzers (DMA), and (5) trapped ion‐mobility spectrometry (TIMS), which involves confinement and selective release of ions. These techniques differ in their application of electric fields and the state of the buffer gas used. DTIMS and TWIMS are time‐dispersive methods wherein all ions traverse the same pathway but exhibit varying drift times. In contrast, FAIMS and DMA function as space‐dispersive methods, separating ions along distinct drift paths based on differences in their mobilities. TIMS operates by initially trapping ions in a pressurized environment before selectively releasing them according to their mobility characteristics. DTIMS allows directly calculating CCS values from drift times. In contrast, other methods necessitate calibrating calibrants with established CCS values to derive the CCS of unknown ions based on their drift times [238, 239].

IM–MS significantly enhances metabolome coverage by improving selectivity and resolution among metabolites, with particularly notable applications in lipidomics. Lipid analysis is particularly challenging due to the structural diversity and numerous lipid isomers within biological samples. Unlike traditional MS/MS approaches, IM–MS allows the differentiation of lipid isomers that vary solely in the acyl chain position, double bond placement, or geometrical configuration. For instance, LTB₄ and 5S,12R‐diHETE, which originate from distinct metabolic pathways and exhibit different biological functions, are diastereomers and geometric isomers [240]. As a result, they present identical mass spectra and similar retention profiles when analyzed using conventional LC–MS/MS. However, incorporating IM–MS (specifically, DMS) facilitates baseline separation of these compounds by employing distinct compensation voltages [241]. Beyond lipid analysis, IM–MS has proven beneficial for examining polar metabolites in various biological fluids, primarily in untargeted applications. Recent advancements in IM–MS applications in metabolomics showcase its capability to enhance analytical throughput, resolution, and the identification of complex metabolites and lipids in various biological contexts. A recent study introduced the rapid HILIC‐Z ion‐mobility MS (RHIMMS) method, which combines HILIC with drift‐tube ion‐mobility‐quadrupole TOF MS (DTIM‐qTOF MS) for untargeted metabolomics. This approach achieved rapid chromatographic separation with a run time of just 3.5 min per sample, maintaining reproducibility with less than 20% variation over 200 injections for selected metabolites. The integration of ion mobility allowed for improved annotation and the ability to distinguish isobaric compounds, enhancing the analysis of complex biological samples [242]. In another most recent study, a fast and comprehensive lipidomics method utilizing ion‐mobility MS was developed. This technique required only 8 min for separation and detected over 1000 lipid molecules in a single analysis of common biological samples. The method demonstrated high reproducibility and accurate quantification, making it suitable for large‐scale clinical lipidomic studies. Application of this method to plasma samples from colorectal cancer patients revealed significant changes in lipid species between preoperative and postoperative states, providing insights into disease‐related lipid metabolism alterations [243].

Metabolite isomerism remains a significant challenge in metabolomics, as many metabolites with similar physicochemical properties can be difficult to distinguish. The resolution of isomeric metabolites is crucial for accurate metabolic profiling, as isomers can have different biological functions and may be implicated in disease pathways. A promising solution to this challenge comes from IM–MS. A new approach combining HR ion‐mobility spectrometry (IMS) with cryogenic infrared (IR) spectroscopy offers a solution. IMS separates isomeric metabolites in milliseconds, whereas cryogenic IR spectroscopy provides distinct molecular fingerprints for precise identification. This method eliminates the need for recurring standards by matching IR fingerprints to a database, significantly improving speed and accuracy. Demonstrated with a database of eight metabolites, this technology enhances metabolomics workflows by offering fast, cost‐effective, and reliable identification of metabolite isomers and isobars [244]. Despite its potential to enhance metabolome coverage and metabolite annotation through CCS values, IM–MS encounters significant challenges in data interpretation. In an LC–IM–MS workflow, in‐source fragments, dimers, and adducts are also separated within the ion‐mobility cell, complicating the accurate regrouping and assignment of these signal features. This complexity remains inadequately addressed by current software solutions, particularly in untargeted metabolomics workflows [238].

5.6. Imaging MS

Imaging MS (IMS) represents a potent method for visualizing the spatial distribution of exogenous drug metabolites and endogenous metabolites such as lipids, amino acids, proteins, and organic acids within tissue sections without the need for radiolabeling [245, 246, 247]. This technique involves the chemical imaging of known and unknown metabolites, thereby providing a comprehensive structural profile of metabolite molecules directly within biological tissues. Figure 3 illustrates the workflow for IMS, showcasing its application to spatially resolved metabolomics analysis of tissue samples. IMS holds significant promise across various stages of the drug discovery process [247]. Applications of IMS encompass metabolite profiling in diverse fields, including lipid evaluation in cancer cells [248, 249], biomarker discovery in diseases like breast cancer [250], and renal cell carcinoma [251]. IMS is a crucial tool for metabolite imaging, filling a critical gap where existing technologies are lacking.

