Neurotransmitters, which are biosynthesized and metabolized through host metabolic pathways, dietary intake, and microbial production, act as messengers and modulators between cells, tissues, and organs in the human body (Yang et al. 2021). Because neurotransmitters regulate various neurological functions via the central nervous system (CNS), distorted levels of neurotransmitters can lead to numerous neurological disorders, such as Alzheimer’s disease, Parkinson’s disease, Huntington’s disease, major depressive disorder, and schizophrenia (Banerjee et al. 2020; Hanada 2020; Bhat et al. 2021). Therefore, the development of analytical methods for detecting neurotransmitters in biological samples is a prerequisite for unveiling the pathophysiological mechanisms underlying numerous neuropsychological disorders. However, developing an analytical method for neurotransmitter profiling in biological samples is challenging because of trace levels of neurotransmitters and complex biological matrices (Bylda et al. 2014). Therefore, for decades, sophisticated sample preparation and sensitive and selective instrumental analysis methods have been developed for detecting neurotransmitters in various biological samples. In this study, we could find 70 results between 2019 and 2023 after searching PubMed using the keywords “neurotransmitter profiling, neurological disease, and biological sample”. Of the results, we reviewed and organized research papers related to analytical methods for profiling neurotransmitters in biological samples. Research on analytical methods to determine neurotransmitters in biological samples published before 2018 was well organized in a previous report (Lee et al. 2018). In particular, we describe recent analytical advances and attempt to provide insights to help discover biomarkers for neurological diseases.

Most neurotransmitters in the human body are vulnerable to the physicochemical conditions of the human body. Neurotransmitter concentration levels in biological samples can be influenced by enzymatic activity (Purves et al. 2001). Furthermore, under high pH, temperature, and illumination, the catechol group of the catecholamine structure is easily oxidized and converted to its corresponding oxidized quinone form (Herlinger et al. 1995; Pham and Waite 2014). Therefore, appropriate storage conditions should be maintained from the time of biospecimen collection to analysis. Generally, to maximally inhibit enzymatic activities, the recommended temperatures for biological sample storage are below −70℃, without additives (Sapcariu et al. 2014). For the stable storage of catecholamine neurotransmitters, the utilization of acid additives is widely accepted, to prevent the oxidation of catechol (Nicolas et al. 2003). However, since acid additives can lead to severe signal distortion during instrumental analysis, the use of acid additives is carefully considered. Sample preparation techniques used for the profiling of neurotransmitters in biological samples are generally categorized into several types, including solvent dilution, liquid-liquid extraction (LLE), and solid- phase extraction (SPE). With the development and advancement of analytical instruments, simple sample pretreatment techniques, such as solvent dilution, direct extraction, and filtration have been widely utilized (Jurado-Sánchez et al. 2009). In particular, with highly sensitive and selective analytical instruments, sophisticated sample preparation methods have been employed to use only small volumes of biospecimens. In this review, we focus on advanced and sophisticated sample preparation techniques, such as liquid-phase microextraction (LPME) and miniaturized SPE. Furthermore, representative microextraction techniques of LPME and miniaturized SPE were summarized in Table 1.

Table 1 . Recent microextraction techniques.

Group		Extraction solution/sorbents	Sample	Reference
LPME	UA-IL-DLLME	BMIMPF6 and acetonitrile	Human urine	Zhou et al. (2020)
	DLLME–derivatization	Chloroform and ethanol	Human urine	Rodríguez-Palazón et al. (2023)
	NLPNE + IFD	Methanol−water	Human plasma	Yu et al. (2022)
	MIL-based LLME	[P6,6,6,14]2[CoCl4]	Human plasma	Ding et al. (2023)
Miniaturized-SPE	TF-SPME	A hydrophilic-lipophilic balance (HLB) coating	Human urine	Chen et al. (2021b)
	Magnetic dispersive SPE	Fe3O4@GO	Rat brain	Zhao et al. (2019)
	MEPS	eVol-MEPS	Cell pellet	Protti et al. (2023)

