2024 Volume 47 Issue 6 Pages 1087-1105
Analysis of endogenous metabolites in various diseases is useful for searching diagnostic biomarkers and elucidating the molecular mechanisms of pathophysiology. The author and collaborators have developed some LC/tandem mass spectrometry (LC/MS/MS) methods for metabolites and applied them to disease-related samples. First, we identified urinary conjugated cholesterol metabolites and serum N-palmitoyl-O-phosphocholine serine as useful biomarkers for Niemann–Pick disease type C (NPC). For the purpose of intraoperative diagnosis of glioma patients, we developed the LC/MS/MS analysis methods for 2-hydroxyglutaric acid or cystine and found that they could be good differential biomarkers. For renal cell carcinoma, we searched for various biomarkers for early diagnosis, malignancy evaluation and recurrence prediction by global metabolome analysis and targeted LC/MS/MS analysis. In pathological analysis, we developed a simultaneous LC/MS/MS analysis method for 13 steroid hormones and applied it to NPC cells, we found 6 types of reductions in NPC model cells. For non-alcoholic steatohepatitis (NASH), model mice were prepared with special diet and plasma bile acids were measured, and as a result, hydrophilic bile acids were significantly increased. In addition, we developed an LC/MS/MS method for 17 sterols and analyzed liver cholesterol metabolites and found a decrease in phytosterols and cholesterol synthetic markers and an increase in non-enzymatic oxidative sterols in the pre-onset stage of NASH. We will continue to challenge themselves to add value to clinical practice based on cutting-edge analytical chemistry methodology.
Metabolites are a phenotype of the central dogma, influenced by both the genome, transcriptome, and proteome, as well as environmental factors such as food and xenobiotics1–4) (Fig. 1). It can also regulate other layers of the central dogma, such as epigenetics and the proteome.5) Metabolites play a critical role in various cellular processes and are therefore essential for cell function.1,2) The composition of the metabolome is altered in various diseases. For example, genetic disorders such as phenylketonuria are representative of metabolic disorders.6–8) Genetic diseases are diagnosed not only by genetic tests but also by metabolite analysis, which is referred to as chemical diagnosis.8–15) The term “biomarkers” refers to objectively analyzed substances that indicate physiological, pathological or treatment responses. They are objective indices that reflect reactions occurring in cells and tissues in the body. In addition, the term “disease biomarkers” is commonly used in chemical diagnostics.16,17) In clinical studies, any laboratory-tested parameters can be considered biomarkers. This review summarizes our achievements in the search for diagnostic biomarkers for screening and analysis of metabolic changes in various diseases using LC/tandem mass spectrometry (LC/MS/MS)18–21) (Fig. 2).
Changes in metabolites, both qualitative and quantitative, that occur during various diseases can be analyzed for chemical diagnosis. Metabolites should be used at the time of abnormality for accurate analysis.8,15,22–25) Metabolites produced in various diseases can serve as indicators of diagnosis and treatment efficacy.24) Metabolite analysis by LC/MS/MS is widely used for chemical diagnostics in various single-gene disorders,26–29) and efforts continue to develop diagnostic biomarkers in combination with metabolites and other modalities for multifactorial diseases such as cancer and dyslipidemia.30–33) This section outlines the authors’ efforts to develop diagnostic biomarkers.
2.1. Highly Sensitive and Precise Analysis of Various Metabolites and Establishment of NPC Diagnostic BiomarkersNiemann–Pick disease type C (NPC) is a rare genetic disorder that affects cholesterol transport, resulting in the accumulation of cholesterol and other lipids in various tissues and organs, particularly the brain and nervous system.34–36) NPC presents with a broad spectrum of symptoms and can affect individuals of all ages, from perinatal to adulthood. These symptoms may include neurological, psychiatric, and systemic manifestations. Diagnosis of NPC can be challenging and is often delayed. Conventional diagnostic methods for NPC, such as filipin staining and DNA sequencing, have limitations, including invasiveness, complexity, cost, and false-negative results.37–42) Current treatments for NPC include miglustat,43–45) the only approved drug, and 2-hydroxypropyl-beta-cyclodextrin (HPBCD),46) an experimental agent.47) Both aim to reduce cholesterol levels and improve neurological outcomes. Biomarker analysis of NPC shows promise in facilitating diagnosis, monitoring disease progression, and evaluating treatment efficacy. LC/MS/MS is used to identify and quantify several lipid metabolites that have been reported as potential biomarkers for NPC.48–51)
2.1.1. Biomarkers Development of Urinary Abnormal Cholesterol Urine Using LC/MS/MSThe diagnosis of NPC disease is challenging due to its variable clinical manifestations and the need for invasive methods such as filipin staining37,39) or genetic analysis.42) Therefore, there is a need for non-invasive biomarkers that can detect NPC disease (Fig. 3). The first goal of the urinary NPC biomarker development was set to develop and validate an accurate, reliable and precise LC/MS/MS technique for the measurement of 3β-sulfooxy-7β-N-acetylglucosaminyl-5-cholen-24-oic acid (SNAG-Δ5-CA) and its glycine and taurine amides (SNAG-Δ5-CG and SNAG-Δ5-CT) in urine. These compounds are unusual C24 bile acids that have been found in very high concentrations in the urine of an NPC patient.52) We hypothesized that these compounds could serve as valuable biomarkers for NPC disease. The synthesized reference standards and internal standard were used for the target analytes.53,54) LC/MS/MS conditions were constructed using a trapping column55–57) and a reversed-phase analytical column. The impact of methanol content in the mobile phase on trapping efficiency and the effect of trapping phase hydrophobicity on peak symmetry were evaluated. The analytical method was validated by assessing its linearity, accuracy, precision, and recovery.58) The detection limits ranged from 0.6 to 2 pg, and the calibration curves were linear from 1 to 300 ng/mL. The accuracy and precision were within acceptable ranges, and the recovery was satisfactory. The method successfully detected and quantified the target analytes in all urine samples. The method was applied to urine samples from two patients with NPC and two patients with 3β-hydroxysteroid dehydrogenase (3β-HSD) deficiency,59) and eight healthy volunteers. The concentration of the sum of all three analytes was 400 times higher in the 3-month NPC patient and 40 times higher in the adult patient compared to healthy volunteers. In all subjects, the non-amidated form accounted for approximately one-third of the total target analytes.60)
Furthermore, we collected urine samples from 23 patients with NPC, 28 healthy controls, and 7 patients with other inherited metabolic disorders. We used receiver operating characteristic (ROC) analysis to evaluate the diagnostic performance of the three metabolites and their total concentration. The concentrations of the three metabolites and their total concentration, corrected for creatinine, were significantly higher in NPC patients compared to healthy controls. The area under the receiver operating curve (AUC) for all metabolites exceeded 0.95, with clinical specificity ranging from 92–100% and clinical sensitivity of approximately 95%. The concentrations of the metabolites were lower in the urine of patients with other inherited metabolic disorders than in the urine of patients with NPC. Conjugated cholesterol metabolites in urine can serve as useful diagnostic markers for noninvasive screening of NPC. These metabolites may reflect the pathology of NPC and the degree of cholesterol accumulation. Urinary SNAG-Δ5-CAs might be useful diagnostic biomarkers for noninvasive screening of NPC.61) However, the study's limitations include a small sample size and the need for further investigation to verify diagnostic performance (Fig. 4).