FIGURE 3.

Overview of imaging mass spectrometry (IMS) techniques used for spatial metabolomics. Following organ excision from a model organism, the tissue is cryosectioned into thin slices using a cryomicrotome and mounted onto appropriate slides: non‐conductive slides for desorption electrospray ionization (DESI) and secondary ion mass spectrometry (SIMS) or ITO‐coated slides for matrix‐assisted laser desorption/ionization (MALDI) imaging. Three IMS techniques are depicted: DESI, where a charged solvent spray desorbs and ionizes analytes directly from the tissue surface, enabling spatially resolved metabolite mapping; SIMS, which uses an ion beam to sputter secondary ions from the tissue, providing high spatial resolution suitable for subcellular imaging; and MALDI, where a thin matrix layer facilitates analyte desorption and ionization when irradiated by a laser, offering broad applicability for mapping large biomolecules such as lipids and proteins. Each technique contributes unique advantages in terms of spatial resolution, sensitivity, and molecular coverage, allowing detailed visualization of the metabolic landscape within tissue sections.

Critical advantages of IMS include its capability to detect and image metabolites without necessitating specialized labeling. It allows for the localization of unexpected metabolites and enables the simultaneous imaging of diverse metabolite types [252, 253]. IMS employs retrospective data analysis to identify and chemically image unknown compounds. It distinguishes between the spatial distributions of metabolites across whole‐body sections. During IMS analysis, analyte molecules are desorbed from the sample surface, typically removing only a few molecular monolayers and leaving sufficient samples for additional studies. IMS relies on retrospective data analysis from experiments to identify and chemically image unknown compounds. It can differentiate the spatial distribution of drugs and their metabolites across entire tissue sections. During IMS analysis, analyte molecules are desorbed from the sample surface, often removing only a few molecular monolayers, which permits additional investigations on the remaining material [247].

Several ionization techniques are employed in IMS to transition analytes from the condensed phase into the gas phase. These techniques include MALDI, secondary ion MS (SIMS), and DESI. In SIMS, the sample surface is irradiated by a beam of monoatomic, polyatomic, or cluster ions (such as Cs⁺, O₂ ⁺, O⁻, Ar⁺, or Ga⁺) at high energy. SIMS requires a high‐vacuum environment and offers the advantage of producing HR spatial images [254, 255]. Among these techniques, MALDI is the most frequently used for IMS due to its gentle ionization process, making it suitable for small molecules (e.g., drugs, lipids, and endogenous metabolites) and large biomolecules (e.g., proteins and peptides). During MALDI‐IMS, samples are irradiated with UV or IR laser pulses. This allows for high‐sensitivity, molecular‐specific imaging of high‐mass molecules, with spatial resolution typically at 25 µm or higher. MALDI is widely applied in the profiling and imaging of peptides and proteins from biological tissues, and there is growing interest in its use for analyzing small molecules, including drugs and their metabolites [250, 256, 257]. DESI, an alternative to MALDI, generates ions through the evaporation of small, highly charged solvent droplets containing analytes [258, 259]. Specialized software allows the plotting of signal intensities for each point in a 2D array, enabling the selection and digital storage of analyte signals from mass spectral data [247]. Currently, IMS is employed to measure small organic compounds, particularly in analyzing exogenous drugs and endogenous metabolites. Nevertheless, IMS has some limitations: it requires high efficiency in targeting specific molecules, and in the lower mass‐to‐charge (m/z) region, multiple compounds can share the same nominal mass, complicating analysis [245].

6. Applications of Metabolomics in Biomedical Research

Nowadays, most of clinical tests involve the analysis of metabolites, underscoring their crucial role in therapeutic applications [260]. A common example is the determination of blood glucose levels in diabetes patients. Notably, over half of today's widely used medications are small molecules, many of which are derived from metabolites. Furthermore, disrupted metabolism contributes to many genetic disorders, and unfavorable interactions between the proteome and metabolites are associated with various human diseases. The increasing relevance of metabolomics in biological research is driven by advancements in MS‐based analytical platforms. One of the earliest applications of metabolomics in diabetes research involved plasma phospholipid metabolic profiling to distinguish Type 2 diabetes mellitus from control groups [261]. This study demonstrated the potential of metabolomics as a source of biomarkers and the use of multivariate statistical techniques, such as principal component analysis and partial least squares discriminant analysis, for classification. Subsequent studies utilizing LC–MS/MS and proton NMR (¹H‐NMR) have examined metabolic pathways affected by insulin deficiency, revealing disruptions in gluconeogenesis, mitochondrial bioenergetics, and amino acid oxidation [262]. Metabolomic profiling has also provided insights into insulin resistance. For instance, Newgard et al. identified a metabolic profile associated with branched‐chain amino acids (BCAAs) that differentiates lean from obese individuals and is linked to insulin resistance [263]. Wang et al. further discovered five amino acids (leucine, isoleucine, valine, and phenylalanine) that can predict the early risk of diabetes by analyzing metabolite profiles in humans [264]. In a 12‐year follow‐up, they analyzed metabolite profiles in control and prediabetic groups, matched for age, sex, body mass index (BMI), and fasting glucose. They observed significant changes in the prediabetic group's BCAA and aromatic amino acid levels. These findings were confirmed in a randomized control group and supported the predictive potential of these metabolites in diabetes risk assessment. Their research employed a targeted metabolomics approach to quantify metabolites related to the urea cycle, nucleotides, and amino acids. In contrast, Zhao et al. [265] used non‐targeted metabolomics to study prediabetic patients' metabolic profiles in plasma and urine, identifying alterations in fatty acid, tryptophan, uric acid, bile acid, and lysophosphatidylcholine metabolism.

