Metabolomics of ischemic stroke: insights into risk prediction and mechanisms

Abstract

Ischemic stroke (IS) is the most prevalent type of stroke. The early diagnosis and prognosis of IS are crucial for successful therapy and early intervention. Metabolomics, a tool in systems biology based on several innovative technologies, can be used to identify disease biomarkers and unveil underlying pathophysiological processes. Accordingly, in recent years, an increasing number of studies have identified metabolites from cerebral ischemia patients and animal models that could improve the diagnosis of IS and prediction of its outcome. In this paper, metabolomic research is comprehensively reviewed with a focus on describing the metabolic changes and related pathways associated with IS. Most clinical studies use biofluids (e.g., blood or plasma) because their collection is minimally invasive and they are ideal for analyzing changes in metabolites in patients of IS. We review the application of animal models in metabolomic analyses aimed at investigating potential mechanisms of IS and developing novel therapeutic approaches. In addition, this review presents the strengths and limitations of current metabolomic studies on IS, providing a reference for future related studies.

Full text

Vol.:(0123456789)
1 3
Metabolic Brain Disease
https://doi.org/10.1007/s11011-022-01011-7
REVIEW ARTICLE
Metabolomics ofischemic stroke: insights intorisk prediction
andmechanisms
RuijieZhang1,2· JiajiaMeng1,2,3· XiaojieWang4· LiyuanPu1,2· TianZhao1,2· YiHuang5,6,7· LiyuanHan1,2
Received: 7 February 2022 / Accepted: 16 May 2022
© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022
Abstract
Ischemic stroke (IS) is the most prevalent type of stroke. The early diagnosis and prognosis of IS are crucial for suc-
cessful therapy and early intervention. Metabolomics, a tool in systems biology based on several innovative technolo-
gies, can be used to identify disease biomarkers and unveil underlying pathophysiological processes. Accordingly, in
recent years, an increasing number of studies have identified metabolites from cerebral ischemia patients and animal
models that could improve the diagnosis of IS and prediction of its outcome. In this paper, metabolomic research is
comprehensively reviewed with a focus on describing the metabolic changes and related pathways associated with IS.
Most clinical studies use biofluids (e.g., blood or plasma) because their collection is minimally invasive and they are
ideal for analyzing changes in metabolites in patients of IS. We review the application of animal models in metabo-
lomic analyses aimed at investigating potential mechanisms of IS and developing novel therapeutic approaches. In
addition, this review presents the strengths and limitations of current metabolomic studies on IS, providing a refer-
ence for future related studies.
Keywords Metabolomics· Stroke· Ischemic stroke· Molecular diagnosis· Metabolism· Biomarkers
Introduction
Stroke is a major cause of adult disability and the sec-
ond leading cause of mortality worldwide (Collaborators
etal. 2018). Epidemiological studies have reported that
ischemic stroke (IS) accounts for 85% of all stroke cases
(Fatahzadeh and Glick 2006). Research has revealed that
early diagnosis and treatment improves the prognosis of
IS patients, but current imaging-based diagnostic tools
for IS, such as magnetic resonance imaging and non-
contrast computed tomography, are time-consuming,
costly, and limited in their sensitivity and availability
(Laborde etal. 2012; Goyal etal. 2020; Montaner etal.
2020). Additionally, the complex etiology of IS means
that its pathogenesis remains unclear (Latchaw etal.
2009; Makris etal. 2018; Shin etal. 2020). In this regard,
the emerging omics science of metabolomics has great
potential to reveal the underlying pathobiology of IS,
identify novel biomarkers, and guide the development
of therapies.
Ruijie Zhang and Jiajia Meng contributed equally to this paper.
* Yi Huang
huangy102@gmail.com
* Liyuan Han
hanliyuan@ucas.ac.cn
1 Hwa Mei Hospital, University ofChinese Academy
ofSciences, Ningbo315010, Zhejiang, China
2 Ningbo Institute ofLife andHealth Industry, University
ofChinese Academy ofSciences, Ningbo315010, Zhejiang,
China
3 Xihu District Center forDisease Control andPrevention,
Hangzhou310013, Zhejiang, China
4 Department ofNeurology, Shenzhen Qianhai Shekou Free
Trade Zone Hospital, Shenzhen518067, Guangdong, China
5 Department ofNeurosurgery, Ningbo First Hospital,
Ningbo315010, Zhejiang, China
6 Key Laboratory ofPrecision Medicine forAtherosclerotic
Diseases ofZhejiang Province, Ningbo315010, Zhejiang,
China
7 Medical Research Center, Ningbo First Hospital,
Ningbo315010, Zhejiang, China

Metabolic Brain Disease
1 3
Metabolomics andits analytical approach
Metabolomics is a powerful tool in systems biology that
is used to qualify and quantify endogenous small mol-
ecules (≤ 1.5ka) in various types of biological samples,
such as feces, biological fluids, and tissues (Nicholson
etal. 2002; Sidorov etal. 2019; Yang etal. 2020; Li
etal. 2021). Metabolomics is positioned below genomics
(DNA), transcriptomics (RNA), and proteomics (protein)
in the hierarchy of systems biology tools and provides the
most accurate representation of the chemical intermedi-
ates or endpoints in cellular processes (Peng etal. 2015;
Ussher etal. 2016; Qureshi etal. 2017; Dang etal. 2018;
Rinschen etal. 2019). Thus, metabolomics best reflects
the actual phenotype of a cell or organism and can depict
its biological status. It can be used to identify biomarkers
for disease diagnosis and monitoring and to delineate the
molecular mechanisms of pathological processes (Laborde
etal. 2012; Sidorov etal. 2019; Donatti etal. 2020).
Recently, metabolomics has been increasingly applied
in biomedical research to investigate various diseases, such
as IS and other cerebrovascular diseases. It has been used
to analyze various biological specimens, such as serum,
plasma, and urine (Nicholson and Lindon 2008; Sidorov
etal. 2019; Donatti etal. 2020). Most analytical platforms
that have been used in such studies are based on nuclear
magnetic resonance spectroscopy and mass spectroscopy,
which offer comprehensive profiles of the metabolic state
of an organism (Liu etal. 2016b; Au 2018).
An appropriate experimental design is essential for the
comprehensive measurement of all metabolites in metabo-
lomic studies and may be used to address related biological
and medical questions (Fig.1). The two strategies applied in
IS metabolomics are untargeted metabolomics and targeted
metabolomics, and each has different objectives (Dang etal.
2018; Saorin etal. 2020; Zahoor etal. 2021).
Untargeted metabolomics involves the detection of all
metabolites in a given sample and generates an unbiased,
Fig.1 General workflow for metabolomics analyses, which can be
regarded as 3 distinct phases. A An experiment is designed, per-
formed, and specimen collection/preparation for metabolomics
assays. B Metabolite features of samples are profiled and detected by
nuclear magnetic resonance (NMR) or mass spectrometry (MS). C
Statistical analyses are implemented to identify significant metabolite
features, metabolic pathway/cluster analyses, and mechanism/model
formulation. GC indicates gas chromatography; LC, liquid chroma-
tography; PCA, principal components analysis and PLS-DA, Partial
least squares discrimination analysis