However, we also observed false negatives in some NPC urine samples (Fig. 4). Therefore, we aimed to identify more accurate diagnostic biomarker candidates. To accomplish this, we reviewed prior reports on deficiencies in enzymes related to cholesterol metabolism,62–65) we focused on the urinary cholesterol conjugate metabolome using MS/MS scan modes such as precursor ion scan and neutral loss scan in their lipidomics study.66)
Bile acids and steroid hormones are metabolites of cholesterol with important biological functions.67–69) They are frequently conjugated to other small molecules, such as amino acids70) sulfuric acid,71–73) or carbohydrates.73–75) This study investigated the effects of conjugations on the ionization and fragmentation behavior of metabolites under electrospray ionization (ESI) MS/MS. Several bile acids and steroid conjugates were analyzed under low-energy collision-induced dissociation (CID) conditions using a triple quadrupole tandem mass spectrometer. Various bile acids and steroid conjugates were synthesized or purchased and dissolved in water/ethanol solutions. An API 5000 mass spectrometer (SCIEX, Framingham, MA, U.S.A.) was used to perform ESI-MS and ESI-MS/MS with low-energy CID. The fragmentation patterns of the conjugates were analyzed by obtaining product ion spectra and analyzing the effects of collision voltages. Different conjugation types produce characteristic product ions and/or neutral losses in the product ion spectra. In this example, metabolites conjugated with glycine consistently produced a product ion at m/z 74, while those conjugated with taurine produced product ions at m/z 124, 107, and 80. Sulfated metabolites produced product ions at m/z 97 and 80, corresponding to HSO4− and SO3−, respectively. To identify carbohydrate conjugates, a combination of product ions and neutral losses was required, such as a product ion at m/z 161 and a neutral loss of 180 Da for glucosides and galactosides. Low-energy CID can provide valuable information about the structure and modification of bile acids and steroid conjugates. The product ion spectra can be used to discriminate among different conjugation types and to identify the conjugated moieties. Low-energy CID is a useful technique for characterizing bile acids and steroid conjugates based on their distinctive fragmentation patterns.76) And then, we proposed and applied the focused metabolomics analysis of urinary conjugated cholesterol metabolites from patients with NPC and 3β-HSD deficiency, using LC/MS/MS. Conjugated cholesterol metabolites are excreted in the urine of patients with metabolic abnormalities and hepatobiliary diseases, and they could serve as biomarkers for these genetic disorders.62,65,77,78) However, many biomarker candidates are still unknown, and reference standards are not available for all of them. We aimed to evaluate the effectiveness of precursor ion scan and neutral loss scan in characterizing conjugated cholesterol metabolites in urine. The objective of this study was to evaluate the effectiveness of precursor ion scan and neutral loss scan techniques in detecting characteristic fragment patterns of different conjugated cholesterol metabolites. Additionally, we applied these techniques to analyze urine samples from patients diagnosed with NPC and 3β-HSD deficiency, two conditions that impact cholesterol catabolism and transport. The paper's methods involved preparing a mixture of authentic standards of conjugated cholesterol metabolites, collecting urine samples from both patients and healthy controls, conducting LC/MS/MS analysis using different scan modes, and identifying detected peaks based on their retention times and m/z values. The study results indicate that precursor ion scan and neutral loss scan were effective in extracting conjugated cholesterol metabolites with common partial structures, such as glycine, taurine, sulfate, glucuronide, glucose, galactose, and GlcNAc. The urine samples of NPC and 3β-HSD deficiency patients contained numerous conjugated cholesterol metabolites, including several disease-specific intense peaks. Intense peaks were found in the urine of the NPC patient, and in the urine of the 3β-HSD deficiency patient, some peaks were tentatively identified as 5-cholenoic acid sulfates and their glycine and taurine conjugates. This paper demonstrates the potential of precursor ion scan and neutral loss scan for focused metabolomics analysis of conjugated cholesterol metabolites in urine. The study reveals novel urinary biomarkers for NPC and 3β-HSD deficiency, two disturbances of cholesterol catabolism and transport. The LC/MS/MS precursor ion scan and neutral loss scan are useful tools for characterizing conjugated cholesterol metabolites in urine79) (Fig. 5).
Next, we aimed to identify and characterize the three unknown peaks they had detected in the urine of an NPC patient in their previous paper,79) and to evaluate their suitability as novel candidate diagnostic markers for NPC. The structures of the unknown compounds were estimated based on their accurate masses, retention times, and possible biosynthetic pathways. Subsequently, the two most likely candidates, 3β-sulfooxy-7β-hydroxy-5-cholen-24-oic acid (S7B-Δ5-CA) and 3β-sulfooxy-7-oxo-5-cholen-24-oic acid (S7O-Δ5-CA), were synthesized and compared to the target compounds using LC/MS/MS with high-resolution mass spectrometry (HRMS). The isotopic patterns and product ion spectra of the deprotonated molecules were analyzed. The two target compounds found in the urine of the NPC patient were confirmed to be the synthesized authentic standards based on the agreement of retention times, accurate masses, isotopic patterns, and product ions. The two sulfated cholesterol metabolites are better diagnostic markers for NPC than the previously reported multi-conjugated cholesterol metabolites because they are not influenced by the activity of UDP-glycosyltransferase 3A1, a gene mutated in some NPC patients.80) Sulfate conjugation at the C-3 position facilitates urinary excretion of these compounds, making them easier to detect. Using a focused metabolomics approach and the synthesis of authentic standards, we identified two novel candidate urinary biomarkers for NPC: S7B-Δ5-CA and S7O-Δ5-CA81) (Fig. 5).
