Recent advances in mass spectrometry-based bioanalytical methods for endogenous biomarkers analysis in transporter-mediated drug-drug interactions

Abstract

Drug-drug interactions (DDI) are a critical concern in drug development and clinical practice. A new molecular entity often requires numerous clinical DDI studies to assess potential risks in humans, which involves significant time, cost, and risk to healthy study participants. Consequently, there is growing interest in innovative techniques to improve the prediction of transporter-mediated DDI. Researchers in this field have focused on identifying endogenous molecules as biomarkers of transporter function. The development of biomarkers is notably more complex than that of exogenous drugs. Owing to their inherent selectivity, sensitivity, and ability to provide absolute quantification, liquid chromatography-mass spectrometry (LC-MS) and liquid chromatography-tandem mass spectrometry (LC-MS/MS) are increasingly being employed for the quantitative investigation of new biomarkers. This review article presents recently developed bioanalytical approaches using LC-MS/MS for putative transporter biomarkers identified to date. Additionally, we summarize the published baseline endogenous levels of these potential biomarkers in a biological matrix to suggest a set of reference values for future research, thereby minimizing errors in biomarker-related data analyses or calculations.

Graphical abstract

Highlights

• •Overview of potential endogenous transporters biomarkers and supporting evidence.
• •Summary of LC-MS(/MS) bioanalytical methods for quantifying endogenous biomarkers.
• •Selection of appropriate approaches for calibration curves and QC samples in biomarker quantification.
• •Summary of individual baseline levels of endogenous transporters biomarkers in biological fluids.

1. Introduction

The absorption, distribution, and excretion of both endogenous and exogenous molecules depend on membrane transporters [1]. Therefore, changes in the expression and/or function of drug transporters critically influence the pharmacokinetics and pharmacodynamics of drug substrates [2]. To ensure safety during polypharmacy in clinical practice, evaluating the transporter-mediated drug-drug interactions (DDI) is essential during the new drug development [1]. The potential of a new drug candidate as a transporter inhibitor in humans is typically assessed using in vitro transporter inhibition studies [3].

Although in vitro data are commonly used to evaluate the potential of DDIs, it is important to recognize the limitations of this approach. These limitations include significant variability across different experimental systems and laboratories, as well as challenges in translating in vitro studies to in vivo situations [4]. To accurately assess the impact of novel drug candidates on co-administered drugs, regulatory agencies, such as the US Food and Drug Administration (FDA), require sponsors of new drug applications to conduct clinical pharmacokinetic DDI studies. If these candidates exhibit evident potential in vitro, regulatory DDI guidelines mandate further clinical studies to prevent false-positive or false-negative results [5].

Over the past several years, various endogenous transporter substrates have been identified as potential biomarkers for predicting functional changes in drug transporters and transporter-mediated DDI during the early stages of new drug development [6]. Analyzing the effects of transporter inhibitors on these endogenous biomarkers in human plasma or urine is an alternate method for evaluating the DDI potential of new drug candidates in vivo. Biomarker-driven studies have been applied in both preclinical and clinical research, as potential techniques for predicting DDI [7,8]. Consequently, the identification of endogenous biomarkers in plasma and urine as early indicators of potential transporter-mediated DDI has gained increasing attention [9].

Various methods have been used in clinical laboratories to investigate endogenous substances as biomarkers. According to the Biomarker Definitions Working Group, funded by the US National Institutes of Health (NIH), a biomarker is defined as a measurable characteristic that can be quantitatively assessed and objectively analyzed, serving as an indicator of normal biological processes, disease-related processes, or responses to therapeutic interventions [10]. Recently, biomarkers have gained recognition as valuable tools in drug research and development. Their use can enhance several aspects of the process, including target identification, risk evaluation, study design, dose escalation, patient stratification, and safety monitoring, ultimately reducing the time and cost associated with drug development. However, these biomarkers often require specialized chemicals and methods that are not commonly available. Consequently, liquid chromatography-mass spectrometry (LC-MS) and liquid chromatography-tandem mass spectrometry (LC-MS/MS) technology have become increasingly popular for the quantitative analysis of emerging biomarkers owing to their inherent sensitivity, selectivity, and ability to provide precise quantification [11].

Developing an assay for biomarkers is substantially more challenging than developing one for exogenous drugs. The complexities of biomarker quantification can be summarized as follows [12]. Fig. 1 summarizes the key challenges in using LC-MS/MS for endogenous biomarker quantification. First, eliminating interference from endogenous levels and accounting for fluctuations arecritical for a successful biomarker experiment. Second, incorporating an analyte into an authentic matrix for calibration curves and quality control (QC) samples can complicate the experiments and lead to less accurate and precise results. Furthermore, determining the specificity of endogenous compounds is difficult due to the absence of a true blank matrix. Third, unlike exogenous drugs, which undergo purification and extensive characterization, the standards for biomarkers are often unavailable. Fourth, naturally occurring analogs, that may share metabolic pathways can also interact because of their structural similarities. Fifth, while exogenous drug concentrations often fall within the detection range of LC-MS/MS methods, biomarker concentrations in preclinical or clinical samples can be exceedingly low. Six, the human or animal body tightly regulates endogenous substances, particularly bioactive substances. This poses an additional challenge in assay development owing to the instability of the biomarkers. If therapeutic agents significantly down-regulate these biomarkers, a more thorough method may be required to achieve sufficient sensitivity. Moreover, the calibration range must account for both the baseline and target levels after drug administration. However, this can be challenging due to discrepancies between healthy and diseased populations and inconsistent reports of baseline endogenous levels in the literature. Seventh, biomarkers can be derived from precursors already present in the biological matrix. If ex vivo synthesis is detected, the target analyte must be isolated from its biological precursor or preventative measures, such as enzyme inhibitors, must be implemented to prevent conversion. Eighth, another important consideration is ensuring sample consistency from the time of collection to analysis, which is known as “vein-to-vessel stability”. Analyte concentrations can be affected by factors such as instability and adsorption loss during sample collection. Therefore, careful assay development and protocol design are required. To accurately detect pharmacological effects, the protocol design should account for diurnal fluctuations, dietary impacts, and emotional fluctuations. In summary, biomarker identification presents numerous challenges for bioanalytical laboratories due to the inherent nature of biomarkers, often necessitating extensive method development to achieve acceptable assay performance.

Fig. 1.

Key challenges of biomarker quantification using liquid chromatography-tandem mass spectrometry (LC-MS/MS).

Several bioanalytical methodologies have been developed to improve the precision and selectivity of endogenous biomarkers [13]. This is particularly relevant given the widespread adoption of LC-MS/MS in both experimental and clinical pharmacokinetic laboratories. Much of the existing literature focuses on the identification and evaluation of promising endogenous biomarker candidates for transporter-mediated DDI. However, reviews of the bioanalytical procedures for these potential endogenous biomarkers, specifically those based on LC-MS/MS and their corresponding endogenous levels in biological fluids, remain limited. Therefore, in this review article, we aim to provide an overview of the evolution of assay methodologies, including LC-MS/MS, for the analysis of endogenous transporter biomarkers over the past two decades, as well as a summary of their baseline levels to facilitate future application studies.

2. Potential endogenous biomarkers for transporters

The liver and kidney play crucial roles in maintaining the delicate homeostasis of salts, water, and nutrients in body fluids through complex metabolic and filtration-reabsorption mechanisms [14]. Simultaneously, the body effectively eliminates unwanted endogenous metabolites and exogenous substances, including toxins and pharmaceutical compounds [15]. However, many solutes have limited permeability across the cellular membranes of the liver or kidney because of factors such as size, charge, and protein-binding affinity. Hence, the primary process of accumulation or active secretion in these organs is mediated by specific transporters belonging to the solute carrier (SLC) 22 family [16]. These transporters include organic cation transporter 1 (OCT1), which is mainly localized in the liver, and three other transporters predominantly found in the proximal tubules of the kidney – OCT2, organic anion transporter 1 (OAT1), and OAT3. These membrane transport proteins share a common genetic lineage and feature twelve membrane-spanning domains [17].

The liver and kidney rely on the combined action of OCT1, OCT2, and OAT1/3, which possess multi-specific binding loops, to facilitate the basolateral uptake of a wide range of cationic and anionic metabolites and drugs [18]. To efficiently eliminate solutes, OAT1/3 and OCT1/2 work in conjunction with other uptake transporters, such as organic anion-transporting polypeptides (OATPs) and efflux transporters located on the luminal cell surface in the liver and kidney [19]. These efflux transporters include multidrug and toxic compound extrusion transporter (MATE) 1/2K, P-glycoprotein (P-gp), breast cancer resistance protein (BCRP), and multidrug resistance proteins 2 and 4 (MRP2/4). OCT1/2 and OAT1/3 collaborate with additional SLC transporters and function as bi-functional transporters that facilitate both uptake and efflux. Notably, transporters such as organic cation carnitine transporter (OCTN1) and OCTN2, located on the luminal side, contribute to this process [20]. However, the rate of solute excretion is primarily influenced by the functional activity of the uptake transporters. Therefore, monitoring and adjusting the levels and activity of these critical uptake transporters is essential to ensure proper kidney excretion and liver elimination.

To date, several endogenous molecules have been identified as potential biomarkers for their associated transporters in important organs, including the intestine, liver, and kidney [6,8] (Fig. 2).

Fig. 2.

Well-known endogenous biomarkers for transporters-mediated drug-drug interactions (DDI) in the intestine, liver, and kidney. Transporters highlighted in green indicate those that are described in detail in this review. Transporters highlighted in pink represent transport proteins that are important but not discussed in this review. Created by using mindthegraph.com. OCT: organic cation transporter, MATE: multidrug and toxic compound extrusion, OCTN: organic cation/carnitine transporter, OAT: organic anion transporter, OATPs: organic anion-transporting polypeptides, BCRP: breast cancer resistance protein, MRP: multidrug resistance-associated protein, OST: organic solute transporter, PEPT: peptide transporter, ASBT: apical sodium-dependent bile acid transporter, MCT: monocarboxylate transporter, NTCP: sodium/taurocholate cotransporting polypeptide, BSEP: bile salt export pump, URAT: urate transporter, IBC: isobutyryl-carnitine, ET: ergothioneine, L-CAR: L-carnitine, m¹A: N1-methyladenosine, NMN: N1-methyl nicotinamide, HA: hippuric acid, GCDCA-S: glycochenodeoxycholate sulfate, CP: coproporphyrin, TDA: tetradecanedioate, HDA: hexadecanedioate, DHEAS: dehydroepiandrosterone sulfate, PDA: pyridoxic acid, HVA: homovanillic acid, KYNA: kynurenic acid, 6β-OHC: 6β-hydroxy cortisol.

A summary of potential endogenous biomarkers and clinical inhibitors associated with specific transporters, used in clinical DDI studies for drug development, is provided in Supplementary Table S1. For many therapeutically important transporters such as P-gp, information about their endogenous substrates is limited, and only a small number of endogenous molecules have been identified as substrates, as is the case for BCRP. Transporters with strong clinical significance and known biomarker data are described in more depth below. Advancements in this field are rapidly progressing, and our understanding of the role of transporters in drug absorption, distribution, therapeutic efficacy, and adverse effects, will significantly improve in the coming years [21].

2.1. Potential endogenous biomarkers of OCTs, MATEs, and OCTNs

2.1.1. Thiamine (vitamin B1)

Thiamine has been identified as a substrate not only for OCT1 (K_m = 0.78 mM) [22], but also for OCT2 (K_m = 59.9 μM–0.75 mM) [22,23], MATE1 (K_m = 3.5 μM–0.83 mM) [22,24], MATE2-K (K_m = 3.9 μM) [23], thiamine transporter 1 (THTR1) (K_m = 2,5 μM) [25], and THTR2 (K_m = 27 nM) [26]. Additionally, OCT3 has been shown to transport thiamine at a lower rate [25], while OCTN1 and OCTN2 were not involved in thiamine uptake [25]. Chen et al. [22] demonstrated that OCT1-mediated thiamine uptake is competitively inhibited by metformin and the biguanide derivative phenformin. These findings also indicate that OCT1 serves as a high-capacity transporter for thiamine and plays a crucial role in hepatocyte lipid and carbohydrate metabolism [22]. Moreover, compared to wild-type (WT) mice, Oct1/Oct2 double-knockout (dKO) mice exhibit 5.8-fold higher plasma concentration levels of endogenous thiamine and 79% reduced renal clearance (CL_R) of exogenous thiamine [27].

2.1.2. Isobutyryl-carnitine (IBC)

The potential of IBC as a biomarker for OCT1 has recently been proposed [[28], [29], [30], [31], [32]]. Multiple studies, including those on genome-wide associations, demonstrated a correlation between plasma IBC levels and OCT1 expression. Therefore, plasma IBC can be considered as a viable biomarker of OCT1 activity. Compared to individuals with OCT1 deficiency, those with high-activity OCT1 genotypes exhibit approximately 3-fold and 2-fold greater plasma IBC concentrations and excretion levels, respectively [29,33]. Clinical evaluations of OCT1 DDI using IBC measurements have proven reliable. For instance, following ritlecitinib administration, the area under the curve (AUC) of IBC from 0 to 24 h decreased by approximately 15%, consistent with OCT1 inhibition [31]. Similarly, another Pfizer development compound 1(PFE1) showed an inhibitory effect, resulting in a 35% decrease in plasma IBC levels [34].

2.1.3. N1-methylnicotinamide (NMN)

NMN has been identified as an endogenous substrate for OCT2, MATE1, and MATE2-K, with K_m values of approximately 300–318 μM, 301 μM, and 422 μM, respectively [35,36]. In women receiving metformin, NMN CL_R was reported to be higher in both mid- and late pregnancies than in the non-pregnant state [37]. In the clinical investigation, trimethoprim, an OCT and MATE inhibitor, increased metformin AUC_0–∞ by 29.5% and reduced the CL_R of metformin and NMN by 26.4% and 19.9%, respectively (P < 0.01) [38]. Furthermore, trimethoprim-induced decreases in NMN and metformin CL_R were strongly correlated (r_S = 0.727, P = 0.010). These findings suggest that NMN can serve as a valuable endogenous probe for renal DDI involving renal cation transporters.

2.1.4. Creatinine

Creatinine has been reported as a substrate for OCT2, MATE1, and MATE2-K with K_m values of 1.9–56 mM, 10 mM, and 22 mM, respectively [24,35]. Transient inhibition of renal transporters (OCT2, OAT2, and MATE1/2K), which are involved in the active secretion of creatinine, can lead to increased serum creatinine levels and reduced CL_R without affecting glomerular filtration rate [39]. The development of biomarker-informed approaches to enhance DDI investigations has been supported by physiologically based pharmacokinetic models for the suggested biomarkers of OCT2 and MATE1 activity, including creatinine and NMN. Furthermore, elevated plasma creatinine levels and reduced urine excretion have been observed in drug-biomarker interactions involving trimethoprim, pyrimethamine, and cimetidine [40].