Metabolic Brain Disease
1 3
comprehensive dataset of metabolites. It examines a wide
range of metabolite classes by comprehensively analyzing all
metabolite sets and thus often discovers previously unknown
metabolites that must be characterized by comparing their
mass spectral data with those of standards and/or with data
in mass spectral databases (Au 2018; Zahoor etal. 2021).
Untargeted metabolomics is used to screen for differences
between the metabolic profiles of diseased and healthy con-
trol groups, which enables the discovery of novel metabolic
biomarkers and reveals underlying metabolic networks (Au
2018; Zhang etal. 2021). However, this approach is only
semi-quantitative at best and is thus not sufficiently precise
or accurate for diagnostic purposes (Tokarz etal. 2017).
Targeted metabolomics is designed for hypothesis-driven
studies and involves the measurement of a predefined set of
metabolites in samples (Tokarz etal. 2020). This approach is
typically used to validate potential biomarkers that are iden-
tified during the discovery phase or to evaluate the response
of a specific metabolic pathway to a treatment (Wishart
2019; Zahoor etal. 2021). Targeted metabolomics is supe-
rior to untargeted metabolomics in terms of precision, accu-
racy, and calibration ranges and enables the absolute con-
centrations of each metabolite of interest to be determined
(Au 2018; Tokarz etal. 2020; Zahoor etal. 2021). However,
the finite number of known metabolites that are analyzed in
targeted metabolomics means that key compounds or path-
ways may remain undetected (Au 2018).
Metabolomic data analysis
Metabolomic datasets are extremely complex and must
be pretreated to reduce systematic biases during analysis
(Fig.1) (Saorin etal. 2020). Then, appropriate statistical
methods—either univariate or multivariate statistical meth-
ods—must be used to reveal the metabolic profiles of dif-
ferent types of samples. Univariate methods, such as a t-test
or a one-way analysis of variance (ANOVA), are used to
measure one response variable in metabolomic data (Au
2018). Multivariate methods, such as unsupervised princi-
pal component analysis and supervised partial least-squares
discriminant analysis (PLS-DA) are widely used to identify
disease biomarkers in metabolomic data (Zhou etal. 2012).
Other multivariate methods, such as multivariate analy-
sis of variance, ANOVA-simultaneous component analy-
sis, orthogonal partial least-squares (OPLS) analysis, and
OPLS-DA, are also used to analyze metabolomic data (Au
2018; Saorin etal. 2020). Metabolomic data have also been
analyzed using machine learning techniques, such as soft
independent modeling of class analogy, hierarchical cluster
analysis, self-organizing maps, support-vector machines, and
random forests (Au 2018; Azad and Shulaev 2019; Tiedt
etal. 2020).
Metabolites can influence various biological systems and
thus affect genetic, epigenetic, and physiological responses
to disease processes and environmental stimuli (Nichol-
son and Lindon 2008; Qureshi etal. 2017; Montaner etal.
2020). Consequently, the metabolic profile of an organism
may directly reflect its physiological and pathological states.
Metabolomics therefore has great potential to facilitate the
integrated mapping of specific processes underlying physi-
ological and pathological states (Johnson etal. 2016). In
the context of IS, the primary goal of metabolomics is
to identify circulating metabolites that provide systems-
level information on the molecular pathways that underlie
IS development and progression, thereby revealing novel
diagnostic and prognostic biomarkers (Qureshi etal. 2017;
Montaner etal. 2020). Metabolite set enrichment analysis
and metabolic network analysis can also be used to provide
key insights into how metabolic changes mediate the biol-
ogy of disease processes or pathological phenomena of IS
(Liu etal. 2016a). Commonly used pathway databases are
the Kyoto Encyclopedia of Genes and Genomes, MetaCyc,
the Small Molecule Pathway Database, and MetaboLights.
These databases are used to meet the increasing demand
for biologically meaningful and informative correlations
between molecular metabolites and a physiological process
or pathological phenotype (Cambiaghi etal. 2017; Hernan-
dez-de-Diego etal. 2018; Li etal. 2021). Additionally, vari-
ous bioinformatics tools (such as MetaboAnalyst, 3Omics,
PaintOmics, integrOmics, and MetScape) are being devel-
oped for determining the relationships between metabolites
and their biological functions (Cambiaghi etal. 2017; Her-
nandez-de-Diego etal. 2018; Li etal. 2021).
Overview ofmetabolomic studies ofIS
Several types of study designs can be used in metabolomics
studies to diagnose and prognose IS. Diagnostic studies have
event-based designs and are used to determine the differ-
ences between control and diseased subjects for differential
disease diagnoses or to investigate the efficacy of treatments
(Fig.2A). For example, in one diagnostic study, a patient’s
post-IS metabolite metrics were measured and compared
with those of a healthy individual, and the differences were
interpreted metabolically (Everett etal. 2019).
Recently, a case–control study of 969 participants was
conducted to identify metabolites that are significantly
associated with IS. It revealed significant differences in the
serum concentrations of 41 endogenous metabolites between
the IS group and the control group (Table1); the AUC for 30
of these metabolites was 0.961, thus confirming significant
differences between the IS patients and the control group.
The study also discovered that the serum concentrations
of pregnenolone sulfate, adenosine, and asymmetric and

Metabolic Brain Disease
1 3
symmetric dimethylarginine effectively discriminated IS
patients from patients with stroke mimics (Tiedt etal. 2020).
Another study analyzed serum samples from 38 IS patients
and 46 control subjects and found that the IS patients had
higher serum concentrations of arginine, vaccenyl carnitine
(C18:1), palmitoylcarnitine (C16), and 3-hydroxylbutyryl-
carnitine (C4OH) and a lower citrulline/arginine ratio than
the controls (Sun etal. 2019a).
In another metabolomic study, the serum samples of 40
IS patients were found to have abnormal concentrations of
amino acids and their metabolites compared with 29 control
subjects. Specifically, IS patients had lower serum concen-
trations of alanine, glycine, isoleucine, leucine, serine, tyros-
ine, methionine, tryptophan, urea, purine, hypoxanthine, and
proline, reflecting disrupted amino acid metabolism (Wang
etal. 2017). In contrast, Goulart etal. (2019) found that
serum concentrations of proline, leucine, and glycine were
significantly higher in IS patients than in the controls (Gou-
lart etal. 2019). These studies have also detailed how serum
concentrations of amino acids differed between acute and
chronic IS phases. Finally, dyslipidemia increases the risk
of atherosclerosis, which in turn increases the risk of IS,
and studies have found that serum samples of IS patients
contained lower concentrations of lysophosphatidylcholines
(lysoPCs) and phosphatidylcholines (PCs) and higher con-
centrations of acylcarnitines than those of healthy subjects,
indicating that IS disrupts phospholipid metabolism and
fatty acid oxidation (Liu etal. 2017; Sun etal. 2017).
Prognostic studies have outcome-based designs and
aim to predict IS outcomes from baseline or longitudinal
metabolic analyses, e.g., the occurrence of future IS or the
prognosis of existing IS in groups of subjects (Fig.2B). A
related type of baseline study involves the identification of
prognostically relevant metabolites based on the metabo-
lomic profiling of biofluids obtained from epidemiological
or biobank studies of IS patients and controls. Longitudinal
studies involve the measurement of metabolic responses in
the biofluid samples of a cohort of IS patients to determine
Fig. 2 Schematic Representations of the Three Key Experimental
Designs in IS Metabonomics research. Subjects are on the left-hand
side followed by the procedures and targets of metabonomics analysis
on the right-hand side. A Case–control design for diagnostic meta-
bonomics approach, differentiating healthy individuals (blue) from
subjects with a disease (red) at a given time point. B Nested case–
control design for predictive metabonomics approach. In this case,
there is a difference (although not complete discrimination) in meta-
bolic profiles before IS between two subgroups(Cambridge blue and
dark blue). C Case-time-control design for diagnostic and prognos-
tic metabonomics approach, differentiating patients with IS in accute
stage (red) from subjects in chronic stage (blue) at a given time point

Metabolic Brain Disease
1 3
Table 1 Metabolomics studies in patients with Ischemic stroke
Study Published
date
Study-
design
Group Validation Average
age
Targeted/
Untargeted
Study
objective
Specimens Analysis
platform
Suggested biomarker
Steffen Tiedt etal.
(Tiedt etal.
2020)
2020 Case–con-
trol
Stage1(IS = 74,NC
= 72),Stage3(SM
= 33, IS = 40);
Stage2(IS = 40,
NC = 40),Stage4(SM = 105,
IS = 105),Stage5(SM = 211,
IS = 289)
70.0 Untargeted Diagnostic
bio-
marker
Serum UPLC-
MS/MS
Asymmetrical and sym-
metrical dimethylar-
ginine, pregnenolone
sulfate, adenosine
Evgeny Sidorov
etal.(Sidorov
etal. 2020a)
2020 Case-time-
control
AS = 20, CS = 20 None 57.0 Targeted Biomarkers
for acute
ischemic
stroke
Serum and
urine
LC–MS Asparagine, tyrosine,
xylose,glycine, acetyl-
carnitine
Laurent Suissa
etal.(Suissa
etal. 2020)
2020 Nested
case–
control
FO = 18,
UFO = 23
None 74.8 Untargeted Predicting
biomark-
ers of
outcome
Cerebral
thrombi
LC–MS Sorbitol
Evgeny Sidorov
etal. (Sidorov
etal. 2020b)
2020 Case-time-
control
AS = 60, CS = 60 None 60.0 Untargeted Discovery
of serum
bio-
markers
related
to infarct
volume
Serum UPLC-
MS/MS
4 unknown metabolites
Nai-Fang Chi etal.
(Chi etal. 2021)
2020 Nested
case–
control
FO = 77, UFO = 73 None 67.4 Untargeted Biomarkers
of func-
tional
recovery
of stroke
Serum UPLC-
MS/MS
147 metabolites
Daokun Sun
etal. (Sun etal.
2019a)
2019 Nested
case–
control
IS = 346,NC = 3558 IS = 114,NC = 112 53.5 Untargeted Predicting
inci-
dent of
ischemic
stroke
Serum GC–MS tetradecanedioate, hexa-
decanedioate
Vânia A. M. Gou-
lart etal.(Goulart
etal. 2019)
2019 Case–con-
trol
Ath = 13, Car = 7,
Lac = 10, Und = 8,
NC = 16
None 61.6 Targeted Stroke
etiology
Plasma GC–MS Proline, lysine, phenyla-
lanine, leucine, glycine,
methionine, alanine
Ruitan Sun etal.
(Sun etal.
2019b)
2019 Case–con-
trol
IS = 38,NC = 46 None 66.6 Targeted Diagnostic
bio-
marker
Serum direct-
infusion
mass
spec-
trometry
Vaccenylcarnitine
(C18:1), palmitoylcar-
nitine (C16),
3-hydroxylbutyrylcarni-
tine (C4OH), arginine,
arginine /ornithine,
citrulline /arginine