Fibrosis, a process that varies between individuals, has no consistent predictors, but mitochondrial dysfunction has been implicated in its development through metabolic disruption. Maeda demonstrated mitochondrial homeostasis issues in a radiation‐induced fibrosis model, with reduced fatty acid oxidation leading to lipid accumulation, including ceramide, diacylglycerol, and sphingomyelin [266]. These findings suggest metabolomics could help identify predictive or diagnostic markers for fibrosis.

Metabolic disturbances, particularly unbalanced lipid metabolism, have been observed in animal models and humans with cystic fibrosis [267, 268]. Disrupted bile acids, essential for cholesterol metabolism and lipid absorption, are characteristic of cystic fibrosis. In human liver fibrosis, sphingosine kinase activation increased sphingosine‐1‐phosphate, contributing to fibrosis and inflammation [269]. Even before the onset of obesity and hyperlipidemia, mice on high‐fat diets developed cardiac fibrosis, with inflammatory lipid mediators such as epoxy‐eicosatrienoic acids, leukotrienes, and prostaglandins likely playing a role. In idiopathic pulmonary fibrosis, changes in bioactive phospholipids such as 1‐acyl‐sn‐glycero‐3‐phosphate and phosphatidic acid, which regulate cell proliferation, migration, and survival, were identified in bronchoalveolar lavage fluid. BCAAs have also been shown to reduce transforming growth factor β, thereby decreasing liver fibrosis. These results suggest metabolomics could be a valuable tool for uncovering disease mechanisms and identifying prognostic or diagnostic biomarkers for fibrosis [270].

Cancer metabolism is primarily characterized by enhanced glycolysis, leading to lactic acid accumulation to fuel rapid cell proliferation, known as the Warburg effect. This metabolic reprogramming is thought to reflect the adaptation of cancer cells to the hypoxic conditions typical of tumors or mitochondrial dysfunction caused by the disease. There is increasing interest in studying the metabolic reprogramming of cancer, particularly pathways related to nucleotide, energy, and lipid metabolism [271]. Bioactive lipids, such as eicosanoids, are implicated in cancer, but the role of lipids in tumorigenesis remains underexplored. Recent studies of global lipid profiles in breast cancer tissues from 267 patients revealed an increase in de novo fatty acid synthesis in tumors compared to normal tissues, likely due to their incorporation into membrane phospholipids. Immunohistochemical analysis of proteins selected from in silico transcriptome database searches has confirmed strong expression of lipid metabolism pathways related to de novo fatty acid synthesis in cancer [272].

7. Conclusion

The recent technological advancements in liquid‐phase chromatography and MS have the potential to enhance metabolomics, particularly in clinical applications, significantly. The drive toward high‐throughput analysis and comprehensive metabolome coverage is essential for discovering novel biomarkers to improve our understanding of disease mechanisms, facilitate early diagnosis, and personalize treatment strategies. Although traditional methods, like RP‐LC–MS and GC–MS, remain prevalent, emerging techniques, such as HILIC, SFC, and IM–MS, offer exciting possibilities for expanding metabolome coverage and improving the discrimination of complex metabolite profiles. However, integrating these advanced methodologies into large‐scale studies is still hindered by challenges such as reproducibility, user expertise, and the need for robust data interpretation. Addressing these issues will require a concerted effort from experienced researchers to mentor the next generation of scientists and to advance technological solutions that ensure reliability and consistency.

Moreover, the distinction between stereoisomers and the exploration of multidimensional chromatographic approaches hold promise for further refining metabolomic analyses. As we look to the future, a collaborative effort to develop standardized protocols, optimize sample preparation techniques, and foster innovation will be crucial. Balancing the complexities of increased metabolite coverage with practical considerations, such as cost and sample size, will be essential to achieving the goals of clinical metabolomics. Ultimately, the continued evolution of metabolomics technologies will pave the way for breakthroughs in understanding metabolic diseases, leading to enhanced diagnostic and therapeutic options for patients. As these innovations mature, they will contribute to a more comprehensive view of human health, enabling personalized medicine and targeted interventions that can significantly improve patient outcomes.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.