2.1.5. N¹-methyladenosine (m¹A)

m¹A has recently been identified as a novel substrate of OCT2 and MATE2-K recently [41], with certain advantages over other OCT2/MATE substrates (creatinine and NMN). Owing to its significant contribution to tubular secretion, the net clearance of m¹A is highly dependent on these transporters. As a result, m¹A plasma levels increase significantly in knockout (KO) mice and cynomolgus monkeys with chemical impairment of OCT2 or MATE2-K (fold change of 2.78 in Oct1/2 dKO mice, and 1.72 in DX-619-treated monkey, respectively). In healthy subjects, m¹A exhibits modest diurnal and inter-individual plasma fluctuations, serving as a superior biomarker for detecting inhibitory drug-induced OCT2/MATE2-K activity [41].

2.1.6. Tryptophan

According to Song et al. [42], tryptophan demonstrates gene-dose effects on transporter activity based on OCT2 genotypes and shows the strongest linear correlation with metformin pharmacokinetic parameters. An inhibition assay revealed that tryptophan inhibits the absorption of 1-methyl-4-phenylpyridinium in a concentration-dependent manner. A subsequent uptake experiment indicated that OCT2 reference and variant (808G > T) oocytes exhibit distinct rates of tryptophan uptake. These findings suggest that tryptophan can be used as an endogenous substrate for OCT2 and as a potential biomarker for assessing variability in the transport activity of OCT2.

2.1.7. Dopamine

Dopamine is a substrate of MATE, and its renal excretion is reduced in Mate1 mutant animals or upon imatinib treatment [43]. Specific transport of tritiated dopamine (³H-dopamine) was induced by the expression of rat OCT2 (rOCT2) in HEK293 cells [44]. A well-known OCT2 inhibitor, 1,1′-diisopropyl-2,4′-cyanine (disprocynium24), prevented ³H-dopamine uptake into 293r_OCT2 cells with a K_i of 5.1 nM [44]. Although dopamine is a substrate of OCT2 in vitro [44], the link between OCT2 activity and urinary dopamine excretion may be weak because dopamine is produced in renal proximal tubule cells [43]. Human studies on the use of dopamine as a MATE biomarker are limited. A recent study revealed that the dopamine D1 receptor (D1DR), the primary mediator of dopaminergic signaling in the brain and kidney, functions on the plasma membrane and is activated in the Golgi apparatus in the presence of its ligand, as demonstrated using novel nanobody-based biosensors [45]. This study showed that OCT2, a dopamine transporter, is required for the activation of the Golgi pool of D1DR, explaining how membrane-impermeant dopamine can access subcellular pools of D1DR. It also showed that the activation of Golgi-D1DR by dopamine in murine striatal medium spiny neurons is OCT2-dependent.

2.1.8. Ergothioneine (ET)

In 2005, ET was identified as a key substrate of OCTN1 [46]. Since ET is transported 100 times more efficiently than tetraethylammonium and carnitine, researchers proposed renaming OCTN1 to “ergothioneine transporter” (ETT). Cells expressing ETT efficiently retain and accumulate ET at high doses, whereas cells lacking ETT do not accumulate ET because of the impermeability of the plasma membrane to this substance. The physiological role of ET as a substrate for rat Octn1 was confirmed in rat Octn1-transfected HEK293 and PC12 cells [47]. Specific transport of ET by human OCTN1 was further validated in HeLa cells [48]. In another study using octn1 gene KO mice (octn1^−/−) and metabolome analysis, a complete deficiency of ET was observed in the mice, with significant reductions across all tissues, particularly in the small intestine [49]. Moreover, our recent study revealed that the expression and function of rat OCTN1 (rOCTN1) are suppressed by 1,25(OH)₂D₃, which elevates the plasma levels of ET and reduces its accumulation in rat tissues [50].

2.1.9. L-carnitine (L-CAR)

L-CAR was first identified as a substrate for OCT2 and MATEs [23,51]. Previous studies have shown a decrease in the urinary excretion of carnitine and acetylcarnitine in Oct1/Oct2 dKO mice compared to their WT counterparts, suggesting that carnitine functions as a substrate for OCT2 [51]; however, in vivo data on this function in human subjects are limited. Additionally, in an untargeted metabolomics study, the administration of pyrimethamine, a MATE inhibitor, led to a significant and consistent reduction in the urinary excretion of carnitine in both mice and healthy subjects [23]. When L-CAR was co-administered with pyrimethamine and metformin at micro- or therapeutic doses, the CL_R of carnitine was significantly reduced by 90%–94% in healthy individuals. Similarly, in mice, pyrimethamine administration resulted in a 62% reduction in carnitine CL_R. However, in vitro studies using HEK293 cells overexpressing MATE1 or MATE2-K showed no increase in carnitine uptake compared with controls. Thus, L-CAR remains a controversial biomarker for OCT2/MATE inhibitors.

OCTN2 is primarily responsible for transporting L-CAR into cells and plays a central role in its pharmacokinetics [52]. While some studies have proposed L-CAR as a potential biomarker for OCTN2-mediated DDI, limited in vivo data have been published to support this claim [7,23,53,54]. Our recent study reported the effects of 1α,25-dihydroxyvitamin D₃ on the biodistribution and pharmacokinetics of L-CAR, confirming that it is a sensitive endogenous biomarker of OCTN2 [39].

2.2. Potential endogenous biomarkers of OATs and OATPs

2.2.1. Taurine

Taurine is an endogenous substrate of OAT1, with a K_m value of 379 ± 58 μM [55]. The K_i value of probenecid (a known inhibitor of OAT1 and OAT3) for the OAT1-mediated taurine uptake is 9.49 μM, which is comparable to the K_i values observed for common substrate drugs. The reduction in the CL_R of taurine caused by probenecid can be effectively explained by the K_i and geometric mean values of the unbound probenecid concentration. Thus, taurine has the potential as a tool for assessing pharmacokinetic DDI involving OAT1 [55]. Additionally, another study showed that in the presence of probenecid, the urinary excretion (X_{e (0–24 h)}) and CL_R of taurine decreased significantly by 3.2- and 3.4-fold, respectively [56].

2.2.2. Pyridoxic acid (PDA) and homovanillic acid (HVA)

The endogenous substrates PDA and HVA were recently identified as being transported by OAT1/3 through DDI investigations and untargeted metabolomics analysis in cynomolgus monkeys administered 40 mg/kg of probenecid [57]. In a subsequent clinical study involving 14 healthy volunteers who received probenecid, a significant increase in the AUC_0–∞ and a decrease in CL_R for both PDA and HVA were observed. However, the data suggested that endogenous plasma PDA may serve as a more effective indicator of OAT1/3 activity than HVA [58], with similar findings reported previously [56]. Additionally, PDA and HVA were validated as endogenous biomarkers of OAT1/3 using a population pharmacokinetic model [59]. This study provided empirical evidence supporting the use of plasma PDA data to detect OAT1/3 inhibitors with varying potencies, from weak to strong. Based on several parameters, PDA demonstrated greater robustness than HVA as a biomarker for OAT1/3 [59].

2.2.3. Glycochenodeoxycholate sulfate (GCDCA-S)

Unlike PDA and HVA, GCDCA-S exhibits greater selectivity, showing an affinity exclusively for OAT3 [56] with K_m values of 64.3 ± 3.9 μM [55]. The K_i values of probenecid for OAT3-mediated uptake of GCDCA-S was 7.4 μM, similar to that of other OAT3 substrate drugs. The plasma exposure of PDA and HVA showed greater sensitivity than GCDCA-S in terms of geometric mean ratio (GMR) of AUC_{0–24 h} with probenecid versus without probenecid, reporting increases of 3.7, 2.1, and 1.9, respectively [56]. However, the GCDCA-S biomarker demonstrated the highest sensitivity among OAT biomarkers in terms of CL_R. The GMR of CL_R with probenecid versus without probenecid showed a decrease of 9.5 for GCDCA-S, compared to 6.0 for PDA, 3.0 for HVA, and 3.4 for taurine [55,56].

2.2.4. Hippuric acid (HA)

HA was initially identified as a substrate for rat OAT1 (rOAT1) with a K_m value of 28 μM, but not for rOAT3 [60]. Similar results were obtained in humans. In addition, p-aminohippurate (PHA), a potent rOAT1 inhibitor, preferentially inhibited HA uptake in kidney slices. The plasma concentration of HA in nephrectomized rats was found to be significantly higher than that in normal rats, while the kidney-to-plasma (Kp) ratio was significantly lower [60]. Furthermore, the CL_R of endogenous HA correlated more closely with PHA clearance than with creatinine clearance. In general, the kidney plays a primary role in the elimination of HA from the bloodstream via active tubular secretion. Thus, the CL_R of endogenous HA is a valuable biomarker for assessing alterations in renal secretion, particularly in cases of chronic renal failure, where the production of OAT proteins is reduced.

2.2.5. Kynurenic acid (KYNA)

KYNA has been identified as a novel biomarker for DDI involving OAT1 and OAT3 inhibition [61]. Probenecid treatment not only increased the C_max of KYNA above the basal level at predose but also elevated its AUC_{0–12 h} in comparison to the furosemide-alone group (both in probenecid-only and probenecid-plus-furosemide groups) [61]. The increase in plasma KYNA exposure was comparable to that of the specific substrate furosemide. Therefore, plasma KYNA has been proposed as a superior endogenous biomarker for DDI mediated by OAT1/3. A recent study by Liu et al. [62] indicated that KYNA serves as a substrate for OAT1/3 and OAT2 but does not interact with OCT2 or MATE1/2K. Notably, KYNA displays equivalent affinities for OAT1 and OAT3. In bile duct-cannulated cynomolgus monkeys, Oat1/3 inhibition was assessed using KYNA as a biomarker, showing C_max and AUC_{0–24 h} values approximately 11.6- and 3.7-fold higher, respectively, in the probenecid-treated group than in the control group. Furthermore, the CL_R of KYNA decreased by 3.2-fold following probenecid administration, although no change was observed in biliary clearance (CL_bile) under the same conditions.

2.2.6. 6β-hydroxycortisol (6β-OHC)

The hepatic enzyme CYP3A4 is responsible for synthesizing 6β-OHC, which is subsequently eliminated through urinary excretion. This compound is a significant endogenous marker of OAT3 [63]. Additionally, 6β-OHC has been identified as a substrate of MATE1 and MATE2-K. In OAT3 KO mice, the CL_R of 6β-OHC was significantly reduced by more than 2-fold compared to WT mice. Similarly, the uptake of 6β-OHC by human kidney slices was markedly reduced following the administration of probenecid. Probenecid administration also led to a significant increase in the AUC_0–∞ for 6β-OHC. However, the injection of pyrimethamine, a strong inhibitor of MATEs, had no discernible effect on 6β-OHC levels. In summary, 6β-OHC is a viable endogenous biomarker for investigating DDI involving OAT3.

2.2.7. Coproporphyrin I and III (CPI and CPIII)

Active uptake of CPI and CPIII was demonstrated in HEK293 cells expressing human organic anion-transporting polypeptide 1B1/3 (OATP1B1/3) [64]. Active transport has also been observed in hepatocytes of both humans and monkeys. Rifampin (RIF), an OATP inhibitor (15 mg/kg, oral), increased the AUC_0–∞ of CPI and CPIII by 2.7- and 3.6-fold, respectively, whereas cyclosporine A (CsA; 100 mg/kg, oral) increased the AUC_0–∞ by 2.6- and 5.2-fold, respectively. In monkeys, the urinary excretion of both isomers increased by 1.6–4.3-fold with increasing systemic exposure. In Oatp1a/1b KO mice, CP concentrations in the plasma and urine were significantly higher (7.1–18.4-fold; P < 0.001) than those in WT animals, consistent with the changes observed in plasma rosuvastatin levels (14.6-fold increase). Another study revealed that changes in CPI and CPIII plasma levels indicate moderate (2–3-fold AUC increase) and strong (≥5-fold AUC increase) OATP1B inhibition in clinical settings [65]. Similarly, in monkeys, RIF significantly elevated the AUC_0–∞ for CPI and CPIII compared to that before RIF administration [66]. Furthermore, a previous study reported a dose-dependent increase in the AUC ratios of CPI-d8, CPI, and CPIII in monkeys [67]. RIF significantly decreased the liver and intestine uptake of [³H]-CPI while increasing kidney uptake in mice. Consistent with the findings in South Asian Indians, the AUC_{0–24 h} values of CPI and CPIII in black, white, and Hispanic populations were elevated compared with predose levels [68]. Additionally, baseline CP concentrations in individuals with the SLCO1B1 c.521 TT genotype were similar to those in individuals with the SLCO1B1 c.521 TC genotype, which is consistent with findings from previous studies using probe substrates. However, individuals with the SLCO1B1 c.388AG and c.388 GG genotypes had lower CPI concentrations than those with the SLCO1B1 c.388AA genotype.

Baseline plasma CPI levels have been linked to the severity of chronic kidney disease (CKD). Notably, unlike the AUCR of pitavastatin, a clinical OATP1B probe, that of CPI, with or without RIF, exhibited a larger difference between CKD patients with CKD and those without renal impairment. A study investigating the mechanisms underlying altered baseline CPI in patients with CKD demonstrated that the CPI model successfully predicted elevated baseline and RIF-mediated AUCR [69]. Collectively, these findings provide strong evidence that hepatic OATPs play a critical role in CPI distribution in animal models, indicating that CPI is a reliable and specific endogenous biomarker of OATP inhibition.

2.2.8. Bilirubin and bile acids

In rats administered 20 and 80 mg/kg of RIF, the plasma levels of bilirubin glucuronides increased then progressively declined, nearly returning to baseline levels within 24 h following administration [70]. An in vivo study using valsartan as an rat organic anion-transporting polypeptide (rOATP) substrate confirmed that RIF inhibits rOATPs. Another study found that Oatp 1a/1b-KO mice (Slco1a/1b^−/− mice) exhibit significantly elevated plasma concentrations of bilirubin glucuronides and unconjugated bile acids [71]. Following intravenous (IV) or oral administration of the OATP substrates methotrexate and fexofenadine, Slco1a/1b^−/− mice showed significantly reduced hepatic uptake and, consequently, increased systemic exposure. Notably, these KO mice did not exhibit any differences in the intestinal absorption of the substrates. Further investigation revealed that RIF acts as a potent and selective inhibitor of Oatp1a/1b, exerting regulatory effects on the pharmacokinetics of methotrexate. These findings suggest that Oatp1a/1b transporters play a crucial role in the hepatic reuptake of conjugated bilirubin as well as the absorption of unconjugated bile acids.