Metabolic Brain Disease
1 3
Table 1 (continued)
Study Published
date
Study-
design
Group Validation Average
age
Targeted/
Untargeted
Study
objective
Specimens Analysis
platform
Suggested biomarker
Xiaofan Guo etal.
(Guo etal. 2019)
2019 Nest case–
control
IS = 66,NC = 66 None 60.8 Untargeted Predicting
inci-
dent of
ischemic
stroke
Plasma UPLC-
MS/MS
9-cis-Retinal, DL-
Indole-3-Lactic Acid,
(2S)-OMPT, 2-methyl-
1-pyrroline, d-pipeco-
linic Acid, 7,7-dime-
thyl-5,8-eicosadienoic
acid, PC(15:0/0:0)
[U], DL-dihydrosphin-
gosine, bufexamic
acid, 10- L-thyroxine,
L-thyroxine
Michael V.
Holmes(Holmes
etal. 2018)
2018 Nest case–
control
MI = 912,IS = 1146
,ICH = 1138,NC
= 1466
None 47.0 Targeted Investigat-
ing the
associa-
tions of
metabolic
makers
with
different
CVD
subtypes
Plasma H.1-NMR Glycoprotein-acetyls,
β-hydroxybutyrate,
glucose, acetoacetate,
docosahexaenoic acid
Yeseung Lee etal.
(Lee etal. 2017)
2017 Nest case–
control
IS = 62,NC = 348 IS = 99,NC = 301 62.1 Targeted Predicting
incident
ischemic
stroke
Serum LC–MS N6-acetyl-L-lysine,
cadaverine,
nicotinamide,2-
oxoglutarate, L-valine,
S-(2-methylpropionyl)-
dihydrolipoamide-
E,ubiquinone,5-
aminopentanoate,
homocysteine, sulfinic
acid, lysine
Peifang Liu etal.
(Liu etal. 2017)
2017 Case–con-
trol
IS = 40,NC = 40 IS = 26,NC = 23 58.1 Untargeted Diagnostic
bio-
marker
Serum UPLC-
MS/
MS,GC–
MS
serine, isoleucine,
betaine, PC(5:0/5:0),
LysoPE(18:2)
Dian Wang etal.
(Wang etal.
2017)
2017 Case–con-
trol
IS = 40,NC = 29 None 63.7 Targeted Diagnostic
bio-
marker
Serum GC–MS Tyrosine, lactate, tryp-
tophan
Hongxue Sun etal.
(Sun etal. 2017)
2017 Case–con-
trol
IS = 30,NC = 30 None 57.6 Targeted Diagnostic
bio-
marker
Serum LC–MS Uric acid, sphinganine,
adrenoyl ethanolamide

Metabolic Brain Disease
1 3
their metabolic trajectories over time (Fig.2C) (Everett
etal. 2019). These prognostic (or predictive) studies aim to
discover prognostic indicators in cohorts of IS patients and
better interpret details of IS progression.
Most of the baseline or longitudinal studies of IS have
examined the metabolic profiles of blood samples to deter-
mine IS incidence or predict its recurrence. Several prospec-
tive studies have reported significantly altered blood concen-
trations of some organic acids and lipids, such as bufexamac
acid, D-pipecolinic acid, dodecanoic acid, and lysoPCs. For
example, one of the first untargeted metabolomic studies
metabolically profiled plasma samples from 293 patients
with recurrent IS. Plasma concentrations of 1-monopal-
mitin, dodecanoic acid, meso-erythritol, threonate, and
lysoPCs were decreased in these patients. Sun etal. (2019a)
quantified 245 metabolites in the serum samples of 3,904
participants (IS = 346, Control = 3558) and compared these
samples to validated samples from 114 IS patients and 112
healthy controls to investigate the association between the
245 metabolites and IS incidence. They found that tetrade-
canedioate and hexadecanedioate were associated with IS
incidence. Holmes etal. (2018) quantified 225 metabolites
in baseline plasma samples from 912 myocardial infarction
(MI) patients, 1,146 IS patients, and 1,466 healthy control
subjects to investigate the association of lipids and related
metabolites with the risks of IS and MI. Their findings
demonstrated that low-density lipoprotein cholesterol and
triglycerides were positively associated with IS (Holmes
etal. 2018). Glutamate functions as an indicator of neu-
ral excitation and the presence of neurotoxins in the brain
and contributes to neuronal injury (Castellanos etal. 2008;
Meng etal. 2015). Serum concentrations of 2-oxoglutarate,
which is a derivative of glutamate, were found to be lower
in patients at risk of thrombotic stroke (a type of IS), which
suggests that glutamate can exhibit increased excitotoxic
activity (Papes etal. 2001; Kimberly etal. 2013; Lee etal.
2017). This finding was supported by another study by Wang
etal. (Wang etal. 2017). A further two longitudinal studies
performed metabolomic profiling of serum samples obtained
from patients over time, which revealed that higher serum
concentrations of asparagine, tyrosine, and xylose were sig-
nificantly associated with acute IS (Sidorov etal. 2020a) and
that four unknown metabolites were associated with infarc-
tion volume in acute IS (Sidorov etal. 2020b).
The studies described above only utilized surrogate bio-
fluids (i.e., serum or plasma) to depict a global overview of
metabolic changes associated with IS. The collection of bio-
fluids is minimally invasive and is feasible for biomonitoring
large samples. However, biofluids are reflective of many bio-
chemical responses across various tissues within the body,
not just the brain, which is the target organ of stroke. Studies
have shown that only a few altered metabolites, caused by
only certain diseases, can be found in both the brain and
Table 1 (continued)
Study Published
date
Study-
design
Group Validation Average
age
Targeted/
Untargeted
Study
objective
Specimens Analysis
platform
Suggested biomarker
P.A. Vorkas etal.
(Vorkas etal.
2016)
2016 Case–con-
trol
SYM = 5,
ASYM = 5
None None Untargeted Diagnostic
bio-
marker
Carotid
Plaques
UPLC-MS butyrylcarnitine, hex-
anoylcarnitine, palmi-
toylcarnitine, TG(58:6),
PC(16:0/20:4),
PC(16:0/18:1),
PE(18:1/18:0), arachi-
donic acid (AA)
Mariona Jové etal.
(Jove etal. 2015)
2015 Nest case–
control
SR = 20,NC = 111 IS = 15,NC = 147 71.7 Untargeted Predicts
stroke
recur-
rence
Plasma LC–MS 1-monopalmitin, dode-
canoic acid, meso-
erythritol, threonate,
LysoPC[16:0], myris-
toyl-ethanolamine,
LysoPC(20:4)
IS, ischemic stroke; NC, normal control; SM, stroke mimics; Ath,atherothrombotic stroke; Car, cardioembolic stroke; Lac,lacunar stroke; Und, undetermined stroke; MI, myocardial infarction;
ICH, intracerebral hemorrhage; AS, acute stage of ischemic stroke; CS, chronic stage of ischemic stroke; SR, stroke recurrence; FO, favorable outcome; UFO, unfavorable outcome; SYM, cer-
ebrovascular symptoms of carotid origin; ASYM, cerebrovascular asymptoms of carotid origin

