Sage Journals: Discover world-class research

Abstract

Transient ischemic attacks (TIAs) before an acute ischemic stroke (AIS) could induce ischemic tolerance (IT) phenomena. with an endogenous neuroprotective role (Ischemic preconditioning. IPC). A consecutive prospective cohort of patients with AIS were recruited from 8 different hospitals. Participants were classified by those with non-previous recent TIA vs. previous TIA (within seven days. TIA ≤7d). A total of 541 AIS patients were recruited. 40 (7.4%). of them had previous TIA ≤7d. In line with IPC. patients with TIA ≤7d showed: 1) a significantly less severe stroke at admission by NIHSS score. 2) a better outcome at 7–90 days follow-up and reduced infarct volumes. 3) a specific upregulated metabolomics/lipidomic profile composed of diverse lipid categories. Effectively. IPC activates an additional adaptive response on increasing circulation levels of structural and bioactive lipids to facilitate functional recovery after AIS which may support biochemical machinery for neuronal survival. Furthermore. previous TIA before AIS seems to facilitate the production of anti-inflammatory mediators that contribute to a better immune response. Thus. the IT phenomena contributes to a better adaptation of further ischemia. Our study provides first-time evidence of a metabolomics/lipidomic signature related to the development of stroke tolerance in AIS patients induced by recent TIA.

Keywords

Transient ischemic attack acute ischemic stroke metabolomics lipidomics ischemic preconditioning

Introduction

Ischemic preconditioning (IPC) consists of brief periods of ischemia that confers protection against later episodes of prolonged ischemia (ischemic tolerance [IT]).¹ The IT-induced mechanisms evoked by previous brief ischemia appear complex and involve unique cellular gene expression changes and metabolic and signaling pathways.¹ The IT could be induced either by local. global. and remote IPC.^1–3

IT could be activated in clinical scenarios by a recent transient ischemic attack (TIA) before an acute ischemic stroke (AIS).^4–7 It has been reported that patients with previous TIA before an AIS had better outcome^4–7 and smaller infarct volumes^6,8 than patients without previous TIA. although data from two studies showed no association between previous TIA and low stroke severity.^9,10

Metabolomics. including lipidomic analysis. is the comprehensive analysis of a biological sample with high-sensitivity to characterize changes in low-molecular-weight molecules. The plasma metabolome study arises as a powerful functional tool to describe phenotypes in both physiological and pathological states.¹¹ Consequently. metabolomics emerges as a promising approach to propose potential prognostic and recurrence biomarkers of cerebrovascular diseases¹² and a plasma biomarker related to neuroimaging findings.¹³

The present study was designed to address the IT phenomenon by the effect of a previous TIA on AIS patients' outcome and to describe its metabolic signature.

Materials & methods

Study participants – cohort description

The present study follows the STROBE statement.¹⁴ A cohort of AIS patients was consecutively and prospectively recruited. Patients were recruited between January 2011 and January 2016 following the same procedure (TIA/IS diagnose. neuroimaging confirmation. blood extraction time) from eight Hospitals in Spain (Hospital Universitari Arnau de Vilanova-Lleida. Hospital Universitario-Pamplona. Hospital del Mar-Barcelona. Hospital de son Espases-Mallorca. Hospital Parc Taulí-Sabadell. Hospital de Sant Pau-Barcelona. Hospital Ramón y Cajal-Madrid. Hospital de Albacete). A neurologist attended AIS patients during the first 24 h after the onset of symptoms. All recruited patients underwent neuroimaging diagnosis. Patients with symptoms that persisted less than 24 h without clear evidence of acute ischemic lesion in the neuroimaging were excluded from the study. Other exclusion criteria were a modified Rankin Scale Score (mRS) greater than 2. more than 24 h after symptoms onset. or patients under 18 years old. Blood from patients was collected immediately upon hospital arrival at the emergency room by standard venipuncture. Plasma was obtained after centrifugation 4°C at 1500 g for 10 min and aliquoted into cryovials for prompt storage at −80°C. In all cases. blood samples were obtained before intravenous fibrinolysis or thrombectomy.

Patients were classified etiologically according to TOAST criteria (Trial of Org 10172 in Acute Stroke Treatment) ¹⁵ as a large-artery occlusive disease (LAA). small-vessel disease (SV). cardioembolism (CE). other cause (OC). or undetermined cause (UND). The following patient's characteristics were collected: age at admission. gender. vascular risk factors. prior treatments. premorbid mRS. and stroke severity on the National Institutes of Health Stroke Scale (NIHSS) on admission and after seven days. Oxfordshire topographic classification was also assessed.

The Ethics Committee approved this study protocol of the Hospital Universitari Arnau de Vilanova (approval code: 928). Written informed consent was obtained from the patients or patients' surrogates in all participating Hospitals. All procedures followed the Declaration of Helsinki guidelines developed by the World Medical Association regarding ethical principles for medical research involving human subjects.

According to previous evidence.⁷ patients were classified and analyzed into two groups according to the time interval between prior TIA and IS: i) TIA within the preceding seven days. and ii) non-TIA or TIA before the preceding seven days. Families of patients with language impairment or disturbances of consciousness and their general practitioners were interviewed to identify prior TIA events.

Previous TIA was defined as a reversible episode of neurological deficit of ischemic origin that resolved totally within 24 h.¹⁶ All baseline data were centrally monitored. and queries were sent to each center's enrolling stroke neurologists. The primary outcome was the degree of functional impairment at three months. measured by mRS ≤1 at 90 days. We employed a face-to-face interview in each hospital to follow up with patients.

Final infarct volume was measured in 396 (73.2%) patients by diffusion-weighted magnetic resonance imaging (DWI) performed within seven days (median 2.8 days: interquartile range: 1.6 to 4.6). One neuroradiologist blinded to clinical features established the presence of DWI abnormalities. Furthermore. OsiriX v4.0 semi-automated segmentation tool and imaging software (Bernex. Switzerland) was used to calculate the total volume of DWI abnormal signal intensity. Polygon tool was used to assess the presence of abnormalities. regions of interest (ROIs) delineated between segmented slices and automatically interpolated. and the infarct volume determined.¹³

Clinical lipid and inflammatory biomarkers panel

Lipid profile (total cholesterol, HDL cholesterol, calculated LDL and triglycerides), fibrinogen, leukocytes, glycemia, and creatinine levels were obtained from clinical history reports by a routine blood test at hospital admission. Interleukin-6 (IL-6), N-terminal pro-B type natriuretic peptide (NT-proBNP), high-sensitivity C-reactive protein (hsCRP), high sensitive troponin (hsTroponin), neuron-specific enolase (NSE), and S-100b protein levels were assessed in plasma samples (Hoffmann-La Roche. Basel. Switzerland) by an electrochemical chemiluminescence immunoassay using the COBAS 6000 e601 (Hoffmann-La Roche. Basel. Switzerland) following previously published protocol.¹⁷

Metabolomic and lipidomic sample processing

Non-targeted metabolomic and lipidomic profiling was performed on an Agilent 1290 LC system coupled to an electrospray-ionization quadruple time of flight mass spectrometer (Q-TOF. 6520 instrument. Agilent Technologies. Barcelona. Spain). All samples were analyzed together. We applied two different protocols to cover as many metabolites as possible. from more polar ones like amino acids to more apolar like cholesterol esters (CE) or triglycerides (TG).

For metabolomics analysis. plasma samples were processed as described previously.¹² Briefly. 30 μl of cold methanol (containing phenylalanine-C13 as an internal standard) were added to 10 μl of plasma. incubated for one hour at −20°C. and centrifuged for three minutes at 12.000 g. The supernatant was recovered. filtered with a 0.2 μm Eppendorf filter. Two µL of the extracted sample were applied onto a reversed-phase column (Zorbax SB-Aq 1.8 µm 2.1 × 50 mm; Agilent Technologies) equipped with a pre-column (Zorbax-SB-C8 Rapid Resolution Cartridge 2.1 × 30 mm 3.5 µm; Agilent Technologies) with a column temperature of 60°C. The flow rate was 0.6 mL/min. Solvent A was composed of water containing 0.2% acetic acid. and solvent B was composed of methanol 0.2% acetic acid. The gradient started at 2% B and increased to 98% B in 13 min and held at 98% B for 6 min. Post-time was established in 5 min.

For lipidomics analysis. plasma samples were processed as described previously.¹⁸ Briefly. to precipitate the protein fraction. five μl of Mili Q water and 20 μl of methanol were added to 10 μl of a plasma sample. After the addition. samples were shaken for 2 min. Then. 250 μl of MTBE plus internal standards were added. and samples were ultra-sounded in a water bath (ATU Ultrasonidos. Valencia. Spain) with a frequency and power of 40 kHz and 100 W, respectively, at 10°C for 30 min. Then. 75 μL of Mili Q water was added to the mixture, and the organic phase was separated by centrifugation (1.400 g) at 10°C for 10 min. The upper phase. containing all the extracted lipid species. was collected and subjected to mass spectrometry. A pool of all lipid extracts was prepared and used as quality controls (QC) as previously described.¹⁹ Ten μl of lipid extract was applied onto 1.8 μm particle 100 × 2.1 mm id Waters Acquity HSS T3 column (Waters. Milford. MA. USA) heated at 55°C. The flow rate was 400 μl/min with solvent A composed of 10 mM ammonium acetate in acetonitrile-water (40:60. v/v) and solvent B composed of 10 mM ammonium acetate in acetonitrile-isopropanol (10:90. v/v). The gradient started at 40% B and reached 100% B in 10 min and held for 2 min. Finally, the system was switched back to 40% B and equilibrated for 3 min. as previously described.²⁰

Metabolomic and lipidomic data acquisition and analysis

Data were collected in positive electrospray mode time of flight operated in full-scan mode at 50–3000 m/z in an extended dynamic range (2 GHz), using N2 as the nebulizer gas (5 L/min. 350°C). The capillary voltage was 3500 V with a scan rate of 1 scan/s. The ESI source used a separate nebulizer for the continuous. low-level (10 L/min) introduction of reference mass compounds 121.050873 and 922.009798. used for continuous. online mass calibration. Mass Hunter Data Analysis Software (Agilent Technologies. Barcelona. Spain) was used to collect the results. and Mass Hunter Qualitative Analysis Software (Agilent Technologies. Barcelona. Spain) to obtain the molecular features of the samples. representing different. co-migrating ionic species of a given molecular entity using the Molecular Feature Extractor algorithm (Agilent Technologies. Barcelona. Spain). as described.¹² We selected samples with a minimum of 2 ions. MassHunter Mass Profiler Professional Software (Agilent Technologies. Barcelona. Spain) was used to select. align. and filter molecular features. Multiple charge states were considered for metabolite profiling only in lipidomics. Compounds from different samples were aligned using a retention time window of 0.1% ± 0.25 minutes and a mass window of 30.0 ppm ± 2.0 mDa. We selected only those present in at least 50% of one group's samples and corrected for individual bias.

Identities were confirmed based on exact mass. retention time. isotopic distribution. and MS/MS spectrum using public databases such as Metlin.²¹ HMDB.²² and LipidMatch.²³

Metaboanalyst platform²⁴ was used to perform multivariate statistics (partial least squares discriminant analyses (PLS-DA) of the extracted features. R software was used to find potential biomarkers of TIA neuroprotection prior to AIS and perform Spearman correlations. Robust linear models were performed using the MASS package's rlm function. considering the metabolite or lipid intensity as a response to the group (previous TIA vs. non-previous TIA). P-values for the regression coefficients were obtained by bootstrapping using the car package's Boot function and were adjusted by Benjamini-Hochberg's False Discovery Rate. In order to adjust for potential confounding. samples were weighted using the overlap weighting method. which has been demonstrated to produce an unbiased treatment effect estimate. with lower variance compared with other propensity score based weighting methods.²⁵ Propensity scores for TIA ≤7 days were estimated using a logistic regression. which included baseline characteristics (age. sex. smoking habits. hypertension. diabetes mellitus. hyperlipidemia. previous atrial fibrillation and ischemic heart disease). previous treatment (anticoagulants. antiplatelets. statins. antihypertensives and antidiabetics) and stroke etiology (LAA/non-LAA). Chord diagram for correlational analyses between selected lipid species was performed using Circos Table Viewer (http://mkweb.bcgsc.ca/tableviewer/visualize/).

Statistical analysis

Sample size was initially calculated considering studies with a clinical endpoint of mRS < 2 or similar: Moncayo et al.²⁶ Schaller et al.²⁷ and Wieh et al.²⁸ We initially calculated a sample size of 490 subjects to achieve 80% power to detect differences of 15% between groups using a two-sided binomial test at significance level of 0.05. assuming that assuming that 56% of non-previous TIA patients will experience good outcome. Finally. we determine a temporal endpoint to increase the number of recruited patient in order to analyse previous TIA patients subgroups.

Data are presented either as the mean ± standard deviation. the median (interquartile range). or percentage. Categorical variables were expressed as numbers and percentages and compared with Fisher's exact or ordinal chi-square tests. as appropriate. Continuous variables were expressed as medians and interquartile ranges or as means and standard deviations and were compared with the use of Student's t-test. ANOVA. and Mann-Whitney test. Univariate analyses were performed to detect variables associated with the presence of previous TIA ≤7 days and with a mRs ≤1 at 3 months. We used multivariable logistic backward stepwise regression analysis to obtain adjusted odds ratios. Variables for which P < 0.10 in univariate testing were included. Statistical analysis was performed using SPSS software version 20 (SPSS. Chicago. IL) or Stata 11 statistics package (StataCorp. College Station. TX). Statistical significance was considered when p < 0.05.

Data availability

Anonymized data will be shared by request from any qualified investigator on reasonable requests.

Results

The cohort included 541 patients (151 excluded patients). Of the 541 enrolled patients. 40 (7.4%) patients had a previous TIA ≤7 days (Figure 1). Previous TIA was ipsilateral to the subsequent stroke in 37 (92.5%) cases and unique in 23 (59.0%) cases (Table 2).

Figure 1.

Flow diagram of recruited cohort. excluded patients. and the number of final included patients in the metabolomic and lipidomic analysis.

Age. basal NIHSS and LAA were independently related to previous TIA: Odds ratio 0.97 (95% CI: 0.95–1.00). P = 0.041; OR 0.88 (95% CI 0.81–0.95). P = 0.002; and OR 2.11 (1.01–4.43). P = 0.048 respectively. Furthermore. the outcome at 7 and 90 days measured by disability or dependence in daily activities (mRS) and NIHSS at seven days was better in previous TIA ≤7d patients than non-previous TIA patients (Table 1). Moreover. previous TIA ≤7d patients' infarct volumes were significantly reduced (75% lower. median volume). Additionally. IL-6. NT-proBNP. S100B. and hs-troponin were decreased in the previous TIA ≤7d group (Table 1).

Table 1.

Clinical characteristics of stratified participants by non-previous TIA vs. previous TIA ≤7 days.

	All	Non-previous TIA	Previous TIA ≤7 days	p-value^a
Total. no. (%)	541 (100)	501 (92.6)	40 (7.4)
Age. (yr)	75.1 (11.6)	75.5 (11.5)	69.4 (11.8)	0.001
Baseline characteristics. no. (%)
Female sex	235 (43.4)	223 (44.5)	12 (30.0)	0.823
Caucasian-Mediterranean	534 (98.7)	494 (98.6)	40 (100)	0.904
Smoking habits
Smoker (>10c/day)	102 (19.0)	88 (17.6)	14 (35.9)	0.005
Unknown	3
Hypertension	409 (76.2)	383 (77.1)	26 (65.0)	0.085
Unknown	4
Diabetes mellitus	157 (29.1)	148 (29.7)	9 (22.5)	0.338
Unknown	2
Hyperlipidemia	232 (43.0)	215 (42.9)	17 (43.6)	0.935
Previous Atrial fibrillation	141 (26.2)	135 (27.1)	6 (15.4)	0.131
Unknown	3
Ischemic heart disease	61 (11.3)	59 (11.8)	2 (5.1)	0.294
Unknown	2
Biochemical parameters. median (IQR)
Cholesterol total	170.0 (145.0–199.0)	169.0 (145.0–197.0)	185.0 (164.0–207.0)	0.100
LDL	103.0 (81.0–125.0)	101.0 (80.0–124.0)	113.5 (91.0–133.0)	0.093
HDL	43.0 (36.0–52.0)	44.0 (36.5–52.0)	39.5 (35.0–48.0)	0.160
Triglycerides	112.0 (85.0–152.0)	113.0 (87.0–150.0)	106.0 (82.0–178.0)	0.871
Fibrinogen	4.5 (3.9–5.1)	4.5 (3.9–5.0)	4.2 (3.6–5.2)	0.379
Leukocytes	8.2 (6.7–10.0)	8.2 (6.7–10.0)	8.4 (6.7–11.2)	0.397
Glycaemia	118.0 (103.0–147.0)	118.5 (103.0–146.0)	117.0 (104.5–148.5)	0.867
Creatinin	0.88 (0.75–1.10)	0.89 (0.75–1.10)	0.85 (0.72–1.06)	0.532

NSE	11.6 (9.4–16.6)	11.7 (9.5–16.6)	10.6 (8.4–14.2)	0.115
IL6	8.65 (4.90–16.20)	8.90 (5.00–16.70)	6.00 (4.20–10.05)	0.015
NT-proBNP	505.2 (145.0–1478.0)	546.5 (154.8–1570.0)	229.5 (65.0–619.7)	0.002
S100B	67.0 (42.4–109.0)	68.8 (44.0–117.8)	54.0 (35.9–74.2)	0.001
Hs-CRP	4.2 (1.9–12.5)	4.2 (2.0–13.0)	4.2 (1.5–8.7)	0.225
Hs-Troponin	14.1 (8.0–26.4)	14.9 (8.8–26.7)	11.0 (4.6–16.6)	0.004
Previous treatment. no. (%)
Anticoagulants	74 (13.7)	70 (14.0)	4 (10.0)	0.482
Antiplatelets	200 (37.0)	190 (38.0)	10 (25.0)	0.101
Statins	184 (34.1)	171 (34.2)	13 (32.5)	0.827
Antihypertensives	362 (66.9)	341 (68.1)	21 (52.5)	0.044
Antidiabetics	120 (22.2)	113 (22.6)	7 (17.5)	0.455
Management. no. (%)
Stroke on awakening	86 (15.9)	79 (15.8)	7 (17.5)	0.773
Stroke code activation	322 (59.5)	309 (61.7)	13 (32.5)	0.002
Revascularization therapy
Rt-PA	105 (19.4)	98 (19.0)	7 (17.5)	0.808
Rt-PA plus Thrombectomy	33 (6.1)	32 (6.4)	1 (2.5)
Thrombectomy	5 (0.9)	5 (1.0)	0 (0)
Admission site
Neurology department	225 (41.6)	201 (40.1)	24 (60.0)	0.021
Stroke Unit	277 (51.2)	261 (52.1)	16 (40.0)
Others	39 (7.2)	39 (7.8)	0 (0)
Stroke severity (IQR)
Premorbid mRS^b	0.0 (0.0–1.0)	0.0 (0.0–1.0)	0.0 (0.0–1.0)	0.992
NIHSS at admission^c	5.0 (3.0–11.0)	6.0 (3.0–13.0)	4.0 (1.0–6.0)	<0.001
Stroke etiologies. no. (%)
LAA	89 (16.5)	77 (15.4)	12 (30.0)	0.005
CE	233 (42.0)	224 (44.7)	9 (22.5)
Lacunar	82 (15.2)	72 (14.4)	10 (25.0)
Other uncommon causes	10 (1.8)	8 (1.6)	2 (5.0)
Indeterminate	127 (23.5)	120 (24.0)	7 (17.5)
Topographic classification. n (%)
TACI	139 (26.2)	135 (27.4)	4 (10.8)	0.166
PACI	200 (37.7)	182 (36.9)	18 (48.6)
POCI	51 (9.6)	47 (9.5)	4 (10.8)
LACI	140 (26.4)	129 (26.2)	11 (29.7)
unknown	11 (2.03)
DWI volume. (IQR) mL	2.8 (0.6–16.1)	3.5 (0.7–17.6)	0.8 (0.2–8.2)	0.004
Clinical evolution. (IQR)
NIHSS at 7 days	2.0 (0–6.0)	2.0 (1.0–6.0)	1.0 (0–2.0)	<0.001
mRS at 7 days	2.0 (1.0–4.0)	2.0 (1.0–4.0)	2.0 (0.0–3.0)	0.003
mRS at 90 days	2.0 (1.0–3.0)	2.0 (1.0–4.0)	1.5 (0.0–2.0)	0.007
mRS <3 at 90 days	308 (58.0)	279 (56.8)	29 (72.5)	0.037
Unknown	10
Mortality at 90 days. no. (%)	75 (14.1)	74 (15.1)	1 (2.5)	0.028
Unknown	10

TIA: transient ischemic attack; rt-PA: recombinant tissue plasminogen activator; mRS: modified Rankin Score; NIHSS: National Institutes of Health Stroke Scale; LAA: large-artery occlusive disease; CE: cardioembolism; TACI: total anterior circulation infarcts; PACI: partial anterior circulation infarcts; POCI: posterior circulation infarcts; LACI: lacunar infarcts; DWI: diffusion-weighted imaging; IQR: interquartile range.

Plus–minus values are means ± SD. Percentages may not total 100 because of rounding.

^aStudent t-test.

^bScores on modified Rankin Score range from 0 to 5. With higher scores indicating major disability or dependence in the daily activities.

^cScores on the National Institutes of Health Stroke Scale range from 0 to 42. With higher scores indicating impairment after stroke.

Table 2.

Comparison of clinical characteristics according to 90-day modified Rankin score outcome.