2.2.9. Tetradecanedioate (TDA) and hexadecanedioate (HDA)

Recent genome-wide association studies have identified TDA and HDA as metabolites that are strongly correlated with genetic variations in the OATP1B1 transporter. This finding is based on data from a clinical DDI trial involving the administration of pravastatin and CsA. The results were further supported by the observation that plasma levels of TDA and HDA increased following the administration of CsA, an OATP1B1 inhibitor [72]. Additionally, HEK293 cells overexpressing OATP1B1 demonstrated significantly higher uptake of TDA and HDA than cells transfected with an empty vector. A recent study confirmed that TDA and HDA are suitable endogenous biomarkers for assessing OATP function in humans [73].

2.2.10. Dehydroepiandrosterone sulfate (DHEAS)

DHEAS is an important endogenous circulating steroid hormone derived from dehydroepiandrosterone and is produced exclusively by the adrenal glands. It has been identified as an OATP1B1 and OATP1B3 substrate in humans [74,75]. In a study where RIF (30 mg/kg, IV) was administered to rats, the volume of distribution (V_d) and K_p,liver of DHEAS decreased by 4.2- and 8.2-fold, respectively, while the plasma AUC_0–∞ of DHEAS increased by 4.2-fold [75]. Increased plasma exposure of DHEAS due to RIF is thought to be primarily caused by the inhibition of hepatic OATPs, as the CL_bile to liver exposure remained unaltered. In vitro experiments have shown that both human and monkey hepatocytes exhibit time- and temperature-dependent uptake of DHEAS, which is inhibited by RIF, further suggesting the involvement of OATPs in DHEAS uptake. In addition, in vivo studies demonstrated a significant increase in the AUC_0–∞ and C_max of DHEAS after RIF administration at a dose of 10 mg/kg [74]. These findings suggest that DHEAS holds strong potential as an endogenous marker for evaluating hepatic OATP function and predicting OATP-related DDI.

2.3. Potential endogenous biomarkers of BCRP and MRP2

2.3.1. Riboflavin (vitamin B2)

To date, riboflavin is the only potential endogenous probe of BCRP that has been reported [76,77]. A previous study observed significant alterations in 130 metabolites in mice lacking Bcrp and P-gp. Specifically, the plasma concentration of riboflavin was significantly elevated in mice with a single KO of Bcrp, as well as in mice with a dKO of Bcrp and P-gp. However, no significant increase in riboflavin levels was observed in mice with a single P-gp KO. Inhibition studies in mice using elacridar, a potent dual inhibitor of BCRP and P-gp, demonstrated a significant and dose-dependent increase in riboflavin AUC_0–∞ in the elacridar-treated group compared with that in the control group. Similarly, a 1.7-fold increase in riboflavin levels was observed in monkeys treated with ML753286, which was comparable to that of the BCRP substrate, sulfasalazine. In vitro investigations have confirmed that riboflavin, as opposed to P-gp, is a specific substrate of BCRP in both monkey and human cells. These findings support the use of riboflavin as a feasible endogenous indicator of assessing BCRP function. However, further research is needed to evaluate the potential of riboflavin as a hematological biomarker for BCRP in humans.

2.3.2. CPI and CPIII

Current studies have focused on CPI and CPIII, collectively referred to as CPs, as intrinsic OATP biomarkers, while previous studies have provided evidence that CPs can serve as substrates for MRP2 [78]. An investigation compared the mechanisms of endogenous CP excretion in WT Wistar rats and transport-deficient (TR^–) rats with impaired MRP2 function. Quantitative targeted absolute proteomic analysis revealed no differences in OATP expression between the two groups. Using LC-MS/MS, CPI and CPIII concentrations were found to be higher in TR^– livers than in WT livers, and the estimated CL_bile was 75- and 840-times lower in TR^– livers, respectively. Additionally, TR^– rats had 14-fold lower CPIII concentrations in their feces than WT rats, but showed no significant difference in CPI concentrations. These findings indicate that the loss of MRP2 function significantly affected the distribution of CPs in TR^– rats.

3. Recent advances in MS-based bioanalytical methods for endogenous biomarker quantification

The accurate quantification of biomarkers is essential for understanding biological processes, diagnosing diseases, monitoring treatment efficacy, and predicting patient outcomes [79]. Various approaches have been used for biomarker quantification, each with its own advantages and limitations. The choice of the quantification method depends on factors such as the nature of the biomarker, sample type, required sensitivity and specificity, throughput, and available resources [80]. The commonly used methods include immunoassays, nuclear magnetic resonance spectroscopy, polymerase chain reaction-based techniques, electrochemical biosensors, lateral flow assays, microarray technology, and MS. Among these, MS is particularly valuable because it enables the simultaneous analysis of multiple biomarkers across a wide dynamic range. In bioanalysis, LC-MS/MS is remarkably useful for studying endogenous biomarkers because it can detect low concentrations of analytes in complex biological samples, such as blood, urine, tissue, and cerebrospinal fluid (CSF). Using highly selective sample preparation methods and optimized chromatographic conditions, LC-MS/MS can achieve exceptional sensitivity and minimize matrix effects, allowing for the accurate quantification of endogenous biomarkers even at trace levels.

The workflow of LC-MS/MS bioanalysis typically involves several key steps, including sample preparation, chromatographic separation, mass spectrometric detection, and data analysis [81]. In the sample preparation stage, biological samples are extracted, purified, and often derivatized to enhance analyte stability and ionization efficiency. Common techniques for determining potential endogenous biomarkers for transporters in biological fluids include protein precipitation, liquid-liquid extraction (LLE), and solid-phase extraction (SPE). Table 1 [22,23,29,30,[36], [37], [38], [39],41,42,50,76,[82], [83], [84], [85], [86], [87], [88], [89], [90], [91], [92], [93], [94], [95], [96], [97]] summarizes LC-MS/MS methods for quantifying potential endogenous biomarkers in biological fluids for OCTs, MATEs, and OCTNs. Table 2 [[55], [56], [57], [58],61,62,[65], [66], [67], [68],73,74,86,[88], [89], [90], [91], [92], [93],[98], [99], [100], [101], [102], [103], [104], [105], [106], [107], [108], [109], [110], [111], [112], [113], [114], [115]] summarizes LC-MS/MS methods for determining potential endogenous biomarkers in biological fluids associated with OATs and OATPs. Table 3 [62,76,92,116] highlights LC-MS/MS methods for determining potential endogenous biomarkers in biological fluids for BCRP and MRP2.

Table 1.

Summary of liquid chromatography-tandem mass spectrometry (LC-MS/MS) methods for determining potential endogenous biomarkers in biological fluids for organic cation transporter (OCTs), multidrug and toxic compound extrusion (MATEs), and organic cation/carnitine transporter (OCTNs).