In recent decades, LPME has been widely used to determine trace levels of target analytes in complicated matrices such as environmental, food, and biological samples (Nováková and Vlcková 2009). Because LPME is characterized by simple and easy operation, less consumption of harmful organic solvents and a demand for a small number of samples, diverse techniques based on LPME have been developed, with increasing attention in many research fields (Psillakis and Kalogerakis 2003; Prosen 2014). Numerous LPME methods have been used to extract neurotransmitters from biological samples. As shown in Fig. 1A, ultrasound-assisted ionic liquid dispersive liquid-liquid microextraction (DLLME) was developed to determine 15 target neurotransmitters, including acetylcholine, dopamine, and epinephrine, in urine samples of patients with dementia (Zhou et al. 2020). In this study, 1-butyle-3-methylimidazolium hexafluorophosphate and acetonitrile were used as the ionic liquid and disperser, respectively, to improve the migration efficiency of neurotransmitters in human urine. Similarly, the DLLME method combined with dansyl chloride derivatization was employed to detect four biogenic amines, synephrine, phenylephrine, tyramine, and octopamine, in human urine (Rodríguez-Palazón et al. 2023). Nanoconfined LPME combined with in-fiber derivatization was employed to detect lung cancer biomarkers, including sarcosine, alanine, gamma-aminobutyric acid, proline, isoleucine, glutamine, glutamic acid, phenylalanine, citrulline, and metanephrine in human plasma (Fig. 1B) (Yu et al. 2022). As shown in Fig. 1C, using [P_6,6,6,14]₂[CoCl₄] as a magnetic ionic liquid, liquid-liquid microextraction (LLME) was developed to detect 20 neurotransmitters, including choline, dopamine, and epinephrine, in human cerebrospinal fluid and plasma samples (Ding et al. 2023).

Figure 1. Advancements of liquid-phase microextraction (LPME). (A) Ultrasound-assisted ionic liquid dispersive liquid-liquid microextraction, (B) nanoconfined LPME combined with in-fiber derivatization, and (C) LPME using magnetic ionic liquid.

As the most popular miniaturized SPE method, solid- phase microextraction (SPME) is a highly selective and sensitive extraction method that does not require chemicals or solvents (Sarafraz-Yazdi and Amiri 2010). Generally, SPME methods require appropriate fiber types, extraction times, and temperatures with/without an automation system (Wei et al. 2021). Although SPME has been widely used to determine neurotransmitters in biological samples for several decades, in this review, numerous SPME methods regarded as traditional SPME were excluded. To overcome the limitations of traditional SPME fibers, thin-film SPME with a hydrophilic-lipophilic balance coating was developed to detect eight neurotransmitters, including epinephrine, norepinephrine, dopamine, metanephrine, normetanephrine, 3-methoxytyramine, serotonin, and histamine, in urine samples (Fig. 2A) (Chen et al. 2021b). Furthermore, to extract 16 neurotransmitters, including dopamine, epinephrine, norepinephrine, glutamic acid, and serotonin, from rat brain microdialysates (Fig. 2B), magnetic dispersive SPE using Fe₃O₄-decorated graphene oxide was developed in combination with stable isotope labeling derivatization (Zhao et al. 2019). Another advancement in miniaturized SPE is microextraction by packed sorbent (MEPS), which uses a small cartridge packed with sorbent materials (Fig. 2C) (Protti et al. 2023).

Figure 2. Advancements of miniaturized solid-phase extraction (SPE). (A) Thin film-coated solid-phase microextraction, (B) magnetic dispersive SPE using Fe₃O₄-decorated graphene oxide, and (C) microextraction by packed sorbent.

In addition to miniaturized extraction methodologies, several advanced pretreatment methods have been developed. To improve the sensitivity and selectivity of detecting free catecholamines in biological fluids, a simultaneous extraction/derivatization strategy was implemented using zirconium dioxide and phenyl isothiocyanate (Chen et al. 2021a). Furthermore, in vivo sampling techniques such as microdialysis have been widely employed because they enable real-time detection of distorted neurotransmitters in living systems (Su et al. 2020). Although real-time in vivo detection techniques can provide insights into the pathological mechanisms of neurological diseases and help diagnose brain diseases, they are difficult to apply directly to patients with neurological diseases. These in vivo sampling techniques have been well-documented in previous reports (Matys et al. 2020; Su et al. 2020; Yu et al. 2022).

As numerous matrix interferences in biospecimens should be separated from neurotransmitters, chromatography has been widely performed to determine neurotransmitters in biological samples. Among various detection techniques, such as flame ionization detection, ultraviolet-visible light detection, fluorescence detection, or electrochemical detection, both gas chromatography (GC) and liquid chromatography (LC) have been conventionally and extensively utilized to separate neurotransmitters from matrix interferences in biological samples. However, GC involves long and tedious sample preparation procedures, while LC generally facilitates to determine polar neurotransmitters in biological samples with ease of sample preparation (Chirita et al. 2010; Nirogi et al. 2010). To identify non-volatile and thermally vulnerable neurotransmitters, GC should be accompanied by chemical derivatization to improve volatility and protect the polar functional groups (Valdez et al. 2018; Mill and Li 2022). Furthermore, because of its suitability for biological samples, LC has been widely employed with and without chemical derivatization (Scalbert et al. 2009). To improve the separation capacity of LC, microchip chromatographic analysis equipped with single-particle frit-based packed columns has been employed to determine neurotransmitters in human urine samples (Li et al. 2020a). Although capillary electrophoresis (CE) is not classified as chromatography, CE has been widely used to detect neurotransmitters in biological samples because of its outstanding separation capacity (Lapainis and Sweedler 2008).