Next, we investigated the diagnostic performance of five urinary cholesterol metabolites, including two newly identified molecules, S7B-Δ5-CA and S7O-Δ5-CA, for NPC. The objective was to evaluate the diagnostic marker performance of five urinary conjugated cholesterol metabolites previously identified as potential biomarkers for NPC. These metabolites are SNAG-Δ5-CA, SNAG-Δ5-CG, SNAG-Δ5-CT, S7B-Δ5-CA, and S7O-Δ5-CA, respectively. A method for analyzing five urinary metabolites simultaneously was developed using LC/MS/MS with a column-switching system. The analytical method was validated for linearity, accuracy, precision, matrix effects, stability, and dilution. The urinary concentrations of the five metabolites were analyzed in 38 healthy controls and 28 patients with NPC. Statistical and ROC analyses were performed to compare the groups and assess the diagnostic performance of each metabolite. The analytical method demonstrated high accuracy and met all validation criteria. Patients with NPC had significantly higher urinary concentrations of the five metabolites, except for SNAG-Δ5-CT, compared to healthy controls. According to the ROC analysis, all metabolites exhibited excellent diagnostic marker performance, with AUC values ranging from 0.916 to 1.0. S7B-Δ5-CA demonstrated the best performance, with 100% sensitivity and specificity, and no overlap between the groups. The five urinary conjugated cholesterol metabolites are reliable biomarkers for NPC diagnosis. The given biomarkers reflect the cholesterol accumulation and metabolic abnormalities caused by NPC pathology.82) They offer advantages over other biomarkers due to their noninvasive, convenient, and stable nature. S7B-Δ5-CA is a particularly valuable biomarker that can prevent false negatives resulting from the UGT3A1 mutation, which affects the GlcNAc conjugation of some metabolites.80) These metabolites, particularly S7B-Δ5-CA, could be utilized for screening and monitoring of NPC in the future.83)
However, the developed LC/MS/MS method requires long separation times and the analysis of large volumes of urine.84) The objective of the next study was to develop a highly sensitive and rapid LC/MS/MS method for the analysis of five urinary cholesterol metabolites using a basic mobile phase additive. This additive can increase the MS intensity85,86) and shorten the LC running time.87–89) The performance of this method was compared with the previously developed method. Five urinary cholesterol metabolites (SNAG-Δ5-CA, SNAG-Δ5-CG, SNAG-Δ5-CT, S7B-Δ5-CA, and S7O-Δ5-CA) and IS were measured by LC/MS/MS with selected reaction monitoring (SRM) in negative ion mode. A column switching LC/MS/MS system with online pretreatment was used. The mobile phases contained a basic additive of 1% (v/v) ammonium solution. We used this method to perform calibration curves, matrix effects, intra- and inter-day reproducibility, and urine analysis of healthy subjects and NPC patients. We compared the results with those obtained using the previous method.83) The basic mobile phase additive increased the peak areas of the analytes and IS by an average of 16-fold compared to the previous method, resulting in a shorter separation time of 7 min instead of 60 min. The method showed good linearity, reproducibility and reliability with no significant matrix effects. The urinary concentrations of the metabolites obtained by this method correlated well with those obtained by the previous method, and the diagnostic performance was similar or slightly better for NPC screening. This study demonstrates the usefulness of the basic mobile phase additive for improving the sensitivity and speed of LC/MS/MS analysis of urinary cholesterol metabolites as biomarkers for NPC. The developed method is suitable for non-invasive diagnostic screening of NPC patients. It requires less urine volume and analysis time, and provides reliable results. A highly sensitive and rapid LC/MS/MS method for the analysis of cholesterol metabolites in urine was successfully developed using a basic mobile phase additive.84) The method is useful for non-invasive diagnostic screening of NPC patients (Fig. 6).
We have attempted to develop blood biomarkers because of the challenges of collecting urine samples from neonates and young infants. Our goal was to determine the chemical structure of lysosphingomyelin-509 (Lyso-SM-509), a lipid metabolite that was found to be significantly elevated in the plasma of NPC patients, but whose structure was previously unknown.90) We aimed to investigate the biosynthetic pathway and biological function of Lyso-SM-509 and related lipids in NPC. Various mass spectrometric techniques, including HRMS, CID, and hydrogen abstraction dissociation (HAD), were used to analyze the fragmentation patterns and partial structures of Lyso-SM-509.91–93) Chemical derivatization experiments were performed to determine the functional groups of Lyso-SM-509. In addition, we synthesized the authentic standard of N-palmitoyl-O-phosphocholine serine (PPCS) and confirmed its identification by comparing its chromatographic and spectral properties with those of serum Lyso-SM-509 from NPC patients. Finally, we performed a targeted lipidomics analysis to measure the levels of Lyso-SM-509 and other N-acyl-O-phosphocholine-serines in the serum/plasma of NPC patients and controls. The results showed that Lyso-SM-509 is a novel class of lipid called PPCS, which has serine as its backbone. It was found that a group of N-acyl-O-phosphocholine-serines with different fatty acid chain lengths were significantly increased in the serum/plasma of NPC patients.94) Conversely, some lysophosphatidylcholines were significantly decreased. It is speculated that N-acyl-O-phosphocholine-serines may be derived from N-acyl-phosphatidylserines, which are rare phospholipids that may accumulate in the lysosomes of NPC cells. We proposed that the addition of phosphocholine to N-acyl-serines may be a novel metabolic reaction occurring in NPC. We identified the structure of Lyso-SM-509 and discovered a new group of lipids that are highly specific and sensitive biomarkers for NPC (Fig. 7). In addition, we provided new insights into the dysregulation of lipid metabolism and the pathophysiology of NPC. These findings may contribute to the development of accurate diagnostic methods and novel therapeutic targets for NPC. It was concluded that Lyso-SM-509 is PPCS, a novel lipid that is significantly elevated in NPC patients. In addition, a group of N-acyl-O-phosphocholine-serines were identified that are also increased in NPC. A possible biosynthetic pathway and biological function of these lipids in NPC were proposed.