Biomarkers (transporters)	Matrix	Sample preparation&IS	Column	Mobile phase	MRM transition (m/z)	LLOQ	Linear range	Ref.
Thiamine (OCT1, OCT2, MATEs)	Mouse plasma	Protein precipitation (0.1% FA in MeOH)	Synergi Hydro-RP-80A (4 μm particle, 50 mm × 4.6 mm)	Gradient elution of A (0.01% trifluoroacetic acid in water, v/v) and B (0.01% trifluoroacetic acid in water, v/v) at 700 μL/min	265.1  →  122.2	N.S	N.S	[22]
		IS: Carbutamide
	Human plasma and urine; mouse urine, plasma, and kidney	Protein precipitation (ACN:MeOH, 9:1, v/v)	XBridge HILIC (3.5 μm, 4.6 mm × 50 mm)	Gradient elution of A (10 mM ammonium acetate pH 5.0) and B (ACN)	265  →  122	1 ng/mL	N.S	[23]
		IS: Thiamine-d3
	Human plasma	Protein precipitation (ACN)	Acquity UHPLC BEH HILIC (2.1 mm × 100 mm, 1.7 μm)	Gradient elution of A (20 mM ammonium formate with 0.1% FA in 9:1, water:ACN) and B (20 mM ammonium formate with 0.1% FA in 1:9 water:ACN)	265.1  →  122.1	0.5 ng/mL	0.5–100 ng/mL	[30]
		IS: Thiamine-d3
IBC (OCT1)	Human plasma and urine	Protein precipitation (mixture of 10% MeOH and 90% ACN, v/v)	Brownlee SPP RP-Amide column (4.6 mm × 100 mm, 2.7 μm) with a Phenomenex C18 pre-column (4 mm × 2 mm)	0.1% (v/v) FA, 0.43% (v/v) MeOH, and 2.57% (v/v) ACN	232.1  →  85.0	N.S	N.S	[29]
		IS: IBC-d6
	Human plasma	Protein precipitation (ACN)	Acquity UHPLC BEH HILIC (2.1 mm × 100 mm, 1.7 μm)	Gradient elution of A (20 mM ammonium formate with 0.1% FA in 9:1, water:ACN) and B (20 mM ammonium formate with 0.1% FA in 1:9, water:ACN)	232.1  →  85.03	0.5 ng/mL	0.5–500 ng/mL	[30]
		IS: IBC-d3
	Mouse plasma, monkey plasma	Protein precipitation (ice-cold acidic ACN)	BEH Amide (1.7 μm, 100 mm × 2.1 mm)	Gradient elution of A (10 mM ammonium formate in water) and B (0.1% FA in ACN) at 0.7 mL/min	231.9  →  173.0	1 ng/mL	1–500 ng/mL	[76]
		IS: IBC-d3
Creatinine (OCT2, MATEs)	Human plasma	Protein precipitation (ACN)	Acquity UHPLC BEH HILIC (2.1 mm × 100 mm, 1.7 μm)	Gradient elution of A (20 mM ammonium formate with 0.1% FA in 9:1, water:ACN) and B (20 mM ammonium formate with 0.1% FA in 1:9, water:ACN)	114.1  →  44.05	20 ng/mL	20–20000 ng/mL	[30]
		IS: Creatinine-d3
	Rat plasma	Protein precipitation (ACN) with	Synergi™ 4 μm polar-RP 80A column (150 mm × 2.0 mm, 4 μm) conjugated with a guard column (SecurityGuard™, 4.0 mm × 3.0 mm)	Isocratic elution of A (0.1% FA in water) and B (ACN) (90:10, v/v) at 0.2 mL/min	114.1  →  85.7	10 ng/mL	10–5000 ng/mL	[39]
		IS: Metformin
	Human plasma and urine, monkey plasma	Protein precipitation (ACN)	PC HILIC (3.0 μm, 2.0 mm × 150 mm)	Isocratic elution of A (10 mM ammonium formate and 0.1% FA in 20% ACN/80% water) and B (10 mM ammonium formate and 0.1% FA in 95% ACN/5% water), 25:75, v/v at a flow rate of 0.4 mL/min	114.1  →  44.1	N.S	N.S	[41]
		IS: Creatinine-d3
	Rat plasma	Protein precipitation (ACN)	Atlantis HILIC Silica (2.1 mm × 150 mm, particle size 5 μm)	Isocratic elution of 0.1% FA in water and 0.1% FA in ACN (13:87, v/v) at a flow rate of 0.5 mL/min	113.98  →  86.1	N.S	N.S	[82]
		IS: Creatinine-d3
	Human plasma	Protein precipitation (ACN)	Luna Silica (4.6 mm × 150 mm, 5 μm)	Isocratic elution of 2 mM ammonium acetate in water and ACN (45:55, v/v) at a flow rate of 1.5 mL/min)	114  →  44	0.5 μg/mL	0.5–150 μg/mL	[83]
		IS: Creatinine-d3
	Human urine	Protein precipitation (ACN)	Zorbax Hilic plus silica (50 mm × 2.1 mm, 1.8 μm)	Gradient elution of 5 mM ammonium formate and ACN at a flow rate of 0.7 mL/min	114  →  44	10 ng/mL	10–10000 ng/mL	[84]
		IS: 13C32H415N1-creatinine
	Mouse plasma and urine	Protein precipitation (ethanol), LLE with chloroform (for plasma)	YMC C18 (2.0 mm × 150 mm, 5 μm)	Gradient elution of A (0.1% FA in water) and B (ACN) at a flow rate of 200 μL/min	114  →  44	0.013 mg/dL	N.S	[85]
		IS: [2H3]-creatinine
m1A (OCT2, MATE2-K)	Mouse plasma, urine, and kidney; human plasma and urine; monkey plasma	Protein precipitation (ACN)	CAPCELL PAK ADME (3.0 μm, 2.0 mm × 50 mm)	Gradient elution of A (2 mM ammonium acetate, pH 5.0) and B (2 mM ammonium acetate in 90% MeOH) at a flow rate of 0.4 mL/min	282.2  →  150.0	N.S	N.S	[41]
		IS: Creatinine-d3
Tryptophan (OCT2)	Human urine	Protein precipitation (ice-cold MeOH and 0.13 N HCl)	Reversed-phase Atlantis dC18 (2.1 mm i.d. × 150 mm, 3 μm particle size)	Isocratic elution of ACN and water containing 0.1% FA (30:70, v/v) at 0.2 mL/min	205  →  188	100 ng/mL	100–200000 ng/mL	[42]
		IS: Tryptophan-d5
	Mouse plasma, CSF, and brain	Protein precipitation (ice-cold MeOH)	GRACE VisionHT C18 column (100 mm × 2.1 mm, 3 μm)	Gradient elution of A (0.1% FA in water add 0.01% TFA) and B (0.1% FA in MeOH add 0.01% TFA) at a flow rate of 400 μL/min	205.1  →  118.0	2.5 nM	Plasma: 100–250000 nM CSF: 100–100000 nM Brain: 10–25000 nM	[86]
		IS: Tryptophan-d5
	Human plasma	Protein precipitation (10% TFA)	Synergi Polar-RP column (50 mm × 2.0 mm, 5 μm)	Gradient elution of 0.1% FA in water and 0.1% FA in MeOH	205.1  →  170.1	0.5 μg/mL	0.5–50 μg/mL	[87]
		IS: Tryptophan-d5
	Rat plasma, urine, and hippocampus	Protein precipitation (ACN)1	Ultimate XB-C18 column (100 mm × 2.1 mm, 5 μm)	Gradient elution of A (0.1% FA and 5 mM ammonium acetate in ACN) and B (aqueous with 0.1% FA and 5 mM ammonium acetate) at a flow rate of 0.3 mL/min	205.1  →  188.2	30 ng/mL	Plasma: 40–12000 ng/mL Urine: 30–3000 ng/mL Hippocampus: 30–9000 ng/mL	[88]
		IS: 2-cloro-l-phenylalanine
	Human plasma	Protein precipitation (ACN)	Luna NH2 (150 mm × 1.00 mm, 3 μm, 100 Å) with a guard column (Security Guard equipped with a Luna NH2 4 mm × 2.0 mm cartridge)	Gradient elution of 5 mM ammonium acetate in water (pH 9.5) and ACN at a flow rate 35 μL/min	203  →  159	0.5 μg/mL	0.5–50 μg/mL	[89]
		IS: Tryptophan-d5
	Rat plasma	SPE utilizing 3 cc waters HLB cartridges, placed on a vacuum elution manifold	Restek C18 aqueous (100 mm × 2.1 mm)	Gradient elution of 0.05% ammonium formate in water, (pH 5.5) and ACN at a flow rate 0.2 mL/min	204.9  →  188.1	0.0352 μM	0.070–1.126 μM	[90]
		IS: Ethyl-4-hydroxy-2-quinolinecarboxylate
	Human CSF and serum	Protein precipitation (ice-cold ACN with 0.1% FA)	Pentafluorophenyl (100 Å, 100 mm × 2.1 mm, 2.6 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in MeOH) at 300 μL/min	205.1  →  118.1	CSF (8.256 nM), serum (3.709 nM)	CSF: 250–8000 nM; Serum: 6.25–200 μM	[91]
		IS: Tryptophan-d5
	Human plasma	Protein precipitation with TCA in water	Zorbax stable-bond C8 reversed-phase (150 mm × 4.6 mm, 3.5 μm)	Gradient elution of A (650 mM acetic acid), B (100 nM HFBA in A), and C (90% ACN in water) at 1 mL/min	206.3  →  189.1	0.5 μM	0.5–400 μmol/L	[92]
		IS: 2H5-tryptophan
	Rat plasma and brain	Protein precipitation with ACN	Kromasil C18 (2.1 mm × 150 m)	Gradient elution of A (0.1% FA and 2.0 nM ammonium acetate in water) and B (ACN) at 0.2 mL/min	309.4  →  263.1 (benzoylated derivatization)	0.1 ng/mL (0.7 nM)	Brain: 0.05–39.2 μM, plasma: 0.01–13.1 μM	[93]
		IS: Caffeic acid
	Rat serum and brain	Protein precipitation with ACN	ACQUITY UHPLC® BEH Amide (50 mm × 2.1 mm, 1.7 μm) with a VanGuard column (5 mm × 2.1 mm, 1.7 μm)	Gradient elution of A (20 mmol/L ammonium formate and 0.25% FA) and B (ACN/water ( 93:7, v/v) containing 20 mmol/L ammonium formate and 0.25% FA) at a flow rate of 0.5 mL/min	205.1  →  188.0	104.2 ng/mL	104.2–3333.4 ng/mL	[94]
		IS: L-2-aminobutanoic acid (α-ABA)
	Rat plasma and urine	Protein precipitation with ACN	Acquity UHPLC BEH Amide (100 mm × 2.1 mm, 1.7 μm)	Gradient elution of A (0.2% FA in water) and B (0.2% FA in ACN) at a flow rate of 0.3 mL/min	204.9  →  187.9	0.8 μg/mL	0.8–40 μg/mL	[95]
		IS: l-phenylalanine-d5 (L-Phe-d5)
Dopamine (OCT2)	Mouse plasma, CSF, and brain	Protein precipitation (ice-cold MeOH)	GRACE VisionHT C18 (100 mm × 2.1 mm, 3 μm)	Gradient elution of A (0.1% FA in water added 0.01% TFA) and B (0.1% FA in MeOH added 0.01% TFA) at a flow rate of 400 μL/min)	154.1  →  137.0	0.25 nM	Plasma: 2.5–10000 nM CSF: 1–10000 nM Brain: 1–1000 nM	[86]
		IS: Dopamine-d4
	Rat plasma and brain	Protein precipitation with ACN	Kromasil C18 (2.1 mm × 150 mm)	Gradient elution of A (0.1% FA and 2.0 nM ammonium acetate in water) and B (ACN) at 0.2 mL/min	466.0  →  241.2 (benzoylated derivatization)	0.5 ng/mL (3.3 nM)	Brain: 0.03–26.1 μM, plasma: 0.049–4.9 μM	[93]
		IS: Caffeic acid
	Rat serum and brain	Protein precipitation with ACN	ACQUITY UHPLC® BEH Amide (50 mm × 2.1 mm, 1.7 μm) with a VanGuard column (5 mm × 2.1 mm, 1.7 μm)	Gradient elution of A (20 mmol/L ammonium formate and 0.25% FA and B (20 mmol/L ammonium formate and 0.25% FA in ACN/water (93:7, v/v)) at a flow rate of 0.5 mL/min	154.0  →  91.1	14.6 ng/mL	14.6–468.8 ng/mL	[94]
		IS: 2,3-Dihydroxybenzoic acid (DHBA)
NMN (OCT2, MATE1/2K)	Human plasma	Protein precipitation (ACN)	Acquity UHPLC BEH HILIC (2.1 mm × 100 mm, 1.7 μm)	Gradient elution of A (20 mM ammonium formate with 0.1% FA in 9:1 water:ACN, (v/v)) and B (20 mM ammonium formate with 0.1% formic acid in 1:9 water:ACN, (v/v))	137.1  →  94.1	1 ng/mL	1–1000 ng/mL	[30]
		IS: NMN-d3
	Human plasma and urine	Protein precipitation (ACN)	PC HILIC (3 μm, 2.0 mm × 150 mm)	Isocratic elution of 10 mM ammonium formate and ACN (30:70, v/v) at a flow rate of 0.4 mL/min	Mass-to-charge ratio of 137.0	10 pg/mL	10–20000 pg/mL	[36]
		IS: [14C]-metformin
	Human plasma, urine	Protein precipitation (ACN)	Cogent Diamond Hydride (150 mm × 2.1 mm, 4 μm) fitted with a Cogent Diamond Hydride precolumn (20 × 2.0 mm, 4 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at a flow rate of 0.3 mL/min	137.1  →  94.0	3.1 nM (plasma) 25 nM (urine)	Plasma: 3.1–500 nM; Urine: 25–1000 nM	[37]
		IS: 1-Methyl-d3-nicotinatmide iodide
	Human plasma, urine	Protein precipitation (ACN)	EC 250/2 Nucleodur 100-3 HILIC with a guard column	Isocratic elution of 8 mM ammonium formate in ACN-water (80:20, v/v) adjusted to pH 4 with FA	137  →  94	50 ng/mL (urine), 2.5 ng/mL (plasma)	N.S	[38]
		IS: NMN-d3
	Human plasma and urine, rat plasma and urine	Deproteinized using 10% TCA followed by ether extraction	Hypersil C18-BDS (150 mm × 2.0 mm, 3 μm)	Gradient elution of A (10 mM nonafluoropentanoic acid (NFPA) in water) and B (100% ACN) at a flow rate of 0.2 mL/min	[M+H]+ at m/z 137	N.S	N.S	[96]
		IS: 2-Cloroadenosine
	Human plasma, urine	Protein precipitation (ACN)	Acquity UHPLC Cortecs HILIC (2.1 mm × 50 mm, 1.6 μm)	Gradient elution of A (20 mM ammonium formate with 0.1% FA in 90% water and 10% ACN) and B (20 mM ammonium formate with 0.1% FA in 10% water and 90% ACN) at 0.6 mL/min	137.0  →  93.9	0.1 ng/mL	Plasma: 0.1–1000 ng/mL; Urine: 0.5–500 ng/mL	[97]
		IS: NMN-d3
ET (OCTN1)	Rat plasma, urine, and tissues	Protein precipitation (ACN)	Synergi™ 4 μm polar-RP 80A column (150 mm × 2.0 mm, 4 μm) conjugated with a guard column (SecurityGuard™, 4.0 mm × 3.0 mm)	Gradient elution of A (0.1% FA in water) and B (ACN) at 0.2 mL/min	230.2  →  127.1	2 ng/mL	2–10000 ng/mL	[50]
		IS: Cefdinir
L-CAR (MATEs, OCTN2)	Human plasma and urine; mouse urine, plasma, and kidney	Protein precipitation (ACN: MeOH = 9:1, v/v)	XBridge HILIC (4.6 mm × 50 mm, 3.5 μm,)	Gradient elution of A (10 mM ammonium acetate (pH 5.0)) and B (ACN)	162.1  →  103.0	1 ng/mL	N.S	[23]
		IS: Carnitine-d9
	Human plasma	Protein precipitation (ACN)	Acquity UHPLC BEH HILIC (2.1 mm × 100 mm, 1.7 μm)	Gradient elution of A (20 mM ammonium formate with 0.1% FA in 9:1 water:ACN, (v/v)) and B (20 mM ammonium formate with 0.1% FA in 1:9 water:ACN (1:9, v/v))	162.1  →  85.0	50 ng/mL	50–20000 ng/mL	[30]
		IS: Carnitine-d3
	Rat plasma	Protein precipitation (ACN) with	Synergi™ 4 μm polar-RP 80A column (150 mm × 2.0 mm, 4 μm) conjugated with a guard column (SecurityGuard™, 4.0 mm × 3.0 mm)	Isocratic elution of A (0.1% FA in water) and B (ACN) (95:5, v/v) at 0.2 mL/min	162.0  →  84.8	2 ng/mL	2–5000 ng/mL	[39]
		IS: Metformin

OCTN: organic cation/carnitine transporter, IBC: isobutyryl-carnitine, m¹A: N1-methyladenosine, NMN: N1-methyl nicotinamide, ET: ergothioneine, L-CAR: L-carnitine, CSF: cerebrospinal fluid, IS: internal standard, ACN: acetonitrile, FA: formic acid, TCA: trichloroacetic acid, TFA: trifluoroacetic acid, LLE: liquid-liquid extraction, SPE: solid phase extraction, BEH: ethylene bridged hybrid, HILIC: hydrophilic interaction liquid chromatography, MRM: multiple reaction monitoring, LLOQ: lower limit of quantification, N.S: data not shown.

Table 2.

Summary of liquid chromatography-tandem mass spectrometry (LC-MS/MS) methods for determining potential endogenous biomarkers in biological fluids for organic anion transporter (OATs) and organic anion-transporting polypeptides (OATPs).