Furthermore, the selection of an appropriate detection technique is important for the selective and sensitive determination of neurotransmitters in biological samples. Although most detectors used in LC are compatible with CE, detectors combined with GC are not compatible with LC or CE, except for those used in mass spectrometry (MS). In particular, MS combined with chromatographic separation techniques has been widely employed to detect neurotransmitters in biological samples. MS can selectively detect different neurotransmitters with the same retention times (RTs). Furthermore, since MS enables sensitive and selective detection with structural identification (Jiang et al. 2010), chromatography equipped with MS enhances identification and confirmation of isobaric neurotransmitters and their metabolites (McNair and Miller 2011). This feature could minimize false positive determination resulted from matrix interferences and analysis time to separate overlapping peaks of analytes with similar RTs (Bicker et al. 2013; Wojnicz et al. 2016). As mentioned previously, MS combined with chromatography is the ‘gold standard’ for detecting neurotransmitters in biological samples. Several advancements in chromatographic separation and detection techniques, such as MS imaging (Xu and Li 2019; Harkin et al. 2022), ion mobility spectrometry (Garcia et al. 2021), and electrochemical biosensors (Lakard et al. 2021; Fredj et al. 2023), have been reviewed in previous reports.

All analytical methods to determine neurotransmitters in biological samples have the potential to be applied to various intractable diseases, including neurological diseases (Jiang et al. 2020). Neurotransmitter profiling techniques have been widely performed to reveal the pathophysiological mechanisms of various disease models and to discover potential biomarkers for various diseases, such as diseases often considered incurable. However, we focused on the discovery of neurotransmitters as potential biomarkers for neurological diseases in this review.

One of the most common and incurable neurological diseases is major depressive disorder. In overall studies, chronic unpredictable mild stress (CUMS)-induced in vivo models are widely used and documented for investigating major depressive disorder. Studies have reported that kynurenine levels were increased, whereas acetylcholine levels were decreased, in brain and serum samples from the CUMS mouse model (Wang et al. 2019). Likewise, in the CUMS rat model, not only urinary kynurenine levels were increased, but also kynurenine/tryptophan ratios in plasma and brain samples were increased (Han et al. 2019). Altered kynurenine pathways in mood disorders, including major depression, bipolar disorder, and schizophrenia, have been well accepted, although further investigation would be necessary to unveil individual characteristics (Marx et al. 2021).

Alzheimer’s disease, the most common cause of dementia among neurodegenerative diseases, has been studied extensively in many fields. Since amyloid-beta plaques accumulated in the brain tissue of Alzheimer’s disease patients have been recognized as a cause or consequence of Alzheimer’s disease (Hampel et al. 2021), it is important to form amyloid-beta plaques in in vivo models. The in vivo Alzheimer’s disease model induced by D-galactose and amyloid-beta-peptide 25-35 has been widely used to examine Alzheimer’s disease. It has been reported that brain levels of norepinephrine and acetylcholine were decreased in Alzheimer’s disease rat models (Sun et al. 2018; Xu et al. 2018). Although the mechanisms of Alzheimer’s disease are not yet known, altered metabolism of neurotransmitters, astrocyte energy and neurons in patients with Alzheimer’s disease (Andersen et al. 2022).

Parkinson’s disease is the second, most prevalent neurodegenerative disorder and eventually leads to dementia (Mhyre et al. 2012). To cause Parkinson’s disease in in vivo models, some neurotoxins such as 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine and rotenone have been popularly used. In particular, in a rotenone-induced rat model, gamma-aminobutyric acid levels were increased in brain samples, while brain levels of several neurotransmitters, including acetylcholine, dopamine, glutamic acid, and glutamine, were decreased (Li et al. 2020b; Dai et al. 2023). However, it was reported that dopamine levels were increased in human plasma samples with Parkinson’s disease (Kremer et al. 2021; Wichit et al. 2021). This suggests that it is challenging to find appropriate targets to develop the drug, since Parkinson’s disease is a multifactorial disease (Yadav and Kumar 2022). Furthermore, to investigate non-motor symptoms of Parkinson’s disease, it was reported that an LC-MS/MS method was developed to profile tryptophan metabolites, including microbial produced neurotransmitters in human urine samples (Chung et al. 2023).

Moreover, Huntington’s disease is also known as a representative neurodegenerative disorder caused by the expansion of abnormal CAG repeats (Ramos et al. 2015). Since the R6/1 transgenic mouse model could express 115 CAG repeats in exon 1 of the human Huntington’s disease gene (Mangiarini et al. 1996), the R6/1 transgenic mouse has been popularly utilized. In a previous study using the R6/1 Huntington’s disease mouse model, it was reported that several neurotransmitters, including dopamine, 3,4-dihydroxyphenylacetic acid, homovanillic acid, serotonin, 5-hydroxy indole acetic acid, were decreased in the mouse brain (Puginier et al. 2019). The overall potential biomarkers described in this review are summarized in Table 2.

Table 2 . Potential biomarkers related to neurological diseases.

Disease model	Experimental group	Sample type	Analyte	Reference
Major depressive disorder	Chronic unpredictable mild stress-induced depression mouse model	Serum	(+) kynurenine, 3-hydroxy kynurenine (−) kynurenic acid, serotonin, 5-hydroxy indole acetic acid, norepinephrine, glutamic acid, acetylcholine	Wang et al. (2019)
		Brain	(+) kynurenine (−) acetylcholine
	Chronic unpredictable mild stress-induced depression rat model	Urine	(+) kynurenine, quinolinic acid (−) kynurenine/quinolinic acid	Han et al. (2019)
		Plasma	(+) quinolinic acid, 5-hydroxy indole acetic acid/serotonin, kynurenine/tryptophan (−) tryptophan, serotonin, xanthurenic acid, tyrosine, kynurenic acid/quinolinic acid
		Brain	(+) quinolinic acid, 5-hydroxy indole acetic acid/serotonin, kynurenine/tryptophan (−) tryptophan, serotonin, tyrosine, gamma-aminobutyric acid
Alzheimer’s disease	D-gal and Aβ25–35 induced rat model	Brain	(+) tryptophan (−) norepinephrine, acetylcholine, glutamic acid	Sun et al. (2018)
	D-gal and Aβ25–35 induced rat model	Brain	(−) serotonin, gamma-aminobutyric acid, dopamine, norepinephrine, acetylcholine, tyrosine	Xu et al. (2018)
Parkinson’s disease	Rotenone induced rat model	Brain	(+) gamma-aminobutyric acid (−) acetylcholine, dopamine, glutamic acid, glutamine	Dai et al. (2023), Li et al. (2020b)
	Human	Plasma	(+) dopamine, norepinephrine (−) epinephrine, serotonin	Kremer et al. (2021), Wichit et al. (2021)
		Urine	(+) homovanillic acid, vanillylmandelic acid (−) 5-hydroxy indole acetic acid
	Human	Urine	(+) indole-3-acetic acid	Chung et al. (2023)
Huntington’s disease	R6/1 mouse model	Brain	(−) dopamine, 3,4-dihydroxyphenylacetic acid, homovanillic acid, serotonin, 5-hydroxy indole acetic acid	Puginier et al. (2019)

To understand the pathophysiological mechanisms of various neurological diseases, development of analytical methods is crucial for detecting neurotransmitters in biological samples. With the advancements in sample preparation and instrumental analysis techniques, sensitive and selective neurotransmitter profiling methodologies have been developed to detect neurotransmitters in complicated biological matrices. In particular, because both LPME and miniaturized SPE methods are simple and easy to operate, consuming less harmful organic solvents and wasting fewer biological samples, they have been widely used to extract neurotransmitters from diverse biological samples. Furthermore, real-time in vivo sampling techniques would provide direct insights into many neurological disorders, while automated techniques could combine many advanced microextraction methods. Chromatographic separation techniques are indispensable for analyzing neurotransmitters, to separate them from endogenous matrix interference. MS combined with chromatography is now considered the ‘gold standard’ for detecting neurotransmitters in biological samples. Although the pathophysiological mechanisms of various neurological diseases remain unclear, with the advancements in analytical methods for detecting neurotransmitters in biological samples, potential biomarkers for the diagnosis, treatment, and prognosis of neurological diseases will be discovered in the near future.

The authors declare that they have no conflict of interest.

This study was supported by the National Research Foundation of Korea (NRF-2021R1C1C1005957).