NBD, 4-fluoro-7-nitro-2,1,3-benzoxadiazole; NPC, Niemann–Pick disease type C. Cited and modified reference number.94)
We have developed a diagnostic screening strategy for Niemann–Pick diseases (NPDs), rare genetic disorders that affect the metabolism of sphingolipids and cholesterol in NPC and acid sphingomyelinase deficiency (ASMD).95–98) Sphingomyelin and sphingosylphosphorylcholine (SPC) accumulate in ASMD, making SPC a diagnostic biomarker. PPCS and SPC concentrations may be useful in differentiating between NPC and ASMD. However, the levels of PPCS and SPC increase differently in NPC and ASMD. Therefore, PPCS and SPC concentrations can be used to detect and differentiate between NPC and ASMD. Two simultaneous assays for PPCS and SPC in serum/plasma were developed. The synthesized PPCS and its isotope-labeled internal standard (PPCS-2H3) were used. LC/MS/MS conditions were optimized for PPCS and SPC, calibration curves and quality control samples were prepared, and serum samples from healthy subjects and patients with NPDs were analyzed using two LC/MS/MS methods: a rapid method (method 1) and a validated method (method 2). The LC/MS/MS methods for PPCS and SPC showed high linearity, accuracy and precision. Serum PPCS and SPC concentrations were significantly elevated in NPC patients compared to healthy subjects. There was a high correlation of PPCS concentrations between methods 1 and 2, but a low correlation of SPC concentrations between methods 1 and 2. PPCS concentration showed high sensitivity and specificity for differentiating NPD patients from healthy subjects. The serum PPCS/SPC ratio differed between NPC patients, ASMD patients and healthy subjects. The combination of the two LC/MS/MS methods for PPCS and SPC may be useful for the diagnostic screening of NPDs. It can detect NPDs at an early stage and differentiate NPC from ASMD. It has been suggested that PPCS and SPC are easier to analyze than cholesterol metabolites, which are other biomarkers of NPD. It is important to maintain objectivity and avoid subjective judgments,82,99) and it was concluded that the proposed screening strategy for NPDs is simple, rapid, and accurate and can be applied in clinical practice and research due to its high disease specificity. We also showed the limitations of the study, such as the small number of ASMD samples, the unclear metabolic pathway of PPCS, and the need for further validation of the cut-off values for PPCS concentration and PPCS/SPC ratio.100)
2.2. Biomarker Discovery and Diagnostic Modeling for Cancer through Metabolomics2.2.1. Development of Diagnostic Biomarkers for GliomaWe have developed a chemical diagnostic method based on metabolite analysis using LC/MS/MS. Initially, we focused on glioma, a malignant brain tumor. Glioma can be differentially diagnosed by a genomic mutation in isocitrate dehydrogenase 1 (IDH1)/IDH2 and chromosome 1p/19q.101) The prognosis of glioma varies depending on the specific type. Therefore, it is crucial to diagnose the genome type in order to determine the appropriate strategy for glioma resection.102) However, in routine medical practice genomic analysis of brain samples is based on the pathological samples and they are applied to PCR analysis,103) immunohistochemical staining,104) magnetic resonance spectroscopy.105)
To detect IDH1 mutations, we aimed to use a mass spectrometry-based method. Resected brain tissue samples from glioma patients with or without IDH1/2 mutations were used. We focused on 2-hydroxyglutaric acid (2-HG), the oncometabolite produced by IDH1 mutation.106) The aim of this study was to establish a method for intraoperative molecular diagnosis of IDH mutation by measuring the level of 2-HG in tumor tissue using LC/MS/MS). To achieve this, we constructed an analytical method for brain 2-HG using rapid pretreatment with bead homogenization and ultrafiltration, followed by LC/MS/MS using a hydrophilic Capcell pak ADME reversed-phase LC column. Our team successfully developed a rapid LC/MS/MS method that can be completed in less than 5 min per run. The aim of this study was to develop a simple, rapid and sensitive method to provide useful information for surgical decision making and integrated diagnosis in patients with diffuse infiltrating glioma. The level of 2-HG was measured in 105 tumor tissues by LC/MS/MS. The study compared the level of 2-HG between IDH mutant and wild-type gliomas and evaluated the diagnostic accuracy of this method using receiver operating characteristic curve analysis. The correlation between 2-HG levels and tumor grade, chromosome 1p/19q codeletion, and glioma subtype based on the 2016 WHO classification was investigated. It was found that IDH mutant gliomas have significantly higher levels of 2-HG than IDH wild type gliomas. Measurement of 2-HG can detect IDH mutation with high sensitivity and specificity. It was found that the level of 2-HG is not related to tumor grade, chromosome 1p/19q codeletion, or glioma subtype, and it cannot predict the integrated diagnosis according to the WHO classification. The study demonstrated that the measurement of 2-HG by LC/MS/MS can rapidly and accurately determine the IDH mutation status in glioma tissues. This information may be valuable for intraoperative diagnosis and surgical strategy. The use of LC/MS/MS to measure 2-HG is a reliable and feasible method for detecting IDH mutation in glioma tissue during surgery. However, it has limited value in predicting tumor grade and glioma subtype. To achieve an integrated diagnosis for glioma surgery, this method can be combined with other molecular and histological analyses.107)
Next, the challenges and limitations of distinguishing astrocytomas from oligodendrogliomas based on histologic and molecular features are discussed. It also highlights the need for a quick and easy molecular diagnostic method that can be used during surgery. To differentiate between gliomas, the author attempted to use 1p/19q codeletion detection. In routine pathological analysis of gliomas, fluorescence in situ hybridization and next-generation sequencing are commonly used to detect 1p/19q codeletion.108,109) We attempted to identify the metabolite changes associated with 1p/19q codeletion using global metabolomics (G-Met) analysis. Our goal was to develop a quantitative molecular diagnostic method based on these findings. Analytical experiments were performed, including sample collection and preparation, LC/MS/MS equipment and conditions, data processing and multivariate analysis, peak identification and quantification, statistical analysis, prediction model development, and performance evaluation of candidate biomarkers. The study started with 26 astrocytomas without 1p/19q codeletion and 23 oligodendrogliomas with 1p/19q codeletion. The brain tissue samples were subjected to homogenization and protein precipitation treatment. The treated samples were then analyzed using G-Met, a non-targeted analysis developed by Saigusa et al.110) The ions were comprehensively detected and applied to Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA) to identify the separating biomarker metabolites. The proposed metabolites were then identified using purchased authentic standards. The paper presents the key findings and results of the study, including the identification of five candidate biomarkers: hypoxanthine, inosine, cystine, proline and uric acid by G-Met.110) We developed an LC/MS/MS analytical method for five candidate biomarkers. We determined metabolite levels in brain tumor samples and found a significant difference in cystine levels between astrocytomas and oligodendrogliomas. In addition, we developed and validated two predictive models to discriminate between the two gliomas using logistic regression analysis. It highlights the discovery of cystine as a potential differential biomarker for astrocytomas and oligodendrogliomas, the feasibility and utility of LC/MS/MS for metabolite analysis and molecular diagnosis of gliomas, and the possible biological roles and mechanisms of cystine and other metabolites in glioma pathogenesis and progression. The paper also acknowledges the limitations and challenges of the study, including the small sample size, single-center design, and the need for further validation and exploration of unknown metabolites111) (Fig. 8).