Biomarkers (transporters)	Matrix	Sample preparation & IS	Column	Mobile phase	MRM transition (m/z)	LLOQ	Linear range	Ref.
Taurine (OAT1/3)	Human plasma and urine; rat plasma and urine	Protein precipitation (ACN/MeOH 9:1, v/v) IS: [D5]-taurine	XBridge HILIC (50 mm × 4.6 mm, 3.5 μm)	Gradient elution of A (5 mM ammonium acetate) and B (ACN) at a flow rate of 1 mL/min	Negative mode 123.9  →  79.9	N.S	N.S	[55]
	Human plasma and urine	Protein precipitation with ACN IS: [D5]-taurine	Acquity UHPLC Amide; 2.1 mm × 50 mm, 1.7 μm column)	Gradient elution of A (0.1 M ammonium acetate in water) and B (ACN)	Negative mode	200 ng/mL for plasma and 100 ng/mL for urine	N.S	[56]
	Rat plasma	Protein precipitation (ACN) with IS: sulfanilic acid	Zorbax SB-Aq colμmn (100 mm × 2.1 mm, 3.5 μm)	Isocratic elution of 0.1% FA in water and MeOH (90:10, v/v)	Negative mode 124.1  →  80.0	2 μg/mL	2–1000 μg/mL	[98]
	Rat urine, bile, and feces	Protein precipitation (ACN) IS: Sulfanilic acid	Atlantis HILIC Silica (150 mm × 2.1 mm, 3 μm)	Isocratic elution of 5 mM ammonium formate and 0.2% FA (75:25, v/v) at 0.3 mL/min	Negative mode 124.1  →  80.0	2 μg/mL (urine); 1 μg/mL (bile and feces)	Urine: 2–500 μg/mL; Bile: 1–500 μg/mL; Feces: 1–250 μg/mL	[99]
GCDCA-S (OAT1/3)	Human plasma and urine Rat plasma and urine	Protein precipitation (ACN/MeOH, 9:1, v/v) IS: [D5]-GCDCA-S	ZORBAX SB-C18 ( 50 mm × 4.6 mm, 3.5 μm)	Gradient elution of 0.1% FA in water and ACN at a flow rate of 1.2 mL/min	Negative mode 528.1  →  448.3	N.S	N.S	[55]
	Human plasma and urine	Protein precipitation with ACN IS: [D5]-GCDCA-S	Acquity UHPLC HSS T3 (2.1 mm × 50 mm, 1.8 μm)	Gradient elution of A (0.1% ammonium acetate in water) and B (methanol)	Negative mode	2.5 ng/mL (plasma), 10 ng/mL (urine)	N.S	[56]
GCDCA-S (OATP1B1, OATP1B3)	Human plasma	Protein precipitation (ACN) IS: GS-A	Acquity C18 BEH (1.7 μm, 2.1 mm × 100 mm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at flow rate of 0.5 mL/min	Negative mode 448.3  →  74	1 nM	1–50000 nM	[61]
	Human plasma	Protein precipitation (ACN/MeOH, 1:1, v/v) IS: Taurocholate-d5	CAPCELL PAK C18 MGII (2 mm i.d. × 50 mm)	Gradient elution of A (5 mM ammoniumacetate in water) and B (MeOH) at 0.4 mL/min	Negative doubly charged 263.6  →  74.1	0.005 μM	N.S	[100,101]
	Mouse plasma and human plasma	Protein precipitation (MeOH) IS: GCDCA-3S-d5	Accucore aQ (50 mm × 2.1 mm, 2.6 μm) with a C18 AQUASIL guard cartridge (2.1 mm × 10 mm, 3 μm)	Gradient elution of A (2 mmol ammonium acetate in water) and B (100% MeOH) at flow rate 0.4 mL/min	Negative mode 528.4  →  448.3	0.5 ng/mL	0.5–100 ng/mL (mouse); 0.5–1000 ng/mL (human)	[102]
	Human plasma	Protein precipitation (ACN) IS: Tolbutamide	Acquity UHPLC BEH C18 (2.1 mm × 100 mm, 1.7 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at 0.5 mL/min	Negative mode 528.3  →  448.3	0.5 ng/mL	0.5–500 ng/mL	[103]
	Rat plasma and liver	Protein precipitation (MeOH) IS: Nor-desoxycholic acid	CORTECS ® C18 (4.6 mm × 100 mm, 2.7 μm)	Gradient elution	Negative mode 528.4  →  448.1	1.1 ng/mL	1.1–137.5 ng/mL	[104]
CPI and CPIII (OATP1B)	Human plasma	Oasis MAX (mixed-mode anion exchange, 96-well) SPE plate following addition of 5% FA in MeOH/ACN (60:40, v/v) IS: 15N4-CPI and CPIII sodium bisulfate salt	Ace Excel 2C18 PFP, (2.1 mm × 150 mm, 3 μm)	Gradient elution of A (10 mM ammonium formate containing 0.1% FA) and B (ACN) at 0.6 mL/min	Positive mode 655.3  →  596.3	20 pg/mL	N.S	[65]
	Human plasma and urine	Protein precipitation (12 M FA and ethyl acetate for plasma and 6 M FA for urine) IS: 15N4-CPI (2.5 nM)	Ace Excel 2C18 UHPLC (2.1 mm × 150 mm, 1.7 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in water:ACN, 2:98, v/v)	Positive mode 655.3  →  596.3	N.S	N.S	[66]
	Monkey plasma and urine; mouse plasma and urine	Protein precipitation (ice-cold ethyl acetate) IS: 15N4-CPI and CPIII-d8	Waters Acquity UHPLC BEH C18 (1.7 μm, 100 mm × 2.1 mm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at 0.6 mL/min	Positive mode 655.3  →  596.3	0.1 nM	0.1–500 nM	[67]
	Human plasma	Protein precipitation (6 M FA), SLE using an Biotage ISOLUTE SLE+ 96-well plate procured IS: 15N4-CPI and CPIII-d8	Acquity UHPLC HSS T3 (2.1 mm × 50 mm, 1.8 μm) with a VanGuard BEH C18 guard column (2.1 mm × 10 mm, 1.7 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in water:ACN, 2:98, v/v) at 0.55 mL/min	Positive mode 655.4  →  596.3	0.078 nM	N.S	[68]
	Human plasma	Protein precipitation (ACN), SLE (Novum SLE 96 well plates), and LLE (acidified ethyl acetate and MTBE) IS: SIL-CPI and SIL-CPIII	Ace Excel 2C18 PFP (2.1 mm × 150 mm, 3 μm)	Gradient elution of A (10 mM ammonium formate containing 0.1% FA) and B (ACN) at 0.6 mL/min	Positive mode 655.3  →  596.3	20 pg/mL	0.02–100 ng/mL	[105]
	Human plasma and urine	LLE with ethyl acetate IS: 15N4-CPI in 12 M FA	Acquity UHPLC BEH C18 (2.1 mm × 100 mm, 1.7 μm)	Gradient conditions of A (0.1% FA in water) and B (0.1% FA in ACN)	Positive doubly charged 328.1  →  238.1	0.05 ng/mL (plasma) 1 ng/mL (urine)	0.05–10 ng/mL (plasma); 1–100 ng/mL (urine)	[101,106]
	Human plasma	Protein precipitation (6 M FA solution), SLE (SLE+ 200 μl 96-well supported liquid extraction (SLE) plate procured from Biotage IS: 15N4-CPI for CPI and CPIII-d8 for CPIII	Acquity™ HSS T3 UHPLC (2.1 mm × 50 mm, 1.8 μm) and a VanGuard BEH C18 guard column (2.1 mm × 10 mm, 1.7 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in water:ACN, 2:98, v/v) at 0.55 mL/min	Positive mode 655.4  →  596.3	0.078 nM	0.078–15 nM	[107]
	Human plasma	Automated-supported liquid extraction (SLE) IS: 15N4-CPI	Waters Acquity UHPLC BEH C18 (2.1 × 100 mm, 1.7 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at a flow rate of 0.8 mL/min	Positive mode doubly charged 328.1  →  238.1	100 pg/mL	100–5000 pg/mL	[108]
	Monkey plasma	Protein precipitation (6 M FA), SLE using Biotage Isolute SLE+ 96-well plate, LLE using ethyl acetate IS: 15N4-CPI and CPIII-d8	Acquity UHPLC BEH C18 (2.1 mm × 150 mm, 1.8 μm) with a VanGuard BEH C18 guard column (2.1 mm × 10 mm, 1.7 μm)	Gradient elution of A (0.1% FA) and B (0.1% FA in ACN) at 0.5 mL/min	Positive mode 655.4  →  596.3	0.078 nM	0.156–40 nM	[109]
TDA (OATPs)	Human plasma	Protein precipitation (MeOH) IS: TDA-d4	Atlantis d C18 (2.1 mm × 50 mm, 3 μm)	Gradient elution of A (10 mM ammonium formate with 2% ACN) and B (ACN) at a flow rate of 0.6 mL/min	Negative mode 257.2  →  239.0	4.1 nM	4.1–1000 nM	[73]
	Human plasma	Protein precipitation (methanol), SLE in 96-well polypropylene plate IS: TDA-d4	Cortecs C18+ (2.1 mm × 100 mm, 1.6 μm)	Gradient elution of (0.1% FA in water) and B (0.1% FA in ACN: MeOH, 90:10, v/v) at a flow rate of 0.5 mL/min	Negative mode 257.1  →  239.2	2.5 nM	2.5–1000 nM	[110]
HDA (OATPs)	Human plasma	Protein precipitation (MeOH) IS: HDA-d28	Atlantis d C18 column (2.1 mm × 50 mm, 3 μm)	Gradient elution of A (10 mM ammonium formate with 2% ACN) and B (ACN) at a flow rate of 0.6 mL/min	Negative mode 285.1  →  223.1	4.1 nM	4.1–1000 nM	[73]
	Human plasma	Protein precipitation (MeOH), SLE in 96-well polypropylene plate IS: HDA-d4	Cortecs C18+ (2.1 mm × 100 mm, 1.6 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN:MeOH, 90:10, v/v) at a flow rate of 0.5 mL/min	Negative mode 285.1  →  223.1	2.5 nM	2.5–1000 nM	[110]
DHEAS (OATPs)	Human plasma	Protein precipitation (MeOH) IS: DHEAS-d5	Waters Atlantis dC18 (2.1 mm × 50 mm, 3 μm)	Gradient elution of A (10 mM ammonium formate with 2% ACN) and B (ACN) at a flow rate of 0.6 mL/min	Negative mode 367.1  →  97.0	68 nM	68–50000 nM	[73]
	Monkey plasma	Protein precipitation (MeOH) IS: DHEAS-d6	Acquity UHPLC BEH C18 (1.7 μm particle size 2.1 mm × 30 mm)	Gradient elution of A (10 mM ammonium formate) and B (ACN) at a flow rate at 0.3 mL/min	Positive mode 367.1  →  97.0	N.S	N.S	[74]
PDA (OAT1/3)	Human plasma and urine	SPE on an Oasis MAX (mixed-mode anion exchange, 96-well) for plasma, protein precipitation (ACN) for urine IS: PDA-d3	Acquity UHPLC HSS T3 (2.1 mm × 50 mm, 1.8 μm)	Gradient elution of A (0.1% acetic acid in water) and B (ACN)	Negative mode	1 ng/mL (plasma) 10 ng/mL (urine)	N.S	[56]
	Monkey plasma	Protein precipitation (0.1% FA in MeOH) IS: Enalapril-d5	Acquity UHPLC BEH130 C18 (2.1 mm × 100 mm, 1.7 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at flow rate of 0.5 mL/min	Negative mode 182  →  138	5 ng/mL	5–5000 ng/mL	[57]
	Human plasma	Protein precipitation (MeOH) with IS: enalapril-d5	XBridge C18 (3.5 μm, 2.1 mm × 100 mm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at flow rate of 0.5 mL/min	Negative mode 182.1  →  108.0	5.6 nM	5.6–14000 nM	[58]
	Monkey plasma, urine, and bile	Protein precipitation IS: PDA-d3	Acquity UHPLC HSS T3 (2.1 mm × 100 mm, 1.8 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at flow rate 0.35 mL/min	Positive mode 182  →  138	N.S	N.S	[62]
	Human plasma	Protein precipitation (TCA in water) IS: PDA-d2	Zorbax stable-bond C8 reversed-phase (150 mm × 4.6 mm, 3.5 μm) with a similar guard column (12.5 mm × 4.6 mm, 5 μm)	Gradient elution of A (650 mM acetic acid), B (100 nM HFBA in A), and C (90% ACN in water) at 1 mL/min	Positive mode 184.1  →  148.0	0.05 nmol/L	0.05–400 nmol/L	[92]
	Human CSF	Protein precipitation (0.1% FA in ACN/MeOH (9:1, v/v)) IS: PDA-d3	2D system: First separation: Waters BEH C18 (100 mm × 2.1 mm 1.7 μm) and second separation: ACE C18-PFP (100 mm × 2.1 mm, 1.7 μm)	Gradient elution of A (0.1% FA in water) and B (MeOH) at 0.2 mL/min for the first separation, A (20 mmol/L ammonium formate with FA) and B (MeOH) for the second separation	Positive mode 184.1  →  148.1	5 nmol/L	5–200 nM	[111]
HVA (OAT1/3)	Human plasma and urine	SPE on an Oasis MAX (mixed-mode anion exchange, 96-well) for plasma, protein precipitation (ACN) for urine IS: HVA-d5	Acquity UHPLC HSS T3 (2.1 mm × 50 mm, 1.8 μm)	Gradient elution of A (0.1% acetic acid in water) and B (ACN)	Negative mode	2 ng/mL (plasma) 50 ng/mL (urine)	N.S	[56]
	Monkey plasma	Protein precipitation (0.1% FA in MeOH) IS: Enalapril-d5	Acquity UHPLC BEH130 C18 (2.1 mm × 100 mm, 1.7 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at flow rate of 0.5 mL/min	Negative mode 181  →  122	5 ng/mL	5–5000 ng/mL	[57]
	Human plasma	Protein precipitation (MeOH) IS: Enalapril-d5	XBridge C18 (3.5 μm, 2.1 mm × 100 mm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at flow rate of 0.5 mL/min	Negative mode 181.0  →  122.0	5.7 nM	5.7–14250 nM	[58]
	Monkey plasma, urine, and bile	Protein precipitation IS: HVA-d5	Thermo Aquasil C18 (2.1 mm × 50 mm, 5 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at flow rate 0.6 mL/min	Positive mode 181  →  122	N.S	N.S	[62]
	Rat plasma and brain	Protein precipitation (ACN) IS: Caffeic acid	Kromasil C18 (2.1 mm × 150 mm)	Gradient elution of A (0.1% FA and 2.0 nM ammonium acetate in water) and B (ACN) at 0.2 mL/min	Positive mode 304.0  →  105.0 (benzoylated derivatization)	0.1 ng/mL (0.5 nM)	Brain: 0.01–5.5 μM Plasma: 0.003–2.7 μM	[93]
KYNA (OAT1/3)	Human plasma	Protein precipitation (ACN) IS: Labetalol	Acquity C18 BEH (1.7 μm, 2.1 mm × 100 mm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at flow rate of 0.5 mL/min	Positive mode 181  →  122	1 nM	1–50000 nM	[61]
	Monkey plasma, urine, and bile	Protein precipitation IS: KYNA-d5	Thermo Aquasil C18 (2.1 mm × 50 mm, 5 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at flow rate 0.5 mL/min	Positive mode 190  →  144	N.S	N.S	[62]
	Mouse plasma, CSF, and brain	Protein precipitation (ice-cold MeOH) IS: KYNA-d5	GRACE VisionHT C18 (100 mm × 2.1 mm; 3 μm)	Gradient elution of A (0.1% FA in water add 0.01% TFA) and B (0.1% FA in MeOH added 0.01% TFA) at a flow rate of 400 μL/min	Positive mode 190.1  →  162.0	0.25 nM	Plasma: 2.5–1000 nM CSF: 0.75–1000 nM Brain: 2.5–250 nM	[86]
	Rat plasma, urine, and hippocampus	Protein precipitation (ACN) with IS: 2-cloro-l-phenylalanine	Ultimate XB-C18 (100 mm × 2.1 mm, 5 μm	Gradient elution of A (0.1% FA and 5 mM ammonium acetate in ACN) and B (aqueous with 0.1% FA and 5 mM ammonium acetate) at a flow rate of 0.3 mL/min	190.1  →  144.1	5 ng/mL	Plasma: 5–3000 ng/mL Urine: 120–12000 ng/mL	[88]
	Human plasma	Protein precipitation (ACN) IS: KYNA-d5	Luna NH2 (150 mm × 1.00 mm, 3 μm, 100 A) with a guard column (Security Guard Luna NH2 4 mm × 2.0 mm cartridge)	Gradient elution of 5 mM ammonium acetate in water, pH 9.5 and ACN at a flow rate 35 μL/min	188  →  144.03	0.5 μg/mL	0.5–50 ng/mL	[89]
	Rat plasma	SPE utilizing 3 cc waters HLB cartridges, placed on a vacuum elution manifold IS: Ethyl-4-hydroxy-2-quinolinecarboxylate	Restek C18 aqueous (100 mm × 2.1 mm)	Gradient elution of 0.05%, v/v ammonium formate in water, (pH 5.5) and ACN at a flow rate 0.2 mL/min	189.9  →  143.7	0.035 μM	0.070–1.121 μM	[90]
	Human CSF and serum	Protein precipitation (ice-cold ACN with 0.1% FA) IS: KYNA-d5	Pentafluorophenyl (PFP) (100 Å, 100 mm × 2.1 mm, 2.6 μm) protected by a PFP guard column	Gradient elution of A (0.1% FA in water) and B (0.1% FA in MeOH) at 300 μL/min	190.1  →  116.1	CSF (0.251 nM), serum (0.474 nM)	CSF: 0.25–8 nM; Serum: 4.688–150 nM	[91]
	Human plasma	Protein precipitation (TCA in water) IS: KYNA-d5	Zorbax stable-bond C8 reversed-phase (150 mm × 4.6 mm, 3.5 μm) with a similar guard column (12.5 mm × 4.6 mm, 5 μm)	Gradient elution of A (650 mM acetic acid), B (100 nM HFBA in A), and C (90% ACN in water) at 1 mL/min	Positive mode 190.3  →  143.9	0.05 nmol/L	0.05–400 nmol/L	[92]
	Rat plasma and brain	Protein precipitation (ACN) IS: Caffeic acid	Kromasil C18 (2.1 mm × 150 mm column)	Gradient elution of A (0.1% FA and 2.0 nM ammonium acetate in water) and B (ACN) at 0.2 mL/min	Positive mode 293.9  →  105.1 (benzoylated derivatization)	0.5 ng/mL (2.6 nM)	Brain: 0.01–10.6 μM; Plasma: 0.011–5.3 μM	[93]
	Human serum	Protein precipitation (MeOH) IS: KYNA-d5	Acquity BEH C18 (2.1 mm × 100 mm, 1.7 μm) column fitted with an Acquity BEH C18 VanGuard pre-column (2.1 mm × 5 mm, 1.7 μm)	Isocratic elution of A (10 mM ammonium formate; pH 4.3) and B (acetonitrile) (85/15) at 0.3 mL/min	Negative mode 190.1  →  144.0	0.01 μg/mL	0.01–0.5 μg/mL	[112]
HA (OAT1)	Human serum	Protein precipitation (MeOH) IS: HA-d5	Acquity BEH C18 (2.1 mm × 100 mm, 1.7 μm) column fitted with an Acquity BEH C18 VanGuard pre-column (2.1 mm × 5 mm, 1.7 μm)	Isocratic elution of A (10 mM ammonium formate; pH 4.3) and B (acetonitrile) (85/15) at 0.3 mL/min	Negative mode 178.1  →  134.6	0.2 μg/mL	0.2–80 μg/mL	[112]
6β-OHC (OAT3)	Human plasma and whole blood	Dried blood spot (DBS) sampling followed by SPE using a Sep-Pak C18 plus short-body cartridge, LLE with ethyl acetate IS: 6β–OHC–d4	Kinetex C8 (75 mm × 3.0 mm I.D., 2.6 μm)	Gradient elution of A (0.0125% FA in water) and B (MeOH) at 0.45 mL/min)	Negative mode 423.2  →  347.2	1.08 pg/50 μL	0.001–0.108 ng/50 μL	[113]
	Human urine	Acidifying with FA, protein precipitation with MeOH IS: 6β–OHC–d2	ReproSilPur C18-AQ (100 mm × 2 mm, 5 μm, Maisch)	Gradient elution of A (0.1% FA) and B (0.1% FA in MeOH) at a flow rate of 240 μL/min	Negative mode 423  →  347	2 ng/mL	2–800 ng/mL	[114]
	Human plasma and urine	LLE with ethyl acetate IS: 6α-methylprednisolone	Inertsil ODS-3 (5 μm, 50 mm × 2.1 mm)	Gradient elution of A (1 mM NH4Cl pH 9.0) and B (MeOH) at a flow rate of 0.25 mL/min	Negative mode [M+35Cl−] 413  →  35	2 ng/mL	2–500 ng/mL	[115]