Metabolic Brain Disease
1 3
the blood. This is disadvantageous for pathological inter-
pretations of disease, as pathway analyses depend on these
altered metabolites. The few studies of stroke that are based
on metabolomic techniques have used brain tissue as a biosa-
mple. The molecular mechanism through which metabolic
dysregulation in brain tissue leads to stroke remains unclear.
Overview ofmetabolomic studies onmiddle
cerebral artery occlusion (MCAO) models
ofIS
Given the inherent ethical and safety concerns regarding the
use of tissue samples from human IS patients for studying
thromboembolic occlusions, various animal models of stroke
have been developed. These models are indispensable for the
following reasons: (1) Unlike human IS, which has diverse
manifestations, experimental IS is a highly reproducible,
controllable, and standardized system, which allows patho-
physiological processes and the effects of potential therapeu-
tics to be analyzed accurately. (2) The affected brain tissue
of animals can be subjected to molecular, biochemical, and
physiological analyses. (3) The pathophysiological changes
associated with IS are detectable by imaging techniques
within a minute of IS occurrence in animal models; such
an assessment is not practical in humans. (4) Perfusion and
vasculature can be modeled in animals, but not invitro (Fluri
etal. 2015; Shin etal. 2020).
The MCAO model is most widely used to investigate the
mechanisms underlying IS and to develop novel therapies.
However, neither the MCAO model nor any other animal
models of IS perfectly replicate the pathological features of
the human disease (Sommer 2017; Kaiser and West 2020).
For example, the MCAO model has several limitations, such
as hyperthermia or hypothermia problems, uncontrollable
infarct location and size, a high risk of vessel rupture with
certain suture types, and unsuitability for neuroprotective
studies (Fluri etal. 2015). In addition, there are macro-
scopic and microscopic differences between humans and the
MCAO model, such that the model fails to fully simulate
human IS; these include differences in brain anatomy, func-
tional organization, genetics and epigenetics, and treatment
reactions (Sommer 2017). Furthermore, subtypes of stroke
in clinical settings can be atherothrombotic or cardioembolic
in origin, while stroke in an animal model is mainly pro-
voked by arterial occlusion. Thus, an animal model cannot
simulate all of the metabolic changes seen in human stroke
subtypes. The results of an animal study therefore cannot be
directly compared to the results of a clinical study. Neverthe-
less, the MCAO model is extremely useful for providing new
perspectives on the diagnosis of IS and for the evaluation of
clinical disease activity.
IS-related metabolites, such as amino acids, lipids, poly-
amines, and nucleotides, have been explored in animal mod-
els to understand disease mechanisms in cerebral ischemia
(Table2). Metabolites have been identified in various types
of biological samples, such as rat plasma, brain tissue,
liver tissue, and cerebrospinal fluid. A liquid chromatog-
raphy–mass spectrometry (LC–MS) analysis by Guo etal.
(2020) found that the serum concentrations of 37 metabo-
lites in an MCAO group were significantly different from
those in a control group (Guo etal. 2020). These metabo-
lites included sphingomyelin, lysoPC, stearidonic acid, PCs,
organic acids, amino acids, and carnitine derivatives, and
pathway enrichment analyses suggested that key metabo-
lites were involved in the biosynthesis of unsaturated fatty
acids, ɑ-linolenic acid metabolism, fatty acid metabolism,
the synthesis and degradation of ketone bodies, and pentose
and glucuronate interconversions (Guo etal. 2020). Simi-
larly, Wu and Liu used LC–MS to reveal that there were
decreased concentrations of PCs, phosphatidylethanola-
mines (PEs), lysoPEs, and sphingosine-1-phosphate in the
blood samples of a rat MCAO model, reflecting perturba-
tions in lipid metabolism pathways (Liu etal. 2016a; Wu
etal. 2020). Paik investigated lipid metabolism in the blood
of rats subjected to MCAO surgery and found that the sur-
gery caused changes in the concentrations of 19 free fatty
acid metabolites (Paik etal. 2009). Additionally, the blood
concentrations of lactate, glucose, glucose 6-phosphate, suc-
cinic acid, malic acid, and citric acid were increased in these
animals compared with control rats; as these metabolites are
associated with energy metabolism, this indicated that the
glucose and anaerobic glycolysis metabolism pathways were
perturbed (Wang etal. 2019). In contrast, Wang etal. (2014)
found that an MCAO rat model had lower blood concentra-
tions of lactate (L-lactic acid) than the control rats, as well as
significantly lower blood concentrations of leucine, isoleu-
cine, and valine, which is consistent with human cohort stud-
ies and indicates perturbed amino acid metabolism (Wang
etal. 2014). We observed changes in the concentrations of
the same metabolites (e.g., lactate, creatine, glycine, alanine,
leucine, and lysine) in the brain tissue, cerebrospinal fluid,
and plasma of MCAO rats, but the changes varied across
these matrices (Kimberly etal. 2013; Wang etal. 2013; Luo
etal. 2019; Wesley etal. 2019).
We collected the identified metabolites published in clini-
cal metabolomics studies and animal studies, with a com-
parison of their consistency in blood between the two types
of studies. A total of 190 metabolites identified in Murine
model studies share 15 ones in clinical studies. We exam-
ined whether their tendency of 15 metabolites is the same
between clinical studies and animal studies. If it showed
the same change in or more than two of three studies, we
determined it in the same trend. Of 15 metabolites, three
have the same tendency in both types of studies, including

Metabolic Brain Disease
1 3
Table 2 Metabolic alterations associated with MCAo in rats model
Study Group Specimens or
subjects
Analysis platform Elevated metabo-
lites
Reduced metabolites Metabolites type pathway
Bingfeng Lin etal.
(Lin etal. 2021)
NC = 15,
MCAo = 15,AEGE = 15,Nim = 15
Plasma LC–MS Tyramine, thy-
mine, lysine,
L-carnitine,
L-Histidine,
L-acetylcarnitine,
deoxycytidine,
lauroyl diethanol-
amide, glutamyl-
phenylalanine,
5′-methylthioad-
enosine, phyto-
sphingosine,
corticosterone,
11b,17a,21-tri-
hydroxypreg-
nenolone, netilm-
icin, LysoPC
[P-18:1 (9z)],
LysoPC(P-18:0)
3-Indolehydracrylic
acid, LysoPC [20:3
(5Z,8Z,11Z)/0:0],
LysoPC [20:1 (11Z)],
LysoPC(20:0/0:0),
LysoPE (0:0/24:0),
LysoPC [22:1 (13Z)/0:0],
arachidonic acid,
5a-tetrahydrocortisol,
lansiumarin C, Scuti-
geral, Mammeigin,
LysoPE (18:0/0:0)
amino acids,
lipids,other
metabolites
phenylalanine, tyros-
ine, and trypto-
phan biosynthesis;
phenylalanine
metabolism; histi-
dine metabolism;
sphingolipid
metabolism; pyrimi-
dine metabolism;
cysteine and
methionine
metabolism; tyrosine
metabolism;
steroid-hormone
biosynthesis
Wenjun Guo etal.
(Guo etal. 2020)
NC = 10, MCAo = 10 Serum LC–MS Acetoacetic acid,
l-acetylcarnitine,
deoxyuridine,
2-hydroxybutyric
acid, leucylpro-
line, hippuric
acid, decenedioic
acid, 3-oxodo-
decanoic acid,
l-palmitoylcar-
nitine, 3-oxotet-
radecanoic acid,
eicosapentaenoic
acid, palmitic
amide, hexa-
decadienoic acid,
alpha-linolenic
acid, palmitoleic
acid, docosap-
entaenoic acid
(22n-3), dihomo-
alpha-linolenic
acid, oleic acid,
eicosadienoic
acid, PC (36:4),
SM (d18:0/18:1)
Proline, molybdate,
trichloroethanol glucu-
ronide, 2-phenylethanol
glucuronide, glycocholic
acid, chenodeoxyglyco-
cholic acid, 2-hydrox-
yhexadecanoylcarnitine,
hydroxyhexadecanoic
acid, goshuyic acid,
LysoPC (20:3), LysoPC
(20:2), myristoleic acid,
LysoPC (20:1), steari-
donic acid, palmitic acid,
SM (d18:1/20:0)
Lipids,organic
acids, amino
acids,other
metabolites
Unsaturated fatty
acids, alpha-
linolenic acid
metabolism, fatty
acid metabolism,
synthesis and deg-
radation of ketone
bodies, pentouse
and glucuronate
interconversions