RCC is typically associated with a poor prognosis and outcome.112) The diagnosis of RCC is usually based on imaging tests, although there are some atypical cases that can make diagnosis difficult.113,114) Another method of obtaining tissue samples is biopsy, although it should be noted that this method is invasive.115) Next, we conducted a search for diagnostic biomarkers for RCC by analyzing metabolites in tissues. Because cancer metabolism, including oncometabolites, has been previously described,106,116,117) we also focused on the metabolites in RCC. As a metabolomic analysis, similar to the previous case, we first applied the G-Met method.111) We used the resected kidney organ to prepare specimens from both tumor and non-tumor regions. A total of 20 pairs of tumor and non-tumor regions were collected from kidney patients. The homogenized samples were analyzed for G-Met and the data were subjected to multivariate analysis to identify candidate RCC diagnostic biomarkers. Fifty-eight metabolites were found to be significantly elevated in the tumor region. Of these, 34 metabolites with AUC >0.8 in ROC analysis by peak intensity were considered useful for potential early diagnosis of RCC. The 34 identified metabolites with ROC >0.8 are mainly categorized in the glycerophospholipid pathway. The text already appears to meet the desired characteristics and is free of grammatical, spelling and punctuation errors. Therefore, no changes were made to the original text. The following metabolites were identified: digalactosyl diacylglycerol (36 : 6) in the lipid pathway, acyl-carnitine (C14 : 1) and acyl-carnitine (C16 : 1) in the carnitine pathway, alpha-tocopherol and gamma-tocopherol in the tocopherol pathway, glucose-1-phosphate and fructose-6-phosphate in the glycolytic pathway, N-formylkynurenine and L-kynurenine in the tryptophan pathway, and trigonelline, S-lactoylglutathione, 3-methoxybenzenepropanoic acid, thymidine glycol, maltotriose, maltotetraose, and 2-hydroxyglutarate in other pathways. Nineteen metabolites were found to correlate with clinicopathologic factors, such as Fuhrman grade, clinical M stage, and coagulation necrosis, and were considered useful for malignancy assessment. Pathways with AUC <0.8 include tryptophan metabolism, tricarboxylic acid (TCA) cycle, glycine metabolism, alanine and aspartate metabolism, nucleotide sugar metabolism, and inositol metabolism. In summary, we have identified 58 metabolites and 21 pathways that show significant differences between tumor and non-tumor tissues of RCC patients using G-Met. We found that 34 metabolites, mainly involved in glycerophospholipid, glutathione, glycoglycerolipid, carnitine and glycolysis pathways, had an AUC greater than 0.8. These metabolites may be useful in the early diagnosis of RCC. In addition, we found that 19 metabolites with AUC less than or equal to 0.8 were associated with clinicopathological factors such as tumor volume, pathological T stage, Fuhrman grade, clinical M stage, and coagulation necrosis. These metabolites may reflect the malignant status of RCC. It was suggested that the TCA cycle, TCA cycle intermediates, nucleotide sugar pathway, and inositol pathway are characteristic pathways for the malignant status of RCC. The study concluded that important metabolites and pathways were revealed that could discriminate between early diagnosis and malignant status of RCC. Further evaluation of multiple metabolites and their pathways are needed in the future.118)
With the goal of developing a diagnostic method for body fluids, we quantified the concentration of each potential molecule in urine and attempted to create a model. The aim of this study is to measure the levels of metabolites in urine samples from patients with clear cell RCC (ccRCC) and controls using the SRM mode of LC/MS/MS with internal standards and urine creatinine adjustment. In addition, predictive models for ccRCC diagnosis and clinical stage based on urinary metabolite concentrations will be developed to evaluate the potential of urinary metabolites as non-invasive biomarkers for diagnosis and malignant status of ccRCC. The concentrations of 33 metabolites were measured in urine samples from 87 ccRCC patients and 60 controls using LC/MS/MS with SRM mode. Four analytical methods were used independently. This paper describes four analytical methods for measuring urinary metabolites associated with RCC using LC/MS/MS. The metabolites and columns used for each method are as follows: Group 1 analyzes glutathione, lactic acid, 2-HG, glutamine, glutamic acid, phosphorylcholine, glycerophosphorylcholine, 2-oxoglutaric acid, acetyl CoA, succinic acid, ophthalic acid, D-glucose 1-phosphate, and D-fructose 6-phosphate. The metabolites D-galactose, S-lactoylglutathione, myoinositol, D-sedoheptulose 7-phosphate, 3-methoxybenzenepropanoic acid, D-saccharic acid, and N-hexanoylglycine were analyzed using two different methods. Group 1 used a CAPCELL PAK ADME column with a water/formic acid and acetonitrile gradient, while Group 2 used a CORTECS silica column with an ammonium formate/formic acid and acetonitrile gradient. Group 1 used a CAPCELL PAK ADME column with a water/formic acid and acetonitrile gradient, while Group 2 used a CORTECS silica column with an ammonium formate/formic acid and acetonitrile gradient. Group 3: This method analyzes various compounds including N-formylkynurenine, cinnabarinic acid, L-tryptophan, L-kynurenine, xanthurenic acid, 3-hydroxykynurenine, 5-hydroxyanthranilic acid, picolinic acid, nicotinic acid, anthranilic acid, 3-hydroxyanthranilic acid, quinolinic acid, indole-3-acetic acid, kynurenic acid, and N-formylanthranilic acid. This method analyzes several compounds including N-formylkynurenine, cinnabarinic acid, L-tryptophan, L-kynurenine, xanthurenic acid, 3-hydroxykynurenine, 5-hydroxyanthranilic acid, picolinic acid, nicotinic acid, anthranilic acid, 3-hydroxyanthranilic acid, quinolinic acid, indole-3-acetic acid, kynurenic acid, and N-formylanthranilic acid. The analysis is performed using a CAPCELL PAK ADME column with an acetylacetone/formic acid in water and acetonitrile gradient. Group 4: This method analyzes carnitine, acetylcarnitine, propionylcarnitine, butyrylcarnitine, pivaloylcarnitine, hexanoylcarnitine, octanoylcarnitine, decanoylcarnitine, lauroylcarnitine, myristoylcarnitine, palmitoylcarnitine and stearoylcarnitine using an Inertsil ODS-3 column with an HFBA gradient in water and acetonitrile.119) After quantifying metabolites using four methods, we performed multiple logistic regression and ROC analyses to develop predictive models for the diagnosis of ccRCC and clinical stage III/IV. We identified five metabolites (L-glutamic acid, lactate, D-sedoheptulose-7-phosphate, 2-hydroxyglutarate, and myoinositol) for the diagnostic model and four metabolites (L-kynurenine, L-glutamine, fructose-6-phosphate, and butyrylcarnitine) for the clinical stage model. The diagnostic model achieved an AUC of 0.966, while the clinical stage model achieved an AUC of 0.837. In addition, the concentrations of four of the five diagnostic metabolites decreased significantly after surgery. In conclusion, we have demonstrated that urinary metabolites can serve as clinically useful biomarkers for the diagnosis and malignant status of ccRCC. Our predictive models may improve patient outcomes by enabling noninvasive screening and facilitating treatment decisions.120) Recurrence after surgical resection is an important issue in RCC. Metabolomic analysis was used to find predictive biomarker candidates. The aim of this study was to identify urinary metabolites that can predict recurrence of RCC after surgery using a precise quantitative measurement method. Data were collected from patients who underwent definitive surgery for RCC and benign urological tumors at the Department of Urology, Tohoku University Hospital, between November 2016 and December 2019. The concentrations of 30 urinary metabolites were measured before and after surgery using an LC/MS/MS method. A comparison was made between the preoperative and postoperative metabolite concentrations of the RCC and benign tumor groups, as well as the RCC recurrence and non-recurrence groups. Multiple logistic regression analysis was performed to identify the metabolites that could predict RCC recurrence, and a prediction model was built based on these metabolites. A significant difference between pre- and postoperative concentrations of 10 urinary metabolites was observed when comparing 27 patients with T1a RCC and 10 patients with benign tumors. Of the 10 metabolites examined, logistic regression analysis identified five (lactic acid, glycine, 2-hydroxyglutarate, succinic acid, and kynurenic acid) as factors in the development of the recurrence prediction model. The prediction model had an area under the ROC curve of 0.894 with a sensitivity of 88.9% and a specificity of 88.0%. When stratified into low and high recurrence risk groups based on this model, we observed a significant decrease in recurrence-free survival rates in the high risk group. We have shown that accurate measurement of urinary metabolites before and after surgery can predict recurrence of RCC. These metabolites may serve as non-invasive biomarkers to improve patient outcomes. We have also suggested that the selected metabolites have oncological significance as they are involved in the metabolic pathways related to RCC progression and metastasis121) (Fig. 9). These studies may contribute to the development of personalized medicine and follow-up strategies for RCC patients by providing new insights into the metabolic alterations of RCC.118,120,121)
Cited and modified reference number.118)
Traditional pathologic analysis relies primarily on morphologic changes in tissues and cells. Although invaluable, this approach can miss subtle biochemical changes that often underlie disease progression. Recent advances in analytical techniques, particularly LC/MS/MS and metabolomic analysis, offer exciting opportunities to delve deeper into the metabolic landscape of diseased tissues and provide a more comprehensive understanding of disease etiology and pathophysiology. LC/MS/MS is a highly effective analytical tool that enables sensitive, high-throughput detection of a wide range of metabolites, including amino acids, lipids, carbohydrates, and small molecules.49,122,123) This detailed metabolite profile, similar to a biochemical fingerprint, can reveal metabolic disorders associated with various pathologies.124–126) Unlike conventional methods, LC/MS/MS provides quantitative data, allowing researchers to identify statistically significant changes in metabolite levels and track their changes over time or in response to therapeutic interventions.
LC/MS/MS has several advantages in pathological analysis of in vitro cell line samples and animal models of disease. (1) First, it can detect subtle metabolic changes that often precede morphological changes, making it a promising tool for early disease detection. (2) Second, it can aid in the characterization of disease subtypes by identifying distinct metabolic profiles, which can aid in personalized diagnosis and targeted treatment strategies. (3) The identification of key metabolic pathways involved in disease progression can provide valuable insights into disease pathogenesis and pave the way for the development of novel therapeutic targets. This can be achieved by using LC/MS/MS to understand disease mechanisms. (4) Tracking changes in the metabolome can provide real-time information on the efficacy of a given treatment, allowing rapid adjustment of therapeutic strategies when needed. (5) Metabolomic analysis: metabolomic analysis involves the comprehensive study of the complete metabolite profile of a cell or organism. This approach provides a deeper understanding of the complex interplay between genetic, environmental and physiological factors that contribute to disease development. By integrating LC/MS/MS data with other omics data, such as genomics and transcriptomics, researchers can gain a more complete understanding of the molecular basis of disease. This can lead to the development of new diagnostic and therapeutic approaches. Our results on pathological samples using LC/MS/MS will be presented.
3.1. Analysis of Variability of Steroid Hormone Metabolism in NPC Model CellsNPC is a rare genetic disorder that affects cholesterol transport and causes various symptoms, especially neurological ones.35,37,127) Steroid hormones, which are bioactive molecules derived from cholesterol and synthesized in mitochondria and endoplasmic reticulum, are involved in many physiological functions.128–130) Previous studies have reported that NPC causes impaired autophagy, oxidative stress, reduced ATP production, and decreased mitochondrial membrane potential in cells.131–134) Therefore, we hypothesized that NPC may affect steroid hormone production due to mitochondrial dysfunction and aimed to develop a comprehensive analytical method to measure steroid hormone levels in NPC model cells. The purpose of this study was to develop a method for the simultaneous analysis of 13 steroid hormones using LC/MS/MS and to investigate the metabolic changes of steroid hormones in NPC model cells. The method of the study can be summarized as follows: A simultaneous LC/MS/MS analysis method was developed for 13 steroid hormones without derivatization and liquid–liquid extraction pretreatment. The MS/MS and LC conditions for 13 steroid hormones were optimized and the reliability of the analytical method was validated for both cells and media. An Npc1-deficient CHO cell line was used as a model cell for NPC. Cell characteristics and morphology were verified by filipin staining and organelle-ID-RGB staining. Steroid hormone levels in the NPC model and wild-type cells and media were analyzed using the developed LC/MS/MS method, and the results were compared. The NPC phenotype and mitochondrial abnormalities in the NPC model cells were confirmed by filipin staining and organelle staining. The levels of six steroid hormones (testosterone, androsterone, progesterone, estrone, cortisone, and aldosterone) were significantly decreased in the NPC model cells and media compared with those in the wild-type cells and media. It is suggested that the decreased steroid hormone levels are related to the impaired mitochondrial function in NPC and that steroid hormone metabolism may be altered in NPC patients135) (Fig. 10). Steroid hormone analysis has the potential to be a novel approach to elucidate the molecular mechanisms and therapeutic targets of NPC, a rare and fatal lysosomal disorder.
Cited and modified reference number.135)
Nonalcoholic steatohepatitis (NASH) is a liver disease characterized by inflammation and fibrosis of the liver.136–138) NASH is a subtype of non-alcoholic fatty liver disease (NAFLD) that affects up to 20–30% of adults worldwide and is associated with obesity, insulin resistance and dyslipidemia. NASH can progress to cirrhosis and hepatocellular carcinoma, leading to increased morbidity and mortality. However, there is no specific and effective treatment for NASH. The exact pathogenesis of NASH is unclear but involves multiple factors such as lipotoxicity, oxidative stress, endoplasmic reticulum stress, mitochondrial dysfunction and inflammation. Bile acids are endogenous molecules derived from cholesterol catabolism in the liver. They play important roles in lipid digestion, cholesterol homeostasis, and metabolic signaling. Alterations in bile acid composition and disposition have been reported in several metabolic diseases, including NASH. Bile acids may act as potential mediators of liver injury and fibrosis in NASH.139,140) The purpose of this paper is to investigate the relationship between bile acid composition and disposition and NASH-associated fibrosis using a mouse model. NASH is a progressive inflammatory and fibrotic disease that can lead to cirrhosis and hepatocellular carcinoma. Bile acids are endogenous molecules that regulate cholesterol homeostasis, lipid solubilization, and metabolic signaling. In this study, we hypothesized that alterations in bile acid profile may be closely related to the pathogenesis of NASH and liver fibrosis. We used a mouse model of NASH induced by feeding a choline-deficient, L-amino acid-defined, high-fat diet containing 0.1% methionine (CDAHFD) for 3 to 15 weeks.141–143) We measured body weight, energy intake, liver weight, plasma and liver biochemical parameters, and bile acid composition in the liver, bile, and plasma of the mice. We performed histopathological evaluation of liver tissue using hematoxylin–eosin (H&E) and Elastica–Masson’s trichrome (EM) staining, and scored the NASH severity (NAS) score for steatosis, inflammation, ballooning, and fibrosis according to the criteria of Kleiner et al.144) We also performed quantitative PCR and Western blotting to analyze the expression of genes and proteins related to inflammation, fibrosis, and bile acid synthesis and transport in the liver and intestine of the mice. LC/MS/MS was used to quantify 54 bile acids and their conjugates in liver, bile and plasma samples. We found that feeding CDAHFD to mice induced NASH-associated hepatic fibrosis and altered bile acid composition and disposition in liver, bile, and plasma. The CDAHFD-fed mice developed hepatic steatosis, inflammation, ballooning and fibrosis within 15 weeks as shown by histological analysis and biochemical markers. We also measured the expression of genes and proteins related to inflammation and fibrosis, such as Tnf-α, F4/80, CD11c, Mcp-1, Mmp-2, Timp-1, Tgf-β1, Collagen1α1, and α-smooth muscle actin (α-SMA), and found that they were significantly increased in the CDAHFD-fed mice compared with the control mice. We analyzed the bile acid profiles in the liver, bile, and plasma of the mice using LC/MS/MS. We found that the CDAHFD-fed mice had altered bile acid composition in all three compartments, with increased levels of hydrophilic unconjugated and conjugated bile acids and decreased levels of hydrophobic secondary bile acids. We also measured the expression of genes related to bile acid synthesis, conjugation, and sulfation, such as Cyp7a1, Cyp27a1, Cyp8b1, Akr1d1, Hsd3b7, Baat, and Sult2a1, and found that they were significantly decreased in the CDAHFD-fed mice compared with the control mice. We also measured the expression of genes related to bile acid transport and reabsorption, such as Asbt, Ntcp, Oatp1b2, Bsep, Abcc2, Abcc3, Abcc4, and Ostβ, and found that they were differentially regulated in the CDAHFD-fed mice compared with the control mice. In conclusion, we suggested that the CDAHFD-fed mice had a shift in the bile acid efflux pathway from the bile to the peripheral plasma, which may be a protective mechanism to reduce hepatotoxicity, but may also promote liver fibrosis.145) Next, we developed an analytical method using LC/MS/MS to measure bile acids and sterols, which are derived from cholesterol and have various physiological functions. We applied this method to liver samples from NASH model mice fed with CDAHFD for different periods of time from the previous work.145) We found