GCDCA-S: glycochenodeoxycholate sulfate, CPI: coproporphyrin I, CPIII: coproporphyrin III, TDA: tetradecanedioate, HDA: hexadecanedioate, DHEAS: dehydroepiandrosterone sulfate, PDA: pyridoxic acid, HVA: homovanillic acid, KYNA: kynurenic acid, HA: hippuric acid, 6β-OHC: 6β-hydroxy cortisol, CSF: cerebrospinal fluid, IS: internal standard, ACN: acetonitrile, FA: formic acid, TCA: trichloroacetic acid, TFA: trifluoroacetic acid, HFBA: heptafluorobutyric acid, SLE: supported liquid extraction, LLE: liquid-liquid extraction, SPE: solid phase extraction, BEH: ethylene bridged hybrid, UHPLC: ultra-high performance liquid chromatography, HILIC: hydrophilic interaction liquid chromatography, MRM: multiple reaction monitoring, LLOQ: lower limit of quantification, N.S: data not shown.

Table 3.

Summary of liquid chromatography-tandem mass spectrometry (LC-MS/MS) methods for determining potential endogenous biomarkers in biological fluids for breast cancer resistance protein (BCRP) and multidrug resistance-associated protein 2 (MRP2).

Biomarkers (transporters)	Matrix	Sample preparation & IS	Column	Mobile phase	MRM transition (m/z)	LLOQ	Linear range	Ref.
Riboflavin (BCRP)	Mouse plasma, monkey plasma	Protein precipitation (ice-cold acidic ACN)	BEH Amide (1.7 μm, 100 mm × 2.1 mm)	Gradient elution of A (10 mM ammonium formate in water) and B (0.1% FA in ACN) at 0.7 mL/min	Positive mode 377.1  →  243.1	1 ng/mL	1–500 ng/mL	[76]
		IS: Riboflavin-13C4, 15N2
	Human plasma	Protein precipitation with TCA in water	Zorbax stable-bond C8 reversed-phase (150 mm × 4.6 mm, 3.5 μm) equipped with a similar guard column (12.5 mm × 4.6 mm, 5 μm)	Gradient elution of A (650 mM acetic acid), B (100 nM HFBA in A), and C (90% ACN in water) at 1 mL/min	Positive mode 377.4  →  243.3	0.05 nmol/L	0.05–400 nmol/L	[92]
		IS: Riboflavin-d8
CPI and CPIII (MRP2)	Monkey plasma, urine, and bile	SLE	Acquity UHPLC BEH C18 (2.1 mm × 100 mm, 1.7 μm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at a flow rate of 0.65 mL/min	Positive mode 655  →  596	N.S	N.S	[62]
		IS: 15N4-CPI
	Rat urine, bile, serum, feces, liver, and plasma	Protein precipitation using 12 M FA, LLE using ethyl acetate, SPE using a Biotage SPE Dry 96	Atlantis T3 HPLC (3 μm, 150 mm × 2.1 mm)	Gradient elution of A (0.1% FA in water) and B (0.1% FA in ACN) at a flow rate of 0.5 mL/min	Positive mode doubly charged 655.3  →  537.2	0.5 nM	0.5–500 nM	[116]
		IS: 15N4-CPI

CPI: coproporphyrin I, CPIII: coproporphyrin III, IS: internal standard, ACN: acetonitrile, FA: formic acid, TCA: trichloroacetic acid, TFA: trifluoroacetic acid, SLE: supported liquid extraction, LLE: liquid-liquid extraction, SPE: solid phase extraction, BEH: ethylene bridged hybrid, HFBA: heptafluorobutyric acid, MRM: multiple reaction monitoring, LLOQ: lower limit of quantification, UHPLC: ultra-high performance liquid chromatography, N.S: data not shown.

Protein precipitation is a critical sample preparation technique for analyzing endogenous biomarkers in complex biological fluids such as plasma, serum, and urine. This method effectively removes proteins that may interfere with biomarker detection, thereby enhancing the sensitivity and specificity. Typically, a defined volume of the biological sample (e.g., 100 μL) is mixed with an equal or greater volume of a precipitating solvent such as acetonitrile, methanol, or ethanol. In some protocols, acidic agents such as trichloroacetic acid (TCA) are included to denature proteins more effectively. For instance, Midttun et al. [92] developed a quantitative method using TCA in water (60 g/L) to deproteinize samples, enabling the simultaneous analysis of riboflavin, five vitamin B6 forms, tryptophan, and six tryptophan metabolites. Protein precipitation is advantageous because of its simplicity, speed, and effectiveness in eliminating interfering substances, making it suitable for high-throughput analyses and enabling the reliable quantification of transporter-associated biomarkers. However, LLE and SPE are often preferred over protein precipitation owing to their ability to achieve better selectivity and cleaner samples. Protein precipitation primarily removes proteins but may leave other contaminants that interfere with the analysis, whereas LLE isolates analytes based on solubility differences, and SPE selectively binds and elutes target compounds, thereby reducing impurities. In complex matrices such as plasma or CSF, LLE, and SPE improve sensitivity and specificity for low-abundance biomarkers such as CPI, CPIII, and 6β-OHC [101,105,106,109,113,115].

During chromatographic separation, the analytes of interest are separated from interfering matrix components based on their physicochemical properties, such as polarity and molecular size. Different types of columns and various changes in LC conditions, including mobile phase composition adjustments, are used. Selecting the optimal column and mobile phase in LC-MS/MS is especially important for accurate separation and sensitive detection of endogenous biomarkers for transporters. Commonly used columns include C18 reversed-phase columns with a particle size ranging 1.7–5 μm and a pore size of 100 Å, which provide high resolution and strong retention of small organic cations. Advanced column types, including biphenyl (e.g., Kinetex Core-Shell Shell Biphenyl high-performance liquid chromatography (HPLC) Columns) and phenyl-hexyl columns (e.g., Luna Phenyl-Hexyl HPLC Columns), offer enhanced selectivity through aromatic interactions, making them particularly useful for biomarkers such as dopamine. Hydrophilic interaction liquid chromatography (HILIC) columns, such as ethylene bridged hybrid (BEH) Amide, are also prominent in endogenous biomarker analysis due to their ability to retain and selectively separate highly polar compounds [23,30,36,38,41,55,82,84,97,99]. These columns, with polar stationary phases and highly organic mobile phases, enhance the sensitivity and resolution of small hydrophilic biomarkers, which are often difficult to retain in reversed-phase columns. HILIC columns are ideal for detecting low-abundance biomarkers in complex matrices, such as plasma and CSF, owing to reduced matrix interference, which improves analytical accuracy. Luo et al. [30] reported that using the HILIC column enables simultaneous quantification of multiple endogenous biomarkers, including thiamine, IBC, creatinine, NMN, and L-CAR. C18 columns are also occasionally used with endogenous biomarkers, as they offer strong retention of less polar or moderately polar compounds, allowing effective separation of biomarkers with varying hydrophobicity, such as bilirubin [117]. While C18 columns are best suited for nonpolar analytes, they can accommodate a range of endogenous compounds by altering the mobile phase conditions, making them versatile for analyzing biomarkers with both polar and nonpolar characteristics across a variety of biomolecular analyses.

Optimization of the mobile phase composition is essential for efficient ionization and improved peak shapes. Common mobile phases are mixtures of water and acetonitrile or methanol with additives such as formic acid (0.1–0.2 %) or ammonium formate (2–10 mM) (Table 1, Table 2, Table 3). These acidic additives mitigate ion suppression and promote the formation of positively charged ions during electrospray ionization. Gradient elution methods are typically used to improve the sensitivity of complex matrices by starting with a high aqueous content and gradually increasing the organic phase. This approach aids in the retention of early eluting compounds and the effective separation of closely related biomarkers. Careful selection of the column and mobile phase ensures maximal signal with minimal matrix effects, enabling the reproducible and robust quantification of endogenous biomarkers across diverse biological fluids.

Modern LC-MS/MS has achieved higher signal intensity and lower limits of quantification (down to 0.1–1 ng/mL or even pg/mL [36,65,105,108]) owing to advancements in sample preparation techniques such as SPE and HILIC, as well as optimized mobile phase compositions. Collectively, these developments support trace-level and accurate quantification of biomarkers in most complex samples. Recent advancements in LC-MS/MS have substantially enhanced clinical diagnostics, proteomics, and metabolomics, particularly for the quantification of low-abundance biomarkers. These improvements are driving progress in personalized medicine, clinical diagnostics, and other fields that require precise and sensitive biomarker quantification.

In summary, as evidenced by the research compiled in Table 1, Table 2, Table 3, LC-MS has become increasingly relied upon for quantifying endogenous biomarkers in transporter-mediated DDI studies. The most common method for sample cleanup and analyte extraction is protein precipitation using MeOH or ACN. Gradient elution with columns packed for HILIC has been frequently employed, enabling the effective separation of polar endogenous biomarkers. To compensate for matrix effects and enhance quantification precision, IS were isotope- or radiolabeled versions of the target biomarkers. The ionization mode varied based on the physicochemical properties of the biomarkers. Most studies used negative ionization for organic anions, while organic cation biomarkers were detected in positive mode, with doubly charged ions improving sensitivity such as CPI and CPIII. Linearity ranges also varied across different biological matrices (plasma, urine, and CSF), reflecting matrix-dependent influences on biomarker quantification. Collectively, these studies highlight the robustness, reproducibility, and versatility of LC-MS-based bioanalytical methods in investigating transporter-mediated mechanisms in pharmacokinetics and DDI.

4. Approaches for managing endogenous levels of target biomarkers in calibration standards and QC solutions

In quantification, selecting an appropriate approach for creating calibration curves and QC samples is essential to accurately replicate the actual study samples and achieve precise measurements. The ideal method for preparing these samples involves adding known amounts of the analyte to the desired matrix. However, for endogenous substances, the original matrix often contains unknown amounts of endogenous compounds, making it unsuitable for creating QC samples or standards. Various approaches tailored to the properties of the target endogenous compounds and specific biological matrices have been developed to address this challenge. They involve the use of surrogate matrices and analytes.

4.1. Surrogate matrices

Currently, the predominant method employed for quantifying endogenous analytes involves the use of a surrogate matrix for standards and QCs to minimize or eliminate the effects of endogenous analytes [118]. This approach has been widely adopted because of its practicality and cost-effectiveness [30,50,57,58,64,73]. In this methodology, the concentrations of analytes in the incurred samples are determined through comparison with the calibration curves established using a surrogate matrix. This approach assumes that the selected methodology is unbiased. However, inherent differences between the surrogate and real matrices can introduce potential bias owing to matrix effects. Bovine serum albumin (BSA) in phosphate-buffered saline (pH 7.4) is commonly used as a surrogate matrix for preparing calibration standards of endogenous compounds, with BSA concentrations ranging from 0.1% to 5% [36,41,50,55,57,58,62,66,71,107,108,116,119]. Charcoal-stripped K₃-EDTA plasma is also frequently used as a surrogate matrix for analyzing endogenous compounds in human plasma [30,61,66,67,73,89,97,103,[106], [107], [108], [109], [110]]. In our previous research, ET and L-CAR concentrations in rat plasma samples were successfully evaluated against calibration curves generated using 1 % BSA in phosphate-buffered saline [39,50].

4.2. Surrogate analytes

The use of stable and radioactive isotopes as surrogate analytes is an advanced technique for addressing issues arising from the baseline amounts of endogenous analyte in the matrix [12,13,120]. Theoretically, stable isotope or radiolabeled analytes, such as those labeled with ¹³C, ¹⁵N, or deuterium, have identical physicochemical properties and exhibit similar behavior to LC-MS/MS their naturally occurring counterparts during LC-MS/MS analysis. However, these labeled analytes are not naturally present in the original matrix, allowing their use in generating QCs and standard calibration curves in matrices free from inherent interference. The concentration of the original analyte can be determined by comparing its mass response with that of the calibration curve constructed using the surrogate analyte. The surrogate analyte approach effectively eliminates interference resulting from endogenous analytes, enabling the preparation of standard calibration curves in the desired matrix without any hindrance. Notably, this method is widely recognized as the only reliable means of quantifying reductions in biomarker levels, as it allows for the preparation of a QC level below the baseline within an authentic matrix. Consequently, the surrogate analyte technique has gained popularity for the quantification of endogenous analytes. For instance, Zhao et al. [83] used a surrogate analyte (creatinine-d3) to validate an LC-MS/MS assay for quantifying creatinine plasma concentration in humans as a biomarker. Additionally, pharmacokinetic studies in monkeys have demonstrated the use of CPI-d8, either alone or with the OATP inhibitor CsA, to assess potential OATP inhibition [67]. This technique enables the study of the pharmacokinetic profiles of endogenous biomarkers without the need for baseline levels. Stable isotopes or radiolabeled substrates are often used to distinguish them from naturally occurring compounds, thereby reducing endogenous interference. Consequently, various pharmacokinetic studies have been conducted in animals using stable endogenous isotopes. Table 4 [22,49,60,67,[121], [122], [123]] summarizes pharmacokinetic studies of endogenous biomarkers using stable isotopes or radiolabeled compounds, detailing the corresponding dose and analytical methods. Although isotopes are suitable for studying endogenous compounds in specific cases, commercial target analytes are not always readily available [124]. Furthermore, the use of radiolabeled compounds is limited by the risk of radioactive contamination. Consequently, stable isotope compounds have recently become less popular than surrogate matrices.

Table 4.

Pharmacokinetic studies of endogenous biomarkers using stable isotope or radiolabeled compounds.