	All	mRS > 1^a	mRS ≤1^a	p-value^b
Total. No. (%)	531 (100)	295 (55.6.0)	236 (44.4)
Age. (yr)	75.1 (11.6)	78.1 (10.9)	71.2 (11.3)	<0.001
Baseline characteristics. N (%)
Female sex	230 (43.3)	146 (49.5)	84 (35.6)	0.001
Caucasian-Mediterranean	524 (98.7)	291 (98.6)	233 (98.7)	0.831
Alcoholism
Yes	32 (6.1)	8 (2.7)	24 (10.2)	0.001
Unknown	5
Smoking habits
Smoker (>10c/day)	100 (8.9)	40 (13.6)	60 (25.5)	<0.001
Unknown	2
Hypertension	404 (76.2)	240 (81.4)	164 (69.8)	0.002
Unknown	1
Diabetes mellitus	155 (29.3)	93 (31.6)	62 (26.4)	0.187
Unknown	2
Hyperlipidemia	225 (42.4)	126 (42.7)	99 (41.9)	0.860
Previous Atrial fibrillation	138 (26.0)	83 (28.1)	55 (23.4)	0.218
Unknown	1
Ischemic heart disease	60 (11.3)	32 (10.9)	28 (11.9)	0.723
Unknown	1
Previous treatment. no. (%)
Anticoagulants	72 (13.6)	44 (14.9)	28 (11.9)	0.308
Antiplatelets	198 (37.4)	121 (41.2)	77 (32.6)	0.044
Statins	180 (34.0)	101 (34.4)	79 (33.5)	0.832
Antihypertensives	355 (66.9)	213 (72.2)	142 (60.2)	0.003
Antidiabetics	118 (22.3)	52 (23.3)	66 (21.5)	0.619
Stroke severity (IQR)
Premorbid mRS	0.0 (0.0–1.0)	0.0 (0.0–1.0)	0.0 (0.0–1.0)	0.836
NIHSS at admission^c	5.0 (3.0-12.0)	9.0 (4.0-17.5)	4.0 (2.0-7.0)	<0.001
Previous TIA
Non previous TIA ≤7d	491 (92.5)	282 (95.6)	209 (88.6)	0.004
Previous TIA ≤24 h	25 (4.7)	10 (3.4)	15 (6.4)
Previous TIA >24 h ≤7d	15 (2.8)	3 (1.0)	12 (5.1)
Number of previous TIA ≤7d				0.929
One	23 (59.0)	8 (61.5)	15 (57.7)
>1 and ≤3	12 (30.8)	4 (30.8)	8 (30.8)
>3	4 (10.3)	1 (7.7)	3 (11.5)
unknown	2
Ipsilateral	37 (92.5)	13 (100)	24 (88.9)	0.213
Stroke etiologies. No. (%)
LAA	89 (16.8)	55 (18.6)	34 (14.4)	0.001
CE	229 (43.1)	138 (46.8)	91 (38.6)
Lacunar	80 (15.1)	28 (9.5)	52 (22.0)
Other uncommon causes	10 (1.9)	4 (1.4)	6 (2.5)
Indeterminate	123 (23.2)	70 (23.7)	53 (22.5)
Topographic classification (IQR)
TACI	137 (26.3)	112 (38.5)	25 (10.9)	<0.001
PACI	194 (37.2)	96 (33.0)	98 (42.6)
POCI	51 (9.8)	24 (8.2)	27 (11.7)
LACI	139 (26.7)	59 (20.3)	80 (34.8)
unknown	10
DWI volume. (IQR) mL	2.5 (0.6–15.3)	7.1 (1.0–37.8)	1.4 (0.3–6.9)	0.064

Plus–minus values are means ± SD. Percentages may not total 100 because of rounding.

^aScores on modified Rankin Score range from 0 to 5. with higher scores indicating major disability or dependence in the daily activities.

^bStudent t-test.

^cScores on the National Institutes of Health Stroke Scale range from 0 to 42. with higher score indicating impairment after stroke.

Table 2 shows variables associated with mRs ≤1 at 3 months. In the multivariate regression analysis. age (OR 0.96 [95% CI: 0.94–0.98]. P < 0.001). basal NIHSS (OR 0.87 [95% CI: 0.84–0.90]. P < 0.001). LAA (OR 0.50 [95% CI: 0.29–0.85]. P = 0.011) and previous TIA ≤7d (OR 2.18 [95% CI: 1.01–4.73]. P = 0.049) were predictors of mRs ≤1 at 3 months.

Untargeted metabolomics and lipidomics analysis were performed to investigate whether previous TIA ≤7d patients had a plasma metabolome signature expressing an adaptive metabolic response with a potential neuroprotective role. The metabolomics approach revealed 18.332 molecular features. whereas in the lipidomic approach. we found 3.215. To define the most robust and reliable molecules. we only considered those present in more than 50% of the samples. Then. metabolomics approach revealed 134 molecules and lipidomics approach 201 molecules.

Multivariate statistical approach (PLS-DA) of plasma metabolites and lipids did not reveal the existence of a specific circulating signature of TIA ≤7d patients in both metabolomics and lipidomics analysis when the global metabolome and lipidome were analyzed (Figure 2(a) and (b)). However, univariate analysis revealed 53 metabolites and lipid species differentially expressed in TIA ≤7d patients (Table 3). and 4 of them obtained FDR p-value <0.05. Among the 53 differential species. 27 (51.0%) were identified based on exact mass. retention time. and/or MS/MS spectrum. mainly lipid metabolism-related species (23/27. 85%). All of the 27 identified species were increased in the TIA ≤7d group. For graphic representation. see https://aitmetllip.pythonanywhere.com/. When identified molecules were evaluated according to their families. we found 5 molecules belonging to glycerolipids (GL). 14 glycerophospholipids (GPL). 3 sphingolipids (SL). Additionally. we identified 3 indoles and derivatives. an isoleucine and an hydroxydodecanoic acid. Regarding GL. we found 2 diacylglycerides (DG(18:1/16:0). DG(34:2)). 2 oxidized triacylglycerides (OxTG(18:1/18:1/18:2);1O. OxTG(55:1);2O) and one triacylglyceride (TG(21:0/18:1/18:2)). Among GPL. we found 4 phosphatidylethanolamine (LysoPE(18:4) LysoPE(20:4). LysoPE(22:4). PE(P-38:6)). 8 phosphatidylcholine (PC(32:2). PC(34:1). PC(38:6). PC(40:6) . PC(42:5). PC(42:6). PC(O-36:4)/PC(P-36:3). PC(O-42:8)). 1 phosphatidylinositol (PI(38:4)) and 1 phosphatidylserine (PS(O-29:0)). Three of these species were lysophospholipids (LysoPE(18:4) LysoPE(20:4). LysoPE(22:4)). Concerning SL we identified two sphingomyelins (SM) (SM(d35:1). SM(d36:2) and one ganglioside. Ganglioside GM3 (d38:0)).

Figure 2.

Multivariate statistics revealed that suffering a TIA ≤7 days before IS determines specific plasma metabolomics (a); and lipidomics (b) signatures. Previous TIA ≤7d patients are represented in green spots and non-previous TIA in red. PLS-DA cross-validation details are: metabolomics (Accuracy: 0.93. R2: 0.054. Q2: −0.05) and lipidomics (Accuracy: 0.92. R2: 0.063. Q2: −0.03). (c) Graphical visualization of significant differential lipid correlation in non-previous TIA compared to prior TIA ≤7 days patients (d). Each ribbon indicates a significant lipid correlation. Ribbon thickness refers to the degree of correlation (p values for Spearman's correlation <0.01).

Table 3.

Metabolite and lipid species differentially expressed in TIA ≤7d patients adjusted by baseline characteristics, previous treatment and stroke etiology.

Class	Feature	m/z	RT	p-value	FDR p-value	Regression Coefficient	Regulation^a
Carboxylic acids and derivatives	Isoleucine^b	132.10374	0.6	0.037	0.272	0.116122025	up
Glycerolipids	DG(18:1/16:0)^b	612.53624	8.4	0.04	0.285	0.435293876	up
	DG(34:2)^b	610.54254	8.1	0.005	0.148	0.150095141	up
	OxTG(18:1/18:1/18:2);1O^b	914.78714	9.9	<0.001	<0.001	0.663651095	up
	OxTG(55:1);2O^b	944.84299	10.2	0.007	0.149	0.164856004	up
	TG(21:0/18:1/18:2)^b	952.87134	10.6	0.016	0.194	0.047775506	up
Glycerophospholipids	LysoPE(18:4)^b	474.25549	0.9	0.034	0.27	0.070198909	up
	LysoPE(20:4)^b	502.28659	0.9	0.02	0.201	0.10714783	up
	LysoPE(22:4)^b	530.31429	1.3	0.005	0.148	0.044700252	up
	PC(32:2)^b	730.54009	6.5	0.017	0.194	0.244202842	up
	PC(34:1)^b	760.58814	7.5	0.046	0.289	0.084488518	up
	PC(38:6)^b	806.57169	7	0.016	0.194	0.151762678	up
	PC(40:6)^b	834.60219	7.5	0.035	0.271	0.060621122	up
	PC(42:5)^b	864.63304	6.6	<0.001	<0.001	0.056937524	up
	PC(42:6)^b	862.62524	6.6	0.027	0.235	0.042149119	up
	PC(O-36:4)/PC(P-36:3)^b	768.58289	7.5	0.026	0.231	0.514944707	up
	PC(O-42:8)^b	844.62564	8.1	0.032	0.263	0.088392924	up
	PE(P-38:6)^b	748.53399	7.3	0.043	0.286	0.117345022	up
	PI(38:4)^b	904.59794	6.6	0.023	0.22	0.087744117	up
	PS(O-29:0)^c	680.48254	8.1	0.032	0.263	0.479873904	up
Hydroxy acids and derivatives	Hydroxydodecanoic acid^b	199.17219	10.2	0.016	0.194	0.468508985	up
Indoles and derivatives	Indole-3-carboxaldehyde^b	146.06264	1.2	0.039	0.283	0.068775765	up
	Indoleacrylic acid^b	188.07489	1.2	0.01	0.164	0.185178092	up
	Tryptophol^b	144.08264	1.2	0.015	0.194	0.09828777	up
Sphingolipids	Ganglioside GM3 (d38:0)^c	1210.81134	6.9	0.013	0.194	0.052732371	up
	SM(d35:1)^b	717.58259	7.1	0.008	0.149	0.331888813	up
	SM(d36:2)^b	729.59199	6.9	0.043	0.286	0.039341587	up
Unknown	100.11089@5.8	100.11089	5.8	0.002	0.129	0.410397654	up
	150.06889@0.7	150.06889	0.7	0.025	0.228	0.208047033	up
	257.06179@5	257.06179	5	0.004	0.147	0.443739639	up
	306.32799@2.8	306.32799	2.8	0.016	0.194	0.458807743	up
	307.33524@2.8	307.33524	2.8	0.01	0.164	0.41567161	up
	349.77374@5.5	349.77374	5.5	0.042	0.286	0.143427371	up
	355.22879@5.9	355.22879	5.9	0.005	0.148	0.172059784	up
	372.25634@4.4	372.25634	4.4	0.023	0.22	0.192514316	up
	400.13639@2.7	400.13639	2.7	0.045	0.289	0.433893866	up
	505.40169@6.3	505.40169	6.3	0.02	0.201	0.037125373	up
	573.46544@7.9	573.46544	7.9	0.046	0.289	0.152046409	up
	575.28819@5.8	575.28819	5.8	0.007	0.149	0.073587699	up
	640.52329@7	640.52329	7	0.008	0.149	0.134167297	up
	722.71024@6.9	722.71024	6.9	0.002	0.129	0.211029338	up
	748.72734@7.4	748.72734	7.4	0.036	0.272	0.059723258	up
	947.28559@9.4	947.28559	9.4	0.017	0.194	0.424719881	up
	1021.30324@9.6	1021.30324	9.6	0.005	0.148	0.060502811	up
	1038.25784@7.2	1038.25784	7.2	0.029	0.249	0.64978914	up
	1217.06494@8.4	1217.06494	8.4	<0.001	<0.001	0.167569121	up
	1342.92484@8.1	1342.92484	8.1	0.003	0.129	0.438353527	up
	1398.08554@7.1	1398.08554	7.1	0.011	0.169	0.326734449	up
	1409.40634@8.8	1409.40634	8.8	0.019	0.201	0.069570968	up
	1412.99434@8.1	1412.99434	8.1	0.006	0.149	0.151652795	up
	1495.43374@7.4	1495.43374	7.4	0.01	0.164	0.627217231	up
	1541.13214@7.1	1541.13214	7.1	0.008	0.149	0.119933639	up
	1562.11044@7.1	1562.11044	7.1	<0.001	<0.001	0.601721278	up

RT: retention time, minutes.

^aPrevious TIA vs. non previous TIA.

^bConfirmed by exact mass, retention time, and MS/MS spectrum.

^cConfirmed by exact mass and retention time.

To discern whether the identified molecules could have a neuroprotective function. we further evaluated the relationship between the differential molecules identified in previous TIA ≤7d group vs. non-previous TIA ≤7d group comparison with AIS severity indicators using Spearman correlations (Figure 3). One of the identified molecules (DG(34:2)) showed a significant inverse correlation (FDR p-value <0.05) with DWI infarct volume. Twenty of them presented a significant inverse correlation with NIHSS at admission (FDR p-value <0.05): DG(18:1/16:0). DG(34:2). Hydroxydodecanoic acid. Indole-3-carboxaldehyde. Indoleacrylic acid. Isoleucine. LysoPE(18:4). LysoPE(20:4). OxTG(18:1/18:1/18:2);1O. PC(38:6). PC(40:6) . PC(42:5). PC(42:6). PC(O-36:4)/PC(P-36:3). PC(O-42:8). PE(P-38:6). PI(38:4). SM(d36:2). TG(21:0/18:1/18:2). Tryptophol. and 5 of them maintained this correlation with NIHSS 7 days after admission (DG(34:2). Indole-3-carboxaldehyde. PC(42:5). PE(P-38:6). PI(38:4)). 13 of them presented a significant inverse correlation with mRS 90 days after admission (FDR p-value <0.05): DG(34:2). Indole-3-carboxaldehyde. Isoleucine. LysoPE(18:4). LysoPE(20:4). OxTG(18:1/18:1/18:2);1O. PC(42:5). PC(O-42:8). PE(P-38:6). PI(38:4). SM(d36:2). TG(21:0/18:1/18:2). Tryptophol. In all the reported cases. those molecules upregulated in the previous TIA group showed an inverse correlation with NIHSS. mRS and DWI infarct volume except for the correlation between hydroxydodecanoic acid and NIHSS at admission.

Figure 3.

Euler diagram of differentially expressed metabolites in TIA ≤7d patients (adjusting by etiology) correlating with IS severity indicators (FDR p values for Spearman’s correlation <0.05). ^aconfirmed by exact mass. retention time. and MS/MS spectrum; ^bconfirmed by exact mass and retention time. Differentially expressed metabolites in TIA ≤7d patients when adjusted by etiology and age are highlighted in bold.

To know whether the relationship between lipid species was also altered in previous TIA ≤7d group. a chord diagram including identified differential molecules between groups was performed (Figure 2(c) and (d)). Correlations between the differential lipids showed a different distribution between the previous TIA ≤7d group and non-previous TIA ≤7d group. For instance. in the previous TIA group. all correlations of OxTG(55:1);2O with other differential metabolites disappear in the previous TIA ≤7d group. Similarly. OxTG(54:4);1O correlates in both groups with many different lipids. though a correlation with SM(d36:2) is lost in the previous TIA ≤7d group. Regarding also SL. while in non-previous TIA group. SM(d35:1) correlates significantly with several aminoacids and related compounds (isoleucine. indoleacrylic acid. indole-3-carboxaldehyde). this relationship is lost in the previous TIA group. Of note. isoleucine relationships with the metabolites of degradation (indoleacrylic acid. and indole-3-carboxaldehyde) seem mainly restricted to the previous TIA group. Reinforcing that these differences could be related to switches in lipid metabolism. for instance. ganglioside GM3 levels are correlated with highly unsaturated lysophosphatidylethanolamines (LysoPE(22:4) and LysoPE(20:4)) in the previous TIA group. but not in the non-previous TIA group. Similarly. these lipids correlate also with an ether containing lipid (PC(O-42:8)). as well as a highly unsaturated phosphatidylcholine (PC(40:6)) in the previous TIA group. but this relationship is not present in the non-previous TIA group. These data supported the idea that previous TIA before AIS induces a specific regulation of lipid metabolism resulting in a specific metabolomics/lipidomic signature.

Discussion

AIS patients with a previous recent TIA had a better clinical and neuroimaging outcome in comparison with AIS patients without prior recent TIA. This clinical translation of the phenomenon of IT induced by an episode of TIA before AIS has been described previously.^4–7 Interestingly. our study provides for the first-time evidence of a metabolomic-lipidomic signature. This signature suggests the induction of a specific regulation of lipid metabolism by recent TIA.

The phenomenon of IPC is well documented in animal models.¹ although IT's physiological mechanisms are still unknown. It depends on complex signaling pathways involving ligands of G-protein-linked receptors. membrane growth factor receptors. mitochondrial ATP-sensitive potassium channels. signaling kinases such as PKC and PKG -elicited by changes in phospholipids-. modulation of reactive oxygen species. TNF-α. and sphingosine-1-phosphate production.²⁹

In our study. IPC impacts the plasma metabolome and lipidome. increasing specific molecules. mainly involved in lipid metabolism. Furthermore. the relationship between these lipid species is also altered by a brief ischemic period. reinforcing the idea of a specific regulation of lipid metabolism associated with preconditioning.

Globally. we described a plasma lipid profile of individuals who had a previous recent TIA characterized by a significant increase in selected GPL. GL and SL species. mainly involved in structural. cell signaling and bioenergetics roles. GPLs are the main components of neural (neurons and glia) cell membranes.³⁰ Specifically. PA is the precursor for “de novo” generation of different types of GPL such as PC. PE. PS. glycerophophatidylinositols (PI). PG. and cardiolipin (CL).³¹ PC are the main components of mammalian cell membranes. but other lipid species such as PE. PS. SM. CL and PI are also important as cell membrane constituents.³² Our findings demonstrated specific changes affecting mainly PC. but also PE. PI and PS. Previous papers have also described changes in brain and plasma phospholipid profile after an ischemic event such as stroke in both humans and animal models^33–36 In addition. we also found three lysophospholipids (specifically lysoPE) statistically different between groups. Lysophospholipids are made from phospholipids via “remodeling pathways” by specific phospholipases generating a wide variety of lysophopholipids which are involved not only in membrane structural properties but also in cell signaling.³¹ In line with our results. LysoPE(20:4) has been previously related to ischemia resistance in a rodent model.³⁷

Among the 14 GPL. 4 of them are plasmalogens. which have antioxidant properties and have been previously related to response to hypoxia in primary neuronal cultures.³⁸ Additionally. the specific lipid species identified contained fatty acids with a high unsaturation degree. in line with the highly unsaturated fatty acid profile characteristic of neural lipid species.³⁹ GPL containing polyunsaturated fatty acids are mainly the result of the “remodeling pathway” and are known as major sources of fatty acid-derived lipid mediators.³¹ Specifically. these species have important roles determining membrane fluidity. stability and activity. and oxidative stress status. among others. Therefore. our results suggest that previous TIA induced changes in lipid composition via “remodeling pathway” to obtain specific lipid mediators potentially important for preconditioning process.

Notably. PC and PS. which are metabolically interconnected maintain a relationship with mitochondria. Thus. PC and PS synthesis involve the mitochondria-associated-membranes (MAM).³² Therefore. it is proposed that previous TIA induces an adaptive response to promote structural membrane biosynthesis. repair and neural cell survival to facilitate functional recovery after AIS. Additionally. this adaptive response probably involves mitochondria for a dual function: improvement of neural bioenergetics and GPL biosynthesis. to maintain the membrane integrity and potential lipid-derived survival signaling pathways.

Furthermore. among these GPL. we described some other lipid forms that are related to neuronal membrane integrity.³⁰ Reinforcing this concept. we also reported an increment of structural SL (two SMs. and the ganglioside GM3). Considering that SM is basically present in plasma membrane and lipid rafts. and that its biosynthesis is mainly derived from ceramide and PC. it can be suggested that the increase in SM content is concomitant with the increase in PC also observed in the previous recent TIA group. GM3 deserves special attention. Gangliosides are glycosphingolipids featured by the presence of sialic acid residues. They also are essential structural components of the lipid rafts and plasma membrane.⁴⁰ Functionally. gangliosides maintain interactions with intra- and extracellular signaling molecules. participates in physiological mechanisms of both differentiation and proliferation. and they are altered in pathological conditions such as neurodegeneration and oncogenesis.⁴⁰ A-series gangliosides – including GM3. GM2. GM1. and GD1a – are the primary. structurally simplest group of gangliosides and represent the most abundant species in brain. In physiological conditions. the complex gangliosides GM1 and GD1a are more abundant in the brain than the simple species GM2 and GM3 and have neuroprotective properties. GM3 is structurally the simplest member of the ganglioside family and acts as precursor to the other more complex ganglioside species and regulates the level of expression of all other gangliosides in cells. being expressed only in very low concentrations in the healthy adult brain.⁴⁰ The accumulation of GM3 in neural tissue generates toxic effects and apoptosis.^40,41 An increase in GM3 expression has been described in both human and animal models of Alzheimer disease.⁴² as well as in rodent models of stroke and in human stroke with previous recent TIA. in line with observations in experimental models. suggesting the activation of an adaptive response from brain injury in order to offer precursors for the biosynthesis of more complex gangliosides such as GM1 and GD1a which have neuroprotective effects to preserve neuronal function.⁴³

Ischemia phenomena is tightly related to oxidative stress.⁴⁴ In this work we described two oxidized triglycerols [OxTG(18:1/18:1/18:2);1O and OxTG(55:1);2O] and four plasmalogens [PC(O-36:4). PC(O-42:8). PE(P-38:6) and PS(O-29:0)] up-regulated in previous TIA group. These results suggest that TIA phenomena induces an increase in cellular oxidative stress that is compensated generating “antioxidant” molecules. This response could contribute to better response to AIS and the better outcome described.

It is important to highlight that the metabolomics analyses revealed the different regulation of three interesting molecules related to indole metabolism: indole-3-carboxaldehyde. indoleacrylic acid and tryptophol. These three molecules are microbiota products of tryptophan and present important immune functions. Specifically. indole-3-carboxaldehyde stimulates the production of interleukin-22 which facilitates mucosal reactivity.⁴⁵ indoleacrylic acid promotes intestinal epithelial barrier function and mitigates inflammatory responses⁴⁶ and tryptophol production is negatively associated with interferon-gamma production (IFNγ) which suggests that tryptophol has anti-inflammatory properties.⁴⁷ The relation between gut microbiota and TIA/AIS is described elsewhere.^48–50 In line with our results. an alteration in the composition of gut microbiota after acute central nervous system injury has been previously described.⁴⁸ This alteration damages the equilibrium of the bidirectional gut-brain axis and leads to poor prognosis by triggering pro-inflammatory responses in both peripheral circulation and central nervous system. Thus. the present results suggest a better response of intestinal immune system in previous TIA patients that should contribute to mitigate the inflammatory process associated with AIS.

Previous studies pointed aminoacids in general and isoleucine in particular as potential biomarkers of ischemic stroke.^51,52 However. there is no information about its role in ischemia preconditioning. Our study demonstrated increased levels of isoleucine in previous TIA group. These results are in line with a previous work where the levels of leucine-isoleucine were related to good functional recovery after suffering stroke.⁵³

Notably. those lipid species that were differentially regulated in the previous TIA group presented a correlation with parameters associated with AIS severity and poor outcome (mRS at 90 days. infarct volume. basal NIHSS and NIHSS at 7 days). indicating that this metabolic adaptation observed in previous TIA patients has a neuroprotective role. Among the differential molecules described in previous TIA patients. DG(34:2) correlates with all the clinical outcomes. and Indole-3-carboxaldehyde. PC(42:5). PE(P-38:6) and PI(38:4) correlate with all the outcomes except from DWI volume. arising as the potential molecules with highest neuroprotective power.

The observation of low levels of inflammatory cytokines and biomarkers of a brain injury like IL-6⁵⁴ and S100b⁵⁵ suggested a neuroprotective environment of preceding TIA IS patients. The phenomena of remote ischemic preconditioning might enhance most of the pathways of focal IPC.

Some limitations of our study should be highlighted. The size of the two cohorts could limit the potential generalization of the obtained results. Further. the lower number of patients with a previous recent TIA contributed to obtaining only 4 molecules with FDR adjusted p-value <0.05. so further studies are needed to confirm our results. We reported significantly fewer biomarkers related to cardioembolism like hs-troponin⁵⁶ and NT-proBNP¹⁷ in previous TIA IS patients than in non-previous TIA. We are aware of the possibility that some information on previous TIA would be lost on severe stroke patients who suffered disturbances of consciousness or aphasia beside our efforts.

Also. we used the time frame of seven days for the preceding TIA before AIS according to the previous evidence.⁷ Again. the low proportion of patients with recent TIA cases in our cohorts prevented us from performing some sub-analyses. Moreover. MRI was not performed. and lesion volume was not analyzed in up to 26.8% of patients. This lack of patients could be a limitation. as those patients who underwent MRI were younger. were less cardioembolic etiology. had a less severe stroke. and had better outcomes than patients who did not undergo MRI. In no cases. MRI was performed within 24 hours of the onset of symptoms. So. infarction volumes were not representative of the early acute phase of the stroke. Outcomes of patients with recent previous ischemic heart disease or peripheral artery disease³ were analyzed. but the number of patients with these antecedents was shallow and statistical analysis was not possible to perform (data not shown).

In conclusion. preceding TIA induces IT which clinical translated as better outcome and smaller ischemic lesion. Previous recent TIA before AIS seems to activate an additional adaptive response based on increasing circulation levels of structural and bioactive lipids to facilitate functional recovery after AIS. Furthermore. changes after TIA event seems to facilitate the production of anti-inflammatory mediators before the definitive AIS that contribute to a better immune response. The IT phenomena contribute to a better adaptation of further ischemia. Our study provides first-time evidence of a metabolomic-lipidomic signature related to the development of stroke tolerance in AIS patients induced by recent TIA and defined by specific regulation of lipid metabolism. Our findings open the way for the future discovery of biomarkers and metabolomics targets for new therapies of neuroprotection.

Footnotes

Funding

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: This study was supported by the Government of Catalonia-Agència de Gestió d'Ajuts Universitaris i de Recerca (FP: 2017 SGR 1628; RP: 2017 SGR 696); Instituto de Salud Carlos III and co-funded by European Union (ERDF/ESF. “Investing in your future” and “A way to build Europe”) (FP: PI17-01725. PI14/01574; M.P-O: PI 17–00134. 20-0155); the Spanish Ministry of Science. Innovation and Universities (RP: RTI2018-099200-B-I00); the Department of Health (RP: SLT002/16/00250); and the INVICTUS plus Research Network (Carlos III Health Institute) (FP. CT and GA: RD16-0019-0017; TS: RD16-0019-0012; AO and JR: RD16-0019-0012; AC: RD16-0019-0006. M.J. is a professor under the Serra Hunter program (Generalitat de Catalunya).

Acknowledgements

We are grateful to all recruited patients and the members of Clinical Neuroscience group at IRBLleida. personnel of Neurology Department at Hospital Universitari Arnau de Vilanova de Lleida for scientific discussions and instrumental help. Rosa Gómez and Ikram Benabdelhak for their technical assistance. We also acknowledge the use of the facility of Biobank (B.0000682) at IRBLleida and Plataforma Biobancos PT17/0015/0027.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Authors’ contributions

GA MJ FP conceived the study. JS RP MPO MJ FP designed experiments. GM. JMS. NA. AO. JR. ST. MGC. JMF. ACC. TS. FP cohorts' recruitment. clinical data and blood samples acquisition. JS LCC CT DGL RP MPO MJ FP sample processing and data analysis. JS RP MPO GA MJ FP participated on data interpretation and draft the manuscript. All authors critically revised the final version of the manuscript. All authors approved the final version to be published. FP procured funding.

ORCID iDs

Ángel Ois

Antonio Cruz-Culebras

Tomás Segura

Francisco Purroy

References

Iadecola

Anrather

Stroke research at a crossroad: asking the brain for directions. Nat Neurosci 2011; 14: 1363–1368.

Purroy

Garcia

Mauri

, et al. Induced neuroprotection by remote ischemic perconditioning as a new paradigm in ischemic stroke at the acute phase, a systematic review. BMC Neurol 2020; 20: 266.

Hess

Blauenfeldt

Andersen

, et al. Remote ischaemic conditioning – a new paradigm of self-protection in the brain. Nat Rev Neurol 2015; 11: 698–710.

Aboa-Eboule

Bejot

Osseby

, et al. Influence of prior transient ischaemic attack on stroke prognosis. J Neurol Neurosurg Psychiatry 2011; 82: 993–1000.

Sitzer

Foerch

Neumann-Haefelin

, et al. Transient ischaemic attack preceding anterior circulation infarction is independently associated with favourable outcome. J Neurol Neurosurg Psychiatry 2004; 75: 659–660.

Castillo

Moro

Blanco

, et al. The release of tumor necrosis factor-alpha is associated with ischemic tolerance in human stroke. Ann Neurol 2003; 54: 811–819.

Weber

Diener

Weimar

Why do acute ischemic stroke patients with a preceding transient ischemic attack present with less severe strokes? Insights from the German stroke study. Eur Neurol 2011; 66: 265–270.

Wegener

Gottschalk

Jovanovic

, et al. Transient ischemic attacks before ischemic stroke: preconditioning the human brain? A multicenter magnetic resonance imaging study. Stroke 2004; 35: 616–621.

Johnston

SC.

Ischemic preconditioning from transient ischemic attacks? Data from the Northern California TIA study. Stroke 2004; 35: 2680–2682.

10.

Della Morte

Abete

Gallucci

, et al. Transient ischemic attack before nonlacunar ischemic stroke in the elderly. J Stroke Cerebrovasc Dis 2008; 17: 257–262.

11.

Jove

Portero-Otin

Naudi

, et al. Metabolomics of human brain aging and age-related neurodegenerative diseases. J Neuropathol Exp Neurol 2014; 73: 640–657.

12.

Jove

Mauri-Capdevila

Suarez

, et al. Metabolomics predicts stroke recurrence after transient ischemic attack. Neurology 2015; 84: 36–45.

13.

Purroy

Cambray

Mauri-Capdevila

, et al. Metabolomics predicts neuroimaging characteristics of transient ischemic attack patients. EBioMedicine 2016; 14: 131–138.

14.

von Elm

Altman

Egger

, et al. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet 2007; 370: 1453–1457.

15.

Adams

Bendixen

Kappelle

, et al. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of org 10172 in acute stroke treatment. Stroke 1993; 24: 35–41.

16.

Special report from the national institute of neurological disorders and stroke. Classification of cerebrovascular diseases III. Stroke. 1990; 21: 637–676.

17.

Purroy

Suarez-Luis

Mauri-Capdevila

, et al. N-terminal pro-brain natriuretic peptide level determined at different times identifies transient ischaemic attack patients with atrial fibrillation. Eur J Neurol 2014; 21: 679–683.

18.

Pizarro

Arenzana-Rámila

Pérez-del-Notario

, et al. Plasma lipidomic profiling method based on ultrasound extraction and liquid chromatography mass spectrometry. Anal Chem 2013; 85: 12085–12092.

19.

Want

Masson

Michopoulos

, et al. Global metabolic profiling of animal and human tissues via UPLC-MS. Nat Protoc 2013; 8: 17–32.

20.

Castro-Perez

Kamphorst

DeGroot

, et al. Comprehensive LC-MS E lipidomic analysis using a shotgun approach and its application to biomarker detection and identification in osteoarthritis patients. J Proteome Res 2010; 9: 2377–2389.

21.

Sana

Roark

, et al. Molecular formula and METLIN personal metabolite database matching applied to the identification of compounds generated by LC/TOF-MS. J Biomol Tech 2008; 19: 258–266.

22.

Wishart

Knox

Guo

, et al. HMDB: a knowledgebase for the human metabolome. Nucleic Acids Res 2009; 37: D603–610.

23.

Koelmel

Kroeger

Ulmer

, et al. LipidMatch: an automated workflow for rule-based lipid identification using untargeted high-resolution tandem mass spectrometry data. BMC Bioinformatics 2017; 18: 331.

24.

Xia

Sinelnikov

Han

, et al. MetaboAnalyst 3.0–making metabolomics more meaningful. Nucleic Acids Res 2015; 43: W251–257.

25.

Thomas

Pencina

MJ.

Overlap weighting: a propensity score method that mimics attributes of a randomized clinical trial. JAMA 2020; 323: 2417–2418.

26.

Moncayo

de Freitas

Bogousslavsky

, et al. Do transient ischemic attacks have a neuroprotective effect? Neurology 2000; 54: 2089–2094.

27.

Schaller

Ischemic preconditioning as induction of ischemic tolerance after transient ischemic attacks in human brain: its clinical relevance. Neurosci Lett 2005; 377: 206–211.

28.

Weih

Kallenberg

Bergk

, et al. Attenuated stroke severity after prodromal TIA: a role for ischemic tolerance in the brain? Stroke 1999; 30: 1851–1854.

29.

Cohen

Downey

JM.

Signalling pathways and mechanisms of protection in pre- and postconditioning: historical perspective and lessons for the future. Br J Pharmacol 2015; 172: 1913–1932.

30.

Jove

Pradas

Dominguez-Gonzalez

, et al. Lipids and lipoxidation in human brain aging. mitochondrial ATP-synthase as a key lipoxidation target. Redox Biol 2019; 23: 101082.

31.

Hishikawa

Hashidate

Shimizu

, et al. Diversity and function of membrane glycerophospholipids generated by the remodeling pathway in mammalian cells. J Lipid Res 2014; 55: 799–807.

32.

Vance

JE.

Phospholipid synthesis and transport in mammalian cells. Traffic 2015; 16: 1–18.

33.

Cinquino

Prontera

Pugliese

, et al. Light-emitting textiles: device architectures, working principles, and applications. Micromachines (Basel) 2021 2021; 12: 652.

34.

Maly

Hajsl

Bechynska

, et al. Lipidomic analysis to assess oxidative stress in acute coronary syndrome and acute stroke patients. Metabolites 2021; 11: 412.

35.

Metabolomics and lipidomics of ischemic stroke. Adv Clin Chem 2018; 85: 31–69.

36.

Nielsen

Lambertsen

Clausen

, et al. Mass spectrometry imaging of biomarker lipids for phagocytosis and signalling during focal cerebral ischaemia. Sci Rep 2016; 6: 39571.

37.

Krupska

Kowalczyk

Beręsewicz-Haller

, et al. Hippocampal sector-specific metabolic profiles reflect endogenous strategy for ischemia-reperfusion insult resistance. Mol Neurobiol 2021; 58: 1621–1633.

38.

Lee

Tang

Kuo

CH.

Differentiating ether phosphatidylcholines with a collision energy-optimized MRM method by RPLC-MS/MS and its application to studying ischemia-neuronal injury. Anal Chim Acta 2021; 1184: 339014.

39.

Naudi

Cabre

Dominguez-Gonzalez

, et al. Region-specific vulnerability to lipid peroxidation and evidence of neuronal mechanisms for polyunsaturated fatty acid biosynthesis in the healthy adult human central nervous system. Biochim Biophys Acta Mol Cell Biol Lipids 2017; 1862: 485–495.

40.

Prokazova

Samovilova

Gracheva

, et al. Ganglioside GM3 and its biological functions. Biochemistry (Mosc) 2009; 74: 235–249.

41.

Nakatsuji

Miller

RH.

Selective cell-cycle arrest and induction of apoptosis in proliferating neural cells by ganglioside GM3. Exp Neurol 2001; 168: 290–299.

42.

Chan

Oliveira

Cortes

, et al. Comparative lipidomic analysis of mouse and human brain with Alzheimer disease. J Biol Chem 2012; 287: 2678–2688.

43.

Kreutz

Frozza

Breier

, et al. Amyloid-beta induced toxicity involves ganglioside expression and is sensitive to GM1 neuroprotective action. Neurochem Int 2011; 59: 648–655.

44.

Chamorro

Dirnagl

Urra

, et al. Neuroprotection in acute stroke: targeting excitotoxicity, oxidative and nitrosative stress, and inflammation. Lancet Neurol 2016; 15: 869–881.

45.

Zhang

Davies

SS.

Microbial metabolism of dietary components to bioactive metabolites: opportunities for new therapeutic interventions. Genome Med 2016; 8: 46.

46.

Wlodarska

Luo

Kolde

, et al. Indoleacrylic acid produced by commensal peptostreptococcus species suppresses inflammation. Cell Host Microbe 2017; 22: 25–37 e26.

47.

Schirmer

Smeekens

Vlamakis

, et al. Linking the human gut microbiome to inflammatory cytokine production capacity. Cell 2016; 167: 1125–1136 e1128.

48.

Yuan

Gut microbiota and acute Central nervous system injury: a new target for therapeutic intervention. Front Immunol 2021; 12: 800796.

49.

Zhao

Wang

, et al. Association between plasma trimethyllysine and prognosis of patients with ischemic stroke. J Am Heart Assoc 2021; 10: e020979.

50.

Yin

Liao

, et al. Dysbiosis of gut microbiota with reduced Trimethylamine-N-Oxide level in patients with large-artery atherosclerotic stroke or transient ischemic attack. JAHA 2015; 4.

51.

Pan

Zhang

, et al. Metabolomics facilitates the discovery of metabolic biomarkers and pathways for ischemic stroke: a systematic review. Metabolomics 2019; 15: 152.

52.

Liu

Antonov

, et al. Discovery of metabolite biomarkers for acute ischemic stroke progression. J Proteome Res 2017; 16: 773–779.

53.

Wang

Liu

Song

, et al. Targeted metabolomic profiling reveals association between altered amino acids and poor functional recovery after stroke. Front Neurol 2019; 10: 1425.

54.

Bustamante

Sobrino

Giralt

, et al. Prognostic value of blood interleukin-6 in the prediction of functional outcome after stroke: a systematic review and meta-analysis. J Neuroimmunol 2014; 274: 215–224.

55.

Wang

Yang

, et al. Serum S100B levels may be associated with cerebral infarction: a meta-analysis. J Neurol Sci 2015; 348: 81–88.

56.

Kerr

Ray

, et al. Elevated troponin after stroke: a systematic review. Cerebrovasc Dis 2009; 28: 220–226.

Ischemia preconditioning induces an adaptive response that defines a circulating metabolomic signature in ischemic stroke patients

Abstract

Keywords

Introduction

Materials & methods

Study participants – cohort description

Clinical lipid and inflammatory biomarkers panel

Metabolomic and lipidomic sample processing

Metabolomic and lipidomic data acquisition and analysis

Statistical analysis

Data availability

Results

Discussion

Footnotes

Funding

Acknowledgements

Declaration of conflicting interests

Authors’ contributions

ORCID iDs

References