that some bile acids and sterols were altered before the onset of NASH and suggested that these metabolites may be involved in the pathological mechanism and progression of NASH. We also proposed that these metabolites may be potential biomarkers for the early detection of NASH. The purpose of this study is to investigate the changes in cholesterol metabolites, such as bile acids and sterols, in the liver of NASH model mice and to evaluate their relationship with the pathological progression of NASH. This paper describes the methods used to analyze cholesterol metabolites in the liver of mice with NASH using LC/MS/MS. Mice were fed a choline-deficient, methionine-reduced high-fat diet (CDAHFD) for 3, 7, 14, or 21 d to induce NASH. Liver tissues were homogenized and subjected to solid-phase extraction for bile acids and liquid-liquid extraction for sterols. The extracted samples were then derivatized with picolinic acid (PA) for 17 sterols or left alone for 46 bile acids. Bile acids and sterols were separated and detected by LC/MS/MS using conditions optimized for each group of metabolites. Quantification was performed using calibration curves and internal standards. Levels of cholesterol metabolites were calculated and compared between control and CDAHFD groups using multivariate and correlation analysis. Changes in cholesterol metabolites were also correlated with pathological scores and biochemical tests of NASH. First, we developed a comprehensive analytical method for 46 bile acids using LC/MS/MS. We optimized the mobile phase composition, gradient elution, and SRM parameters to achieve good separation and sensitivity for different bile acids, including conjugated, sulfated, and glucuronidated forms. We validated the analytical method for bile acids using activated charcoal-treated mouse liver as a surrogate matrix.146,147) We evaluated the linearity, accuracy, precision, and recovery of the method for each bile acid. We found that most of the bile acids showed good linearity (R2 > 0.99) and reproducibility (RE and CV <15%) within the calibration range. We developed an analytical method for 17 sterols using LC/MS/MS with PA derivatization.148–150) We optimized the PA derivatization reaction and SRM parameters to increase the ionization and fragmentation efficiency of sterols. We detected [M + PA-H2O + Na]+ or [M + 2PA-2H2O + Na]+ ions from the PA sterol derivatives. We optimized the LC conditions for PA-derivatized sterols using a SunShell C30 column, which has high separation selectivity for sterically bulky compounds. Separation of the sterol isomers was achieved. We validated the analytical method for sterols using activated charcoal-treated mouse liver as a surrogate matrix. We evaluated the linearity, accuracy, precision, and recovery of the method for each sterol. We found that most of the sterols showed good linearity (R2 > 0.99) and reproducibility (RE and CV <15%) within the calibration range. We established NASH model mice by feeding them CDAHFD for 3, 7, 14, and 21 d. We confirmed the development of NASH by measuring the body weight, liver weight, hepatic lipid levels, and histopathological findings of the mice. We found that the CDAHFD-fed mice showed significant weight loss, liver enlargement, lipid accumulation, inflammation and fibrosis compared to the control mice. We analyzed the changes in cholesterol metabolites, including 46 bile acids and 17 sterols, in liver samples from NASH model mice using the developed LC/MS/MS methods. We found that the levels of bile acids and sterols changed significantly with the progression of NASH. We observed a decrease in desmosterol, campesterol, sitosterol, and some secondary bile acids and an increase in some primary bile acids, conjugated bile acids, and oxysterols in the CDAHFD-fed mice compared to the control mice. We performed multivariate analysis, including principal component analysis and orthogonal projection to latent structure discriminant analysis, to identify the characteristic changes in cholesterol metabolites at different stages of NASH. We found that the levels of triglycerides, cholesterol, conjugated bile acids, and oxysterols were the most characteristic metabolites of the CDAHFD group, and that these metabolites were associated with NASH pathology (Fig. 11). We proposed that these metabolites may be useful as biomarkers for early detection of NASH.151)
This review presents LC/MS/MS methods for the analysis of endogenous metabolites in patient and disease model samples. Diagnostic biomarkers for NPC, glioma and RCC are discovered and metabolic alterations in NPC and NASH are analyzed. The author and collaborators identified urinary conjugated cholesterol metabolites, serum PPCS and SPC as useful biomarkers for NPC and ASMD. They demonstrated the clinical utility of these biomarkers for the chemical diagnosis of NPC and the potential for neonatal screening for NPC and ASMD. In addition, 2-HG and cystine were found to be useful differential biomarkers for glioma subtypes. In the analysis of RCC, some urinary metabolites were found to reflect the malignant status and recurrence risk of RCC. We showed changes in steroid hormones and cholesterol metabolites in NPC model cells and NASH model mice, indicating their involvement in the pathophysiology and progression of these diseases. We concluded that their analytical methods and findings could contribute to the advancement of clinical practice and basic research based on metabolomics.
I would like to thank Prof. Nariyasu Mano (Tohoku University), Prof. Emeritus Junichi Goto (Tohoku University), Prof. Tomoyuki Oe (Tohoku University). I would like to thank all the members of the Laboratory of Biomolecule and Pathophysiological Chemistry, Graduate School of Pharmaceutical Sciences, Tohoku University and the Department of Pharmaceutical Sciences, Tohoku University Hospital. We thank all the collaborators in our study. The studies in this review were supported in part by Grants from the Japanese Society for the Promotion of Science (JSPS) for Young Scientists (B) [KAKENHI: 16K20900], Young Scientists [KAKENHI: 18K15699], and Scientific Research (C) [KAKENHI: 21K07814], respectively. The studies in this review were also supported in part by Grants from the Kurozumi Medical Foundation and the Chemical Evaluation and Research Institute and the Kawano Masanori Memorial Public Interest Incorporated Foundation for Promotion of Pediatrics, the Japanese Society of Inherited Metabolic Disease/Sanofi LSD Research Grant, and the Clinical Research Promotion Program for Young Investigators of Tohoku University Hospital and the Research Promotion Program for Collaboration of Young Investigators of Tohoku University Graduate School of Medicine.
The author declares no conflict of interest.
This review of the author’s work was written by the author upon receiving the 2023 Pharmaceutical Society of Japan Award for Young Scientists.