Compound	Stable isotope/radiolabeled form	Reference dose	Analytical method	Pharmacokinetics parameters	Species	Refs.
Thiamine	[3H]-thiamine	iv (2 mg/kg) with 0.2 μCi/g of [3H] thiamine	LSC	N.S	Mouse	[22]
ET	[3H]-ET	PO 330 μg/kg	LSC	N.S	Mouse	[49]
HA	[14C]-HA	0.1 mg/kg	LSC	AUC0–∞ (% of dose × min/mL): 25.1 ± 1.1 CLR (mL/min/kg): 18.1 ± 0.7	Rat	[60]
CPI	CPI-d8	0.2 mg/kg, iv	LC-MS/MS	AUC0–24h: 188 ± 31.4 nM × h	Monkey	[67]
	[3H]-CPI	0.04 mg/kg [3H]-CPI (5 mCi/kg)	QWBA	N.S	Mouse
L-CAR	[3H]-L-CAR	250 ng/kg iv and po	LSC	AUC0–4h,iv: 29.2 ± 1.3 ng × min/mL AUC0–4h,po: 19.7 ± 1.0 ng × min/mL	Mouse	[121]
Creatinine	[14C]-creatinine	IV 1 μCi (approximately 20 nmol) in 100 μL saline	LSC	Kidney/serum ratio [14C]-creatinine 2.05 ± 0.067	Mouse	[122]
Taurine	[3H]-taurine	40 μCi/kg po 30 mg/kg (10 mL/kg)	LSC	AUC0–last: 1181 ± 62 (μg × min/mL)	Rat	[123]

ET: ergothioneine, CPI: coproporphyrin I, L-CAR: L-carnitine, HA: hippuric acid, iv: intravenous, po: oral, LSC: liquid scintillation counting, LC-MS/MS: liquid chromatography-tandem mass spectrometry, QWBA: quantitative whole-body autoradiography, AUC: area under the curve, CL_R: renal clearance, N.S: data not show.

Surrogate matrices and analytes areeffective approaches for the challenges in biomarker quantification when authentic matrices or analytes are unavailable or unstable. A surrogate matrix is an alternative, such as saline or artificial plasma, that resembles the sample matrix but lacks endogenous interference, enabling straightforward calibration and analysis. Conversely, a surrogate analyte is a chemically related compound used when the actual analyte is unstable or difficult to obtain, thus modeling the behavior of the target compound in the matrix. These two approaches enhance the robustness of the method; surrogate matrices improve background clarity, whereas surrogate analytes help assess the stability and response similarity in quantification methods (Fig. 3) [125].

Fig. 3.

Liquid chromatography-tandem mass spectrometry ( LC-MS/MS) chromatograms illustrating the different approaches used for analyte quantification. (A) Background peak of the endogenous analyte in the blank matrix. (B) Neat standard with a known concentration of the analyte. (C) The same matrix after spiking with a standard of known concentration using the background subtraction method. (D) The study sample was split into several aliquots, each spiked with a standard of different concentration using the standard addition method. (E) Analyte standard spiked into a surrogate matrix containing no background, using the surrogate matrix method. (F) The surrogate analyte standard spiked into the original matrix, demonstrating the surrogate analyte method. Reproduced with permission from Ref. [125].

5. Applications of LC-MS-based biomarker quantification in pharmacokinetic studies and baseline levels of endogenous biomarkers in biological fluids

In pharmacokinetic studies, the quantification of LC-MS-based biomarkers is essential for determining endogenous biomarker levels in biological fluids. In such studies, LC-MS offers high sensitivity and specificity for monitoring drug metabolism, absorption, distribution, and elimination, enabling the precise measurement of drug concentrations and metabolites in biological matrices. This precise quantification helps to accurately model drug behavior in the body, facilitating dose adjustment and therapeutic monitoring. Additionally, LC-MS is crucial for quantifying the individual baseline levels of endogenous biomarkers, including hormones, metabolites, and proteins, in biological fluids such as blood, urine, and saliva. Measuring these biomarkers at low concentrations provides valuable insights into their physiological and pathological states, thereby supporting personalized medicine, disease diagnosis, and therapeutic efficacy monitoring.

Basal levels of endogenous biomarkers refer to their typical baseline levels naturally produced within the body [126]. Fluctuations in these levels are common and can be influenced by factors such as the time of day, diet, stress, activity level, and age [127]. Understanding the normal range and expected fluctuations in these biomarkers is essential for interpreting diagnostic tests and monitoring health conditions. For example, hormones such as cortisol and insulin exhibit diurnal variations with predictable fluctuations throughout the day. Similarly, certain enzyme levels in blood can vary depending on recent meals or physical activity. Monitoring these fluctuations can provide insights into an individual’s health status, aiding healthcare professionals in making informed decisions regarding treatment or intervention [128]. However, significant deviations from typical fluctuations may signal health issues and warrant further investigation.

Several methods of assessing endogenous substrates have been documented owing to their importance in clinical and pharmacological studies. Techniques for measuring endogenous agents in biological fluids (such as urine, bile, serum, and CSF) include enzymatic methods, the Jaffe method, HPLC with fluorescence or UV detection, and LC-MS/MS. However, each technique has limitations that can affect the accuracy of endogenous level measurements. A comprehensive review of the baseline levels of each endogenous biomarker across various biological matrices is lacking. The inconsistent reporting of endogenous concentration values poses a significant challenge, leading to uncertainty for researchers studying these analytes. Factors contributing to this inconsistency include standardization issues, variations in sample collection and handling, population differences, analytical sensitivity and specificity, and publishing bias [12,13,125]. Minor procedural differences can lead to substantial variations in results, and performance discrepancies among laboratories contribute to further inconsistencies. For example, creatinine concentrations vary significantly across species, and even within the same species, depending on the method used. The rat plasma creatinine levels were reported as 12.8 ± 0.1 μM and 20–25 μmol/L in two different studies both using LC-MS/MS, respectively [82,129]. Similarly, among humans, creatinine levels differ during nonpregnancy, pregnancy, and lactation: 62 μmol/L in a nonpregnant woman, 44 μmol/L during pregnancy, and 62 μmol/L during lactation [130]. However, a study with a larger sample size (n = 135) reported creatinine levels of 72 μmol/L for women, measured using the enzymatic method. In another example, free-form tryptophan levels in rat plasma have been reported with significant variation across different studies. Rat plasma tryptophan levels were found to be 16.5 ± 1.8 μg/mL using HILIC-UHPLC-MS/MS and varying levels when measured using LC-MS/MS (1872.19 ± 868.08 ng/mL, 8941 ± 1734 ng/mL, 31.61 ± 1.54 μM). Although the accuracy of these published values cannot be definitively determined, each study provides a reference range that allows the estimation of the actual concentration range of each endogenous biomarker within biological samples.

To address these challenges, it is necessary to foster collaboration among researchers, establish standardized procedures, implement strict QC measures, adopt transparent reporting practices, and carefully analyze demographic characteristics. Creating reference ranges and comparing them to established controls can enhance the uniformity and reliability of the reported values for endogenous biomarkers in scientific studies. Therefore, in Table 5 [22,23,29,30,36,37,39,41,50,55,57,58,[60], [61], [62],65,68,74,76,82,83,[85], [86], [87], [88], [89], [90], [91], [92],94,95,101,[103], [104], [105], [106],[108], [109], [110], [111], [112], [113], [114],119,[129], [130], [131]], we present a summary of these values for transporters published in previous studies to facilitate convenient reference and enable comparison across various methods.

Table 5.

Basal levels of endogenous biomarkers in biological fluids.

Biomarkers (transporter)	Matrix	Basal level	Analytical method	Refs.
Thiamine (OCT1, OCT2, MATEs)	Mouse plasma (n = 3)	≅ 400 ng/mL	LC-MS/MS	[22]
	Human plasma (n = 8)	≅ 6 nmol/L	LC-MS/MS	[23]
	Human plasma (n = 4)	0.57–2.47 ng/mL	UHPLC-MS/HRMS	[30]
IBC (OCT1)	Human plasma (n = 22)	22.6 ± 2.6 ng/mL	HPLC-MS/MS	[29]
	Human plasma (n = 4)	10.88–34.67 ng/mL	UHPLC-MS/HRMS	[30]
Creatinine (OCT2, MATEs)	Human plasma (n = 4)	5816.01–8786.61 ng/mL	UHPLC-MS/HRMS	[30]
	Rat plasma (n = 6)	2420–3700 ng/mL	LC-MS/MS	[39]
	Human serum (n = 15)	-Human: 0.87 mg/dL (0.77–1.05) - Monkey: 40 μM	LC-MS/MS	[41]
	Monkey plasma (n = 3)
	Rat serum (n = 15–20)	0.758 ± 0.024 mg/dL; clearance: 2.83 ± 0.09 (mL/min/kg)	LSC	[60]
	Rat plasma (n = 3–7)	12.8 ± 0.1 μM, AUC0–4 h = 5373 ± 62 μM × min, CLR = 10.9 ± 1.2 mL/min/kg, Kp,kidney = 10.2 ± 0.7	LC-MS/MS	[82]
	Human plasma (n = 10)	5.83–11.19 μg/mL	LC-MS/MS	[83]
	Mouse plasma (n = 65)	0.076 ± 0.002 mg/dL, total clearance: 329 ± 17 μL/min	LC-MS/MS	[85]
	Rat plasma	20–25 μmol/L	LC-MS/MS	[129]
	Human plasma (n = 50)	-Nonpregnant (n = 17): 62 (51–71) μmol/L; CL: 97 ± 29 mL/min -Pregnant (n = 17): 44 (42–53) μmol/L; CL: 161 ± 62 mL/min -Lactating (n = 16): 62 (62–71) μmol/L; CL: 100 ± 37 mL/min	Enzymatic method	[130]
NMN (OCT2, MATE1/2K)	Human plasma (n = 4)	8.89–27.34 ng/mL	UHPLC-MS/HRMS	[30]
	Human plasma (n = 8)	6–8 ng/mL CLR: 403 ± 61 mL/min	LC-MS/MS	[36]
	Human plasma (n = 25)	165.7 nM (8.0–1237.0 nM)	LC-MS/MS	[37]
m1A (OCT2, MATEs)	Mouse plasma (n = 3)	69.0 ± 6.2 nM, AUC0–120min: 165 ± 24 μmol/L × min, total clearance: 2.06 ± 0.26 mL/min, CLR: 1.06 ± 0.13 mL/min	LC-MS/MS	[41]
Tryptophan (OCT2)	Mice plasma (n = 8)	90 ± 13 μmol/L	LC-MS/MS	[86]
	Human plasma (n = 6)	4.5–11 μg/mL	LC-MS/MS	[87]
	Rat plasma (n = 6)	8941 ± 1734 ng/mL	LC-MS/MS	[88]
	Human plasma (n = 100)	10.5 ± 2.2 μg/mL	LC-MS/MS	[89]
	Rat plasma (n = 5)	31.61 ± 1.54 μM	LC-MS/MS	[90]
	Human CSF and serum (n = 14)	-CSF: 1847 ± 305.6 nM -Serum: 52043 ± 11057 nM	LC-MS/MS	[91]
	Human plasma (n = 94)	63.1 μmol/L	LC-MS/MS	[92]
	Rat serum and brain (n = 6)	-Serum: 1872.19 ± 868.08 ng/mL -Brain: 141.05 ± 13.47 ng/g wet tissue	LC-MS/MS	[94]
	Rat plasma and urine (n = 6)	-Plasma: 16.5 ± 1.8 μg/mL (free-form tryptophan) -Urine: 0.8 ± 0.1 μg/mL	HILIC-UHPLC-MS/MS	[95]
Dopamine (OCT2)	Mice plasma (n = 8)	23 ± 6 nmol/L	LC-MS/MS	[86]
	Rat serum and brain (n = 6)	-Serum: 23.89 ± 0.08 ng/mL -Brain: 137.52 ± 12.42 ng/g wet tissue	LC-MS/MS	[94]
ET (OCTN1)	Rat plasma (n = 5)	200–500 ng/mL	LC-MS/MS	[50]
L-CAR (OCTN2)	Human plasma (n = 4)	3889.61–5606.32 ng/mL	UHPLC-MS/HRMS	[30]
	Rat plasma (n = 6)	1975–3351 ng/mL	LC-MS/MS	[39]
CPI and CPIII (OATP1B)	Cynomolgus monkey plasma (n = 3)	CPI: 0.000476 ± 0.000165 μM, AUC0–24h: 0.00814 ± 0.00232 μM × h	LC-MS/MS	[62]
	Human plasma (n = 13)	CPI: 0.72 ± 0.22 nM CPIII: 0.11 ± 0.04 nM	LC-MS/MS	[65]
	Human plasma (n = 16)	CPI: 0.72 ± 0.15 nM CPIII: 0.14 ± 0.03 nM	LC-MS/MS	[68]
	Human plasma (n = 12) Human plasma (n = 6)	CPI: 0.598 ± 0.147 ng/mL; CPIII: 0.078 ± 0.025 ng/mL CPI: 0.710 ± 0.300 ng/mL; CPIII: 0.089 ± 0.031 ng/mL	LC-MS/MS	[105]
	Human plasma Human urine	CPI: 0.45–1.1 ng/mL; CPIII: 0.05–0.5 ng/mL CPI: 5–35 ng/mL; CPIII: 1–35 ng/mL	LC-MS/MS	[101,106]
	Human plasma	CPI: 350–550 pg/mL	LC-MS/MS	[108]
	Monkey plasma (n = 3)	CPI: 1–5 nM; AUC0–6h: 13.2–16.0 nM × h CPIII: 0.5–3 nM; AUC0–6h: 2.2–2.5 nM × h	LC-MS/MS	[109]
GCDCA-S (OATP1B)	Human plasma (n = 6)	Cmax: 100 ng/mL	LC-MS/HRMS	[103]
	Rat plasma (n = 6)	4–12 ng/mL	LC-MS/MS	[104]
DHEAS (OATP1B)	Monkey plasma (n = 4)	Cmax: 83.3 ± 38.3 ng/mL AUC0.5–8h: 530 ± 282 ng/mL × h	LC-MS/MS	[74]
TDA & HDA (OATP1B)	Human plasma (n = 20)	TDA: 7.0–172.6 nM HDA: 10–106.5 nM	LC-MS/MS	[110]
Bilirubin (OATP1B)	Rat serum (n = 10) (pregnant rats)	0.67 ± 0.14 mg/dL	N/A	[131]
PDA (OAT1/3)	Monkey plasma (n = 3)	1.5 ± 0.3 μM	LC-MS/MS	[57]
	Human plasma (n = 14)	18.1–19.6 nM	LC-MS/MS	[58]
	Cynomolgus monkey plasma (n = 3)	0.0482 ± 0.0273 μM, AUC0–24h: 1.47 ± 0.762 μM × h	LC-MS/MS	[62]
	Human plasma (n = 94)	24.3 nmol/L	LC-MS/MS	[92]
	Human CSF (n = 148)	<5 nmol/L	LC-MS/MS	[111]
	Rat plasma (n = 6)	50 nmol/L	LC-MS/MS	[129]
	Human plasma (n = 50)	-Nonpregnant (n = 17): 43 ± 17 nmol/L; CL: 232 ± 94 mL/min -Pregnant (n = 17): 34 ± 10 nmol/L; CL: 337 ± 140 mL/min -Lactating (n = 16): 40 ± 11 nmol/L; CL: 215 ± 103 mL/min	Cation-exchange HPLC	[130]
HVA (OAT1/3)	Monkey plasma (n = 3)	85 ± 32 nM	LC-MS/MS	[57]
	Human plasma (n = 14)	37.3–56.4 nM	LC-MS/MS	[58]
	Cynomolgus monkey plasma (n = 3)	0.101 ± 0.0388 μM, AUC0–24h: 2.20 ± 0.652 μM × h	LC-MS/MS	[62]
Taurine (OAT1/3)	Human plasma (n = 6)	5000–8000 ng/mL, AUC0–∞: 53.1 ± 2.6 μg/mL × h, CLR: 0.372 ± 0.054 mL/min/kg	LC-MS/MS	[55]
	Rat serum, urine (24h accumulation), liver (n = 8)	-Serum: 0.28–0.33 μmol/mL -Urine: 307–485 μmol/kg bw/24 h -Liver: 8.3–9.1 μmol/g wet weight	HPLC FLD	[119]
	Rat serum (n = 10) (pregnant rats)	50.35 ± 5.30 μmol/L	N/A	[131]
GCDCA-S (OAT1/3)	Human plasma and urine (n = 6)	150–500 ng/mL, AUC0–∞: 1.79 ± 0.29 μg/mL × h, CLR: 0.0517 ± 0.0140 mL/min/kg	LC-MS/MS	[55]
KYNA (OAT1/3)	Human plasma (n = 14)	39.7 ± 16.5 nM	LC-MS/MS	[61]
	Cynomolgus monkey plasma (n = 3)	0.0361 ± 0.0210 μM, AUC0–24h: 0.853 ± 0.305 μM × h	LC-MS/MS	[62]
	Mouse plasma (n = 8)	47 ± 18 nmol/L	LC-MS/MS	[86]
	Rat plasma (n = 6)	11.6 ± 4.7 ng/mL	LC-MS/MS	[88]
	Human plasma (n = 100)	8.3 ± 3.4 ng/mL	LC-MS/MS	[89]
	Rat plasma (n = 5)	0.07 ± 0.01 μM	LC-MS/MS	[90]
	Human CSF (n = 14)	0.9 ± 0.6 nM	LC-MS/MS	[91]
	Human serum (n = 14)	32.2 ± 10.4 nM
	Human plasma (n = 94)	35.4 nmol/L	LC-MS/MS	[92]
	Human serum (n = 1, healthy volunteer)	0.05 μg/mL	LC-MS/MS	[112]
HA (OAT1)	Rat serum (n = 15–20)	12.3 ± 2.4 μM	LSC	[60]
	Human serum (n = 1, healthy volunteer)	1.89 μg/mL	LC-MS/MS	[112]
6β-OHC (OAT3)	Human plasma (n = 3) Human whole blood (n = 3) Human DBS (n = 3)	0.38–0.97 ng/mL	LC-MS/MS	[113]
		0.25–0.71 ng/mL
		0.018–0.041 ng/50 μL
	Human urine (n = 5)	107–361 ng/mL	LC-MS/MS	[114]
Riboflavin (BCRP)	Mouse plasma (n = 3)	-Mouse: 23.3 ± 2.8 nM (average Cmax), AUC0–24 h = 419 ± 44.7 (nM × h), CLR: 0.0517 ± 0.0140 mL/min/kg -Monkey: 25.4 ± 4.0 nM (average Cmax), AUC0–24 h = 407 ± 44.1 (nM × h) -Human: 106–638 nM	LC-MS/MS	[76]
	Monkey plasma (n = 3)
	Human plasma (n = 64)
	Human plasma (n = 94)	11.7 nmol/L	LC-MS/MS	[92]
	Rat plasma (n = 6)	30 nmol/L	LC-MS/MS	[129]