Metabolic Brain Disease
1 3
Table 2 (continued)
Study Group Specimens or
subjects
Analysis platform Elevated metabo-
lites
Reduced metabolites Metabolites type pathway
Lanlan Wu etal.
(Wu etal. 2020)
NC = 11, MCAo = 11 Serum LC–MS L-lactic acid,
isobutyrylgly-
cine, sebacic
acid, linoleic
acid, gluco-
sylgalactosyl
hydroxylysine,
docosahexaenoic
acid, oxoglutaric
acid
Hydroxyphenylacetyl-
glycine, N-methyl-
nicotinium, N2-suc-
cinyl-L-ornithine,
tauroursodeoxycholic
acid, sphingosine
1-phosphate,sphinganine
1-phosphate,2-hydrox-
yestradiol-3-methyl ether,
glutamylphenylalanine,
vitamin D2 3-glucuron-
ide, neoxanthin, stearoyl-
carnitine, petroselinic
acid
Lipids, organic
acids, amino
acids, other
metabolites
Di-unsaturated fatty
acid β-oxidation;
glycine, serine,
alanine, threo-
nine metabolism;
glycolysis and
gluconeogenesis;
glycosphingolipid
metabolism;
linoleate metabo-
lism; TCA cycle;
urea cycle, metab-
olism of arginine,
proline, glutamate,
aspartate, and
asparagine
Umadevi V.
Wesleya etal.
(Wesley etal.
2019) ξ
NC = 3,MCAo2d = 3, MCAo3d = 3,
MCAo7d = 3, MCAo14d = 3
Plasma 1H-NMR Glycine, isoleu-
cine, o-acetylcar-
nitine,, lysine
Methionine,Choline, Cit-
rate, Proline, Threonine,
Tyrosine
Amino acids,
lipids, other
metabolites
None
Umadevi V.
Wesleya etal.
(Wesley etal.
2019).ψ
NC = 3,MCAo2d = 3, MCAo3d = 3,
MCAo7d = 3, MCAo14d = 3
Liver 1H-NMR Choline,
3-hydroxybu-
tyrate, Creatine,
Taurine
Tyrosine, glutathione, tryp-
tophan, serine, alanine,
histamine, PC
Amino acids,
lipids,other
metabolites
None
Umadevi V.
Wesleya etal.
(Wesley etal.
2019)ξ
NC = 3, MCAo2d = 3, MCAo3d = 3,
MCAo7d = 3, MCAo14d = 3
Brain tissue 1H-NMR 3-hydroxybutyrate,
creatine, taurine,
glycine, methio-
nine, isoleucine,
ethanolamine,
fumarate, glyc-
erol, pantothen-
ate, tryptophan
Glutathione, tryptophan,
serine, alanine, hista-
mine, N-acetylaspartate
Amino acids,
lipids,other
metabolites
None
Lan Luo etal. (Luo
etal. 2019)
NC = 16, MCAo = 16 Brain tissue 1H-NMR Lactate, acetate
(Ace), myo-
inositol(m-Ins),
γ-aminobutyric
acid (GABA),
aspartate (Asp),
alanine (Ala),
leucine (Leu),
isoleucine (Ile),
choline (Cho)
Creatine/phosphocreatine
(Cr/PCr), N-acetyl-aspar-
tate (NAA), N-acetylas-
partylglutamate (NAAG)
Organic acids,
amino acids,
other metabolites
None

Metabolic Brain Disease
1 3
Table 2 (continued)
Study Group Specimens or
subjects
Analysis platform Elevated metabo-
lites
Reduced metabolites Metabolites type pathway
Lan Luo etal. (Luo
etal. 2019)
NC = 16, MCAo = 16 Urine 1H-NMR creatinine (Crn),
acetoacetate,
acetone (Aco),
N-acetyl-glyco-
protein (NAG),
dimethyl-
amine (DMA),
3-hydroxybu-
tyrate
_ Organic acids,
polyamines,
other metabolites
None
Lan Luo etal.(Luo
etal. 2019)
NC = 16, MCAo = 16 Plasma 1H-NMR Lactate, trimethyl-
amine-N-oxide/
betaine
VLDL/LDL, 3-hydroxybu-
tyrate (3-HB), ace-
toacetate (AcAc), poly
unsaturated fatty acid
(PUFA)
Organic acids,
lipids, other
metabolites
None

Metabolic Brain Disease
1 3
Table 2 (continued)
Study Group Specimens or
subjects
Analysis platform Elevated metabo-
lites
Reduced metabolites Metabolites type pathway
Yang Wang etal.
(Wang etal.
2019)
NC = 10, MCAo = 10 Serum GC–MS 4-Hydroxy-proline,
allantoic acid,
beta-alanine, cre-
atine, creatinine,
homocysteine,
isoleucine, ser-
ine, methionine
sulfoxide, orni-
thine, Arabitol,
glucose, ribose,
xylose, glucose
6-phosphate,
ribonolactone,
caproic acid,
docosahex-
aenoic acid,
dodecanoic acid,
heptadecanoic
acid, linoleic
acid, myristic
acid, myristoleic
acid, oleic acid,
Palmitic acid,
palmitoleic acid,
pentadecanoic
acid, stearic
acid, cholesterol,
o-phosphoe-
thanolamine,
pregnenolone,
guanosine,
inosine, pipecolic
acid, malic acid,
Petroselinic acid,
Picolinic acid,
succinic acid,
Dopamine, pan-
tothenic acid
Arachidonic acid Lipids, organic
acids, amino
acids, carbo-
hydrates, other
metabolites
Linoleic acid metab-
olism, homocyst-
eine degradation,
fatty acid biosyn-
thesis, amino acid
metabolism

Metabolic Brain Disease
1 3
Table 2 (continued)
Study Group Specimens or
subjects
Analysis platform Elevated metabo-
lites
Reduced metabolites Metabolites type pathway
Mengting Liu etal.
(Liu etal. 2016b)
NC = 10, MCAo = 10 Plasma UPLC/TOF–MS Cholic acid,
pseudouridine/
uridine, pyruvate,
taurine, lactate,
tryptophan,
glutamine, his-
tidine, cytidine,
creatine, alanine,
phenylalanine,
5-hydroxyin-
doleaceticacid,
norepinephrine,
palmitoyl-
L-carnitine,
N-stearoylserine
Acetone, PC(14:0/0:0),
PE(17:0/0:0),
PS(20:0/0:0), PC(14:1),
LysoPE(0:0/16:0),
LysoPE(20:1),
LysoPE(0:0/24:1(15Z)),
LysoPE(20:0/0:0),
LysoPE(24:0)
Nucleotide, lipids,
organic acids,
other metabolites
Monoamine
neurotransmitter
metabolism,amino
acid
metabolism,energy
metabolism, lipid
metabolism
Yun Wang etal.
(Wang etal.
2014)*
NC = 10, MCAO0 = 10,
MCAO0.5 = 10, MCAO1 = 10,
MCAO3 = 10, MCAO6 = 10,
MCAO12 = 10, MCAO24 = 10
Serum 1H-NMR Malonic acid,
citric acid, gly-
cine, ornithine,
pyruvic acid,
glutamic acid,
succinic acid
L-lactic acid, betaine,
L-serine, acetic acid,
L-valine, L-alanine
Organic acid,
amino acids
TCA
Yun Wang etal.
(Wang etal.
2013)*
NC = 10, MCAO0 = 10,
MCAO0.5 = 10, MCAO1 = 10,
MCAO3 = 10, MCAO6 = 10
Cerebrospinal fluid 1H-NMR L-lactic acid,
L-Alanine,
glutamic acid
None Lipids, organic
acids, amino
acids
None
W. Taylor
Kimberly etal.
(Kimberly etal.
2013)
NC = 7,SMCAo = 6,LMCAo = 7 Cerebrospinal fluid LC–MS Xanthosine, lysine Leucine, isoleucine, valine Amino acids, other
metabolites
None
W. Taylor
Kimberly etal.
(Kimberly etal.
2013)
NC = 7,SMCAo = 6,LMCAo = 7 Plasma LC–MS Xanthosine, carno-
sine, glutamate
Leucine, isoleucine, valine,
niacinamide, phenyla-
lanine
Amino acids, other
metabolites
None

Metabolic Brain Disease
1 3
Table 2 (continued)
Study Group Specimens or
subjects
Analysis platform Elevated metabo-
lites
Reduced metabolites Metabolites type pathway
Man Jeong Paik
etal.(Paik etal.
2009)
NC = 8, MCAo = 6 Plasma GC–MS Caproic
acid(C6:0),
caprylic
acid(C8:0), dece-
noic acid(C10:1),
capric
acid(C10:0), lau-
ric acid(C12:0),
myristoleic
acid(C14:1),
myristic
acid(C14:0),
γ-linolenic
acid(C18:3n6),
eicosapentaenoic
acid(C20:5n3),
docosahexaenoic
acid (C22:6n3),
docsapentaenoic
acid (C22:5n3),
erucic acid
(C22:1), nervonic
acid (C24:1),
lignoceric acid
(C24:0), hexa-
cosanoic acid
(C26:0)
Palmitic acid(C16:0), lin-
oleic acid(C18:2n6), ara-
chidonic acid(C20:4n6),
eicosenoic acids(C20:1)
Lipids None
ξ To investigate metabolite changes in cerebral ischemia/reperfusion (I/R) injury, 3-month rats were euthanized at day 2, 3, 7 and 14 of I/R. ψ For study of ischemia/reperfusion injury related to
time, 12-month rats were grouped to reperfusion time of day 2, 3, 7 and 14. *To investigate metabolite changes in cerebral ischemia/reperfusion injury, 0, 0.5, 1, 3, 6, 12, 24h of I/R

Metabolic Brain Disease
1 3
lower levels of proline, tyrosine, and valine. It suggests that
cerebral artery occlusion will cause turbulent amino acid
metabolism, both in the rat of MCAO model and patients of
IS. On the contrary, the four inconsistent ones were isoleu-
cine, glycine, lysine, and alanine, with higher concentrations
in the rat of MCAO model rat while lower levels in patients
of IS.
Benets andlimits ofthemetabolomic
analyses ofISinthe current studies
Metabolomic studies of stroke focus on changes to metabo-
lites in blood samples because the collection of these bio-
markers is easy and minimally invasive, making for a speedy
and convenient choice for clinical diagnoses. Inflammatory
mediators released in the early stage of IS occurrence dam-
age the blood–brain barrier, resulting in the leakage of pro-
teins and metabolites from the brain to peripheral blood,
where they induce changes in metabolites. This makes
metabolomic studies ideal for early diagnosis and predic-
tion of functional outcomes. However, the majority of serum
or plasma biomarkers may have low specificity due to their
lack of organ-specificity. Furthermore, expression of specific
biomarkers could be the result of altered conditions in spe-
cific cellular context of an organ. This issue can be remedied
by the application of a combination of multiple biomark-
ers to improve diagnosis and prognosis efficiency. When
blood samples are used for metabolic pathway analysis, it
is difficult to achieve a comprehensive and direct metabolic
pathway, compared with the samples of the central nervous
system. In addition, for many metabolites, the plasma lev-
els are highly regulated, and a local change could lead to a
non-detectable change in the blood concentration. However,
in some cases, blood concentrations are poorly controlled.
Other types of samples (e.g., urine) should be analyzed to
obtain a more comprehensive understanding of metabolic
pathways for IS patients.
In clinical studies, most subjects are patients within the
first 24h of stroke onset. Levels of metabolites change in
the serum or plasma of patients because of brain injury. The
altered metabolites are not only identified as biomarkers but
also investigated through metabolomic analyses of the serum
or plasma to reflect the real-time pathological condition result-
ing from the trauma of stroke. The identified pathological con-
ditions are related to clinical signatures and help to understand
the molecular mechanisms underlying clinical features. Several
days after stroke onset, biomarkers are collected and analyzed
to predict outcomes. These biomarkers reflect the biological
processes that occur when the patient is recovering from an
IS; different biochemical processes are related to favorable
or unfavorable patient outcomes. Such analyses further our
understanding of the molecular mechanisms underlying IS.
Cerebral ischemia/reperfusion leads to a series of biological
reactions such as oxidative stress, inflammation, mitochon-
drial dysfunction, reduction of antioxidative agents, and ATP
depletion(Wu etal. 2018; Liao etal. 2020). In clinical settings,
it is not feasible to confirm cell death in the ischemic region or
detect metabolic alteration in brain tissue. The animal model
is thus a good substitute for exploring the biological processes
of ischemia/reperfusion. Studies using animal models have
shown that some metabolite levels change differently, depend-
ing on whether they are in blood or brain tissue after ischemia/
reperfusion. One example is methionine, which is downregu-
lated in plasma but upregulated in brain tissue (Wesley etal.
2019). This demonstrates that changes in metabolites in the
ischemic/reperfused region of the brain do not necessarily lead
to corresponding changes in the blood. Important information
would be lost if we performed analyses of dysregulated path-
ways according to disordered metabolites in the blood alone.
This suggests that the use of blood analysis for the interpreta-
tion of biological mechanisms occurring in brain tissue during
IS and after reperfusion is limited.
Given the ethical and safety concerns involved with har-
vesting tissue samples from human IS patients, the use of
tissue samples in metabolomic studies of IS is rare. To cir-
cumvent this issue, scientists use MCAO models to imitate
human stroke. In animal studies, metabolomic techniques
are used to explore the potential biomarkers for early diag-
nosis of ischemic stroke and novel therapeutic targets. For
this purpose, blood samples are collected before and during
cerebral artery occlusion and after reperfusion to record the
integral metabolic trajectory. In contrast, blood samples of
stroke patients are generally collected several hours or days
after stroke onset; thus, obtaining metabolomic data upon
stroke onset in humans is difficult. However, published study
have demonstrated that acylcarnitines and adipokines are
associated with the destabilization and rupture of plaque
(Vorkas etal. 2016). Low levels of neurosteroids were
related to poor outcome in many brain pathologies (Adib-
hatla and Hatcher 2008). Furthermore, sorbitol and glucose
levels in the thrombi are correlated with functional outcomes
(Suissa etal. 2020). Through metabolomic analyses, we can
speculate how the metabolites change and obtain informa-
tion on the active pathways at stroke onset, which can pro-
mote improved etiological diagnosis of stroke and patient
outcomes (Jia etal. 2021; Chumachenko etal. 2022).
Conclusion
IS is a complex and multifactorial disease that accounts
for many deaths and cases of serious disability worldwide.
Thus, there is an urgent need for new approaches to disease-
specific diagnosis and prognosis, which will facilitate the
development of more precise and personalized treatment.

Metabolic Brain Disease
1 3
Metabolomics is a powerful tool in systems biology that
affords new insights into the pathological status of disease,
and studies in the past decade have demonstrated its poten-
tial utility for diagnosing and prognosing IS. In particular,
findings from the vast majority of studies suggest that many
metabolomic signatures could serve as biomarkers for the
early-stage diagnosis of IS, provide better prognostic evalu-
ations, and enable better evaluation of drug responses. How-
ever, despite the many IS-associated metabolites that have
been identified thus far, a clinical biomarker has yet to be
identified. Moreover, studies have been affected by con-
founders, such as drug-taking, diet, and comorbidity, which
profoundly influence metabolomic profiles and biomarker
discovery. Furthermore, the results of studies are inconsist-
ent in some areas because of study heterogeneity and the
range of experimental conditions and analytical techniques
used for metabolic profiling. Thus, the results are far from
being applicable to real-world practice.
Metabolomic profiling has been used to investigate IS for
more than 10years, with recent research having focused on eti-
ological, predictive, diagnostic, and prognostic aspects. Despite
the aforementioned limitations of these studies, the large vol-
ume of exploratory data has enhanced understanding of disease
mechanisms and generated a list of candidate biomarkers.
Future metabolomic studies should use standardized
experimental processes, upgraded analytical platforms, and
standard cohorts and perform refined data analyses. Insights
into disease mechanisms and disease biomarkers should be
achieved in part by integrated approaches based on a combi-
nation of metabolomics, other omics fields, and analyses of
the gut microbiota. The resulting candidate metabolites should
be subjected to validation studies before clinical application.
Authors’ contributions Meng Jiajia, Zhang Ruijie and Han Liyuan
wrote and revised the manuscript. All authors participated in critical
revision of the manuscript and approved the final version.
Funding This study is supported by the National Natural Science
Foundation of China (82173648), Innovative Talent Support Plan
of the Medical and Health Technology Project in Zhejiang Prov-
ince(2021422878), Internal Fund of Ningbo Institute of Life and Health
Industry, University of Chinese Academy of Sciences(2020YJY0212),
Sanming Project of Medicine in Shenzhen (SZSM201803080), Zhe-
jiang Provincial Public Service and Application Research Foundation
(LGF20H250001 and GC22H264267), Public Welfare Foundation of
Ningbo (2021S108), Ningbo Science and Technology Innovation 2025
Specific Project (2020Z096),and Shenzhen Nanshan District Science
and Technology Bureau (2020075).
Data Availability The datasets supporting the conclusions of this article
are available from the corresponding author.
Declarations
Ethics approval Not applicable.
Consent to participate Not applicable for animal studies.
Consent for publication Not applicable.
Competing Interests The authors have no relevant financial or non-
financial interests to disclose.
References
Adibhatla RM, Hatcher JF (2008) Altered lipid metabolism in brain
injury and disorders. Subcell Biochem 49:241–268. https:// doi.
org/ 10. 1007/ 978-1- 4020- 8831-5_9
Au A (2018) Metabolomics and Lipidomics of Ischemic Stroke. Adv
Clin Chem 85:31–69. https:// doi. org/ 10. 1016/ bs. acc. 2018. 02. 002
Azad RK, Shulaev V (2019) Metabolomics technology and bioin-
formatics for precision medicine. Br Bioinform 20:1957–1971.
https:// doi. org/ 10. 1093/ bib/ bbx170
Cambiaghi A, Ferrario M, Masseroli M (2017) Analysis of metab-
olomic data: tools, current strategies and future challenges for
omics data integration. Br Bioinform 18:498–510. https:// doi. org/
10. 1093/ bib/ bbw031
Castellanos M, Sobrino T, Pedraza S etal (2008) High plasma gluta-
mate concentrations are associated with infarct growth in acute
ischemic stroke. Neurology 71:1862–1868. https:// doi. org/ 10.
1212/ 01. wnl. 00003 26064. 42186. 7e
Chi NF, Chang TH, Lee CY etal (2021) Untargeted metabolomics
predicts the functional outcome of ischemic stroke. J Formos Med
Assoc 120:234–241. https:// doi. org/ 10. 1016/j. jfma. 2020. 04. 026
Chumachenko MS, Waseem TV, Fedorovich SV (2022) Metabolomics
and metabolites in ischemic stroke. Rev Neurosci 33:181–205.
https:// doi. org/ 10. 1515/ revne uro- 2021- 0048
Collaborators GBDLR of S, Feigin VL, Nguyen G etal (2018) Global,
Regional, and Country-Specific Lifetime Risks of Stroke, 1990
and 2016. N Engl J Med 379:2429–2437. https:// doi. org/ 10. 1056/
NEJMo a1804 492
Dang VT, Huang A, Werstuck GH (2018) Untargeted Metabolomics
in the Discovery of Novel Biomarkers and Therapeutic Targets
for Atherosclerotic Cardiovascular Diseases. Cardiovasc Hematol
Disord Drug Targets 18:166–175. https:// doi. org/ 10. 2174/ 18715
29X18 66618 04201 70108
Donatti A, Canto AM, Godoi AB, etal (2020) Circulating Metabolites
as Potential Biomarkers for Neurological Disorders-Metabolites in
Neurological Disorders. Metabolites 10. https:// doi. org/ 10. 3390/
metab o1010 0389
Everett JR, Holmes E, Veselkov KA etal (2019) A Unified Conceptual
Framework for Metabolic Phenotyping in Diagnosis and Progno-
sis. Trends Pharmacol Sci 40:763–773. https:// doi. org/ 10. 1016/j.
tips. 2019. 08. 004
Fatahzadeh M, Glick M (2006) Stroke: epidemiology, classification,
risk factors, complications, diagnosis, prevention, and medical
and dental management. Oral Surg Oral Med Oral Pathol Oral
Radiol Endod 102:180–191. https:// doi. org/ 10. 1016/j. tripl eo.
2005. 07. 031
Fluri F, Schuhmann MK, Kleinschnitz C (2015) Animal models of
ischemic stroke and their application in clinical research. Drug
Des Devel Ther 9:3445–3454. https:// doi. org/ 10. 2147/ DDDT.
S56071
Goulart VAM, Sena MM, Mendes TO etal (2019) Amino Acid Biosig-
nature in Plasma among Ischemic Stroke Subtypes. Biomed Res
Int 2019:8480468. https:// doi. org/ 10. 1155/ 2019/ 84804 68
Goyal M, Ospel JM, Menon B etal (2020) Challenging the Ischemic
Core Concept in Acute Ischemic Stroke Imaging. Stroke 51:3147–
3155. https:// doi. org/ 10. 1161/ STROK EAHA. 120. 030620

Metabolic Brain Disease
1 3
Guo X, Li Z, Zhou Y etal (2019) Metabolic profile for prediction of
ischemic stroke in Chinese hypertensive population. J Stroke Cer-
ebrovasc Dis 28:1062–1069. https:// doi. org/ 10. 1016/j. jstro kecer
ebrov asdis. 2018. 12. 035
Guo W, Wang Y, Fan M etal (2020) Integrating metabolomics and
network pharmacology to explore the protective effect of gross
saponins of Tribulus terrestris L. fruit against ischemic stroke in
rat. J Ethnopharmacol 263:113202. https:// doi. org/ 10. 1016/j. jep.
2020. 113202
Hernandez-de-Diego R, Tarazona S, Martinez-Mira C etal (2018)
PaintOmics 3: a web resource for the pathway analysis and visu-
alization of multi-omics data. Nucleic Acids Res 46:W503–W509.
https:// doi. org/ 10. 1093/ nar/ gky466
Holmes MV, Millwood IY, Kartsonaki C etal (2018) Lipids, Lipo-
proteins, and Metabolites and Risk of Myocardial Infarction and
Stroke. J Am Coll Cardiol 71:620–632. https:// doi. org/ 10. 1016/j.
jacc. 2017. 12. 006
Jia J, Zhang H, Liang X etal (2021) Application of Metabolomics to
the Discovery of Biomarkers for Ischemic Stroke in the Murine
Model: a Comparison with the Clinical Results. Mol Neurobiol
58:6415–6426. https:// doi. org/ 10. 1007/ s12035- 021- 02535-2
Johnson CH, Ivanisevic J, Siuzdak G (2016) Metabolomics: beyond
biomarkers and towards mechanisms. Nat Rev Mol Cell Biol
17:451–459. https:// doi. org/ 10. 1038/ nrm. 2016. 25
Jove M, Mauri-Capdevila G, Suarez I etal (2015) Metabolomics pre-
dicts stroke recurrence after transient ischemic attack. Neurology
84:36–45. https:// doi. org/ 10. 1212/ WNL. 00000 00000 001093
Kaiser EE, West FD (2020) Large animal ischemic stroke models:
replicating human stroke pathophysiology. Neural Regen Res
15:1377–1387. https:// doi. org/ 10. 4103/ 1673- 5374. 274324
Kimberly WT, Wang Y, Pham L etal (2013) Metabolite profiling
identifies a branched chain amino acid signature in acute cardi-
oembolic stroke. Stroke 44:1389–1395. https:// doi. org/ 10. 1161/
STROK EAHA. 111. 000397
Laborde CM, Mourino-Alvarez L, Akerstrom F etal (2012) Potential
blood biomarkers for stroke. Expert Rev Proteomics 9:437–449.
https:// doi. org/ 10. 1586/ epr. 12. 33
Latchaw RE, Alberts MJ, Lev MH etal (2009) Recommendations for
imaging of acute ischemic stroke: a scientific statement from the
American Heart Association. Stroke 40:3646–3678. https:// doi.
org/ 10. 1161/ STROK EAHA. 108. 192616
Lee Y, Khan A, Hong S etal (2017) A metabolomic study on high-risk
stroke patients determines low levels of serum lysine metabolites:
a retrospective cohort study. Mol Biosyst 13:1109–1120. https://
doi. org/ 10. 1039/ c6mb0 0732e
Li JT, Zeng N, Yan ZP etal (2021) A review of applications of metabo-
lomics in osteoarthritis. Clin Rheumatol 40:2569–2579. https://
doi. org/ 10. 1007/ s10067- 020- 05511-8
Liao S, Apaijai N, Chattipakorn N, Chattipakorn SC (2020) The pos-
sible roles of necroptosis during cerebral ischemia and ischemia
/ reperfusion injury. Arch Biochem Biophys 695:108629. https://
doi. org/ 10. 1016/j. abb. 2020. 108629
Lin B, Chen R, Wang Q etal (2021) Transcriptomic and metabolomic
profiling reveals the protective effect of Acanthopanax senticosus
(Rupr. & Maxim.) harms combined with Gastrodia Elata Blume
on cerebral Ischemia- Reperfusion injury. Front Pharmacol
12:619076. https:// doi. org/ 10. 3389/ fphar. 2021. 619076
Liu M, Liu X, Wang H etal (2016a) Metabolomics study on the effects
of Buchang Naoxintong capsules for treating cerebral ischemia
in rats using UPLC-Q/TOF-MS. J Ethnopharmacol 180:1–11.
https:// doi. org/ 10. 1016/j. jep. 2016. 01. 016
Liu M, Tang L, Liu X etal (2016b) An Evidence-Based Review of
Related Metabolites and Metabolic Network Research on Cerebral
Ischemia. Oxid Med Cell Longev 2016:9162074. https:// doi. org/
10. 1155/ 2016/ 91620 74
Liu P, Li R, Antonov AA etal (2017) Discovery of Metabolite Bio-
markers for Acute Ischemic Stroke Progression. J Proteome Res
16:773–779. https:// doi. org/ 10. 1021/ acs. jprot eome. 6b007 79
Luo L, Kang J, He Q, etal (2019) A NMR-Based Metabonomics
Approach to Determine Protective Effect of a Combination of
Multiple Components Derived from Naodesheng on Ischemic
Stroke Rats. Molecules 24. https:// doi. org/ 10. 3390/ molec ules2
40918 31
Makris K, Haliassos A, Chondrogianni M, Tsivgoulis G (2018) Blood
biomarkers in ischemic stroke: potential role and challenges in
clinical practice and research. Crit Rev Clin Lab Sci 55:294–328.
https:// doi. org/ 10. 1080/ 10408 363. 2018. 14611 90
Meng XE, Li N, Guo DZ etal (2015) High plasma glutamate levels
are associated with poor functional outcome in acute ischemic
stroke. Cell Mol Neurobiol 35:159–165. https:// doi. org/ 10. 1007/
s10571- 014- 0107-0
Montaner J, Ramiro L, Simats A etal (2020) Multilevel omics
for the discovery of biomarkers and therapeutic targets for
stroke. Nat Rev Neurol 16:247–264. https:// doi. org/ 10. 1038/
s41582- 020- 0350-6
Nicholson JK, Connelly J, Lindon JC, Holmes E (2002) Metabonomics:
a platform for studying drug toxicity and gene function. Nat Rev
Drug Discov 1:153–161. https:// doi. org/ 10. 1038/ nrd728
Nicholson JK, Lindon JC (2008) Systems biology: Metabonomics.
Nature 455:1054–1056. https:// doi. org/ 10. 1038/ 45510 54a
Paik MJ, Li WY, Ahn YH etal (2009) The free fatty acid metabolome
in cerebral ischemia following human mesenchymal stem cell
transplantation in rats. Clin Chim Acta 402:25–30. https:// doi.
org/ 10. 1016/j. cca. 2008. 12. 022
Papes F, Surpili MJ, Langone F etal (2001) The essential amino acid
lysine acts as precursor of glutamate in the mammalian central
nervous system. FEBS Lett 488:34–38. https:// doi. org/ 10. 1016/
s0014- 5793(00) 02401-7
Peng B, Li H, Peng XX (2015) Functional metabolomics: from bio-
marker discovery to metabolome reprogramming. Protein Cell
6:628–637. https:// doi. org/ 10. 1007/ s13238- 015- 0185-x
Qureshi MI, Vorkas PA, Coupland AP etal (2017) Lessons from Meta-
bonomics on the Neurobiology of Stroke. Neuroscientist 23:374–
382. https:// doi. org/ 10. 1177/ 10738 58416 673327
Rinschen MM, Ivanisevic J, Giera M, Siuzdak G (2019) Identifi-
cation of bioactive metabolites using activity metabolomics.
Nat Rev Mol Cell Biol 20:353–367. https:// doi. org/ 10. 1038/
s41580- 019- 0108-4
Saorin A, Di Gregorio E, Miolo G, etal (2020) Emerging Role of
Metabolomics in Ovarian Cancer Diagnosis. Metabolites 10.
https:// doi. org/ 10. 3390/ metab o1010 0419
Shin TH, Lee DY, Basith S, etal (2020) Metabolome Changes in Cer-
ebral Ischemia. Cells 9. https:// doi. org/ 10. 3390/ cells 90716 30
Sidorov E, Bejar C, Xu C etal (2020a) Potential Metabolite Biomarkers
for Acute Versus Chronic Stage of Ischemic Stroke: A Pilot Study.
J Stroke Cerebrovasc Dis 29:104618. https:// doi. org/ 10. 1016/j.
jstro kecer ebrov asdis. 2019. 104618
Sidorov E, Sanghera DK, Vanamala JKP (2019) Biomarker for
Ischemic Stroke Using Metabolome: A Clinician Perspective. J
Stroke 21:31–41. https:// doi. org/ 10. 5853/ jos. 2018. 03454
Sidorov EV, Bejar C, Xu C etal (2020b) Novel Metabolites as Potential
Indicators of Ischemic Infarction Volume: a Pilot Study. Transl
Stroke Res. https:// doi. org/ 10. 1007/ s12975- 020- 00876-z
Sommer CJ (2017) Ischemic stroke: experimental models and real-
ity. Acta Neuropathol 133:245–261. https:// doi. org/ 10. 1007/
s00401- 017- 1667-0
Suissa L, Guigonis JM, Graslin F etal (2020) Metabolome of cerebral
thrombi reveals an association between high glycemia at stroke
onset and good clinical outcome. Metabolites 10:1–15. https:// doi.
org/ 10. 3390/ metab o1012 0483

Metabolic Brain Disease
1 3
Sun D, Tiedt S, Yu B etal (2019a) A prospective study of serum metab-
olites and risk of ischemic stroke. Neurology 92:e1890–e1898.
https:// doi. org/ 10. 1212/ WNL. 00000 00000 007279
Sun R, Li Y, Cai M etal (2019b) Discovery of a new biomarker pattern
for differential diagnosis of acute ischemic stroke using targeted
metabolomics. Front Neurol 10:1011. https:// doi. org/ 10. 3389/
fneur. 2019. 01011
Sun H, Zhao J, Zhong D, Li G (2017) Potential serum biomarkers and
metabonomic profiling of serum in ischemic stroke patients using
UPLC/Q-TOF MS/MS. PLoS ONE 12:e0189009. https:// doi. org/
10. 1371/ journ al. pone. 01890 09
Tiedt S, Brandmaier S, Kollmeier H etal (2020) Circulating
Metabolites Differentiate Acute Ischemic Stroke from Stroke
Mimics. Ann Neurol 88:736–746. https:// doi. org/ 10. 1002/ ana.
25859
Tokarz J, Adamski J, Rizner TL (2020) Metabolomics for Diagnosis
and Prognosis of Uterine Diseases? A Systematic Review. J Pers
Med 10. https:// doi. org/ 10. 3390/ jpm10 040294
Tokarz J, Haid M, Cecil A etal (2017) Endocrinology Meets Metabo-
lomics: Achievements, Pitfalls, and Challenges. Trends Endo-
crinol Metab 28:705–721. https:// doi. org/ 10. 1016/j. tem. 2017.
07. 001
Ussher JR, Elmariah S, Gerszten RE, Dyck JR (2016) The Emerging
Role of Metabolomics in the Diagnosis and Prognosis of Cardio-
vascular Disease. J Am Coll Cardiol 68:2850–2870. https:// doi.
org/ 10. 1016/j. jacc. 2016. 09. 972
Vorkas PA, Shalhoub J, Lewis MR etal (2016) Metabolic Phenotypes
of Carotid Atherosclerotic Plaques Relate to Stroke Risk: An
Exploratory Study. Eur J Vasc Endovasc Surg 52:5–10. https://
doi. org/ 10. 1016/j. ejvs. 2016. 01. 022
Wang D, Kong J, Wu J etal (2017) GC-MS-based metabolomics
identifies an amino acid signature of acute ischemic stroke.
Neurosci Lett 642:7–13. https:// doi. org/ 10. 1016/j. neulet. 2017.
01. 039
Wang Y, Wang Y, Li M etal (2013) (1)H NMR-based metabolomics
exploring biomarkers in rat cerebrospinal fluid after cerebral
ischemia/reperfusion. Mol Biosyst 9:431–439. https:// doi. org/
10. 1039/ c2mb2 5224d
Wang Y, Wang YG, Ma TF etal (2014) Dynamic metabolites profile
of cerebral ischemia/reperfusion revealed by (1)H NMR-based
metabolomics contributes to potential biomarkers. Int J Clin Exp
Pathol 7:4067–4075
Wang Y, Zhao H, Liu Y, etal (2019) GC-MS-Based Metabolomics to
Reveal the Protective Effect of Gross Saponins of Tribulus ter-
restris Fruit against Ischemic Stroke in Rat. Molecules 24. https://
doi. org/ 10. 3390/ molec ules2 40407 93
Wesley UV, Bhute VJ, Hatcher JF etal (2019) Local and systemic
metabolic alterations in brain, plasma, and liver of rats in response
to aging and ischemic stroke, as detected by nuclear magnetic
resonance (NMR) spectroscopy. Neurochem Int 127:113–124.
https:// doi. org/ 10. 1016/j. neuint. 2019. 01. 025
Wishart DS (2019) Metabolomics for Investigating Physiological and
Pathophysiological Processes. Physiol Rev 99:1819–1875. https://
doi. org/ 10. 1152/ physr ev. 00035. 2018
Wu L, Chen C, Li Y etal (2020) UPLC-Q-TOF/MS-Based Serum
Metabolomics Reveals the Anti-Ischemic Stroke Mechanism of
Nuciferine in MCAO Rats. ACS Omega 5:33433–33444. https://
doi. org/ 10. 1021/ acsom ega. 0c053 88
Wu MY, Yiang GT, Liao WT etal (2018) Current Mechanistic Con-
cepts in Ischemia and Reperfusion Injury. Cell Physiol Biochem
46:1650–1667. https:// doi. org/ 10. 1159/ 00048 9241
Yang L, Wang Y, Cai H etal (2020) Application of metabolomics
in the diagnosis of breast cancer: a systematic review. J Cancer
11:2540–2551. https:// doi. org/ 10. 7150/ jca. 37604
Zahoor I, Rui B, Khan J etal (2021) An emerging potential of
metabolomics in multiple sclerosis: a comprehensive over-
view. Cell Mol Life Sci 78:3181–3203. https:// doi. org/ 10. 1007/
s00018- 020- 03733-2
Zhang Y, Li J, Zhang X etal (2021) Advances of Mechanisms-Related
Metabolomics in Parkinson’s Disease. Front Neurosci 15:614251.
https:// doi. org/ 10. 3389/ fnins. 2021. 614251
Zhou B, Xiao JF, Tuli L, Ressom HW (2012) LC-MS-based metabolomics.
Mol Biosyst 8:470–481. https:// doi. org/ 10. 1039/ c1mb0 5350g
Publisher's note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.