OCT: organic cation transporter, MATE: multidrug and toxic compound extrusion, OCTN: organic cation/carnitine transporter, OAT: organic anion transporter, OATP: organic anion-transporting polypeptide, BCRP: breast cancer resistance protein, MRP2: multidrug resistance-associated protein 2, P-gp: P-glycoprotein, MRD1: multidrug resistance protein 1, m¹A: N1-methyladenosine, NMN: N1-methyl nicotinamide, ET: ergothioneine, L-CAR: L-carnitine, HVA: homovanillic acid, KYNA: kynurenic acid, PDA: pyridoxic acid, HA: hippuric acid, 6β-OHC: 6β-hydroxy cortisol, GCDCA-S: glycochenodeoxycholate sulfate, CPI: coproporphyrin I, CPIII: coproporphyrin III, DHEAS: dehydroepiandrosterone sulfate, HDA: hexadecanedioate, TDA: tetradecanedioate, CSF: cerebrospinal fluid, C_max: peak plasma concentration, LC-MS/MS: liquid chromatography tandem mass spectrometry, HPLC: high-performance liquid chromatography, GC-MS: gas chromatography-mass spectrometry, UHPLC-HRMS/MS: ultra-high performance liquid chromatography coupled to high-resolution mass spectrometry, LSC: liquid scintillation counting, HILIC: hydrophilic interaction liquid chromatography, FLD: fluorescence detection, CL: clearance, CL_R: renal clearance, AUC: area under the curve, N/A: not available, bw: body weight.

6. Emerging trends in application of MS-based biomarker quantification in drug development

The high specificity and sensitivity of LC-MS/MS have made it a widely adopted method for quantifying small-molecule biomarkers. The FDA requires regulatory submissions, including verified biomarker assays [132], with timely bioanalytical guidance for their validation. An emerging trend in biomarker quantification is driven by advancements in high-resolution MS such as Orbitrap and quadrupole time-of-flight MS [133], improved ionization sources, and novel separation methods such as HILIC. These technologies, which enable the detection of low-abundance biomarkers with greater sensitivity and resolution, have become indispensable for the analysis of complex matrices. Ambient ionization techniques, such as desorption electrospray ionization [134] and matrix-assisted laser desorption ionization [135], are also being integrated into workflows to reduce the need for extensive sample preparation. Coupled with advances in hardware, bioinformatics, and machine learning, data processing has improved, enabling pattern recognition in complex datasets and enhancing quantification accuracy. Together, these developments have made high-throughput biomarker analysis faster and more insightful, streamlining transporter-related research [136].

Mass spectrometry imaging (MSI) has become an essential tool for studying transporter-related biomarkers, offering some advantages over traditional mass spectrometry techniques [137]. Unlike LC-MS/MS and ambient MS, which require homogenization of samples, MSI preserves the spatial distribution of analytes within biological tissues. This allows researchers to investigate transporter activity and biomarker localization with precise spatial resolution, providing valuable insights into drug transport and tissue-specific distribution. One of MSI’s key strengths is its ability to perform label-free, multiplex analysis of biomolecules in their native state [138]. This enables a comprehensive understanding of transporter-mediated processes within a single experiment, as MSI can simultaneously detect multiple analytes. By mapping transporter substrates and endogenous biomarkers in organs such as the liver, kidney, and brain, MSI reveals regional variations in transporter activity. This spatial information is valuable for evaluating the role of transporters in drug absorption, distribution, metabolism, and excretion. Several MSI techniques have been successfully applied in the field of biomarker research. Among these, Matrix-Assisted Laser Desorption/Ionization (MALDI-MSI) has gained popularity due to its high sensitivity, and versatility, particularly for imaging of small molecules, lipids, and peptides in tissue samples. Desorption Electrospray Ionization (DESI-MSI) operates under ambient conditions with minimal sample preparation and is well-suited for rapid tissue imaging [139]. Secondary Ion Mass Spectrometry (SIMS), with its extremely high spatial resolution, is ideal for imaging subcellular transporter activity. Each technique offers unique advantages, allowing researchers to select the most suitable methodology based on their specific study objectives. The applications in transporter-mediated DDI research have proven highly valuable. By enabling the direct visualization of DDI’s effects on the spatial distribution of drug metabolites and biomarkers, MSI contributes considerably to predicting adverse drug effects [140]. This technology also facilitates understanding transporter involvement in the biodistribution of new drug candidates during development. Moreover, its application in toxicology and pathology provides critical insights into tissue-specific biomarker accumulation, furthering our knowledge of disease mechanisms and toxicological outcomes related to transporter activity. Collectively, MSI represents a major breakthrough in transporter-related biomarker research, offering unparalleled spatial and molecular insights. As instrumentation and data analysis techniques continue to advance, MSI is poised to play an even greater role in DDI prediction, biomarker discovery, and precision medicine.

Miniaturization and high-throughput biomarker analysis are becoming increasingly relevant in both preclinical and clinical applications of MS-based bioanalytical methods for quantifying endogenous biomarkers. However, these topics remain underrepresented topics in the literature [141]. Miniaturization involves the development of compact and efficient sample preparation and analysis workflows that require smaller sample volumes and reduce the overall experimental footprint [142]. This is particularly important in applications such as pediatric or neonatal pharmacokinetic studies, where sample availability is limited, or in high-stakes drug development studies, where resource optimization is crucial. The integration of sample preparation, separation, and detection into a single, streamlined workflow is showing promise with miniaturized systems, such as microfluidic platforms and lab-on-a-chip technologies [143]. However, their development and adoption have been slow, partly due to challenges in maintaining analytical sensitivity and reproducibility in scaled-down formats. Additionally, the ability to analyze multiple biomarkers simultaneously in large sample sets can significantly enhance the efficiency of biomarker discovery and validation workflows.

While robotic automation and multiplexing capabilities in mass spectrometry have advanced, their integration into workflows for transporter-mediated DDI studies remains limited. Often, most workflows are optimized for sensitivity and specificity in detecting a small number of biomarkers, rather than being designed for high-throughput analysis [144]. In an era where the pharmaceutical and clinical industries increasingly demand more efficient and cost-effective analytical workflows, MS has lagged in addressing these practical considerations. Bridging this gap requires the development of more robust, miniaturized, and high-throughput MS platforms tailored specifically for transporter-mediated DDI studies [145]. Accelerating advancements in this area would not only enhance biomarker analysis efficiency but also expand the accessibility of these techniques for routine clinical applications.

7. Future perspectives

With the increasing frequency of clinical investigations into DDI and the critical need to understand how drug candidates can interfere with transporters, exploring innovative methods to assess the potential of DDI in drug development has become essential [146]. This requirement arises from the complex nature of DDI, which can significantly affect the efficacy and safety profiles of therapeutic agents. One promising approach is to use endogenous analytes, which are naturally occurring substances that offer valuable insights into the pharmacokinetics and pharmacodynamics of drug candidates, as biomarkers. By monitoring changes in the levels of these biomarkers, researchers can better understand how new drugs interact with existing ones, thereby enhancing the drug discovery and development process. This method not only enhances the precision of DDI assessments but also streamlines the development pipeline, potentially reducing the time and cost of bringing new drugs to the market [147]. Consequently, integration of endogenous analytes as biomarkers is becoming increasingly common in the pharmaceutical industry, underscoring their growing importance in modern drug development. Nevertheless, one limitation of using biomarkers for transporters is that they may not completely replace traditional DDI clinical studies, likely because of the potential need for DDI sensitivity testing. Unlike well-established specific drug transporter substrates, which are commonly used as “victim” drugs in DDI studies, biomarkers may not yet be reliable enough to serve as substitutes for traditional clinical DDI assessments. Therefore, further studies are needed to validate and establish the reliability of biomarkers in conventional assessments.

The accurate quantification of endogenous biomarkers in biological samples requires thorough validation, including stability tests under various storage conditions. This is crucial for certain unstable biomarkers, such as HVA. This present review provides an in-depth analysis of bioanalytical methods for quantifying endogenous biomarkers of drug transporters using LC-MS/MS. Reviewing the literature from the past two decades, we emphasize that LC-MS/MS offers significant advantages over older techniques such as HPLC, including greater accuracy, simpler sample preparation, and smaller injection volumes. Additionally, this review highlights the importance of considering factors that may influence analytical data, such as the baseline concentration of target analytes, extraction techniques, and LC-MS/MS system conditions. To further advance this field, clinical investigations employing efficient LC-MS/MS methods are essential.

Integrating multi-omics data, proteomics, metabolomics, and genomics is crucial for obtaining a comprehensive view of transporter functions in health and disease [148]. Cross-omics integration is expected to reveal novel biomarker targets and deepen our understanding of transporter pathways. Analytical platforms, such as lab-on-a-chip devices, are likely to be miniaturized to enable point-of-care testing, allowing for rapid and minimally invasive biomarker monitoring. Another key goal is to develop standardized protocols and accessible biomarkers databases to improve reproducibility and foster consistency across studies. These innovations aim to make transporter biomarker analysis more accessible, accurate, and impactful, supporting its integration into personalized medicine and early diagnosis, particularly in clinical cancer diagnosis.

8. Conclusion

Recent advancements in LC-MS-based bioanalytical techniques have significantly enhanced the quantification of endogenous biomarkers, especially for transporter-mediated DDI. Innovations in LC-MS/MS, high-resolution MS, and refined sample preparation methods have enabled more sensitive, selective, and precise detection of low-abundance biomarkers associated with drug transport and metabolism. These approaches offer critical insights into transporter activity, aiding in the assessment of transporter roles in drug disposition, efficacy, and safety. In pharmacokinetic research, accurate biomarker quantification using MS supports the early identification of DDI, deepens the understanding of transporter mechanisms, and provides dose adjustment strategies in clinical practice. These advancements promote safer and more individualized pharmacotherapy by enabling early detection of transporter-mediated DDI risks during drug development. This review presents updated methods for the analysis of biological samples for the quantification of endogenous biomarkers of drug transporters. Additionally, we present the concentrations of these biomarkers in biological fluids as reported in various previous studies. By that we hope to shed light to facilitate rapid access to biomarkers for preclinical and clinical experiments, thereby optimizing cost and time efficiency in new drug development.

CRediT authorship contribution statement

Dang-Khoa Vo: Writing – review & editing, Writing – original draft, Visualization, Software, Conceptualization. Han-Joo Maeng: Writing – review & editing, Visualization, Supervision, Resources, Project administration, Methodology, Funding acquisition, Conceptualization.

Declaration of competing interest

These authors declare that they have no conflict of interest.

Appendix A. Supplementary data

The following are the Supplementary data to this article: